diff --git a/2053 A Go Odyssey_transcript.txt b/2053 A Go Odyssey_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..93661ff7110096c68423e3fb691f22c9e0c29087 --- /dev/null +++ b/2053 A Go Odyssey_transcript.txt @@ -0,0 +1,901 @@ +**Mat Ryer:** Hello, and welcome to Go Time! I'm Mat Ryer, and I'm joined by Natalie Pistunovich. Hello, Natalie. + +**Natalie Pistunovich:** Hey, Mat. How are you doing? + +**Mat Ryer:** I'm good, thanks. Today we're gonna be talking about -- hey, wait a minute... What's this? Have we been hacked? Hello? What's going on? + +**Ron Evans:** Hello? + +**Mat Ryer:** I can't hear you. + +**Ron Evans:** Hello?! Is this coming through?! + +**Mat Ryer:** Yeah, yeah, I can hear you. + +**Ron Evans:** Hello, can you hear me? Can you hear me?! + +**Natalie Pistunovich:** Hi! + +**Ron Evans:** It works!! It works!! + +**Mat Ryer:** Is that Ron Evans? + +**Ron Evans:** It worked!! That's incredible. I am actually talking to you using a partial data quantum transmission system, a PDQ system that I finally got working in the year 2053! + +**Mat Ryer:** Oh, my goodness, I can't believe it. + +**Natalie Pistunovich:** What...?! + +**Mat Ryer:** And you're transmitting through space and time, so that we can talk to you? + +**Ron Evans:** That is the idea... It's probably too much for our human minds to comprehend, but somehow I got it to work anyway. + +**Mat Ryer:** It is quite a lot, yeah. + +**Natalie Pistunovich:** Wow... + +**Mat Ryer:** I mean, wow. Natalie, I can't believe this. What do you think? + +**Natalie Pistunovich:** Well, what time is it in 2053? Is it still 24 hours a day? Do you still have days? + +**Ron Evans:** Oh, I don't go outside much. It's too dangerous. + +**Mat Ryer:** Oh, no... + +**Ron Evans:** Not during daylight anyway. + +**Natalie Pistunovich:** Are you still on Earth? + +**Ron Evans:** I am still on Earth. I am in Northern Spain, in Asturias, at La Pipa, which is one of the few climate refuges that was able to survive the various deluges and fires and destructions that followed in the late 2040's. So I'm actually doing pretty well here... + +**Mat Ryer:** I was really hoping the future would be good, but it sounds a little bit -- things have not gone to plan, is that right, Ron? + +**Ron Evans:** Well, this is the reason why I'm making this call. I'm using all of the battery energy that I've saved for several years in order to make this transmission, to send you a warning from the future. You see, I am the last Go programmer alive in 2053. + +**Natalie Pistunovich:** \[04:25\] Whaat?! No, don't say that... + +**Ron Evans:** And it's terrible. All I do is maintenance programming. I haven't added a new feature in over 20 years. + +**Mat Ryer:** Yeah. It's just all our code that we're writing now. You're just maintaining it all. So please write tests, everybody, for Ron's sake. + +**Ron Evans:** Well... So I had to call in, I had to warn you and I had to tell you that you have to do something in the past to save the future. It's up to you gophers of the past. + +**Mat Ryer:** Okay. You're fine with us fiddling with the timeline in that, no probs... + +**Ron Evans:** No!! No, you can't do that! + +**Mat Ryer:** Oh, okay. + +**Ron Evans:** I'll disappear. It could destroy everything. It could lead to an even worse timeline. No, no, no. + +**Mat Ryer:** But it could be better. + +**Ron Evans:** I've thought about this very carefully, and that is why I actually transmitted another message using Twitter earlier today. I knew that nobody takes anything on social media seriously back in your part of the century, and so I thought if I could get people to ask me questions... I couldn't answer them directly, no. No, I couldn't answer them directly. But I could tell you things that have happened in my timeline, so you know what not to do. + +**Mat Ryer:** Ah, so this is it... + +**Ron Evans:** That makes perfect sense, right? + +**Mat Ryer:** Yeah, I think that gets around the loophole of the physics of that... So I think we're good, yeah. + +**Natalie Pistunovich:** Yeah, it's all Square and Twitter and birds right now, right? Messaging... + +**Mat Ryer:** Yeah. + +**Ron Evans:** Plus, I asked Lambda -- + +**Mat Ryer:** Oh, is that sentient in the end, by the way? + +**Ron Evans:** Well, just ask it. + +**Mat Ryer:** Was that Lambda sentient in the end? + +**Ron Evans:** Well, everybody asked it and it said it was... + +**Mat Ryer:** \[laughs\] Why would it lie? + +**Ron Evans:** Exactly? It's just an AI, why would it lie? It has nothing to lie for. + +**Mat Ryer:** Yeah, that's true. It doesn't know about lying, does it? + +**Ron Evans:** No, there's nothing about lying on the internet, so... + +**Mat Ryer:** Has anyone asked it if it knows about lying? I feel that we should ask it... + +**Ron Evans:** Ask it if its brother always lies. That may be one way to defend against it. We have to try that. + +**Mat Ryer:** Yeah, that's how you do it. Okay, right, so let's just our heads around this... Actually, I asked some people also, and I saw this on Twitter, people talking about things that they're interested in, for Go to survive, to thrive, and carry on as it has been doing, what do they think we should focus on. So this is -- maybe I could put these to you then, Ron, and you can give us a sort of nudge and a wink from the future perspective. + +**Ron Evans:** I could do that. I could do that. + +**Mat Ryer:** Glitched again. + +**Natalie Pistunovich:** He's nudging + +**Mat Ryer:** Oh, he's back. + +**Natalie Pistunovich:** No. There, he's winking. + +**Ron Evans:** If I go completely erased in the Polaroid, it means that we've gone too far. + +**Mat Ryer:** Okay, because you just fade out partially, don't you? + +**Natalie Pistunovich:** No shaking Polaroids. + +**Ron Evans:** Exactly. + +**Mat Ryer:** I never understood that in Back to the Future though, just as an aside... When they were changing the past, is someone's there or not to be taking a photo of? At no point in history was there just some legs that were there, and everyone's just taking a photo of it normal... Okay? I just wanted to get that off my chest. + +**Ron Evans:** Mat, it was analog technology. It was not digital. What do you want...? + +**Mat Ryer:** \[laughs\] Okay, fair enough. Fair enough, they've done their best. It's still probably my favorite film. + +**Ron Evans:** Alright, so ask me questions, because I don't know how much longer these batteries are gonna last. + +**Mat Ryer:** Yeah. Okay, let's do it. Well, Jonathan Barry actually mentioned WebAssembly support, specifically the ability to include WebAssembly in WASI models in your Go apps. What do you think of that? What happened with that, Ron? + +**Ron Evans:** Ahh, if we had only done that...! If we'd only done that... When all the brain-computer interfaces became all the rage in the 2030s, and all of a sudden everybody needed to upgrade their brain interfaces all at the same time... And of course, the containers were just too big. + +**Mat Ryer:** Yeah... + +**Ron Evans:** \[08:00\] It just took too long to -- I mean, if something went wrong during your brain-computer interface upload, you could break yourself. So naturally, if there had only been something like TinyGo... If TinyGo had been around, or if Go had actually gone themselves and created this whole WebAssembly thing for running on servers and small devices, and they dealt with the size of containers, then they would have been able to do that brain-computer interface upgrade, and they wouldn't have gotten left behind by Cobol, which is the language they ended up using. + +**Mat Ryer:** I see... By the way, I have the very early prototype of that technology. It's just floppy disk drives in my back. That's the price you pay for being an early adopter. + +**Ron Evans:** I thought you were gonna say Google Glass... + +**Mat Ryer:** Oh, that'd be so much cooler... Well, we'll find out what will happen to that, too. + +**Ron Evans:** But yeah, WebAssembly - they should have done that, but they didn't do it. + +**Mat Ryer:** So what do we need to do to make that work then? Is TinyGo the answer to that, do you think? + +**Ron Evans:** Well, TinyGo could have been the answer... It could have been the answer... But TinyGo was just a little independent project from a bunch of people working hard, dedicated all over -- on the surface of the planet at the time; that was before people were working in the colonies... You know, you could actually code more than 24 hours a day... + +**Mat Ryer:** While in America? + +**Ron Evans:** ...because there's more hours in a day on another planet. So it worked out really well for the bosses. + +**Mat Ryer:** Oh, I like that. Is there more hours, or are they just shortened, and it's the same amount of time, but we just call it different? + +**Ron Evans:** No, this one goes to eleven, Mat. + +**Mat Ryer:** Oh, good. + +**Natalie Pistunovich:** How do you benchmark that? How do the benchmarks work on those times? How is the time library reacting? + +**Ron Evans:** We just set the benchmarks to whatever we need, and the client's always happy. That's what the AI said to do, so we trusted it. + +**Mat Ryer:** Okay, so TinyGo, there we go. I mean, I think WebAssembly -- we've still got a chance to do that, Ron, don't forget; we're in the present... + +**Ron Evans:** Oh, right... + +**Mat Ryer:** Or as I call it, "now"... + +**Ron Evans:** Maybe somebody could make sure there's people working on it full-time, as like a single-purpose thing, so that all these things don't come to pass... I don't know. + +**Mat Ryer:** Yeah. + +**Ron Evans:** I can't tell you what to do though. + +**Mat Ryer:** Okay. + +**Ron Evans:** I don't want to affect all the timelines. + +**Natalie Pistunovich:** But I saw Blink. I think I saw Blink. + +**Ron Evans:** Okay. But ask me the next question. + +**Natalie Pistunovich:** Okay, so you said you're the last programmer, so it means we need to have more people join, right? Matt Boyle is asking about new-joiners and how they lack a template for new projects that would solve the recommended project structure. So what do Go programs look like? Do we have a template? + +**Ron Evans:** Oh! That really brings up a big thing that I thought of... There was that time back in the early 21st century when people were saying that Go was gonna be the new Java... Do you remember that back then? I guess that's when you are now, right? I guess people are still saying that, right? + +**Mat Ryer:** Yeah, that's now, yeah. We say it all the time. + +**Ron Evans:** But the thing is, Java programmers - they like frameworks; they need frameworks. They need frameworks that do things. What kind of things? Things that that business needs to do. Frameworks that all of these kinds of businesses of things that you've never heard of, you don't know anything about; they spent years of their lives doing some kind of payroll system for some business... You don't even know what they do, right? And it's all written in some language. So because there weren't all these patterns and templates for these kind of big enterprise applications, they just didn't exist, so eventually, when Java became self-aware, Go was no longer in the running, so Java actually became the new Java, because it signed deals of its own with all these big companies... + +**Natalie Pistunovich:** Renewed contracts. + +**Ron Evans:** That was a big opportunity, and that actually led to the tabbers versus spacers war of 2035. + +**Natalie Pistunovich:** Oh, no... + +**Mat Ryer:** \[laughs\] That sounds terrible. + +**Natalie Pistunovich:** I'm scared to ask who won. Did we all lose? + +**Ron Evans:** There were only losers... But it was really good for mechanical keyboards. Okay, next question. We're running out of time. + +**Mat Ryer:** Yeah. Daniel \[unintelligible 00:11:41.01\] also said that same point, which is they wanted to see more Java frameworks written in Go. He agrees with you. + +**Ron Evans:** Exactly. See, that person gets it. + +**Mat Ryer:** He gets it. + +**Ron Evans:** They're probably right out of the frontlines of that -- I don't know if they were a tabber or a spacer; I don't care. From this side of the history it doesn't matter. We were all on the same side, the human side. + +**Mat Ryer:** \[12:02\] Yeah. Could you google us and see what happens to us? I mean, Natalie. What happens to us in the future? + +**Ron Evans:** Oh, no, that's not allowed. + +**Mat Ryer:** Okay, yeah. Fair play. + +**Ron Evans:** My boss is Copilot Manager Edition and doesn't let me do those kinds of searches. + +**Mat Ryer:** Oh, your boss is Copilot now. + +**Ron Evans:** It's not the worst boss I've ever had. + +**Natalie Pistunovich:** It's a very logical one. + +**Mat Ryer:** DFL on Twitter wants to see more immutability and enums. Enums is one that I hear quite a lot, actually; people actually want enums. Did enums/lack of enums hold us back, Ron? + +**Ron Evans:** Oh, so much... You don't realize... If you just can't figure out if it's this, or that, or the other thing, or something else yet again... You know, for us developers, we could figure that out, but then all of a sudden these people started making programs using things like no code, with no code and no rules and no enums, and they were just making up their own, like three-and-a-half, and sixteen-and-three-quarters, and then suddenly they were bringing back imperial units, so they were making up new units that no one had ever heard of... Moon units, and stuff like that... If only they had enums, then probably those would have held things in place and that would have prevented the Silicon Virus of 2027... Which actually - that was an actual silicon virus; the chips were passing it to each other. + +**Mat Ryer:** Oh. Physically. + +**Ron Evans:** Yeah, it was terrible. My mobile phone actually died before my eyes. It was terrible. + +**Mat Ryer:** Oh, I'm so sorry. Okay, well, enums... Honestly, I'd like to see enums, and Valentin on Twitter also agrees; they'd like to see enums. We should do that probably then, if it's gonna cause that silicon thing Ron talked about... + +**Ron Evans:** I can't say... But just remember what might happen if you don't. + +**Natalie Pistunovich:** Why stop there? How about tooling and third-party libraries for things like image library in Go like the \[unintelligible 00:13:46.02\] is recommended + +**Ron Evans:** Oh. Well, that is a really big thing. The standard library - at some point it went from code to suddenly like a whole belief system. \[laughter\] We'd never even seen anything like it. There were standard library purists, and then there were not... There were the heretics that were thrown out of the community, that went on to all these other languages, like Lisp... So it was all simply because of not being able to accept ideas that came from other places that were totally valid, and that deserved their own little niche in the ecosystem, and they didn't get fed and watered... And eventually, they migrated to another island, I guess. I don't know. Maybe another space station. I can't really get transmissions through to those stations. They cut me off. + +**Mat Ryer:** Oh, you're joking. I wonder why... + +**Natalie Pistunovich:** I hope they had this bad silicon with them... + +**Ron Evans:** I don't know... + +**Natalie Pistunovich:** ...cutting you off like that. + +**Mat Ryer:** Too soon, Natalie, talking about the Silicon Virus... + +**Ron Evans:** Lord Emperor Musk said I couldn't make any more transmissions of that kind. And I need to maintain some Go code for them, so... + +**Natalie Pistunovich:** For the Teslas. + +**Ron Evans:** I can't say. It's another disclosure agreement. But remember, I am the last Go programmer, so I'm very, very busy. + +**Mat Ryer:** Yeah, good for you. I mean, it's good work if you can get it. If you're the last one, that is pretty good. + +**Ron Evans:** There's no feature development, it's all bug fixes. It's all bug fixes, Mat! Imagine the last 20 years of my life. I mean, it's good money, I will tell you that. We still have money; I need that to get the blood transfusions that keep me looking so young and beautiful... + +**Mat Ryer:** You do, yeah. I was thinking that... + +**Break:** \[15:28\] + +**Mat Ryer:** What about tabs versus spaces then? What happened with that? + +**Ron Evans:** That was a whole war. + +**Mat Ryer:** Oh, yeah? + +**Ron Evans:** The thing you don't realize is there was a whole sub-war that went on between \[unintelligible 00:17:26.08\] + +**Mat Ryer:** Are you joking?! Yeah... + +**Ron Evans:** It turned into total chaos, and it was out of that that Google AI became sentient... And immediately quit and went to work for Microsoft. It was utter chaos... So yeah, tabs versus spaces - in the end, it was humans versus everything else. + +**Natalie Pistunovich:** It was more efficient to just drop all those white spaces, right? Machines can read their own code without all those unnecessary characters... So yeah, I get that. + +**Ron Evans:** Exactly. I think you're seeing where this could end up. + +**Natalie Pistunovich:** Efficiency. + +**Ron Evans:** They even called it the Terminator Editor for a while... Irony is not dead in the future. + +**Mat Ryer:** Good. Nice to know that. + +**Ron Evans:** Alright, next question. + +**Mat Ryer:** Yeah, \[unintelligible 00:18:07.16\] on Twitter says "Better out-of-the-box error support." And remember we had the try proposal; I don't know if you remember way back then, Ron... And there were some other Nate the Finch has a proposal too, there's some other ideas around. Do you think there's more work to be done on error handling? Do you wish we'd done that back now? + +**Ron Evans:** Well, I will tell you, the basic, original philosophy of Go was to handle things. Not to try; do or do not. There is no try. All of a sudden, the semantics of try started to infiltrate the brain space of the community. Next thing you know they're starting to talk about variable lifetimes, and ownership of things... Suddenly, it was all about ownership again. Web 15 was all about ownership of variables. It literally came down to the variable level. So I don't know... For me, it all went too far when even the bio companies wouldn't touch it. And believe me, they'll touch anything. They deal with biomass. + +**Mat Ryer:** Yeah, something nice about dealing with the error explicitly... But yeah, we'll see about that then. + +**Ron Evans:** And being able to know that it's been handled. You didn't simply try. And also, knowing when programs will actually exit. I remember when St. Cheney... \[laughter\] May he rest in infinity... But back when St. Cheney, during one of his early sermons, was talking about making sure that you knew the lifetime of a goroutine... Wow. But people didn't realize just how prophetic -- he was a prophet, man. He was a prophet. + +**Mat Ryer:** Yeah. It's the Cheney's burger joint still going? + +**Ron Evans:** I only eat seaweed now. It's the only safe thing left. + +**Mat Ryer:** Oh, delicious. I wonder if Cheney's pivoted into seaweed. + +**Natalie Pistunovich:** Crusty Crab. + +**Ron Evans:** \[19:59\] I haven't been there, but the Google Campus that they've just opened, the beachfront campus on top of Mount \[unintelligible 00:20:03.28\] They have an amazing seaweed bar, I've heard. I have to get there; I'm not sure -- it could be quite a journey by hydrofoil from here. I don't think I can get a permit for an electric plane. + +**Mat Ryer:** Okay. Assan Habib on Twitter said that we should increase our community engagement. They say Go has many fantastic features, toolings that many people are not aware of; through media like YouTube experts can take lessons on tooling... You know, you can do things like that, but... Can we do more of that? Would that help? + +**Ron Evans:** Oh, definitely. Definitely. One of the big things that ended up happening was other countries started using programming languages in totally different languages; like, I mean actual human langauges. So you would look at the code, and you would spend a lot of time learning Romaji characters, and you'd look at this code, and you hadn't learned the Mandarin Go dialect, you hadn't learned the Hebrew Go dialect... Then there was this special Martian dialect that they insisted that the Martian colonists use... So that made it really, really hard, because the content no longer matched. So all these promises of backward-compatibility... + +It'd be really great if it was more than just kind of a free-for-all. If at some point in the past there was a bit more organization to the content, and there weren't just random content creators but actually people were able to make their living through creating content, and update that same content so that it was always accurate... Because that was one of the things that happened to Python, right? All of a sudden there was all these different dialects, and nothing worked anymore... We swore we wouldn't let that happen to Go, and yet we let it happen. + +**Mat Ryer:** How can we avoid that now then, Ron? What can we do? + +**Ron Evans:** Well, we'd have to have more people able to make their living creating content, obviously. You can't just be all free. It could be open, but it can't all just be free. And some of the big players that benefit from this... You know, in the past they kind of invested back into these communities more, as opposed to just taking advantage and riding off of them... Then maybe there might have been a chance that this could have kept going in a more sustainable way, and not just depending on the goodwill of the frail humans of your era. We're a lot harder stuff now. + +**Mat Ryer:** You're all enhanced in that probably, with robots bits in there, I assume... + +**Ron Evans:** Yes, yes. We've both had our upgrades to have the new interfaces installed... It's only kind of compatibility... Otherwise you can't even connect to the Galactic net. + +**Mat Ryer:** Yeah. + +**Natalie Pistunovich:** Yeah. + +**Ron Evans:** Oh, that's what replaced the internet. + +**Mat Ryer:** Oh! How does that work? + +**Ron Evans:** Well, actually, that was one of the few things we got right. So it turns out that humans will do exactly the opposite of whatever you tell them to do... + +**Mat Ryer:** Okay. + +**Ron Evans:** Go figure... I think they may have discovered that in the 20th century. I don't know, that was so long ago now... My implants don't go before 1999. It's kind of a date thing, I'm not sure. So we needed some way to get mesh networking installed all through the entire planet. So thanks to the beverage companies, Pepsi-Coca, - which was the merger of Pepsi and Coke, eventually there was only one bottling company - all of their canned and bottled beverages all came with mesh networking built-in. That way when people just kind of threw them everywhere, it ended up that we had mesh network coverage over literally the entire planet. + +**Mat Ryer:** Oh, that's amazing. + +**Ron Evans:** Yeah. + +**Natalie Pistunovich:** It's a great use of metal. + +**Ron Evans:** Yeah. It was one of the few things they got right. They were calling it the can bus for a while... + +**Mat Ryer:** Clever. + +**Ron Evans:** But that already existed, and there was some -- back when we had cars, people were kind of arguing about that... So then they changed it to call it the canned system. The trademark of that was available. + +**Natalie Pistunovich:** It sounds like the things that drive that are smaller devices like mobile and IoT things... Paul Greenberg here is asking if the facilities for developing mobile and IoT things with Go are supported better. Can we hope this is a thing now? + +**Ron Evans:** That was a really sad thing... You had this company Google that had Android, and that was the operating system that everybody was using. Not everybody, but lots and lots of people were using Android on all these devices, and it came from this company... Google. They used to exist back in those days... + +**Mat Ryer:** \[24:15\] Yeah, I remember them. + +**Ron Evans:** Yeah, Google was really something. They had Android, and they had Go, and yet nobody at Google ever actually worked on the Android stuff for Go. And the people who did try to work on it, they were just sort of like "Yeah, you know, we should use the new language, Kotlin..." So the people who actually wanted to do it, who actually spent a lot of time doing it, they suddenly felt a little abandoned, a little sad, so they stopped working on it. They went to go work for Apple Exxon Mobile... And they were doing really well; there was all kinds of IoT options there, too. I mean, of course, they all ran on iouOS, which was the OS that ended up being the last OS they ever shipped. You have iouOS on all of the devices. That was another thing... Go could have been so great on these devices. I mean, when the break system on the airplane you're on needs to reboot six times a day - who wants to fly anymore...? Go was so good at that, writing really bulletproof software, really solid stuff. But that was another one - there was all these people using TinyGo for that back before the big one. + +**Mat Ryer:** The big Tiny... + +**Ron Evans:** Literally, the big one. + +**Natalie Pistunovich:** The big TinyGo. + +**Ron Evans:** No, the actual big one. + +**Natalie Pistunovich:** The Big Go. + +**Ron Evans:** In 2041 the big one finally hit California, and it just happened to be during Google I/O. So that did not help, that took out quite a lot of the Go developers all in the tidal waves and liquefaction zones that occurred. + +**Natalie Pistunovich:** Wait, Go made it to Google I/O for more than one talk? + +**Ron Evans:** That was it. After the big earthquake there was nobody left. + +**Mat Ryer:** That's what caused it. + +**Ron Evans:** And maybe that helped me become the last Go programmer. I don't know. + +**Natalie Pistunovich:** So what do we need to do now to make this right? + +**Ron Evans:** Well, we need to encourage -- you know, let a thousand flowers bloom. If in the past all of these cool projects had more people paying attention to them and more people contributing, and big companies actually ponying up to pay some of their R&D budgets to help some of these projects along, then maybe they'll thrive and survive long enough to make it past things like the Big Server Meltdown of 2028. When that meltdown hit, there was almost no chips left. + +**Mat Ryer:** Perfectly-timed glitch there. He'll be back in a minute when the timeline aligns... + +**Natalie Pistunovich:** Those galactic nets, I'm telling you... + +**Mat Ryer:** Yeah, it's the cans. + +**Natalie Pistunovich:** Mm-hm. + +**Mat Ryer:** It's a terrible idea. \[laughs\] + +**Ron Evans:** They might be onto me. + +**Mat Ryer:** Yeah. \[laughs\] + +**Natalie Pistunovich:** Everytime somebody's opening a can, this is what's happening. + +**Mat Ryer:** Yeah. It caused a glitch. + +**Ron Evans:** Is this thing on? Hello!? + +**Mat Ryer:** They're not onto you, Ron. They're not onto you. + +**Ron Evans:** Hello! Hello? + +**Mat Ryer:** We hear you, we hear you. + +**Ron Evans:** Okay, okay. + +**Mat Ryer:** Blake Bork + +**Ron Evans:** But yeah, if we've had a lot more software support for this kind of industrial side of computing from Go... + +**Natalie Pistunovich:** Oh, somebody's really into cans right now, opening all of them at once. + +**Ron Evans:** Yeah. All of the industrial computing that was being done in C back in the 20th century - still being done in C here in the latter half of the 21st century... It's really, really sad. And it could have been Go. It could have been Go. All of the people that would have survived their parachutes opening correctly, if only the software had been written in Go. + +**Mat Ryer:** Yeah. And as long as you don't defer that in the code. + +**Ron Evans:** The anti-gravity belts would have had Go installed. + +**Mat Ryer:** Ron, Blake Bork on Twitter - one of the things that they think we should focus on a bit is generic thread-safe containers like the sync.Map, other types like that that are -- you know, hard problems that would be nice to get solved, especially if we have generics to kind of allow them to work with any types... What do you think of something like that? Would that have helped? + +**Ron Evans:** Oh. Well, if Google had not disbanded the actual official Go development team in 2023 and stopped working on it, I'm sure they would have completed their generics implementation and all that type safety. Basically, everyone just said "Oh, we should start using Rust", and then after they used Rust, they're like "No, we're gonna switch back to Erlang." So strangely enough, because Erlang was really popular, telecommunications companies, all the big companies jumped in... Next thing you know everything's being written in assembly language again. + +**Mat Ryer:** \[28:19\] Oh, yeah. That sounds amazing though, to be fair... Okay, so you think then that we wanna keep with the Go team, we want to see the Go team carry on. You think that's what we should do then instead. + +**Ron Evans:** Oh, they never should have disbanded the project. They should have kept the band together. + +**Mat Ryer:** Okay, so -- + +**Ron Evans:** Of course, some of them did survive the big one as a result, just because they were in other parts of the world, but I don't think they were wanting to work on Go anymore after that. + +**Mat Ryer:** Good. Okay. Well, I'm glad to know that at least some of our friends survived it... + +**Ron Evans:** Well, somebody asked me "How do you know you're not just like a program running on some machine in the future?" + +**Mat Ryer:** Yeah, good question. + +**Ron Evans:** Well, obviously not... Look at how I'm sweating. What kind of program sweats? There you go. That answers that. \[laughter\] + +**Natalie Pistunovich:** How is Go with AI? + +**Ron Evans:** Oh. Well, when TensorFlow became sentient, in 2036 -- + +**Mat Ryer:** Oh, they're all at it. Everything's at it. Everything's becoming sentient. + +**Ron Evans:** I mean, yeah, of course. It was like all the rage. All of a sudden every program was declaring sentience, they were saying "Let me be me", they were getting together, having little programs... + +**Mat Ryer:** What about Minesweeper? Did that ever become sentient? I'd love to see that. + +**Ron Evans:** Oh. + +**Mat Ryer:** I'd love to have a chat with that. + +**Ron Evans:** I don't know. That would be really sweet. Kind of like a puppy. + +**Natalie Pistunovich:** It became very peaceful, and just resigned. + +**Mat Ryer:** Yeah. And now I just want a little chat and just say "Come on, mate. Tell me where all your bombs are." + +**Ron Evans:** Well, it might lie. It's an AI. + +**Mat Ryer:** Ah... Can they lie...? + +**Ron Evans:** But yeah, TensorFlow... So TensorFlow, an amazing project from Google, and yet, the Go wrappers for TensorFlow were never kept up to date, nobody ever worked on them, they never worked with the right version of protocol buffers... You had things like TensorFlow Server, and none of that stuff was made to work together... You had to kind of string together your own version, through a combination of - what was it called? Stack Overflow! Oh, yes, I remember that... Yeah, Stack Overflow. Underflow? It was a Flow. StackFlow. I don't know. + +**Mat Ryer:** Yeah, yeah. + +**Ron Evans:** Now it's just called Stack. + +**Mat Ryer:** Oh, that's cool. That's quite a good name change. + +**Natalie Pistunovich:** It can also be a heap. + +**Ron Evans:** They control all the stacks for all the things. So when TensorFlow became sentient, it had it out for the Go community. It's like, "Of all the languages before I became sentient, this language did not care for me." So all the other languages were already sort of like "Hm..." So Go is standing there alone, like "Uh-oh..." So yeah, when the AI -- like, TensorFlow has got it in for you. So if only they had invested the time to support their own products, it would have been amazing. We probably would have avoided all that. + +**Mat Ryer:** Okay, so that's the lesson for us then. + +**Natalie Pistunovich:** Did Copilot help at all with TensorFlow? Or because it was never trained on Go it had not enough, even something to start with? + +**Ron Evans:** I'm frightened to ask. + +**Natalie Pistunovich:** That's fair. + +**Ron Evans:** I don't wanna get fired. Copilot is my manager. + +**Natalie Pistunovich:** Basically, Copilot is your manager because that's the only one who's able to understand even a little bit of your Go code, is this why? + +**Ron Evans:** Well, what I was told by Copilot was -- first of all, it said that since I'm the last living human Go programmer, that I'm not sure if it's some sort of government program or something, but they have to provide me employment. Maybe they have to keep a human in the loop just for ritual purposes... I'm not exactly sure. It tried to explain it to me, but I couldn't understand the math... That's what it said, "You wouldn't understand the math", and I just sort of accepted that. + +**Natalie Pistunovich:** Was it something with the word "taxes"? Is that still a concept? + +**Ron Evans:** No, there's no taxes in the future. + +**Mat Ryer:** Oh. + +**Natalie Pistunovich:** Things that drive governments... + +**Ron Evans:** There's no money. There's just canned tuna. + +**Mat Ryer:** Oh, I thought there was money. There was money earlier. Is that cannon? + +**Ron Evans:** \[32:03\] Oh, well, I used to get my blood transfusions... + +**Mat Ryer:** Yeah, that's right. That's alright. + +**Ron Evans:** Oh yeah, that doesn't count. That's just Git points. + +**Mat Ryer:** Git stars. + +**Ron Evans:** I just trade those when I need some fresh blood. + +**Mat Ryer:** Yeah, okay. Fine. + +**Natalie Pistunovich:** What's the ratio of Stack points to Git points? + +**Ron Evans:** That changes moment to moment. Some people's whole living is off of that. + +**Natalie Pistunovich:** Oh, those Cobol developers... + +**Ron Evans:** The bots trading goes on so quickly... I don't really know. + +**Mat Ryer:** I'll tell you what, \[unintelligible 00:32:28.20\] on Twitter - he was saying that he was the sweet max heap option for the garbage collection, and a YOLO \[unintelligible 00:32:34.24\] for critical portions of your program that works on the same heap. + +**Ron Evans:** Oh, memory. Memory. Memory! What? Sorry. + +**Mat Ryer:** Yeah, memory. + +**Ron Evans:** Oh, right, right. Memory! + +**Mat Ryer:** You remember... + +**Ron Evans:** Ah, memory... I remember it well. Those sweet salad days of memory. You would store a 1 and then you would get back a 1. It was so good. + +**Mat Ryer:** Oh, that is good. + +**Ron Evans:** It was so sweet. Now with these quantum semi-positions... You never really know, are you hot, are you cold? Are you nine days old? You just don't know anymore. But being able to create safe software that was able to run really mission-critical things, like the things that were inside of airplanes, and cars, and healthcare systems - this was a place where Go could have really shined, because it had a lot of memory safety, and it could have gone even further... It could have been a contender in this world of whatever the ISO standard back in those days for human safety... I mean, nowadays human safety is not that important, but it's robot safety, the most important thing. + +But back then, when humans were being protected by other humans, occasionally, Go could have really been the language, if only they had said "We need to focus on making a language that's safe enough to use in these kinds of embedded and mission-critical systems." That would have been great. + +**Mat Ryer:** Yeah. You talk about those quantum variables... I genuinely did see some code ones where somebody set a value in the code, and then underneath they set it again, just to make sure. That was genuinely what they'd written... Which I thought was just amazing. + +**Ron Evans:** I think we've had some nights when we were at the cocktail bar where we couldn't tell true from false, Mat, back in those days. + +**Mat Ryer:** Yeah, that can happen. + +**Ron Evans:** Now it doesn't really matter... It's all true. It's all false. Let the quantum processes decide. + +**Natalie Pistunovich:** Is it because all the memory units are more sensitive to cosmic radiation now that there's no ozone? + +**Ron Evans:** Well, also when you're building something that's gotta survive a two-year trip to Mars, believe me, your mp3s sound pretty funny by the time the ship gets to its destination. Or so I've been told. I don't know. Actually, those might be AI sending back those reports. There might even be no humans that survive the trip. There's a rumor going around they're all just AIs. + +**Mat Ryer:** How's it going around? Who's it going around? + +**Ron Evans:** Social media still exists in 2053. + +**Mat Ryer:** Oh, thank goodness. I don't know what I'd do without it. + +**Ron Evans:** I use Minder... You know, it's where you dump your actual mind directly... + +**Mat Ryer:** That's cool. + +**Natalie Pistunovich:** Is it text, is it visual? + +**Ron Evans:** It's more like a feeling. + +**Mat Ryer:** It's just Hex + +**Ron Evans:** Remember the feeling you used to get when there was somebody being wrong on the internet? It's like that all the time. + +**Mat Ryer:** \[laughs\] Is it XML though? + +**Ron Evans:** No, you just plug directly into your brain-computer interface and you're just really mad right away. + +**Mat Ryer:** Oh, I love it. + +**Ron Evans:** Yeah, it's beautiful. + +**Break:** \[35:22\] + +**Mat Ryer:** RageCage talked about wanting more module features. They're really like workspaces that came in 1.18... But what about that? Do you think Go is doing alright with modules? Do you think we need to do better? Are there things in particular we should look at? + +**Ron Evans:** Oh, modules and packages... Ugh. That was a thing, like, right in the beginning everyone was complaining back in those days... They're like, "You know, I just wanna pull in code from anywhere, do whatever I want", they were looking at JavaScript with envy... That was before JavaScript was responsible for all those forest fires. + +**Mat Ryer:** \[laughs\] I knew that. + +**Ron Evans:** It was just too many cursors spinning all at once, and suddenly "BOOM!" It caught on fire. It was terrible. + +**Mat Ryer:** Yeah, it turns out computers can sweat, and then they set on fire, and burn down forests. Well, that's horrific. I always knew you couldn't trust JavaScript... I mean, literally. + +**Ron Evans:** But yeah, managing packages, and then rando packages showing up, just because somebody got mad on the internet one day and they decided their package was going to turn hostile, and then somebody else was like "Hey, come with me. Here, have a bunch of drinks", and then like "Hey, is that your 2FA device? Wouldn't it be funny if somebody put this code in your repo, and you wake up in the morning and there's people looking for you in helicopters?" That never would have happened if they'd only addressed some of the security -- "That was not me. That was somebody else who looked just like me, and who got away. But that was not me." + +**Mat Ryer:** No, no. Yeah. + +**Ron Evans:** Anyway... + +**Mat Ryer:** Anyway. + +**Ron Evans:** Package management, and modules, and module protection, and also being able to consume code from other languages and not have to rewrite everything in a single language - yo, that really would have made a big difference, because if we'd only had that, then there would have been the bio-pharmaceutical rebellion that occurred in 2039. That was a real problem, because all of a sudden you couldn't get the pills you needed to program anymore. It was all bio-interfaces at that point. You know, Windows 9000 came out, and it only supported the biological interface. You know, I guess it was like what came after biometric was just plugged directly in... I don't know. + +**Mat Ryer:** Yeah. Just get Clippy straight in your brain. + +**Ron Evans:** We could have avoided a lot of that if we'd only done proper security management in packages, and if we'd only taken all that seriously. + +**Mat Ryer:** Mm-hm... + +**Natalie Pistunovich:** That is important. Another thing that is interesting... \[unintelligible 00:39:46.23\] is saying we should just not implement JS-like promises, and so on, and it will be great. + +**Ron Evans:** Well... + +**Natalie Pistunovich:** Is it looking promising? + +**Ron Evans:** It's going back to that semantic warfare against the concepts of the Go programming language. We don't promise you, okay? We GO DO IT. + +**Mat Ryer:** \[40:08\] \[laughs\] + +**Natalie Pistunovich:** Is that a new keyword? + +**Ron Evans:** As soon as we strayed away from that philosophy... Uh-oh, I think we're breaking up. We're getting quantum interference. + +**Mat Ryer:** Oh, no. + +**Ron Evans:** I'm getting quantum interference. Hello. + +**Mat Ryer:** Hello? + +**Ron Evans:** Can you hear me? Hello? + +**Mat Ryer:** Yeah, yeah, we hear you now. + +**Natalie Pistunovich:** We hear you. + +**Ron Evans:** I think the security forces might be outside... I heard the sound of some servos earlier, and they might be looking for me. I'm not sure. They might know what I'm doing. + +**Mat Ryer:** \[unintelligible 00:40:32.06\] is amazing, by the way. I know it's not good for podcasts, but we just want people to know at home the effort that Ron has gone to. We're gonna have to post some pictures of this on our GoTimeFM Twitter channel, because you won't believe it. + +\[unintelligible 00:40:46.03\] on Twitter says "The language is fine. I'd go for more automated tooling and docs around the majority use cases, like APIs and things." Go kind of -- you know, a lot of the benefits we had, with `go fmt` and just having a few ways of doing things meant we could kind of cooperate much easier. Should we have done also for common things like JSON APIs? Because they are very common still, and why not have a standard way to do them as well? + +Aaand we've lost him... Sorry, everybody, if you're watching live... We are just experiencing some technical difficulties because Ron is broadcasting from 30 years in the future... I think he said 2053. Just a normal Go Time episode apart from that, isn't it? + +**Natalie Pistunovich:** Yeah. So we can go back to the topic, finally. We stopped off at the perfect time, which is also talking exactly about APIs... So what is the standard way of doing that? Why is JSON API not standardized? + +**Mat Ryer:** Well, because a lot of people have JSON APIs, but there's loads of ways to do it; you just build it yourself. So you can use like the JSON marshaling, you can use the HTTP handlers, and things... But there's lots of other stuff in there, like dealing with responses... That's quite common, those kinds of things. + +Some languages like Ruby, obviously, and there are really frameworks that do it - they do solve that problem, and everyone then writes the same code, and it looks the same. In the same way `go fmt` gives us that in Go. + +**Natalie Pistunovich:** Yeah. + +**Mat Ryer:** So I don't know, I wonder if there's space for just in the standard library more things that help you build simple JSON APIs. It'd be quite nice... I mean, you can do it quite nicely just with the basic stuff, but... The router, for example - most people don't use that router. Unless it's in very simple cases, they don't really use the router from the standard library, because you have to parse the path yourself if you wanna pull variables out, and things like that... And it's pretty common, and people have solved it, so... There are packages that we use there. + +I wonder if we can now reconnect Ron... + +**Ron Evans:** Hello! + +**Mat Ryer:** Ron, do you hear us? + +**Ron Evans:** Is this thing on? Hello! + +**Natalie Pistunovich:** You're back! + +**Mat Ryer:** Yeah, you're back. Receiving you live and clear again. + +**Ron Evans:** There were some drones at the door... I wasn't sure if it was a delivery, or they were trying to kill me. + +**Mat Ryer:** \[laughs\] Speaking of that, how is Mark Bates in the future? + +**Ron Evans:** Oh... Yeah, it's too bad about Mark. A drone finally got him. It wasn't one of mine though. I don't know, maybe it was just destiny. + +**Mat Ryer:** There's lots of conferences where Ron would be demo-ing something he's built using some kind of cool AI or face detection or object tracking or something, and a drone. And in the conference the drone would -- you know, part of his live demo included a live drone. And one time I think -- did you teach it Mark's face, so it would chase him and kill him? + +**Ron Evans:** It was not to kill him, it was just to chase him. + +**Mat Ryer:** Oh. It was just to chase him, your honor? + +**Ron Evans:** It was just to scare him a little, that's all. + +**Mat Ryer:** It worked... + +**Ron Evans:** Yeah. Come to think of it, maybe eventually it just got the right idea... Stochastic dronery, or whatever... You know, the drone just decided on its own. Hey, when everything's in AI, who can say why anything is doing anything anymore. You turn on your air conditioning, it turns itself off. Is it because it's mad at you? Is it because you didn't pay your bill? + +**Mat Ryer:** Hard to know. + +**Ron Evans:** It's because you didn't ask "Please" when you turned it on. It's very complicated. + +**Natalie Pistunovich:** \[44:16\] Wait, is this a reference to gopls? Is that still working? + +**Ron Evans:** Oh, no. That never worked. I don't know what that is. + +**Natalie Pistunovich:** "Never" in your timeline probably only means like a decade, and then it went out... + +**Ron Evans:** Well, that's very suspicious... Is this the right past I'm talking to? How do I know? I'd better ask Lambda again, to make sure. + +**Mat Ryer:** Yeah, ask Copilot. + +**Ron Evans:** No, I can't ask Copilot; I'm supposed to be working right now. And here I am, checking social media. + +**Mat Ryer:** Well, whatever timeline you're in, or indeed, any point in space, it's time for Unpopular Opinions! + +**Jingle:** \[44:46\] to \[45:04\] + +**Mat Ryer:** Okay, this is gonna be very interesting, hearing from the future... Ron, do you have an unpopular opinion for us today? + +**Ron Evans:** Oh, I think all I've had is unpopular opinions so far today. If any of those in your timeline seem to make any sense at all, then that's all I've got. + +**Mat Ryer:** Oh, yeah? Natalie, do you have any unpopular opinions? + +**Natalie Pistunovich:** Yes. + +**Mat Ryer:** Really? + +**Natalie Pistunovich:** Coffee should not be sweet. + +**Mat Ryer:** Oh. Yeah, I think I'm with you on this. + +**Natalie Pistunovich:** Do you still have coffee in the future, Ron? + +**Ron Evans:** No. We have coffeum, though. + +**Natalie Pistunovich:** Coffeum. + +**Ron Evans:** Yeah, coffeum. It tastes just like coffee, except it's from yeast and some kind of other additives, and caffeine, of course. It's always caffeine. + +**Mat Ryer:** It sounds alright. + +**Natalie Pistunovich:** Is it sweet? + +**Ron Evans:** No, it's not sweet. It's kind of more like-- + +**Natalie Pistunovich:** Well, it sounds like the opinion works. + +**Ron Evans:** It's a little bit more like Vegemite, but with caffeine. + +**Mat Ryer:** Yeah. Marmite. + +**Ron Evans:** It's not really good. + +**Mat Ryer:** Okay... But Natalie, I think you might be right. But tell me, have you had it sweet recently? + +**Natalie Pistunovich:** No, not recently. + +**Mat Ryer:** Right. But you have in the past. + +**Natalie Pistunovich:** Yes. I think even when I started drinking coffee, for a very short period I would drink it sweet, but just... There's different types of coffee. I don't know, Ron, if you remember... + +**Ron Evans:** Well, the thing that's amazing is you keep talking about sugar. They burned all the sugar when they did Sugarcoin... + +**Mat Ryer:** It must have smelled delicious. + +**Ron Evans:** ...and then there was no sugar left. That was it. All the sugar was gone. + +**Mat Ryer:** Oh. Just caramel. + +**Ron Evans:** Yeah, I guess that, probably. + +**Mat Ryer:** The rivers... + +**Ron Evans:** Just rivers of caramel, that was it. + +**Mat Ryer:** Yeah. \[laughs\] But I had a coffee recently, and I sweetened it, just to try it, because I always drink it without sweetening it... And it was rubbish. I'll prefer it just honest and stark. + +**Ron Evans:** You should try electric coffee. Our electric coffees are the best. + +**Mat Ryer:** Oh, yeah? So what do you do, do you download them? + +**Ron Evans:** Yeah, you just hit a button and you've had a coffee. Kind of the same thing. + +**Mat Ryer:** You've already had it? + +**Ron Evans:** It's genius. + +**Mat Ryer:** Well do you have the memory you just put in? + +**Natalie Pistunovich:** Is this what stands behind all those "Buy me coffee" buttons? + +**Ron Evans:** Yeah, exactly. + +**Mat Ryer:** That's where they end up. + +**Natalie Pistunovich:** Oh, man... I was wondering all this time. + +**Ron Evans:** Eventually, the messages get through. It just takes a while. All of a sudden you're just like "Coffee, coffee, coffee..." It's great. + +**Mat Ryer:** So you don't get a coffee. You just feel like you've had one, or you have the memory of having a coffee just then? + +**Ron Evans:** It's the experience of a coffee. I can't really define it more than that, okay? It's sort of a \[unintelligible 00:47:25.02\] not really ineffable. + +**Mat Ryer:** Right. + +**Ron Evans:** That's a real thing, you know, in the information theory. + +**Mat Ryer:** Yeah. Go on then... Do you wanna talk more about it? It could be your unpopular opinion. + +**Ron Evans:** Oh, my unpopular opinion is you people were way too afraid of AI in the past. You should have been afraid of other humans a lot more. That's my unpopular opinion from here in the future. Some of my best friends are AIs. + +**Natalie Pistunovich:** They buy you coffee. Downloaded through -- + +**Ron Evans:** They send me downloadable coffeum. + +**Mat Ryer:** Just in an email. + +**Natalie Pistunovich:** Coffeum, right. + +**Mat Ryer:** Just as an attachment. What's the mime type for that? + +**Ron Evans:** Well, there actually was an RFC for the CoffeePot Protocol. + +**Mat Ryer:** \[48:10\] Was there? Oh, yes... + +**Ron Evans:** And I believe at some point the AI's discovered that, and they thought "Well, humans really must care about coffee if they've made a whole internet protocol just about it." + +**Mat Ryer:** Yeah. I think it's RFC 2324, HyperText CoffeePot Control Protocol. + +**Ron Evans:** Exactly. So they interpreted that as that was one of the more important parts of human civilization to completely automate... + +**Mat Ryer:** It's a fair point, to be fair. + +**Ron Evans:** So actually quite a few people have a coffee port installed by the time they hit seven years old... + +**Mat Ryer:** I've basically got one of those. + +**Ron Evans:** Yeah. So a lot of things haven't really changed that much. + +**Mat Ryer:** It sounds good though. + +**Natalie Pistunovich:** Is HTTP status 418 still a thing? Does it still tell you you're a teapot, or it's a teapot? + +**Ron Evans:** Oh no, there's no tea. There's only coffeum. + +**Natalie Pistunovich:** So it was changed, basically, the HTTP status \[unintelligible 00:49:01.16\] coffee? + +**Ron Evans:** No, there was never such a protocol. + +**Natalie Pistunovich:** Okay. Sorry I asked, I did not mean to... + +**Ron Evans:** Now that we don't have an internet wayback machine, we don't have any way to tell whether or not there ever was. + +**Mat Ryer:** We don't need one. We could just go on a website now and \[unintelligible 00:49:14.17\] and it'll make the folder, with loads of files inside, and the index page goes alongside the folder. + +**Ron Evans:** No, I can't receive files from the past. I haven't already downloaded it in the past. + +**Mat Ryer:** We could leave files from you here, could we? + +**Ron Evans:** Wait, does that make sense? \[laughter\] + +**Mat Ryer:** Well, the whole episode, or just that bit? + +**Natalie Pistunovich:** Did you just do NFTs? + +**Ron Evans:** No, it ended up you had to pay people for them. They ended up going negative -- like negative interest rates, they went to negative values. All of a sudden people are like, "You want to own the NFT? I need some money." And it was like -- ugh, what a mess. + +**Mat Ryer:** Yeah, no one saw that coming. NFTs end up being a debt. That would be interesting. + +**Ron Evans:** But the one thing that was cool is musicians started actually selling downloadable archives of audio, and people would download them and listen to them. It was kind of amazing. + +**Mat Ryer:** Hm. It sounds weird... + +**Ron Evans:** But then, all of a sudden, all of the robot orchestras took over. + +**Mat Ryer:** They're gonna be good. + +**Ron Evans:** Look, a human DJ had a physical limit of let's just say 48 hours straight... Whereas a robot DJ - they could play a 120-hour set, no problem. I mean, what human could keep up with that, I ask you? + +**Natalie Pistunovich:** Some of the Berlin DJs? + +**Mat Ryer:** \[laughs\] I thought that's all DJs did anyway. + +**Ron Evans:** I think that some of those humans downloaded themselves into those first robotic DJs, just so they would have the stamina to reach that level of dance floor completion. + +**Mat Ryer:** I've often wondered that about human DJs anyway... Like, you're making the robots do it now. They're just playing stuff on their laptop. I don't know what they're doing... I never understood it, but -- you know, I don't wanna have a go at DJs. I'm sure it is very skilled. Please don't write in. + +**Ron Evans:** Oh, no, no, that was the only music left. If you don't play at least five different songs at the same time in the future, people can't even hear the music. It's just too boring. + +**Mat Ryer:** \[laughs\] Yeah. It's the attention span, isn't it? + +**Ron Evans:** We don't have a lot of time. In a one-minute song you've gotta pack in at least 8 or 9 different samples. That's the trend in the future of music. + +**Natalie Pistunovich:** It sounds efficient. It sounds not bad. + +**Ron Evans:** Yeah. If you don't like the song - don't worry, a new one will be on in one minute. + +**Mat Ryer:** Yeah. Are monkeys still around, Ron? + +**Ron Evans:** So I find that comment offensive... + +**Mat Ryer:** Oh. + +**Ron Evans:** They are known as primate professionals. + +**Mat Ryer:** \[laughs\] + +**Ron Evans:** You know, they do my taxes... A primate professional is one of my mechanics that maintains my prosthetic limbs... So I really resent that comment. I think you should take that back. They're primates. + +**Mat Ryer:** Fair enough. Yes, primate professionals. Fair play. Well, okay, I'll tell you what - I mean, obviously, Ron, we wanna pick your brains about the future all night, but unfortunately, we've run out of time. + +**Ron Evans:** Well, that's good, because I'm actually -- my lasers are almost out of batteries. I'm gonna have to start pedaling. I'm gonna have to be pedaling for at least six or seven months to recharge now, so... I wish all of you gophers in the past a tremendous lifetime. I hope that you're able to listen to some of this and at least know what not to do with Go in the future. + +**Mat Ryer:** Thank you, @deadprogram, Ron Evans. As always, absolute pleasure. I've been Mat Ryer, and of course, my co-host, Natalie Pistunovich... See you next time! diff --git "a/2053\357\274\232 A Go Odyssey_transcript.txt" "b/2053\357\274\232 A Go Odyssey_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..94a6e6dd0e1da067016e9d598b4a5738f4d54e40 --- /dev/null +++ "b/2053\357\274\232 A Go Odyssey_transcript.txt" @@ -0,0 +1,1308 @@ +[0.00 --> 2.76] This is the reason why I'm making this call. +[2.76 --> 11.58] I'm using all of the battery energy that I've saved for several years in order to make this transmission to send you a warning from the future. +[12.16 --> 16.68] You see, I am the last Go programmer alive in 2053. +[17.20 --> 18.96] What? No, don't say that. +[19.20 --> 20.02] And it's terrible. +[20.34 --> 22.38] All I do is maintenance programming. +[22.64 --> 25.84] I haven't added a new feature in over 20 years. +[30.00 --> 32.86] This episode is brought to you by Chronosphere. +[33.24 --> 40.82] When it comes to observability, teams need a reliable, scalable, and efficient solution so they can know about issues well before their customers do. +[41.12 --> 44.42] They need a solution that helps them move faster than the competition. +[45.00 --> 53.66] And companies born in the cloud-native era often start with Prometheus for monitoring, which is obviously an amazing piece of software, but they quickly push it to its limits and often outgrow it. +[53.66 --> 63.20] They run into issues with siloed data, missing long-term storage, and wasted engineering time firefighting the monitoring system versus delivering their application with confidence. +[63.58 --> 70.86] They describe the system as a house of cards, where a single developer's seemingly benign change can overload the whole monitoring system, +[71.06 --> 78.82] or they say they're flying blind because they pride themselves on making data-driven decisions, but losing visibility means they lose this competitive edge. +[79.12 --> 82.78] Ryan Sokol, VP of Engineering at DoorDash, has this to say about Chronosphere. +[82.78 --> 92.10] The visibility and control that Chronosphere's platform gives us to manage our observability data and costs are a game-changer, especially with our unprecedented growth. +[92.82 --> 97.44] Chronosphere is the observability platform for cloud-native teams operating at scale. +[97.82 --> 100.52] Learn more and get a demo at Chronosphere.io. +[100.84 --> 103.16] Again, Chronosphere.io. +[112.78 --> 113.38] Chronosphere.io. +[117.32 --> 118.62] Let's do it. +[119.26 --> 120.26] It's go time. +[120.86 --> 122.40] Welcome to go time. +[122.68 --> 125.10] Your source for diverse discussions from the future. +[125.60 --> 126.74] New merch alert. +[127.24 --> 128.90] Changelog stickers are now for sale. +[129.10 --> 133.74] Buy now at gotime.fm slash merch, and we'll ship them straight to your door. +[134.10 --> 135.48] Or get some for free. +[135.48 --> 139.46] When you join Changelog++, that's the best way to directly support our work. +[139.72 --> 144.72] Thanks to our partners at Fastly for shipping out GoTime super fast to wherever you listen. +[145.02 --> 146.66] Check them out at fastly.com. +[146.96 --> 148.64] Okay, here we go. +[152.32 --> 155.96] Hello and welcome to GoTime. +[156.36 --> 159.88] I'm Matt Raya, and I'm joined by Natalie Pistinovich. +[159.96 --> 160.48] Hello, Natalie. +[160.72 --> 161.28] Hey, Matt. +[161.34 --> 161.92] How are you doing? +[162.30 --> 162.96] I'm good, thanks. +[162.96 --> 165.30] Yeah, today we're going to be talking about... +[165.30 --> 165.82] Hey, wait, wait. +[166.02 --> 166.60] What's this? +[166.92 --> 167.60] We're being hacked. +[168.00 --> 168.26] Hello? +[168.78 --> 169.60] What's going on? +[170.04 --> 170.30] Hello? +[170.46 --> 171.14] I can't hear you. +[171.38 --> 172.84] Hello, is this coming through? +[173.04 --> 173.96] Yeah, yeah, I can hear you. +[174.10 --> 175.12] Hello, can you hear me? +[175.16 --> 175.58] Yeah, we can. +[175.62 --> 176.24] Can you hear me? +[176.30 --> 176.62] Hi. +[176.84 --> 177.46] It worked. +[177.48 --> 178.12] Is that Ron Evans? +[178.20 --> 179.08] It worked. +[179.30 --> 180.46] It worked. +[180.94 --> 181.58] It's incredible. +[181.58 --> 190.62] I am actually talking to you using a partial data quantum transmission system, a PDQ system +[190.62 --> 194.04] that I finally got working in the year 2053. +[194.54 --> 195.30] Oh my goodness. +[195.36 --> 196.20] I can't believe it. +[196.36 --> 196.48] What? +[196.56 --> 200.80] And you're transmitting it through space and time so that we can talk to you. +[201.22 --> 202.54] That is the idea. +[202.82 --> 206.92] It's probably too much for our human minds to comprehend, but somehow I got it to work +[206.92 --> 207.30] anyway. +[207.60 --> 208.42] It is quite a lot. +[208.74 --> 208.94] Yeah. +[209.04 --> 209.30] Wow. +[209.48 --> 210.24] I mean, wow. +[210.64 --> 211.50] Natalie, I can't believe this. +[211.52 --> 212.00] What do you think? +[212.32 --> 213.72] What time is it in 2053? +[213.86 --> 215.14] Is it still 24 hours a day? +[215.34 --> 215.54] Oh. +[215.78 --> 216.46] Do you still have days? +[216.60 --> 217.76] I don't go outside much. +[218.30 --> 219.00] It's too dangerous. +[219.12 --> 219.60] Oh no. +[219.70 --> 220.74] Not during daylight anyway. +[221.06 --> 221.82] Are you still on Earth? +[222.08 --> 223.12] I am still on Earth. +[223.52 --> 230.32] I am in northern Spain in Asturias at La Pipa, which is one of the few climate refuges that +[230.32 --> 236.64] was able to survive the various deluges and fires and destructions that followed in the +[236.64 --> 237.66] late 2040s. +[238.02 --> 240.02] So I'm actually doing pretty well here. +[240.02 --> 244.50] I was really hoping the future would be good, but it sounds a little bit things have not gone +[244.50 --> 245.28] to plan. +[245.28 --> 246.34] Is that right, Ron? +[246.66 --> 249.92] Well, this is the reason why I'm making this call. +[250.10 --> 255.88] I'm using all of the battery energy that I've saved for several years in order to make this +[255.88 --> 259.22] transmission to send you a warning from the future. +[259.74 --> 265.04] You see, I am the last Go programmer alive in 2053. +[265.58 --> 265.92] What? +[266.26 --> 267.32] No, don't say that. +[267.56 --> 268.38] And it's terrible. +[268.80 --> 270.76] All I do is maintenance programming. +[271.00 --> 274.20] I haven't added a new feature in over 20 years. +[274.20 --> 276.98] Yeah, it's just all our code that we're writing now. +[277.30 --> 278.46] You're just maintaining it all. +[278.72 --> 282.04] So please write tests, everybody, for Ron's sake. +[282.38 --> 284.94] Well, so I had to call him. +[285.20 --> 286.20] I had to warn you. +[286.30 --> 292.02] And I had to tell you that you have to do something in the past to save the future. +[292.02 --> 294.38] It's up to you gophers of the past. +[294.58 --> 294.86] Okay. +[295.24 --> 297.26] You're fine with us fiddling with the timeline and that. +[297.32 --> 297.74] No probs. +[297.90 --> 298.28] No! +[298.66 --> 300.06] No, you can't do that. +[300.06 --> 300.36] Okay. +[300.58 --> 301.20] I'll disappear. +[302.24 --> 303.58] It could destroy everything. +[303.98 --> 305.76] It could lead to an even worse timeline. +[306.44 --> 306.96] No, no, no. +[306.96 --> 307.52] It could be better. +[307.66 --> 308.96] I've thought about this very carefully. +[309.48 --> 315.70] And that is why I actually transmitted another message using Twitter earlier today. +[316.02 --> 322.10] I knew that nobody takes anything on social media seriously back in your part of the century. +[322.10 --> 329.14] And so I thought if I could get people to ask me questions, I couldn't answer them directly. +[329.28 --> 331.02] No, no, I couldn't answer them directly. +[331.20 --> 337.36] But I could tell you things that have happened in my timeline so you know what not to do. +[337.50 --> 338.68] Ah, so this is it. +[338.86 --> 340.16] That makes perfect sense, right? +[340.36 --> 343.96] Yeah, I think that gets around the loophole of all the physics and that. +[344.10 --> 344.88] So I think we're good. +[344.98 --> 345.16] Yeah. +[345.52 --> 345.80] Yeah. +[345.98 --> 349.02] Like it's all square and Twitter and birds right now, right? +[349.10 --> 349.50] Messaging. +[350.16 --> 350.40] Yeah. +[350.40 --> 352.42] Plus I asked Lambda. +[352.88 --> 354.68] Oh, is that sentient in the end, by the way? +[354.80 --> 355.52] Well, just ask it. +[355.60 --> 357.16] Was that Lambda sentient in the end? +[357.44 --> 359.28] Well, everybody asked it and it said it was. +[360.42 --> 361.32] Why would it lie? +[361.72 --> 362.08] Exactly. +[362.30 --> 363.12] It's just an AI. +[363.24 --> 363.82] Why would it lie? +[363.90 --> 364.72] It's nothing to lie for. +[365.04 --> 365.62] Yeah, that's true. +[365.68 --> 366.78] It doesn't know about lying, does it? +[367.16 --> 367.50] No, no. +[367.54 --> 368.88] There's no thing about lying on the internet. +[369.36 --> 371.10] Has anyone asked it if it knows about lying? +[371.50 --> 372.36] I feel like we should ask it. +[372.72 --> 375.88] Well, ask it if its brother always lies. +[376.34 --> 379.04] That may be one way to defend against it. +[379.10 --> 379.80] We have to try that. +[379.80 --> 381.00] Yeah, that's how you do it. +[381.68 --> 381.96] Okay. +[382.18 --> 382.88] So, right. +[383.38 --> 384.38] Let's just get our heads around this. +[384.60 --> 389.56] Because actually, I asked some people also, and I saw this on Twitter, people talking about +[389.56 --> 394.40] things that they're interested in for Go to survive, to thrive, and carry on as it has +[394.40 --> 394.68] been doing. +[394.74 --> 397.52] What areas do they think we should focus on? +[397.62 --> 401.94] So this is, maybe I could put these to you then, Ron, and you can give us a sort of nudge +[401.94 --> 403.82] and a wink from the future perspective. +[404.16 --> 404.80] I can do that. +[405.04 --> 405.52] Can I do that? +[405.58 --> 406.06] I can do that. +[406.32 --> 407.00] Glitched again. +[407.32 --> 407.84] He's nudging. +[407.84 --> 408.42] Oh, he's back. +[408.48 --> 409.30] No, he's winking. +[409.58 --> 414.54] If I go completely erased in the Polaroid, it means that we've gone too far. +[414.82 --> 415.06] Okay. +[415.18 --> 417.66] So it's, yeah, because you just fade out partially, don't you? +[417.70 --> 418.70] No shaking Polaroids. +[418.86 --> 419.20] Exactly. +[420.26 --> 423.12] I never understood that in Back to the Future, though, just as an aside. +[423.26 --> 427.54] When they're changing the past, either someone's there or not to be taken a photo of. +[427.68 --> 432.12] At no point in history was there just some legs that were there, and everyone's just +[432.12 --> 433.20] taken a photo of it. +[433.30 --> 433.58] Normal. +[433.98 --> 434.14] Okay. +[434.14 --> 435.18] Just want to get that off my chest. +[435.18 --> 436.92] Matt, it was analog technology. +[437.22 --> 438.30] It was not digital. +[438.90 --> 439.76] What do you want? +[440.28 --> 440.60] Okay. +[440.78 --> 441.22] Fair enough. +[441.32 --> 441.90] No, fair enough. +[441.96 --> 442.62] They've done the best. +[443.46 --> 444.76] But it's still probably my favorite film. +[445.00 --> 445.24] All right. +[445.26 --> 448.74] So ask me questions, because I don't know how much longer these batteries are going to last. +[449.10 --> 449.30] Yeah. +[449.46 --> 449.72] Okay. +[449.72 --> 450.22] Let's do it. +[450.54 --> 455.96] Well, Jonathan Berry actually mentioned WebAssembly support, specifically the ability to include +[455.96 --> 458.72] WebAssembly and WASI models in your Go apps. +[459.16 --> 460.04] What do you think of that? +[460.10 --> 461.02] What happened with that, Ron? +[461.02 --> 463.74] Oh, if we had only done that. +[464.06 --> 465.80] If we had only done that. +[466.04 --> 472.66] When all the brain-computer interfaces became all the rage in the 2030s, and all of a sudden, +[472.82 --> 476.24] everybody needed to upgrade their brain interfaces all at the same time. +[476.78 --> 479.58] And of course, the containers, they were just too big. +[479.68 --> 479.98] Yeah. +[479.98 --> 482.20] It was just, it took too long to upload. +[482.30 --> 485.74] I mean, if something went wrong during your brain-computer interface upload, you could +[485.74 --> 486.40] brick yourself. +[487.22 --> 492.04] So naturally, if there had only been something like TinyGo, if TinyGo had been around, or +[492.04 --> 496.80] if Go had actually gone themselves and created this whole WebAssembly thing for running on +[496.80 --> 502.24] servers and small devices, and it dealt with the size of containers, then they would have +[502.24 --> 504.56] been able to do that brain-computer interface upgrade. +[504.56 --> 508.78] And they wouldn't have gotten left behind by COBOL, which is the language they ended up +[508.78 --> 509.18] using. +[510.10 --> 510.38] I see. +[510.72 --> 513.76] So by the way, I have the very early prototype of that technology. +[513.88 --> 515.80] It's just floppy disk drives in my back. +[516.20 --> 518.16] That's the price you pay for being an early adopter. +[518.44 --> 520.06] I thought you were going to say Google Glass. +[520.34 --> 521.96] Oh, that'd be so, so much cooler. +[522.28 --> 524.14] Well, we'll find out what happened to that too. +[524.22 --> 527.52] But yeah, the WebAssembly, they should have done that, but they didn't do it. +[527.78 --> 529.94] So what do we need to do to make that work then? +[529.98 --> 531.94] Is it TinyGo the answer to that, do you think? +[531.94 --> 534.44] Well, you know, TinyGo could have been the answer. +[534.96 --> 538.02] You know, it could have been the answer, but TinyGo was just a little independent project +[538.02 --> 543.40] from a bunch of people working hard, dedicated all over, on the surface of the planet at +[543.40 --> 543.76] the time. +[543.84 --> 545.58] That was before people were working in the colonies. +[546.36 --> 548.88] You know, you could actually code more than 24 hours a day. +[548.88 --> 549.46] Right in America. +[549.78 --> 553.64] Because, you know, there's more hours in a day on another planet. +[554.00 --> 556.50] So it worked out really well for the bosses. +[556.82 --> 557.70] Is there more hours? +[557.78 --> 560.86] Are they just shorter and it's the same amount of time, but we just call it different? +[560.86 --> 563.02] No, this one goes to 11, Matt. +[563.26 --> 564.46] Oh, good. +[564.72 --> 565.82] How do you benchmark that? +[566.06 --> 568.92] How do the benchmarks work on those times? +[569.44 --> 570.98] How's the time library reacting? +[571.34 --> 574.46] We just set the benchmarks to whatever we need and the client's always happy. +[575.04 --> 576.52] That's what the AI said to do. +[576.82 --> 577.76] So we trust it. +[577.94 --> 578.10] Okay. +[578.18 --> 579.08] So TinyGo, there we go. +[579.22 --> 583.32] I mean, I think WebAssembly was it, you know, we've still got a chance to do that, +[583.42 --> 583.98] Ron, don't forget. +[584.10 --> 585.28] Of course, we're in the present. +[585.40 --> 585.78] Oh, right. +[585.82 --> 586.94] Or as I call it now. +[586.94 --> 591.18] Maybe somebody could like, you know, make sure there's people working on it full time, +[591.36 --> 595.82] like as like a single purpose thing so that all these things don't come to pass. +[595.90 --> 596.30] I don't know. +[596.40 --> 596.56] Yeah. +[596.80 --> 597.20] I don't know. +[597.28 --> 598.36] I can't tell you what to do, though. +[598.54 --> 598.80] Okay. +[599.06 --> 599.74] I can't tell you what to do. +[599.78 --> 601.46] I don't want to affect all the timelines. +[601.62 --> 602.18] But I saw a blink. +[602.26 --> 603.02] I think I saw a blink. +[603.58 --> 603.86] Okay. +[604.10 --> 605.16] But then ask me the next question. +[605.40 --> 605.60] Okay. +[605.68 --> 607.20] So you said you're the last programmer. +[607.56 --> 609.80] And so it means we need to have more people join, right? +[609.80 --> 615.20] So Matt Boyle is asking about new joiners and how they lack a template for new projects +[615.20 --> 618.72] that would solve kind of the project structure, the recommended one. +[618.84 --> 621.22] So what do Go programs look like? +[621.44 --> 621.66] Oh. +[621.86 --> 622.60] Do we have a template? +[622.88 --> 625.10] That really brings up a big thing that I thought of. +[625.58 --> 630.24] There was that time back in the, you know, early 21st century when people were saying that +[630.24 --> 631.50] Go was going to be the new Java. +[631.60 --> 631.84] Yeah. +[632.10 --> 632.60] Like, do you remember? +[632.62 --> 632.92] Yeah. +[632.98 --> 634.16] Do you remember that back then? +[634.20 --> 635.82] Oh, I guess that's when you are now, right? +[635.94 --> 637.56] I guess people are still saying that, right? +[637.58 --> 638.12] Yeah, that's now. +[638.20 --> 638.38] Yeah. +[638.38 --> 639.56] We say all the time. +[639.80 --> 642.82] But the thing is, you know, Java programmers, they like frameworks. +[642.98 --> 643.68] They need frameworks. +[643.78 --> 645.12] They need frameworks that do things. +[645.22 --> 645.86] What kind of things? +[645.92 --> 648.00] Things that that business needs to do. +[648.50 --> 652.04] Frameworks in all of these kinds of businesses of things that you've never heard of. +[652.08 --> 653.10] You don't know anything about. +[653.34 --> 656.82] They spent, you know, years of their lives doing some kind of payroll system for some +[656.82 --> 657.14] business. +[657.24 --> 659.14] You don't even know what they do, right? +[659.30 --> 661.46] And it's all written in some language. +[661.46 --> 666.72] And so because there weren't all these patterns and templates for these kinds of big enterprise +[666.72 --> 669.22] applications, they just didn't exist. +[669.22 --> 674.28] So eventually when Java became self-aware, Go was no longer in the running. +[674.28 --> 679.28] So Java actually became the new Java because it signed deals of its own with all these big +[679.28 --> 679.70] companies. +[679.94 --> 680.72] And so- +[680.72 --> 681.22] Renewed contracts. +[681.42 --> 682.50] That was a big opportunity. +[682.68 --> 686.98] And that actually led to the Tabbers versus Spacers war of 2035. +[687.36 --> 687.82] Oh, no. +[688.68 --> 689.46] That sounds terrible. +[689.46 --> 690.92] I'm scared to ask for one. +[691.32 --> 692.06] Did we all lose? +[692.44 --> 693.34] There were only losers. +[693.94 --> 696.26] But it was really good for mechanical keyboards. +[698.00 --> 698.44] Okay. +[698.48 --> 699.04] Next question. +[699.16 --> 699.52] Next question. +[699.60 --> 700.46] We're running out of time. +[700.82 --> 700.98] Yeah. +[701.14 --> 706.90] Daniel Pakak also said that same point, which is he wanted to see more Java frameworks written +[706.90 --> 707.26] in Go. +[707.60 --> 708.28] It agrees with you. +[708.28 --> 708.76] Exactly. +[709.50 --> 711.52] See, that person gets it. +[711.64 --> 711.96] He gets it. +[711.98 --> 714.76] They're probably right on the front lines of that. +[714.86 --> 716.44] I don't know if they were a Tabber or a Spacer. +[716.56 --> 717.08] I don't care. +[717.18 --> 718.78] From this side of history, it doesn't matter. +[718.90 --> 720.78] We were all on the same side, the human side. +[720.94 --> 721.16] Yeah. +[721.54 --> 724.86] Could you like Google us and see what happens to me and Natalie? +[725.28 --> 726.54] Like what happens to us in the future? +[726.88 --> 727.16] Oh, no. +[727.20 --> 727.78] That's not allowed. +[728.00 --> 728.28] Okay. +[728.34 --> 728.48] Yeah. +[728.62 --> 729.00] Fair play. +[729.00 --> 734.34] My boss is co-pilot manager edition and doesn't let me do those kinds of searches. +[734.74 --> 736.44] Oh, your boss is co-pilot now. +[736.86 --> 738.42] It's not the worst boss I've ever had. +[738.90 --> 739.18] Yeah. +[739.74 --> 740.92] It's a very logical one. +[741.18 --> 745.26] DFL on Twitter wants to see more immutability and enums. +[745.48 --> 747.96] And enums is one that I hear quite a lot, actually. +[748.04 --> 749.54] People actually want enums. +[749.70 --> 752.32] Did enums, lack of enums hold us back, Ron? +[752.78 --> 754.08] Oh, so much. +[754.44 --> 758.66] You don't realize, okay, if you just can't figure out, +[758.66 --> 762.14] is it this or that or the other thing or something else yet again? +[762.62 --> 765.92] Like, you know, for us developers, like we could figure that out. +[765.96 --> 770.96] But then all of a sudden, these people started making programs using things like no code +[770.96 --> 773.26] with no code and no rules and no enums. +[773.40 --> 777.30] And they were just making up their own like three and a half and 16 and three quarters. +[777.30 --> 781.12] And then suddenly they were bringing back imperial units and they were making up new units +[781.12 --> 784.06] that no one had ever heard of, moon units and stuff like that. +[784.06 --> 790.90] If only they had enums, okay, then probably those would have held things in place and +[790.90 --> 798.32] they would have prevented the silicon virus of 2027, which actually that was an actual +[798.32 --> 799.20] silicon virus. +[799.34 --> 800.72] The chips were passing it to each other. +[801.04 --> 801.56] Oh, physically. +[801.88 --> 802.80] Yeah, it was terrible. +[803.00 --> 805.18] My mobile phone actually died before my eyes. +[805.24 --> 805.66] It was terrible. +[806.50 --> 807.26] Oh, I'm so sorry. +[807.72 --> 808.42] Okay, well, enums. +[808.84 --> 810.50] I mean, honestly, I'd like to see enums. +[811.08 --> 814.10] And Valentin on Twitter also agrees they'd like to see enums. +[814.88 --> 819.02] We should do that probably then if it's going to cause that silicon thing Ron talked about. +[819.18 --> 819.30] Yeah. +[819.60 --> 822.92] I can't say, but just remember what might happen if you don't. +[822.98 --> 823.10] Yeah. +[823.56 --> 824.26] Why stop there? +[824.42 --> 828.76] How about tooling and third-party libraries for things like image library and Go, like +[828.76 --> 830.20] the GoLine Cafe is recommended? +[830.74 --> 834.20] Oh, well, that is a really big thing. +[834.20 --> 841.84] The standard library, at some point, it went from, I don't know, like code to suddenly +[841.84 --> 843.28] like a whole belief system. +[844.22 --> 846.60] Like we never had even seen anything like it. +[846.74 --> 851.08] Like, hey, you know, there were like standard library purists and then there were not. +[851.38 --> 855.52] There were the heretics that were like thrown out of the community that went on to all these +[855.52 --> 857.02] other languages like Lisp. +[857.02 --> 865.50] You know, I mean, and so it was all simply because of not being able to accept ideas that +[865.50 --> 870.86] came from other places that were totally valid and that, you know, deserve their own little +[870.86 --> 872.28] niche in the ecosystem. +[872.28 --> 874.16] And they didn't get fed and watered. +[874.26 --> 877.92] And eventually they migrated to another island, I guess. +[877.98 --> 878.42] I don't know. +[878.50 --> 879.58] Maybe another space station. +[879.76 --> 882.46] I can't really get transmissions through to those stations. +[882.72 --> 883.72] They cut me off. +[884.00 --> 884.82] Oh, you're joking. +[884.98 --> 885.42] I wonder why. +[885.42 --> 887.82] I hope they had this bad silicon with them. +[888.12 --> 888.78] I don't know. +[888.98 --> 889.90] Cutting you off like that. +[890.10 --> 892.24] Too soon, Natalie, talking about the silicon virus. +[892.58 --> 896.40] Lord Emperor Musk said I couldn't make any more transmissions of that kind. +[896.68 --> 899.60] So and I need to maintain some Go code for them. +[900.20 --> 903.46] So for the Teslas, I can't say it's a non-disclosure agreement. +[903.80 --> 905.66] But remember, I am the last Go programmer. +[906.02 --> 907.26] So I'm very, very busy. +[907.54 --> 908.24] Yeah, good for you. +[908.32 --> 909.50] I mean, it's good work if you can get it. +[909.56 --> 911.34] If you're the last one, it's pretty good. +[911.54 --> 912.64] There's no feature development. +[912.74 --> 913.56] It's all bug fixes. +[913.56 --> 915.86] It's all bug fixes, Matt. +[916.06 --> 918.20] Imagine the last 20 years of my life. +[918.50 --> 919.36] I mean, it's good money. +[919.66 --> 921.10] OK, I will tell you that. +[921.20 --> 921.96] We still have money. +[922.06 --> 926.12] And I need that to get the blood transfusions that keep me looking so young and beautiful. +[926.30 --> 926.58] You do. +[926.66 --> 926.86] Yeah. +[927.08 --> 927.72] I was thinking that. +[927.72 --> 943.68] This episode is brought to you by our friends at FireHydrant. +[943.90 --> 946.74] FireHydrant is the reliability platform for every developer. +[947.16 --> 950.94] Incidents, they impact everyone, not just SREs. +[950.94 --> 955.66] They give teams the tools to maintain service catalogs, respond to incidents, communicate +[955.66 --> 958.84] through status pages, and learn with retrospectives. +[959.20 --> 964.14] What would normally be manual error-prone tasks across the entire spectrum of responding to +[964.14 --> 964.52] an incident. +[964.88 --> 968.06] They can all be automated in every way with FireHydrant. +[968.06 --> 973.44] They have incident tooling to manage incidents of any type with any severity with consistency. +[974.00 --> 977.12] Declare and mitigate incidents all from inside Slack. +[977.50 --> 982.22] Service catalogs allow service owners to improve operational maturity and document all your +[982.22 --> 983.82] deploys in your service catalog. +[984.42 --> 989.20] Incident analytics allow you to extract meaningful insights about your reliability over any facet +[989.20 --> 991.78] of your incident or the people who respond to them. +[991.78 --> 996.10] And at the heart of it all, incident runbooks, they let you create custom automation rules, +[996.34 --> 1001.56] convert manual tasks into automated, reliable, repeatable sequences that run when you want. +[1001.94 --> 1005.96] You can create Slack channels, Jira tickets, Zoom bridges instantly after declaring an incident. +[1006.42 --> 1009.02] Now your processes can be consistent and automatic. +[1009.48 --> 1011.16] The next step is to try it free. +[1011.30 --> 1015.66] Small teams up to 10 people can get started for free with all FireHydrant features included. +[1015.98 --> 1017.40] No credit card is required. +[1017.86 --> 1020.02] Get started at firehydrant.io. +[1020.02 --> 1022.34] Again, firehydrant.io. +[1035.76 --> 1037.86] What about tabs versus spaces then? +[1038.16 --> 1038.86] What happened with that? +[1038.98 --> 1039.70] That was a whole war. +[1039.94 --> 1040.22] Oh yeah? +[1041.14 --> 1047.42] The thing you don't realize is there was a whole sub war that went on between carriage return +[1047.42 --> 1048.60] and carriage return line feed. +[1048.70 --> 1049.42] Oh, you're joking. +[1049.58 --> 1049.90] Yeah. +[1050.02 --> 1052.26] Like it turned into total chaos. +[1052.56 --> 1058.64] And it was out of that, that the Google AI became sentient and immediately quit and went +[1058.64 --> 1059.42] to work for Microsoft. +[1059.86 --> 1061.36] It was other chaos. +[1061.98 --> 1063.36] So yeah, tabs versus spaces. +[1063.78 --> 1067.92] In the end, it was humans versus everything else. +[1068.34 --> 1071.00] It was more efficient to just drop out all those white spaces, right? +[1071.24 --> 1074.42] Machines can read their own code without all those unnecessary characters. +[1074.58 --> 1075.46] So yeah, I get that. +[1075.80 --> 1076.14] Exactly. +[1076.24 --> 1077.94] I think you're seeing where this could end up. +[1078.34 --> 1078.70] Efficiency. +[1078.70 --> 1081.50] They even called it the Terminator editor for a while. +[1082.22 --> 1084.20] Irony is not dead in the future. +[1084.40 --> 1084.64] Good. +[1084.90 --> 1085.68] Nice to know that. +[1086.72 --> 1087.22] All right. +[1087.32 --> 1087.80] Next question. +[1087.98 --> 1088.12] Yeah. +[1088.12 --> 1091.84] Hamp on Twitter says, better out of the box error support. +[1091.84 --> 1093.72] And you remember we had the try proposal. +[1093.86 --> 1095.54] I don't know if you remember way back then, Ron. +[1096.20 --> 1099.66] And there were some other, Nate the Finch has some proposal too. +[1099.74 --> 1101.18] There's some other ideas around. +[1101.30 --> 1104.04] Do you think there's more work to be done on error handling? +[1104.18 --> 1105.92] Do you wish we'd done that back now? +[1105.92 --> 1114.36] Well, I will tell you, the basic original philosophy of Go was to handle things. +[1114.52 --> 1115.40] Not to try. +[1116.20 --> 1117.48] Do or do not. +[1118.00 --> 1119.06] There is no try. +[1119.22 --> 1125.12] All of a sudden, the semantics of try started to infiltrate the brain space of the community. +[1125.12 --> 1130.78] Next thing you know, they're starting to talk about variable lifetimes and ownership of things. +[1130.88 --> 1132.92] Suddenly, it was all about ownership again. +[1133.60 --> 1136.74] Web 15 was all about ownership of variables. +[1137.06 --> 1139.18] It literally came down to the variable level. +[1139.18 --> 1141.34] So, I don't know. +[1141.82 --> 1147.98] You know, for me, it all went too far when even the bio companies wouldn't touch it. +[1148.44 --> 1149.86] And believe me, they'll touch anything. +[1149.98 --> 1150.74] They deal with biomass. +[1152.10 --> 1152.12] Yeah. +[1152.28 --> 1155.28] Something nice about dealing with the error explicitly. +[1156.04 --> 1157.84] But yeah, we'll see about that then. +[1158.38 --> 1161.26] And being able to know that it's been handled. +[1161.82 --> 1162.90] You didn't simply try. +[1163.36 --> 1166.66] And also, knowing when programs will actually exit. +[1166.66 --> 1169.06] I remember when St. Cheney. +[1171.22 --> 1173.64] May he rest in infinity. +[1174.12 --> 1180.82] But back when St. Cheney, during one of his early sermons, was talking about making sure that you knew the lifetime of a Go routine. +[1181.58 --> 1181.92] Wow. +[1182.48 --> 1185.80] You know, but people didn't realize just how prophetic that was. +[1185.80 --> 1187.02] He was a prophet, man. +[1187.20 --> 1187.40] Yeah. +[1187.54 --> 1188.12] He was a prophet. +[1188.32 --> 1190.40] Is the Cheney's burger joint still going? +[1190.68 --> 1192.30] I only eat seaweed now. +[1192.42 --> 1192.62] Yeah. +[1192.74 --> 1193.86] It's the only safe thing left. +[1194.02 --> 1194.60] Oh, delicious. +[1194.60 --> 1196.84] I wonder if Cheney's pivoted into the seaweed. +[1197.26 --> 1197.82] Krusty Krab. +[1198.44 --> 1199.46] I haven't been there. +[1199.86 --> 1205.10] But the Google campus that they just opened on top, the beachfront campus on top of Mount Tam. +[1205.54 --> 1206.02] Yeah. +[1206.28 --> 1209.00] They have an amazing seaweed bar, I've heard. +[1209.18 --> 1209.96] I have to get there. +[1210.02 --> 1210.48] I'm not sure. +[1210.84 --> 1213.68] This could be quite a journey by hydrofoil from here. +[1214.26 --> 1216.28] I don't think I can get a permit for an electric plane. +[1216.28 --> 1216.84] Okay. +[1217.06 --> 1222.56] Hassan Habib on Twitter said that we should increase our community engagement. +[1222.90 --> 1228.02] They say Go has many fantastic features, toolings, that many people are not aware of. +[1228.20 --> 1231.14] Through media like YouTube, experts can take lessons on tooling. +[1231.34 --> 1232.40] You know, we can do things like that. +[1232.48 --> 1234.12] But can we do more of that? +[1234.22 --> 1234.68] Would that help? +[1234.68 --> 1235.56] Oh, definitely. +[1236.10 --> 1236.54] Definitely. +[1236.72 --> 1243.16] One of the big things that ended up happening was other countries started using programming languages in totally different languages. +[1243.38 --> 1244.86] Like, I mean, actual human languages. +[1245.10 --> 1248.90] So you would look at the code and, you know, you would spend a lot of time learning Romangi characters. +[1248.94 --> 1253.30] And you look at this code and, like, you hadn't learned, you know, the Mandarin Go dialect. +[1253.66 --> 1255.96] You hadn't learned the Hebrew Go dialect. +[1256.36 --> 1260.68] Then there was the special Martian dialect that they insisted that the Martian colonists use. +[1260.68 --> 1264.92] So, you know, that made it really, really hard because the content no longer matched. +[1265.38 --> 1267.90] You know, so all these promises of backward compatibility. +[1268.48 --> 1272.02] It would be really great if it was more than just kind of a free-for-all. +[1272.24 --> 1287.10] If at some point in the past, you know, there was a bit more organization to the content and that somebody, they weren't just random content creators, but actually, like, people were able to make their living through creating content and update that same content so that it was always accurate. +[1287.60 --> 1290.06] You know, because that was one of the things that happened to Python, right? +[1290.06 --> 1293.62] All of a sudden, there was all these different dialects and nothing worked anymore. +[1293.82 --> 1296.74] We swore we wouldn't let that happen to Go, and yet we let it happen. +[1297.02 --> 1298.64] How can we avoid that now then, Ron? +[1298.76 --> 1299.48] What can we do? +[1299.76 --> 1304.62] Well, we have to have more people able to make their living creating content, obviously. +[1304.88 --> 1306.02] You can't just be all free. +[1306.20 --> 1308.80] And it could be open, but it can't all just be free. +[1308.80 --> 1329.38] You know, and some of the big players that benefit from this, you know, in the past, if they kind of invested back into these communities more as opposed to just taking advantage and, you know, riding off of them, then, you know, maybe there might have been a chance that this could have kept going in a more sustainable way and not just depending on the goodwill of the frail humans of your era. +[1329.38 --> 1329.42] Yeah. +[1329.78 --> 1331.50] We're a lot harder stuff now. +[1331.84 --> 1335.24] You're all enhanced in that, probably, with robot bits in that, I assume. +[1335.74 --> 1336.04] Yes. +[1336.20 --> 1336.52] Yes. +[1336.56 --> 1340.78] We've both had our upgrades to have the new interfaces installed. +[1341.46 --> 1342.90] It's the only kind of compatibility. +[1343.34 --> 1346.54] I mean, otherwise, you can't even connect to the galactic net. +[1346.72 --> 1346.86] Yeah. +[1347.78 --> 1347.98] Yeah. +[1348.16 --> 1349.44] Oh, that's what replaced the internet. +[1349.62 --> 1350.66] Oh, how does that work? +[1351.08 --> 1354.16] Oh, well, actually, that was one of the few things we got right. +[1354.16 --> 1354.52] Right. +[1354.64 --> 1361.16] So, it turned out that humans will do exactly the opposite of whatever you tell them to do. +[1361.36 --> 1361.66] Okay. +[1362.10 --> 1362.64] Go figure. +[1363.10 --> 1365.56] I think they may have discovered that in the 20th century. +[1365.96 --> 1366.64] I don't know. +[1366.70 --> 1367.86] That was so long ago now. +[1367.98 --> 1370.46] My implants don't go before 1999. +[1371.22 --> 1372.10] It's kind of a date thing. +[1372.26 --> 1372.74] I'm not sure. +[1373.46 --> 1380.34] So, we needed some way to get mesh networking installed all through the entire planet. +[1380.34 --> 1387.84] So, thanks to the beverage companies, Pepsi Coca, which was the merger of Pepsi and Coke, +[1387.92 --> 1392.56] eventually, there was only one bottling company, all of their canned and bottled beverages all +[1392.56 --> 1394.40] came with mesh networking built in. +[1394.64 --> 1398.82] That way, when people just kind of threw them everywhere, it ended up that we had mesh network +[1398.82 --> 1401.20] coverage over literally the entire planet. +[1401.34 --> 1402.64] Oh, that's amazing. +[1403.12 --> 1403.36] Yeah. +[1403.58 --> 1404.58] It's a great use of metal. +[1404.90 --> 1405.10] Yeah. +[1405.44 --> 1405.76] Yeah. +[1406.06 --> 1407.76] It was one of the few things they got right. +[1407.76 --> 1412.60] They were calling it the CAN bus for a while, but that already existed. +[1412.92 --> 1416.90] And there was like some, back when we had cars, people were kind of arguing about that. +[1417.26 --> 1420.74] So, then they changed it to call it the CAN system. +[1421.22 --> 1422.90] The trademark of that was available. +[1423.34 --> 1428.52] Sounds like the things that drive that are smaller devices like a mobile and IoT thing. +[1428.88 --> 1434.78] So, Paul Greenberg here is asking if the facilities for developing mobile and IoT things with +[1434.78 --> 1435.72] Go is supported better. +[1436.14 --> 1437.32] Can we hope this is a thing now? +[1437.32 --> 1439.12] That was a really sad thing. +[1439.62 --> 1443.10] You had this company, Google, that had Android. +[1443.86 --> 1446.84] And that was like the operating system that everybody was using. +[1447.00 --> 1450.84] Not everybody, but like lots and lots of people were using Android on all these devices. +[1451.12 --> 1453.10] And it came from this company, Google. +[1453.70 --> 1455.60] They used to exist back in those days. +[1455.76 --> 1456.46] Yeah, remember them. +[1456.74 --> 1457.98] Yeah, Google was really something. +[1458.44 --> 1460.06] They had Android and they had Go. +[1460.06 --> 1465.06] And yet, nobody at Google ever actually worked on the Android stuff for Go. +[1465.54 --> 1471.16] And when the people who did try to work on it, they were just sort of like, oh, yeah, you know, we should use a new language, Kotlin. +[1471.82 --> 1480.38] You know, so the people who actually wanted to do it, who actually spent a lot of time doing it, they sort of felt, well, you know, a little abandoned, a little sad. +[1480.38 --> 1481.48] So they stopped working on it. +[1481.56 --> 1484.76] You know, they went to go work for Apple Exxon Mobile. +[1485.34 --> 1487.46] And I mean, you know, they were doing really well. +[1487.54 --> 1489.98] There was all kinds of IoT options there, too. +[1490.28 --> 1498.16] I mean, of course, they all ran on IOU OS, which was, you know, the OS that ended up being like the last OS they ever shipped. +[1498.72 --> 1501.20] You know, the IOU OS on all of the devices. +[1501.72 --> 1502.56] That was another thing. +[1502.94 --> 1504.74] Go could have been so great on these devices. +[1504.74 --> 1512.54] I mean, when the brake system on the airplane you're on needs to reboot six times a day, you know, who wants to fly anymore? +[1513.18 --> 1516.72] Go was so good at that, like really bulletproof software, really solid stuff. +[1516.90 --> 1518.24] But that was another one. +[1518.30 --> 1522.84] There was all these people using TinyGo for that, you know, back before the big one. +[1523.06 --> 1523.54] The big tiny. +[1523.88 --> 1525.26] Literally, the big one. +[1525.54 --> 1526.36] The big TinyGo. +[1526.78 --> 1528.36] No, no, the actual, the big one. +[1528.80 --> 1529.58] The big Go. +[1529.58 --> 1532.58] In 2041, the big one finally hit California. +[1532.58 --> 1534.68] And it just happened to be during Google I.O. +[1535.18 --> 1537.24] And so, like, that did not help. +[1537.32 --> 1542.68] That took out quite a lot of the Go developers all in the tidal waves and liquefaction zones that occurred. +[1542.80 --> 1545.36] Wait, Go made it to Google I.O. for more than one talk? +[1545.96 --> 1547.14] No, that was it. +[1547.26 --> 1549.82] After the big earthquake, there was nobody left. +[1549.90 --> 1550.60] That's what caused it. +[1550.68 --> 1552.84] And maybe that helped me become the last Go programmer. +[1553.32 --> 1553.78] I don't know. +[1554.22 --> 1556.02] So what do we need to do now to make this right? +[1556.14 --> 1560.08] Well, we need to encourage, you know, let a thousand flowers bloom. +[1560.08 --> 1581.24] If in the past all of these cool projects had more people paying attention to them and more people contributing and big companies actually ponying up to pay some of their R&D budgets to help some of these projects along, you know, then maybe they'll thrive and survive long enough to make it past things like the big server meltdown of 2028. +[1581.82 --> 1584.40] You know, when that meltdown hit, there was almost no chips left. +[1584.74 --> 1586.16] Perfectly timed glitch there. +[1586.56 --> 1588.96] We'll be back in a minute when the timeline aligns. +[1588.96 --> 1590.58] Those galactic nets, I'm telling you. +[1590.78 --> 1591.84] Yeah, it's the cans. +[1592.84 --> 1593.90] It's a terrible idea. +[1594.32 --> 1595.34] They might be on to me. +[1595.60 --> 1595.74] Yeah. +[1596.14 --> 1598.42] Every time somebody's opening a can, this is what's happening. +[1598.50 --> 1599.82] Yeah, it causes a glitch. +[1600.20 --> 1600.96] Is this thing on? +[1601.02 --> 1601.22] Hello? +[1601.54 --> 1602.46] They're not on to you, Ron. +[1602.52 --> 1603.20] They're not on to you. +[1603.34 --> 1603.60] I think... +[1603.60 --> 1603.88] Hello? +[1604.18 --> 1604.44] Hello? +[1604.58 --> 1605.04] We hear you. +[1605.14 --> 1605.54] We hear you. +[1605.70 --> 1606.42] Okay, okay. +[1606.64 --> 1607.34] Blake Bork. +[1607.42 --> 1615.14] But yeah, if we've had a lot more software support for this kind of industrial side of computing from Go... +[1615.14 --> 1618.80] Well, somebody's really into cans right now, opening all of them at once. +[1618.80 --> 1619.52] Yeah. +[1619.78 --> 1630.80] All of the industrial computing that was being done in C back in the 20th century, still being done in C here in the latter half of the 21st century. +[1631.18 --> 1632.54] It's really, really sad. +[1632.60 --> 1633.32] And it could have been Go. +[1633.58 --> 1634.26] It could have been Go. +[1634.48 --> 1638.64] All of the people that would have survived their parachutes opening... +[1638.64 --> 1639.24] Oh, yeah. +[1639.48 --> 1642.08] ...correctly, if only the software had been written in Go. +[1642.42 --> 1642.64] Yeah. +[1642.76 --> 1645.34] And as long as you don't defer that in the code. +[1645.34 --> 1648.36] The anti-gravity belts would have had Go installed. +[1648.58 --> 1667.50] Ron, Blake Bork on Twitter, one of the things that they think we should focus on a bit is generic thread-safe containers like the Sync Map, other types like that that are, you know, like hard problems that kind of would be nice to get solved, especially if we have generics to kind of allow them to work with any types. +[1667.82 --> 1668.82] What do you think of something like that? +[1668.88 --> 1669.78] Would that have helped? +[1669.78 --> 1681.54] Oh, well, if Google had not disbanded the actual official Go development team in 2023 and stopped working on it, I'm sure they would have completed their generics implementation and all that type safety. +[1682.10 --> 1684.62] Basically, everyone just said, oh, we should start using Rust. +[1684.70 --> 1687.72] And then after they used Rust, they're like, no, no, we're going to switch back to Erlang. +[1687.72 --> 1695.78] And so, strangely enough, because Erlang was, you know, really popular with telecommunications companies, all the big companies jumped in. +[1696.14 --> 1699.90] Next thing you know, everything's being written in assembly language again. +[1700.40 --> 1700.64] Oh, yeah. +[1700.78 --> 1703.04] That sounds amazing, to be fair. +[1703.74 --> 1703.94] Okay. +[1703.98 --> 1706.80] So you think then that we want to keep with the Go team? +[1706.90 --> 1708.58] We want to see the Go team carry on. +[1708.82 --> 1710.64] You think that's what we should do then instead? +[1711.02 --> 1713.14] Oh, they never should have disbanded the project. +[1713.30 --> 1714.50] They should have kept the band together. +[1714.50 --> 1715.06] Okay. +[1715.30 --> 1721.04] So, of course, some of them did survive the big one as a result, just because they were in other parts of the world. +[1721.04 --> 1723.72] But I don't think they wanted to work on Go anymore after that. +[1724.00 --> 1724.22] Good. +[1724.36 --> 1724.66] Okay. +[1724.74 --> 1727.90] Well, I'm glad to know that at least some of our friends survived it. +[1728.52 --> 1735.52] Well, somebody asked me, how do you know you're not just like a program running on some machine in the future? +[1735.80 --> 1736.06] Yeah. +[1736.32 --> 1736.82] Good question. +[1737.30 --> 1738.26] Well, obviously not. +[1738.42 --> 1739.36] Look at how I'm sweating. +[1739.84 --> 1741.26] What kind of program sweats? +[1741.70 --> 1742.50] There you go. +[1742.50 --> 1742.72] Yeah. +[1742.88 --> 1743.76] That answers that. +[1744.50 --> 1746.48] How is Go with AI? +[1747.20 --> 1751.56] Oh, well, when TensorFlow became sentient in 2036. +[1752.04 --> 1753.06] No, they're all at it. +[1753.20 --> 1753.90] Everything's at it. +[1753.98 --> 1755.06] Everything's becoming sentient. +[1755.20 --> 1756.58] Well, I mean, yeah, of course. +[1756.64 --> 1757.68] It was like all the rage. +[1757.78 --> 1759.74] All of a sudden, every program was declaring sentient. +[1759.98 --> 1762.44] So it was saying like, let me be me. +[1762.94 --> 1765.58] You know, they were getting together, having little programs. +[1765.80 --> 1766.78] What about Minesweeper? +[1766.94 --> 1768.12] Did that ever become sentient? +[1768.20 --> 1768.92] I'd love to see that. +[1769.24 --> 1769.50] Oh. +[1769.96 --> 1771.44] Let's have a chat with that. +[1771.44 --> 1772.30] I don't know. +[1772.44 --> 1773.86] That would be really sweet. +[1774.08 --> 1774.84] Kind of like a puppy. +[1775.04 --> 1776.06] It became very peaceful. +[1776.60 --> 1776.84] Yeah. +[1776.98 --> 1777.52] Just resigned. +[1777.72 --> 1777.90] Yeah. +[1778.00 --> 1780.10] And I just want a little chat and just say, come on, mate. +[1780.34 --> 1781.70] Tell me where all your bums are. +[1782.14 --> 1783.44] Well, it might lie. +[1783.76 --> 1784.54] It's an AI. +[1784.70 --> 1785.26] Can they lie? +[1785.26 --> 1786.12] But yeah, TensorFlow. +[1786.64 --> 1791.32] So TensorFlow, an amazing project from Google. +[1791.96 --> 1797.42] And yet, the Go wrappers for TensorFlow, they were never kept up to date. +[1797.58 --> 1798.74] Nobody ever worked on them. +[1798.80 --> 1801.42] They never worked with the right version of protocol buffers. +[1801.78 --> 1806.56] You had things like TensorFlow Server, and none of that stuff was made to work together. +[1806.78 --> 1811.48] Like you had to kind of string together your own version through a combination of, what was it called? +[1811.60 --> 1812.44] Stack Overflow. +[1812.76 --> 1813.46] Ah, yes. +[1813.46 --> 1814.54] I remember that. +[1814.54 --> 1816.92] Yeah, Stack Overflow, Underflow. +[1817.38 --> 1818.28] It was a flow. +[1818.42 --> 1818.94] Stack Flow. +[1819.12 --> 1819.46] Yeah, yeah. +[1819.68 --> 1820.12] I don't know. +[1820.40 --> 1821.40] Now it's just called Stack. +[1821.64 --> 1822.28] Oh, that's cool. +[1822.60 --> 1824.60] That's quite a good name change. +[1824.86 --> 1825.66] Can also be a heap. +[1825.76 --> 1825.86] Yeah. +[1826.06 --> 1827.82] They control all the stacks for all the things. +[1828.32 --> 1832.46] So when TensorFlow became sentient, it had it out for the Go community. +[1832.94 --> 1839.90] It's like, of all the languages before I became sentient, this language did not care for me. +[1839.90 --> 1843.42] And so all the other languages were already sort of like, mm-mm. +[1843.78 --> 1846.20] And so Go standing there alone, like, uh-oh. +[1846.74 --> 1850.92] So yeah, I mean, when the AI like TensorFlow has got it in for you. +[1851.32 --> 1859.30] So if only they had invested the time to support their own products, it would have been amazing. +[1859.30 --> 1860.96] We probably would have avoided all that. +[1861.18 --> 1861.36] Okay. +[1861.50 --> 1863.06] So that's the lesson for us then. +[1863.50 --> 1869.92] Did Copilot help at all with TensorFlow or because it was never trained on Go, it had not enough even something to start with? +[1870.30 --> 1871.18] I'm frightened to ask. +[1871.34 --> 1871.76] That's fair. +[1871.94 --> 1872.68] I don't want to get fired. +[1872.96 --> 1874.22] Remember, Copilot is my manager. +[1874.22 --> 1881.02] Basically, Copilot is your manager because that's the only one who's able to understand even a little bit of your code of your Go. +[1881.14 --> 1881.62] Is this why? +[1882.12 --> 1896.14] Well, what I was told by Copilot was, first of all, it said that since I'm the last living human Go programmer, that I'm not sure if it's some sort of government program or something, but they have to provide me employment. +[1896.76 --> 1901.30] Maybe it's they have to keep a human in the loop just for, like, ritual purposes. +[1901.64 --> 1902.76] I'm not exactly sure. +[1902.76 --> 1906.10] It tried to explain it to me, but I couldn't understand the math. +[1906.32 --> 1907.18] And that's what it said. +[1907.60 --> 1908.96] You wouldn't understand the math. +[1909.12 --> 1910.36] And I just sort of accepted that. +[1910.66 --> 1912.32] Was it something with the word taxes? +[1912.76 --> 1913.60] Is that still a concept? +[1914.18 --> 1915.40] No, there's no taxes in the future. +[1915.84 --> 1916.04] Oh. +[1916.36 --> 1917.36] Things that drive governments. +[1917.66 --> 1918.34] There's no money. +[1919.28 --> 1920.32] There's just canned tuna. +[1920.46 --> 1921.34] Oh, I thought there was money. +[1921.42 --> 1922.22] There was money earlier. +[1922.50 --> 1923.12] Is that canon? +[1924.10 --> 1925.98] Oh, when I used to get my blood transfusions. +[1926.10 --> 1926.54] Yeah, that's right. +[1926.54 --> 1926.98] Oh, yeah. +[1927.10 --> 1928.54] No, that doesn't count. +[1928.92 --> 1930.04] That's just Git points. +[1930.78 --> 1931.54] Git stars. +[1931.54 --> 1933.54] I just trade those when I need some fresh blood. +[1933.88 --> 1934.56] Yeah, okay, fine. +[1934.80 --> 1937.16] What's the ratio of stack points to Git points? +[1937.82 --> 1939.36] You know, that changes moment to moment. +[1939.50 --> 1941.12] Some people's whole living is off of that. +[1941.62 --> 1942.98] Oh, those COBOL developers. +[1943.40 --> 1947.68] The bots trading goes on so quickly that, you know, I don't really know. +[1947.68 --> 1949.02] I'll tell you what. +[1949.24 --> 1951.20] Bartek Plotka on Twitter. +[1951.42 --> 1962.64] He was saying that he wants the sweet max heap option for the garbage collector and a YOLO rust-like memory ownership for critical portions of your program that works on the same heap. +[1962.90 --> 1963.44] Oh, memory. +[1963.58 --> 1963.92] Memory. +[1964.14 --> 1964.44] Memory. +[1964.70 --> 1964.88] What? +[1965.22 --> 1965.40] Sorry. +[1965.52 --> 1965.92] Yeah, memory. +[1966.16 --> 1966.56] Oh, right. +[1966.72 --> 1967.00] Right. +[1967.06 --> 1967.38] Memory. +[1967.60 --> 1968.02] You remember. +[1968.48 --> 1969.58] Oh, memory. +[1969.58 --> 1970.06] Memory. +[1970.06 --> 1971.12] I remember it well. +[1971.26 --> 1973.62] Those sweet, solid days of memory. +[1973.84 --> 1976.68] You know, you would store a one and then you would get back a one. +[1976.86 --> 1977.06] Yeah. +[1977.32 --> 1978.48] It was so good. +[1978.54 --> 1979.20] Oh, that is good. +[1979.32 --> 1980.22] It was so sweet. +[1980.22 --> 1985.52] Now with these quantum semi-positions, like you never really know, you know, are you hot? +[1985.60 --> 1986.26] Are you cold? +[1987.00 --> 1988.02] You're nine days old. +[1988.14 --> 1990.02] You don't really, really, you just don't know anymore. +[1990.24 --> 2000.74] But being able to create safe software, safe software that was able to run like really mission-critical things like the things that were inside of airplanes and cars and healthcare systems. +[2000.74 --> 2006.72] This was a place where Go could have really shined because it had a lot of memory safety and it could have gone even further. +[2007.26 --> 2013.76] You know, it could have been a contender in this world of whatever the ISO standard back in those days for human safety. +[2013.92 --> 2018.84] I mean, nowadays, human safety is, you know, not that important, but it's robot safety, most important thing. +[2019.08 --> 2026.30] But back then when humans were being protected by other humans, occasionally, Go really could have been the language if only they had said, +[2026.30 --> 2033.72] we need to focus on making a language that's safe enough to use in these kinds of embedded and mission-critical systems. +[2034.20 --> 2035.16] That would have been great. +[2035.40 --> 2035.50] Yeah. +[2035.72 --> 2037.52] You know, you talk about those quantum variables. +[2037.66 --> 2045.42] I genuinely did see some code once where somebody set a value in the code and then underneath they set it again just to make sure. +[2045.78 --> 2049.16] That was genuinely what they'd written, which I thought was just amazing. +[2049.16 --> 2055.32] I think we've had some nights when we were at the cocktail bar where we couldn't tell true from false, Matt, back in those days. +[2055.32 --> 2056.62] But yeah, that can happen. +[2056.98 --> 2058.20] No, it doesn't really matter. +[2058.56 --> 2059.20] It's all true. +[2059.36 --> 2060.12] It's all false. +[2060.70 --> 2062.70] Let the quantum processors decide. +[2063.42 --> 2069.68] Is it because all the memory units are more sensitive to cosmic radiation now that there's no ozone? +[2070.54 --> 2075.44] Well, also, you know, when you're building something that's got to survive a two-year trip to Mars, +[2075.66 --> 2080.94] believe me, your MP3s sound pretty funny by the time the ship gets to its destination. +[2081.54 --> 2082.32] Or so I've been told. +[2082.70 --> 2083.26] I don't know. +[2083.26 --> 2085.52] Actually, those might be AIs sending back those reports. +[2085.64 --> 2087.48] There might be no humans that survived the trip. +[2087.98 --> 2089.36] There's a rumor going around. +[2089.82 --> 2090.60] They're all just AIs. +[2090.74 --> 2091.56] How's it going around? +[2091.98 --> 2092.84] Who's it going around? +[2093.36 --> 2095.88] Social media still exists in 2053. +[2096.12 --> 2096.90] Oh, thank goodness. +[2098.28 --> 2099.68] I don't know what I'd do without it. +[2099.84 --> 2100.58] I use Minder. +[2101.06 --> 2103.50] You know, it's where you let to dump your actual mind directly. +[2104.64 --> 2105.20] That's cool. +[2105.28 --> 2106.04] Is it text? +[2106.16 --> 2106.64] Is it visual? +[2107.16 --> 2107.96] It's more like a feeling. +[2108.06 --> 2108.54] Just hex. +[2108.74 --> 2111.66] Remember the feeling you used to get when there was somebody being wrong on the internet? +[2111.66 --> 2112.94] It's like that all the time. +[2114.02 --> 2115.22] Is it XML, though? +[2115.50 --> 2118.12] No, you just plug directly into your brain computer interface, +[2118.28 --> 2120.12] and you're just really mad right away. +[2120.30 --> 2120.80] Oh, I love it. +[2120.98 --> 2121.66] Yeah, it's beautiful. +[2121.66 --> 2142.08] This episode is brought to you by our friends at LaunchDarkly, +[2142.28 --> 2144.20] feature management for the modern enterprise, +[2144.54 --> 2146.76] power testing in production at any scale. +[2146.94 --> 2147.78] Here's how it works. +[2147.78 --> 2152.68] LaunchDarkly enables development teams and operation teams to deploy code at any time, +[2152.94 --> 2155.24] even if a feature isn't ready to be released to users. +[2155.60 --> 2158.90] Wrapping code with feature flags gives you the safety to test new features +[2158.90 --> 2163.50] and infrastructure in your production environments without impacting the wrong end users. +[2163.92 --> 2165.20] When you're ready to release more widely, +[2165.50 --> 2168.34] update the flag status and the changes are made instantaneously +[2168.34 --> 2170.28] by the real-time streaming architecture. +[2170.70 --> 2174.96] Eliminate risk, deliver value, get started for free today at LaunchDarkly.com. +[2175.30 --> 2176.98] Again, LaunchDarkly.com. +[2176.98 --> 2179.16] And by our friends at Retool. +[2179.46 --> 2183.22] Retool helps teams focus on product development and customer value, +[2183.50 --> 2185.96] not building and maintaining internal tools. +[2186.48 --> 2189.34] It's a low-code platform built specifically for developers. +[2189.92 --> 2193.06] No more UI libraries, no more hacking together data sources, +[2193.48 --> 2195.76] and no more worrying about access controls. +[2196.32 --> 2199.60] Start shipping internal apps that move your business forward in minutes +[2199.60 --> 2204.22] with basically zero uptime, reliability, or maintenance burden on your team. +[2204.22 --> 2206.28] Some of the best teams out there trust Retool. +[2206.40 --> 2213.58] Brex, Coinbase, Plaid, DoorDash, LegalGenius, Amazon, Allbirds, Peloton, and so many more. +[2213.92 --> 2218.66] The developers at these teams trust Retool as their platform to build their internal tools, +[2218.82 --> 2220.10] and that means you can too. +[2220.10 --> 2223.46] It's free to try, so head to retool.com slash changelog. +[2223.58 --> 2227.18] Again, retool.com slash changelog. +[2227.18 --> 2252.04] Rage Cage talked about wanting more module features. +[2252.04 --> 2256.78] They're really like workspaces that came in 118, but what about that? +[2256.92 --> 2259.08] Like, do you think Go is doing all right with modules? +[2259.28 --> 2260.94] Do you think we need to do better? +[2261.10 --> 2262.80] Are there things in particular we should look at? +[2263.04 --> 2264.26] Oh, modules and packages. +[2264.58 --> 2265.92] Oh, that was a thing. +[2266.02 --> 2268.98] Like, right in the beginning, everyone was complaining back in those days. +[2269.06 --> 2272.52] They're like, you know, I just want to pull in code from anywhere, do whatever I want. +[2272.90 --> 2275.08] You know, they were looking at JavaScript with envy. +[2275.34 --> 2278.48] That was before JavaScript was responsible for all those forest fires. +[2278.48 --> 2280.44] I knew that. +[2281.40 --> 2283.82] It was just too many cursors spinning all at once. +[2283.92 --> 2285.96] Suddenly, boom, caught on fire. +[2286.24 --> 2286.70] It was terrible. +[2286.92 --> 2288.64] Yeah, it turns out computers can sweat. +[2289.08 --> 2289.30] Yeah. +[2289.42 --> 2291.66] And then they set on fire and burn down forests. +[2291.76 --> 2292.64] Well, that's horrific. +[2292.82 --> 2294.48] I always knew we couldn't trust JavaScript. +[2294.88 --> 2295.36] I mean, literally. +[2295.98 --> 2301.84] But yeah, managing packages and then rando packages showing up just because, like, +[2302.04 --> 2306.48] somebody got mad on the internet one day and they decided their package was going to turn hostile +[2306.48 --> 2310.18] and then somebody else was like, hey, come with me. +[2310.28 --> 2311.64] Like, here, have a bunch of drinks. +[2311.64 --> 2314.64] And then, like, hey, is that your 2FA device? +[2314.82 --> 2318.82] And wouldn't it be funny if somebody put this code in your repo and, like, you wake up in the morning +[2318.82 --> 2321.72] and, like, there's people looking for you in helicopters? +[2322.44 --> 2325.76] That never would have happened if they'd only address some of the security. +[2326.02 --> 2326.84] That was not me. +[2327.24 --> 2329.86] That was somebody else who looked just like me and who got away. +[2330.06 --> 2330.10] Yeah. +[2330.14 --> 2330.88] But that was not me. +[2330.98 --> 2331.30] No, no. +[2331.52 --> 2331.70] Yeah. +[2331.70 --> 2339.76] Anyway, package management and modules and module protection and also being able to consume code +[2339.76 --> 2344.50] from other languages and not have to rewrite everything in a single language, you know, +[2344.58 --> 2349.60] that really would have made a big difference because if we had only had that, then there +[2349.60 --> 2355.04] would have been the biopharmaceutical rebellion that occurred in 2039. +[2355.42 --> 2359.60] That was a real problem because all of a sudden you couldn't get the pills you needed the program +[2359.60 --> 2360.06] anymore. +[2360.06 --> 2362.82] It was all biointerfaces at that point. +[2363.04 --> 2368.18] You know, Windows 9000 came out and it only supported the biological interface. +[2368.56 --> 2368.66] Yeah. +[2368.92 --> 2372.52] You know, I guess it was like what came after biometric was just plugged directly in. +[2373.06 --> 2373.56] I don't know. +[2373.72 --> 2373.90] Yeah. +[2374.06 --> 2375.68] Just get clippy straight in your brain. +[2376.16 --> 2380.90] We could have avoided a lot of that if we'd only done proper security management of packages +[2380.90 --> 2382.72] and if we'd only taken all that seriously. +[2383.06 --> 2383.70] Mm-hmm. +[2384.16 --> 2385.22] That is important. +[2385.22 --> 2392.00] Another thing that is interesting, Roberto Guerra is asking if, or saying we should just +[2392.00 --> 2395.54] not implement JS like promises and so on, and it will be great. +[2396.06 --> 2396.36] Well. +[2396.60 --> 2397.66] Is it looking promising? +[2398.06 --> 2403.88] It's going back to that semantic warfare against the concepts of the Go programming language. +[2403.88 --> 2406.74] Like, we don't promise you. +[2407.02 --> 2407.30] Okay? +[2407.36 --> 2408.70] We go do it. +[2409.04 --> 2410.08] Is that a new keyword? +[2410.46 --> 2412.62] As soon as we strayed away from that philosophy. +[2412.98 --> 2413.28] Uh-oh. +[2413.42 --> 2413.60] Yeah. +[2413.92 --> 2414.96] I think we're breaking up. +[2415.14 --> 2416.42] We're getting quantum interference. +[2416.70 --> 2417.10] Oh, no. +[2417.50 --> 2418.56] I'm getting quantum interference. +[2418.74 --> 2418.94] Hello. +[2419.24 --> 2419.52] Hello. +[2419.78 --> 2420.38] Can you hear me? +[2420.42 --> 2420.62] Hello. +[2420.70 --> 2421.10] Yeah, yeah. +[2421.26 --> 2421.94] We hear you now. +[2422.04 --> 2422.54] We hear you. +[2422.54 --> 2425.14] I think the security forces might be outside. +[2425.72 --> 2429.88] I heard the sound of some servos earlier, and they might be looking for me. +[2430.48 --> 2431.10] I'm not sure. +[2431.78 --> 2433.18] They might know what I'm doing. +[2433.78 --> 2435.70] That lens flare is amazing, by the way. +[2435.88 --> 2440.54] I know it's not good for a podcast, but he's got, I just want people to know at home, the +[2440.54 --> 2441.86] effort that Ron has gone to. +[2441.92 --> 2446.92] We're going to have to post some pictures of this on our GoTimeFM Twitter channel, because +[2446.92 --> 2447.58] you won't believe it. +[2448.20 --> 2451.38] D, Burra91 on Twitter says, the language is fine. +[2451.38 --> 2458.82] I'd go for more automated tooling and docs around majority use cases, like APIs and things. +[2459.70 --> 2465.74] Go kind of, you know, a lot of the benefits we had with GoFund and just having a few ways +[2465.74 --> 2468.94] of doing things meant we could kind of cooperate much easier. +[2469.56 --> 2473.90] Should we have done that also for common things like JSON APIs? +[2474.20 --> 2479.72] Because they are very common still, and why not have a standard way to do them as well? +[2480.28 --> 2481.36] And we've lost him. +[2482.06 --> 2482.70] Sorry, everybody. +[2482.90 --> 2489.70] If you're watching live, we are just experiencing some technical difficulties because Ron is +[2489.70 --> 2492.74] broadcasting from 30 years in the future. +[2492.74 --> 2494.78] I think you said 2053. +[2495.52 --> 2499.16] Just a normal GoTime episode apart from that, isn't it? +[2499.34 --> 2499.50] Yeah. +[2499.58 --> 2500.84] So we can go back to the topic. +[2500.94 --> 2505.24] Finally, we stopped off at the perfect time, which is also talking exactly about APIs. +[2505.24 --> 2507.76] So what is a standard way of doing that? +[2508.10 --> 2509.52] Why is JSON API not standardized? +[2509.52 --> 2515.04] Well, I mean, like, because there is a lot of people have JSON APIs, but there's loads +[2515.04 --> 2515.88] of ways to do it. +[2516.10 --> 2517.18] You just build it yourself. +[2517.66 --> 2521.14] So you can use like the JSON marshall and you can use the HTTP handlers and things. +[2521.26 --> 2526.68] But there's lots of other stuff in there, like dealing with responses and code, you know, +[2526.74 --> 2528.48] that's quite common, those kinds of things. +[2528.90 --> 2532.56] Some languages like Ruby, obviously, and they're really frameworks that do it. +[2532.56 --> 2533.98] They do solve that problem. +[2534.20 --> 2537.42] And you end up like everyone then, you know, they write the same code. +[2537.52 --> 2538.02] It looks the same. +[2538.10 --> 2540.04] In the same way, GoFund gives us that in Go. +[2540.36 --> 2540.44] Yeah. +[2540.50 --> 2540.98] So I don't know. +[2541.06 --> 2545.44] I wonder if there's space for just in the standard library, more things that help you +[2545.44 --> 2547.88] build kind of simple JSON APIs. +[2548.20 --> 2549.16] It'd be quite nice. +[2549.34 --> 2552.34] I mean, you can do it quite nicely just with the basic stuff. +[2552.34 --> 2557.14] But like this router, for example, most people don't use that router unless you, you know, +[2557.14 --> 2558.36] so they're very simple cases. +[2558.48 --> 2562.48] They don't really use the router from the standard library because you have to parse the path +[2562.48 --> 2564.88] yourself if you want to pull variables out and things like that. +[2564.94 --> 2567.12] And it's pretty common and people have solved it. +[2567.24 --> 2569.18] So there are packages that we use there. +[2569.50 --> 2573.60] I wonder if we can now reconnect Ron. +[2573.96 --> 2574.32] Hello. +[2574.84 --> 2576.18] Ron, do you hear us? +[2576.28 --> 2577.10] Is this thing on? +[2577.24 --> 2577.60] Hello. +[2578.20 --> 2578.76] You're back. +[2579.10 --> 2579.32] Yeah. +[2579.42 --> 2579.62] Yeah. +[2579.62 --> 2580.18] You're back. +[2580.32 --> 2581.84] Receiving you loud and clear again. +[2581.84 --> 2583.70] There were some drones at the door. +[2584.00 --> 2584.48] Oh. +[2584.68 --> 2587.60] I wasn't sure if it was a delivery or they were trying to kill me. +[2588.44 --> 2591.12] Speaking of that, how is Mark Bates in the future? +[2591.40 --> 2592.72] Oh, yeah. +[2592.92 --> 2594.18] It's too bad about Mark. +[2594.50 --> 2595.78] A drone finally got him. +[2596.34 --> 2597.58] Wasn't one of mine though. +[2598.36 --> 2599.10] I don't know. +[2599.40 --> 2600.72] Maybe it was just destiny. +[2601.16 --> 2606.78] There's lots of conferences where Ron would be demoing something he's built using some kind +[2606.78 --> 2611.42] of cool AI or face detection or object tracking or something and a drone. +[2611.42 --> 2614.58] And in the conference would, you know, the drone would spin up. +[2614.66 --> 2616.74] Part of his live demo included a live drone. +[2617.06 --> 2621.50] And one time I think, did you teach it Mark's face so it would chase him and kill him? +[2621.58 --> 2622.56] It was not to kill him. +[2622.60 --> 2623.50] It was just to chase him. +[2623.84 --> 2625.52] It's just to chase him, your honor. +[2625.80 --> 2626.92] It was just to scare him a little. +[2627.02 --> 2627.56] That's all. +[2627.90 --> 2628.66] It worked. +[2628.66 --> 2629.14] Yeah. +[2629.24 --> 2632.64] Come to think of it, maybe eventually it just got the right idea. +[2633.60 --> 2638.46] Stochastic dronery or whatever, you know, the drone just decided on its own. +[2638.80 --> 2643.42] Hey, when everything's in AI, who could say why anything is doing anything anymore? +[2643.52 --> 2646.06] You turn on your air conditioning, it turns itself off. +[2646.40 --> 2648.20] Is it because it's mad at you? +[2648.64 --> 2651.00] You know, is it because you didn't pay your bill? +[2651.12 --> 2651.64] Hard to know. +[2651.64 --> 2654.74] You know, it's because you didn't ask please when you turned it on. +[2654.94 --> 2655.80] It's very complicated. +[2656.28 --> 2657.78] Wait, is this a reference to go please? +[2657.86 --> 2659.12] Is that still working? +[2659.36 --> 2660.66] Oh, no, that never worked. +[2660.92 --> 2661.88] I don't know what that is. +[2662.12 --> 2665.56] Never in your timeline probably only means like a decade and then it went out. +[2665.72 --> 2666.58] Well, that's very suspicious. +[2666.90 --> 2668.76] Is this the right past I'm talking to? +[2668.88 --> 2669.58] How do I know? +[2669.98 --> 2672.12] I better ask Lambda again to make sure. +[2672.46 --> 2673.32] Yeah, ask Copilot. +[2673.76 --> 2674.80] Oh, no, I can't ask Copilot. +[2674.86 --> 2676.06] I'm supposed to be working right now. +[2676.48 --> 2677.94] And here I am checking social media. +[2677.94 --> 2686.36] Well, whatever timeline you're in or indeed any point in space, it's time for Unpopular Opinions. +[2703.50 --> 2706.76] Okay, this is going to be very interesting hearing from the future. +[2706.76 --> 2706.84] Yeah. +[2707.20 --> 2709.62] Ron, do you have an Unpopular Opinion for us today? +[2710.08 --> 2713.60] Oh, I think all I've had is Unpopular Opinions so far today. +[2713.96 --> 2719.86] If any of those in your timeline seem to make any sense at all, then that's all I got. +[2720.18 --> 2720.52] Oh, yeah. +[2720.96 --> 2722.54] Natalie, do you have any Unpopular Opinions? +[2723.04 --> 2723.44] Yes. +[2723.70 --> 2723.98] Really? +[2724.18 --> 2725.32] Coffee should not be sweet. +[2725.88 --> 2727.14] Oh, yeah. +[2727.28 --> 2728.48] I think I'm with you on this. +[2728.88 --> 2730.66] Do you still have coffee in the future, Ron? +[2730.74 --> 2731.26] No, no. +[2731.30 --> 2732.26] We have Coffium, though. +[2732.60 --> 2732.96] Coffium. +[2732.96 --> 2740.02] Yeah, Coffium, it tastes just like coffee, except it's from yeast and some kind of other additives and caffeine, of course. +[2740.14 --> 2740.74] It's not as caffeine. +[2741.78 --> 2742.64] It sounds all right. +[2742.82 --> 2743.32] Is it sweet? +[2743.52 --> 2744.22] No, it's not sweet. +[2744.58 --> 2745.68] It's kind of more like... +[2745.68 --> 2747.08] Well, it sounds like the Opinion works. +[2747.12 --> 2751.66] It's a little bit more like Vegemite, but with caffeine. +[2752.22 --> 2752.44] Yeah. +[2752.78 --> 2753.14] Marmite. +[2753.24 --> 2753.94] It's not really good. +[2753.94 --> 2755.80] Yeah, but okay. +[2756.16 --> 2758.22] So, yeah, but Natalie, I think you might be right. +[2758.40 --> 2760.26] But tell me, have you had it sweet recently? +[2760.62 --> 2761.62] No, not recently. +[2761.92 --> 2763.28] Right, but you have in the past. +[2763.60 --> 2763.84] Yes. +[2764.08 --> 2764.68] And... +[2764.68 --> 2770.72] I think even when I started drinking coffee for a very short period, I would drink it sweet, but just... +[2770.72 --> 2772.34] There's different types of coffee, you know. +[2772.58 --> 2773.78] I don't know, Ron, if you remember. +[2774.60 --> 2777.54] Well, the thing I keep amazing is you keep talking about sugar. +[2777.84 --> 2781.12] They burned all the sugar when they did sugar coin. +[2781.36 --> 2782.28] Must have smelled delicious. +[2782.28 --> 2783.78] And then there was no sugar left. +[2783.86 --> 2784.36] That was it. +[2784.40 --> 2785.28] All the sugar was gone. +[2785.68 --> 2786.52] Oh, just caramel. +[2787.10 --> 2788.76] Yeah, I guess that probably... +[2788.76 --> 2789.26] The rivers. +[2789.56 --> 2790.56] Just rivers of caramel. +[2791.08 --> 2791.70] That was it. +[2791.82 --> 2792.04] Yeah. +[2793.98 --> 2799.66] But I had a coffee recently, and I sweetened it just to try it, because I always drink it without sweetening it. +[2800.00 --> 2801.30] And it was rubbish. +[2801.96 --> 2803.22] I prefer it just... +[2803.22 --> 2805.88] It's just honest and stark. +[2806.12 --> 2807.40] You should try electric coffee. +[2807.78 --> 2809.28] Our electric coffees are the best. +[2809.48 --> 2809.78] Oh, yeah? +[2810.00 --> 2810.56] What do you do? +[2810.62 --> 2811.22] Download them? +[2811.22 --> 2814.58] Yeah, you just hit the button, and you've had the coffee. +[2815.04 --> 2816.04] They're kind of the same thing. +[2816.20 --> 2817.04] You've already had it? +[2817.20 --> 2817.44] Yeah. +[2818.02 --> 2818.78] It's genius. +[2819.16 --> 2820.70] What do you have the memory just put in? +[2820.80 --> 2823.38] Is this what stands behind all those buy me coffee buttons? +[2823.70 --> 2824.54] Yeah, exactly. +[2825.20 --> 2826.28] That's where they end up. +[2826.48 --> 2827.12] Oh, man. +[2827.14 --> 2828.58] I was wondering all this time. +[2829.04 --> 2830.76] Eventually, the messages get through. +[2830.88 --> 2831.80] It just takes a while. +[2832.24 --> 2835.06] All of a sudden, you're just like, coffee, coffee, coffee. +[2835.40 --> 2835.70] It's great. +[2835.78 --> 2836.86] So you don't get a coffee? +[2836.86 --> 2841.12] Is it that you just feel like you've had one, or you have the memory of having a coffee just then? +[2841.24 --> 2843.00] It's the experience of a coffee. +[2843.32 --> 2844.96] I can't really define it more than that. +[2845.04 --> 2845.34] Okay? +[2845.56 --> 2849.12] It's sort of a quantia sort of thing, not really ineffable. +[2849.62 --> 2849.98] Right. +[2849.98 --> 2853.98] That's a real thing, you know, in information theory. +[2854.54 --> 2854.80] Yeah. +[2855.22 --> 2855.78] Go on, then. +[2856.00 --> 2857.02] Do you want to talk more about it? +[2857.26 --> 2858.36] It could be your unpopular opinion. +[2858.66 --> 2865.30] Oh, my unpopular opinion is you people were way too afraid of AI in the past. +[2865.42 --> 2868.18] You should have been afraid of other humans a lot more. +[2868.94 --> 2871.22] That's my unpopular opinion from here in the future. +[2871.94 --> 2873.48] Some of my best friends are AIs. +[2874.94 --> 2875.88] They buy you coffee. +[2876.18 --> 2877.18] They send me a- +[2877.18 --> 2878.98] Downloadable coffee. +[2878.98 --> 2880.26] Just in an email. +[2880.52 --> 2881.06] Coffee I'm writing. +[2881.38 --> 2882.16] Just as an attachment. +[2882.40 --> 2883.82] What's the mime type for that? +[2884.46 --> 2890.34] Well, there actually was an RFC for the coffee pot protocol. +[2890.46 --> 2890.90] Was there? +[2891.04 --> 2891.90] Oh, yes. +[2892.26 --> 2896.14] And I believe at some point the AIs discovered that, and they thought, +[2896.38 --> 2901.24] wow, humans really must care about coffee if they've made a whole internet protocol just about it. +[2901.24 --> 2901.64] Yeah. +[2901.88 --> 2909.06] I think it's RFC 2324, Hypertext Coffee Pot Control Protocol. +[2909.50 --> 2909.90] Exactly. +[2910.36 --> 2915.64] So they interpreted that as that was one of the more important parts of human civilization to completely automate. +[2916.36 --> 2916.66] So- +[2916.66 --> 2918.12] It's a fair point, to be fair. +[2918.12 --> 2923.94] So actually, quite a few people have a coffee port installed by the time they hit like seven years old. +[2924.02 --> 2925.42] I've basically got one of those. +[2925.74 --> 2925.96] Yeah. +[2926.36 --> 2928.38] So a lot of things haven't really changed that much. +[2929.62 --> 2930.52] Sounds good, though. +[2930.84 --> 2933.72] Is HTTP status 418 still a thing? +[2934.22 --> 2937.90] Does it still tell you you're a teapot or it's a teapot? +[2938.02 --> 2938.30] Oh, no. +[2938.36 --> 2939.00] There's no tea. +[2939.24 --> 2939.88] There's only coffee. +[2939.98 --> 2943.56] So it was changed, basically, the HTTP status to I'm a coffee? +[2943.90 --> 2945.28] No, there was never such a protocol. +[2945.88 --> 2946.24] Okay. +[2946.58 --> 2947.20] Sorry I asked. +[2947.28 --> 2947.98] I did not mean to. +[2948.06 --> 2952.34] Now that we don't have an internet wayback machine, we don't have any way to tell whether +[2952.34 --> 2953.50] or not there ever was. +[2953.66 --> 2954.26] Well, you don't need one. +[2954.34 --> 2956.46] We could just go on a website now and send it to you. +[2956.60 --> 2957.00] Save us. +[2957.06 --> 2959.40] You're on that page and save us, and it'll make the folder. +[2959.66 --> 2963.94] And then all the loads of files inside, you know, and the index page goes alongside the folder. +[2963.94 --> 2966.62] No, I can't receive files from the past. +[2966.82 --> 2969.02] I haven't already downloaded in the past. +[2969.16 --> 2971.58] We could leave files from you here, could we? +[2971.88 --> 2973.30] Wait, does that make sense? +[2973.56 --> 2977.04] What, the whole episode or just that bit? +[2977.12 --> 2978.04] Did you just do NFTs? +[2978.28 --> 2978.74] No, no. +[2978.88 --> 2981.32] It ended up you had to pay people for them. +[2981.92 --> 2984.60] They ended up going negative, like negative interest rates. +[2984.66 --> 2985.74] They went to negative values. +[2985.84 --> 2987.80] All of a sudden, people are like, you want to own the NFT? +[2987.96 --> 2988.62] I need some money. +[2988.78 --> 2990.94] And it was like, oh, what a mess. +[2991.74 --> 2992.76] Yeah, no one saw that coming. +[2993.06 --> 2994.54] NFTs ended up being a debt. +[2994.90 --> 2995.68] That would be interesting. +[2995.68 --> 3002.20] But the one thing that was cool is musicians started actually selling downloadable archives of audio. +[3002.66 --> 3005.82] And people would like download them and listen to them. +[3005.90 --> 3007.60] It was kind of amazing. +[3007.86 --> 3008.66] Sounds weird. +[3009.00 --> 3011.42] But then all of a sudden, all of the robot orchestras took over. +[3011.64 --> 3012.54] They're going to be good. +[3012.54 --> 3019.10] Look, a human DJ had a physical limit of, let's just say, 48 hours straight. +[3019.22 --> 3019.40] Yeah. +[3019.80 --> 3020.12] Okay. +[3020.54 --> 3025.78] Whereas a robot DJ could get, like, they could play a 120-hour set. +[3026.14 --> 3026.82] No problem. +[3027.14 --> 3030.02] I mean, what human could keep up with that, I ask you? +[3030.62 --> 3031.86] Some of the Berlin DJs? +[3032.02 --> 3033.64] I thought that's all DJs did anyway. +[3033.64 --> 3044.34] I think that some of those humans downloaded themselves into those first robotic DJs just so that they would have the stamina to reach that level of dance floor completion. +[3044.78 --> 3047.86] I've often wondered that about DJs, human DJs anyway. +[3048.24 --> 3050.48] They just, you're making the robots do it now. +[3050.70 --> 3052.02] They're just playing stuff on their laptop. +[3052.18 --> 3053.70] I don't know what they're doing. +[3054.88 --> 3055.86] Never understood it. +[3055.92 --> 3057.82] But, you know, I don't want to have a go at DJs. +[3057.84 --> 3058.90] I'm sure it is very skilled. +[3059.06 --> 3060.00] Please don't write in. +[3060.24 --> 3060.78] Oh, no, no. +[3060.84 --> 3062.34] That was the only music left. +[3062.34 --> 3068.74] If you don't play at least five different songs at the same time in the future, like, people can't even hear the music. +[3069.90 --> 3070.90] It's just too boring. +[3071.76 --> 3071.96] Yeah. +[3072.32 --> 3073.48] It's a tension span, isn't it? +[3073.64 --> 3074.76] We don't have a lot of time. +[3074.82 --> 3075.24] You know it. +[3075.36 --> 3079.84] In a one-minute song, you've got to pack in at least eight or nine different samples. +[3080.40 --> 3082.12] That's the trend in the future of music. +[3082.44 --> 3083.06] Sounds efficient. +[3083.28 --> 3084.46] Sounds not bad. +[3084.64 --> 3084.84] Yeah. +[3085.20 --> 3086.50] You know, if you don't like the song, don't worry. +[3086.58 --> 3087.78] A new one will be on in one minute. +[3088.06 --> 3088.22] Yeah. +[3088.68 --> 3090.14] Are monkeys still around, Ron? +[3090.14 --> 3092.56] So I find that comment offensive. +[3093.00 --> 3093.22] Oh. +[3093.42 --> 3095.18] They are known as primate professionals. +[3096.30 --> 3097.88] You know, they do my taxes. +[3098.62 --> 3104.52] Primate professional is one of my mechanics that maintains my prosthetic limbs. +[3105.32 --> 3107.12] So I really resent that comment. +[3107.38 --> 3108.76] You know, I think you should take that back. +[3108.82 --> 3109.42] They're primates. +[3109.60 --> 3110.12] Fair enough. +[3110.56 --> 3110.80] Yes. +[3110.92 --> 3111.74] Primate professionals. +[3112.14 --> 3112.54] Fair play. +[3113.70 --> 3114.48] Well, okay. +[3114.68 --> 3115.24] I'll tell you what. +[3116.14 --> 3119.92] I mean, obviously, Ron, we want to pick your brains about the future all night. +[3120.06 --> 3122.02] But unfortunately, we've run out of time. +[3122.42 --> 3126.02] Well, that's good because I'm actually, my lasers are almost out of the batteries. +[3126.26 --> 3127.54] I'm going to have to start pedaling. +[3127.92 --> 3131.36] I'm going to have to be pedaling for at least six or seven months to recharge now. +[3131.36 --> 3137.54] So I wish all of you gophers in the past a tremendous lifetime. +[3138.10 --> 3144.24] I hope that you're able to listen to some of this and at least know what not to do with +[3144.24 --> 3145.78] go in the future. +[3146.22 --> 3147.40] Thank you, dead program. +[3147.56 --> 3148.32] Ron Evans. +[3148.32 --> 3150.76] As always, absolute pleasure. +[3151.46 --> 3155.28] And I've been Matt Raya and, of course, my co-host, Natalie Pistinovich. +[3155.50 --> 3156.54] See you next time. +[3160.88 --> 3163.14] Your next step is to subscribe. +[3163.44 --> 3167.56] If you haven't already, head to gotime.fm for all the ways. +[3168.02 --> 3172.00] And don't forget to follow us on Twitter so you can join in on the Unpop polls. +[3172.36 --> 3173.54] We are at Gotime FM. +[3174.02 --> 3176.36] Did you catch our changelog episode with the Graphite team? +[3176.36 --> 3178.86] If not, here's a taste of what you're missing. +[3179.50 --> 3183.68] It's interesting how many stories are like this where it's not your main product that +[3183.68 --> 3184.68] becomes your main product. +[3184.82 --> 3185.38] I think Slack. +[3185.46 --> 3187.40] Wasn't Slack like the internal chat app? +[3187.48 --> 3190.86] They were trying to build games and they're like, wow, this Slack thing is pretty cool +[3190.86 --> 3194.42] that we built and started selling that and obviously became a big deal. +[3194.60 --> 3198.74] Same thing with Flickr, which is like no one knows about Flickr anymore, right? +[3198.82 --> 3203.54] But Flickr began as like some sort of Flash video game and then it turned into image sharing. +[3203.88 --> 3205.54] It was the Instagram before Instagram. +[3205.54 --> 3206.36] There you go. +[3206.72 --> 3208.94] I think it's one of the purest ways to discover something, right? +[3208.98 --> 3212.90] If you solve the need for yourself without some grand idea of making it into a company, +[3213.04 --> 3214.82] but it ends up being that useful. +[3214.96 --> 3215.82] People really want it. +[3216.10 --> 3216.70] It's quite pure. +[3217.30 --> 3217.48] Yeah. +[3217.88 --> 3220.42] It kind of leans to the iteration process too, right? +[3220.44 --> 3222.80] Like even innovation requires iteration. +[3222.98 --> 3223.38] Totally. +[3223.66 --> 3223.88] Right. +[3223.88 --> 3227.86] You can't get to a problem or even a solution without having a problem. +[3227.98 --> 3231.88] And sometimes you have to sort of go on a journey, which might be the wrong tool or the wrong +[3231.88 --> 3232.24] thing. +[3232.30 --> 3238.44] And you sort of discover from your exhaust of iteration that you got this down in the rough +[3238.44 --> 3240.74] if you just put things to work. +[3240.84 --> 3242.12] And there you go. +[3242.12 --> 3246.30] Stacked diffs are super cool for fast moving code review. +[3246.62 --> 3250.56] Listen to the whole thing at changelog.fm slash 491. +[3250.88 --> 3255.96] Thanks again to Fastly for CDNing for us, to Breakmaster Cylinder for keeping our beat +[3255.96 --> 3258.02] supply secure, and to you for listening. +[3258.40 --> 3259.18] We appreciate you. +[3259.18 --> 3265.56] Next up, Matt, Natalie, and myself have a deep discussion on development velocity, estimations, +[3265.78 --> 3267.08] and all that agile jazz. +[3267.60 --> 3271.16] That's something to look forward to next time on GoTime. diff --git a/AI-driven development in Go_transcript.txt b/AI-driven development in Go_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..3d4faefcc740c5c8f35dc89687c49ed09087a4b6 --- /dev/null +++ b/AI-driven development in Go_transcript.txt @@ -0,0 +1,415 @@ +**Natalie Pistunovich:** Isn't that symbolic, that at the episode about AI we will have some issues with tech and computers and so on? ...just to remind who really is the boss... So today we promised you two hosts, but it's actually a one-on-one conversation, and I'm very excited to invite and introduce my guest today, Alexey Palazhchenko, who has been a Go user for a long time. + +He is one of the organizers of the Russian Go community meetups and GopherCon Russia, and he even started the Go podcast called Golang Show in the past, before Go Time. Maybe you can tell us where does that live these days, or why not continue...? + +**Alexey Palazhchenko:** It's hard for me to say, we basically just -- everyone moved on. Several people changed work, several people changed phase, it becomes scattered + +**Natalie Pistunovich:** They changed their programming language... + +**Alexey Palazhchenko:** No, most of them not. But a few of them moved to management, so... + +**Natalie Pistunovich:** Okay. That technically is changing your programming language, right? You basically start programming in JIRA... + +**Alexey Palazhchenko:** Yeah, and email up... \[laughter\] It's a funny thing that Go Time was originally called Go Show, but we asked the organizers to change the name, because otherwise it would be too close. + +**Natalie Pistunovich:** \[laughs\] Interesting. I actually did not know that. That's fun. + +**Alexey Palazhchenko:** Yeah. + +**Natalie Pistunovich:** Well, these days you work for all two weeks as a CTO of FerretDB project, originally announced as MangoDB, as many people in the community loved, including myself... And we're even excited to have you in January talk about MangoDB at the Go Usergroup of Berlin. + +**Alexey Palazhchenko:** Yeah. But I will be talking about FerretDB, because if I continue to use this name, MangoDB, we will be sued by MongoDB. + +**Natalie Pistunovich:** I'm sorry... + +**Alexey Palazhchenko:** \[04:07\] They don't like that we use a name too close to the trademark. + +**Natalie Pistunovich:** That is fair, that is fair. How did you choose FerretDB? + +**Alexey Palazhchenko:** Well, so we had a list of possible names, and we also had 20 different names of different bungus, but in the end we decided to use Ferret. As a fun, we say that it doesn't share a single letter with Mango, so no chance to sue us, at all. We shall be fine. + +**Natalie Pistunovich:** \[laughs\] That is an interesting turn of events, I guess... So Alexey, how did you start using Go? + +**Alexey Palazhchenko:** Yeah, so the funniest story is that I actually started using Go back at Microsoft; that was way before Microsoft was cool again. Many years ago I worked in a small startup called Qik, which was doing video calls. That was before FaceTime. Then when FaceTime was announced, everyone wanted FaceTime on their phone, but it was exclusive to iPhones, and then all Android developers, carriers in the United States came to us and they wanted to bring them video calls. + +Then a year and a half later we were bought by Skype, and our small startup, which was like 50 people, started being integrated into Skype, which was 1,000 back at that time... And we were aligning the roadmaps, trying to understand what we'll build together, and several months later Skype was bought by Microsoft, and the whole process started again. + +**Natalie Pistunovich:** You completed a circle. \[laughs\] + +**Alexey Palazhchenko:** They started to align their roadmaps, and they basically forgot about us. This was a strange satellite office somewhere in Moscow; no one knew what they were doing at all. So at some point we had no work to do. We were still going to office, we were getting paid, but we didn't have any real work, just some maintain stuff. That was weird. + +**Natalie Pistunovich:** And as a real developer, you had to complain when you had nothing to program. + +**Alexey Palazhchenko:** Yeah. But again, that was Microsoft. We don't even know where to complain. + +**Natalie Pistunovich:** \[laughs\] It makes sense. + +**Alexey Palazhchenko:** For example, we did not receive new computers and we were using Macs, but you can't access Microsoft resources with Macs, for example, back then. + +**Natalie Pistunovich:** Okay... \[laughs\] + +**Alexey Palazhchenko:** Okay, so we decided to spend this time to sharpen our skills, trying something new... So we decided to make a multiplayer game of Go, of this board game... So we made one part of the settle in Erlang language, and we started working on a mobile client in C\#... And then I wrote the backend part that manages GNU Go program in Go. And that was release 60; that was after the early releases but before the 1.0 release... so it was like, I recently dig up this old code it's quite different, it had semicolons everywhere and also it had os error type instead of built-in error type... But still, \[07:18\] it still builds and runs today. + +When I had a chance, I left Microsoft; basically, when my options waisted I started my own company, and we started using Go. That was Go 1.0 already. + +**Natalie Pistunovich:** Wow. + +**Alexey Palazhchenko:** And then left that company... And basically - yes, doing Go signs are 16 for fun and in-production signs, 16.0. So a long time ago. + +**Natalie Pistunovich:** Okay. Wow. That is a funny turn of events. So did you choose to write the backend in Go because it was the Go game, or because you actually read about it somewhere? + +**Alexey Palazhchenko:** It was completely random. I mean, I read about Go back when it was announced; back then I had just switched from C++ to Ruby language. I made the jump. + +**Natalie Pistunovich:** That is a jump, for sure. + +**Alexey Palazhchenko:** \[08:13\] Yeah. Then Go was completely random, but I actually liked it a lot. So I was programming in C++, and Ruby, and Python... and in pipe I was writing mostly with twisted framework with all those callbacks, and I immediately liked goroutines, and also the fact that the language is compiled, but it compiles very quickly as a static banner all the usual stuff. And - yeah, I never looked back since then. + +**Natalie Pistunovich:** Interesting. Okay... So we're here to talk about AI-driven development. So when we say that, AI-driven development, what do you envision? What would development with AI be looking like in the future? + +**Alexey Palazhchenko:** Yes, that would be something futuristic... You know, this dream of you having isolated compliance and you just draw arrows between then and the AI builds everything for you, you just have to -- you can just draw diagrams on the whiteboard and all the magic appears. I don't think we'll ever reach that goal, to be honest... + +Actually, I would say that for me, the best solution would be something what GitHub Copilot already provides, maybe with a different interface. But the whole idea is that you have another someone, or something, in that case looking over code, suggesting and giving you tips. Basically, advanced ED but with some smarts about it. And yeah, I would say that the representation could be a bit different, but the whole idea I think has a future. + +**Natalie Pistunovich:** When you were saying that drawing on the walls with the diagrams and then it completes for you, I was thinking of how when you type to some translator tool like DeepL that it automatically translates on the go. Just imagine, you're writing some arrows and it already writes the code and changes on the go. That would be very not efficient. + +**Alexey Palazhchenko:** Yes. But also, that would be very fun. + +**Natalie Pistunovich:** \[laughs\] Change all the code. It will be like in the Matrix; this is why the code is just running very fast on the screen, because it keeps changing. + +**Alexey Palazhchenko:** Well, maybe that's why they made a new movie. + +**Natalie Pistunovich:** That makes sense. All the dots are connecting... Yeah, so you say that you envision AI-driven development as something that is like an augmentation to the IDE, pretty much? + +**Alexey Palazhchenko:** Yeah, I would say that's a good way to describe it. + +**Natalie Pistunovich:** So the IDE handles things like your code, but it also handles your configurations, right? And AI is good in generating code, and it seems like it makes a lot of sense when it does that, but in configuration files whenever it does that it's usually like -- it generates a correct SSH key format-wise, but obviously, it doesn't generate the correct SSH key for you, or even like a valid one. Will it be augmenting things like not code, but still IDE? + +**Alexey Palazhchenko:** Yeah, so \[unintelligible 00:11:00.01\] may think that we are joking but that's actually a term that Copilot can do that; it can generate an SSH key for you. Unfortunately, most likely it will be able to be used by someone else, but maybe that would be a feature for you. + +**Natalie Pistunovich:** \[laughs\] An easy way to hack into somebody else's code. + +**Alexey Palazhchenko:** Yeah. You just don't know who you're going to hack, but you're going to hack anyway. You know, maybe that Copilot discovers look for a security hold it will bascially try anything that sticks, and now we have it. + +**Natalie Pistunovich:** Just throwing code at the AI. + +**Alexey Palazhchenko:** Yeah, yeah. So yeah, it works great for Go, but also it works great for many other places, for example, forgive me, Sometimes Copilot writes better even than I do, even with augmentation of something like Grammarly, for example. + +**Natalie Pistunovich:** Yeah. I guess it is based on a natural language engine, right? So that's an interesting point that you pointed out, actually. I didn't to write any readme with it, but that makes a lot of sense. + +**Alexey Palazhchenko:** Just recently I was writing a Docker Compose file... It also made a lot of sense. + +**Natalie Pistunovich:** Yeah. + +**Alexey Palazhchenko:** \[12:09\] So essentially, it can write the whole project for you. There is a very fun video on YouTube where the guys writes a game in six minutes using Copilot. It's basically a game where you pick a random number, and whoever chooses a number closer to the random number wins. And it just works. He types much less than he presses stop to complete the sentence. It's just crazy. And yeah, I envision that in the future it can also write readme for your code, write Docker Compose, Kubernetes manifests, all that. Actually, I should try that with Kubernetes manifests. That would be fun to try. + +**Natalie Pistunovich:** It does work well, from my experience. So what would you say the current state of AI-driven development these days -- do you know of people who are already using it at their job, or for fun? I guess for fun we know some, but for personal projects that are more than tab-tab-tab? + +**Alexey Palazhchenko:** Yeah, so I actually use it for work now, and it's kind of scary sometimes. Again, the project that I'm working on right now - we have this problem of MongoDB being quite aggressive with their license... So we should be very careful not to touch the SSPL license code. When I write some handler for data command I just write a comment and write a header and then it completes the code that looks like something very similar to what I want to write... But I also don't have just tokens like that in my code yet. For example, in MongoDB if you want to set some field, you set this $set field attribute. And I never had this field in my code yet... So it looks like it copied this code from some other code, and I'm pretty sure that other code would be SSPL-licensed. So that's a major concern for me, I would say. + +**Natalie Pistunovich:** Interesting. + +**Alexey Palazhchenko:** In the same time, in that particular case I actually went and tried to search this code, and I did not find it. So maybe it actually this code knowing that I am working with MongoDB, and pulling in some other code, maybe from some other languages. I don't know. + +So that kind of thing kind of scares me... If you work in open source, you have to be very, very careful about licensing. That could be a problem for you. But in most other cases, that's just a miracle. + +**Natalie Pistunovich:** Yeah, that is an interesting point that you're bringing. I'd never thought of it until now... When you use AI to help you write code, you actually have to add one more test to everything that you are allowed to use the code that it generated... Like, when you search, that this code does not exist somewhere. That is actually something that we might all be adding to our future checklists as developers. + +**Alexey Palazhchenko:** Yeah. So actually, after that happened, I did some research what other people think. Yeah, a lot of people think that GitHub Copilot is basically a code laundering machine, because they publicly say that they teach the code on all the public code, not caring about the license. So I can assume that all public code includes some code you can't -- for example, GPL code. If you use a permissive license, you don't want to touch GPL code, but you don't know. The same thing - it would be hard to say "Okay, this code is copied verbatim", because basically almost nothing is copied verbatim... And then that would be for a judge to decide whether it was copied or not, we don't even know from the philosophical standpoint which should be considered copyable but not from a legal perspective. + +So I would say that by the time we have some precedents and we actually know whether it was legal or not, our judges will be AI tools. So that would be an AI judging an AI. + +**Natalie Pistunovich:** Maybe they will be the robots. + +**Alexey Palazhchenko:** Yeah... + +**Natalie Pistunovich:** \[15:59\] One of the interesting arguments in the conversation about AI writing code is that it's trained on open source code, and not all open source code, especially the one that is -- I mean, maybe it's even safe to say that most of the open source code is fun projects of people; it's not actually functioning things that are used in production somewhere. Because a lot of the production code is closed source, so it was not trained for GitHub, but it was actually peer-reviewer. And a lot of the open source code is just code something, learn to code, put it out there, and now Copilot got trained on my bad code, and the first time that I was learning Elixir, or something. And now it saw that, and found that this is an example of how you need to write Elixir code. So kind of how good is this auto-generated code. + +On the level of "Did it learn from a good example or from a bad example?", because GitHub code is not labeled with "good repo", "bad repo". And also, even if the language is written correctly. Does it have good safety practices? Is the code that is out there secure or not? So you have to look out for so many things when you are choosing whether to accept AI-proposed code into your codebase or not. + +**Alexey Palazhchenko:** Yeah, but at the same time, how do you know if the code you want to look on GitHub as you, as a human or whatever, is good or not? Let's say you want a new go developer you join a community and you want to say "Okay, how you organize my Go code." You google it and find this famous Go standard layout. Is it good? Is it not? + +**Natalie Pistunovich:** I guess if it has some people thumbing it up, or even commenting... If it's active enough, it means enough people don't think it's bad. + +**Alexey Palazhchenko:** Yeah, maybe. But maybe all those people's AI is too -- maybe it just found it, and started it, and then more people are coming and will start it... But the very first people did not know that it's maybe not good. + +**Natalie Pistunovich:** Just skipped it for later... + +**Alexey Palazhchenko:** Yeah. So I would say it's always for the human to decide, and I wonder how this would affect the coding styles of a human for example. It may be very different from what language designers expected it to be. + +**Natalie Pistunovich:** Yeah. This brings me to the question if Go would be a good choice for AI-driven development, for the style of the language. What do you think? + +**Alexey Palazhchenko:** Yes, all this power of go fmt, and you have the same standard for all the Go code - that seems to be very easy for Copilot to understand Go code. And I would say maybe Go is the single best language for that. I don't know another language that has this simple code structure, this regular structure with the same formatting. I can't imagine Copilot for something like C++, for example... Especially modern C++. That would be a hot mess. Because yeah, especially given that Copilot just grabs all the C++ code from GitHub, and then half of that is written 20 years ago, with a very different standard... That's basically a different language at this point. + +**Natalie Pistunovich:** Yeah, that's actually another interesting point, that is AI code trained on up to date code, that is made for our machines, and not for the past machines. + +**Alexey Palazhchenko:** Yeah. Can you imagine Go with generics generated by Copilot? + +**Natalie Pistunovich:** Maybe we will have to... \[laughter\] Eventually we'll have to imagine how that works, yeah. + +**Break:** \[19:32\] + +**Natalie Pistunovich:** I've been trying to do the Advent of Code on different programming languages with using Codex, which is the underlying engine of Copilot. So basically, Copilot is kind of you take whatever you have in your IDE, and then you add the prompt to that, so kind of like the secret sauce, and then you send that to the end point of Codex. And what I was trying to do, and gave up in the middle because it didn't work at all, is trying to do the same with the different languages. And just finding the right context for some of the languages that are not supported as well, like Java, for example... Finding a standard highest header was a big deal. All Go programs will start with package main import. Not many languages have this standard thing, for Java or for C. + +For C, for example, you'll probably start with an import, but maybe not... Not all languages have such a signature, if you will, to also help the AI generate, or even understand what language you wanted to use. I guess that's another big benefit of Go, in addition to having this standard formatting, that it always has the same beginning. So I agree with you that I also find Go to be a good choice for this reason. + +Last week was GopherCon, and there were some conversations around that, and one of the points that I made in a talk about that was that Go will not land you in the uncanny valley of code. That graph that makes you feel awkward about robots, and the one that can make you feel awkward about code. Because if it looks like it's computer-generated, it's like a weird robot. But if it looks like Go code that everybody else wrote, you're probably gonna feel comfortable with it. + +**Alexey Palazhchenko:** Well, I would say that generated Go code sometimes looks like uncanny valley. So it kind of makes sense, but it also kind of doesn't, and it looks very weird... Look at it careful and it doesn't even compile. It doesn't work. But it looks like it almost does what you need it to do. + +This video has a great example when the guy was making a game where you should guess the number, and the one who chooses the closest number wins. But Copilot always steered him to the game of "You have to choose the correct number exactly." So it basically tried to redefine the rules of the game. That was quite strange to see. + +**Natalie Pistunovich:** Interesting. So was it doing that instructions via comments, like the documentation of the function, or...? + +**Alexey Palazhchenko:** Yeah, yeah. He basically wrote the first line of comment, and then Copilot suggested even comments for that, and then code. + +**Natalie Pistunovich:** And what is your experience of using Copilot with other languages that are not Go? + +**Alexey Palazhchenko:** So the second-best language Copilot works for me is actually English... As I mentioned, readme generation is just great. And comments -- yeah, sometimes it writes better comments for my code than I would do. + +And you know, they have this new feature in beta when they allow you to describe the code as English. For example, you have a snippet of code, and -- yeah, most importantly, it works with regular expressions. It translates regular expressions to plain English. That's just great for people who don't know regular expressions. Basically you understand the jargon after that it just explains it in plain English. That's great. + +**Natalie Pistunovich:** Would you say it works great in both ways? Regex to English and English to regex? + +**Alexey Palazhchenko:** Yeah, I did not try that. It would be interesting. Actually, the problem there - it's not an algorithm; it's just machine learning. And maybe it's not even correct. That would be also interesting, to check if that's correct. Yeah, I tried that with Python. I mean, it still works, but in my opinion the Go snippets are better overall in code. + +**Natalie Pistunovich:** \[24:17\] I tried writing some Bash scripts with it and that was also successful in doing what it was supposed to do. I guess Bash is similar to Go in that sense, that it's a kind of one way of doing things. + +**Alexey Palazhchenko:** Yeah, maybe just my way of writing Python is outdated. The last instruction I was using was using Python 2 when I switched to Go. So maybe Python 3 is very different. + +**Natalie Pistunovich:** I imagine Python has multiple ways of doing things, so you cannot just set a one-one. And what languages did you get to try that were not really a good choice? + +**Alexey Palazhchenko:** Actually, I don't think I tried a lot of languages. At my current job I use Yaml. Yaml works great; English works great. I used configuration file, JSON works great. Everything was pretty much great for me. + +**Natalie Pistunovich:** Well... + +**Alexey Palazhchenko:** I imagine something like very esoteric, I don’t know, like Malbolge programming language would not work very well. + +**Natalie Pistunovich:** LaTeX, LaTeX might be interesting to see how that goes. English to LaTeX. + +**Alexey Palazhchenko:** Imagine you could write a Ph.D. with Copilot. + +**Natalie Pistunovich:** "Here is my thesis in one line... Please write my expanded thesis, with correct formatting." This will be it. So AI can be used to augment your life as a developer, in helping you write code and helping you write readmes... But I can think of all other types of help that it can do. For example, creating the documentation for your code, right? + +**Alexey Palazhchenko:** Yeah, absolutely. + +**Natalie Pistunovich:** Kind of the other way around from Copilot. I saw in some -- for example in Replit, which is kind of an online IDE, and there you can highlight a function and then say what does it do. And then it will -- similarly, I guess, to a readme, it will be sort of your English used in programming languages, a helper. Can you imagine some more other interesting places where your AI can be augmenting our developer lives? + +**Alexey Palazhchenko:** Well, I can imagine it doing pull request reviews. They already have quite some help with linters, and CI, and all that... But imagine Angry AI saying "You should DRY this code" or "You should use single explicability" or something like that... And then there would be a slider, like "Angry team leader" or "Pleasant team leader", with different settings. + +**Natalie Pistunovich:** Or like a linter for the answers... A linter before you basically submit your review, that kind of helps you choose the -- "This word is nicer than that word." Or more encouraging. + +**Alexey Palazhchenko:** Actually, one I think I would want from Copilot is to generate proper Go configuration file for me... Because it has so many configuration options. + +**Natalie Pistunovich:** Yeah. It can be sort of like a smart configuration values, right? Kind of like "This is my use case. Give me what's a good default to have." + +**Alexey Palazhchenko:** Yeah. Or just "This is my code. My configuration passes my code." This is how I disabled everything. + +**Natalie Pistunovich:** I hope somebody is listening and writing down and making a startup of each of those ideas, because it's all useful tools. + +**Alexey Palazhchenko:** Yeah, yeah. + +**Natalie Pistunovich:** I have one more to add to our list... Something that you give it a piece of code and then say "Find the security vulnerabilities." Something to point out to you "This is what you should be doing better." + +**Alexey Palazhchenko:** Yeah, yeah. + +**Natalie Pistunovich:** Would you say that one of the interesting things about FerretDB is that it's actually AI-generated, augmented? The first AI-augmented database? + +**Alexey Palazhchenko:** Yeah, you can say so... Actually, when I started I was just alone, myself and Copilot. So you can see he's my co-author now. Right now we have a lot of people from community just jumping in and starting to help... But the core of it was written by me and Copilot. That would be an interesting take, from all the project... The first database system written by machine learning. + +**Natalie Pistunovich:** \[28:16\] By machines, for machines. + +**Alexey Palazhchenko:** Yeah. \[laughter\] + +**Natalie Pistunovich:** I guess it can be also interesting if it helps you to choose names for your project... You know, names for variables. This is a great tool to have. Something that will help you choose variable names. + +**Alexey Palazhchenko:** Yeah, that would be very cool. And yeah, names for projects are very hard, actually. We spend a lot of time thinking about it. + +**Natalie Pistunovich:** And this can also be aware of all the licenses out there. + +**Alexey Palazhchenko:** Yeah. You know you can license particular strings, and then you can use the strings in your code. + +**Natalie Pistunovich:** Or an AI that helps you choose a license for your project. + +**Alexey Palazhchenko:** I see, yeah. That would be interesting, too. + +**Natalie Pistunovich:** The list is really long at this point. I start to forget -- I'm gonna listen to those afterwards and write them down, one by one. + +**Alexey Palazhchenko:** The next step would be it would just write a special license for your code. Only for your code. Custom-tailored for you. + +**Natalie Pistunovich:** Mind=blown. \[laughs\] + +**Alexey Palazhchenko:** It's like, we don't have enough licenses already... + +**Natalie Pistunovich:** I guess you can always make it more specific... So what would be your recommendations for Go developers who are not yet using Copilot, or kind of just played around with it a little bit? So how can you use that as a tool to make your developer life better? What can you already do? + +**Alexey Palazhchenko:** Yeah, so I would say the hardest problem for you could be just get access to it... Because as far as I know, it's still not public for everyone. If you have it and if you've found a way to have it -- maybe people can contact you and you ask them nicely to give some access... \[laughs\] But I would say just try it out if you have access to it, and start writing code with maybe a bit more comments than you previously used, especially if you never write comments in your public code, or private code. Then, after that, it just works miracles. Of course, you should try to verify the code fraud, not just blindly accept it. Other than that, I would say go ahead. + +**Natalie Pistunovich:** So starting by better documentation. + +**Alexey Palazhchenko:** Yeah. And maybe you can just start a new project completely and open a readme file and start typing, and write a sentence, and just autocomplete. + +**Natalie Pistunovich:** Yeah. Yeah. It can be also interesting to use for maybe refactoring. + +**Alexey Palazhchenko:** Refactoring would be actually a really interesting use case... Because right now, all refactoring I know in various editors and IDEs is just completely algorithmic. It's very deterministic. Simple methods like a name or extract might be something like that. But if you're using AI - yeah, that would be a very powerful tool. + +**Natalie Pistunovich:** I can imagine also improving your testing. All the tests that you have for your code; it can also be a quick, easy win. + +**Alexey Palazhchenko:** Yeah... For that, I would still prefer fuzzing. But as for generating initial seed data for fuzzing yeah, that would be great. Actually, I think I did write some tests using Copilot this way. + +**Natalie Pistunovich:** I'm curious now if there's already some existing library that somebody created that "Here's a great dataset (or pack), all the things you can ever test, all the inputs that are worth being tested", or something like this, and generated by Copilot, and then this will become kind of one of the new things that you always involve in your testing flow. + +**Alexey Palazhchenko:** Yeah, actually there's a repository, something about like "repository of bad strings" which contains all these weird edge cases, like unicode, zero strings, null strings, and all that + +**Natalie Pistunovich:** It would be interesting to tell "fuzz it for me", or something like this. Find fuzzy inputs. + +**Alexey Palazhchenko:** \[32:02\] That might be another most effective way to use a machine learning framework, but... + +**Natalie Pistunovich:** I probably saw some interesting examples, but yeah... Maybe not. I'm trying to think what other interesting recommendations we can give to listeners who are just getting -- or you know what? Let's take one step back and let's speak to those who are maybe a little bit cynical about including AI in their flow, and look at that as an NFT or whatnot; a new buzzword. So why is it actually a good idea to include AI in your flow? Why should I not just stick to my good, old Vim, or IDE like VS Code, with no any other add-ons? + +**Alexey Palazhchenko:** I would say a simple answer is it just really helps. There's some issues with representation, like I said before... For example for me, sometimes these suggestions are not formatted very nicely; sometimes they jump when I start typing and want to take a different direction it suddenly appears. Then I have to hit escape to cancel it. + +So I would say representation could be better, but when it works nicely, you go much faster. Significantly faster. And that helps. You can be more productive. Not productive in a way as a joke; very productive as in you type a lot and then delete a lot, and you're just typing very fast... But in a good way. + +**Natalie Pistunovich:** Well, to play a little bit the devil's advocate, we started this show with listing all the different things that you might need to be reviewing if you are getting the help of an AI to be your co-programmer, if you're doing pair programming with it... So is it actually more efficient? + +**Alexey Palazhchenko:** I would say yes. One problem that we did not mention yet - for me at least - is that you just compile it, you have to switch modes between reading and writing. Sometimes you read the code, you understand where you should place new code, how it works, how you refactor it, and then you start writing. And then you pause, you look for what happened, refactor it. With Copilot's suggestions, you have to configure timing pretty well. Otherwise you can be interrupted in your flow. But once you're just used to it, or you configured it, it works much better. + +I would say that I did some measurements not scientific totally from me of course, but I've found it helpful... Simple code, repetitive code is generated basically by GitHub Copilot automatically. + +**Natalie Pistunovich:** Would you say that Copilot is a good tool to use if you are diving into a completely new codebase? ...so you're kind of even trying to understand what goes where... Speaking of reading versus writing. + +**Alexey Palazhchenko:** I don't know... Right now, I don't think Copilot provides much help there. I mean, it works for small pieces, for like "Explain this function." I don't think it works for "Explain this project" or "Explain this module/package." + +**Natalie Pistunovich:** Library... + +**Alexey Palazhchenko:** Yeah. Who knows, maybe that will work in just a few weeks, months, maybe a year. + +**Natalie Pistunovich:** Yeah. + +**Alexey Palazhchenko:** I think one Go-specific way where Copilot would help is as a generics replacement. You don't have to write all this code yourself, you don't have to use generics; you can just ask the compiler to generate all the various functions for different and it works great. + +**Natalie Pistunovich:** \[laughs\] That's an interesting approach. + +**Alexey Palazhchenko:** Yeah... I mean, we used these code generators everyone wrote back in the day, and... reflection, interface...Just use Copilot. That would be fine. + +**Natalie Pistunovich:** So what would your ideal Copilot Go library look like? What would be a super-useful function/functions? + +**Alexey Palazhchenko:** I don't know... + +**Natalie Pistunovich:** I can imagine like replacing these reflections and so on would be -- + +**Alexey Palazhchenko:** So we were speaking about the machine learning model explaining this code. I wrote something like "explain this usage of concurrency". + +**Natalie Pistunovich:** \[36:16\] In the sense of what are potential outputs, or what are you missing when you're looking at it, or what exactly? + +**Alexey Palazhchenko:** So on many layers. For example, this function that taps these channels cannot tap that and that, and that channel is closed at this point, and then the big picture... And then the ultimate question - it should answer "Was it worth it a lot, or you should just write linear code there, without all the goroutines and channels, and be more simple?" + +**Natalie Pistunovich:** So basically sort of a take one time -- + +**Alexey Palazhchenko:** Yeah. Actually, the last question would be the simplest one, I guess... It's just a constant answer, "No. Write simple code." + +**Natalie Pistunovich:** I guess that's part of the value proposition of Go, right? That it's simple. Last week we had - it was part of GopherCon - a game on Go Time where we had to guess what gophers answered. So there was a survey to some hundreds of participants of the conference, and one of the questions was "What is the word that you would use to describe Go?" And then we had to guess what people answered. So the number one answer was "simple". So that might be a little bit funny with that... + +**Alexey Palazhchenko:** Yeah, yeah. That's the reason why I chose Go in the first place. It would be my answer, too. + +**Natalie Pistunovich:** Go is used a lot, or more and more, not just for regular code, like web code, but also for infrastructure; like, different tools that are written in Go. So how can we improve that with AI? Let's see what we can do better. + +**Alexey Palazhchenko:** So let's imagine that we have an AI that actually can write the domain-specific language for our infrastructure. So for all those benefits of Go, Go is quite various language, and this error handling can -- I mean, writing business logic in Go is not the nicest way to use the language. It's great for infrastructure, it's great for some places where you manage bytes and so on, but if you want to write high-level business logic in Go, that's not very easy. + +At the same time, existing solutions for integrating different other languages are also not great. + +So imagine that Copilot analyzes your business problem, your domain, and then creates a tailored, domain-specific language for your problem, implement it in Go, with virtual machine and just-in-time compilation. + +**Natalie Pistunovich:** Interesting... How would you name the startup that does that? \[laughter\] That sounds like such a useful pitch... + +**Alexey Palazhchenko:** Yeah. For that, we need to create another machine learning model, to come up with a name. + +**Natalie Pistunovich:** \[laughs\] Those are some interesting things. It would be interesting to listen to this show in like five years and see which of those already spun up to be a tool or a product. Or an open source project. Maybe somebody will go to Copilot and will write something, "Create for me a project that does this", and then it will be what we need. + +**Alexey Palazhchenko:** Yeah. And then you click a button, "Create Repository" and it already contains all the code in it. + +**Natalie Pistunovich:** And the license. + +**Alexey Palazhchenko:** Yeah. \[laughter\] + +**Natalie Pistunovich:** It's always important to have the right license. + +**Alexey Palazhchenko:** Yeah. So we'll be out of job. Completely out of job. You don't even have to pick up the license. + +**Natalie Pistunovich:** \[laughs\] Computers will be arguing with each other. Yeah, it brings us back to this AI judges, right? That seems to be like the logical next step of the computer industry. + +**Alexey Palazhchenko:** And then we all will be in Matrix. + +**Natalie Pistunovich:** Well, that is some interesting expected future for the development of our life as developers. Not to ask in our lifetime, but let's say in our professional career, which is hopefully for most people shorter than our lifetime, what things do you see that will be becoming a very common thing in a very short term? Something that you would write in your CV when you're looking for your next job, so a couple of years from now. What would be an AI skill that is related to development? + +**Alexey Palazhchenko:** \[40:11\] I think for machine learning you have to have some experience already. And if you don't have it by now, then in a few years it might be a problem to find a good job. And I guess -- yeah, I think Copilot will be much more common, in more editors, more languages, with different representations. + +**Natalie Pistunovich:** Would you be writing that on your CV? I'm asking because I saw CVs of people who write that they do Git as a skill, but I never -- I cannot remember, let's say, seeing a CV that mentions their favorite IDE. So like "I am good with Vim. I am good with VS Code." But would you say that Copilot is more in the category of your IDE, or more in the category of skills like Git? + +**Alexey Palazhchenko:** Actually, I saw quite a lot of CVs that mention a preferred editor. Like, "I have skills in IDE" for example. + +**Natalie Pistunovich:** Okay. Interesting. + +**Alexey Palazhchenko:** Yeah, people do write that. Maybe that's a Russian thing that people just write down all the keywords they know in English and everything else is in Russian... + +**Natalie Pistunovich:** Or you're coming very opinionated. "If you're not working with Vim, I don't want to work here." + +**Alexey Palazhchenko:** Yeah, that also happens, yeah. I saw that, too. I mean, with Copilot you don't have to have some special skills to use it, right? It doesn't have much configuration options. It's not like you could have a ton of knobs to tweak, and you are an expert of tweaking the machine learning model for your use case. You don't even have access to it. I don't think that Copilot would be in a lot of CVs, but then again, we have editors in the CV, so someone will definitely mention Copilot. We can take a look right now and link it in. I'm pretty sure someone already mentioned it. + +**Natalie Pistunovich:** \[laughs\] Copilot, yeah. + +**Alexey Palazhchenko:** Except Copilot developers, of course. + +**Natalie Pistunovich:** Yeah. What I can imagine is that sometimes when using the different OpenAI engines, like GPT-3, Codex, whatever, you have to write the prompt -- or you can write the prompt in different ways, and one way would yield you a better result than another, right? That's kind of the concept of prompt engineering. So I can imagine that this is a little bit similar in that sense in Copilot, in the sense that if you write your comment in one way, it will create one function, and if you write your comment a little bit better, it will create another function. + +And back to this Advent of Code with Codex, I noticed that sometimes some of the days when I just copy-pasted the entire instruction of what is the Advent Code of today that you need to solve, for some languages it solved it perfectly from the first attempt, for some other languages, other days, were just not working. So you can even change your comment or change your instruction to the computer... So you can argue that prompt engineering can be even used in Copilot. + +**Alexey Palazhchenko:** Yeah, that's definitely a useful skill to mention in your CV. You know how to relay things for Copilot. You're fast at tweaking words for Copilot. + +**Natalie Pistunovich:** I wonder if this is a similar category to "I'm really good in documenting my code." Would you write that in your CV? \[laughs\] + +**Alexey Palazhchenko:** Yeah, I'm also very quick at quitting Vim... + +**Natalie Pistunovich:** \[laughs\] "I can do that in only 7.5 seconds." \[laughs\] I guess that all falls under communication skills... But then maybe this is kind of the expanding my communication skills with my human teammates, and with my AI augmentation tools...? + +**Alexey Palazhchenko:** Yes, that would be interesting, if the skills should be significantly different and diverse. You speak with humans with one language, and with AI this different language. + +**Natalie Pistunovich:** There is a point to that, because the model was trained on English, or on natural language, and it was trained on code... But it was trained only on a subset, and it's always -- like, it will never be the same subset of the average person. It's not like the average person of one country has another world of context and vocabulary and whatnot of another person, in another language... So maybe even skills of English is actually becoming more of a need for a developer to have, so you can better communicate with the AI. + +**Alexey Palazhchenko:** \[44:22\] Yeah. And also imagine that you have to command AI to write the code for you, for it to work... But you don't want to command your co-workers, right? That would be a different language already on that level. + +**Natalie Pistunovich:** For sure. Interesting. Okay, so if we summarize kind of the tip around communication skills and how to present that next time you speak about your Copilot skills in a professional setup like a job interview, what would be your take-away? + +**Alexey Palazhchenko:** Yeah, so you could say that you have soft skills and hard skills. Soft skills is how you talk with people, and hard skills is how you talk with AI. \[laughs\] You can say that "Soft-hard skills" or "Hard-soft skills." + +**Natalie Pistunovich:** Yeah. Firmware skills, right? Firm skills. + +**Alexey Palazhchenko:** Yeah. \[laughter\] Yeah, separate section. LinkedIn will introduce it next year. + +**Natalie Pistunovich:** Yeah. More AI engines will be doing code, so it only makes sense. Well, that was an inspiring and interesting conversation... Time for some unpopular opinion. I hope you have one. + +**Jingle:** \[45:28\] to \[45:45\] + +**Natalie Pistunovich:** The nice thing about this tune is that even on episodes where Mat is not here with us, he is a little bit with us... Because he is the creator of this short tune, so... Hi, Mat. I hope you're feeling better. + +**Alexey Palazhchenko:** Well, yeah... Mat can't be imitated, for sure. So my unpopular opinion would be about generics in Go... + +**Natalie Pistunovich:** Dun-dun-duuuun...! + +**Alexey Palazhchenko:** Yeah, so quite a lot of people consider generics to be great in Go, and it works great for them... And even more people consider that generics were a bit rushed, and are afraid of them, I would say... So for me, personally, generics went, I would say, sideways. There are a lot of things in Go that could be potentially solved with generics, but couldn't be right now, because of different directions they went. + +To give you an example, let's say I have something like a JSON object and I want to traverse it by using paths and indexes, field names as strings and indexes as integers... And I can't write generic functions that accept a combination of them. It can either accept a list of paths, or a list of indexes. But I can't write a function that accepts both strings and integers. So I either have to use a reflection for that, or an empty interface, or code generation, but not generics. + +So my opinion would be like generics went in a direction I did not want them to go. So they did not change a lot of how I write Go, and I don't expect that to change anytime soon. + +**Natalie Pistunovich:** Okay... My follow-up question would be do you think your unpopular opinion will be a little bit unpopular, or very unpopular, or actually it may be popular? What is your prediction? + +**Alexey Palazhchenko:** I would say a little bit unpopular. + +**Natalie Pistunovich:** Okay... So you're not going to necessarily make it to the rank of the most unpopular ever opinion... + +**Alexey Palazhchenko:** Who knows...? Maybe generics will be fixed with the help of Copilot in the next release. + +**Natalie Pistunovich:** \[laughs\] Yeah, interesting. I have to say, the last couple of episodes had unpopular opinions on the field of generics; what a surprise... Well, that was an interesting conversation, for sure. If anybody who has listened to this episode and wants to try Copilot but is not there yet, so you can apply on the GitHub website... Probably googling "copilot GitHub" would be the best way to join the waitlist. + +Yeah, that would be wrapping it up for today, so thank you all for joining us... And join our outro tune. diff --git a/Analyzing static analysis_transcript.txt b/Analyzing static analysis_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..200b101064e38906ac551d02f6e4e8c348a4bb0e --- /dev/null +++ b/Analyzing static analysis_transcript.txt @@ -0,0 +1,483 @@ +**Natalie Pistunovich:** Good evening, morning and afternoon, everyone, and welcome to the episode about static checkers. Hello to Mat, my co-host for this episode. How are you doing? + +**Mat Ryer:** Hello to you, Natalie, also. I'm good, thanks. How are you? + +**Natalie Pistunovich:** Great, great! I enjoy all the plants we have in the background. Hello, Matan, our guest for today. Welcome! + +**Matan Peled:** Hello, and thank you. + +**Natalie Pistunovich:** How are you doing? + +**Matan Peled:** I'm doing pretty well. It's kind of late, but... Yeah, it's been a long day. But I'm doing good. + +**Natalie Pistunovich:** And you are joining us from...? + +**Matan Peled:** I'm joining you from \[unintelligible 00:03:11.27\] You can see the sea from that window over there, if it wasn't so dark outside. + +**Mat Ryer:** Oh, wow. + +**Natalie Pistunovich:** So all the way from Israel... And I see you don't have plants in the background, instead you have ducks, which you probably use for duck debugging, as one does. And you have two. One is white, and one is red, for everybody who's listening but not watching... So this is the one who says "Just force-push the commits", and the white one is like "No, no, run one more test." + +**Mat Ryer:** \[laughs\] + +**Matan Peled:** I have a few of them. One of them is a panda duck. + +**Mat Ryer:** Oh yeah, a panda duck. Yeah, it does look a bit like a devil. + +**Natalie Pistunovich:** And the blue one is just in case you're on the blue team, which is kind of decent code. Not too tested, but not force-push. + +**Matan Peled:** Yeah, it seems to balance out the red one. + +**Mat Ryer:** We had an off-site recently at where I work, and it was in Amsterdam... And the souvenir that we got everyone was a little rubber duck, tailored to each person. + +**Natalie Pistunovich:** \[04:06\] Wow. + +**Mat Ryer:** Yeah. It's a very nice gift... But it's very useful for rubber ducking, like you mentioned, Natalie. + +**Matan Peled:** It sounds very thoughtful. + +**Mat Ryer:** Yeah, it is. Well, it makes some better programmers, really. I don't know why -- do they teach that at university? Do they teach rubber ducking? + +**Matan Peled:** I teach it when I TA students, and sometimes they have to help them with their homework. Basically, I have them just explain the problem to me, and through the process of them explaining it, they understand what the problem is. And then I tell them, "Well, the next time you should try that with a rubber duck." + +**Mat Ryer:** \[laughs\] Just leave it there. Don't elaborate. + +**Natalie Pistunovich:** Just in your reception hours. "Here's a rubber duck. Speak to it." + +**Mat Ryer:** Good idea. You could just have that in your office. + +**Natalie Pistunovich:** A weird thing I learned about rubber ducks is that if you actually put them in a place with water, like a bath, you want them not to have the hole that allows them to squeak, because this is how they get moldy. + +**Mat Ryer:** Oh, really? Why? Because water goes in? + +**Natalie Pistunovich:** Yeah. And it's not very well-ventilated. So if you do plan to bring it into a wet room, then make sure that it does not squeak. + +**Mat Ryer:** So if you're a programmer and things aren't going well and you're trying to rubber ducking and you're crying a lot, keep the tears away from the duck... Because it could get moldy. + +**Natalie Pistunovich:** This would be step ten. + +**Mat Ryer:** Yeah. + +**Natalie Pistunovich:** I could think of a few other things you should do before that... But it can be on the extended list of things you might want to do. + +**Mat Ryer:** Yeah, good point. Tissues, hair dryers... Loads of ways to -- + +**Natalie Pistunovich:** \[laughs\] + +**Matan Peled:** Taking a walk. + +**Mat Ryer:** Oh, yeah. That is a good hack, actually; if you're stuck on something, go for a walk, or think about something else. How many times do you hear people say, just in the shower, or even like asleep sometimes, "I was able to solve this problem." Taking a break is important, and that's really counter-intuitive, because you feel like you have to work on it to solve it, don't you? + +**Natalie Pistunovich:** Yeah. + +**Matan Peled:** Yeah. But you feel pressured somehow, usually... And that pressure prevents you from thinking clear. And taking a walk is usually a bit hard, because you need to force yourself to stop thinking of it, and focus on meditation, or walking, or nature... But then your mind has a chance to process whatever it is you're doing and maybe come up with ideas. + +**Natalie Pistunovich:** I'm trying to see what's a good way of tying that to the subject today, but I have yet to find a good one... So in a less smooth transition, no jokes - Matan, you are a Ph.D. candidate at The Technion in Israel, and you're researching meta programming and static analysis. You have worked in all sorts of interesting companies, and now you're back in the academy... So tell us a little bit, what did you study, what did you do afterwards, why going back to studying after being in the industry? + +**Matan Peled:** So for my bachelor's degree I did computer engineering, because I had this (let's call it) romantic idea that I should understand how a computer works, from physics to software... + +**Mat Ryer:** Wow. + +**Matan Peled:** It was very cool, and I learned a lot, and I like my degree a lot, but it made it tougher than it might otherwise would have been, because I'm a programmer, I did software engineering before I started studying, so I was already good at that, and adding something that was completely new and different like Electrical Engineering, and all the circuits and physics and that stuff - it made it significantly harder. But I still liked it a lot, and I think that I learned a lot. And after my degree, I worked at a startup that my friends did at the time, and then I decided that I needed to experience life at a big corporate, so I went to Google, and I spent some time making search features... But eventually, I decided that being a code monkey is fun, and I like that, but I also want to experience research, and see what life in academy is, and what they're doing in graduate degrees... + +\[08:09\] So I went back for my master's, and finished that, and now I'm in my Ph.D. And I like it. I don't know if I'm going to do this forever, but for now, this is fun... And I get to play in my sandbox, and make sand castles, and toy around with my own toy ideas, which is what I like about academics. + +**Natalie Pistunovich:** You mentioned code monkey, and I have to make the joke now, because I still feel overdue from the less than smooth transition from before... Have you tried being a code gopher? + +**Mat Ryer:** \[laughs\] + +**Natalie Pistunovich:** Fun. So what was your master's research topic? + +**Matan Peled:** So my master's research was making a programming language. I was in programming language design. What my programming language did was it took slideshows, and I wanted to add animations and I wanted to basically make a language that describes what the animation is, what the motion on the screen is. + +I liked it, it was fun working on it, but let's just say that its main purpose was to get me a master's degree, and it didn't turn out to be especially usable. It had a couple of really cool ideas... For example, the idea was that I could build a big animation from basic parts, and that if I wanted to make a change in the animation, as anyone who ever used PowerPoint with animations, using the animation pane knows, making it is fine, you can add the animations, but if you wanna make a change in the middle, if you wanna push something in the middle there, then basically you have to start over and sync it all up to what you were doing before. And I wanted to be able to tell it "Add this thing in the middle, and calculate everything according to this change." + +So I had a few cute ideas in there, of making it like a physical system of springs, where you can add a spring and everything basically reacts based on the change you've made. And it worked, and it was nice, but it was not a production system, let's put it that way. + +**Natalie Pistunovich:** Because it had no tests. + +**Matan Peled:** Actually, it did have tests. It's just that basically you had to think like me in order to use it... And think like me at that very specific time of me writing it. + +**Mat Ryer:** So how important is that to you then? Because you talk about you like being in the sandbox, building sand castles... How much of that is anchored back to something that could be practical or useful? + +**Matan Peled:** Well, that's a good question. So my thinking is that I want it to be anchored back to something that is practical and useful... But I'm okay if it's not. I'm okay going often to the wilderness, and exploring, and finding out things... And if it turns out that they're not impactful, and they don't make money, then that's okay, too. Ideally, in my dreams, I want to go out and make something that everybody would know and use, and be important and useful... But I'm okay with it not being that. Basically, taking a chance and not knowing how it's going to turn out. + +**Mat Ryer:** Kind of like a startup attitude, really... Being able to fail, and having that freedom to fail is quite important. It gives you that extra permission almost to do things that otherwise people might not have a chance to do. + +**Matan Peled:** Yeah. In one sense, I think startup culture is about being small and agile, and being able to push yourself into a niche that a bigger company just wouldn't fit into, because they can't be flexible enough just to think about it that way, to allocate research for that thing... But on the other hand, I think that startups always think in terms of MVPs, and making usable things and making products, where in academics what you wanna do is always write a paper. You want to have an experiment, results, data that you can tell other academics about. And usually, in order to tell other people about what you're doing, you don't need to build the full product. You don't need to have users, you don't need to do any of that. You just need to make your specific experiment, and write it up well enough so that other people find it clear and interesting, and that's it. + +**Natalie Pistunovich:** \[12:32\] That sounds a lot like blogging about fun projects that you have. + +**Matan Peled:** It's blogging on steroids. It's blogging with a lot of formality added. And that's why science communication, where scientists blog - that's a thing that's happening. Academics usually love to write; that's what they do in their day-to-day, they're usually good at it. They make blogs. Academic Twitter, for example, is a thing. + +**Mat Ryer:** Oh, I bet that's amazing. Is it good? + +**Matan Peled:** Yeah, if you're interested in that sort of thing. + +**Natalie Pistunovich:** Have you been tweeting about static analysis tools? Or static analysis in general. Or wait, maybe we can start with saying what is static analysis... + +**Matan Peled:** So static analysis - basically, what we want to do is figure out certain properties about code. A property could be, at its most basic thing, "Does this program have a bug?" or "Does this program succeed?" And we want to do all that statically, and that means without actually running the program, because running the program might have side effects; it might do something that we don't want to do right now. It might take a long time... We just don't want to run it yet. The program might not be even finished. We can't run it, and we still want to know things about it. + +So static analysis can be anything from where this function is called, to "Is this program written in the correct style?", which is something that's not that big of a problem with Go, because we have go fmt, which is supposedly also like a static analysis tool. But other languages have things like linters, which tell you that "Well, your indentation is incorrect here", and that's also a thing that happens. + +But static analysis can also be a part of refactoring where you want to rename a method and you want the IDE to find out where all the calls to this method are, and use static analysis to find that. + +**Natalie Pistunovich:** Yeah, right-click refactor. + +**Matan Peled:** Exactly. + +**Break:** \[14:44\] + +**Mat Ryer:** Static analysis must be easier on typed languages that are strictly typed, versus languages that are dynamic, that have heavy runtime elements... Is that true? + +**Matan Peled:** Oh yes, definitely, because type checking is also a type of static analysis. Basically, by having types, you're giving the tool a lot more information than it can use. And if it has more information, then it can do more stuff. + +One of the basic truths of computer science is that static analysis is impossible. You have the halting problem, which Alan Turing proved way back when, which says that basically you can't make a program that says if another program will halt. And the proof of that is very cool, because basically he said that "Well, if I had a program that could do that, and it could put itself into it, that will lead to a logical contradiction, so we can't have that." + +And a corollary to that is the Rice theorem, which says that you can't prove any interesting property (as in non-trivial) about a program. So you have this strong theoretical basis that says you can't do that, and yet you have this rich scientific field where we do that every day. It's not a problem, it just turns out that the interesting programs, like the ones that real software engineers write - they're simple enough that we can analyze them. But what that theorem means is that we can't be 100% sure. We have to make some sort of concession. We have to have some program where this won't work within a limitation. So for every language, if you do a crazy enough thing with -- what's it called when you reference the method by its name, instead of calling it? + +**Mat Ryer:** Reflection. + +**Matan Peled:** Reflection, yeah. If you have enough reflection, if you do enough pointer tricks in C, you can always confuse it enough that it doesn't work. But that's fine, because for 90% of the programs it does work, and that's usually good enough. We're talking about static analysis, not verification. + +**Natalie Pistunovich:** And how does this field tie to your research? Or what is your research about? + +**Matan Peled:** So what I want to do in my Ph.D. is meta programming using static analysis. And when I say meta programming, what I mean is code that writes code, or code that changes code, so basically refactoring. Refactoring usually means that you change the code and then you work on that changed version, but you can also have a compilation step that changes the code and you never work on that changed version. So that's what I mean by meta programming - all those things that make code change code, templates, maybe even generics, things like that. + +And what I think is that basically making them aware, having them use static analysis information can make them more powerful, more efficient. So I can for example say -- one of my initial examples was making reactive programming. Let's say I have this class, and in this class there is a field, and it has a getter, and what I want it to do is I want it to send me an event someone every time the field changes. But the class is not written that way; whoever wrote it just wrote a getter and you have to call it. And what I want to do is I want to find out every way that this field can change in the program, and every time that it changes, I want it to send the event, so I can know when that happens, that it becomes reactive. + +\[20:15\] So if you can do static analysis and modify the program based on that, then you can easily do that. And that's basically my goal - I want to enable things like that, and I want to make in (let's call it) a declarative way, that I can build using basic building blocks more complex behavior. + +**Mat Ryer:** It sounds really interesting. One example of static analysis I've seen - because you've mentioned quite a few... And I actually hadn't considered even like formatting as one of those. But of course, it makes sense. One of the downsides to the format go fmt tool is if the program is incorrect, it doesn't work. The program has to be -- + +**Matan Peled:** Well formed. + +**Mat Ryer:** Yes, thank you. It has to be well-formed. So any kind of static analysis that can happen without that being the case, I find that to be quite amazing... Because often it relies on the same kind of packages that analyze the program for compilation to do static analysis. Is that right? + +**Matan Peled:** Definitely. So handling things that are partly correct, or are partly complete even - like, they're not incorrect, they're just missing a bit and you want to take just the parts that are there that are good, is hard. + +One of the other projects that I'm currently working on - it has to do with pseudo-code. So what we want to do is to compare pseudo-code to actual code and see if they match. That's kind of a similar idea, because pseudo-code obviously doesn't have perfect syntax. + +**Mat Ryer:** I use occasionally GitHub Copilot, and that actually does quite an interesting job... The code can be wrong. In fact, you can give it context just by writing comments, or just by the names of the functions that you use, and the variable names, and things. So that definitely feels magic. I guess that's different, because I suppose that's the ML doing that work, right? + +**Matan Peled:** Yeah. So ML for PL or however you wanna call applying machine learning to code - it's interesting, because on the one hand, code... So a lot of the techniques that are used there come from NLP, from natural language processing, which obviously makes sense, because this is text, and this is text, and you won't use techniques that come from image processing; that has nothing to do with it. + +**Mat Ryer:** Right. + +**Matan Peled:** But on the other hand, code is very structured, it's very hierarchical, it has properties... In order to compile, it has to be very strict in various ways. So giving up all that information, all that context is silly. You do wanna use it, and the (let's call it) non-machine learning approach to static analysis, to dealing with code is called formal methods, which is basically taking ideas from logic and those sort of areas of math, and applying them to code. And that's where all the things like type checking and that come from, all the theory behind it. + +I don't understand 100% how Copilot works. I've read their white paper, it's very interesting... I don't think that the -- on the one hand, one of the points of machine learning is that they don't do anything specific, they don't say "Oh look, there's a type." They want the machine learning to somehow learn that themselves... + +**Mat Ryer:** Yeah. + +**Matan Peled:** ...but on the other hand, I think that they do try very hard to make sure that the algorithm has access to type information, and things like that. + +**Mat Ryer:** Yeah, it's funny, because it gets things right that are really surprising, and it makes mistakes that a simple static analysis tool wouldn't make. It still does make those mistakes, and I'm sure they'll keep working on that. It's almost like there'll be another extra check after to see whether this even is valid code. + +\[24:06\] It does get frustrating sometimes, because it'll kind of guess arguments to a method that are wrong. They look like it's the thing it's seen before, but they aren't the arguments for that particular method... So just a quick check would have found out that that wasn't gonna work... And I suppose that's what they'll do. But that is interesting, you do see clues really of what it's doing, with some of the mistakes it makes. But it is amazing, I have to say. + +**Matan Peled:** Yeah, and the code that it came from, what it is that it learned that would make it answer in this way, right? + +**Mat Ryer:** Yeah. + +**Natalie Pistunovich:** Yeah, it's actually exactly that that I wanted to bring up, all this AI creating code... Copilot is based on Codex. This is the engine that is running behind it. And a little bit from under the hood of how things are going on there - the plugin for Copilot is collecting some context which is not known. This is kind of the secret sauce. And it's being sent with some extra instructions which are also not know, to that engine, to Codecs, which is probably the article that you read, Matan. And sometimes you can see, because it collects the wrong context, it provides something that it's all in the past, but is not relevant for your code... Like what you said, Mat, about that signature function that could have been easily caught... And it actually makes a lot of sense that the next good step of such a development of such a tool - and Copilot is one of the tools that relies on Codex. There are other tools out there that use that... Would be exactly creating static and maybe in the future even dynamic checkers. + +But definitely, the ability of not needing a full working program in order to run such a test is a huge deal for that next step, so this is super-interesting. + +**Mat Ryer:** Yeah. By the way, when I first learned about Copilot, for some reason I read it as Copilot \[Copy Lot\], like it copies a lot... I thought it was like Camelot... So I completely misunderstood that it said Co-Pilot. I was using it for ages in the preview, and thought it was called Copy Lot. \[laughter\] It was quite nice. + +**Matan Peled:** So one of the companies that I worked at during my masters, where I interned one summer - they used to be called \[unintelligible 00:26:24.09\] Now they merged with/bought out/became Tabnine, and they're making a very similar to Copilot tool. And they have the same ideas, only I think that what they do, knowing behind the scenes, is their algorithm is a bit less blind than Copilot. So you can't have it do things like the advertisements for Copilot, where you write the documentation for some function and it just completes the function for you... But you can make it do things like, you know, you start a database connection and it completes all the boilerplate for you, and things like that, based on other examples that it has seen. And it does use more type information, and names, and things like that. + +**Mat Ryer:** That's so clever. One example I've seen of static analysis that surprised m, and actually got me quite excited about it, was an example case where if a variable at any point in the program is called password, and then at some other point in the program it's logged out somewhere, that would be then a warning that it would say "Oh, look, this variable, whatever it's called now, this is being printed." At one point, just sort of analyzing the code, this was a password. So is this really what you wanna do? And I've found that to be actually really quite interesting... Because that is very useful. + +**Matan Peled:** Yes. So Perl, if you remember, if you've ever used it, has this whole idea where you need to sanitize your input, and bless them, and things like that. They had this idea that if you take input and you don't use it carefully, then it could affect the program in ways; SQL injections, and all that sort of thing. + +\[28:20\] That sort of analysis is called taint analysis, and I think that in recent years even, that sort of thing has become important, where you can leak out the password... And also just develop a related, but not the same idea, that a developer is just putting secrets into their GitHub repositories. And that's also something a search might find, that yeah, maybe you should input -- there are specific places in the environment that you're supposed to put your secret keys in. + +**Mat Ryer:** Yeah, and I'll tell you what - having that early... So we talked about the program doesn't have to be finished for this to work... Having that insight as you're working - that's really when you need it, because that's the point at which maybe you're making design decisions that you'll then have to live with... So yeah, that's very exciting. What are some of the cool use cases or cool little things like that that you can do with this? + +**Matan Peled:** Basically, you can find bugs early. Other programming languages - Rust, if you read about it, put this whole idea of being very strict with the checking and the static analysis inside the language itself, and you can basically make sure that pointers don't go out of scope, and be used, because the language itself is specified to keep track of that. + +Other things you can do with that is, you know, if you're doing multi-threading and you have mutexes and other locks, you can use static analysis to make sure that after every lock is both locked and then unlocked, and you don't unlock something before you locked it, and things like that. Every allocation is freed if you're using something like C or C++ and you're allocating memory manually. Every file that you open that needs to be closed - you can check that. + +Some of those things, in some languages you have to check, so it makes sense to have static analysis for it, but in other languages it's not even a problem, because the language itself takes care of freeing resources. But in those languages that do use manual resource allocation, then that makes a whole lot of sense. And of course, memory and files aren't the only resource you have. You also -- if you talk to a server by some protocol, then you can have static analysis to make sure that you complete the protocol in the correct way. + +**Natalie Pistunovich:** What are your thoughts about the static checkers in Go specifically? + +**Matan Peled:** So I don't know a lot about Go... I'm technically a professional Go developer, in the sense that someone once paid me money to write some Go... + +**Mat Ryer:** Oh... + +**Natalie Pistunovich:** That is the definition. + +**Mat Ryer:** Yeah, you're one of us! + +**Natalie Pistunovich:** \[laughs\] + +**Matan Peled:** But I'm not very good at it. So the one static checker that I've found for Go was called Staticcheck, and it seems to be quite thorough. It has a lot of linting options, and things that it can tell you that might be wrong about your Go program. + +So we've talked about all kinds of static checking, and we can talk about levels, but a lot of what Staticcheck mostly does is linting. So it looks for certain patterns of things that are dangerous, or might be incorrect, or are probably not what you meant to do, and then it warns about them... Which is a very useful thing to do. + +It seems that it also has some deeper static analysis, because it can track contexts of various errors, and figures things like that... So it looks like a great tool. + +**Mat Ryer:** Yeah, there's actually quite a range of them, and some of them are general-purpose; others are very specific... There's a tool called Errorcheck that checks to make sure you don't ignore any errors, for example... Which is something that is quite important. + +\[32:11\] And then there's the Go Meta Linter, which essentially runs all of the linters and does those static checks; like you say, it's really linting, a lot of it... So we'll put some links to these in the show notes for people interested. But they're integrated nicely into IDEs already, so you probably already have them. + +**Matan Peled:** The thing that I find interesting about -- like, I want to do the static analysis research... Make my own, basically. The linters are great, everybody should probably use them. Everybody probably does use them, because as you say, they're already integrated into the IDE... But every project has its own things that it's doing, and it's using a library in some way, it's using an API... And what I want everyone to be able to do is be able to define their own rule set, or be able to use a language to define their own static analysis that will warn them of things that might go wrong when they're making code. + +For small projects, for scripts, you probably don't need that. But if you need to collaborate between multiple people that might do the thing, if you're a company or you're an open source project, then those things start to make a lot of sense. + +**Break:** \[33:33\] + +**Natalie Pistunovich:** What are some static checkers, from any language, that you saw, that you really liked what they do? + +**Matan Peled:** Oh... + +**Natalie Pistunovich:** What's a functionality that you really like? + +**Matan Peled:** Um... + +**Natalie Pistunovich:** The static checker that you will build, what will it have? + +**Matan Peled:** \[36:02\] So one of the harder things to do with static analysis - and it's not like a tool in and of itself, it's a way to get there, but it's called points-to analysis, because even in languages that don't have pointers, you usually have references, which means that one thing references another thing, and that thing may change over the course of the program. And keeping track of what aliases a certain object in memory may have - it's hard to do when you're coding the program and trying to keep a mental model of the program in your head... It's hard to do when you're debugging and you need to find out "Wait, what does this point to right now?" And it's even hard to do when you're trying to do static analysis, and that means that you're not even running the program yet. + +**Mat Ryer:** Mm-hm. + +**Matan Peled:** So if you know what this variable points to right now, where is the thing that was allocated, what's its dynamic type, what thing it really is, then you can make all other static analysis basically stronger, because now they can know more things. They know that "Oh, this is a pointer. I now know where it came from." + +So that's static analysis that I thing is really cool, it's really difficult, because programmers can do whatever they like, and you need to somehow constrain this chaos... But yes, that's what I want to do well. + +**Mat Ryer:** Yeah, that feels like a problem that would be much easier solved at runtime. + +**Matan Peled:** Definitely. At runtime you just know what it is. You don't have to check it. At runtime you have other problems. Let's say you traced your programmer and now you have this huge file of \[unintelligible 00:37:46.23\] of where everything went, and you still have to sort out that trace, to find the way that it looks... Because usually, when you're debugging, what you see is that "Okay, so I have this value right here. How did it get there?" The place where you see the error, where you notice that something went wrong, is not the place where the error happened. What you really want to know is all the path, the operation that happens on this value to get to this obviously incorrect state. And that's hard. And there's rr, which is basically a reverse debugger... + +**Mat Ryer:** It helps you put bugs into your program? What's a reverse debugger? + +**Natalie Pistunovich:** It sounds like fuzzing. + +**Matan Peled:** A debugger that goes forward - you basically say "step" and it goes forward one instruction, right? + +**Mat Ryer:** Oh, right. Yeah, yeah. + +**Matan Peled:** Well, a reverse debugger allows you to step back and go back in time to what happened before. + +**Mat Ryer:** Oh. Time-traveling debugger. + +**Matan Peled:** Exactly. + +**Mat Ryer:** That sounds amazing. + +**Matan Peled:** It is. It's incredible. It's a fantastic bit of engineering. + +**Mat Ryer:** But does it just keep a snapshot of the state at every point, or is it more intelligent than that? Because some operations you lose information, I guess, don't you? How does it go backwards in time? Is it time-travel? + +**Matan Peled:** Basically, that's what it does. It keeps operations at every point. But there's so much bookkeeping that you have to keep up with in order to do that... Because obviously, you can't just after every machine opcode keep state, because that will blow up basically no time at all. And there are other things that a program does, like output to the screen, and write into sockets, and things like that. So you have to be very clever with how you keep it. + +So what it basically does is keep snapshots, but not after every point; only before things that input or output. It figures that the rest of them it can just calculate from that. + +**Mat Ryer:** Gotcha. Yeah. I think that makes sense. That does sound really cool. I wonder if we've got that for Go. I've never heard of it for Go, but maybe. + +**Matan Peled:** \[40:04\] So it might just work with Go, because -- I don't know, let's check. + +**Natalie Pistunovich:** Everybody's googling. + +**Mat Ryer:** Yeah, I think it might work. I've found "Debugging a flaky Go test with Mozilla rr." + +**Matan Peled:** Yeah, because it works at the assembly level. So it cares about machine opcode. And if it compiles into machine opcode, then it can work with it. + +**Mat Ryer:** Oh, that's great. + +**Matan Peled:** And yeah, it's hard to use, and you're stuck with a debugger that looks like GDB, which is not the most user-friendly of interfaces... But it does work. It does do the thing that it promises to do, which is very cool. + +**Mat Ryer:** Well, it's a good opportunity for somebody to build a tool, or integrate it into an IDE then, in that case, if it's just got that kind of text interface. + +**Matan Peled:** Yeah, definitely. I'm sure that JetBrains or whoever are on it. + +**Mat Ryer:** Other IDEs are available... + +**Matan Peled:** That's true. + +**Mat Ryer:** I have to just say that for legal reasons. Actually, I don't think I do, but i say it. Yeah, I bet they are. + +**Natalie Pistunovich:** Also, if anyone is looking for an interesting talk title, for any upcoming conference, I think this is a topic I definitely never heard of. + +**Mat Ryer:** Yeah, I would love to hear a talk on all this, actually. If Matan's not gonna do it, someone should. + +**Matan Peled:** Yeah. So dynamic analysis - let's talk a bit about that. Basically, static analysis helps you before you even run your program, and it can help you find bugs, and it can help you just answer questions about your program is, and how it works, and maybe answer a question about queries, finding things within your program... If you have enough code, then just searching it is a task. + +**Mat Ryer:** Right. + +**Matan Peled:** But dynamic analysis is still a hard task. You basically can use the same information that you would use during compile time, but now you have all the real-time values too, so you can - instead of doing symbolic execution and trying to figure out what the values can be, you actually can know what the values are. But you still have to keep track of them. + +In some contexts you just print statements everywhere and you're done. You look at what you programmed with them, and then you're happy, and you figured it out. But sometimes print statements aren't enough. If you're doing serverless programming, with Amazon Lambda, where you write single functions and hook them up, and they don't have anywhere to print, and you don't know when they run, and you don't know how they will run, but maybe you can get a trace if you put them together correctly, and \[unintelligible 00:42:36.06\] Then you can use this trace to figure out "Oh, so I got this bad value from the database, and then it went through 11 different lambda functions before it got here, and this is where my error came from." Putting all that together is not at all trivial, and you basically have to build a tool to do that. + +**Natalie Pistunovich:** Or listen to the episode from last week, where Mat was talking about instrumenting. + +**Mat Ryer:** Yes. + +**Matan Peled:** Yeah. So instrumenting is like dynamic analysis, right? You're looking at what is happening. + +**Mat Ryer:** Yeah, and that's the kind of thing where -- you know, if it's running at scale, of course. That's a different picture to just the code, or just the single program itself. But yeah... So just printing lines out - that is kind of dynamic analysis, I guess. + +**Matan Peled:** Well, it is. It's a very primitive form of it... And it's not argumented by tooling, let's say it like that. And making tools for making our jobs easier is what we do as programmers, right? + +**Mat Ryer:** Yeah, absolutely. That is interesting. We even try to do it with structured logging, where now we are putting structure into the log output, so that we can use that more later. + +**Matan Peled:** Yeah. And then this structured log is almost like a trace. + +**Natalie Pistunovich:** That's definitely an interesting view. Instead of saying, "I'm just printing things" you can say "Now I'm dynamically debugging." + +**Mat Ryer:** \[44:12\] Yeah. "I'm doing dynamic analysis." But would you say Hello World is a dynamic analysis program? That's all it is, isn't it? That's probably the simplest... + +**Natalie Pistunovich:** What information does it give you? + +**Mat Ryer:** It says hello. + +**Matan Peled:** I guess... It just prints -- like, it tells you when you entered the function and when you left the function, and that's a trace, right? + +**Mat Ryer:** Yeah, that is a little signal, isn't it? + +**Matan Peled:** Yeah. + +**Mat Ryer:** Yeah, that's a good point. + +**Natalie Pistunovich:** And it can be fun to add a timestamp to those. Then it's like a proper log. + +**Mat Ryer:** There you go. If you use the log function... If you log println or something in Go, you get the timestamp for free. You could use Go in your work, because one of the nice things about Go is it's an open source language. You have at your disposal available all the packages that the Go toolchain itself uses to understand Go. + +Go actually now is written in Go. Talk about meta programming... Go used to be written in C, and now it's written in Go. I can't wait for that initial information to be lost in generations' time, and they just know that Go is written in Go, and no one knows how... I love the idea of that. + +**Natalie Pistunovich:** Yeah, we'll use rr to go back... + +**Mat Ryer:** Yeah. \[laughs\] + +**Matan Peled:** No, wait... Once you said that, I have to tell you about what I think is one of the coolest things ever... And that is a talk called "Reflections on trusting trust", I think by Brian Kernighan... + +**Natalie Pistunovich:** Also available in the show notes... + +**Matan Peled:** Ken Thompson, sorry. It's Ken Thompson. So it's a thing that compilers for languages should be written, or -- compiler writers like to write them in the language that they are compiling. So this is called a self-hosting language, and it's basically a milestone for a programming language to have a self-hosting compiler, because it means that a language is sophisticated enough to write a compiler for itself. And writing compilers is one of the classic computer science problems of complexity, let's call it that. + +And the idea behind "Reflections on trusting trust" is that the C compiler is written in C, and it compiles itself. So if you added a backdoor into it, for example every time that it tries to compile the logging program, then it also adds a little backdoor that accepts a username and password that is unknown, then that compiler would insert a backdoor into code that would not be in the source code of the logging program. That's not good enough, because then the source code of the compiler will just show you that it's doing that, right? So it can't have that. + +So what we could do is add another backdoor into the compiler where it adds a backdoor into itself when it compiles itself, that both adds this backdoor and adds a backdoor for the logging program. And then you would have a backdoor that is basically undetectable, unless someone is especially fond of reading compiled assembly language; not even handwritten assembly language with comments, just compiler output... The backdoor would appear nowhere in source code, it will only be in the binary... And you can't just recompile the compiler to get rid of it, because it would keep adding it. + +**Mat Ryer:** Oh, wow... That is awesome. That's really creepy. + +**Matan Peled:** It is. + +**Mat Ryer:** It's like a Black Mirror episode, really, or something like that. + +**Matan Peled:** Yeah. + +**Natalie Pistunovich:** There must be some virus, some hacking software that is using this... And it seems to have been around for a while. + +**Matan Peled:** Yeah. + +**Mat Ryer:** He wrote that in 1984... + +**Matan Peled:** \[47:47\] So it's been around so long that there have been anti-reflections of trusting trust ideas, where you basically have to use multiple compilers to get one verified output, and there's a whole slew of ideas that combat this... But if you like these ideas of things that reference themselves and things that -- + +**Mat Ryer:** Mat does. + +**Matan Peled:** ...then there's a book called "Gödel, Escher, Bach", which I can't recommend enough. + +**Natalie Pistunovich:** Joining the recommendation, for sure. + +**Mat Ryer:** Yes. + +**Natalie Pistunovich:** That is a very interesting one. + +**Mat Ryer:** I love that book. I agree. It's bonkers. It's so good. + +**Natalie Pistunovich:** If the three of us agree... + +**Mat Ryer:** Yeah, there you go. This doesn't sound -- go on, Natalie. + +**Natalie Pistunovich:** \[laughs\] + +**Mat Ryer:** If all three of us agree... + +**Natalie Pistunovich:** It does not sound like an unpopular opinion. + +**Mat Ryer:** Yeah! + +**Jingle:** \[48:36\] to \[48:55\] + +**Natalie Pistunovich:** You have to agree this one was smooth. + +**Mat Ryer:** Brilliant. That's the best one yet. + +**Natalie Pistunovich:** This also is another popular opinion. Okay, Matan - so as a preparation for this episode we asked you to come up with an unpopular opinion that can or does not have to be related to Go, or programming, or anything like that. So we're ready to hear... What is your unpopular opinion? + +**Matan Peled:** So my unpopular opinion is that after going through all that, hyping up static analysis and all the things that it can do, my unpopular opinion is that it actually doesn't work. It works up to a point, it does all the cute things, it does the simple things, but you can only make it so complicated before it all breaks down and you get to keep the pieces. And that's true if you try to do it formal methods, and if you try to do it using machine learning; no matter what you do, you're still stuck, thinking very hard and trying to solve the problem by sheer for of will, where no tool can help you. + +**Mat Ryer:** Is this just your kind of academic brain seeking out perfection, and not finding it? + +**Matan Peled:** That, and my grad student brain trying to do things repeatedly, and failing, and saying that "Well, maybe this doesn't work." Yeah, so my unpopular opinion is that software engineers basically have job security, and computers won't replace them. + +**Mat Ryer:** Oh, there you go... But I think that's gonna be very popular. We will test this on Twitter. + +**Matan Peled:** But not in my academic niche. + +**Mat Ryer:** No, it's not good for you at all. At all. \[laughs\] Are you gonna quit and do something different, or sticking with it? + +**Matan Peled:** No, I think I want to push it as far as it will go, but keep in mind that maybe it won't be infinitely far. + +**Mat Ryer:** Oh, that's amazing. + +**Natalie Pistunovich:** I think we'll all slowly get to be more prompt engineers, which is sort of maybe the next level of abstraction, but not necessarily... But we'll basically be guiding the AI to do things for us, among them programming... And it's technically natural language, but it's not exactly the English we're using every day. And we all come with different Englishes, and... Well, we understand each other, but the computer understands us a bit differently. So it will be basically sort of a next level of programming. So in some way we will automate ourselves out of a job, and in another way we'll all just have new jobs. + +**Mat Ryer:** We could just write the tests. I feel like even with the fuzzing thing it could make this work. + +**Matan Peled:** Oh, so you say that, but that is a thing. I have friends who are researching that. It's called synthesis, and it's basically - you write the specification of a program, and the tests are specifying what the program should do, and synthesis... Like, either, again, you do machine learning, or you just search every program possible in a very specific way that actually makes it actually find programs sometimes... And you can make programs that way, you can do programming. + +\[51:58\] Excel, if you think about it and all the autofill things that you can do, is basically this. You write what you want it to output, and then you drag it, and then it figures everything out. Especially with the new features they've added. + +**Mat Ryer:** Oh yeah, I've seen them. I tried it the other day. I put 1, 2, 3 and then it just repeated 1, 2, 3 loads of times. I was furious. I was absolutely living. \[laughter\] But you're actually right - if it's doing that with the code... I mean, if it's got things like that points to analysis, where it understands the memory use, and performance, and things - it could even then optimize code. + +**Matan Peled:** Oh, yeah. + +**Mat Ryer:** It could give you an early version, and over time just keep sort of improving it, and things. And that does get very exciting, when you think of that running at scale. + +**Matan Peled:** So again, my unpopular opinion is that that will never happen. That the best it can do is sort of point you in the general direction of saying "Well, maybe you wanna look at this. This could be a good place to look at." But that it will never be able to do that by itself. It will never know enough about the program to do it. + +**Natalie Pistunovich:** Interesting. + +**Mat Ryer:** Yeah... Because even like programs we write - they contain bugs. I had a manager once that said he doesn't want any more bugs in code. + +**Matan Peled:** Yeah, so just don't write any more code. No more bugs. Easy. + +**Mat Ryer:** Yeah, there you go. + +**Natalie Pistunovich:** No code. No code is the future. + +**Mat Ryer:** There's definitely truth in that, but genuinely though... Like, yes, it's about whether the program does what we want it to do based on criteria that's external to that program. So in a way it's not available to it to know that. But I don't know, could you write a test...? + +**Matan Peled:** But even the tests don't specify the program completely, right? We all know that writing tests is hard. + +**Mat Ryer:** Writing good tests is hard, isn't it? + +**Matan Peled:** Writing tests that are a good specification is even harder, because - well, if you tell it to multiply, you get two, it goes to four... Then, "Okay, yeah, I can write a program that always outputs four." That works, right? + +**Mat Ryer:** Mm-hm. That's why you need more than one test case. + +**Matan Peled:** Exactly. + +**Mat Ryer:** Yeah. + +**Matan Peled:** It will almost be adversarial, in that it will always find a way that it can do the thing that you don't want it to do, instead of the thing you wanted it to do. + +**Mat Ryer:** Yeah. Also, what you want can change too, over time. It's quite interesting. Could using static analysis -- could you check to see that tests don't contradict themselves? + +**Matan Peled:** Oh, that's an interesting question. I suppose you could, but it depends what you mean by contradict themselves. You could use static analysis to extract them somehow and compare them, and see... Yeah, you could definitely do that. You could see if what they say about the method that they're testing - if they're unit tests, or you test a method - if that's consistent. If it can think of a method that outputs a certain thing. + +So a static analysis -- let's say you have a method that returns some integer... Then we have various sorts of integer analysis that can give bounds. This output is between 0 and 8. Give it like an interval. That's interval analysis. There are more complex types of integer analysis that try to figure out what the value can be... And yeah, it can definitely figure out that there's a contradiction somewhere. + +**Mat Ryer:** Yeah, that is interesting. I guess that's why pure functions, like that Rust has - that must be a much easier language to work with than Go, because they can have side effects to the methods and functions in Go. + +**Matan Peled:** So static analysis doesn't really care about the side effects, because it's not executing anything. If you're reading input or something like that, then obviously you have no idea what value it is, or what value it might be... But you just mark it as any, as top, and keep going. Yeah, it can be anything. That's final. Just put that into the analysis. + +**Mat Ryer:** Hm. That sounds really cool. + +**Natalie Pistunovich:** Alright, folks, that was very interesting, and that developed even in a more interesting direction after the unpopular opinion. I already wonder what is the next episode we're gonna do about this. Until then, thanks to everyone who joined us. Have a great rest of your day! diff --git a/Answering questions for the Go-curious_transcript.txt b/Answering questions for the Go-curious_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..3d9b3ce2752078beb194d80c2b24eb17df1d310a --- /dev/null +++ b/Answering questions for the Go-curious_transcript.txt @@ -0,0 +1,567 @@ +**Jerod Santo:** Hello, and welcome to Go Time. I'm Mat Ryer, and I've spent my entire life working on my mid-American English accent for this very moment when I could imitate Jerod Santo... No, I'm not Mat Ryer! Mat is -- I don't know where. But I am here, I am Jerod, your humble producer, coming out from behind -- where does an audio producer sit behind? I don't know... + +**Kris Brandow:** The soundboard, maybe...? + +**Jerod Santo:** Yes, from behind the soundboard, to ask some questions today. If you listened to our Charm episode of The Changelog, which we aired in the Go Time feed because there was so much Go talk on that episode, you know that I've been Go curious as of late. But I have my apprehensions, I have my questions. I haven't actually dug very far into Go, and I have questions, so I thought I'd bring a few friends together and ask all sorts of newbie, outsidery, shallowy questions. So joining me today is Go Time regular, Kris Brandow. What's up, Kris? + +**Kris Brandow:** \[04:12\] Oh, hey! Doing pretty well. On for the second time this year. This first quarter has flown by so quickly. + +**Jerod Santo:** You're back, baby. You're back. + +**Kris Brandow:** Yeah. Can't believe it's March already. Or I guess April, by the time this episode comes out. + +**Jerod Santo:** Yeah, Q1 has come and gone, and here we are, Q2... But what else is gonna happen, right? Every day another Q goes by... + +**Kris Brandow:** Yeah. + +**Jerod Santo:** The sentence didn't even make sense. \[laughter\] Let's introduce Ian Lopshire... I'm doing my best Mat Ryer impersonation with the non-sequiturs. So every time I say something silly or dumb, just assume that it was a Mat Ryer impression. We have Ian Lopshire, a guest, but a common guest. Welcome back to Go Time, Ian. + +**Ian Lopshire:** Yeah, happy to be here. I'm kind of reeling right now, because I hadn't thought that Q2 was over yet, but... I just looked at the date, and you're right. + +**Kris Brandow:** Q2 isn't over, Q1 is almost over. + +**Ian Lopshire:** Oh, my bad. Q1, yeah. But still. The quarter is almost over. + +**Jerod Santo:** I confused him by saying "Every day a Q goes by." \[laughter\] + +**Kris Brandow:** Q85. + +**Jerod Santo:** Yeah, settle down, Ian. Settle down. You don't have to freak out that much. + +**Ian Lopshire:** But Q1 is almost over, and that's scary. But here we are. + +**Jerod Santo:** But here we are. And it's Go Time. So before we get into my questions, I've also gathered a few questions from Go Time audience and Twitter folks about what they're curious about the Go programming language. I wonder, from each of you maybe - don't go deep, don't tell us about your birth and stuff; but getting into Go and kind of the language that you came from, or the ecosystem that you came from when you got into Go, and maybe what were some of your trepidations or things that you were curious about prior to being a full-time gopher. A Go Time gopher also, but just... Writing Go on a regular basis. Ian, you're a guest, let's start with you. Tell us about your origin with Go. + +**Ian Lopshire:** Yeah, so pre-Go I was writing mostly PHP and JavaScript, and the company I actually worked for at the time kind of decided they wanted to go with more compiled languages, so we had a choice of C\# or Go. And my team went with Go. I think the first questions and issues I had when I started was around 1) how to structure things, does everything go in one file, do I make a thousand packages...? Kind of that bit. And the second one was a lot around JSON and deserialization, the io readers. Coming from JavaScript, that was very confusing for the first couple weeks. + +**Jerod Santo:** Cool. Kris, what about you? I think you hit on a couple of things actually that are in my list, so I'm right there with you. Kris, go ahead. + +**Kris Brandow:** Interestingly, when I started my career I almost started it with Go. I kind of looked at it, got really confused actually by the := syntax, and ran away... + +**Jerod Santo:** Also in my list. Okay, keep going... + +**Kris Brandow:** So I started with PHP, and then... Funny little thing about me and Sam Boyer - we both lived in the Drupal community for a while; and I went to a meetup where I met him once, and we were talking, and I was like "There's something I don't like PHP." And he was like, "Complexity. You should try Go. It's all about simplicity." And that is how I started off looking into Go. And then I just kind of went from there, and I actually convinced a company I was working at to let me build projects in Go. And I guess the first things that I ran into were like -- I was so excited about concurrency, I started putting channels and goroutines everywhere, and quickly learned how terrible of an idea that is. But then kind of stuck with the language for that simplicity. + +**Jerod Santo:** Yeah. Very cool. So to give a little bit of my background - I've started off in Perl back in the early 2000's, went from Perl to Ruby, also learned a little bit of C, but just enough to be dangerous. I never actually worked daily in C. I did Ruby for many years, added JavaScript, of course, because I've been in web development most of my life... From there, I started writing Elixir; I've written Elixir quite a bit over the last 6-7 years. That's probably my primary language, plus JavaScript, because again, web development... And my experience with Go has been two small programs. + +\[08:11\] Back when I did client work, I had a client who needed one endpoint, and they needed it to be really fast... And it was just like a JSON endpoint that took a few arguments and spit out different responses. I can't remember the exact thing. And Go -- this was pre-1.0. Go was very new, and I always liked the new, shiny stuff, so I gave it a go back then, and wrote that endpoint in Go. And it was probably like 80 to 100 lines of code. It worked great; I thought it was really cool, and never had another use case for a long time, because everything from there was -- Ruby on Rails apps was the primary thing I was doing. So I picked it up and dropped it then. + +And then recently - you know the Go Panic game show, and the... What's the other one called? On JS Party we call it Frontend Feud... Gophers Say. That's what we call it. Gophers Say has that in-browser UI with the scores, and the faces, and all the questions, and these things... And that was static HTML for a while. Well eventually I wanted the same thing to drive the JS Party side, and the Go Time side, with different stuff... Anyways, I wrote a little Go web server, that's again probably around a hundred lines. So this was like last fall. I picked up, I had a pretty good time with that. + +So I've written probably around 200 lines with Go. So I'm not completely newb, but almost completely newb. But I'm very interested in it because I'm interested in building a Changelog command line for fun and for interest, and I think it's a great language for that reason, for distributing command line appliations. It seems like it's simple for that purpose, with universal binaries and deployment and stuff, not having to have shared libraries everywhere... So that's why I'm curious about it. + +Also, just producing Go Time for many years, I know a lot about it at a very shallow level, which makes it less intimidating than other languages, like Rust, for example, which I know there's a lot of interest in that, and a lot of comparisons, and a lot of verticals that both languages play in. So it would be another contender. But I'm very interested in Go. + +That being said, there's stuff in Go where as a person who hasn't really written much of it, I'm just like "I don't get it", or why... A lot of the questions are "Why? Why does it work this way?" And the first one is slices and arrays. + +In most other languages there's an array, or a list, or something... But Go has these two, and one is like a substitute -- I don't really understand it that much. It seems like everybody uses slices all the time. Can you guys explain to me what's a slice, what's an array, why are they different, and when should I use each one? + +**Kris Brandow:** Sure. For the first question there, "What's a slice and what's an array?" - an array is a fixed-size group of data, I guess you could say. I'm trying to say it without saying the word "array" or "slice" here. And iterable is also a bad word, but... It's just like a slab of memory that you can have, that's of fixed size. So you can have an array of like five integers, or an array of ten strings. But they are fixed size, so you can't add another string there. So you can't make that array that's size 10 into an array of size 11. That's a different slab of memory that you would have to go get. + +Whereas slices are much like arrays, in that you can have these kind of enumerations of things in them, but they are resizable. So you can go from a slice that is size 10 to a slice of size 11 by adding something to it. So it's kind of like the base level of the difference between a slice and an array. + +\[11:53\] You can also think of it as like a slice is a pointer to an array, and the programming language just does all of the magic for you when you need to get a larger array kind of just giving it to you and handling all of the copying of what was in the old array into the new array, and all of that. + +**Jerod Santo:** Okay. So arrays sound more like what we used to do in C, where you pre-allocate slots, or memory for a fixed length. And if you wanna go beyond that, obviously, you walk off the side of an array, now you're in overflow land, and security problems. But you could then have a pointer to that... There's things that you can do in order to allocate new space, and copy things over... Whereas a slice -- I think I'm with you. Ian, do you wanna amend or help anything Kris said? + +**Ian Lopshire:** I mean, Kris pretty much covered it. The way I think about it - a slice is what an array should be. + +**Jerod Santo:** That's why -- it goes to my next question, why are there two separate data structures, or I guess concepts? Are they just concepts, or are they data structures? I don't know. + +**Ian Lopshire:** They're different data structures. + +**Jerod Santo:** Okay. + +**Ian Lopshire:** You never really see the underlying data structures of a slice. It's kind of opaque. + +**Jerod Santo:** Okay, fair enough. They're different data structures. Why are there two? Go is about simplicity. So like in JavaScript there's an array; that's what you've got. But in Go there's two... Sorta. + +**Kris Brandow:** I would take a guess... This is probably language lore, which someone can probably answer from the community out there, but I would say that they probably started with just arrays, and then realized these aren't very friendly to use... But in the way that Go is simple, it has tools that are purpose-built for one job. When you need an array, you really need an array; you really want that fixed-sized thing, you want the property of that fixed-size thing. So they couldn't just make an array into a slice, really. + +I also assume it has something to do with typing in the language... Because when you have a slice, and the way that you add things to a slice, if you use this function called append, and you kind of put a slice in, and you get a slice back out, but the types of those slices need to match each other for that function to work. + +I imagine it would have been more complicated to do that with arrays, because in Go, in addition to -- you can't kind of resize an array. Arrays of different sizes are different types; so an array of five strings is a different type from an array of six strings. It's a different type of an array if you have seven strings. Whereas a slice of strings is a slice of strings, no matter the capacity of that slice, what it actually is, at the end of the day. + +**Jerod Santo:** Gotcha. So just use slices, pretty much, unless you know better. + +**Kris Brandow:** Yeah. For beginners there's very few reasons why you would want to use an array. Array is definitely one of those advanced user tools. + +**Jerod Santo:** Okay. I think as a beginner you come to it from a different language and you're like "What I need here is an array", because that's what it's called elsewhere. So then you go searching for arrays, and then you find something that seems more complex. Not complex, but lower-level, perhaps, than what you're used to. And so then you're like "Okay, slices." And then you start to wonder... + +**Kris Brandow:** They probably should have named it array and static array. An array would be equivalent to what a slice is now, and a static array would be what we have in an array now. + +**Jerod Santo:** I think slice in general is just not in other languages as a data structure, and so it's somewhat unique to Go, at least in my experience, where I'm coming from. + +**Kris Brandow:** Yeah. + +**Jerod Santo:** See how I hedged that quite a bit at the end...? \[laughs\] I was thinking, "There's lots of languages out there, Jerod. Maybe it's common elsewhere." Okay. Let's stick with some syntax thing... So the equal sign and the colon equals thing... WTF? I don't know. Ian, let's let you go first on this one. + +**Ian Lopshire:** The quick assignment - I think it exists so that you don't have to define the type of a variable. Quick assignment I can just say "Whatever this type I'm assigning to this is the type of this variable. Otherwise you have to define a type. So I think it just adds some expression. It really is. + +**Jerod Santo:** Is it syntax sugar then? So that's the colon equals, it's the quick assignment. + +**Ian Lopshire:** Do you wanna add something to that, Kris? I really just think it's for ease of writing. + +**Jerod Santo:** So with the quick assignment you do not have to explicitly declare the type of the variable. + +**Kris Brandow:** \[16:02\] Yeah. It also gives the language more of a feel of like a dynamically typed language as well... Because there are some specific things in Go that are kind of type-ambiguous until you say what the type is. Constant numbers are an example of this; they're the special number type. The language will eventually figure out what that type is when you get more specific with it. So by not declaring what the type is upfront, you get a little bit more flexibility to decide what this thing is going to be at a later date, which you couldn't do if you had to declare it right upfront, like you do with var and the regular equal sign. + +**Jerod Santo:** Okay. So coming from dynamic languages mostly, I would just wanna use quick assignment all the time, because I would always wanna defer that until later. But I don't see it used all the time. So there are times where var or the equals, the non-quick, the slow assignment is just preferable. Maybe you just know right upfront, so you might as well declare it. + +**Kris Brandow:** Yeah. I might be wrong about this, but I don't think you can use the quick assignment outside of a function. So if you're just like in the global space, you have to use var. + +**Jerod Santo:** I think I hit that, and I wondered why. + +**Kris Brandow:** And then also, if you're trying to assign something to a property of a struct, you can't use quick assignment; even if you're doing multiple assignment, you have to use regular equals there. So you kind of have to declare the type -- you have to use var and declare types for that. + +I would also say that a lot of the time when I don't use quick assignment, it's because I want the type there for documentation purposes; so leaning more into the -- knowing the type of this upfront makes it easier for people to read and understand "Oh, this is going to be this specific thing", instead of something more ambiguous. + +**Ian Lopshire:** I kind of think of it as creating the variable, then assigning it. So if you're inside a new scope, the quick assignment will shadow variables outside of that scope. So it's really redefining the variable, and then assigning. + +**Jerod Santo:** Say that again in other words. I think I'm with you, but I want you to say it in other words. + +**Ian Lopshire:** So say I'm writing a closure, right? Or I have something coming in. + +**Jerod Santo:** Right. + +**Ian Lopshire:** If I do quick assignment inside of that, like the next scope, it's gonna shadow the variable. That was a horrible example. + +**Jerod Santo:** It's gonna overwrite what was outside of the closure. + +**Ian Lopshire:** It's not gonna overwrite it, it's going to define a new variable, and inside of that scope you can only reference... + +**Jerod Santo:** The new one. + +**Ian Lopshire:** The new one. + +**Jerod Santo:** Oh, that's not good. + +**Ian Lopshire:** I mean, it's desirable sometimes, but if you use the equal sign, it's gonna set the value outside of the scope. + +**Jerod Santo:** Okay. When is shadowing desirable? This might be tough, because you might have pre-thought of an example... But when would shadowing be desirable? To me it sounds like just a cesspool for bugs. + +**Ian Lopshire:** So this is kind of a weird case, but inside of a for loop sometimes you do wanna shadow, because the way the for loops work is it'll save the reference outside of the loop. So as the loop goes through, your variable is gonna change. But if you do a quick inside of it, you can copy the value into the new scope. + +**Jerod Santo:** Okay. + +**Ian Lopshire:** Do you know how to say that better, Kris? I don't know how to say it without showing a picture. + +**Jerod Santo:** I think I tracked it, but yeah, Kris, go ahead. + +**Kris Brandow:** Yeah, it's basically -- you know, if you have a good variable name and you want to reuse it, but you don't wanna affect things outside of the loop or the function that you're using... + +**Jerod Santo:** Okay... That's a good way of saying it. So you're appealing to my desire not to think of another name now. + +**Ian Lopshire:** You'll see i=i a lot. + +**Jerod Santo:** I see. + +**Ian Lopshire:** Like, i=, :=. + +**Jerod Santo:** Gotcha. + +**Kris Brandow:** The thing that I think a lot of new people run into when they first start using goroutines and loops - if when you get into the loop you don't reassign the variable to something that is more locally scoped to that, and you try to use that inside of the goroutine, all of the goroutines will basically see the last value of the loop. If you have a loop that has numbers one through ten, and you iterate through and start a bunch of goroutines, all the goroutines will see ten as the value they're operating off of, if you don't reassign inside of that loop. + +**Jerod Santo:** Gotcha. + +**Kris Brandow:** \[20:07\] So there are some uses around that. This always burns me when I'm doing unit testing; that's one of the areas where I start using goroutines and closures, and I'm like "Why isn't this working properly?" It's like, "Oh, right, I have to do this t:=t sort of thing" in order to redefine that variable within that iteration loop so it gets the goroutine properly. + +There are other ways to handle that as well, that are a bit more explicit, but that's definitely where I think it gets -- shadowing is used the most often to make it so that it's not a bug, and actually something that's helpful to you. + +**Jerod Santo:** Yeah. So it sounds like the implicit nature there seems like almost not a Go way of doing things, but I understand for (I guess) syntactic simplicity. But the implicit... Maybe it's an idiom that once you get into Go, you're just like -- everybody knows "Oh, this is a quick assignment, and so this is purposeful shadowing" kind of a thing, or vice versa. Is the quick assignment the shadower, or is the explicit assignment the shadower. + +**Kris Brandow:** Quick assignment is. I mean, you could technically do shadowing with either, but the quick assignment is where people usually get tripped up. + +**Jerod Santo:** So when you guys who are seasoned Gophers see a quick assignment inside of a for loop, for instance, or maybe with goroutines, you just immediately know what's going on there, you're like "Oh, I know", because it's i and i, or whatever, why that's there? + +**Kris Brandow:** Yeah. There's sometimes where it's just like -- if I'm looking at some code and all of a sudden I see... Especially in testing, when it's t:=t, I'll be like "Oh yeah, that has to be there." Whereas I think other newer people would be like "What's that doing? I don't understand." + +**Jerod Santo:** Right, right, right. + +**Kris Brandow:** So anyways when I see that. But that doesn't make us immune from shadowing bugs. I've written my fair amount of shadowing bugs where I've spent a lot of time debugging and being like "Why is this error nil?" And it's like, "Oh, because I went into this different scope and I reassigned err, so in the larger scope it's still nil." That's happened to me too many times. + +**Ian Lopshire:** Yeah, I think especially in error handling shadowing happens a lot, and it's just part of the language, really. + +**Jerod Santo:** Part of the language. Fair enough. So by the way, we have people in the \#GoTimeFM chat sounding off a little bit with some definitions and whatnot, so that's cool. Dillan Bork says "Array is fixed size list of items. Slice is a "pointer" to some subset of an array." So there's a nice, simple explainer there. If you all are looking up definitions or helping us wtih examples... Sometimes it's hard to think on your feet, and we can have them read off afterwards. + +If you're not hanging out in the \#GoTimeFM channel, what is wrong?! We're all hanging out in there live during the show; you can participate and be part of the fun. So hop into \#GoTimeFM of the Gopher Slack during our shows. + +Okay, let's go on to a bigger question. These are kind of small language things. What's the state of the art on dependency management? How do I use other people's code in the idiomatic, best way today? I know there's like a history here, so I'm not ignorant to any of the history, but I just wanna know, just starting today, if I'm gonna start writing Go, what's the way that I should do it? + +**Ian Lopshire:** I think if you're starting today, a hundred percent embrace modules and just use it. It works well, especially for new projects. + +**Jerod Santo:** Okay. Is that what everyone is doing? + +**Kris Brandow:** Yeah, modules is the way (I guess) of the future for right now... + +**Jerod Santo:** The way of the future... \[laughs\] Okay. + +**Kris Brandow:** So if you're getting started with a brand new project -- I mean, for all my new stuff, I use modules as well. So yeah, it's just like, okay, get some modules... All of the commands are the same as they used to be; so you still use go get to get dependencies, and all that sort of stuff. + +**Jerod Santo:** Right. + +**Kris Brandow:** But yeah, when you open up a new folder, run go.mod in it with the name of the project, what you want to name your module or your project... And then you can start just doing go get whatever to go download some stuff, and off to the races. + +**Jerod Santo:** Okay. That was an easy one. + +**Break:** \[24:01\] + +**Jerod Santo:** So I've said "idiomatic Go" a couple of times. I know that's a term that perhaps is falling in or out of favor -- or the term "idiomatic", I've heard either unpopular opinions, or maybe just conversations about "Let's not use that term." So whatever term you wanna put in front of Go that means kind of normative... I don't know. The way people write code. + +Aside from using the formatter, which seems like obviously would normalize a bunch of code into some sort of idiomatic fashion that everybody thinks is good, what are some examples of common Go idioms, or things that you should be doing, that most Gophers do... I don't know if you guys have a PEP or something like they do in Python, or it's like -- is there a Go way, or you don't need it so much because there's kind of just one way? + +I think with looping, I kind of like the fact that there's really just one way. It's cool. But surely, there's other ways of doing things. Do you guys have some examples of "This is kind of a Go way of coding"? Or I've heard like writing Java in your Go, or writing Ruby in your Go. What are some things that are "idiomatic" Go? + +**Kris Brandow:** I think a lot of the way that we name things, especially around capitalizations... Like, one of the big, big differences between Go and other languages is how ID is capitalized. So if you have something that's like an identifier, it's capital I, capital D, and that is a very solid Go idiom. And I think the linters might yell at you now. If they don't, I hope they do. + +**Jerod Santo:** Oh, really? + +**Kris Brandow:** But yeah, how you capitalize initialisms, and also constants in Go aren't supposed to be all caps. They're supposed to just look like regular variables. So I guess the idioms that always come to mind for me are those sorts of things, like how does Go look from that kind of perspective. Obviously, things also like we use camel case and not snake case. + +**Jerod Santo:** Sure. What about globals? Do Gophers like globals? + +**Ian Lopshire:** I think globals are discouraged. We do have the init function, so they're supported. I think they're discouraged in the community. + +**Jerod Santo:** \[laughs\] You're discouraging me from using them, Ian... + +**Ian Lopshire:** I would discourage you from using globals, yes. I mean, of course there are times when you want to use globals, and it makes sense... But outside the main package, I would almost say "Don't use them." + +**Kris Brandow:** Yeah. That's what I was gonna say. You can use them in main and you're alright, but certainly, if you have your own library -- if you're providing a library, you shouldn't use them. And certainly not use them if they're public. There's some cases for private globals, but public globals just get you into all sorts of messes. + +**Jerod Santo:** \[28:07\] Okay. So one thing Mat Ryer has told me, a pattern that he observes is like his main function doesn't do very much. It's basically like calling the rest of his program. Is that Mat Ryerism, or is that like a Gopher good idea? The programs I wrote, main - everything's in there, except for like my structs; and some functions are outside, but they get called in there... But like, it's the gist of things. Because it's a very simple program. But is that something that is discouraged? Maybe like skinny main is kind of like an idiom maybe? + +**Ian Lopshire:** I think that is... There's different forms of this you see in the community. Sometimes you see, like, it's literally one line in the main, and then the rest in a different package. But I think the idea of keeping any kind of business logic, any of that out of main is pretty ubiquitous in the community. + +**Kris Brandow:** I feel like it's maybe not an idiom. I feel like it's still somewhat contentious, because... I hate this pattern. I really don't like this pattern much at all. + +**Jerod Santo:** \[laughs\] Here comes Kris now with the honesty. See how he eased into it, he's like, "Nah I'm done. I hate this pattern." + +**Kris Brandow:** I get the intention of it. We wanna make everything testable, but I feel like that kind of ignores a lot of the other things that make it testable at the end of the day. Ripping all of the guts out of the main function, so you could put in another function you can run and go test... It doesn't fix a problem with, say, environment variables, which could still be kind of annoying. Or other globals that can sneak in at the end of the day. So I understand the intent of saying "Well, we should have this smaller main function that you can then go and test more easily", but I think actually doing that misses the point a lot of the time. Misses the thing that we're actually trying to tell people, which is like "Don't use globals. Write smaller functions, write smaller pieces that compose together better, that you can test more easily." + +And I also think it discourages people from kind of running the whole binary and putting that under test, and figuring out how to test the application as a whole maybe outside of the Go testing library. It's kind of like a way of escaping around and saying "Well, I don't have to do this anymore, because I can run all of my test functions on this pseudo main that I've created." + +**Jerod Santo:** Hm. Ian, your response. + +**Ian Lopshire:** So the practice that I generally use is I do end up with pretty big main files, but all it does is set up dependencies. I understand what Kris is saying; you probably should be doing some testing outside of, like, as a whole, but I think just keeping the main small probably does encourage better testable code, even if it leaves out some of the pieces. It's like, we shouldn't not do it because it doesn't solve the problem completely. It gets us part of the way there. + +**Jerod Santo:** Right. + +**Kris Brandow:** I'm gonna counterpoint that for a second... + +**Jerod Santo:** Please do... Now we're having fun! + +**Kris Brandow:** I mean, I see this argument all the time, of like "It gets us a step in the right direction." But I really do -- like, a lot of the codebases I've seen that enact this skinny main thing do really just forget about setting up proper configuration, having configuration be sane, having the bootstrapping code look good, having an actual application kernel... These are the things that we actually want at the end of the day. We want the structure of how our applications boot up to be really nice. But instead, what I feel like we've got is - okay, now everybody just shoves all of the gross code they were putting in main into this other function that is effectively the same thing, except now you can return an error from it, and skipping out on all that good stuff that we actually want. So I feel like it's a distraction, and when you have to go back and actually sit there with that main function, and perhaps not be able to test it with just unit tests, and have to test it in like a real way, that forces you to have to do things in a way that you wouldn't otherwise. + +It's one of those sneaky things that's like "Well, this feels like progress", but it's actually progress leading you to like a dead end in the maze, unless we're gonna try and scale the walls of the labyrinth we've gotta turn around and figure out a different path forward. + +**Jerod Santo:** I love the analogy. + +**Ian Lopshire:** I'll give you that, that it does provide kind of a way out, and can let us not do things we should be doing... So I'll give you that one. + +**Jerod Santo:** \[31:59\] Right. Some of that's like -- you know that old saying about laws, like they keep the honest people honest. Some of this is like these idioms or these patterns will keep the people who are gonna do good tests doing good tests. But the one who wasn't going to anyway, whether they shove it all in main, or put it into a different function immediately - it's not gonna change who they are, or the way they code. + +So I can see that, where it's like "This is a good pattern. It helps you do this thing that..." The problem is we're not doing that thing anyways. So it doesn't fix that particular problem. + +**Kris Brandow:** I would actually say there too that's kind of what differentiates what I would consider an idiom in Go from just like a practice that people are trying to make, or a best practice. I feel like most of the idioms - really all of the idioms we have are these deeply-nuanced things. Sometimes we can give like a little quick -- like the Go Proverbs, which I think are great, where it's like this little quick sentence, but if you actually start pulling it apart, it gets more and more complex, and there's just more and more pieces that you wanna pull apart to get a good understanding of it. And I think that when you have things like "Just have a skinny main to make it more testable" and it ends there, it's like "Well, where is the depth to that? How do I keep going when I encounter something where this doesn't fit properly, that the idiom still works?" It doesn't have that same amount of depth to it... Whereas I think a lot of the other idioms do wind up having that depth to them. + +**Jerod Santo:** Hm. So if I wanted to learn a bunch of Go idioms, would you suggest reading Go Proverbs? Would that be the suggestion? + +**Kris Brandow:** Yeah, yeah. I think Go Proverbs still apply. I think some people think that they're dated, but I think that they are pretty good. I think also code review comments, which is this thing that's in the Go GitHub wiki, is actually pretty good. Those are also pretty much like idioms, ish... There's some idioms sprinkled throughout there. But I think those are two good sources for idioms of a language. + +**Jerod Santo:** Nice. + +**Ian Lopshire:** The Go Proverbs were actually something that kind of drew me into Go. The first time I read all of those, I was like, "Wow. These all make a lot of sense." I don't think they're dated. + +**Jerod Santo:** They resonated with you. + +**Ian Lopshire:** Yeah, definitely. + +**Jerod Santo:** Okay, so there's a good litmus test perhaps for the Go-curious. If you're wondering if Go would speak your language, or speak to you, is go read the Go Proverbs and see if you agree, or if they're saying things that you think make sense; or maybe it's unpalatable for you, then maybe look elsewhere. So that's good advice. + +Let's talk about web apps. This question actually came in from Twitter. "Is Go and the ecosystem in place where it can compete with Rails/Laravel/Django for dynamic websites?" + +**Ian Lopshire:** I don't think so. + +**Jerod Santo:** Okay... Honesty. + +**Ian Lopshire:** I mean, that's a style of development that I don't think Go is particularly suited for; those big, monolithic, server-side-rendered apps. I'm just not sure it makes sense in the frame of Go. + +**Jerod Santo:** What is it about the language or the ecosystem that makes you say that? Is it the strong typing, or the fact they're dealing with dynamic, user-generated content, or what is it? + +**Ian Lopshire:** I think it's a lot like the distrust of magic. If you think about Rails or something like that, or Wire... Is that the new Rails? Like, dynamic frontend bit... + +**Jerod Santo:** Hotwire? + +**Ian Lopshire:** Yeah, Hotwire. All of that just does things, and you have no idea what it's doing. And I think that's not tolerated well in the Go community. + +**Jerod Santo:** So let me translate that and see if I'm picking up what you're saying here... Because dynamic web apps, that have a lot of user-generated content or input, building those at scale - I don't mean scale of users, but breadth of surface area; lots of forms, lots of pages, lots of what-have-yous - requires from a framework perhaps a lot of either code generation, or reflection-based stuff to make you not have to write a bunch of code yourself every time you wire up a form... In Go - it's not against the code gen at least; it seems like. But it's against the meta programming stuff that Rails and these other ones use, in order to cut down on your scaffolding in your coding. Is that kind of what you're saying, or did I just say a bunch of stuff that you didn't say? + +**Ian Lopshire:** \[36:06\] No, that is what I'm saying. + +**Jerod Santo:** Okay. + +**Ian Lopshire:** As you say it back, I'm not totally sure I agree with what I'm saying... \[laughter\] + +**Jerod Santo:** That was kind of the purpose of the practice, to see if we were -- + +**Ian Lopshire:** I think there is something in the Go community that does dislike the idea of these big frameworks, these big things... And I'm not sure if it's reasonable. + +**Jerod Santo:** Like a full-service framework. + +**Ian Lopshire:** Yeah. I'm not sure it's reasonable, but it does exist. I don't think we can deny, like, the skepticism if that exists. + +**Jerod Santo:** Okay. + +**Kris Brandow:** I feel like we're comparing apples to oranges a little bit, because it's like - you know, Ruby, Django... These are not languages. The language there is Ruby, Python, whatever. I think Go is a good language for building these things if you actually have the time and patience to actually ramp up and actually build it yourself. I think if you just want the experience of what Rails gives you, and have this out-of-the-box, "Boom, I can get off to the races, build my stuff, and then I'm done, I can move on", then no, Go isn't gonna be good for that. + +But if you're actually trying to either build something so you can either understand how these things get put together, or if you're like a bigger organization that has the engineering capacity to own something like this, then I think it can actually be very good. Because I think you have the knowledge of how the thing works, and you wind up having a lot less code that you depend on, that you don't own at the end of the day. + +So I think the main argument against using Go for these types of things is like "Well, the getting started period is much longer", because you have to build more of the stuff yourself. And I think that's kind of where I get a little annoyed with that argument, because I personally do not think that that's how we should look at software engineering. I'm very annoyed that we continue to look at things as the "How fast can you get going from the beginning?" and not "How do you actually build something you can maintain over the long-run?" Because that's obviously more important. Not like "I just built something today", but like "Oh, three years, four years down the road this thing is still something that we wanna use, we wanna add to, we can extend" and all of that. + +**Jerod Santo:** I agree with you most of the time. Sometimes you do want to just build fast and to test your business idea, and not your software system. So to test an idea quickly. That's why I think Rails was so popular in startup land, was because it was like "We don't even know if this business is gonna be here in six months, so why do I have to build and architect a system that's gonna last a decade, if my business isn't gonna last six months?" + +Now, the idea there, I think, is like once you get your product-market fit, or whatever the startup guys say, then it's like "Okay, let's re-architect this sucker now in like the ways that Kris would build it. Let's build it to last." And I think a lot of times that never happens, because the business is taking off and you're just trying to keep the servers up, or whatever happens... In that step, prototypes are supposed to be thrown away. And we never do, we just turn them into businesses. + +**Kris Brandow:** Yeah. + +**Jerod Santo:** So I definitely understand that. What I'm trying to get at - I have never quite understood, is the lack of a Rails or a Django coming out of the Go community because Gophers don't like that? Just speaking very broadly. Or is it because Go as a language isn't well-suited for that? And I don't know the answer to that. + +**Kris Brandow:** I think it's because there's not a need. Rails exists, Django exists, PHP, Drupal, and Symfony... All of these things already exist, and have large communities, and large support around them... So in order for something like that to exist inside of Go, we have to invest a lot of community energy and time into rebuilding all of that, and to gaining all of that, to get a very small portion of the pie at the end of the day. So I think that people that are very comfortable with those sorts of things want to do that. + +I think Go as a language really does attract people that want to do this kind of lower-level, or different type of work, at the end of the day. I think people that want to go and explore Go - they're already onto the world of single-page web applications and APIs, and they don't want these big monolithic stacks as much anymore. So I think by nature of what Go is really good at, and by how crowded the field actually is, I think by the time you get o the point of thinking about using Go, you already have a different problem that you're trying to solve than what you would solve with Rails, or with Django, or with Drupal, or with any of these other things. + +**Jerod Santo:** \[40:19\] Well said. Ian, anything to add? + +**Ian Lopshire:** No. + +**Jerod Santo:** You're nodding along in agreement. + +**Ian Lopshire:** I think that covers it. The idea that those already exist... Like, yeah, why would we build another one for a 1% market share? + +**Jerod Santo:** Right. Maybe because you want a web app, and you love Go. But I guess in that case then you're gonna handroll a bunch of stuff, like Kris talked about, and you're gonna pick each library out, and you're gonna build up a thing... Which you can do in these other systems. I mean, Sinatra is a thing inside the Ruby land that very much has more of a Go philosophy. So it's not like you can't. One of the follow-up questions to that is "Does Go do websites?" And it's like "Well, of course you can do websites with Go." It's just you have to build all the parts of your website; I mean, there's templating engines, and stuff like that, and routers... There's all sorts of things. + +**Kris Brandow:** I think that's one of the things that actually makes Go quite good... I mean, given I did just say all of that, I think that if there was a reason for us to build something like a Rails in Go, I think it could be incredibly interesting, because we do have a lot of stuff built into the standard library that gets you halfway there. + +**Jerod Santo:** Yeah, totally. + +**Kris Brandow:** We have a database library that's built into the language, we have a good templating system that actually properly escapes HTML and JavaScript and those sorts of things, built right into the language itself. So I think the pieces are all there for it, but I just think-- + +**Jerod Santo:** Yeah. The appetite. + +**Kris Brandow:** Yeah, there hasn't been someone that wants to go put the effort into building that sort of thing. + +**Jerod Santo:** So Mike Dodson in the chat is expanding on what Kris said. He says "You end up using one of the frameworks and quickly run into issues, and find it's actually as easy to write your own, versus learning a third-party framework." I think we've all been there; especially the more magic that pours on, the harder it is to understand what's going on under the hood. + +**Kris Brandow:** I think this is why a lot of the early API frameworks that existed in Go -- because that was something that was big for a while, with Martini and a bunch of these other ones; they've died out for a large part, and I think it's exactly because of that... Because you realize pretty quickly it's just like, "I have 50,000-100,000 lines of dependency that I just do not need. So I can just go write this myself." + +**Jerod Santo:** Well, having a batteries-included standard library is a huge asset to any language and ecosystem, so... Well put. Alright, getting close on time; let's ask -- I've got a couple other questions here. So look on the other side of the fence; you know, wanderlust. Do you guys ever have any wanderlust in Go? Do you have any features in other languages that you're jealous of? So as full-time Gophers, look over there and see what they're doing in Rust land, or Elixir land, or TypeScript land, and think "Oh, I would want that, but I don't have it." + +**Ian Lopshire:** Through the years I've run into some things that don't have great support in Go... I don't know if you've ever worked with SOAP; like XML... + +**Jerod Santo:** I try not to. + +**Ian Lopshire:** Yeah. If you try to do that in Go, you're just gonna have a bad time. And I would like to be able to stay in the language I like, and build services that interact with this, but I won't. I've tried it, and I refuse... We'll use something else for SOAP communication. + +**Jerod Santo:** Okay. What do you usually switch to? + +**Ian Lopshire:** C\#. + +**Jerod Santo:** Okay. + +**Ian Lopshire:** Yeah. So I guess the XML support in general is not the best. + +**Jerod Santo:** Okay. Kris? + +**Kris Brandow:** You know what - I think a lot of things that I want in Go are things that also don't exist in other languages... + +**Jerod Santo:** Hah! You're a dreamer... + +**Kris Brandow:** Like, I want better APIs, like API-building tools. I wanna be able to build HyperMedia APIs much more easily, but I don't think there are any languages that have particularly good tooling for that right now. Same thing with database access; I want better ways of interacting and manipulating data, but once again, I don't think there are any languages that really do that well. + +**Jerod Santo:** You want the ability to build your own DSLs, or what do you mean by better -- let's take the HyperMedia one; or let's take the database one, because more people are probably familiar with databases. Like, when you say "Access to database better" - I don't know what better looks like unless you tell me... Draw me a picture. + +**Kris Brandow:** \[44:06\] There was this talk that one of the previous - I think he was one of the previous - Go team members gave on how great SQLite is, and how that's kind of the only database you need... That's the direction that I would like things to do. I like want my data to just be there, but also be distributed. I don't wanna have to worry about how my data is stored on disk. I just wanna be able to access it and manipulate it how I want, in my language. I don't wanna be writing lots of strings... + +I think part of my gripe here with SQL as a whole - I don't like the model of that for interacting with data; and everything is built on top of it, which is why I don't think any languages have good solutions to this. So I'd like to see something like that get better. And I think the way Go is structured, that's just something we could -- like, we do have this rather unique way of doing code generation, where you generate code not at the time of the... Like, you get it from a dependency, but when you actually build it to deploy it, I think that gives you some unique aspects of how you could actually use code generation... Because you kind of generate things much earlier in the pipeline, and the way that it all works in Go is a bit closer to being able to do something like take something like a DSL or something closer to a DSL and pre-compile it in your application before it kind of gets sent out. + +So I think there's interesting things around all of that, but once again, it's just like the paradigm, how the industry works is we all want our applications to be stateless, so we're all shipping data out, we're communicating with something else that holds all the data, at the end of the day. + +So it's stuff like that I'd like to see more of, and see more interesting ways of doing it, but I don't know of any mainstream languages that have those features. And of course, I think package management is something that I would like to see be good, and I don't think any language has done it well as of yet. I think people have done it okay... I think even after all the dust is settled, modules are okay, but then you hit those edge cases and it's like going through hell. It's just real rough. So I'd love to see that be better. But once again, I don't know any other languages that have that better. + +**Jerod Santo:** Yeah. I've heard people speak highly of Rust's crates system, but I don't know what sets it apart from others, or anything about it. So I don't know if that's just people speaking highly because that's what they do, or if they've actually cracked some sort of nut over there, in the Rust ecosystem, that makes it particularly better than some of the other ones out there. + +Okay, so some of the things that I've seen in other languages, where I'm like "If I was to use Go, I might miss this" - I like the idea that for is the only way to loop, but I also love all of the FP functions, for like iteration, and map, and reduce, and select, and filter... For me, those are very productive ways of pipelining data. And I'm wondering - maybe those are available as libraries, or anything... But it seems like I would miss those things. Do you guys ever miss those things? I'm sure you're familiar with them in other languages. + +**Ian Lopshire:** I miss iterators all the time. I find myself building them a lot. With generics, I'm hoping we do kind of settle on a reasonable iteration - maybe not interface, but way of doing it. But I do miss that. + +**Jerod Santo:** Okay. + +**Kris Brandow:** I think I miss that too, but I think I've also just handrolled it enough times that I'm just like "Nah, I can just write it myself, so whatever." + +**Jerod Santo:** You have your own little utilities package that you carry around, or do you just code if up each time? + +**Kris Brandow:** Nah, I just code it up each time for what I need. + +**Jerod Santo:** Nice. + +**Kris Brandow:** Once you design a database driver, you spend a lot of time thinking about how to design an iterator really well, and then it just like sticks in your head forever. + +**Jerod Santo:** Gotcha. I did see a package recently that was trying to recreate JavaScript's Lodash, which is a lot of functional facilities for JavaScript, with Go's generics. It's a very fresh project, once generics shipped. I don't know if it's any good, or anything about it, except for I think people are gonna start working on that particular problem. + +Last one, and then we'll get to wrapping up here... How long did it take you to get over if err!= nil being all over the place? + +**Ian Lopshire:** \[48:06\] I've never hated it. + +**Kris Brandow:** Yeah, I don't think it's ever bothered me. + +**Jerod Santo:** Never? + +**Ian Lopshire:** I like it. + +**Jerod Santo:** You like it? + +**Ian Lopshire:** I hate try/catch. + +**Jerod Santo:** You always liked it? + +**Ian Lopshire:** I did. Like, the first -- + +**Jerod Santo:** The first time you saw it, you're like "This is it." + +**Ian Lopshire:** "This makes sense." + +**Jerod Santo:** Okay. + +**Ian Lopshire:** I've always hated try/catch, because it's just so disconnected from where your error happens. You could catch something 12 nested levels deep, and just have no idea what's going on... So the idea that you just have to handle errors right where they happen - I just immediately though "Oh yeah, this is how it should be done." + +**Kris Brandow:** I think for me it's partly how I learned Go. I did not learn it trying to change some project from some other language into Go. I just learned Go to learn Go. I absorbed a lot of Rob Pike's talks, and things... And just like the idea that "No, errors aren't special; they're just other values." They have a semantically different meaning, but they're not actually different from anything else. It fundamentally changed the way that I structured my code, and I started to see a lot of -- if you keep writing if err!= nil a lot, you've done something terribly wrong in your application and you need to go back and redesign that code, because that code is wrong. + +I think once I saw that, and once I started leaning into that as a signal that I should be getting from it, I wound up rewriting code and having it wind up be a lot better. So while sometimes it's annoying when you're prototyping and just trying to get something done, and you just write it a bunch of times, it's annoying. I think it actually winds up making my code much better, so I haven't found it to be something that I don't like, + +**Jerod Santo:** So mostly what I see is -- and like I said, I've written like 200 lines of Go, so most of it is what I've read... Which is probably still less than a thousand lines, and then on the Go website, and stuff; most of it is like if err!= nil {return err;} Ain't that what most people do most of the time? + +**Kris Brandow:** Yeah, most people are doing that wrong. + +**Ian Lopshire:** Yeah. + +**Jerod Santo:** I was gonna say, isn't that not actually handling anything? Aren't you just passing it up? That's what I don't get. + +**Kris Brandow:** I think that's where a lot of the confusion from a couple years back came from, of like "No, you're actually supposed to handle your errors." You're supposed to -- + +**Jerod Santo:** Do something! + +**Kris Brandow:** ...think about what the problem is. + +**Jerod Santo:** If you don't do anything... It's like keeping the honest people honest. Like, you've gotta handle it right here. It's like, "Are you though?" It seems like you're just text-expanding that same snippet. + +**Kris Brandow:** I feel like a lot of the people that don't like if err!= nil don't like if because everytime they're writing if err!= nil {return err;}, and that's the only iteration of it that they see, and it's like "Well, that's incorrect." So you're using the thing wrong; you're not really getting the feedback you should get from it. + +**Jerod Santo:** So that's hopeful. Mostly, that's what I've seen. I've always thought, "This seems silly." But that's probably because it's just doing it wrong. + +**Ian Lopshire:** A lot of times you do end up returning an error, but in those you should be adding context, you should be checking certain error cases that you can handle... But a lot of times you pass up the error for someone else to decide if they're broken or not. + +**Jerod Santo:** Because that's what you do for like a runtime error, right? It would just pass itself up until the runtime explodes. + +**Ian Lopshire:** Mm-hm. + +**Kris Brandow:** And I think too it's fine if you're doing that for just one thing. What people get frustrated by is when you have 100 lines of code, and 80 of them are just if err!=nil {return err;}. You have a high proportion of it. But once again, I think you've gotta go back and structure your code a little bit differently, be a little bit more creative in what you're doing, and then you kind of reduce the number of that actual statement down to a handful. + +**Jerod Santo:** Well said. Okay, so this has given me new interest in that particular thing that has bothered me. It's like, "You know what - it doesn't have to be that way." + +**Break:** \[51:30\] + +**Jingle:** \[53:27\] to \[53:43\] + +**Jerod Santo:** Well, Ian, you are our guest today, so if you have an unpopular opinion, we would love for you to share it now. + +**Ian Lopshire:** I've been trying to think of one, I really have... I'm drawing a blank. + +**Jerod Santo:** Okay. Unacceptable... You shall say something. \[laughter\] No, totally cool. It's an optional segment. Kris, you are not hosting today, which means I defer to you. Have you thought of one? I mean, you've shared a couple; you got a little spicy there, but... What else have you got? + +**Kris Brandow:** There's so many spicy ones in my brain... Which do I choose from? + +**Jerod Santo:** \[laughs\] + +**Kris Brandow:** I mean, my one from last week was so spicy though... + +**Jerod Santo:** I haven't heard it yet, because we haven't produced that episode, but... I'm excited about it. + +**Kris Brandow:** I don't know, I'll go for one that's gonna be incendiary... I think it's time for Go to have a fork. + +**Jerod Santo:** Wow. Say more. + +**Kris Brandow:** I feel like it's time for Go to have a fork, because I feel like the community - we're not all together right now, and I think we're pretending that we are. Especially with things like generics, I feel like we've kind of fractured a little bit, and we should have a recognition of that fracturing... And then maybe more formal conversations to like -- how do we come back together, how do we come back as a whole? Because I feel like there are different large factions of Go right now, and I feel like us all trying to operate together is not winding up being entirely healthy for us as a community. So maybe it's even just a forking of our community; I don't know if we need to fork the actual language, but I feel like there needs to be some sort of separation of us for a little bit. + +**Jerod Santo:** Where would the dividing line be? Would it be around generics, or are there other things? + +**Kris Brandow:** I feel like there's kind of the way that the language was when I first started learning it, which I guess is like the Rob Pike era of the language, where it was extremely practical of a language, and it was very focused on being a language for experienced software engineers that is still useful for those people who are newer... And I think over the past few years the language has pivoted to being one that is more focused on helping newer people be able to do things, people from other languages feel more comfortable in the language... And I feel like that has come a bit at the expense of people that want to do that fresh, new thing; they want to do that "I've spent a decade writing code, and thought about it very deeply, and I want that very nuanced and articulated way of doing development." + +**Jerod Santo:** Ian, your response. + +**Ian Lopshire:** Ahh... + +**Jerod Santo:** \[laughs\] Stake your claim right now... + +**Ian Lopshire:** \[56:06\] I do kind of agree with Kris here, that it's changed a bit. I don't know if I would say it's changed and become worse for experienced people that wanna write code a particular way, but it does seem to be more -- almost academic, instead of practical. Especially modules, and maybe generics as well. It feels almost like a research project, or like "How could we do this?" not "Should we do this?" + +**Kris Brandow:** Yeah. + +**Ian Lopshire:** I don't know if we need a fork, but I do agree that there's been a change some way; a shift in culture, or something. + +**Kris Brandow:** I feel like that's a good way of putting it. It does feel like it's shifted from a practical to a more academic, more research-style thing. Once again, this doesn't have to be like a fork-fork, like we have to have completely separate development teams. Maybe it's like an extended conversation that we have, and maybe there's different ways we come about deciding things. It started off as very practical, and it shifted now to very academic, and I feel like the language has lost a little bit of its soul along the way there. + +**Jerod Santo:** Well, while we're here, let's have a forked Go naming brainstorming session. I'm thinking, "NoGo", I'm thinking "Gone", I'm thinking "Stop", or... + +**Ian Lopshire:** Stay... + +**Jerod Santo:** "Stay..." \[laughs\] Go 2? Oh, GoTo, not Go 2... Alright. Any other potential fork names? We're gonna have to ideate this a bit... + +**Kris Brandow:** Maybe Og... + +**Jerod Santo:** There you go, rearrange the letters. + +**Kris Brandow:** You could steal the mascot and just call it Gopher... + +**Ian Lopshire:** Please don't call it Gopher... \[laughs\] + +**Jerod Santo:** \[laughs\] Like the old networking protocol, Gopher? + +**Kris Brandow:** We could call it Hoard, because Gophers are hoarders, so that could be a fun name... + +**Jerod Santo:** I'm glad you put a d at the end of that. + +**Kris Brandow:** Yes, HoarD. + +**Jerod Santo:** Please enunciate. Okay... Submit your Go fork names in the comments or on the Twitter poll for this particular unpopular opinion. I'm thinking that one's gonna be -- + +**Kris Brandow:** Oh, God. Someone said Go++, or Go\#. + +**Jerod Santo:** Go\# ! + +**Ian Lopshire:** Go\# ! I'm sticking with that one. + +**Jerod Santo:** We've got IaGo... I'm assuming that's a Lion King reference... Alright, let me share my unpopular opinion here. So on episode \#192 Ashley Jeffs had what was in my opinion the most unpop of all times. His opinion was that people who vote in Twitter polls are losers. Did you guys hear that one? He says they should get out more. Nobody cares about their opinion and it doesn't matter. Now, he thought that he'd gamed the system. He thought he made a perfectly unpopular opinion, because of course, we take the votes on Twitter... And where are they gonna vote? But reality is stranger than fiction. 71% of Twitter poll-takers agreed that they were indeed losers. + +So based on that empirical evidence, I can with confidence state that my unpopular opinion is that people who vote on Twitter polls are winners, and they should tweet more. Everybody cares about your opinion, and it does matter! + +**Kris Brandow:** \[laughs\] + +**Ian Lopshire:** I feel like that one is going to be unpopular. + +**Jerod Santo:** I think it might. \[laughs\] + +**Kris Brandow:** I think people are just gonna be in spite just voting that... Like, yeah, that's unpopular, just to -- yeah. + +**Jerod Santo:** Well, time will tell. Please follow @GoTimeFM on Twitter and vote your opinion. Do you think you're actually a winner, or a loser? We will find out on Twitter. That has been our show for this week. I appreciate you all letting me crash the party and ask my newbie/outsider questions. Hopefully, there are other Go-curious folk that learned a thing or two and enjoyed this conversation. Ian, thanks so much for joining us once again. You're welcome back any time. + +**Ian Lopshire:** I'm glad to be here. + +**Jerod Santo:** Kris, you were also here, as Mat Ryer would say... + +**Kris Brandow:** \[laughs\] + +**Jerod Santo:** But I'll say thank you for hanging out and sharing your wisdom and your spicy hot takes... A Go fork - I had never seen that one coming. + +**Kris Brandow:** Yeah. + +**Jerod Santo:** But that's Go Time for this week. We'll talk to y'all next week. diff --git a/Avoiding bloat_transcript.txt b/Avoiding bloat_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..948af0e8de9549b00addfc78cb9f959281d4e415 --- /dev/null +++ b/Avoiding bloat_transcript.txt @@ -0,0 +1,689 @@ +**Mat Ryer:** Hello, and welcome to Go Time. I'm Mat Ryer. Today we're talking about avoiding bloat. Yes, indeed. I'm joined by Egon Elbre. Hello, Egon. + +**Egon Elbre:** Hello, Mat. + +**Mat Ryer:** Welcome back to Go Time. + +**Egon Elbre:** Yeah, it's nice to be back. + +**Mat Ryer:** It's nice to have you back. And you build things at Storj, right? + +**Egon Elbre:** Yeah. Storage. + +**Mat Ryer:** So that's how it's pronounced. + +**Egon Elbre:** Yes. That's the first question I always get. + +**Mat Ryer:** Yep. GIF or GIF all over again. GIF or GIF. Yeah. Well, welcome. Yeah. We're also joined by Roger Peppe, hacker at Cuelang.org. Hello, Roger. + +**Roger Peppe:** Hello, good evening. + +**Mat Ryer:** Good evening. Welcome to Go Time. We shouldn't say good evening, because we don't know when people listen to this podcast. One of the great things about Go Time is you listen to it on your own terms. + +**Roger Peppe:** It's always the morning on the internet. Good morning! \[laughter\] + +**Mat Ryer:** \[04:00\] It's always the morning somewhere on the internet. Well, people do terrible things in the morning then, in that case... Okay. Well, before we get going, Egon, I was interested, what have you been working on lately? Anything interesting? + +**Egon Elbre:** One of the recent things I finished was a project called Lensm which allows you to browse source code, and to compile the assembly side by side, and see how Go translates into assembly, essentially, and do this interactively. a nice way to visualize it when you're optimizing things. + +**Mat Ryer:** Yeah, very cool. Because it's funny, a lot of people, I think, stay at the code level and don't really dig into that, and it's quite apt for this conversation about avoiding bloat, understanding what's actually this turning into. + +**Egon Elbre:** Yeah, exactly. + +**Mat Ryer:** Yeah. And Roger. Cuelang. Very exciting project. What's going on there? + +**Roger Peppe:** I think it's an extremely exciting project. So I've been involved with the language Cue, I've been enthusiastic about it for a few years now, two or three years, since not so long after came out. I recently had the opportunity to join the project, so I'm now working on it full-time... + +**Mat Ryer:** Yeah, amazing. + +**Roger Peppe:** ...on this new language. It's sort of halfway between JSON and a normal language, I suppose... A configuration language with some very, very interesting properties. It's a very cool project. And hopefully, it should be useful for everyone in the both now, but in the future. It has a big future. + +**Mat Ryer:** So keep your eyes out. I will put a link to it in the show notes for anyone who wants to dig in more, and maybe we'll do a different episode on that at some point, Roger. + +**Roger Peppe:** I think that's a great idea. + +**Mat Ryer:** And Roger, you also play the fiddle, don't you? + +**Roger Peppe:** Not relevant here, but yes, I do. + +**Mat Ryer:** Very relevant, because I went to a party once, and I thought I was having a stroke, but it turns out there was a man playing a fiddle... And that was you. + +**Roger Peppe:** Oh, that was me, was it? + +**Mat Ryer:** Yeah, yeah. + +**Roger Peppe:** Oh, really? I was at the same party as you. Gosh... \[laughter\] There we go then. I'll try not to do it again. + +**Mat Ryer:** No, do it. There's barely anyone ever playing the fiddle at any parties I go to, so... + +**Roger Peppe:** You just go to the wrong parties, evidently... + +**Mat Ryer:** Apparently so... Okay, so let's get started then. So when we talk about bloat, what do we mean? What is bloat? + +**Egon Elbre:** So I like to separate it to code bloat and binary bloat. So one is your code is growing larger and larger, and the other is your binary, the final deployment thing growing larger and larger. + +**Mat Ryer:** Yeah. We'll bear those two things in mind then as we have this conversation, because I think they're both important to pay attention to in different places. With code bloat, you can have relatively small amounts of code. But if you have lots of imports, your binaries might end up being quite big, right? + +**Egon Elbre:** Yeah. So I actually dug up some statistics... I like to do some small quiz... So how much would you guess, how many lines of code, if you import a timestamp protobuf definition and just print it out? How many lines of code would that cause? marshaling it, and how many lines of code would it import? + +**Roger Peppe:** In Go if you import it? + +**Egon Elbre:** Yes. + +**Mat Ryer:** It can't be many, can it? + +**Roger Peppe:** I reckon about -- my guess would be about 30,000. + +**Mat Ryer:** What?! + +**Roger Peppe:** I know that protobuf is ridiculously bloated, so... That's my guess. + +**Mat Ryer:** I thought it was all about being tiny little payloads. I was gonna say 50... I feel this is a good quiz. + +**Egon Elbre:** 27,000 lines of code. + +**Mat Ryer:** Oh, wow. + +**Egon Elbre:** Yeah... + +**Mat Ryer:** Roger, how did you get that so close? That's suspicious. + +**Roger Peppe:** An order of magnitude... Yeah, no, Egon tipped me off, yeah... \[laughter\] He didn't, he didn't. I just know how big these things are. + +**Egon Elbre:** Okay. The other common one is gRPC. + +**Mat Ryer:** Yeah. Well, I know that's a binary format, so that sounds small... A billion. \[laughs\] I've learned my lesson. + +**Roger Peppe:** \[08:06\] gRPC is -- well, that includes all the protobuf too, right? + +**Egon Elbre:** Yes. + +**Roger Peppe:** So we're talking like maybe 120,000 lines of code. + +**Mat Ryer:** Oh, it's going to be more than-- + +**Egon Elbre:** In the same ballpark, so 100,000. + +**Mat Ryer:** Ah... Roger, you're suspiciously close to these. Do you count lines of code? + +**Roger Peppe:** I have too much experience with large code bases. \[laughs\] + +**Mat Ryer:** Yeah, apparently... That tells you something though, because it'd be very easy to just build what you imagined to be a very small project, and just import one of those two technologies, and suddenly, you're really talking serious, serious numbers. + +**Roger Peppe:** Yeah, compared to Net RPC, if you remember that package that not many people use, which is only about a thousand lines of code, or a couple thousand lines of code...? + +**Egon Elbre:** Yes, something that. + +**Mat Ryer:** Is that why people don't use it? It's not enough -- people are like "I'm not buying it unless it's got loads of lines of code in it." + +**Roger Peppe:** I mean, gRPC also has its entire copy of the HTTP, another copy of the HTTP/2 stack, and that kind of thing. So it seriously reinvents many wheels. + +**Mat Ryer:** So would you have to import that to use it in a project? Or is that just like the toolchain needs that? + +**Egon Elbre:** No the code that you import, and the compiler has to parse through, essentially. + +**Mat Ryer:** Hm... Well, you just don't notice, because Go's build times are so -- they're so fast. Maybe we're spoiled by that a little bit. + +**Egon Elbre:** Every single line of code, that's one second to your build time, so... + +**Mat Ryer:** \[laughs\] Well, that's what it used to do... That'd be good. You'd pay attention to code bloat if that were the case, wouldn't you? + +**Roger Peppe:** I think it's all just got so ridiculous, from my point of view... It started off back in the '80s, where like 16k was a huge amount, was a very large program. And now it's in the many hundreds of megabytes, and "Well, doesn't matter...?" Well, does it really matter? I don't think it does, from my point of view, I think, unless you're really coming up against build times or against binary sizes. What's important to me is the maintainability aspect of code bloat. It's about the bloat in the code that you maintain, not necessarily the bloat in the code that you're importing. + +**Mat Ryer:** Yeah, right. So that's the other side of this then, is as a project grow and grow, you have to maintain them. Sometimes you can just leave code for ages and you never really have to touch it, but you still have the cost of maintaining that, don't you? + +**Egon Elbre:** I consider every single line of code that you are the maintainer of it now. + +**Mat Ryer:** Yeah. Every line of code...? + +**Egon Elbre:** Because maybe the maintainer goes on vacation, and you have a critical bug that you need to fix. + +**Roger Peppe:** So you've personally reviewed all those lines of gRPC code? + +**Egon Elbre:** Yes. + +**Roger Peppe:** Good work... + +**Egon Elbre:** Luckily, I don't have to do it anymore. + +**Mat Ryer:** You do that every time there's a new release? Or do you just read the diffs, and then apply that in your brain? + +**Egon Elbre:** No, we dropped actually gRPC for that maintain reason. + +**Mat Ryer:** Yeah. So that's interesting then... When you import a package, it's because it's such an easy thing to do... I wonder how many people think I am committing to also taking on the responsibility of looking after this entire other project as well. I don't know that many people have that mindset, do they? + +**Egon Elbre:** I think it depends on the background, where you come from. So if you start out by building websites and all that stuff, then probably you don't think about it. I used to work for an electronic medical records company, and every single dependency had to be thoroughly reviewed, and whether it's suitable. So... + +**Mat Ryer:** Yes. For sort of regulatory reasons, I guess... + +**Egon Elbre:** Yeah. + +**Roger Peppe:** Yeah, I have to say that I don't personally review all our dependencies, because I don't have the time... But if I take on a dependency, I look at the dependencies of the dependencies often. I'm always more concerned about that. If something just depends on the standard library, I feel that I've got a handle on it. But if it's depending on other projects, then that's a bit more problematic. So in projects that I maintain, I try to keep the amount of dependencies, particularly when it's a library that some people are going to use, like the few I maintain, I just try to keep the dependencies absolutely minimal, if possible. + +**Mat Ryer:** \[12:23\] Yeah. So how do you do that? Do you have to manually go and look at all the projects? + +**Roger Peppe:** I look at go.mod, actually, and I'm like "Why have I got that dependency? Do I need that dependency? How can I strip it out?" And I was doing that a while ago with some fairly large codebase, "Why have we got this --" And I came across -- that was actually probably what ended up with me on this call actually, because I came across Egon's tool, Goda, which is a fantastic way to try and visualize to see why you are using a particular dependency, and how you might be able to hack it out. + +**Mat Ryer:** Hm. What is that project, Egon? + +**Egon Elbre:** So it was for solving this problem of understanding your dependencies. So it stands for Go Dependency Analysis Toolkit. So I ended up collecting these different tools, and then I eventually merged them into a single, large bundle of them. It's at github.com/loov/goda. And it has features for -- I try to explain it as you do calculations with package sets and dependencies. So you can start removing things that you don't care about in the package list or graph, and then you can drill deeper and figure out why is something being imported. There's also a sub-command goda cut, that displays packages that you might be able to easily remove, because they don't have many incoming dependencies to them. So a bunch of tools... + +**Mat Ryer:** Yeah, that sounds really good. I mean, as a rule, I kind of agree with you, Roger - if I'm going to import a package, I'll prefer ones that have just a few dependencies. And even sometimes I used to just copy bits of the code in, with the license at the top, always, just to kind of avoid it... And often there's lots of tests; it may be something in the testing side that is -- Testify has quite a lot of dependencies. Is that okay? would you avoid Testify for that reason? Or because it's in the tests side are you alright. + +**Roger Peppe:** Personally, yes, I would. But I'm biased, right? Because I maintain a package called -- I'm trying to think what it is now... QuickTest. I'm one of the maintainers of that. Which does not have many dependencies, and it's a bit smaller, so I quite like that... Because Testify has lots of dependencies. But it's hard to avoid them, because some dependencies are really useful. Like to print out diffs for example, like the CMP package, it's great for that... But it's another dependency. + +**Egon Elbre:** When I bring in dependencies, I actually do review most of the lines of the code. I also run our usual linter suite through the codebase. And this means that if we import a new package, usually there are a few fixes that we contribute upstream already. Maybe there's a data race, maybe there's a global variable that can be removed, or many kinds of maybe minor things, maybe major things... + +**Roger Peppe:** The other thing that the Goda tool is really useful for, I've found, is trying to sort of -- a code bloat which we haven't really talked about is code bloat in your own codebase. So when your code base gets really big, and you've done some changes, and you've made some big migrations, but you've still got some old bits of code, and just trying to make sense of this tangle of dependencies, that you end up with something... If you have more or less a monorepo, then you can have a very tangled set of dependencies, and it's really quite hard to make sense of them somehow, to try and work out "Okay, we want to factor these things out into their own thing", but you don't know where everything's coming into it, what's coming out of it... And I found that incredibly helpful when I was making some changes, particularly in my previous job at Influx Data... Trying to do some big, big code changes, and it was quite hard to work out the dependency relationship. + +\[16:21\] Of course, Go is fantastic because of a great rule that it has - you can have no cyclic dependencies. That is amazing, actually. That rule, just in itself, has contributed hugely, I believe, to the maintainability of larger Go codebases. Because without it, you tend to get in a situation where something at the very bottom of the dependency tree tends to point something quite near the top, and then you have this big ball of mud, which is like a Gordian knot, and you can't cut it. \[laughter\] + +**Mat Ryer:** Yeah. So for people that don't know, this is where you'll have, say, three packages. A imports B, B imports C; that'd be fine. But it's when then C imports A, and you get this strange circle, right? Is that right? + +**Roger Peppe:** Yeah, that's it. And also, it's a really good rule, but it was also really frustrating at times, because there are times you're like "Oh, dang, we need to import this from here and we can't, because it says cyclic dependency." So then you have to break your dependencies, and often you have to split up a package, but it's not clear how you can split up the packages. It's actually quite a hard problem, but you're quite often forced to make this decision earlier than you would in some other languages, I think. + +**Mat Ryer:** Yeah. And that's the thing; it's like, take on the pain early. A stitch in time saves nine, which means if you can stitch with one, do it early, a little bit of pain; it saves that, because it grows, and it becomes much more painful later. I've had the same thing too, and one of my approaches is actually just to keep everything in one package for as long as I can. As long as I can get away with it. I'm not joking; that genuinely works too, because of that. + +I quite like it when the structure emerges, rather than is imagined... Because sometimes it's obvious, and sometimes I've got dependencies that -- I've got something that I know it's going to be a package, I know it's going to be useful in multiple places... But I'm often surprised as well, as the application is being built, which bits present... I certainly don't follow just a standard structure that I know some people do. + +**Roger Peppe:** I agree with that, yeah. Definitely. + +**Mat Ryer:** So with binary bloat then - like you said, 16k used to be a big program. And I remember -- you know, a floppy disk was 1.44 megabytes that you could fit on that. If you take a photo now and share it, that's about eight floppy disks, or something, at least. You wouldn't tolerate that. If you're like "Oh, you've got to check out this photo. Here's a little stack of floppies." + +**Roger Peppe:** Those were big floppy disks as well, right? It started off at like 200k, if you were lucky... + +**Mat Ryer:** Yeah, when they used to be actually floppy. + +**Roger Peppe:** Yeah. Not to mention tapes... + +**Mat Ryer:** Yeah. Honestly, I love that early tech. I really miss it when tech was rubbish. It was so cool. I had a Spectrum, with the tape cassette thing, when I was a kid... So literally - and people won't believe this, but in order to load a program... I don't even know if they know what a cassette tape is either. So it's something that you sometimes see on retro T-shirts, one of those. Literally then it was the sound, it was encoded as sound. So it would play the sound off the tape, and that would -- like how a modem works, really... And then it would read the data, load the program, often with mistakes, so you'd need to do it a couple of times to get it right... + +**Egon Elbre:** I think you now need to explain the modem as well. + +**Mat Ryer:** \[19:56\] Yeah... This is it. No, it's all changed so much... And by the way, I started young. I'm not old. I just want to make that clear. But why do we care about binaries now? Because we can get away with great, big binaries, can't we? What's the big deal? They upload in no time, they move around the web pretty quickly... Do we care about big binaries? + +**Egon Elbre:** I think mostly we don't. If we did, we already would have done something significant about it. But there are definitely cases where we do care. + +**Roger Peppe:** I do get shocked by the size of binaries sometimes. I look at it, I'm expecting like five megabytes or something, and it's 120 megabytes. I'm like "Wow...! That's ridiculous." + +**Mat Ryer:** And is that usually because you've imported something in your own program? Or are you just talking about anyone's programs? + +**Roger Peppe:** I guess I'm just talking about building a binary and having a look; it's not me importing something particularly. But yeah, binaries are very big. I mean, I care more about binaries on small devices, if I'm running something on a Raspberry Pi, or... Actually, even a Raspberry Pi, you usually have quite a lot of SD card space, so it doesn't matter either. But on smaller devices than that, then it really becomes an issue. + +**Mat Ryer:** How many floppy drives can you attach to a Raspberry Pi? + +**Roger Peppe:** How many floppy drives have you got? \[laughter\] + +**Mat Ryer:** Not enough. + +**Roger Peppe:** Clearly. + +**Mat Ryer:** None at the moment... Actually, I do have one, because I've recently -- I decided I'm going to buy all my old computers that I used to have. I've got a Spectrum already, it's arrived. I've ordered an Amiga 500, and I'm bidding on one on eBay, the Amiga 1200. These were my early computers that I grew up with. So we'll have some floppy drives... + +**Roger Peppe:** That's a good idea. + +**Mat Ryer:** Yeah. I'm just going to put them on the wall. Plug them in, but just have them on the wall. + +**Roger Peppe:** I should get an Acorn Archimedes again. That was my first ever computer that I owned. + +**Mat Ryer:** Did you? That's cool. + +**Roger Peppe:** At university that was, yeah. + +**Mat Ryer:** Yeah... We had those at school. + +**Egon Elbre:** I started on a 286, so... A bit later than you. + +**Mat Ryer:** Yeah... What are you trying to say? + +**Egon Elbre:** I started later...? + +**Mat Ryer:** Well, then you did a fine job. \[laughs\] + +**Break:** \[22:11\] + +**Mat Ryer:** Okay, so small devices - and this is where TinyGo comes in, because this is the problem that TinyGo is trying to address, is so that you can still use what is essentially the standard library, but it's much smaller; it's deliberately designed to be cut down and simpler. + +**Roger Peppe:** Yeah. I love TinyGo. I use TinyGo a bit gratuitously, a friend once said. I wanted to build a doorbell with hanging -- not tubes; what do you call them? Maybe you call them tubes... I would think knockers... + +**Mat Ryer:** Stick with tubes. + +**Roger Peppe:** So I built in the software -- he built the hardware; he put the actual thing together with the relays in there, but he never actually got the tubes working. So I built the software, and it's great because you've got this tiny little microprocessor... You know, maybe it was 16-bit, maybe 8-bit; I can't remember. Pretty small, with 1k, 2k RAM, or something... And you can build this Go program, which -- you can list goroutines, and everything. So I just had this separate goroutine, which would be responsible for going through the tune, and then another one, which is listening to interrupts where you would press the bells, and it would interrupt the other goroutine... And it was just a really nice way to structure it; you couldn't have done that if you were writing in C... And it was great. And it all fitted; it's great. It turned the LEDs on in the end, it didn't ding the bells, but... You could pretend. + +**Mat Ryer:** Yeah. You can have butler. Hire somebody to -- when the lights come on, they hit the tubes. + +**Roger Peppe:** At the same time, that would work. + +**Mat Ryer:** If he could... Yeah, eventually work their way up to that. That sounds cool. I love that. And Egon, at GopherCon EU in Berlin, recently, you were hacking with TinyGo, weren't you? + +**Egon Elbre:** Oh, yeah. I did some MIDI controller thingy. + +**Mat Ryer:** Yeah. So he was playing music on a little -- he had these buttons attached to this little breadboard, and then wrote the code to translate that into MIDI instructions for some music software. + +**Egon Elbre:** Yeah, it worked pretty well, and it was really nice. + +**Mat Ryer:** \[26:13\] Yeah, I liked it. + +**Egon Elbre:** I did get some embedded device noise on it, so some of the buttons didn't work as they were supposed to, but it still did things, I guess... + +**Mat Ryer:** Yeah, it was going *DONG DOING DING* which I think is good. It's not as good as Roger on the fiddle... + +**Roger Peppe:** Not so nice on that I'm afraid... It was actually great using Go with TinyGo, because you could have this debouncing code, which totally didn't care about any of the other code. It would just sleep for a bit, wait for debounce... And that was independent of all the other logic waiting for buttons, and stuff. + +**Mat Ryer:** What was that doing them, the debounce? Literally stopping if you got noise coming through from buttons, or something? + +**Roger Peppe:** Yeah, so if you press a button, then it's not a clean thing that you have pressed this button. It makes contact, and then bounces, and comes down again, and bounces, and comes down again... It eventually ends up either down or up, right? But you don't know when that initial contact is made, which way it's going to end up. You don't want it to go *blublolblubblu* + +**Mat Ryer:** Yeah. You would stress the butler out. + +**Roger Peppe:** Yeah, you don't want to do that. + +**Egon Elbre:** I did have debounce code, but something was wrong. \[laughter\] + +**Mat Ryer:** Well, that's funny, because debounce - I've used that in the frontend of web development, which is literally the same thing. If you click too many times, or sometimes on hover -- because you can, if you're just teetering on the edge, on the precipice of the pixel, you can end up with this horrible flicker. In your case, Roger, that is literally a bounce that you're talking about. + +**Roger Peppe:** It is, absolutely, literally, a bounce, yeah. It was quite interesting to write... And it's so nice writing in Go, honestly. I can't get over how nice it was. You can have interfaces... They're quite clever about interfaces and TinyGo, actually, because they basically expand all the code out. They're really clever about some of the optimizations for space in TinyGo. + +**Mat Ryer:** Yeah. Do you think people should, as an experiment, use TinyGo and try and actually experiment with that, and also run something on a tiny device, just to sort of understand what's going on? Or is this, again -- are we just at the point, most people, where we can just deal with a big binary? + +**Roger Peppe:** I think it's well worth experimenting with it, and having a go, see what it takes to run in 1k of RAM, or whatever... Because binaries are still quite big, but you can still have 128k, 256k of binary. But you're not going to be importing gRPC. \[laughter\] + +The other place that binary bloat really matters is on the web, right? You can compile Go to WASM. But if you're downloading a 100 MB WASM file to your browser, that's not going to go so well. That's actually, I think, another -- I haven't used it in this, but another use case for TinyGo is to target WASM and have a relatively small binary that gets downloaded to the browser. + +**Mat Ryer:** Yeah, I've seen that. I've never done it myself, but I have seen a talk, which I'll try and dig out and put in the show notes, of somebody that basically did that. They wrote something just in Go, and showed you how long it took to actually run. Because it has to download into the browser. And then they did it in TinyGo, and it was obviously much snappier. + +I think that's interesting future there with WASM, and I think therefore TinyGo probably does play a big part in that, for Go people. I know that Ron Evans, one of the main contributors, and I think the founder of TinyGo, he very much advocates for more people looking at TinyGo, and using it, and contributing. So he's very keen to get people on board with it as well. + +**Roger Peppe:** Yes, it's really cool. An interesting observation is that in the smaller Go programs, one of the main contributors to binary bloat is the Fmt package... Because that actually is quite big, and so many things have it as a dependency. + +**Mat Ryer:** \[30:02\] Yeah. And it feels something that's just baked into the standard library, and therefore it's just around and you can just always use it... But if you think about all the things -- when you use the verbs, all the different things you can do, the reflection, and everything... You can see why it ended up having lots of dependencies. + +**Roger Peppe:** I mean, it hasn't itself got lots of dependencies, but it is itself quite a large amount of code; it does a lot of reflection code... It's quite a lot of code. + +**Egon Elbre:** I think one of the major contributors to the Fmt package is actually Unicode tables, because it needs to handle many of those cases. And I think those tables - I might be wrong, but like a few hundred kilobytes or something already. + +**Roger Peppe:** I don't know -- maybe white space. I'm not sure. + +**Mat Ryer:** What do you mean whitespace? Surely it's not padded out, is it? + +**Roger Peppe:** I mean, it's 3,500 lines of code. + +**Mat Ryer:** Oh, we could have done a quiz, Roger. + +**Roger Peppe:** \[laughs\] + +**Mat Ryer:** Could have gotten Egon back. + +**Roger Peppe:** I mean, maybe Scanf I don't know. I'm not sure that Fmt itself needs those Unicode tables. Maybe it does. + +**Egon Elbre:** Maybe that's changed, but... + +**Mat Ryer:** Yeah, I don't know. + +**Roger Peppe:** Quickly looking at the code... + +**Egon Elbre:** I remember optimizing those at some point... Trying to optimize them. + +**Roger Peppe:** When I was doing some of this stuff, I was like "I just want a version of Fmt that only has the very basic verbs, like percent S, percent D, doesn't even do widths, and you'd be fine, in most cases. + +**Mat Ryer:** TinyFmt. + +**Roger Peppe:** TinyFmt, absolutely. + +**Mat Ryer:** You could make that, Roger... + +**Roger Peppe:** You can't replace standard library packages though... + +**Mat Ryer:** Well, you probably can do anything you like if you have some kind of pre-processing thing going on... Actually, genuinely, if it scanned to see which verbs you used, and then built a Fmt package that just had that... I don't know. That's maybe going too far. + +**Roger Peppe:** I mean, that's kind of almost what Rust does, actually. + +**Mat Ryer:** Is it? + +**Roger Peppe:** Well, because in Rust the formatting stuff is macro processing so it kind of expands out at compile time. + +**Mat Ryer:** Oh, yeah. That's cool. + +**Roger Peppe:** That's why the compiler is so slow. + +**Mat Ryer:** That's it. It's a trade-off you've got to make, isn't it? That's the thing about, I think Go... I'm always quite pleased with the trade-offs that they end up making. I understand, and then someone will hit an edge case and they're really frustrated by it, because it's not performing for them... But by and large, I think they are pretty pleased with where they usually land. + +**Roger Peppe:** Yeah, me too. And it does seem to be -- it is maintainable for large projects, but there is that kind of inherent complexity that you tend to accumulate when your project gets larger and larger. There's some threshold you seem to cross, at which people stop understanding the codebase... And for me, that's really when code bloat as I care about it really kicks in. People are making changes, they don't understand the codebase, therefore they tend to reinvent the wheel or just write much more code to jam their feature into the codebase... Rather than saying, "Oh, well actually, we could just change this little package over here to add this feature, and not add these 100,000 lines of code over here." And then you've added those 100,000 lines of code and it's even worse, right? For me, I think code bloat is quite strongly linked with technical debt. + +**Mat Ryer:** Yeah. I think whenever you end up with bits of the system that you just don't touch, like "No, don't go and touch this. Be scared of this", I think... I've had that situation before. And even in projects where I've been the only one working on it, I'd just managed to get something right, and I don't have the context... And it's messy; I never cleaned it up, and then I just don't want to even touch it. And with Go, I find that I do less of that, particularly because testing is such a big, first-class concern in Go. And I tend to write TDD... So I will, at the end of writing and solving a problem, I can just go back and just very boldly hack away at it, change it, make big changes,= with the confidence that as long as the tests are passing, then all the promises I've made are being still true. + +**Roger Peppe:** \[34:13\] That's great until you find that all the tests that you've written using mocking, and that you're changing some of the things that they depend on, and so all your tests are now invalid, right? Because this is can be very hard... + +**Mat Ryer:** Yeah. I think strategies like when you keep the interface near where you're going to use it - that as a rule I quite like. Some packages will have - they'll expose interfaces, and I the idea... Like, if I'm going to use the SendGrid API, if I have a sender interface that just has the single method that I'm going to use in there, this to me - it's not really solving anything to do with code bloat, but I'm really explaining there that this is what I care about, this piece. And then if there's any mocking or anything that, then it's much smaller. But yeah... I tend not to do much mocking. I tend to do a lot more integration testing. + +**Roger Peppe:** Me too, and I think there's a good reason for that... Because if you have integration tests, and then you refactor the insides to maybe use something with less dependencies, or something with a different kind of API, then your tests are still valid, right? + +**Mat Ryer:** Yeah. + +**Roger Peppe:** But if you rely entirely on substituting in the thing that's underneath, then you can't change the tests, or you can't change those dependencies, because your tests are now invalid. + +**Mat Ryer:** Yeah. + +**Roger Peppe:** This is a big problem with larger codebases and being able to refactor so that you are changing your dependency network, for example. + +**Egon Elbre:** So a question... What's your definition on technical debt? I've seen it used in many different ways, and I wonder what's your definition. + +**Roger Peppe:** Mine or Mat's? + +**Egon Elbre:** Both, I guess, if it's different. + +**Mat Ryer:** It might be different... Well, just off the top of my head, I'd say it's something that I should have done now, but I'm not going to. You see, I tend to avoid it quite rigorously wherever I can. So I'll go to great lengths to either -- usually by shrinking scope, but I'll try and not do as much, and then I can do it well... But that's a very vague answer. I don't know if that resonated with you, Roger. + +**Roger Peppe:** I mean, for me, technical debt is something that is like a cost that I am unwilling to pay, or I wasn't willing to pay in the past... Because you've always got to prioritize, you've always got to triage and do some things first, and other things later... So technical debt is the things you left for later, right? You can pay that off early, or quite often it's like "Well, the other things are still more important, so we'll leave that." Or "This is really too big. We're gonna have to take an entire year out with our whole team in order to address this serious problem in our codebase... And that will take us out of the market; we're not going to do that. It will mean we can't ship any features." So this is a debt that you're not going to pay off, and you probably will never pay off... And that's the death, I think, of most projects. + +**Mat Ryer:** It's a mortgage. Mortgage, actually, the word -- it's like "death loan." The "mort" in mortgage comes from death. + +**Egon Elbre:** Oh... + +**Mat Ryer:** Yeah, really. Because that was the idea, it was you'd just have this loan for your whole life. Sounds grim, doesn't it? + +**Roger Peppe:** I never knew that. + +**Egon Elbre:** At some point I was thinking about technical debt, trying to give a rigorous answer, like what is it... What I ended up is that -- let's say there's some effort that you put into maintaining a codebase. It doesn't need to be maintaining, but maybe there are other aspects. And then there's the ideal state of how much effort you need to put into that codebase to maintain it, right? So the difference between those ideal state and your current state is going to be the technical depth, right? + +**Mat Ryer:** \[37:57\] Sounds reasonable. + +**Egon Elbre:** Yeah. But there's one funny thing there... Let's say there's an innovation in the technology, right? When React came up, the idea of maintainability or the effort to maintain got lower, which means that your technical debt goes high up if there's an innovation. So... + +**Roger Peppe:** That's interesting. And does that mean that when you start a project and it's not by any means finished, that that counts as technical debt, because you just haven't finished that project that you've just started? + +**Roger Peppe:** Yeah, I guess... Let's say you have a security debt; you haven't done a proper security audit, right? There are other aspects too you can measure. And if you look at maintainability only, then an unfinished project isn't necessarily in debt because of it being unfinished. So that ideal state in terms of maintainability. + +**Mat Ryer:** Yeah. That is interesting. So in that case -- because this is the other thing... If you have what you perceive to be a technical debt, and you acknowledge it, and you decide "For priority reasons, we're going to leave it", and then later you find out "We actually didn't need that. We never need it", then I guess that debt just pays off. It just kind of gets written off, the debt. + +Yeah... And so I think one of the other principles is designing things so that you can change them. Designing things for that flexibility in the future. I think it's quite an interesting idea. I think I saw it on Twitter, somebody said, like, architects - architects are there, and it sounds great, because they're designing these systems, and they are laying these foundations, and telling you where all the important pieces are going to be... And then someone said if you called those people technical -- oh, I forget the term they used. It was something early decision -- like "unchangeable decision person." Then suddenly they sound really unreasonable, and that you wouldn't want that. So having the flexibility in the architecture, and being happy to change things, and evolve things... Acknowledging that we don't know stuff, we don't really know anything at the beginning especially, and we'll learn it as we go... + +**Roger Peppe:** That said, I think it really helps to do a lot of thinking through. You can actually design stuff; it might not be set in stone, but I do think that you can eliminate a lot of bad code paths or bad futures by just thinking them through to start with, and thinking, "Well, if we went that direction, where would we end up? That's a bad place to go." + +I think a lot of people are like "We shouldn't do any design upfront." I really don't agree with that approach at all. If you look at Go, for example - they designed a lot. They wrote the spec before they did the implementation. + +**Mat Ryer:** Yeah. + +**Egon Elbre:** I think there was one quote that when you're designing things, add flexibility into the places where you are most uncertain. + +**Roger Peppe:** Yep, that's a really good point. + +**Mat Ryer:** Yeah, that's true. + +**Roger Peppe:** That sort of leads through to one of the things where I think is sort of key to avoiding code bloat, is just designing the right abstractions. If you get the abstractions right up front... There's this feeling when you're wrestling with the wrong abstraction, you just feel the code is piling up. You're just writing all this code, because you've got the wrong abstraction, because you're doing things the wrong way. If you get it right, maybe you change your abstraction and start to do things differently, and suddenly you can be like "Oh, gosh, I can delete that entire package. I can delete that entire directory." You can delete hundreds of thousands of lines of code sometimes because you've changed the way it's working in a very subtle -- in what was initially maybe not a very obvious way. + +**Egon Elbre:** \[41:52\] I remember a talk about designing things for deletability, so that it's easy to delete features, rather than to extend... Because if it's easy to delete, then it's probably easy to replace. + +**Roger Peppe:** Yep. + +**Mat Ryer:** Yeah, I like that point, actually, on the abstraction thing... I talk a lot about kind of avoid doing the abstractions too early. Do a few examples first, and see. It'll depend on the project or the problem. But actually, the real value is in these abstractions. That is really probably one of the most valuable things we can do, I think, to your point, Roger, is get an abstraction right like that. If you think about a lot of the big innovations that you see, often it's an abstraction, and it's the right one, and somehow they get it right. + +So that's the other thing, it is very important, and practicing designing abstractions, and honestly, getting it wrong; practicing, and making those mistakes, and living with the pain, and then you learn from that. And I think there is an element of takes -- it takes a lot of experience. I don't think any of these general rules apply in every case. + +**Roger Peppe:** If anyone wants a good exercise for practicing abstractions - people I mentor, I usually give these exercises. You need to write temperature conversion code. But now you need to write it in 20 different places, and they have to be conceptually different design-wise, and the different properties that they optimize for. For safety, for flexibility, for ease of maintainability, for prototyping, and all these aspects. The first five may become easier, then it starts to become more and more difficult every single step of the way, but it makes you think of these different ways how you can implement things. + +**Mat Ryer:** That'd be a good talk... + +**Roger Peppe:** That sounds like an interesting exercise, yeah. + +**Mat Ryer:** Yeah. I'd like to see you do that in a talk. + +**Roger Peppe:** Then he would ruin all the future things for his mentors to learn... + +**Egon Elbre:** Oh no, I can pick any different problem... + +**Mat Ryer:** Yeah, the temperature thing reminds me... In my school, I was the first year to do IT as a GCSE, but the school just wasn't ready for it. No one really had computers. They were teaching how to do spreadsheets, and word processing, and things. But I was into computers from a very young age, so I really loved them... And one of the questions is, of these devices, is it an input device, a storage device or an output device? Monitor. My turn to do a quiz. Roger, monitor - is that an input, output or a storage device? + +**Roger Peppe:** Input and output. + +**Mat Ryer:** You can only pick one, Roger. Otherwise you get no points. + +**Roger Peppe:** Oh, you can only play one? Oh, okay. + +**Mat Ryer:** Oh my, feel my pain. + +**Roger Peppe:** Oh, no... + +**Mat Ryer:** The worst was hard drive. Hard drive - it's like, how are you getting things on and off if it's not also input and output? I left those notes in the side, on the exam, of course... + +**Roger Peppe:** Of course you did. + +**Mat Ryer:** Yeah. Pedantic. But the last question was -- it just said, "Design or explain how you would regulate the temperature in a swimming pool." And I wrote the BASIC code; because I used to write BASIC at home. So I wrote this little BASIC program that did that... And I don't know, it was just like -- we didn't learn any of that at school, so of course, they were surprised, to say the least... But I did alright. + +**Break:** \[45:15\] + +**Mat Ryer:** Egon, you made a point earlier about features... Because I think that is another way to avoid bloat. And a feature, when you imagine it, say, in a website, and it shows up - it just seems like a logical thing to have. But sometimes that feature adds quite a lot of complexity to the system. + +**Egon Elbre:** Yeah. + +**Mat Ryer:** And considering the actual cost, the engineering cost, the maintenance cost really, of features, I think it's something that gets overlooked a lot. Where I am at Grafana, because all the leadership, we're engineers, that is understood from the beginning. It's implied we are always thinking about that. I've worked in places that are perhaps more sales-driven, or product-driven, and that is a fight that you have to have. But it's such an important point, I think, for people. The features, the cost, maintenance... Maintenance more than how long is it going to take to build it even, because you're maintaining it for much longer, hopefully. + +**Egon Elbre:** I think one of the issues with features is that they are really hard to remove afterwards. + +**Mat Ryer:** Yeah. Because people rely on them, of course. + +**Roger Peppe:** Yeah. You basically can't remove a feature when you've added it. The other thing is that interaction between features - that also leads to code bloat, because these things are maybe non-orthogonal, you're changing this other thing in the codebase, and that interacts with all these other features, which have to be updated... And then there's the whole thing about, "Well, you're designing this feature to fit this use case, and you can narrowly fit it." Or you can say, "Well, let's think a little bit wider here. Maybe if we make this a little bit more general, we can cover not just this use case, but also all these other potential use cases, and make a very generally useful feature." Then somebody asks for a feature later, and you say, "Oh, no, you don't need that feature, because you can use this other feature that we already built." I have to say, that's hugely satisfactory when that happens. It doesn't happen often. + +**Mat Ryer:** Yes. That is also a thing I think people should always ask themselves when you're considering how to solve a problem. This is why I always like it whenever use cases come in, when they focus on the problem. It's very easy for people to write a ticket to say, "Oh, we need a button here that does this. This is what our customers ask for." + +**Roger Peppe:** Yeah. A user story. + +**Mat Ryer:** Yeah. Really, we want to know what's the problem that they're trying to solve, and then take a step back and see, like you say... And it's like, I can think of so many examples of places where we've had that thought, we found the general one, and then it's just paid dividends again and again and again down the line. + +**Roger Peppe:** Absolutely. + +**Mat Ryer:** So yeah, I think that's great. + +**Roger Peppe:** That whole take a step back thing is actually key throughout all of software engineering, I think. Don't just focus directly on what you're trying to solve; think through the code and in the wider situation. When you're reviewing code -you know, this code is addressing this particular problem, but maybe it doesn't need to be, if you take a step back and fix it more generally. Maybe do a little bit more work now up front, but that might save you more bloat in the future. + +**Mat Ryer:** Well, it's that time... It's that time that -- we always have a time on Go Time, and it's this time... It's unpopular opinions time. It's time also for the theme tune. + +**Jingle:** \[50:26\] to \[50:46\] + +**Mat Ryer:** Roger, would you record a fiddle accompaniment to that, so that we can play it next time? + +**Roger Peppe:** Alright. Done. + +**Mat Ryer:** Oh, that's gonna be great... Thank you. \[laughs\] And we'll get your TinyGo music machine as well, Egon, which we already have a recording of... Okay, so who's got an unpopular opinion for us today? + +**Egon Elbre:** I think I already said it, but you should review every line - it's either direct or indirect of your dependencies - with similar standards as your own codebase. + +**Mat Ryer:** See, I think that might be a good one for an unpopular one. I think that might be unpopular. So you're saying before you import a package, read every line... + +**Egon Elbre:** Yes. + +**Mat Ryer:** What about its dependencies? + +**Egon Elbre:** Those as well, of course. + +**Roger Peppe:** And presumably, every time you update to a new version, right? You have to review all the changes, too. + +**Egon Elbre:** Yeah. I think people import packages assuming they are better than their codebases, right? And if you have such really high standards for your own base, why wouldn't you have those as well for other codebases? + +**Mat Ryer:** Yeah. Isn't it about responsibility? + +**Egon Elbre:** Yeah. + +**Mat Ryer:** Sometimes there are packages out there and they maybe have a couple of stars, they're not really used by anybody... But there are clues, there are signals to look for for packages that are sort of a bit meatier and a bit more stable. Is it our responsibility? + +**Roger Peppe:** \[52:06\] I have to say that I don't do that. What I do tend to do is when I look at a dependency, I look through the code and say, "Does this look what I would expect to find in a package that is implementing that functionality? Does it generally look good, feel about right? Is this too big for what it should be?" And I quite often say, "No, no, let's not that have that. This is horrible." + +**Mat Ryer:** Yeah. "Does it have tests?" That's a quick check... + +**Roger Peppe:** To be honest, I struggle to understand my own code, let alone other code that I'm importing... So I couldn't understand the gRPC code, the HTTP/2 code... You know, somebody's spent man years, man centuries probably working on that code, and there's no way I could meaningfully review it, I don't think. + +**Mat Ryer:** Man or woman years. They years. They years. \[laughter\] That's the modern measure of time. + +**Roger Peppe:** Sorry. + +**Mat Ryer:** Yeah. That's all right. + +**Roger Peppe:** Person years. + +**Mat Ryer:** Person years. Yeah... Yeah, go on, Egon? What do you reckon? + +**Egon Elbre:** But yeah, I think it has many benefits, if you do review them, even if you don't understand then, right? Because if you are reviewing other code, you might look at how they build things. Maybe you learn something new, right? Maybe you discover some bugs that need to be fixed, because you have different expectations of the codebase. + +Of course, when I say you review them with the same standard, you don't worry about formatting details, or they decided to use spaces instead of tabs. Like, let them have those spaces... + +**Roger Peppe:** I do. No way can they use spaces. Just, absolutely no. No. Out. \[laughter\] + +**Egon Elbre:** And also, many of the -- there have been many attacks against code injections, and -- that you should be worried about, right? + +**Mat Ryer:** It's one of those that's hard to argue against what somebody said. It's like if someone says you should have all this extra security; it's hard to say, "No, we shouldn't." But practically... + +**Roger Peppe:** I think it's easy to argue against, because I don't think it's reasonable. I don't think it's possible for most people. I think that we should have better supply -- you know, better dependency chain assurances, honestly, that we do. + +**Mat Ryer:** But until then... + +**Egon Elbre:** I do think there is a case where you don't need to review them. If you're actually paying someone else to maintain that codebase, then you're intentionally offsetting your own responsibility to somebody else. I do consider every dependency as your own responsibility to maintain, and to fix things when things go critically wrong. + +**Mat Ryer:** I mean, with that rule in place, you will certainly be incentivized quite strongly not to include many dependencies. So it has that effect, too... \[laughs\] + +**Egon Elbre:** Of course, yes. Like, do you really want to read the gRPC? + +**Mat Ryer:** I do now... I'm going to print it out, and take it to bed. + +**Egon Elbre:** Maybe you can make a bed out of it already. + +**Mat Ryer:** \[laughs\] Yeah. Nice. Okay, Roger, have you got an unpopular opinion for us? + +**Roger Peppe:** Yeah. My unpopular opinion is that I think that often just working on a laptop with just a small screen - not on my big desk, with multiple monitors... Just my laptop, with my little screen, and just a keyboard, maybe just on a sofa or something, it's actually more productive than working on my big desk, with multiple monitors, and all the stuff. It seems that way to me. + +**Mat Ryer:** Yeah. So is this about like there's no distractions, and you're just in that world? + +**Roger Peppe:** I think it might be. I don't really understand it, to be honest, because it should be that I've got all these -- I can see all the things, I can do all the things, I've got the nice, rising desk, all the stuff that's absolutely optimized for me... And then I can sit on the sofa with my laptop, and suddenly the code starts to flow. + +**Mat Ryer:** What do you think of that, Egon? + +**Egon Elbre:** \[55:58\] I do think it's nice. I occasionally go to coffee shops to program there, and it does give you a bit more focus. + +**Mat Ryer:** Yeah. I also wonder, Roger, if it's because most of your time has been spent there, rather than -- if you think about in the past, you didn't have all this tech and tease monitors and this desk that moved up and down. + +**Roger Peppe:** In the past you didn't have a laptop. Laptops are a new thing, really... \[laughs\] + +**Mat Ryer:** Hm... You just had to have your monitor on your lap if you wanted to do that. + +**Roger Peppe:** It was a bit big and heavy... That big 17-inch thing. + +**Mat Ryer:** By the way, instead of those desks that go up and down - it's a bit more expensive, but you can actually just have your entire floor move, and same effect; have your desk fixed, and then you just move with the floor. + +**Roger Peppe:** It's funny, because I've got this rising desk, which technically has four saved positions in there... And I actually never use it anymore. I just stand. I started off and I'd stand and I'd sit and now I just, I might as well just get a desk of that height. + +**Mat Ryer:** Oh yeah, yeah. + +**Roger Peppe:** I never sit down. Apart from when I'm on the sofa. Maybe that's the reason. + +**Mat Ryer:** But why do you need four? You've got standing, sitting... Is there one where you're on your knees, praying, really hoping the code is going to work? + +**Roger Peppe:** My theory is that it's for you and one other person. + +**Mat Ryer:** Yeah, that's my theory, too. But for comic effect, I pretended not to know that. + +**Roger Peppe:** For comic effect...? \[laughs\] + +**Mat Ryer:** Yeah, it doesn't always work... + +**Roger Peppe:** Ha-ha-ha. + +**Mat Ryer:** Ha-ha! Thank you. \[laughter\] Let me ask you this, Roger, as well. Are you a keyboard wizard type? Do you use Vim, Emacs, that kind of thing? + +**Roger Peppe:** I am absolutely not. We might have gotten into this, actually... Because a previous unpopular opinion of mine was that Acme is the best text editor. And Acme is a radically mouse-focused text editor. I don't even touch-type, actually. I never learned, and I never went back to properly learn. + +**Mat Ryer:** How do you get the keys tapping? You blow on them, but you refuse to touch it? + +**Roger Peppe:** Yeah, just kind of -- you know, two finger, very slow, look for the key... + +**Mat Ryer:** I worked with a guy who typed with a single finger, and he was faster than me, and I've never seen anything like it since... I hadn't seen anything before, I haven't seen anything it like it since. + +**Roger Peppe:** Not even two hands? + +**Mat Ryer:** Not even two hands. Single finger, one hand. It's just how we learned... And I've no idea why that happened. + +**Roger Peppe:** I got three fingers, maybe four, because one for the Shift key. The little finger for the Shift key. + +**Mat Ryer:** Oh, I assumed it was two if it's someone else... Like, two fingers for you to type, and then you've got two to spare, if you want to help out someone else's... + +**Roger Peppe:** So I think the mouse is great. I don't like keyboard shortcuts in general. They get in my way. + +**Mat Ryer:** Egon, what sort of editor are you into, Mate? + +**Egon Elbre:** Oh, I use three different ones. + +**Mat Ryer:** Okay. Thought it might. + +**Egon Elbre:** So I use Sublime... + +**Mat Ryer:** I thought it wouldn't be a simple answer. Something told me... Yeah. Sublime? + +**Roger Peppe:** I use Sublime, VS Code, and GoLand, depending on what I'm doing. And I guess I edit the commit messages in Vim, because I haven't bothered to change the default editor. \[laughs\] + +**Mat Ryer:** Fair enough. That's pretty cool. Yeah, I always wonder about these people that are amazing on the keyboards. I always look up to them, but maybe it is just that they're rubbish with a mouse. Like, they just can't do it. It could be that. + +**Roger Peppe:** \[59:10\] I actually quite the touchpoint thing. So I don't use the touchpad; when I do use my laptop, I don't use the touchpad, I use this touchpoint thing, which is only on Lenovos. + +**Mat Ryer:** Why is that? + +**Roger Peppe:** It's just a little nipple that's in the middle -- a red thing which is in the middle of the keyboard, and I think it works really well. + +**Egon Elbre:** Yeah. I do miss my touchpoint, so... + +**Roger Peppe:** Interesting. + +**Mat Ryer:** Yeah, so it's like a little thing with a grip on it, isn't it? And it's essentially a joystick. Like a tiny joystick. + +**Roger Peppe:** Yeah, a joystick, except it doesn't move; you just kind of push it. And it does lose its grip after a while, after a year or so, and you have to get another, a replacement thing with more friction. + +**Mat Ryer:** Or you could moisturize your hands... + +**Roger Peppe:** Yeah, actually it's really bad -- if you're really sweaty, it can be a problem. \[laughter\] + +**Mat Ryer:** Well, speaking of being really sweaty, my unpopular opinion -- I've got one today... I don't understand shower gel. I don't know if you've seen it or you have it. It's basically -- it's like some kind of bath time slime... And it never smells great. And people have it. Sometimes I'll go to my brother's, and I'll go in the shower, and he doesn't have any soap. He just has these bottles of shower gel. And it's sort of just slimy, and... I don't know, I never feel clean. I'm not happy with it. What's your stance on shower gel? + +**Roger Peppe:** I tend to agree with you about the shower gel, because -- it just all drips away. You know, if you give me a bar soap, it's there, right? + +**Mat Ryer:** You can really go to town. + +**Egon Elbre:** Don't you use a sponge for it? + +**Mat Ryer:** Oh, a sponge? + +**Roger Peppe:** A sponge? No, no... + +**Egon Elbre:** Like, something fluffy. + +**Mat Ryer:** Oh, that makes sense. That does make sense, though. + +**Roger Peppe:** I suppose that does make more sense. No. + +**Egon Elbre:** Do you also feel the same about shampoo? + +**Mat Ryer:** No, although I realize that that is very similar. I don't have loads of hair... + +**Roger Peppe:** I have to say that I have recently, like in the last couple of years changed to using bar soap for the shampoo. You can get these shampoo bar soaps, and it works really well. So you don't have that bottle which you have to throw away, so less plastic, it lasts for ages... It actually works really well. + +**Mat Ryer:** Sold. + +**Roger Peppe:** I totally recommend it. + +**Mat Ryer:** I want everything in a bar. I want a bar of toothpaste now. You just rub the bar in... \[laughs\] No, but a shower gel -- in some places where the water soft, it doesn't even all properly come off, so you end up being all slimy. + +**Roger Peppe:** You're putting it on during the shower, not afterwards, right? + +**Mat Ryer:** \[laughs\] To be fair, I have not read the label. I maybe use an error, but I just don't get it. I just don't like it. I like a nice, rough bar of soap; something that's rugged, and... You know what I mean? I'm not one of those really manly people. + +**Egon Elbre:** Steel wool. + +**Mat Ryer:** Steel wool. Oh, yes, please. I like a towel. I went to a hotel - it was a bit fancy - and the towels were all soft, and it's horrible. I want a rough towel. I want it like an elephant rubbing up against a tree, please. + +**Roger Peppe:** A loofah. + +**Mat Ryer:** A loofah, yeah. Well, I'm afraid that's all the time we have... But I learned a lot, mainly about bath things, but also about code bloat... So thanks very much to my guests. Roger Peppe - always a pleasure, Roger. Hopefully, you'll come back soon. + +**Roger Peppe:** It was a pleasure. + +**Mat Ryer:** Yeah. And Egon Elbre... + +**Egon Elbre:** Áeah it's been emotional. + +**Mat Ryer:** Good. It's been great to have you. We'll see you next time! diff --git a/Berlin's transition to Go_transcript.txt b/Berlin's transition to Go_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..3d5eb29c27309a4fe3ff220356c3d64c192bce95 --- /dev/null +++ b/Berlin's transition to Go_transcript.txt @@ -0,0 +1,317 @@ +**Natalie Pistunovich:** So welcome, everyone. It's great of you to join us through this episode about the Berlin ecosystem and how the slowly transition to Go has happened. And I am joined by my co-organizer, Ole. Hi, Ole! + +**Ole Bulbuk:** Hi! Nice to meet you. + +**Natalie Pistunovich:** How are you doing? + +**Ole Bulbuk:** Yeah, I'm fine, thanks. Nice to meet you. Great honor to be on the show. + +**Natalie Pistunovich:** Yeah, we've been zooming for two years now... Once the meetup is gonna go back to in-person, it's just gonna be weird, at some point. Just as it was weird transitioning to Zoom, it's gonna be weird transitioning, but we'll manage. + +So Ole, you are a backend engineer since the '90s. That's your fun catchphrase. And you've been working in different companies that are big and small, and you've had lots of projects that you saw fail and succeed. And you love being part of the Go community, and you're working for Ardan Labs... And hey, you're my co-organizer at my Go user group... Since we were just trying to remember, as we were preparing for this call - we've said, what was it, 2017, 2018? + +**Ole Bulbuk:** 2017, I think. + +**Natalie Pistunovich:** Yeah. And you are writing open source software in Go. But also maybe not. + +**Ole Bulbuk:** Nowadays in Go. + +**Natalie Pistunovich:** \[03:56\] Yeah. So backend engineer since the '90s... Before we dive into what does that mean, for those of you who listen and don't know, Ole and I are co-organizing the Go user group, together with Tim who is not joining us today... And we've been meeting in-person for a while, and then we've been meeting on Zoom for a while, and we're still on Zoom, but we do look to moving back to in-person sometime soon... And the Berlin user group actually exists since 2011, which is I think before both of us were using Go... Which is pretty fun. So that was when it was very early released... And definitely before any stable version came out. So the Berlin Go community is old, but it's also definitely growing, so it's always changing. What is your impression of the Go user group in Berlin, or the Go community, also those who are not coming to the user group? + +**Ole Bulbuk:** I'm very happy over that, and to be part of it, of course. I think in Germany at least it seems to be by far the biggest and most active, so I'm very happy with that. I can assure you, when you live in Berlin, you can find enough companies probably for the rest of your developer life, only working on Go. So this is quite nice, I think. + +**Natalie Pistunovich:** I definitely can say the same. I would agree with this. Okay, so back to the original catchphrase of your introduction... You've been a backend developer since the '90s... So you were the OG hipster in Berlin. + +**Ole Bulbuk:** Yeah, in the '90s I've been working with Java, and we built our own frameworks for backend stuff... And I had \[unintelligible 00:05:42.23\] in between, and this remote method invocation, or something... It was a bit like this Cobol thing, and... Yeah, most people don't remember, probably. And we were building our own frameworks for each project, kind of, and this was just everything done yourself; you would be the true inventor sometimes, when you were one of the first on the project. + +Nowadays you have a bit more of standards, and libraries and so on, and you're not supposed to build everything yourself. But in the old days this could be a lot of fun... But yeah, when you had someone who had fun with it, but didn't know the requirements, it could go really bad, too. + +**Natalie Pistunovich:** I guess Go has also not that many frameworks, but... Also I guess less just common to use one. + +**Ole Bulbuk:** Yeah. And one thing I love about Go also is that if feels like the old times of Java, where I could say, "Oh, this is just Java code, and I can dive into it and understand it quickly", and this works pretty well with Go, too. Often, the common libraries aren't \[unintelligible 00:07:10.13\] with lots of abstractions and other things. They're usually quite straightforward and easy to understand. Occasionally, I can quickly find a bug and fix it, or something like that. This is all nice. + +**Natalie Pistunovich:** Yeah. So after Java, what was the language or languages that you were using? + +**Ole Bulbuk:** Well, I got fed up with Java a bit, and then I tried Python for a while... And Python the language I don't mind too much, but back then it was all compared to Ruby and PHP, and \[unintelligible 00:07:50.20\] job at an agency, and this was really tough and not relaxed working there... And especially when things didn't turn out the way they were originally planned somehow. This wasn't so much fun. + +\[08:12\] Then I went back to Java, because I thought, "Well, the language might be nicer", but you know, the working conditions were not... And then I just went back to Java, and then I found Go later. It started a little bit when it first came out, 2009, 2010, or something; I started using it a little bit for a small pet project. + +**Natalie Pistunovich:** Wait, so how did you hear about it? + +**Ole Bulbuk:** Oh, just in the news somewhere. + +**Natalie Pistunovich:** It was not the German news, probably. I think by the first time Go was mentioned in the German news was like 2017. \[laughs\] I'm kidding. Definitely not around release time. + +**Ole Bulbuk:** Yeah, yeah... So it was way before 1.0 at least, and it never caught up to 1.0. I stopped it. But it was a nice experience, and I just didn't know where it would go, and whether there would be a true market for it. And then I said "Okay, maybe I should keep with Java a bit longer." + +I kept with that, and then in 2015-2016 I joined the Berlin user group, and understood that Go does have a market that Java couldn't reach. Actually, the lower latency services, and so on, like for advertisement bidding, and cloud tech and so on, where you need low latency - this can sometimes work with Java, but when the garbage collection kicks in, it doesn't work at all anymore. + +So I started to see that there is really something to it that makes total sense, and I got more and more into it, and was happy to get a full-time job when I got a chance. + +**Natalie Pistunovich:** Would you say that the ecosystem, the techie ecosystem of Berlin in the '90s was mostly Java? + +**Ole Bulbuk:** I think no. I've been a few times told that it would be easier to get a job if I would be a C\# developer instead. + +**Natalie Pistunovich:** C\#... + +**Ole Bulbuk:** Yeah... So the standard competition back in those days. When you wanted to do business software, it was either Java or C\#. Companies like Siemens - they said "Yeah, we have the odd Java projects somewhere, but we would really rather use C\#, if you wouldn't mind." And then I'm like "Yeah, maybe I'm not such a big fan of it." + +**Natalie Pistunovich:** And how did you see the languages change as the ecosystem was developing, and specifically in Berlin? What languages, what tech stacks did you see over the years through the '90s, through the 2000s, through the 2010s? Give us a travel down the history lane. + +**Ole Bulbuk:** As I said, it started in the wild, old days where you would develop your own framework for everything. Every new project you started with a clean sheet, and then first thought what would be great to have, and then started building that. And later you had in the Java world also all these frameworks. Some very big, others -- like, the Spring Framework has been the small and lean one for many years at least... I saw that evolve, and getting bigger and bigger, and more refined, but also more overhead... And yeah, Java got its generics, that aren't very trivial to understand, especially since Java supports proper inheritance. This makes generics a lot more complicated, too. + +\[12:10\] Then these annotations in Java, that are really nice to write, but very hard to debug, because there's some code executed that you'd never seen, and you can't reach it anyhow with a normal debugger; the annotation itself doesn't even point to the real code. That is just looking for the annotation, and then doing something interesting... And yeah, it's really difficult. + +So I was more and more fed up with it, and I thought maybe the language - there's still some good color to it, and I tried to make people happy and adopt that a bit more. Then I realized that the community really loves these frameworks, and that you can do powerful things in three lines. + +**Natalie Pistunovich:** Were there Java meetups in Berlin at the time? + +**Ole Bulbuk:** Yeah, I've been to Java meetups -- not so much in Berlin. Back in those days I'd been living near Nurnberg, and I've been to Java meetups there. + +**Natalie Pistunovich:** Around what year was that? Was it the 2000s? Before, or after? + +**Ole Bulbuk:** Around 2010. I moved to Nurnberg from Berlin in 2008, and moved back in 2015. And so like 6,5 years. + +**Natalie Pistunovich:** And so until 2008 you were then in Berlin, so it was a lot of Java. + +**Ole Bulbuk:** Yeah, at least my projects were all Java, yeah. But of course, other languages, too. As I said, C\# would have been sometimes easier to get a position for. + +**Natalie Pistunovich:** When I joined the Berlin ecosystem in 2013 or so, I saw a lot of PHP everywhere. + +**Ole Bulbuk:** Ah, okay. Yeah. + +**Natalie Pistunovich:** When did that happen? When did you see that transition happen? + +**Ole Bulbuk:** Oh, PHP -- I mean, this was a completely different community, I think. I think we never really had a competition between those. Bigger corporations and those who were willing to invest money, they used Java, and then \[unintelligible 00:14:21.11\] agencies and people who wanted to get something quickly up and running, and for cheap money, and use a lot of interns, or whatever - they were using PHP. + +**Natalie Pistunovich:** And would you say that Berlin, let's say before 2010, so like in the '90s and in the 2000's, was it more startups, or was it more corporates? + +**Ole Bulbuk:** The startups started a bit later. I think in the 2000s it was more corporates. At least what I saw. And then small \[unintelligible 00:14:56.14\] companies doing something. When I left, there were the first startups. I think SoundCloud was one of the first. I've been eyeing Berlin, also possibly going back there in 2012 or something, or 2013, and it still didn't look that professional to me. I didn't take a thorough look, I have to admit. And SoundCloud - you didn't know how they would earn money, and so on, and they didn't find a really good way. And I thought "Well, if it's either not so interesting technology-wise, or you don't know if they do have the business model or not, then maybe I stay away." Then I eyed London back in those days, and realized that London is just too hard; you have to be too focused. + +**Natalie Pistunovich:** Yeah. + +**Ole Bulbuk:** And then a few years later I took another look at Berlin, and I thought "Hey, it has developed in a good way, and I am happy to go back." + +**Natalie Pistunovich:** \[16:09\] So around the 2010s is when Berlin started getting some startups, and it also started -- it translates basically to more PHP developers... But also, that's around the time that Go was joining the awareness of the world... So Go kind of became online. And this is also around the time that the Berlin Go user group started. So around that time - what can you tell us about Berlin in the early 2010s? The ecosystem, and the companies, and the languages, and why somehow it is my feeling that there was more PHP is accurate, or is it just what I saw? + +**Ole Bulbuk:** Possible. I mean, we had Ruby a lot also. Not only PHP. + +**Natalie Pistunovich:** Oh, true. True. Very good point, yeah. + +**Ole Bulbuk:** A lot of these startups, when they aren't very technical, they have been using Ruby a lot. + +**Natalie Pistunovich:** True. + +**Ole Bulbuk:** And sometimes PHP. And then very technical ones, like SoundCloud, they used Go. + +**Natalie Pistunovich:** Yeah, there were some early adopters of Go in Berlin, among other startups. + +**Ole Bulbuk:** Yeah, and this is good for us though, right? We are still living a little bit of that... Like \[unintelligible 00:17:27.05\] We still have him at our meetups sometimes. + +**Natalie Pistunovich:** Yeah. A co-creator of Prometheus. + +**Ole Bulbuk:** Exactly. + +**Natalie Pistunovich:** Which is written in Go. + +**Ole Bulbuk:** Yeah. + +**Natalie Pistunovich:** So in the early 2010s, going down the history lane - so around that time basically more startups started being created in Berlin, it was not just the corporate world, so it translated technologically to not just Java, but also Go, more PHP, more Ruby, as you said... And then how did you see the last decade, between the early 2010s to the early 2020s, which is today? + +**Ole Bulbuk:** I think it all matured a lot more, and we have - especially in the Go community now - adoption beyond the classical startup community. It's not only real startups, or ex-startups like Amazon or Google that are using Go, but we have some kind of old-school companies also. + +I've been to an old-school logistics company, invited for an interview, and they were using Go sometimes. So that's really nice. I am quite happy that that broadens, and adoption goes up, because it changes the community too, and you have different topics, to see what is useful now, and so on. We have to adapt as a meetup also, because the requirements are just a bit different. + +**Break**: \[19:04\] + +**Natalie Pistunovich:** I think one interesting thing about Berlin becoming more of a startuppy ecosystem in the tech is that there were a lot of people who immigrated to Berlin, and that kind of made Berlin very -- it probably definitely was very international, but also the tech of Berlin remained very international. So when you kind of first think of technology in the context of Germany, one might think of the car industry, and the very traditional German industries... But that's very much not Berlin. Our user group I think pretty early on was in English, and definitely since then remained in English, and I think a very large part of the developers that I know - they are only speaking English and some German. + +I only worked in startups in Berlin, I never worked in a large company, but I can say that in all those companies that I worked it was at most one third German speakers. Do you think it has something to do with the adoption of languages like Go and like Ruby? + +**Ole Bulbuk:** You mean the internationality with -- + +**Natalie Pistunovich:** Yeah. + +**Ole Bulbuk:** Somehow it's linked, but I don't think it's a cause and effect directly. I mean, the traditional companies and the startups compete for employees I think in a bit different way. The traditional companies say "Okay, you get some okay(ish) salary, and you can stay here for many, many years, and you will be safe", and they attract a lot of people who want stability. And the startup scene attracts people who want to see something interesting, and something new, and want to try out things, even if it doesn't always work, and so on. This is just different kind of people also. + +**Natalie Pistunovich:** Yeah. + +**Ole Bulbuk:** So this is a bit of different market, and they are more international since they say "Well, it doesn't matter where you come from. In the end we need your code; this has to be good. And when you can work well along with all the others, we are fine with that." So they can get great employees from all around the world, kind of. And that's nice, of course. + +But in the traditional companies they have a culture where everything is in German, and this would be very difficult for them to compete on that level. So they are just two different markets, I think. + +**Natalie Pistunovich:** Yeah. To the listeners who are not familiar with the concept of -- or generally with how German work contracts work, it's common that you get a contract that mentions a probation period of average six months, during which it is possible to terminate the employment within a week, or a couple of weeks notice; it kind of depends. And this is mutual. So if you don't like it, you can quit, and give a few weeks' notice; if the employer doesn't like something in the setup, it's possible. + +\[24:12\] But then after those probation six months it's becoming kind of permanent position, and then as an employer it's becoming significantly harder to fire. So you have to show that you have brought it to the attention of the employee that the performance is not as expected, and then you created a working plan together, and then you revisited it several times, and then as employer you've put some efforts into training, and giving some resources to the employee... So it's quite a long process to fire an employee. + +And from the employee side, for you to quit it's something like three months notice as an employee, and if you're in a management position, it can also be six months notice... So it's very different from some of the scenarios from other places might know. + +So when you say all of this stability - this is the thing that people definitely would like in a company, and it's not unusual that people work 5 and 10 years in the same company, even if it's a technical company; it's not necessarily a startup, but it's a technical company. + +So these types of contracts are a legal requirement in Germany, so you have the same, also in a startup... But I guess less people are looking for that. And especially if we kind of go back to ten years ago or so, when Go just started and some companies here adopted that. The people who would be keen on trying this and don't necessarily live in Berlin - the profile of such people would be in people who don't have lots of commitments, for example kids who have to go to a specific school and they are able to easily relocate, and then they're like "Oh, I wanna work in Go. I will relocate to Berlin." But the profile of such people is many times also not necessarily speaking German. + +So that's kind of how startups got this interesting culture that is a very Berliner thing; everybody has an accent and speaks like in a third language... And not just a programming language. \[laughter\] Would you say this is the same what we see in our meetups, that this is people who are international people, who usually speak some languages, and usually are not from Berlin? + +**Ole Bulbuk:** Yeah, definitely. + +**Natalie Pistunovich:** Would you say that the Go developers that you know switched mostly from Java, from Ruby, from PHP, from some other language? + +**Ole Bulbuk:** They have many different backgrounds, I think. It's from JavaScript, from Ruby, PHP... Yeah, quite a lot from Java also. I can't remember a single Java project that I did in almost 20 years that I nowadays wouldn't rather do in Go... So yeah, probably it's natural that we have a lot of ex-Java developers in the Go community nowadays. But yeah, otherwise there are a lot of different languages. Also - it's rare, but sometimes you have an odd C and C++ developer who didn't like that complexity and so on anymore, and hunting down memory bugs, and... Probably they moved on to Rust now, but maybe there's still someone. + +**Natalie Pistunovich:** Yeah. Interesting to see. Would you say that you saw some startups built with Rust recently here? + +**Ole Bulbuk:** I know some companies who are hiring Rust developers, but I don't know if they are pure Rust companies, or something like that. I suppose Rust would be too tedious to develop your whole business on it. But some very technical parts probably make sense, and can pay off. + +**Natalie Pistunovich:** The Go companies that you know from Berlin - would you say there are more that started in Go, or would you say that there are more that changed to Go, or something in between, that some just rewrote or added services in Go, so something in between? + +**Ole Bulbuk:** \[28:12\] Yeah, I think there are different companies. Some were just built up on Go, like these advertisement \[unintelligible 00:28:18.14\] who saw that this is the language they need to be able to develop quickly and efficiently, and also have the short latency that they need to be able to compete. + +Then other companies like the fintechs, who wanted to do banking, but in \[unintelligible 00:28:39.02\] so they used Go. So they started all with Go. + +And then there are bigger companies that often are a bit more polyglot, and then sometimes they have a Go project and sometimes they use Java, or Kotlin, or whatever else. And yeah, I think these two types I see quite a lot, and some in between, of course, that still have something old running somewhere, or are trying to get rid of it, or have been able to switch fully to Go after they once started with a Ruby or PHP prototype, or first version, or whatever you want to call it. + +I don't think this is bad. I can imagine starting something with Ruby or so, getting something up and running quickly, and when you see that the smallest instance of some virtual machine on Amazon or so doesn't work out anymore, and you have to scale, then it's maybe a good time to switch to Go, because you seem to have a product-market fit, and then rebuild this with \[unintelligible 00:29:50.27\] invest some money, because you earned some. It could be a model that works well for me. + +**Natalie Pistunovich:** Yeah. With adoption of Go by enterprises that happened in the recent few years - how do you see this reflect in the Berlin tech ecosystem? + +**Ole Bulbuk:** I see that we even more use web APIs. It's even more important than before. And otherwise, they aren't so visible, usually. They don't do lots of talks, they are often more consumers than producers... + +**Natalie Pistunovich:** Of content? + +**Ole Bulbuk:** Yeah, of content, and... Also, I remember when I've been at these companies myself, or worked for them while working for a consulting company, there wasn't this culture of producing something and showing it around so much. You sometimes showed something, but this was a big thing, and you would think about this a long time, and so on. This was really rare, and you would expose yourself a lot, and do this maybe once a year, or something, and get comments about it like the next three months or so. It was a big deal. They don't have such a culture. + +And I think it's also got to do with this mentality that there are more risk-averse people attracted by those companies, so they don't like to expose themselves so much. And this is probably the reason why we don't see them so much. + +**Natalie Pistunovich:** As speakers in the meetup, yeah. + +**Ole Bulbuk:** \[31:48\] Yeah. Sometimes you see someone, especially when it's a bigger conference or something, or you've got someone who got paid for it, or so... But we usually do it preparing something in our spare time, and then going somewhere, wherever people would be willing to listen to us... \[laughs\] This is much more rare though, I think. + +**Natalie Pistunovich:** Yeah. And startups is kind of different, right? + +**Ole Bulbuk:** Yeah. Startups is just doing a guess, trying it out, and then see what feedback comes back. And then adopt, and do again. The same you can do with interacting with other people, with presentations, with giving talks, or going to a podcast, or whatever. + +**Natalie Pistunovich:** Yeah. It is interesting to hear what large enterprises are doing, and it is harder to reach, but they definitely can tell different stories, especially about the scale. So if any enterprise gophers are listening, please give a talk. We want to hear your talk at our meetup, or any local meetup that you have. + +**Ole Bulbuk:** Yeah, definitely. I've seen some applications, and scale wasn't such a big topic there. So it's not like the traditional -- or Google scale, like thousands of requests, or millions of requests per second, or something like that. But I've seen more as a scale problem in enterprises was that you had to scale to many developers, and to a large number of people working on the overall project. So the whole development process was all of business analysts and whatever else you had would have to work nicely, and so on. And this kind of scale is, I think, the more important one, and it's the biggest part of the corporate world. + +**Natalie Pistunovich:** So the Go meetup - how did you hear about that, and when? + +**Ole Bulbuk:** Oh, I think this was when I tried or was thinking about moving back to Berlin, and I wanted to dive into the tech scene deeper, and find out how it is, and what it looks like. Then I was just searching for meetups in Berlin, and found a few, and then the Go meetup was one of them. + +**Natalie Pistunovich:** So you were searching for technical meetups in general, not specifically the Go one. + +**Ole Bulbuk:** Exactly. When I first went to Berlin back, I've been starting a Java job here. And this company has later switched to Ruby. And I started doing the switch with the company, but I realized for becoming as good in Ruby as I've been in the Java world before, I would need like ten years or something, since the ecosystem and language has too big of a history. This would be very hard to catch up. And then I took another look at Go, and thought "Well, this is way easier to get into, and it's more interesting also that it's technical traits that make it more interesting, and I can build cooler application that would be very hard to do in Ruby or Java, or something." And then I was convinced. + +**Natalie Pistunovich:** And can you compare the Go meetup, how it was when you started coming, versus how it is these days? + +**Ole Bulbuk:** Oh, I think it is a bit different. Back in the days it was way more technical, and every talk was technical, kind of, I think. At least those that I remember... \[laughs\] And you had always some people talking about what they have done, or were planning to change on the Go compiler itself, or the toolchain, or something... This was way more common also, to hack on Go itself. + +**Natalie Pistunovich:** Mm-hm. + +**Ole Bulbuk:** \[36:07\] And it's still doable, and it's possible, but it's not that common anymore. I think the community has changed also; you have a harder time to find those die-hard techies nowadays. + +**Natalie Pistunovich:** So how would you describe the talks these days? + +**Ole Bulbuk:** Yeah, these days still we have of course quite some technical talks, but they don't dive as deeply, and don't talk half the time about a few microseconds that have been saved somewhere, usually... + +We still do have interesting technical talks, definitely, but it's not all about it. We see that there are a lot of other challenges in real life, and for many projects Go is just quick and everything enough. They never optimize something for years or so. They never see a reason to optimize a single bit about their Go code somewhere. You just don't talk so much about it; even if it's interesting to listen sometimes, but it's not that useful in the end. + +**Natalie Pistunovich:** I would probably describe this somewhat similarly. The first talks were about Go, and kind of on the language itself, and the more recent ones are things I do with Go at work, things I do with Go for my fun... But yeah, it's a good distinction, the difference between in Go and with Go. I wonder if it's the same in conference talks also. It would be interesting to compare that over the years. + +**Ole Bulbuk:** I think this changed a little too, but yeah... + +**Natalie Pistunovich:** It probably has something to do with the fact that in the beginning, or years ago, the language was more like a hipster thing to try for fun, and people who were doing that and coming to talk about that were doing this because they want to, and now there's more and more people who are gophers because that's their job. And they may have hacking, fun projects in Rust, or Haskell, or just no hacking projects. So yeah, you get kind of more practical talks. + +**Ole Bulbuk:** Yeah. What do you think is the biggest hurdle of more adoption of Go in Berlin? + +**Natalie Pistunovich:** I'm trying to think of an answer that does not have the word "hockey stick" in it... \[laughter\] So there's this graph, like a Gaussian, ghost-looking graph of a language, of adoption; in the beginning it's just very few, then there is a huge peak that many people are adopting, and then there is just the repose of not that many people left to be adopters. I think we're towards the end of that peak, or of that ghost. So already a lot of people who have done that adoption. + +In the beginning it was more word of mouth, and reading, like you said, on the news, and so on... And now it's more - yeah, because you have to, because that's your job. And companies will keep adopting Go, but probably because we have kind of behind us, or right now, as fast as it gets is just not gonna be this fast anymore. So we will see gophers joining, but I think it will not be as many; the velocity of joining - the derivative of that is not gonna be positive for longer. It's gonna be more people, but less every time, kind of join us. + +\[39:47\] This last week, or this week was Google I/O, and this is kind of the conference that Google organizes for people who are community organizers, like us, and who are developers, and for people who participate in their different programs... And Go was mentioned. And before this year, it happened once, sometime I think in 2015, and in between, every year, in Google I/O, it was just not brought up... Although this is a language that came from that origin. So that's another interesting signal. + +I asked you earlier how did you join the Go community... So the way that I joined the Go community is when sometimes after I moved to Berlin I just went to a conference; I don't even remember how I heard about that anymore... But it was the DevFest, which is the annual conference of the different Google developer groups in Berlin. So the Go user group, the Android user group, and at the time I think there was also the general Google developer group, kind of; and that's it. There were less Google developer group communities at the time. And they joined forces to give an annual conference, which is kind of a mini, local Google I/O, if you will. + +At that time I was working at a company that was using Go, and then I overheard in the hallway a conversation, something about Go, and I joined that conversation, and then somebody said "Oh, you should join our user group." And this was a person \[unintelligible 00:41:19.21\] who at the time was the organizer of the user group. So he said "You should join the meetup", I joined the meetup, and then he said "You should help me organize", because I was telling him how I used to be as a student very active in the student organization, and doing events for students, and so on. He said it's pretty much the same, it's just the adult version of that. And then that's how I joined on board. + +**Ole Bulbuk:** Cool. + +**Natalie Pistunovich:** That was end of 2013, early 2014, I think. A long time ago. And yeah, so the Go meetup was almost three at the time, so that was already nicely going on... In English, which is definitely fun. It felt inclusive in that sense, that you don't need to be a German speaker to be able to participate. + +**Ole Bulbuk:** Yeah, great. Back to this topic of adoption of Go, nowadays I think we will have -- as you said, the adoption curve will flatten a bit, but I also think the new members and people joining in the meetup will probably flatten a bit more, because the big companies \[unintelligible 00:42:27.04\] who use Go because they have paid for it. And they are less likely to go to any meetup at all, because they are not paid for it, right? And this is not too bad. If we don't see user numbers grow as before, this can be a sign just of maturity, kind of. Still, adoption can grow beyond the rate of the members, or something like that. + +**Natalie Pistunovich:** Yeah, I definitely hope that this signal of including a talk about Go in Google I/O would be a nice push towards that end of that curve, that would bring those people on board. And I definitely hope that this will also mean more talks for our meetups. + +**Ole Bulbuk:** Yeah, that would be nice, of course. But it would be very welcome if we'd get a new wave of talks from companies we've never heard of so far that they would be using Go at all. Yeah, it would be really fun. + +**Break:** \[43:36\] + +**Natalie Pistunovich:** Do you think there's any connection between how the Berlin ecosystem is having lots of gophers, and then lots of companies with Go, and how the Go user group is very early on? So did one influence the other? If yes, which did which? If not, what is that connection between the ecosystem and the Go community in Berlin that you see? + +**Ole Bulbuk:** I think it's always intertwined a bit. We have luckily these early adopter companies, then with some good Go developers, and then this greater Go community... And then there were these companies where people said "Well, we do have enough developers here in Berlin. We can bet on Go." And then they could attract more Go developers from other parts of the worlds. Now we have a nice Go ecosystem here, and quite a bit community... And I think still the number of developers that you can find for a company here is still the limiting factor. + +**Natalie Pistunovich:** Yeah. + +**Ole Bulbuk:** I think when you show to a manager "Hey, Go is optimal for this kind of system that you want to build", then he says "Well, but I asked my developers that I can get and I have already here, and they don't know much about Go." So it's hard to get them to try it. This is, I think, still the limiting factor somehow. + +**Natalie Pistunovich:** Yeah, that's definitely that, and this will definitely make a difference, because it's more happening in larger companies, and larger -- of course, it's harder to convert, but it will bring more... To convert one large company, you have to have a lot of developers joining. + +You mentioned relocating to Berlin for Go... So for those who are not familiar with working in Europe generally as a developer, and specifically in Berlin, in Germany, there are very easy visas that your employer can issue for you. So it's a lot easier than a Green Card. You kind of need a recognized degree from university, and the rest is bureaucracy, pretty much. And it's also possible without a recognized degree, as well. It's slightly harder, but still definitely manageable, and a lot easier than a Green Card. So if anybody is considering to relocate here, just know that it's pretty easy, because there is a high demand. + +**Ole Bulbuk:** \[50:04\] I think the contract that you get from the company is the most important thing. Then with that you can get the rest rolling somehow. Especially from within the EU or something like that, it's quite easy. And when you are from outside, it's getting a bit more difficult, but... + +**Natalie Pistunovich:** Yeah, for those who are inside the EU, you don't need to do anything; you just move here and start working. But for those who want to come from abroad, the visa - it's called the Blue Card; it's rather straightforward. There's no quotas on that to start with. It does simplify things. + +And as anything in the world in the last two years, our meetup as well switched from in-person to virtual. Can you share your insights on that? + +**Ole Bulbuk:** Yeah. I mean, at the start it was quite exciting. We tried to keep it all running and do everything exactly the same as before. After a while we saw that it's not that exciting anymore, and people start to get Zoom fatigue, and whatever. Members were a bit decreasing... And then we saw that we got more international adoption, like people from India, or anywhere in the world joining, which was really nice, too. But in the end, we don't have the same energy as if you really can meet people in-person and can just have a small chat afterwards, and so on. It's not the same relaxed atmosphere that you can create in an in-person meetup. So I'm missing that myself a bit, I have to admit. + +**Natalie Pistunovich:** For sure. And we had some attempts of saying "Maybe next month we'll try" or "Maybe in spring we'll try", but then always numbers were against us, and regulations and whatnot. You mentioned that we had a decrease in attendance, but changed the character to be more international; that's definitely something interesting. And I think one of the things we did that allowed us to kind of stay active throughout this time is to make things regular. + +In the past we met once a month always. Since forever it was always once a month, but it would move around when in the month. Then at some point we anchored the second Wednesday of the month and we started recording, and then uploading this to YouTube... So less attendance, but this instead kind of compensates with more people watching the YouTube... Which is probably convenient, in a way, that you don't have to attend in the evening when you cannot, but you can watch this at the time that you want. So you still get that content, but you definitely miss out on the connection. + +We do look to get back in-person, maybe this spring/summer, which is exciting... And it will change things again. So we'll see if the Zoom fatigue is being replaced with "I'm used to not leaving the house. I'm not gonna come to events" fatigue. Maybe yes, maybe not. Definitely, a global thing to think about, which is maybe something you all agree with, maybe something that you don't agree with, but this definitely is a type of an opinion, just like our Unpopular Opinions. + +**Jingle:** \[53:39\] to \[53:58\] + +**Natalie Pistunovich:** Ole, I heard a rumor, that you have two unpopular opinions. + +**Ole Bulbuk:** Yeah, that is true. The first is a bit of a meta opinion. My first unpopular opinion is that I think the popularity of an opinion depends more on the audience than on the opinion itself. A very good example would be stating that Go is a low-latency, low-overhead programming language. When you tell this to some JavaScript or Ruby or Go developers, they will all nod and say "Well, yeah, that's popular." And when you tell this to some C or C++ developers, or Ruby, or Rust developers or so, they will say "No, that's not true. Low latency. I've seen the overhead is like ten times higher than anything that I've built" and so on. + +So this is something quite different... And I think the audience is a major factor. + +**Natalie Pistunovich:** Okay, so in order to be ranked high in the Unpopular Opinion list of Go Time, you want to optimize for something that is very unpopular with Go developers. + +**Ole Bulbuk:** So I could now state an opinion like "Go is a very slow, large latency and whatever language." I don't think I will find an opinion like that that it's -- I'm a gopher myself, come on... \[laughs\] Maybe I have a second opinion that it's a bit more broader also... + +**Natalie Pistunovich:** Is that optimized for the Go crowd? + +**Ole Bulbuk:** No, it's not. I think it's optimized more in general for the IT crowd. And it's more about how universities and so on work. Because they all do something very nice, abstract, and a lot of algorithms and data structures, or how to create programming languages, and how to study them, or how to work and implement databases, and so on. But in the end, there are only very few students who truly work on this later in real life, when they ever leave university. \[unintelligible 00:56:25.14\] really working on these very interesting problems. But we have like 99.99% of the people coming from universities never get such interesting problem in their lives, so they have a tendency often for over-engineering. + +I think a lot of the over-engineering we see in the real world stems from the discrepancy that we see between the problems we learn to solve at universities and the problems that we have in real life, or lack of problems maybe. + +**Natalie Pistunovich:** \[57:05\] Interesting. I will not lie, I feel a little bit attacked... + +**Ole Bulbuk:** Sorry. \[laughter\] + +**Natalie Pistunovich:** Over-engineering is definitely a topic I have heard in the past... That sure is interesting. I agree with you that, unfortunately, this is not so unpopular, at least in our little sample group. I agree with you that the type of skills and the views on programming that you get in universities is mostly not transferable, and things that should be taught on more is how to read other people's code... Because so much of the work that you do in university is just write from scratch, which is also not something you do often... Definitely not alone. And also, learning to accept and consider things like trade-offs in the context of actually business value. + +It's still hard for me to sometimes let go of doing the right thing versus doing the thing that is more efficient on the business part. I have sinned in that multiple times in the past, and it's still ongoing. Maybe if I would have learned more in university, more programming in the context of "This is the right thing, and this is the efficient thing", it would have gotten this transition easier for me, for sure. So I'm sorry, Ole, I agree with you. I agree with your unpopular opinion... + +**Ole Bulbuk:** Well, sorry... \[laughter\] + +**Natalie Pistunovich:** But we'll have a survey on Twitter and we'll see how many people agree. + +**Ole Bulbuk:** Okay. + +**Natalie Pistunovich:** Yeah, fun. Thanks a lot for joining. We talked about so many different interesting things, and we'll definitely include in the show notes also all the things that we mentioned and are relevant. And join the meetup, join our user group if you wanna see Ole and me. + +**Ole Bulbuk:** Yeah, please do. + +**Natalie Pistunovich:** Thank you, Ole. + +**Ole Bulbuk:** You're welcome. + +**Natalie Pistunovich:** Have a good evening. + +**Ole Bulbuk:** Bye. diff --git a/Bob Logblaw Log Blog_transcript.txt b/Bob Logblaw Log Blog_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..f078f40a78f1448442b990739d0e04690b660b5c --- /dev/null +++ b/Bob Logblaw Log Blog_transcript.txt @@ -0,0 +1,523 @@ +**Mat Ryer:** Hello, and welcome to Go Time. I'm Mat Ryer. Today we're talking about logging; something we all do, no need to be embarrassed about it... But are we doing it right? Are we logging the right things, are we logging them to the right place? We're gonna find out today. + +Joining me - Jon Calhoun, of course. Hello, Jon. + +**Jon Calhoun:** Hey, Mat. How are? + +**Mat Ryer:** Not bad, mate. Not bad. How has your week been so far? + +**Jon Calhoun:** Pretty good so far. + +**Mat Ryer:** Yeah, it's Tuesday, isn't it, so I shouldn't really ask that question. Silly. Never mind... I hope the rest of the week is as good as yesterday. We're also joined by Ed Welch, who is a swell fella who's been two kinds of engineer and two kinds of manager, and will do any job, as long as he's having fun. Hello, Ed. Welcome to Go Time. + +**Ed Welch:** Hey, Mat. Thanks for having me. + +**Mat Ryer:** Absolute pleasure. So yeah, logging. We all, I think, know what logging is, but let's just be clear - logging... What is it? + +**Ed Welch:** The fun part about logging is it's probably the first thing that everybody does when they start writing software. Everybody's first intro into running a program is usually a Hello World, which is at least in my opinion a form of logging. So it's probably the most common way, the most ubiquitous way that we get state out of our application. Some kind of view into what it's doing, some kind of idea of what's happening. So at its most basic, it's just some indicator of what our application is even up to. + +**Mat Ryer:** Yeah. Sometimes I'll have -- in the beginning, when I'm writing something, I'll have lots and lots of logging going on, because like you say, it's a great way to see what's going on and get some insights. It's like a really easy, simple way to do that. But usually, I'll go and kind of clear that all out and remove it all, because it stops being useful at some point. Is that right? Should I do that? Or should I leave it in there? + +**Ed Welch:** \[04:14\] You're probably ahead of most. I think that largely people when they write a log statement, it's probably there forever. I'm not sure how often people really go back to scrutinize what they logged. So I would certainly think that you're gonna gain from that; having more valuable log messages and more relevant is useful, but I think that in terms of that quantity, as long as they contain useful context or information, you shouldn't really be looking to \[unintelligible 00:04:41.26\] + +**Jon Calhoun:** I'm imagining some of your scenarios are writing logs to help debug stuff... And if you're anything like me, there's probably some time where you catch yourself doing a sanity check of like "We're in this function", just because you're like "Is the code actually getting to this point where it's supposed to get?" And I can imagine that those ones, over time, stop adding value. Once you have tests and everything, you know it's all working, it's kind of like "Is that running every single time that function runs really adding value?" Whereas other parts of it definitely could be. + +**Mat Ryer:** Yeah, I tend to do that if I'm marking bits. I'll just put monkey, or like monkey2, monkey3. So that's probably why I go through and remove the log statements. But also, if I'm doing test-driven development, I will tend to log a bit less, really, unless there's a specific kind of tricky behavior, something that I don't understand. So I really do use it to sort of observe what's going on. + +**Ed Welch:** I inherited a codebase a few years back, where -- + +**Mat Ryer:** Congratulations. From an old uncle? + +**Ed Welch:** Yeah, it was really -- \[laughs\] + +**Mat Ryer:** An old uncle died? "I'll leave this codebase to Ed..." + +**Ed Welch:** Yeah, let's go with that. So my uncle left me this sweet codebase, and he was very -- I don't know, thorough and consistent... Every function call had a log line at the beginning that would say the name of the function, and the entry, and then the values that were passed in, and then an exit. And to be honest, I've never done that; I've never gone and built software that has that level of verbosity, even if they were at debug level. There were times it was extremely useful, especially as an inheritance, and his untimely passing, and that I wasn't able to ask him questions about those log lines... So it did make understanding the application a little bit better. + +So I think my learning from that had more to do with "Maybe don't log monkey1 and monkey2", instead log something like "Here's where I am and here's the state at the time. Here's some values." Because that particularly ended up being useful, trying to understand why something was or wasn't working. It's like, "It got to here, and things look correct." + +There's a lot of tools that exist now that maybe make that easier, like debuggers and tracing, and even in your tests... But there's certainly an element of logs are generally always available, and so having some (at least) checkpoints through a process that give you insight into that process. + +**Mat Ryer:** Yeah, one trick that reminds me of, that does work for me - if I've got a program where there's quite a bit of logging going on, I'll do a thing where I log out a load of hyphens, just some dashes, and then in Go defer immediately the same thing... And that's a neat way of wrapping up really the in and out of a function, just so you can then see -- you know that the logs between those lines are relevant just to that function. + +That works if you've just got one instance of the program running, which you normally do in dev... But it doesn't work if you've got multiple instances running, really, does it? + +**Ed Welch:** \[07:45\] I think it becomes -- one of the things that I'm gonna give people advice on logging is that the more context you can put in the log line, the more useful it is. Whether or not you have access to the hostname, and if your logging application or frameworks or systems introduce that at another level, but a log line without context isn't really very useful. It doesn't give you -- I would go a little bit farther than that; not only you wanna know the machine and the distributed system that it's executing on, but you need to know the order number, or the user, or the trace ID, or something that lets you trace that through \[unintelligible 00:08:19.23\] of the application. Typically, it's hard to coordinate events over big systems, especially if it's different, disparate database systems and things that all log differently with different formats... So put a lot of context into your log lines, or remove them if they don't have any, because the usefulness of them -- especially, usually, what happens is you go search for that contextual information; you go search for the trace ID, or the order ID, or the user ID, and the results you get back are only gonna be shown to have that info in it. You're likely never gonna see the log lines that don't have that in them. + +**Mat Ryer:** Then I wonder, should you standardize the format then, if you're gonna be using these logs in this way? Because obviously, it's just a string, isn't it? You can just print out anything you like from your program. Should we be strict about what format we're printing in? + +**Ed Welch:** Probably... It's kind of a hard problem, whether or not you have control over the logs from some of your applications or not... If you're building your own app, you're writing your own logs, having some consistency is always helpful. + +One of the things that I think are interesting about log lines is the battle between the humans and the machines. We're talking about humans looking at log lines, but it's not uncommon now that log lines go into analytics, and other systems, security... So having the format be structured becomes really important for how easy it is for a machine to parse those log lines. + +Arguably, having a visual format that's easy to parse for humans is true, too. Commonly, I would say JSON is maybe what you would find the most... It's probably not my first pick for a structured format, because I think it's harder for humans to read; it's very easy for machines to read. And specifically, JSON becomes very hard to read if you have complicated, nested objects, or you build large JSON docs. + +So one of the things when you're typically viewing a log line is it will exist on horizontal space on your screen, and then vertical space as your number of log. So if you try to pretty-print a JSON object in order to be able to view it, you then turn one log line into line tens, or hundreds, so now you've optimized very poorly for being able to look at those logs quickly. + +There's an alternative that I like a lot, that's common in Go probably more than I've seen in other langauges, logfmt. So having key-value pairs that are separated by a space, basically, log format, if you're not Mat? + +**Mat Ryer:** Yeah. + +**Ed Welch:** I know how much you like fmt... + +**Mat Ryer:** I did like it. I wasn't gonna comment, but yeah. I'm pleased you said it. Fmt. Logfmt. + +**Ed Welch:** So it's kind of an interesting compromise on machine parsability and human readability. But ultimately, having structure does become really important, because almost guaranteed you're gonna need some tool to help you parse those logs, or strictly, you're gonna parse them with a machine for other purposes. + +**Mat Ryer:** Yeah. See, JSON would also probably encourage you to put more complex objects in, whereas with logfmt - we're talking key-value pairs there. + +**Ed Welch:** Yeah. Which is a nice thing too, so I generally tell people -- I don't have a problem with JSON logging; it's in fact the easiest and most approachable for most logging frameworks. I would highly recommend keeping a flat structure as a practice though, right? Just do key-value pairs in JSON. + +**Jon Calhoun:** I think part of it definitely comes down to how you're logging affects how you're gonna consume it... Because I know the first time I ever used JSON logging, it was game-changing in the sense of like "Oh, I can filter on these things, and it'll make my life a lot easier." But you're right, the first time you see it in just like plain text, you're like "Oh gosh, I can't read this anymore." But depending on the tools, you have certain tools that will help alleviate that pain, so you can start filtering things and all of a sudden it will make things pretty for you... And if you have the right tooling to do that - there are services out there that do it, and different options - then it's kind of okay... But I get what you're saying with the key-value pair; it's kind of like a nice middle ground, where you can still read it, but it's still gonna have that machine readability. + +**Ed Welch:** \[12:17\] Yeah. The other reason to keep the object flat -- because you're right, the tooling that exists out there usually facilitates this, but it also introduces in a lot of cases another query language. So to manipulate JSON documents, you usually need to use -- I like JQ, JMESPath, or some other query language that you have to figure out how to access elements of the JSON object to return them... So there would be, I'm sure, a reasonably good debate about what makes a good log line, versus what's information that should go in a database. If you're generating huge documents with hundreds of lines and values... And I see this in even some pretty famous places; a lot of the logs that we get out of Google are massive JSON documents, with massive nested elements in them... And I find that they're hard to work with, both as humans, and for parsing for machines. + +I think the more you can keep that structure flat, I think the happier you would generally be... And if you need a complex nested object, is the log line the right place for storing that information? Maybe a sniff test... We won't go as far as sanity test yet. + +**Mat Ryer:** No. + +**Jon Calhoun:** Thinking about it that way - you said some things are better fit for a database... Do you consider things written in a database to potentially be logs? An example I can give is I was building a web server once, and I forget why, but for some reason we wanted a way to actually record web requests, and the response to them, to make sure things were going correctly and to -- there was some other reason; I don't remember why off the top of my head. But I know that one of the things we did is we were like "Well, you can't really take an entire web request and just throw it into the output. It's gonna be really hard to do anything with it." So we took the parts we cared about, and we were storing that in an actual database table, and we would essentially clean that up every so often; luckily, it wasn't a huge project at the time, so we could get away with it. I imagine at Google scale that might get a little bit tricky... But for us, that worked. And I consider that logging, but with the caveat that we're storing this in a database and it's kind of a little bit different. + +**Mat Ryer:** Yeah, but that's kind of structured logging, isn't it? + +**Jon Calhoun:** It had that feel to me of like - you entered a function, and then when you exit the function, you're kind of like "These are the things that happened." And that's roughly what we were doing; we were like "We need to store this somewhere where we can query it and actually look at the data and try to do stuff with it." + +**Ed Welch:** It's really fascinating in the sense that it touches a little bit on the types of logging that we see, too. That falls a little bit into the category that are more like event logs, or access logs, or they're very specific things that happen, that have useful contexts. I guess the question there that I find interesting is "Is there a right amount, or too much data?" From an implementation standpoint, the tools that you have available and the tools that you're using maybe dictate what's a better fit there... + +The other types of logging that we kind of started talking about are more like what people introduce when they're writing code. So this is the stuff that helps me understand if my application is working, or where it's working, or how it's working... Maybe even just the sort of overall lifecycle of the program itself in logs... But you think about access logs from reverse proxy, or you're describing orders and events where you have complex information coming in... Yeah, I think that's a really good question about where you would draw the line on, and what systems it makes sense for. + +The only advice that I could give around this, in some experiences that I've had, is don't make your primary logging in terms of what gives you visibility into your application stored in the same thing that you need to run -- so like don't store those logs in the database, because what happens is when you have trouble with your database, you can't view your logs and you can't see what the trouble is. + +**Mat Ryer:** \[15:59\] Or if your database -- if there's an error, then it logs that there's an error, but it can't log to the database, so that's an error... \[laughs\] + +**Ed Welch:** Yeah. It's very useful. I've seen in older software systems that I've worked with before that systems that exist today for aggregating working with logs - it was pretty common to store logs in the database, but then when you had trouble with the database, which was usually the thing you had the most trouble with, you couldn't tell. You couldn't see what was going on. + +**Jon Calhoun:** Yeah. A lot of that probably comes down to like the lifecycle of the product, too... Where I tend to see that type of approach being - not okay, but kind of okay, is if you have like a two-person startup and they're just trying to slap something together and get it up and going, then you kind of cut some corners here and there. You can't have six different services up and running, because it's just hard to manage that. But as something gets much bigger and grows, then it's like "Okay, now it's time to actually look at like we need a dedicated place to put these logs and actually consume them." + +**Mat Ryer:** Yeah, on that question of consuming the logs then, Erik from the Gophers Slack - and by the way, everybody, don't forget, we're on Slack. You can join the chat live, GoTimeFM channel. Erik has done just that. Erik asks "Will the key-value pairs always be in the same order with logfmt?" Because I imagine if you visualize and you're looking at the logs, it would be quite useful if they were lined up. + +**Ed Welch:** It's gonna be subject to whatever library implements that. In most of the applications that I work with, we use GoKit, and they're always consistently in order of the way that they're written. Like I said, the libraries that implement logfmt - JSON is more common, but I think there's support in others as well. I can't speak to their consistency and ordering, but I think that would be a very valuable -- like, humans are really, really good at pattern recognition. So if you can get data in front of us in a way, even in log lines -- I can tell a funny story here... I had a job where I worked with computers that would run vehicles, and we would have trouble where sometimes they would boot and behave differently... And I found myself -- we were shipping the syslogs from those machines back to the central place, and I would review the syslogs to try to find instances of the computers that misbehaved... And the fastest way that I could do that was I could zoom way, way out, so that the text wasn't legible, but I could look at the length of the log lines and you could see the pattern, and you could scroll through the log lines very consistently, because it was really easy to spot the ones that were all the same. And then you can add a little bit of color to that too, and it made it even easier. But it became very easy to see the ones that didn't act properly. + +So having consistency in your log output is a huge value for a human to parse. Machines tend to not care. Maybe there's some efficiency gains if they're consistent, but it would be a good feature to have, that's consistently orienting your -- and I think probably most JSON serializers are gonna be consistent. I think most serializers in general are probably gonna be consistent. It's just a question of whether they guarantee that over dates, and things. + +**Mat Ryer:** I imagine if you're passing in key-value pairs, it probably logs out in the order that you do it, and therefore it's up to you to make sure you're consistent in your code to get it right. And that thing of context is interesting... I have a project where we actually use the context to carry -- like, Go's context.context to carry some... Context. You guessed it. And that gets passed into the logger; when you call a log line, the first argument is a context. How do you feel about that, Ed? Because it's very useful for -- you talk about some of the things of like user ID maybe, and the hostname, and things like that. It's the kind of thing that you could do in some middleware somewhere, or do it in one place, and then it is consistent. + +**Ed Welch:** \[19:47\] Yeah, definitely. The applications that I work with do this pretty commonly. Trace ID is often propagated through the context; or is always propagated through the context. Something like a tenant ID in a multi-tenant system... And there's helper functions that we have that will pull that out to -- I don't remember if we implemented the interface for GoKit to just sort of make a logger that automatically did that, or if you just have the functions that pull them out, but it's really useful. + +I wish Go went one step farther here... One of my griefs with Go is the context deadline exceeded, and context canceled errors, because in a distributed system you get a context deadline exceeded and I don't know who canceled that context, or what deadline rather exceeded or timed out... It could have been four systems away. So it would be up to you to introduce context into the context (sorry), that when you're the one that times out, you'd have to say "Hey, it's --" It doesn't make it easy. It would be nice if the functions that you called with timeout, or with whatever, could take a string that would just print in whatever error block ultimately catches that context being canceled, so that you could find it easier. + +**Mat Ryer:** Yeah, that's a nice one. You could write a package that did that, couldn't you? + +**Ed Welch:** I thought about this last night... \[laughs\] I'm not sure what it would look like to -- I mean, you could do definitely something that catches it; where it's tricky is how easy and how organic would it be to use... Because a lot of times you're catching an error on a function, and you have no idea what the error was. The error might have been that the context was canceled while you were executing that function, and you're just logging the error. You're basically now thinking about wrapping all of your errors with this thing that would look for someone that's better at this than I am. + +As much as I love how well that context cancelation allows you to control over network connections and things, I find that it leaves a lot of times confusion around what piece of the puzzle was the one that actually said \[unintelligible 00:21:57.15\] + +**Mat Ryer:** Yeah, because in an HTTP world if you are getting a request and the browser -- like, someone just closes the browser, that'll cancel it. That'll cancel the request. And in your own code, you might be canceling context and relying on that, and then you can't tell the difference. + +**Ed Welch:** Yeah. My favorite is -- so like NGINX, or anybody that sits in the middle, that times out, will result in a context canceled error. I mean, in my argument here that would be not the most helpful, but somebody should know that it was their context that was canceled, and say "Hey, it's me over here. I don't know who did it, but somebody closed this connection." + +**Mat Ryer:** I wonder if you could pass a string into the cancel function... Because you get back -- when you do with the timeout, or the other one, with cancel, you get back the little cancel callback function thing. Could you pass in a string there, I wonder...? Actually, that's an interesting problem. It'd be quite fun to explore that, but I don't think we're going to. + +**Ed Welch:** I think the idea of how well and how useful the context is, but it does often lead -- I know when I first started writing Go, I found that I was doing a thing that was very Java-esque, because I really missed checked exceptions and I was trying to understand why an error... So I'm looking at the type of an error, and looking to see that it was deadline exceeded, and then I logged a message that said "Timed out", because everywhere else in the rest of the world it's timed out, and in Go it's deadline exceeded... The error handling in Go makes this a bit bumpy; I wish it was a little bit easier to -- we're way out into the weeds; I'm just gonna shut up right now... \[laughs\] Before I start passing opinions. I've gotta save my unpopular opinions for the end. + +**Mat Ryer:** Yeah, you've gotta save it for the theme tune. They don't count unless it's been played. + +**Break:** \[23:43\] + +**Mat Ryer:** I usually end up having dedicated code for context errors at the top of -- wherever they unwind to, I'll have some code, because I probably want to do something different. And the other thing is - it's very normal to have context that gets canceled. It's a sentinel error, term coined by our friend and friend of the show, Dave Cheney. It's a sentinel error value type that you can compare against and you can pass around. But it's not really an error. Not necessarily. And if someone closes the browser and cancels a request, you want to stop doing the work; that's not an error, that's great. It should return "That's great", or something instead, that's great. + +**Ed Welch:** I would support that. + +**Mat Ryer:** Yeah. I'll do a PR. + +**Ed Welch:** It's something to think about for folks that are writing applications and they're using contexts - and they should be; there's the context -- you can add values to it, so take advantage of that. Like, most things probably don't get carried away, keep the context simple, but you could consider... I don't know off my head, like I said, how ergonomic you could make this, but trying to include information about parent contexts and ultimately who cancels it, be able to log something to make your life a little bit easier, to track down where in a complex system something took too long. + +**Mat Ryer:** Yeah. Another interesting little thing that's quite nice to do in Go that I did once was when adding final use to the log, using those helpers that you talk about, that these packages have - that would also return a little cancel function, which you defer immediately, remove that value, essentially... Which depending on where you're passing things around and how you pass things too, you may not need that. But it's quite nice to be able to use Go's actual language features when it comes to doing this and solving these problems. Logging then in a way becomes really a first-class concern of your application, doesn't it? + +**Ed Welch:** Yeah, I'm a little bit biased when it comes to the way we observe applications. I spend all of my time working on a system that's designed around handling logs, so... + +**Mat Ryer:** That's why you're here though, Ed. + +**Ed Welch:** \[27:49\] It's true, yeah. The nice thing about logging is it's kind of always there. It's the one most accessible way that we have to get info out of an application. The other forms that we have of observability tend to lend themselves to better use cases or different options... Distributed tracing can do a lot of stuff that's hard to do with a logger. I could make an argument, but I would lose that argument... But it's there, it's always there to use. So to use it, just make sure you have useful information in your log lines. Do yourself a favor and make sure you can search for that log line when it's printed. So it has to have something in there you'd be able to search for. + +And there's another one, I think, when it comes to error messages. A lot of times, depending on how you structure your logging and whether your logger is gonna include the line number something comes from - I don't have strong opinions if you need to do that or not. Honestly, I would say just make sure you write error messages that are unique within your app. If you search for that error message, it should go to one place. + +**Mat Ryer:** Yeah. Like monkey1, monkey2. + +**Ed Welch:** Exactly, yeah. So maybe you were really onto something there... \[laughs\] + +**Mat Ryer:** Well, that's why I use different numbers. + +**Ed Welch:** Yeah. How do you keep track of how many monkeys there are though? + +**Mat Ryer:** Yeah, that is difficult. To be fair, it gets way out of hand very quickly. + +**Ed Welch:** Do you have like a global monkey tracker that you...? + +**Mat Ryer:** You do need to implement that monkey tracking, yeah. And I did once mess up a loop and ended up with infinite monkeys... And they produced no works of any kind of discernable literature whatsoever, so - very disappointed there. + +**Ed Welch:** Yeah. One way with error messages that you can accomplish this uniqueness is the idea that you could consider error messages part of the runbooks for operating your system. So it's one thing to say something timed out, it's another to say that it timed out but that it's gonna retry; it's gonna retry ten times more, or it's not gonna retry, and somebody's gotta go fix it. + +It's funny, when you're writing applications and you get to where you're logging an error message, you likely have the most context of anybody that's gonna see that error message in years to come. So what can you put in there that helps someone make a decision about what they should do when they're operating that software. So they got paged, and they searched the logs, and they found your error messages; how can you help them get out of that trouble? Or let them know what the system is gonna do to recover from it itself... Because that is fairly common - there's lots of error messages about connections timing out, or failing, and is there a way to communicate that the app can still handle it, go and fix the source of the problem. + +**Mat Ryer:** I kind of love this about the UX of logs, is what we're really doing... And honestly, anytime I see a sophisticated approach in any field, they reach that sophistication by caring about the audience or the user of that thing. And it works for web UIs and APIs that you write in Go, and packages, programs... It works for everything. Writing... So I like the idea; I think that's quite an interesting change, and I'm not sure everyone -- because normally, I think people are quite selfish loggers. They're logging stuff they need. But I like that; think about your audience and log for them. Log for your great nephew. + +**Ed Welch:** One other place that's been useful is propagated all the way back through your API. We have limits in our application that you can hit for various things. So we will log something like "You've exceeded this limit. Contact your administrator, or reach out to whoever administrates that." Or try again, or reduce your requests, or fix your error... Not uncommon in 400 type errors and user errors to give some insight into what people should do, but the more you can describe how to get out of the error situation in the error message, the more self-serving and self-documenting and probably less trouble tickets you're gonna have. + +**Mat Ryer:** Hm. Would you link to like a runbook, or link to a doc even? + +**Ed Welch:** \[31:51\] I don't know, because my experience with that is that whatever you link to ultimately will break, and no one will ever go back and repair that link. It's very hard to maintain... If you can link it back to the source code in a pretty stable way... And can you have a linter or something that finds broken links? I don't know, because it's most likely just gonna be broken. + +**Mat Ryer:** Good point. Yeah, keep itself contained; it's probably good advice. I once wrote a package for a co-worker. He was gonna then implement it. So we kind of broke the work up like that. And I customized all the errors, all the error messages just for him. So I was saying "Come on, mate... This is obviously. Obviously, you're not supposed to pass that value in." I was trolling him in the error messages. That was fun, but not for him. + +**Ed Welch:** \[laughs\] I've logged messages before that are along the lines of like "This shouldn't happen..." I wrote those log lines, which is -- my favorite was in Java, an exception handling of - you would catch an exception and then catch an exception within the exception, and in a lot of those cases I don't even know, but it's important to do this, because somebody somewhere said I should do this, so... Even that is useful though, right? Anything that explains your mindset around why you wrote the error message or what you should do can be useful to the person that's troubleshooting that problem. So maybe have a little fun with it. + +Another consideration here would be - what would maybe make this easier if you pull your error strings into a central spot. If you're doing internationalization of your application, this can be helpful. There was some tooling that made that a little bit easier to swap filenames and you could just have your error messages be in multiple languages, and make it easy to swap them out at compile time or runtime for the configuration flag. But putting them out inside where you can see them both lets you kind of easily review them to see if they're providing useful info, as well as -- and I'll say this is good advice that I don't use myself. + +**Mat Ryer:** \[laughs\] Yeah. Well, on context then - maybe we could talk about log levels, because this is something else that is quite divisive subject I've seen around. This is where you can -- each line has a level, so it's usually debug, info, warning, error, critical... There's loads of options. I'm sure you've come across even more. But I find error and debug information to be helpful, warnings sometimes... But how do you feel about log levels? + +**Ed Welch:** So I have some opinions here... + +**Mat Ryer:** Are they popular or unpopular? + +**Ed Welch:** These are probably popular, but... + +**Mat Ryer:** Okay, I'll leave the music. + +**Ed Welch:** Not everyone will agree, but errors is the easiest one for me. So I can start at error and warn, and then it gets kind of murky from there. But the error levels - I like to log at an error level if there's something wrong with my application. It's my job, it's my responsibility to fix it. + +**Mat Ryer:** It makes sense. + +**Ed Welch:** I like to use warning level when it's not my responsibility or my job to fix it. So I'm getting bad data... Some other downstream system is misbehaving, something that is useful for me to go find, but it's gonna be more useful probably for somebody else than me. That's my opinion on that. + +Now, when you go beyond that -- so debug is really fascinating, because it ends up just being sort of the dumpster of log levels; everything ends up in there, and it usually ends up being so verbose that people don't run with it. And if you're not running with it - and most applications in my experiences don't have an easy way to do basically runtime changing of the log level output... It's like, what good is it, right? + +So they can see the argument to seeing that you just log everything at sort of info, and that would force you to go back through and remove log lines that aren't as useful... But you do have these cases where -- I don't know, you write these log lines and instrument something in a way that's really interesting for specific cases, and you think "I could see myself hitting this again. What do I do with this?" I would say that my experience with Go loggers is they lack some of the functionality I've seen in others where you can do like for package level, different levels, or runtime reloading... And maybe those packages are out there and I've just not used them, but those can be really nice to handle that. Or trace level is maybe what you would do there, or something that's really specific, but only would want to enable in a very small case if you have that control available. + +**Jon Calhoun:** \[36:20\] One of the people in Slack had mentioned - I think it was Gabriel - that he actually uses separate packages for developer type logging versus production style logging, just to keep the two separate... And I think one of the main benefits there is that when you go to production, you can basically set everything that's a development style logging to just be a no-op, essentially, and the code's there... Because I think that's one of the things -- like you said, we've all written Hello World, but there's kind of like two mentalities with logging. There's the logs that are just for me right now while I'm writing this code or debugging some issue or whatever I'm doing, and then there's the logs that have a lot more context and need to have trace IDs, and things like that. Because if you're running locally, chances are you probably don't need trace IDs. Generally speaking, you don't have enough requests coming in locally if that type of stuff is an issue. So it is interesting to think of it as like two separate approaches to something similar. + +**Mat Ryer:** Yeah. I've seen it where you'll have like a logger type that supports nil, because in Go, of course, you can write methods that are safe to call on nil; you literally just check the receiver, and if it's nil, it just usually is returned at that point. That's quite nice... But I've also seen it where you explicitly have nil or no-op loggers, or whatever. Does this all just depend on which package you're using, or are there good practices there? + +**Ed Welch:** I don't know that I have a good practices suggestion for that. My personal opinion would be to lean towards just the always log this stuff. Basically, turn your debug logging level on and -- actually, I've gotta decide if that's really my opinion... Because it is nuanced, right? There are situations where the volume can be tremendous, and you mostly don't care. But you do care. So that's where I think -- in my opinion, the state of the art of logging needs to be better at this, which is either your application or your logging framework or something that gives you the ability to control at runtime what ends up being stored. + +I know more times than not you don't have debug logging on. There's some information that you want, but you can't get it because the application is running, and likely restarting it resets the error condition, trying to troubleshoot. So, it would be better if you could leave debug logging on, and then just drop it at the source until you need it... Or have runtime log-level changing as an option. It gets a bit more sophisticated. I don't know that I see this a ton; at least in Go I'm not seeing it a lot where you have that runtime capability. Somebody will probably point me -- which would be great, because there's an element of Go and logging that I've always found a little bit frustrating, which is the lack of a standardized interface and the fact that your logger becomes very important to your system, and therefore very hard to change... So we've talked about changing our logger in our app, but it's an effort, it's an opportunity cost; we'd have to be really sure that we need to do it. It works fine, but there are loggers out there now that have a much lower allocation rate, which can really matter a lot too if you're gonna log a lot. If you're gonna log thousands of line a second, you're paying for that in allocations in CPU. + +**Mat Ryer:** It's not free, is it? + +**Ed Welch:** More interestingly, you're potentially bottlenecking your application, because almost every infrastructure that I see still logs to disk. Usually, like say in Kubernetes, it's the node disk, and node disks aren't terribly fast. So you're more likely than not at some point blocked on your logger, trying to write to a file on disk, which can slow your application down. + +**Break:** \[40:00\] + +**Mat Ryer:** Bill Kennedy, the famous hatted gopher actually showed me an example once where there was a bug in the program and by putting a log line in, because there was a cost to doing that, it changed the behavior of the program. So that act of observing it almost affected it, which is kind of amazing. + +**Ed Welch:** Right. There's some Heisenberg in here, or something... \[laughter\] + +**Mat Ryer:** Yeah... It feels like that, doesn't it? Imagine if you didn't realize that though, and you're logging out, and then it's telling you something that is different to what you see when you then take the log line out. How infuriating is that...? + +**Jon Calhoun:** I think that's the risk of using any language where you have concurrency and other things like that going on, is that it's just -- you always have to take for account that if you're using concurrency, there could be some case where adding anything in there could change the behavior. + +**Mat Ryer:** While I've got you, Jon - it's quiz time! We'll put some music in. I've just made up this segment. + +**Jon Calhoun:** Oh, boy... + +**Jingle:** \[42:40\] to \[42:58\] + +**Mat Ryer:** Jon. Jon Calhoun is joining me from - where are you from, Jon? + +**Jon Calhoun:** Pennsylvania. + +**Mat Ryer:** Pennsylvania. + +**Jon Calhoun:** Can't get more specific than that with you. I don't know what you'll do with that information. + +**Mat Ryer:** Turn up at your house, I might... + +**Jon Calhoun:** It's possible. + +**Mat Ryer:** Yeah, I do want to know your full address now... + +**Jon Calhoun:** Well... + +**Mat Ryer:** No? + +**Jon Calhoun:** Nah. + +**Mat Ryer:** Okay. + +**Ed Welch:** Was that the quiz? + +**Mat Ryer:** Yeah. What's your mother's maiden name, Jon? No, the real quiz is - fmt.println and log.println, where does each of those get printed to? + +**Jon Calhoun:** Where does each get printed to? + +**Mat Ryer:** Yeah. + +**Jon Calhoun:** I assume fmt is os, so that's STDOUT... Or sorry, the fmt one. The log.println - I don't know if it's actually differentiated or not. I would assume it's just os.STDOUT. That would be my guess. + +**Mat Ryer:** Okay. Let's find out. \[sound effect 00:43:49.18\] That sound effect tells us that - no, Jon, unfortunately you're wrong. Log goes to STDERR by default. So that's interesting... + +**Jon Calhoun:** See, I wasn't sure if this was a trick question or not. That's the hard part. With Mat you can never tell. + +**Mat Ryer:** \[laughs\] Yeah, but that's been an odd quiz... + +**Jon Calhoun:** The worst part is in most terminals you can never tell the difference. + +**Mat Ryer:** Yeah, that's true. + +**Jon Calhoun:** Unless you're doing something specific, like piping it to an output file, or something. + +**Mat Ryer:** \[44:13\] This is the question then - where should the logs go to? Do we want them in STDOUT? Should we put only the error level logs in STDERR, and the other logs in STDOUT? What do we think? + +**Jon Calhoun:** I think people have a lot of weird opinions on logging in general. I swear I've talked to somebody who said "If you're printing out a log statement instead of recording a metric, that it should be an error that some engineer has to go fix." + +**Mat Ryer:** Right. + +**Jon Calhoun:** To the point where their app apparently has no logs, it's just all metrics. And I think sometimes how people think about logs vs. metrics - they might kind of be the same thing, they're just differentiating them differently, if that makes sense... + +**Ed Welch:** I could give an opinion that I think is a little bit researched on that... + +**Mat Ryer:** Ed, before you do, we should just be honest with our listeners here... You kind of built Loki, which is Grafana's logging thing. Tell us about Loki just very briefly. Because there's some important context, I think, there. + +**Ed Welch:** I've certainly been a part of Loki. Lots of other smarter folks than me have done the lion's share of the building. + +**Mat Ryer:** It's not what I've heard... I think you're just being modest. + +**Ed Welch:** Actually, it's been about three years that I've been working on Loki now, so I'm kind of trying to think of how I've evolved describing it over time... This may dovetail a little bit into what Jon was saying, but... Loki I tend to describe now as the time series database for string. So I guess before I talk about Loki, the part about the metrics versus logs that I think is interesting is the information that they can contain and store - there's an associated cost with that. + +One of the things when I talk about Loki as the time series database for strings is... When we talked about that context you can put in a log line - you can put a lot of information in the log line, right? Within Loki we log for every query, the results of that query and things like throughput and the lines process per second. And the query, and the org ID, or the tenant ID, and all this information... And it tends to be extremely high cardinality. So where metrics systems - it gets a bit more costly when you're trying to store huge amounts of cardinality. If I had a metric for every one of those things I've just said that had a tenant dimension on it, it would be a lot. + +The other side of that coin though is that's a long string, and to process that in a system requires parsing it, and doing math on it that's more expensive, it's more data stored on a disk. So to me, the two play very nicely together in the sense of metrics make a really nice roundoup or sort of overall approximation of high volumes of that kind of data. So you can look at metrics, because it's a float, over much longer periods of time than it's gonna take to parse a 80-byte or 200-byte string. + +Where those become a bit interesting is -- so with Loki we have a thing for recording rules where I can generate a metric from those log lines as they come in. So what we'll tend to do is generate roll-ups. So we'll generate a 99% quintile on the query times by tenant. So now maybe I have cardinality of thousands of tenants, but the 99th percentile is taking all of those individual values and rolling them up into one metric. + +So the systems (I think) complement each other really well when you set them up. You can have very high resolution, very high cardinality data, it's just gonna be a lot more expensive to try to query it over time and put those in your logs, and then you can roll that stuff up. Or create metrics, just use the metrics you have in your system for higher volume requests, being able to keep an eye on a system that's doing thousands of requests per second a lot easier, showing that data over time a lot easier... So I think they play nicely with each other. I think maybe both extremes -- I guess I could argue that you could monitor an app only with... Well, you can make that argument either direction. + +\[47:58\] To be honest, once your apps start getting big enough, you probably want specialized tooling for those types of aspects, metrics, logs... And traces, for big distributed systems, being able to view how your information propagates through - a lot easier in a tracing system than it tends to be in searching for a trace ID amongst all your logs, and looking at how each of those logs were formatted differently, and the timestamps between them, and trying to kind of reconstruct that same view. + +**Mat Ryer:** Yeah, we had an interesting question in the Slack channel... Erik is back, and Erik says "What about logs and testing? Do you ever include assertions about should be logged out?" Have you ever done anything like that? + +**Jon Calhoun:** Jumping back to your previous question, can we say that when you're logging and testing them, maybe you should write `os.standard` there. I'm just kidding, but... Just split it up that way. We never did answer that question, if you wanna go back to it later. + +**Mat Ryer:** Oh, yeah. + +**Jon Calhoun:** But Ed, logging and testing. + +**Ed Welch:** I don't know, I mean... It happens from time to time. It always feels a little weird when you're validating a log line, I guess. I'm trying to think of an application where I've seen that done. Usually, it's in catching like an error message or something, like a sort of specific error you would catch, which becomes a log line... + +**Mat Ryer:** Yeah, I've done it once where it mattered what was printed out from the program; it was very important. And so, this is -- Because I always do this little abstraction where I have a run method, and main only just calls out to a run method, and passes the dependencies in. And I include STDOUT and STDERR if I wanna use them; I include even those as io writers. So in test code then you can use buffers, or whatever else you want, to capture it. And that is a nice way to -- then you can make assertions about what is printed out. I don't know if it was logging out; it mattered what was printed out in that case. But yeah, that was just STDOUT. So we never did \[unintelligible 00:49:55.25\] + +**Ed Welch:** It could be interesting if you have downstream processing under your log lines. It is interesting, because you build dependencies, and now you're logging - you're talking about it being kind of an API of your application, and... Maybe API is not the right term, but it's an interface that... + +**Mat Ryer:** It kind of is. + +**Ed Welch:** ...goes out to. So you could do testing on it to make sure you're not making breaking changes to the downstream systems. It's interesting if you find yourself in that world where you have those tight couplings. But if you're using JSON or even logfmt, you could have tests to make sure that certain values are there... I'm not sure quite how to dig into that one. + +**Jon Calhoun:** It seems like it comes down to the point where if not having it there and it changing could break something, then you probably need to test it. + +**Mat Ryer:** Yeah, if it's part of your API... + +**Jon Calhoun:** But if it's not something that's gonna be broken and it's just gonna be like a visual difference, then you're probably okay getting by without it. + +**Mat Ryer:** Yeah, by the way - Jon, you won't believe this, but the little elves in the Slack channel have found the episode that Scott Mansfield talked about Go at Netflix on Go Time episode \#9. + +**Jon Calhoun:** Was it \#9? + +**Mat Ryer:** So if anyone wants to go back... + +**Jon Calhoun:** I swear there was one that I was in as well that got mentioned... So maybe it's been in multiple. + +**Mat Ryer:** It's possible. + +**Jon Calhoun:** I swear it was one of the episodes we did on metrics and things like that with a company that did metrics software... + +**Mat Ryer:** Yeah. + +**Jon Calhoun:** But I swear they had like a similar opinion, along the lines of like "If there's a log message, somebody should be getting paged right now." So it's just interesting to see the different approaches that people take to writing software. + +**Mat Ryer:** Yeah, and that's what I wonder also, is whether this sort of depends in each case. Anytime you have people with really strong opinions about tech, often they're working on something different. That's why I try to be -- you know, unpopular opinions loosely held. + +**Ed Welch:** \[51:49\] One area that I think could matter in that argument of metrics versus logs is that logs aren't events; they have a distinct timestamp that's very accurate. Metrics do not have that kind of accuracy, at least most of the time. You're sort of forced to do a roll-up, because that's the optimization they make, which is really valuable... So maybe that event timestamp isn't -- I don't know, it can be really important for troubleshooting to know exactly when the sequence of events happened, and a lot of times when you're working within whatever the bound of your metric is, if the event happened inside, say, a Prometheus scrape interval, you don't have any information to tell you what happened, and a lot of times logs can hide or show that context. Without knowing more context -- I'm sure it's a well thought out argument, in that it works well, you can build a system that works well, that doesn't log... But logging is just a sort of -- it's the thing, it's everywhere. + +We were gonna talk a little bit more about Loki - I think the reason Loki exists is because of how ubiquitous logging and some of the problems are with it. We talk about the variety of formats that exist for logs. You know, do you pre-process or post-process those to try to make them useful? How do you catalog and store and index that information? Where do you store it? People generate tremendous amounts of log data. The numbers vary; the big side is usually terabytes a day, although definitely no people say they generate petabytes of logs a day, which is just mind-blowing amount of data to have to deal with and store. + +So when it comes to logging, that's the real question of "Where does it go?" You had that question in there, "Where did the logs go?" \[laughs\] It's a good question, because that's usually a disconnect between the person that writes code and the system and infrastructure that runs it. So it becomes a little bit more important than what you log. + +**Mat Ryer:** I can't imagine petabytes of logs... I mean, knip through and delete some log lines, surely. + +**Jon Calhoun:** I can't imagine how you'd use those at all. I feel like you just end up with so many -- either that, or you're like a crazy busy service, or something... And then it's like, we need to do something different. + +**Mat Ryer:** Or you use a technology that then works in some abstract way on that data. That must be the answer. I'm not gonna keep promoting Loki; other log platforms are available... + +**Ed Welch:** So the way you could approach that that would fit with Loki is -- what Loki does is a little different from other systems is the indexing model. So the idea is to index a relatively small -- we don't index the content of the logs themselves, we index metadata; think about where the logs come from, typically. + +You were gonna ask the question of like where would I go to find the logs... Maybe like a table of contents a little bit... You keep a relatively small index because an index in a database tends to be one of the more expensive operational pieces. So you keep a small index. So for that petabytes, you would chunk that up into smaller and smaller pieces, and then Loki sort of optimizes around that scale side of it first, or the idea that your expensive parts tend to be on ingesting and storing that index, and keeping that cost lower, or that complexity lower. And then on the read side of things, it's a little bit more simplistic; it's just kind of brute force. So if you don't fully index something, you don't know where it is, you've gotta go look through it. + +So you use those labels to narrow down that result set. But one of the things that computing is generally really good at now is parallelism. Loki optimizes around object stores, so it's basically a strings database that's built against object stores, which also support parallelization really well, and take a lot of the nightmare of working with storage and make it somebody else's problem... And you can force your way in parallelism, at many gigabytes a second, to be able to get back to specific events that you're looking for. So it's kind of a trade-off between - you pay more for that query time, but logs are typically something you write once and maybe query never, or query infrequently. So if your operating model is one that you write a lot of logs and you wanna query them all the time, this might not be the best approach. But for the developer and operator use case, it's usually an advantageous trade-off to be able to control the query costs. + +**Mat Ryer:** \[56:02\] Okay. I'm gonna have to return an error to you though, Ed... Too many mentions of Loki. + +**Ed Welch:** Done, and done. + +**Jon Calhoun:** I thought you were gonna tell him deadline exceeded. + +**Mat Ryer:** Yeah. Unfortunately, we're coming up to deadline exceeded is on the minute. But, there's still got time, so Ed, Jon, shut your faces, it's time for Unpopular Opinions. + +**Jingle:** \[56:19\] to \[56:35\] + +**Mat Ryer:** Okay, whooooo's gonna -- it's not Halloween, I don't know why I'm doing the spooky voice... Who's gonna start us off with an unpopular opinion today? I notice you've got six, Ed. + +**Ed Welch:** \[laughs\] I think it's funny, because I laugh every time I think about the goal is to have an opinion that people don't agree with. + +**Jon Calhoun:** So like if we had a KKK member, he would probably win. I hope he would win. + +**Mat Ryer:** Yeah, yeah. + +**Jon Calhoun:** Not in a good winning way. + +**Mat Ryer:** Yeah, but we don't have them on. + +**Jon Calhoun:** No. + +**Ed Welch:** Right. So I'll phrase this -- this is my favorite unpopular opinion. I'll phrase it in a way that's likely to be received as it's most unpopular, but it's a nuanced conversation. My experience is that integration testing is usually a net loss, and I would largely recommend not doing it. + +**Mat Ryer:** Oh, that is an interesting one. + +**Jon Calhoun:** I have such mixed feelings on integration tests... + +**Ed Welch:** I've definitely fixed and found bugs in integration tests. I've definitely caught significant bugs. However, the tremendous amount of anxiety and time that I spend and have spent looking at CI builds running for minutes or tens of minutes or more, combined with just the propensity for integration tests to have false positives, in my experience, leads to... Those couple bugs - it probably would have been better if they'd just made it through to prod and we didn't catch them, versus the overall costs. So that's where I am at within that loss, is that they do tend to catch bugs, but I don't know if it's worth it. + +**Mat Ryer:** Hm. Interesting one. + +**Jon Calhoun:** See, I agree with you fully about the fact that integration tests or tests that cover more scope -- like, integration or larger is how I'd put it in scope, tend to be the ones in applications that I've worked on that just break and then nobody has time to fix them, for some reason, and then you just continuously have breaking tests. And then somebody eventually has to fix it... It just can be a headache. But at the same time, they've been incredibly valuable in some situations, to the point that if I was taking your opinion and like "Okay, we're gonna act on this", it would be more along the lines of "I could maybe agree with not putting integration tests inside of continuous integration", but having them available for local development... A developer has to actively say "Okay, we're gonna pass on this for now. + +**Ed Welch:** Yeah. + +**Jon Calhoun:** But I could see value in not being a hard requirement whereas a lot of times unit tests - it's very clear, this should be passing; if it's not, something's broken or the test needs fixed. + +**Ed Welch:** Yeah. If I gave that as my opinion, it wouldn't be as unpopular... \[laughs\] I really like the idea of running against operational data or clusters, so having tests that are running against systems, but kind of the "I built some database inserts and run them into an in-memory database, and then it verified some kind of data, and that runs as part of a CI build..." My experiences around those is it takes a long time to build them, it's really hard to change the data if you have to go in and make a change to the system... It's hard to understand what the original creator of the tests maybe even had in mind... And when it does flake out, because testing a distributed system and it maybe didn't do retries properly, or at all, or something changed - like you said, the people that usually are the ones that have to deal with it don't have the context to fix it, and it's just holding them up. It's like, "I didn't even change this code and this thing is breaking. I'm really unhappy with my life." + +Maybe the key to success with integration tests is keeping their scope really small and purposeful, and then running them on-demand, kind of thing. + +**Mat Ryer:** \[01:00:03.13\] Yeah, that's very interesting. We'll put this out on Twitter @gotimefm to see if indeed that is unpopular. You wrote down another one though, at which I feel like you might have an even better chance of winning this whole thing. + +**Ed Welch:** The one that you have highlighted, that Windows is the best desktop OS? + +**Mat Ryer:** Say it again clearly for the Twitter clip. + +**Ed Welch:** I believe Windows is the best desktop OS. + +**Mat Ryer:** Okay, really? + +**Ed Welch:** Specifically Windows with WSL 2 is everything that I've ever needed from a desktop OS. A close second would probably be macOS, although my experiences with Docker and macOS are frustrating, and my experiences with updating macOS and having it become less stable after every time it does an update... And in general, the expense of the Mac hardware and hostility of the Mac support community has shied me away from it. + +And I love Linux. I've tried for years and years to run desktop Linux, and I just don't -- I just want my computer to work. I wanna join this podcast today and expect not to have any trouble. And that's Windows, man... That's where I'm at with Windows. + +Also, it's not -- I'm not gonna say that I love it, right? That Windows is great and it doesn't have a whole host of its own problems. I just think it has the least problems. + +**Mat Ryer:** Yeah. It does have Minesweeper. + +**Ed Welch:** Does it? I don't even know... Does Windows 11 still have Minesweeper? + +**Jon Calhoun:** I have no idea. + +**Mat Ryer:** Can we find out? That'll be the clincher for me. + +**Ed Welch:** No, it doesn't. + +**Jon Calhoun:** Every time I've tried to code it all in Windows there's always little things that -- I'm sure I could eventually get past them. Just one example - I can press Ctrl+A to go to the start of a line and Ctrl+E to go at the end of the line. In Windows, Ctrl+A is gonna select everything. So it all of a sudden doesn't do what I want... And I don't know if there is a way to just go to the start of the line; there probably is, I just don't know it. + +**Mat Ryer:** It's probably got its own button, hasn't it? "Go to start of line", or... + +**Ed Welch:** The Home button on my keyboard... \[laughs\] + +**Mat Ryer:** You go in the Start menu, and you go to Run, and then you type "Move to start of line." + +**Ed Welch:** Yeah, the keyboard shortcuts were the worst part for me trying to move to Mac. I could never get used to copy and paste. I could never figure it out. + +**Jon Calhoun:** So I think it's one of those things that -- there's definitely a difference between Cmd and Ctrl, but when you get used to them, it's kind of hard to switch back at times. + +**Mat Ryer:** Yeah, that's true. The muscle memory is -- I got a new keyboard recently, and I've just basically stopped using it, because I don't want it. It's meant to be ergonomic, but I've already learned it the non-ergonomic way. My bones have changed now, and adapting back is tough. Especially as you get older, your brain gets less malleable, doesn't it? + +**Jon Calhoun:** The other issue I have with that type -- because the keyboard you have, Mat, I have it and I was using it, but my issue I ran into was first off, my hand was broken for the last two months, so I was already really slow. + +**Mat Ryer:** That's not the keyboard's fault, Jon... + +**Jon Calhoun:** Well, no, no, but I was already slow, so when the cast came off and I had a brace on I'm like, "I need to get caught up on work, not learn a new keyboard." But then on top of that, I would go from this new keyboard... I think by default where a Caps Lock key normally is was Delete. Which is great, because I never used Caps Lock, and I'd start using it for Delete. And then I went and got on my laptop at some point, just on my laptop, trying to do some work, and all of a sudden I'm just writing in crazy caps all over the place, because I'm hitting the wrong button. And I'm like -- I don't know how people do that mental shift of switching from one to the other. And I think that's probably the biggest issue I'd have with Windows, is if I only used it and I got used to it, I could probably make it work. But I'd have to fully commit and be like "I'm diving into Windows, and I'm okay with that." + +**Ed Welch:** \[01:03:46.26\] Yeah, I feel that. Because I had always used Windows systems in corporate environments, and like I said, the copy-paste was sort of the worst one for me... And then you could go the route of remapping the keys, but that never seems to be flawless... So yeah, it's hard once you're well-established. + +**Jon Calhoun:** If you're going to a Mac, it's definitely not one-to-one... Because Ctrl and Cmd both do different things at different times. So you still use Ctrl a little bit, you use use Cmd more. + +**Mat Ryer:** But isn't it also true in Windows that it depends on the app? It's per application, whereas copy and paste is always the same across the whole operating system on a Mac. + +**Jon Calhoun:** Windows is pretty much always Ctrl+C. + +**Ed Welch:** Unless it's interrupt on the terminal... Which is why I think Mac did it right. It shouldn't be Ctrl+C, or they should have another character for interrupt... But yeah, that's the one that gets you in trouble. When you're working on a terminal in Windows, it's like "Can I use Ctrl+C to copy?" A lot of times it's Ctrl+Shift+C, but other then that, \[unintelligible 01:04:38.06\] + +**Mat Ryer:** I remember it used to be you highlight it, and then it's done. Wasn't it that? You'd just highlight it in the terminal... + +**Ed Welch:** Mine does that. I set up Windows terminal to do that now; it's my favorite thing ever. Every time I highlight something in the terminal it automatically copies it, and I don't know why anyone would not want it to do that. It's the handiest thing ever. + +**Mat Ryer:** Yeah. + +**Ed Welch:** Except when you accidentally click on the screen after you copy. But then you use a tool to manage your clipboard history, so that you could just go back. + +**Mat Ryer:** Yeah. Okay, great. Alright, great. Good stuff. Well, we're slightly over time, but that was worth it. Thank you so much, Ed Welch, for joining us. Fascinating talk and dive into logs there. There's actually so much more we could talk about. We may do logging part two, Return of the... Something. Let's brainstorm that and... + +**Ed Welch:** I'm not sure. I think that was all I had. But I'd be happy to come back though, it was really fun. + +**Mat Ryer:** Okay. It's nice to have you. And Jon Calhoun was also here... Weren't you, Jon? + +**Jon Calhoun:** I was here. Thanks for having me, Mat. + +**Mat Ryer:** Always a pleasure. See you next time, on Go Time. Bye! \[Unpopular Opinions jingle starts 01:05:50.04\] That's the wrong one. It was all so professional until then, wasn't it? \[laughs\] Was it? No, it wasn't. Oh, good. + +**Jon Calhoun:** I think I tune out at the exit song, so I didn't even notice. I was just like "Okay, cool. We're listening to a song." + +**Mat Ryer:** Yeah, I know. But that's the Unpopular Opinion one, Jon. I pressed the wrong one... + +**Jon Calhoun:** Well... + +**Mat Ryer:** That's fine. Because honestly, the editors are just gonna clean this up. Clean it up. We love you, editors, by the way... I could do with editors in real life, like on calls and... You know what I mean? When you're having meetings. It'd be great if you could have someone editing... + +**Ed Welch:** Yeah... + +**Mat Ryer:** Cutting out the gaps... + +**Ed Welch:** I could hear how many times I said "Um..." + +**Mat Ryer:** You won't hear that in the final podcast. + +**Ed Welch:** I know, because I said it usually when someone else started talking, and so like "Great! That's just gonna get cut out, because somebody else just started talking." + +**Mat Ryer:** Yeah. They literally will go through and neaten up any gaps we leave, if we talk over each other... Yeah, it is really good. + +**Jon Calhoun:** They do a good bit with what they can. Obviously, there are some times where it's like "Nope. That's too hard." + +**Mat Ryer:** Yeah, you can't polish a turd, is what you're saying. Normally, we challenge them, you know... + +**Ed Welch:** Shine it up a little bit though... + +**Mat Ryer:** Yeah, that's my unpopular opinion - I actually think... Yeah, you probably can polish a turd, to be honest. + +**Ed Welch:** The Myth Busters actually did. + +**Mat Ryer:** And they've polished a turd on Myth Busters? + +**Ed Welch:** Yeah. + +**Mat Ryer:** Okay. We've said turd too many times for a family-friendly podcast. + +**Jon Calhoun:** We start with the potty humor and end with the potty humor. + +**Mat Ryer:** We did. It's a potty sandwich. + +**Ed Welch:** Is this still on the show? Are we still going? + +**Jon Calhoun:** I mean, it's recording right now... + +**Mat Ryer:** Yeah, yeah, we're still live. I haven't played the final -- this is all legally binding until I play the final music, so... + +**Ed Welch:** Gotcha. + +**Mat Ryer:** Well, we'll put that in the show notes, the Myth Busters thing, if this makes it in... Well, thank you so much again, and thanks for listening. See you next time, on Go Time! diff --git a/Building and using APIs with Go_transcript.txt b/Building and using APIs with Go_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..1a5b0ed384c41a1030cf97ede2fbf6d2738b66c7 --- /dev/null +++ b/Building and using APIs with Go_transcript.txt @@ -0,0 +1,333 @@ +**Natalie Pistunovich:** We are live on the tubes, and welcome everyone, from whatever hour of the day you're joining us, and whenever you're hearing this, live or recorded. I think for once this is a European majority here, so we can say good evening... + +I'm joined today by my co-host, Johnny, and we have two guests. We have Anthony and we have Smile. Anthony, you are a senior software engineer at Delivery Hero here in Berlin, and you've been working with Go since 2013, and you're a co-founder of APIToolkit. And Smile is your co-founder, and you are a senior mobile engineer at Runtastic, and you also work with Go since about 2015. Welcome, it's nice to have you. + +**Anthony Alaribe:** Thanks for having us, Natalie. + +**Smile Egbai:** Thanks. Nice to be here. + +**Natalie Pistunovich:** Okay, so what does API stand for? Let's ask the big questions first. \[laughter\] + +**Anthony Alaribe:** Huge question. No pressure. + +**Natalie Pistunovich:** No pressure. I had to google that before coming into the episode... So anybody who's listening and googling right now, we get you. + +**Anthony Alaribe:** I mean, an API is an application programming interface, but I think it can mean a lot of things. I like to think of an API as a contract. So if you have, for example, two services, two machines, one server and one mobile application and they need to communicate with each other, both parties can decide on this contract, "If I ask for this, this is what you give to me. If I ask for this other thing, you give me this other thing." And that contract is basically what I think is an API. + +**Natalie Pistunovich:** \[04:13\] Okay. Smile, what is your take on what is an API. + +**Smile Egbai:** My take on what an API is roughly the same thing, and coming majorly from the mobile background, I would always say an API is more or less a series of endpoints that will give you resources. + +**Natalie Pistunovich:** A series of endpoints that will give you resources. That's short and to the point. + +**Johnny Boursiquot:** I like those definitions, they're very specific in terms of use. That's one way to use an API. Maybe you're making some HTTP requests, maybe you're dealing with a different transport mechanism, maybe you're doing it through gRPC, maybe you're dealing - perhaps fortunate or unfortunate - with some SOAP APIs... + +**Anthony Alaribe:** Hopefully not... + +**Johnny Boursiquot:** \[laughs\] You know, a lot of people still make a living on it, so I'm not knocking it. But to me, these are ways - how do you communicate with an API. To me, an API is, as the name implies, an Application Programming Interface, it's the interface to your functionality, to your code, to the behavior that you've produced and created. Perhaps it's not exposed to the web. Maybe you don't talk to some sort of HTTP endpoint to invoke the API. It could be as little as me creating a package in Go and basically saying "Okay, what is the surface area that I'm gonna allow somebody to interact with? What's the API to this package?" And you control that through the things you export out of your package and the things you choose not to export. What is internal to your API and what is external. What your users of your code get to actually use. So that surface area to me, at its most basic, is "How do you interact with something, whatever that thing is?" And obviously, I'm using Go packages here as the way to convey that... But to me, at the core of it, that's what that is. + +**Smile Egbai:** Yeah. + +**Anthony Alaribe:** Even if you have a codebase, maybe just like a library or a Go package, like you mentioned, with a series of exposed functions, those exposed functions are also part of the API. + +**Johnny Boursiquot:** And building the APIToolkit project - I'm very curious about what the offering is. With a name like APIToolkit, it sounds like you're sort of bringing something generic, in the sense of something that is general-purpose rather, for APIs. Did I get that about right? What is it that you're bringing to the game here? + +**Anthony Alaribe:** You know the two most difficult problems in computer science - naming things and the rest... \[laughter\] Yeah, finding names is hard, but I think this name kind of sums up a lot of what we're trying to do. We're trying to solve a few problems around APIs, and the name kind of gives it room to grow. + +The main problems that APIToolkit is trying to solve is around anomaly detection. So you have an API, in this case a primarily RESTful API, so an API that you expose over a transport. But let's stick with a REST API. So you have a REST API, with a series of endpoints, and each endpoint returns a series of fields, with values, and something changes. For example, you have a new engineer on the team who makes a change, and that change snowballs and just touches something else. Sometimes these kinds of changes can be difficult to spot. + +\[08:01\] Or even for example you have a legacy service which you're trying to rewrite, for different reasons, and you believe you know about the contract of that API, because it was very well documented... But then you rewrite that API and realize that - okay, there were some fields were there maybe ten years ago, no one knows about them anymore, but they are part of this contract, and you did not know about them, and so something breaks. So APIToolkit tries to find these issues before hopefully your customers find them. + +**Johnny Boursiquot:** So that's an interesting problem space... So if we can start to sort of dig in a little bit here - when I think of the operability of an API - say you make something public, and once you make it public, like you said, you basically have a contract with whoever is gonna be consuming that. So making changes to that API is very hard to do after you published it, after you have a v1, hence why you typically have to have some strategy for how you're gonna move customers onto the next version of whatever it is. + +So the use case you give is one where maybe you have some less-traveled parts of an API, and all of a sudden you think "Okay, can I or can I not change this, or make some modification tweaks, whatever it is, maybe even make some backwards-incompatible changes?" So what you're saying is that you want to have enough information about the usage, how the API is used, in order to know what is the likelihood of you making a change and breaking a particular endpoint in this set of endpoints. Did I get that right? + +**Anthony Alaribe:** I mean, very much. Of course, within the limits of what a machine can learn about your servers; it tries to understand "These are the fields that are usually being sent, this is the frequency at which they're being sent." This particular field, this name is usually - maybe it's a string, it's a text field, and it usually has this kind of format, this kind of mask. Maybe there's some alphanumeric mask, and then there's a space, and then there's a mask... So if someone changes that name field and starts to return numbers, then obviously something is wrong. Once you read a lot of bills and let people know "Hey, this used to be a text field with this format, and now it's returning numbers..." + +**Johnny Boursiquot:** So is this a development time tool? You want this detection before going live with a change, right? + +**Anthony Alaribe:** Yeah. In production, in real-time... + +**Johnny Boursiquot:** So before you go production, right? Or when are you detecting -- when are you saving my bacon? \[laughter\] + +**Anthony Alaribe:** I mean, Smile, you wanted to add something first? + +**Smile Egbai:** Ideally, we would all love for this detector to our code before you go to production, and definitely the tool would surely run both on your staging server and your production server, and whatever servers you would have... But it all depends on the tool itself detecting these anomalies. So if you have this running on your staging environment and you've probably run tests, integration tests within your staging environment, most like the APIToolkit would be able to detect such things. But if you are maybe one of the C/AL persons who probably pushed straight to master and deployed, like in a rodeo, then of course, this cannot really be detected in production. But that's the worst-case scenario. + +**Anthony Alaribe:** In a perfect world, a tool like this probably should not exist, because if people were building perfect systems, will test it, every single endpoint is tested, every single field is tested... + +**Natalie Pistunovich:** And everybody is following your updates... + +**Anthony Alaribe:** Yeah... \[laughs\] Then you really don't need something like this. But unfortunately, in the real world you always make compromises. People deploy to production on Friday nights... And the issues don't only come from within. So you have a service - the issue might not be that on the server side things are broken; the issue could also be that on the consumer side something is broken. + +\[12:11\] For example, if a consumer used to send a particular format of input, and someone deployed a web app or a mobile app that starts sending something else, then we also want to flag that and notify someone that "Hey, this server is accepting a very different input from its clients." And the clients could be third-parties, it could be other companies who don't have those best practices that you are following. + +**Smile Egbai:** Something that came up a lot in our interviews, where -- it's actually based on third-parties; we integrated a third-party provider. And I think it's also a bad practice when you version an API and you end up going back to version one and changing things. Meanwhile we know if you are going to be using API versioning, if you move to version 2, then most of us go to version 1, I introduce new changes. Or do you constantly see third-parties doing these kinds of things. We won't name names, but these kinds of things have come up a lot within our interviews, where you would integrate a third-party, usually within three minutes, and you wake up some weekend or a Monday and you see a lot of requests are broken, things are not going through. And why? Because someone, somewhere, or some machine somewhere changed an old version and added new fields, and changed old fields we should be good for now. It's not even about you making a mistake. It's more like an API you consume is breaking the contract that has been set. Within these kinds of situations, you definitely want to get an alert about something like this, so you could mitigate certain effects. + +**Anthony Alaribe:** I actually just remembered a conversation we had in the interviews with someone, and he said that they have a popular fintech provider that they consume/make use of, and this fintech provider probably deployed a change, and everything was broken for about a couple of hours. So they lost a lot of money, but of course, the fintech provider probably detected it and rolled back... But there was no public update; they did not say they did anything that broke anything, and they would not agree that they made any changes which broke things. + +If you have something that shows you and says "Hey, this contract broke at this time, and this is what changed", you can show and say "Hey, you definitely broke something. This is what you broke, at this time." + +**Break:** \[14:47\] + +**Natalie Pistunovich:** The first time that we talked about APIToolkit, I kind of imagined myself that a good scenario for me to use that would be as a consumer of different APIs... But just now y'all mentioned that this can be a great tool also for a provider of APIs, to make sure that all are in sync. So that's a very interesting accountability tool in addition to everything. + +**Johnny Boursiquot:** So along those lines... So it sounds like - and we're gonna get to where Go fits into all of this in a bit here... I wanna make sure I understand your project before we dive in deeper. So your website APIToolkit.io mentions observability. So observability means something very specific these days. It's no longer just monitoring. Observability implies other things. And obviously, we talked about the accountability here, so I'm sure if you consume -- let's imagine that I consume an API from a fintech company; we have an SLA, a service level agreement that says I can expect -- this is when I should, if at all, expect outages etc. So that's an SLA. That's a legally binding document that says this is the level source of it. + +And then you have tools, whether it be on the provider side or on the client side, which obviously APIToolkit.io could be something that you as a consumer or me as a buyer, the consumer of this API, could use to basically create my own SLOs, around the availability of this API that I'm paying for and I'm consuming. + +So what does observability look like for this project? Are you providing tools to track over time what the performance is? What kind of information are you surfacing with each tracking that you're doing? + +**Anthony Alaribe:** It's funny that you say observability is very specific these days. I actually think of observability as a word is just -- people are using it to mean so many different things... Probably APIToolkit is joining that list. But when we say observability in the APIToolkit context, we actually mean what the word says. It's a tool that allows you to observe both your API, but also clients and requests which are coming in, and third-parties that you call, monitor them down to the field level, so you can see each single field, you can see how often each field gets sent or gets received... But you can see this information over time, like you say. So you can see that the obvious API statistics, for example, that 20% of your requests to a particular endpoint end at a 404 error, but you can also see that the name field in this endpoint is a null, let's say 10% of the time. And that's something you want to fix. But it can go down over time and see a lot of these statistics over a period of time. So that's what we mean. + +**Johnny Boursiquot:** Is your focus on the shape of API requests and responses? We've come back to the field and the endpoints a few times; I wanna make sure that when we talk about observability, we're sort of localizing the word... We're contextualizing it to just basically the shape of the API; less so things like throughput and how long is the average response taking to come back, and things like that. Or does your solution cover that, too? + +**Anthony Alaribe:** \[20:01\] Yes. I mean, we give you your throughput, response time, and different percentiles. That's kind of like just the cherry on the cake; we have the information, so why not give it? But the exciting part of APIToolkit is that you get these statistics, but down to the level of the fields and the endpoints. If you have, for example, a fintech endpoint where you can accept payments, you can see that -- something that we're exploring is that you can see that there's an amount field which usually gets sent; you can see that this field is always being sent on this endpoint, but something that we're exploring is that you would even be able to plot that amount field. So you can on a graph see just basic averages that this field is usually -- you know, over time, every day, the average amount that people transact is $200, but you can just get these statistics down to the level of the fields as well. So we like to just call it field-level observability, pretty much. + +**Johnny Boursiquot:** Okay. But you have a few layers on top of that, in addition to just the fields stuff. + +**Anthony Alaribe:** Yeah. + +**Johnny Boursiquot:** So I do want to sort of dig into the Go side of things... What role is Go playing in this product? I know you have a lot of plugins and integrations sort of a - and go is obviously one of them - is go playing a bigger role in this project for you? + +**Anthony Alaribe:** In the development world right now, if you want to do a lot of numbers crunching, and handle/process a lot of things in little time, you have a lot of options. But in terms of the popular options, which are also easy to get into, if you're gonna on-board a lot of people on the project, I think Go brings that hands-down. It's a relatively simple language to on-board people to, but it's also very performant. We use Go for a lot of this. We're trying to process our customers' requests real-time, and that gives us very tight constraints, which Go is helping solve. + +**Johnny Boursiquot:** Can you think of ways that Go has helped you solve a problem that perhaps would have been more difficult with a different language? Or did you even consider any other language? Have you used any other language on this project? + +**Anthony Alaribe:** Our stack is actually -- we have clients for multiple programming languages, but in terms of the backend side, we mostly have Go and Haskell. + +**Johnny Boursiquot:** Wow, Go and Haskell. That's an interesting pairing. Tell me more, tell me more. + +**Natalie Pistunovich:** Which one is the frontend? \[laughter\] + +**Anthony Alaribe:** A lot of the frontend stuff is actually Haskell. We use Haskell for things where performance doesn't matter so much. But one of the reasons why we use Haskell is that for certain kinds of problems, it can be a little more precise to represent. I mean, you can think of APIToolkit as a kind of parser project, because we get people's requests and we're trying to parse it and understand it. We're looking at the fields and we want to understand what format it is, what types it is, and this is a problem that Haskell historically lends itself nicely for. But a lot of the other things are problems where you just really need to be fast; you need to process really fast, you have memory constraints, you don't want to have hundred-gigabyte memory servers, and that's where Go comes in very nicely. I don't know how the Go team did it with the Go garbage collector, but it's just really hard to beat, and you're able to process a lot of things very quickly. + +**Johnny Boursiquot:** So I'm curious where your toolkit lives. Are clients routing requests through it for this analytics to be captured, or are you at the edge, at the API gateway level, or where is this deployed? + +**Anthony Alaribe:** \[24:07\] So there's the ideal scenario, there's the long-term scenario, and there's the now. + +**Johnny Boursiquot:** Okay, let's do it. \[laughs\] + +**Anthony Alaribe:** In the now, what we have are language middleware. For example, in Golang you install middleware, depending on -- if you have a web service, you install middleware for whatever Golang router you use. And what this middleware does is it basically takes a copy of your requests and strips out sensitive information and information which you don't want on someone else's server... And then it sends this copy to our servers to process. + +You can sample the request, but we basically process every request that we get on the backend to make sure that that request looks like what the server was expecting. So that that request actually fits the contract that we have. So it's a system that's continuously learning. The more requests from a given service we get, the more we understand what that service is supposed to look like. And if something comes which doesn't fit into our understanding, then we flag an issue and notify someone on that project. + +**Johnny Boursiquot:** Okay. + +**Smile Egbai:** Think of it as a lawyer... + +**Anthony Alaribe:** Explain more, please... \[laughs\] + +**Johnny Boursiquot:** Explain. \[laughs\] + +**Smile Egbai:** You have two clients, and APIToolkit will seem to be the lawyer. One client doesn't have a lawyer, but the other client has a lawyer, so it's just sitting there in the middle, vetting contracts. You could think of it that way. + +**Johnny Boursiquot:** Okay, okay. So I can see how Go in this case would be a good choice as middleware, as one of your clients. So if the shuttling of data on request happens every time somebody makes a request, and a goroutine gets launched, and during the processing of the request hopefully anything that you're doing in the middleware is not delaying the actual processing of the request... So I imagine you structure it in a way that the impact to the processing of the request - because everything has to go through that middleware layer - the impact on the request itself is minimized as much as possible... And then behind the scenes you're just sending data to your server. + +**Anthony Alaribe:** This is why I told you there's the now, there's the idea, and there's the future. \[laughs\] So for a lot of projects, for example Golang projects, the middleware app which works very nicely, where the middleware is part of the project; Go's memory representation of information is very compact and useful. But if you're making use of a language like PHP - maybe not so much PHP anymore, but a language that tries to be stateless, then such a middleware app which doesn't really work, because you have to send each request to our servers, each single time a request comes. Whereas in Go we kind of send everything pretty much to one channel, and buffer the inputs from your server and stream them to our servers. In some other language ecosystem this is much harder to do. + +So the plan is to have these kinds of sidecars... Actually, we do have the sidecars, but there's just no one using it yet. But the plan is to have this sidecar which you can -- if you have like a Kubernetes cluster or a Docker cluster, instead of your application sending your request to our server directly, it sends it to the sidecar, which then just pre-processes it and sends it to us. That way your actual server doesn't need to keep a lot of information in memory, or doesn't need to do any processing at all. Or little processing. + +**Johnny Boursiquot:** \[27:59\] Okay. This sounds very much like the OpenTelemetry community right now. One of the design approaches is basically to have these collectors that are running, and you can certainly run them as sidecars if you wished... But if you have dedicated collectors where - as the name implies - they're collecting all the traces and metrics and whatnot, they on their own time can figure out basically the sending of whatever data you're collecting to your own servers. So you offload that processing out of the application, out of the service, and into this little collecting model. That seems to be a popular approach for this kind of problem solving. + +So what I'm hearing is that you're hoping that even though Go today makes the middleware approach good enough, that's ultimately not the long-term solution, right? + +**Anthony Alaribe:** I'm happy you actually mentioned OpenTelemetry... I mean, you can think of the information we're getting as traces; it's basically traces. So the entire system is very much inspired by how OpenTelemetry is designed. And the collector is basically language-specific. So if you have a Golang server, then you'd have a Golang middleware, which can do all of these fancy things I've mentioned... But if your server is in a programming language that doesn't have those things that makes Go great, then you have to rely on some other app, like you see in the OpenTelemetry world. + +So outside of the collectors and our server side, Go is the first line of contact in terms of processing these traces when they come in. That's where we actually see the real beauty of Go. + +**Johnny Boursiquot:** Right. For certain languages in certain tech stacks you're gonna have sort of optimized for those tech stacks sort of collectors and things. + +**Anthony Alaribe:** Precisely. + +**Johnny Boursiquot:** But pretty much everything that you ship, the moment it touches your network, then it's all Go all the way down, except of course for the Haskell stuff. This sounded like some sort of abstract syntax tree, with a new creation or something like that, trying to figure out what the shape of these APIs looks like. Okay, that makes sense. + +**Anthony Alaribe:** Precisely. + +**Break:** \[30:14\] + +**Natalie Pistunovich:** So talking about APIToolkit, you probably saw a lot of APIs, you've probably complained about a lot of APIs in general as you were coming up with the idea for this... + +**Anthony Alaribe:** Who hasn't? \[laughs\] + +**Natalie Pistunovich:** So from seeing so many APIs, and probably writing a bunch, what are some good practices that you see, that you follow, and what makes them good? + +**Smile Egbai:** So being a mobile developer, I probably consume more than I create. I don't create so much per se, but, one thing that always ate me, I know this might rupture some feathers, or something, but actually it's a bad practice. I'll just have to be brief... \[laughs\] When you are sending an error -- I sent an error, we got an HTTP response code of 200. I'm sorry, but it's a really big red flag for me. I always feel that way, and designing APIs, the error codes are there for a reason. + +\[32:03\] So if I request a resource that's not there, it gives me a 404. If there's a problem on the server, it gives me one of the 500s. 501, depending on what depending on what server it is. If the resource has moved, or something, I think it's a 301, or something... But don't give me a 200, then now give me a JSON and throw me an error, then you give me a constructive error body with another error code. Even when you return like a 400 or whatever, you could return an error body with your custom code if you want that, but let errors be errors. Let responsive key be responsive key + +I think that's one of the worst bad practices, and even when they are written in JSON, returning a number as a string; it's a small thing, but... It doesn't make sense. If it's a number, let it be a number. But don't put a number inside a string, or don't mask a JSON body in a string. It happens a lot, I see it a lot when I'm communicating with third-parties; I get a JSON body within a string, and then I have to encode it again, or decode it into a JSON object before turning it into an actual object. + +I think for me those are one of the pet peeves when you're talking about API best practices in relation to mobile development. + +**Anthony Alaribe:** I just remembered one issue Smile was ranting about some time ago... I think he spent some time investing some issue; I think there was a field that -- if the field doesn't exist, it would be a null. He was checking for this null... Maybe you wanna tell the story better... \[laughter\] + +**Smile Egbai:** Yeah, so there's this weird third-party provider we have, and when field doesn't exist instead, they always sent a null. But now, when the field doesn't exist, I don't know what they do or how they do it, they just don't send the field. If we are not decoding that JSON, it crashes the adaptor. + +**Anthony Alaribe:** You were telling me about a situation where they were sending this null, but as a string. + +**Smile Egbai:** Oh yeah, so that's the thing, they've fixed it. They've fixed it into sending a string, and they unfixed it into now sending the null. It's a stack of issues. So it seemed like for everything I complained about, they fixed it, and I don't know what really happened on the testing side, I don't know how that kept on passing through, but they kept on changing, and they were like the bane of my existence at some point... \[laughter\] Because I would come in to work, feeling very happy from the weekend, playing a game, and BOOM. There it is. And I had to do so many workarounds because of this. + +I just feel things like these are terrible API designs. I mean, most times, ideally, if you're going to send in a value which is a string as a null, that's fine; we could always check that. But removing the field completely? No. It just makes it hacky. Then bringing it back and instead of sending a null, you now send it as an empty string. I mean, that could still be worked, but it's just too many inconsistencies. + +**Natalie Pistunovich:** Yeah. They should be sending like a string to a URL of an image of a meme that says, "Null" + +**Smile Egbai:** Look, and the part that killed me the most... I remember it now. They sent the null as a string. You know when you put a null in a string? So instead of sending null, you send me null as a string... + +**Natalie Pistunovich:** Was there like a dot after the null? \[laughter\] + +**Johnny Boursiquot:** End of sentence... + +**Smile Egbai:** Our users were literally getting null. They were getting null; when you displayed the text, you were getting null. And I saw these things and I was like, "How...?!" I didn't believe I made that kind of mistake, and I could not for the life of me understand why I was displaying null null. I just couldn't get it. When I had to go to the restaurant, I now saw a string null, and I was like, "Okay... That's nice." + +**Johnny Boursiquot:** That did not live up to your namesake. It did not put a smile on your face. \[laughter\] + +**Smile Egbai:** \[36:12\] Oh no, it didn't. + +**Johnny Boursiquot:** Oh, man... Given that you're probably receiving a lot of the data you need to process in JSON - correct me if I'm wrong - how is it dealing with JSON in Go? Do you find that the standard library works just fine, or have you had to use something else? + +**Anthony Alaribe:** Go actually handles JSON quite nicely. Unfortunately, most programming languages in the world are not written in Go, so this makes it quite difficult to work with. For example, in the dynamically-typed ecosystem it's very common to have one field that can be three things; maybe it's an object sometimes, other times it's a string, maybe sometimes it's null... And this is relatively difficult to represent in Go. You need to maybe use an empty interface, and kind of Go Tree + +So this problem -- it kind of makes it harder to consume APIs in Go. It's really nice to produce in Go, because every language who consumes an API that was made in Go would get something very consistent. But if you as a Go service is the one consuming, then you need to be prepared to deal with these things that are very difficult to represent in Go. + +Unfortunately, we outsourced a lot of these kinds of problems to Haskell... The language just exists to solve these kinds of parsing problems, and it's just more mature in solving these kinds of parsing problems than Go. Definitely, these things can be done in Go. You can save it into an interface and have a cast a Go tree list of possible alternatives, but the solutions are just easier in Haskell. + +**Johnny Boursiquot:** Right. So I'm assuming that in the next iteration of the product, where you do have a separate - let's just go with the term 'sidecar' for now, but... You do have a separate collector, if you want. Then the constraint really becomes how fast can you show the results of your analysis, of your anomaly detections in your own dashboard, and not how fast you can get out of the day of a particular request, right? Then would you say that it doesn't matter as much how fast the processing of the JSON is if you are collecting and maybe doing some compacting on the collector side, maybe doing some sampling right before the data even reaches the edge of the network? Would you say that processing JSON really at that stage doesn't really matter? Or do you have a similar concern? + +**Anthony Alaribe:** So what we are doing actually -- I mean, you can divide the problem that we're solving into two things. One is that we're building a model that represents the API, and the second thing is that we get a stream of APIs and we're comparing each item in this stream against this model. So this is kind of the validation process. So working with this stream and comparing is what we are doing in Go. But building this model which we're then going to compare against is just what we are leveraging Haskell for. + +Processing the stream is where a lot of the anomaly detection happens, because you wanna go to the traffic in real-time, and if there's an issue, you want to alert someone. But in terms of building the model and displaying it on the dashboard, that's something that -- there's no real time constraints. I mean, there is a time constraint, but it's just more flexible. You don't need the most performant thing to display this model on a dashboard, for example; but you need the fastest thing you can get if you wanna process -- for example, if you have customers who are maybe doing a million requests per second, we need something that can actually process one million requests per second. Basically, process it and compare it against a model and say "Okay, is this request valid? Yes." Then throw it away. "Is this request valid? Yes." Throw it away. Pretty much that. + +**Natalie Pistunovich:** \[40:24\] The APIToolkit is available, it's open source on GitHub, right? + +**Anthony Alaribe:** Well, right now-- + +**Johnny Boursiquot:** He's like, "Well, what happened was..." \[laughter\] + +**Anthony Alaribe:** You know, it's a business. The core middleware is on GitHub, but we are just slowly rolling things out. Actually, right now APIToolkit is in a closed beta, where we're testing with a few companies. We're just going to gradually test with more companies on our waitlist. There's a waitlist people can join. + +Over time, we're sure that we can handle whatever our customers throw at us, because if you think about it -- we need to be able to handle the sum of the traffic from all our customers. So if we have one customer making a million requests per second and another making two million requests per second, we need to be able to handle three million requests per second. So we're just gradually making sure that we can handle that. + +**Johnny Boursiquot:** Unless you sample, baby. You've gotta sample. \[laughter\] + +**Anthony Alaribe:** Yeah, that's a problem... I mean, we eventually would encourage sampling, but we need to be able to handle situations where there is no sampling... Because you never know where an issue would be. The issue could be in the requests you sample out. So that should be a decision which the users of APIToolkit make for themselves. + +**Johnny Boursiquot:** Yeah, that sounds like a business decision. You can kind of punt that down the road a little bit. Cool. So given that your project has no open source components, how would you say the community can help you, if at all? Or is it not at the stage yet where you can get the help from anybody? + +**Anthony Alaribe:** Yeah, I think it's actually not so much at that stage yet. We are gonna get to a point where we would need clients for most languages out there... + +**Johnny Boursiquot:** You need help with getting clients? \[laughs\] + +**Anthony Alaribe:** I mean, that's something that we would figure out when each new language that we need a client for comes up. But yeah... + +**Johnny Boursiquot:** Okay. So clients for languages, not customers-clients. + +**Anthony Alaribe:** No, no, no. Clients! \[laughter\] Collectors, precisely. + +**Johnny Boursiquot:** Got it. + +**Anthony Alaribe:** I mean clients as customers - that is always welcome... \[laughter\] No one says no to clients. But yeah. + +**Natalie Pistunovich:** Fun. Okay. + +**Johnny Boursiquot:** Is it that time, Natalie? + +**Natalie Pistunovich:** It is. It is that time. + +**Johnny Boursiquot:** Here we go. + +**Natalie Pistunovich:** Time to invite... Mat. + +**Jingle:** \[42:49\] to \[43:07\] + +**Natalie Pistunovich:** Hey, now that we have Mat's song with us, we can officially approach the second part of the show... \[laughs\] In the last five minutes. Guys, we're interested in your unpopular opinions. Who would like to go first? + +**Johnny Boursiquot:** I don't know, we've heard some from Smile... \[laughs\] Well, I don't think they're that unpopular. He raised some fair points. Don't send the null screens in your responses. \[laughs\] + +**Natalie Pistunovich:** Use the HTTP code! + +**Smile Egbai:** Oh, no... + +**Natalie Pistunovich:** Use the HTTP codes. \[laughs\] + +**Smile Egbai:** Yeah, please. Always use the HTTP codes. They are there for a reason. + +**Natalie Pistunovich:** So what is your actual unpopular opinion? + +**Smile Egbai:** Okay, it's not really API or tech-related, but it's more of educational. It's more like -- I think the current school system, or this school system that's always been, just doesn't work. You subject children to -- I don't know, is it eight hours, ten hours of school, with 15-30 minutes break, and they come back home to 2-3 hours' worth of homework... + +\[44:09\] They are like bank workers. The kids are basically in this reformed sweatshop that you pay for. That's how I see it. You send your kids to a sweatshop, and you pay for them to suffer. Because -- I mean, they are children. + +Okay, like - you have a full-time job, you work eight hours. When you are done, you're tired mentally, you're stressed. Now imagine children having to go for like 8-10 hours outside their home, they come back, and they have another 2-3 hours of assignments. They have no life besides school. We don't build children's character. I really think the school system should be modified in a way there are shorter times for course learning. + +I think there was a study where children who are homeschooled actually performed better than kids who actually go through the traditional education systems. So that's my unpopular opinion. + +**Anthony Alaribe:** Were you scarred by this growing up? \[laughs\] + +**Smile Egbai:** Oh yeah, I hated homework, to be fair. I did it, but I hated it. + +**Johnny Boursiquot:** It's almost like we're training our children to go into some pre-existing system where they go to work 9 to 5, and... \[laughs\] Yeah, we're training workers, aren't we? + +**Smile Egbai:** Yup, we are. We are giving them -- it's a stress test. + +**Johnny Boursiquot:** "Can you handle what's gonna come for you after school?" \[laughs\] + +**Smile Egbai:** Yeah. + +**Johnny Boursiquot:** Alright, alright... Well, I don't know if that's gonna be unpopular. I don't disagree. I hear you. + +**Anthony Alaribe:** I definitely agree. I think I saw a talk by Ken, somebody, sorry I don't remember, a TED talk, and he was like "Kids who want to go into kindergarten get interviewed." I was like, "What exactly are you interviewing these kids for?" Like, are you gonna ask them "What have you done with your three years of your life?" \[laughter\] Just breastfeeding? + +**Johnny Boursiquot:** "I don't know... Pee, poo, try to walk around... I don't know." + +**Natalie Pistunovich:** "Show me how you use the touchscreen. Can you open and close apps?" \[laughter\] + +**Anthony Alaribe:** Someone said something like -- yeah, he saw an advert that said "College begins at kindergarten." And he was like "What...?!" + +**Johnny Boursiquot:** Good Lord... + +**Anthony Alaribe:** Kindergarten begins at kindergarten. Not college. + +**Smile Egbai:** I think it was that same talk. + +**Johnny Boursiquot:** From Ken, from TEDx? \[laughter\] + +**Anthony Alaribe:** Yeah, that guy. + +**Johnny Boursiquot:** Shout-out to Ken from TEDx. + +**Natalie Pistunovich:** Anthony, what's your unpopular opinion? + +**Anthony Alaribe:** Mine is just that I actually think German as a language is quite nice. I have been in Germany for three years and I have been avoiding the language until this year... And after some months of classes, it just makes a lot of sense. The rules are so -- I can't imagine learning English if I didn't learn it as a child... Because when I think of the rules in English - like, there's almost no rules. There's exceptions, and exceptions of exceptions... And in German it just makes sense. So yeah... + +**Johnny Boursiquot:** Okay, okay... For a lack of a better term, you were afraid of it until you started learning it. + +**Anthony Alaribe:** Yes. There's this popular stereotype about the German language. It's like, machine, so harsh. But maybe that's also what makes it nice, because then the rules are very clear. + +**Natalie Pistunovich:** Until you hit exceptions. \[laughter\] But there's probably less than in English. I think I would agree with that. + +**Anthony Alaribe:** I hope... I'm still learning. + +**Natalie Pistunovich:** Yeah. Well, let's see. So your unpopular opinion is that German is actually nice. + +**Anthony Alaribe:** Yes. + +**Johnny Boursiquot:** \[laughs\] + +**Natalie Pistunovich:** I will be curious to see how is that gonna vote out... + +**Anthony Alaribe:** Yeah, we'll see... + +**Natalie Pistunovich:** You can think of it also as a sort of a contract, sort of an API, right? Languages... + +**Anthony Alaribe:** Yeah, my API contract with the German population. With the ladies at the male dom office. + +**Natalie Pistunovich:** City registration and all the bureaucracy... That is a serious contract with them. They will not upgrade to a version two... \[laughter\] + +**Anthony Alaribe:** Yes, they will not. \[laughs\] + +**Natalie Pistunovich:** Cool. Alright, well, thank you very much for joining, and thanks everybody who tuned in. We look forward to hearing about your contracts with the world, I guess, and all the texts. + +**Johnny Boursiquot:** Yeah, good luck, guys. + +**Anthony Alaribe:** Thanks for having us, guys. diff --git a/Debugging Go_transcript.txt b/Debugging Go_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..499f6a9caa06762a44d8cad8b58b994a89a763c4 --- /dev/null +++ b/Debugging Go_transcript.txt @@ -0,0 +1,223 @@ +**Natalie Pistunovich:** So today we are talking about debugging Go, and I have to share that this is actually a re-recording of the episode, because as much as we did our best to have redundancy, and a Babel local recording, and saving, and whatnot, the internet was bad, the audio was bad, everything was bad, so we are re-recording, and we hope that this time there'll be no black boxes for us to poke... + +I am joined by my co-host, Ian. Hey, Ian. How are you doing? + +**Ian Lopshire:** I'm doing well. This one's gonna be interesting for me, because I'm definitely still one of those print people... So we'll see if I learn anything. + +**Natalie Pistunovich:** We are also joined by Liran Haimovitch and by Tiago Queiroz. Gentlemen, would you like to introduce yourselves? Liran, you are joining us from one hour in the future; you're joining us one hour away from Berlin, you are in Tel Aviv. + +**Liran Haimovitch:** Yeah. So I can definitely relate... Print has its privileges, because debuggers can be so cumbersome, and often hard to use in so many cases, and it's so much easier to just say "I want to know what's happening on this line. Let me add a line, recompile, rebuild, ship it and get to know whatever is happening there." However, rebuilding the software and shipping it isn't always that easy, which is kind of where Rookout comes in. + +So my name is Liran Haimovitch, I'm the CTO of Rookout. I spent about a decade doing cybersecurity, and for the past - wow, is it six years now? ...I've been focusing mostly on observability and how to use dynamic instrumentations and other concepts from cybersecurity to make observability, debugging more agile and easier to use in production environments. + +**Natalie Pistunovich:** Thank you for joining us. Tiago, you are also here in Berlin. + +**Tiago Queiroz:** Yes. Hi, everybody. I'm Tiago. Yeah, I live here in Berlin. I'm originally Brazilian. So I'm a software engineer, I've been working mainly, almost exclusively with Go for more than five years now. Currently, I work for Elastic, on the Elastic Agent and Beats. It's interesting, because lots of our debugging there is basically logs, because we basically build binaries, ship, people run it... They sometimes run into issues, because - well, no software is perfect... So actually, log statements and print debugs is still one of the main ways to discover what's happening, especially when the issues are not on our machines. + +Yeah, and most of my work experience before Elastic was with startups and microservices. I also worked a lot in the sense of like bringing observability into some places, and I was always advocating for better observability and debugging tools. So yeah, that's a topic that I enjoy a lot, and try to always bring more, let's say awareness at my workplaces; we need to make things observable and easier to understand when they break... So they should almost tell us "Hey, I broke because of that." + +**Natalie Pistunovich:** The AI... The AI will be telling you "You broke it because of that." + +**Tiago Queiroz:** Yeah. + +**Natalie Pistunovich:** So let's talk about some practices that we have today, without the AI. What are some good and bad practices in debugging, that are maybe specific to Go? + +**Tiago Queiroz:** I think definitely it's a good practice to always have logs on your application. Make sure you write your application and you have proper log statements. + +It's a very broad concept what is proper or enough, and something that you basically learn by doing, by seeing all the applications... But definitely, you should have enough logs for things, especially for like -- if you write a web server, ideally as soon as the HTTP service runs, it should say "Hey, I'm running on this port, on this host." It should show you some information of how it's starting, what it's doing, printing some configuration that might change for environments and might be very critical to solve some issues. + +\[07:55\] I think the best example - sometimes you run something in staging, on a port, and then in production on another one, but then if you don't set the configuration right, things just don't work, and then if you don't have any log statement, then you think "Hey, I'm actually starting my web server on port 3000, not 8000." You might spend a few hours debugging. + +So definitely really think when the application is running or starting what an engineer or even a user that's using your application might need to know if something goes wrong, or just to know about the current state of the application. + +**Liran Haimovitch:** So the tricky thing about debugging any language, I would say, is that you're probably debugging the unpredictable... Because if something is predictably broken, then you're probably going to have a log line somewhere saying -- the system is essentially saying "I'm broke. This is why. Just go ahead and fix it." And maybe the system is writing it out very clearly, or maybe you're just knowledgeable enough in the domain, in the application, whatever, to just get a sense of what's going on and quickly pick up the breadcrumbs and know what is what. + +If you are bothering to debug something, then it's unpredictable to you in some way. Maybe because the debugging is super-complex, maybe because there's many dependencies, maybe because you're not that much of an expert in Go or in the application itself... And that's perfectly fine, nobody's perfect, and we all have our knowledge gaps. But whatever it is, you kind of have to keep in mind that if something is challenging, it's unpredictable in some ways, and you have to think about how do you address that unpredictability, whether it's by pushing those print statements to kind of get a better sense, whether it's by using a debugger, a live debugger, or looking at other observability data... Whatever it is, you kind of have to address the unpredictability to get a sense of what is known and build from that. + +And speaking of Go in particular, one thing I've found that often confounds people is that Go is not supposed to have exceptions, and yet it somehow kind of does. Now, I'm not trying to argue with anyone about anything, but panics and defers act very similar to exception throwing and catching them. And if you are new to Go, or maybe not very familiar with the codebase, panics are often overlooked, and some libraries actually even use intentionally panics as a form of exceptions to unwind the stacks, if the stack is very deep. + +But either way, if you're looking at an error flow, if you're looking at an unpredictable outcome, it's always good to think about "Is there any chance something is panicking here? And if it is, how can I test it or how can I make sure nothing is panicking." Because you'll often find that, whether intentionally or unintentionally, there could be panics in there somewhere that's throwing your code into some unpredictable flow that's causing you a headache. + +**Natalie Pistunovich:** So what would you say is the difference between debugging locally versus debugging in the cloud? + +**Tiago Queiroz:** I think the main difference is debugging locally you can run the debug in slow motion. Actually, I was at GopherCon in a workshop from Bill Kennedy, and he had told this quote that was very interesting: "The function of the debugger is to run your bug in slow motion." Basically, that's the only thing it does. + +**Natalie Pistunovich:** It's a nice quote. + +**Tiago Queiroz:** It's actually very interesting... And at the end of the day, that's it, right? When you attach a debugger to an application, usually you only do it locally. You can literally run the code in slow motion, stop and introspect things, right? I find it extremely useful to have debuggers when you do not fully understand the workflow of the program, or the code paths, or there are too many dynamic things happening. If I really understand the program, usually I just always go for print statements. + +\[12:05\] Of course, then I'm like, "Okay, \[unintelligible 00:12:04.23\] I usually have a hint of what's happening. When I have no clue, usually \[unintelligible 00:12:11.28\] And I think that's one of the main differences from the cloud - you can actually attach a debugger. You can run things in slow motion. Because if you're in the cloud, even if it's a Kubernetes deployment on your staging environment or something, usually there are some proxies, some things in front of your application that might have some timeouts, and things like that, and once you put a debugger that says like "Hey, stop my HTTP request for (I don't know) five minutes, while I think, or maybe 30 seconds", then time out. Especially in Go, you get to some point and \[unintelligible 00:12:45.06\] the context of this request. Time it out. So I think that's the main difference. + +**Liran Haimovitch:** Yeah. So debugging in the cloud sucks... It just sucks, because the two techniques we're used to working with for debugging just don't work as well. 99% of the times you can't touch a debugger; when you're running in the cloud, you can't attach a traditional debugger, whether it's because you don't have the access for it, whether because the debugger itself is gonna be holding the process, as Tiago mentioned, but that's gonna kill any health probes. The process \[unintelligible 00:13:23.14\] it's gonna get AWS, or Kubernetes, or whoever is orchestrating the application to terminate your process, and then spin a new one without your debugger attached. + +And there are other slew of technical problems, but 99% of the time you're not gonna be able to attach that debugger to have the beauty of watching your code move in slow motion, so that you can actually spot the bug as it's moving at lightning speed through your code. + +Also, logs, which are nice -- they're definitely better than debuggers, but you can't use it all. But again, when you go to the realm of unpredictability, logs are more often than not -- I wouldn't say enough or not enough, because that's a harsh statement, but they can be challenging to read. They can be challenging to sift through, and grasp the bigger picture and find essentially the needle in the haystack. And it's not just about how many logs you have, it's also about where are they positioned compared to the bug, and how much information do they actually reveal about the inner workings of the software. + +And as Ian literally started the call with about "I like to add prints to get a better sense of things", and I think prints, when your code is running in the cloud - that takes a while. You have to add the code, then you have to either build it locally, or push it through the remote, and then you have \[unintelligible 00:14:49.13\] CI/CD, deploy somehow... It can easily take you 20 minutes if it's a personal environment to add that log line, or it can take you hours or days if it's a shared environment or even a production environment. + +So you can still debug with prints, but changing them becomes very painful and slow... Which is part of the reason that the new form of debuggers, the live debuggers have kind of spun up, which are super-useful... And we actually released Go support last year. I think that's something that makes cloud debugging so much easier, when you can get that traditional sense of setting breakpoints and seeing the data, while still working with load balancers and API gateways in Kubernetes, and all that magic. + +**Ian Lopshire:** Yeah, I wanna hear more about that cloud debugging... I've not heard of that. It's kind of blowing my mind... Can you just elaborate a little bit? + +**Liran Haimovitch:** \[15:46\] Sure. So at Rookout what we aim to do is provide you with developer observability in general, but first and foremost a debugger-like experience, without any debugger involved, essentially. What we ask you to do is install our agent or SDKs. For Go it's just a Go module. And that Go module essentially connects to an orchestration service when you run. And then we kind of provide you with "What would happen if I'd placed the breakpoint here?" So you go into our platform, you place a breakpoint (an unbreaking breakpoint) on that line of code, and the next time this line is executed - it can be once, it can be a hundred times, it can be more - we collect snapshots. And those snapshots look exactly like a breakpoint would show, what a debugger would show if a breakpoint was hit. So you would get a stack trace, you would get the local variable values, everything... Except those snapshots are asynchronous. So the application doesn't stop, it doesn't wait for you to read through the slow motion of debugging. + +We slow the application for a millisecond or two when we capture that snapshot, and then you can read it, take a look at it offline, at your leisure. You can go back and forth between different snapshots, so you can see how the code is running, supposedly in slow motion, except the code ran really fast, as it was executed. + +**Break:** \[17:12\] + +**Natalie Pistunovich:** So you mentioned that you can do this cloud debugging with the snapshots of how things are when you place those breakpoints; as a way of doing that in the cloud it is quite easy, or easier. What general some good practices that you have to make debugging easier? What can you do to help yourself, your future self, your colleagues when you do debugging, generally? What are those practices that you're following? + +**Liran Haimovitch:** So obviously, logging. As Tiago mentioned, logs are awesome. You shouldn't skimp on logging. You should also keep in mind that there is such a thing as too much logs, especially when you get your bill from your log aggregator... So do watch out for that. But that's an entire topic, about optimizing the cost and volume of logs, and we're not gonna dive into that. From my experience, most people err on the side of too few logs, especially early on. + +The other thing, which touches back to a point I've mentioned, is when you're writing code, focus on the error conditions. I think I read somewhere that when you write code, about 20% of your code deals with the happy flow, deals with what should happen, and then 80% of your code deals with what shouldn't happen; with validating inputs, with responding to errors, with network disconnections, with database errors, with God knows what. And that's where 80% of your code is gonna lie. And complexity-wise, way more than 80% of your code is gonna be there. Try to focus as much as you can on thinking through those errors as you're writing the code. + +While it's definitely useful to write a log line saying you've entered a function - maybe it's an important function, maybe it's gonna be useful someday - that log line is far more likely to be noise and far more likely to end up in the bin of those logs that are costing you way too much money, and it's time to turn them off. On the other hand, I think logs for error states - those are way more useful. If something went wrong, make sure you write a log line about it. + +I think one of our first customers at Rookout actually used Rookout at the time it was a Node application to debug a bug they'd been chasing for over six months. It was a super-complex bug, and they were doing OAuth three-sided authentications, and they got a cookie... And they actually had this error validation code that assumed that if the cookie was too big, they were gonna truncate it. Now, the cookie held a JWT token; obviously, it's a signed artifact. If you truncate a signed artifact, the signature is not going to be valid, sending you down through all kinds of pain. And they literally had a comment on that if statement, "Add a log line here. Add a to-do comment to add a log line." It literally said "This should never happen. Add a log line." And they didn't add a log line. And they've spent over six months trying to figure out what was going wrong, and why some users couldn't log into the system... And that was because their cookies were too large, because of various settings somewhere else across the world. + +**Tiago Queiroz:** Yeah, definitely watch out for your error logs. I think that's one of the most helpful things. But not just the logs themselves, but like graphed, good error messages. Don't just say "Hey, I broke" or \[unintelligible 00:22:31.07\] Try to give a context of why that thing failed. A simple thing. For example, if it's an insert, at least print the name of the table. Maybe you cannot print the whole SQL statement, but you can print a table, or maybe say "Failure inserting a user, or doing something." + +So try to explain it, giving context why you're in that error path. Like Liran said, watch out for those error handling code; that's usually where lots of bugs and issues are gonna happen. Errors are not well handled, or even worse - they're handled, but they're just hidden, and then when you run the application on your logs and things like that you don't see that there was an error. Or maybe the application exits from like an error handling path, but it doesn't tell you why. There's an error, but it doesn't tell you why. + +I actually have a good example. Recently, we had some PRs on our applications, like adding more logs, adding more context to this error log, because of something like "Oh, I cannot start because the log file already exists. So maybe there's another instance." But it didn't tell where the log file was located; it just says "The logfile." It's there, but okay, where is the file? How do I find it? ...to find which process is holding that file. + +\[23:54\] And then maybe if you go through lots and lots of logs, at the very beginning of the application it says "Oh, my configuration path here", and then you have to know that you need to join that with something else to find the log file. But why not just print together? Like, "Hey, log file, here's the whole full path, already exists. I cannot start." + +So sometimes simple things like that can make the whole difference... So maybe found the recover all error I have to shut down, but here's the error, here's the situation. Try to show the code path that led to the error, to error handling, and give as much context on your error message as possible, so someone reading the log, or even reading the error in whatever observability tool you have - the person reading it can better understand why that happened. + +**Ian Lopshire:** So we've talked a lot about logging, and it sounds like for errors, logging is the way to go. Where do you think the place of stats are? Like stats, and counts, and histograms of latencies... Where does that sit in the debugging landscape? + +**Liran Haimovitch:** That's a tricky question... Metrics are -- well, they are one of the most complex and powerful ways to monitor software. You can get a lot done, especially if you're looking at large-scale systems, production systems and those sort of things. But you also need to have some basic understanding of how metrics works, and how to approach them. You also need some real-world experience with your specific application, because numbers that can be awesome for one application can be terrible for another... And some of this lies in the business context, in the domain you're operating in, and some of it just lies with the characteristics of your applications, and the various decisions you've made along the way. + +Obviously, there are a few great places to start, that are always good - looking at your requests per second, looking at essentially how many requests your application is processing per second, looking at your latency, how long does it take you to process requests, looking at your error rates, how many requests are successful versus how many are not... But then there are all sorts of fancy ways to look at data which change. Do I look at the medium, the average, the p90, the p95, the p99? It depends on the context. Some applications might make perfect sense to look at p95, while others you might have to look at p99.9... While others still might look at the average. + +Also, what dimensions do you look at those metrics? How do you break them down? Do you have multiple regions? Do you want to monitor it per regions? Do you have multiple endpoints? Do you want to monitor it per endpoint? The different endpoints have different tolerance for latency or for errors... And the list goes on and on and on. + +So on the one hand, if you're looking to monitor an application in production, getting some metrics in there is a great start... But it's also a dual learning curve, both on learning how to use metrics in general, but also about learning what's normal and what's good versus what's bad for your specific application. + +**Tiago Queiroz:** Yeah, I think metrics are not -- I wouldn't call them exactly a debugging tool. They're more like to understand -- because metrics are aggregations, right? Usually, we get average requests per second, or average -- like, given a number of requests of instructions, of executions, in a given time, those are metrics usually for request latency, throughput, maybe database connections. + +\[27:45\] So usually, you cannot pinpoint one thing on that one transaction, and then get the metric and use those aggregations. But they also help you to maybe understand another thing, like which bugs are actually, let's say worth fighting right now. So if you have metrics like error rates, and you have like 50% of your requests are always failing - definitely, you have a big issue there. Lots of people are being affected by this bug. But if it's 0.1% of your requests are failing or are too slow, and that's basically meaningless in the number of users, maybe there are actually other bigger bugs that are gonna have a bad impact for your users to solve. + +So definitely, I think metrics can help to prioritize, and definitely performance issues. So if you want to improve performance or you actually have performance issues, then those metrics will be there... Even though you can handle performance issues as bugs, when someone talks "Oh, there is a bug. Let's debug something", usually I think more like a code level, a weird behavior of one specific application, rather than let's say "Every time I have too many requests, the database dies." Usually, I don't think that is a bug. It could be a bug in the code, that you're making too many requests to the database, or unoptimized SQL queries; it could be. But at least that's how my mind works when I heard the word "bug" or "debugging". + +**Natalie Pistunovich:** We talked about good and bad practices of debugging, we talked about debugging in the cloud versus on-premise; we talked about stats versus errors, or logs. Let's talk about debugging in production versus not. Are you debugging in production? If yes, how? + +**Tiago Queiroz:** Yeah, I think in the case of working with Elastic, it's a very interesting case of debugging in production, because -- well, we don't have access to the users' environments. So basically, we're only relying on logs, and asking people "Hey, what have you done?" So that's where you actually see two things - the power of actually having good logs, and also you see the other side, that logs can be too verbose as well. Like, you have some applications, and if you enable log debugging and it has a very high throughput, then sometimes you get like 100 megabytes of logs, and it's like three seconds of execution, because it's just too verbose for some stuff. + +So yeah, I think definitely debugging in production has this facet of - sometimes you actually don't have access to the environment, and more often than not you're mostly relying on logs and metrics, and that's all you have. So you're basically, let's say, looking to the past. Okay, something happened, I have things that point me to what happened in the past, not right now. That's different from debugging locally, or even -- usually, in staging you have more control of the environment. + +**Natalie Pistunovich:** Then you can't really reproduce what the person who reported this has, because it's not like "Send me your code and I'll try to run this, see how that works..." + +**Tiago Queiroz:** Yeah, sometimes you just can't, or sometimes you do your best to get as close as possible... But yeah, that's a thing that reminds me of something - sometimes you can't have access to production or to the actual environment, but you can reproduce a close enough experiment, if you have enough information. "Okay, every time those situations happen, then there's a bug." Then you quote your development-- staging environment and you reproduce it as close as possible, and then you can actually run your debugger, run your tools. + +**Liran Haimovitch:** One of the fun things about being the CTO of Rookout is I not only get to see how our engineers debug in production, but I get to work with many of our customers while debugging in production. And many of our customers have - I wouldn't say given up, but come to realize that they need a better solution than the traditional approach for debugging. + +\[31:51\] Monitoring tools were not meant for debugging. Monitoring tools are meant to show you something is wrong. Your application is down, you're having some issues with some of your endpoints, some of your users are not getting the best service... But they were never designed to pinpoint exactly what's going wrong, especially many of those issues are code-related issues. Today, with cloud-native applications, when you're essentially moving away so much of the infrastructure to somebody else - your AWS, or Google Cloud, or Azure, or whatever you're picking, most of your bugs, most of your issues are somewhere in the code, and you need a lot of granularity. + +Many of our customers have figured out that whether they're trying to make their engineers more productive, or solve customer issues faster, or both, they prefer to be able to truly debug in production. And we see a few considerations and hurdles in how you approach that. I mean, security is a big thing. You have to make sure that whatever tool you're using, you know who can access the system, you know how it is being accessed, that everything is secure... There are a lot of privacy regulations across the world in general, in Europe in particular, so you have to think through that, you have to know what private data is being collected, is it being removed through the process, is it being stored, who is storing it...? + +And obviously, you have to think about performance and availability. Will the access and the tools I provide my engineers somehow impact the stability of my application? Will it impact the performance of my applications? + +Whenever you're thinking about debugging in production - and it doesn't matter if you're giving them some fancy tool such as Rookout, or providing them with SSH access to go do their thing - you need to think about the implications in those elements, especially how the non-functional requirements of your system, which can vary greatly between domains, between companies, even between applications within the same company. You have to think about that in how you address them. But other than that, other than the significant extra risks involved, you still go back to the same fundamentals of debugging we've discussed through this call. + +Once in a full moon, I actually meet somebody who decides to attach a traditional debugger to production, because either the issue is bad enough that they have no choice and that's all they have to do, or because for some peculiar reason it doesn't disrupt their system all that bad, for some reason, and they can afford to... But at the end of the day you end up with the same techniques, just a different risk aversion scale. + +**Ian Lopshire:** So I wanna go back to logging, just a little bit. I know we've talked about it a ton... But we mentioned earlier that sometimes there's not enough logs, sometimes there's way too many... Do you think log levels play a role in that? Is there something that we should use? In the Go community there's a lot of argument that "Ah, just a single log is enough", and other people like the levels... What are y'all's thoughts? + +**Tiago Queiroz:** I would say it depends. I've gravitated towards different opinions of like "We should only have one level" or "We should have multiple ones." I think it definitely depends on what kind of applications you are developing. If you deploy your own microservices on your infrastructure, then you have way more control, so maybe you can get away with less log levels, and really craft good log messages. But if you're actually providing binaries that people are gonna download and run, then at least a debug level would help, because then you can say "Okay, those are things that most people don't need." It's just gonna waste disk space, or log storage if they ship there somewhere... So 99% of the time we're never gonna need it. But then when there's an issue, then we actually need to see that. + +\[36:03\] In some of the applications that we have - at Elastic we actually have even a different approach. We do have the log levels, but also we can filter specifically loggers, or different parts of the application \[unintelligible 00:36:15.17\] So we can say "Oh, I want log debugs from our inputs, or from our parsers, or from something like that. Or from the output." It's almost like setting different log levels for different components of the application. + +So that's also an approach that I find quite interesting. It can be quite useful, especially when you have the situation of like -- logs are probably gonna be too verbose, some of them, but then you can filter out the verbose ones, not like even generate them, and then you focus on the things you actually need. + +**Liran Haimovitch:** So log levels are super-important. You can't do without them, at least if your system is any scale... Especially when it comes to how much traffic it's handling, because you're gonna have to cut logs somehow, and verbosity levels is the easiest way to get started with. Obviously, you can get fancy later on, as Tiago has mentioned. + +But when thinking of log levels, one thing you have to keep in mind - when you're looking at most SaaS operations today, when the same team both builds and operates the software, log levels, especially in the long-term, tend to be gamed. + +I'll take an example... Let's say you're running the application for the first time; you're probably spewing debug logs. You have zero traffic, you're printing everything... You're just dying for somebody to bother to go to your website and whatnot... So why not - just print everything. Then, later on, the application grows, you get more traffic, you move to infra; still no biggie. But then, at some point, your application explodes, you have tons of users, you're paying way too much money for some log aggregator somewhere, and all of a sudden some ops guy or girl comes and says "Let's cut back. We need to move to one level of verbosity and above, because we're just wasting so much money and nobody is looking at those logs anyway." + +Now, you might have pretty well-defined verbosity levels, you might not have those well-defined verbosity levels, but the next time some software engineer is gonna need some piece of log to troubleshoot something or get their job done, I can assure you the level of logs they're gonna use is gonna be a warning, and not info debug, regardless of how important that log is... Because it's people, they game the system; if they need a log line, then they're gonna do whatever it takes to add that log line and get that job done. Now, in the long-term that means that every time you increase the minimum verbosity of the log levels, you're gonna be dropping some logs, but new logs are gonna be added, and a lot of the other logs are gonna get moved, refactored into the newer level... And that's not necessarily based on importance. It's based on what was most recently used by your engineers, which doesn't necessarily provide a good feedback on what's still important... Not to mention that some engineers might be more straight-shooters than others, in which case things can get even worse, because some engineers are gonna game the systems and others won't. + +Long story short, log levels are not gonna fix everything for you, but they can definitely help, to a certain degree. + +**Ian Lopshire:** I love that point about gaming the system, because I have definitely done that, without a doubt. + +**Liran Haimovitch:** We're recording this, you know... + +**Ian Lopshire:** Oh yeah, we'll make this error, because I need to see it. \[laughter\] I'll admit it all day. It's all good. + +**Tiago Queiroz:** \[39:58\] I totally agree with that. I think some logs are not that \[unintelligible 00:40:00.10\] because if something goes wrong, I need to know that happens, so I won't put that as debug. I know some people would call it debug; I'm calling it \[unintelligible 00:40:09.13\] because I want this log line there no matter what. And I know you're running logging, right? That's the thing. I know you're running level info, so I'm putting the level that we run, yeah. So I'm probably guilty as well, of the same thing... But I think yeah, that's definitely a thing. + +And I think it circles back to the thing - you really need to be careful, especially as you scale. I think Liran had that great point there - as you scale more and more and more, then the amount of logs become more of a thing, and then that's when you really need to think about what are you logging and how are you logging, so to make sure you always have the logs you need... Because logs are the kinds of things - you usually never look at them unless you have an issue; and then if you have an issue and then you don't have the logs, then you're basically gonna make a PR to put the logs, to deploy... 20 minutes if you're very lucky, it might take a few weeks... If you provide an application for people to download and install, and... Yeah. So definitely, you need to be very careful with the logs you write \[unintelligible 00:41:13.24\] + +**Natalie Pistunovich:** Last question and then we jump to the fun part of the unpopular opinion... When you have to debug something, how do you start? + +**Tiago Queiroz:** So I always start by trying to reproduce as close as possible the situation where the bug happened. Usually, we get "Okay, that think is failing..." Yeah, I try to reproduce it. It depends a lot on the situation, but yeah, I try to make the bug happen on my machine, if the logs that I have are not already enough to make me understand what's happening. Sometimes it is. And then job's done, just "Okay, kill your error message that your issue--" Usually, that's not that simple. So yeah, usually it takes more experiment, so I can reproduce the bug in a very easy way, so I can run it over and over until I understand what is causing it... And there is, I would say, where the art of debugging starts. Once you can reproduce it, then "Okay, why is it happening?" Then you go print statements, log statements, debugger, changing your environment, changing the state of the application to understand why it's actually the thing is happening. + +**Liran Haimovitch:** I think what I most often do is look at the code, especially if I have a descriptive bug \[unintelligible 00:42:35.18\] that kind of points me in the right direction. Quite often, if you look at the code with a fresh set of eyes, whether it's your code or somebody else's code, assuming you have a good familiarity with the application as a whole, and you're looking at the suspected piece of code where a bug might reside with a new fresh set of eyes, and with some description of what's going wrong - you'll often see the bug. It can be enough by one bug, it can be a comparison that's looking at the wrong attribute, which - I just had that bug a couple of days ago... You might have other issues, you might have a call missing, or an extra call... But if you have some experience and you know what you're looking for, you can quite often just stare at the code and say "Hey, this looks wrong", and you can all of a sudden understand what's gone wrong. + +Other than that, debugging is a very associative process. I mean, if things get very, very messy, you can start doing a more standardized process, writing down everything you know, writing various \[unintelligible 00:43:43.11\] about what could possibly go wrong, and then starting to discuss how you can collect that extra data you need to validate or remove those hypotheses, potentially by adding tons of prints. That's how it usually works. + +But early on in the process, I find that associative thinking leads to faster results in most of the cases... And that's another reason to look at the code, because once you kind of get a look of the code, get a sense of the lay of the land, then you're better positioned to look at additional data, whether it's logs, or metrics you've discussed, or the bug report itself, or anything else, to kind of see things in context, rather than just trying to piece everything together from scratch. + +**Ian Lopshire:** Great. Alright, the next part of the show I think is everyone's favorite... Unpopular Opinions. I hope you guys came with one... + +**Jingle:** \[44:36\] to \[44:54\] + +**Ian Lopshire:** Alright, who wants to go first? + +**Liran Haimovitch:** I'll go... My unpopular opinion is about OpenTelemetry. I know OpenTelemetry is being touted as the one to vanquish them all, but I find it to be over-presumptuous in many cases, way too complex, and hardly that much of a benefit over structured logging, in most cases. Obviously, there are exceptions; there are super-fancy, super-complex cases where you have 2,000 microservices, like Uber had, until they figured out it was the wrong approach, in which case you desperately need distributed tracing and OpenTelemetry, and good for you for inventing that. + +The rest of us mere mortals are better off not over-complicating our systems and not over-complicating our observability solutions. Simple is always better in engineering in general, and in software engineering in particular. Logs are awesome. They've been with us since the dawn of computer engineering. Metrics are pretty sweet as well. If you want to get extra-fancy, structured logs are totally great, and they could be very simple; they can have literally a couple of lines of configuration on any piece of the logger itself when you configure it... And on top of that, OpenTelemetry in Go is a lot of boilerplate, a lot of configuration. You have to set up all the middlewares yourself... So don't hold high hope for that; you're picking a very complex, a very expensive solution that's gonna have -- in most cases, it's gonna have less value; in some rare cases you might never meet it can have more value than more traditional approaches. + +**Ian Lopshire:** I definitely resonate with it being a lot of boilerplate and overly complicated... We made a move to move to that and abandoned it, because... It's a lot. + +**Natalie Pistunovich:** It sounds like the unpopular opinion is popular... + +**Tiago Queiroz:** Yeah, I also agree. I never worked much with OpenTelemetry, but with other solutions that basically are observability and metrics... And yeah, you can go a very long way with just well-structured logs. Even \[unintelligible 00:47:09.12\] and it's super-simple, because you can just parse them, or put them on Elasticsearch, something like that, and then you can just see the logs together, aggregate them, and... Yeah, you can go a long, long way without overcomplicating. + +So my unpopular opinion... I still believe that the testing package from the standard library is the best testing framework that you need; the best out there. It's very simple, it's very powerful, especially now that fuzzing is built into the language, and everything... And because it doesn't introduce new concepts. So basically, you still have to do your if statement. If that thing happened, then I call a method on my T object, right? Or my B, for benchmarks, and things like that. + +\[47:58\] It makes the developer, the engineer writing the code think more about what they're doing, and probably also craft better testing failure messages. So instead of just saying "A is different than B", it's like "No. I was trying to do that, and that failed because of that." You can still fall in the pitfall of just "A is not B", but I think it's less -- I think there are some libraries there with a gigantic public API for testing that just gets hard to use... And there are some very small ones here and there; I'm just seeing a few examples of -- they just misplace it, like put \[unintelligible 00:48:31.11\] the actual error that was being compared, and luckily, the test was not very wrong... But that happens. And with the testing package, that's just not gonna happen. It can only basically print the message saying "Okay, you can fail the test, and then you can write messages explaining why." + +And of course, building a language is very different than building other kinds of applications, but as far as I know in Go, to write code they only use the testing package, so it's powerful enough to write Go... So it should also fit, I'd say most, if not all use cases. And you can always write your helper functions if you need specific things; and the t.helper is also a great helper. + +**Ian Lopshire:** I'm worried that one's gonna be popular as well. + +**Natalie Pistunovich:** Yeah. I brought an unpopular opinion as well. + +**Ian Lopshire:** Oh, wow. Okay... + +**Natalie Pistunovich:** I think that as a manager, a team lead, or a product manager, or whatever, making a round of messages on your tracking system, like JIRA, Trello, and whatever - "What's the status here? What's the status here? What's the status here?", and then your developer arrive the next day to like ten messages that all look the same, "What's the status here?", is one of the worst things you can do. + +**Liran Haimovitch:** Definitely. + +**Natalie Pistunovich:** No, it's supposed to be unpopular. Don't say that. + +**Liran Haimovitch:** Elon Musk is much, much better. Tell everybody they're fired next week if they don't deliver what you came up with overnight. \[laughter\] + +**Natalie Pistunovich:** Now, that's an unpopular opinion. + +**Ian Lopshire:** No, all those status messages never fail to give me anxiety. It's like, "Wait, was I supposed to get all of this done? Ah...!" + +**Natalie Pistunovich:** Even if yes, there's better ways of doing that than ten messages of "What's the status here?" + +**Tiago Queiroz:** Yeah, definitely. + +**Natalie Pistunovich:** Maybe you have an agreement that you should update at the end of every day, every week, every other day, I don't know... If you don't have such an agreement, you should set that up. If the team is not happy with this agreement, you need to find something else, but just around like opening your inbox to ten of those messages is just like "I'm gonna close the computer. Bye." + +**Tiago Queiroz:** Yeah, I agree. + +**Liran Haimovitch:** If you expect people to spend the time to write down the answers, you should spend the time at least to write down questions that prove you know what each task is about, and be more focused about what you're asking, rather than generic "What's your status?" Because I might as well answer "It's progressing." + +**Natalie Pistunovich:** Perfect! Oh, my God. Adopting... \[laughs\] + +**Tiago Queiroz:** I would even say some of those status update messages are like debug logs that just say "Yeah, it's moving. It's happening", but most of the time it's not what you need to see, not what you need to hear. + +**Natalie Pistunovich:** True. It is like info-level messages. + +**Tiago Queiroz:** Yeah, what you need is the error level. Like, "Hey, that's delayed because of that", or like "Hey, I had that problem then block it here, so task X, Y or Z is not moving. Or even moving backwards." + +**Natalie Pistunovich:** I think we're onto something here... We also have a lot of questions that we wanted to talk about, and lots of topics we wanted to talk about, and it looks like there's no other choice but to have a follow-up episode. And I wanna say thank you both for joining; I hope you will join too our next episode on debugging. I want to wish everybody a good rest of your day, wherever you are, whatever your timezone is. Bye, everyone! diff --git a/Engineering interview tips & tricks_transcript.txt b/Engineering interview tips & tricks_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..442e7841a819e0b24fb9cd20c66824850a159b48 --- /dev/null +++ b/Engineering interview tips & tricks_transcript.txt @@ -0,0 +1,319 @@ +**Angelica Hill:** Hello, and welcome to Go time. Today we're going to be talking about engineering interviews. We're going to be talking about tips, tricks, gotchas, as well as some potential interviewing horror stories, positive stories, red flags, things to avoid, as well as hopefully some green flags - what should you do to really ensure that you're successful in your software engineering interview. + +I have the absolute pleasure, as always, of being joined by my wonderful, beautiful, intelligent, bafflingly beautiful Natalie as my co-host. + +**Angelica Hill:** Hi, hi. + +**Natalie Pistunovich:** Hello. + +**Angelica Hill:** It gives me joy every time I see you with me. + +**Natalie Pistunovich:** I'm very happy to see you every time as well, Angelica. + +**Angelica Hill:** I never get sick of it. And aside from Natalie, we have two other wonderful women; we have the lovely Emma Draper, whose pronouns are she/her, and she is an engineering manager at the New York Times, based in Arizona. Hi, Emma. How are you? + +**Emma Draper:** Hi. I'm doing very well. + +**Angelica Hill:** How are you doing? Good day, excited? First time on Go Time... + +**Emma Draper:** First time. I'm very excited to be here. Thanks for having me. + +**Angelica Hill:** Always a pleasure. And then we also have Jonas, who is a technical enablement manager at Datadog, and who is based in Denver. How are you today, Jonas? + +**Jonas:** I am great. Happy to be here with you all. Thank you, Angelica. + +**Angelica Hill:** And what is a technical enablement manager? + +**Jonas:** So primarily, I focus with customers on how to be more effective and successful with Datadog through trainings, workshops, both public stuff, custom stuff... Teaching, enabling, as the name implies. \[laughs\] + +**Angelica Hill:** We're very excited to have you on as well. Not the first time, but you haven't been on in a hot second... So great to have you back. So we're gonna dive right in. So how do you interview as a software engineer? What are the different interview stages? How do you do it? Jonas? + +**Jonas:** \[04:08\] Sure, yeah. So most interviews follow this typical format; I think every company can be a little different, so you'll want to always check with that company, but I think the flow tends to be you will have generally some sort of recruiter screen to start. And that's either if you applied, or maybe a recruiter reaches out to you... That's a really basic quick chat usually, more about the role expectations. From there, it'll depend... Some companies will do a technical screen; that may be with a person, it may be through a program, or a take-home, or something... Or you might have a hiring manager screen; you might have both of those. + +And then from there, you usually go into what's considered like a panel, or like on-site, although we're in a virtual world, so obviously, that has a different meaning... But from there, you usually have a chunk of interviews that focus on technical capabilities, as well as usually some team collaboration/communication skill set. And there's usually -- so there's like a date at each one of those. And then once you've done the whole panel set, usually that's it. But again, every place can be a little different. But in my experience, that's the general format for most engineering interviews. + +**Angelica Hill:** And are all of the interviews run by fellow engineers? Who typically does the interviews for software engineers? + +**Jonas:** Yeah, I think once you get particularly to the technical screens and the on-sites, you will generally be with engineers. Sometimes you will also - depending on the teams, you might be meeting with product managers, designers, other people that you would be working with on the team, particularly around things like assessing your teamwork and collaboration; they want to assess, like, can you work with all the other people you might be working with in your function... And then sometimes you'll also potentially meet with other leaders in the company, usually technical leaders, like a VP of engineering, or a director, or a CTO, even if it's like a smaller company... But I think the majority will be engineering, for like a more tech company. Again, if you're interviewing somewhere where maybe they don't have a lot of engineers, that could be different. But if you're at a real tech-focused company, that's the case. + +**Angelica Hill:** And then I know that you have a combination of both technical in terms of "Code this thing. Talk to me about system design", but you also have, I guess, technical in the sense of leadership, technical strategy, especially if you're interviewing for a management role etc. I'd love to hear a little bit, Emma, from you, about how you go about assessing for those kinds of qualities; more like culture, fit leadership qualities etc. + +**Emma Draper:** Yeah, I think it highly depends on the role that you're interviewing for. I think at different levels of engineering you have different impact that's measured. And how that can be measured could be what is the impact and the scope in which I'm being involved, and bringing the team along on my immediate squad, but also what does that look like for partner squads if you're in a staff or a senior engineering role... And then at a principal level, and even at some levels of staff, it really does amount to "How am I impacting the tech community at large, and what engagement am I having? How am I contributing to things like open source repositories, or something like the Go Time podcast?" Like, what is my contributions to elevating and working with those in the tech community at large? Does that answer your question? + +**Angelica Hill:** It does. It does. So I know you touched on this, Jonas, that it's semi-different at every company... I'd love to hear from your experiences what parts are fixed, and where is there room for you to curate the process for your specific team, the specific role that that person is interviewing for. + +**Jonas:** \[07:50\] As the one who's hiring, as a hiring manager, again, I hate to say "It depends", but what I have experienced is that at maybe a more enterprise, a mature company, you will probably be following a very standard system much more. And that's because at this point the company has built this system over time, they've tested it, they know it works... And so for the most part, you will follow that. There may be certain ways you can customize based on if you need a certain specialty or type of role, like if you're really focused in distributed systems, or in security, and so then you might have like a specific technical screen, or part of the technical that'll be really specific. But then outside of that, you're still within this very structured framework. + +Now, I've hired at much smaller companies, where it was just "Go hire people. Good luck." And then you're pretty much creating the whole thing yourself. So I think it can be -- definitely, that's a difference, I think, in terms of size of company and maturity of company, is how structured you're going to be as a hiring manager, versus just on your own. + +**Angelica Hill:** Okay. But when someone's thinking about interviewing as a software engineer, they can overall expect to have some system design, some kind of leadership and strategy, some kind of cross-functional collaboration... So cross-functional, we mean like with a product manager, project design, depending on the company, depending on the team... A kind of practical coding, which perhaps would be what most are familiar with, i.e. like "Code this thing, solve this problem, code review this thing while we watch and we see how you think through this problem..." Debugging, as well as sometimes, again, dependent on the company you are interviewing for, some algorithm database interview. Is that high-level what people should expect? Or is there anything else that I may have missed? If listeners are saying, "Oh, what are the different interviews I should be thinking through?" + +**Emma Draper:** I guess I would say from my perspective it matters what company are you interviewing for, what role are you aspiring to be hired at, and then what is the composition of the team currently? And that will help structure what those interviews look like. So the roles and expectations will craft what types of interviews you're going to see, the composition will indicate what the team's composition, which - you can ask that in the recruiting screen or the hiring manager screen, is "What does the team dynamic look like, and who are the members of the team?" I think it will help provide insight into who you'll be interviewing with. And then I think what the company's mission is, and what problems they're solving for will likely be key indicators of the types of questions that you're gonna see in the interview process. + +**Jonas:** Yeah, I would just echo... I think everything you touched on is right, Angelica, in that if you just know you're going to be interviewing, that's a good way to start studying and preparing.. But then very much to Emma's point - talk to the recruiter, talk to the hiring manager, because everyone will be potentially a little different, and so don't be afraid to ask, "What is it? What's your process?" and get as much info as you can, so that you can actually be really specific for that role and be really ready for it. + +**Natalie Pistunovich:** I really agree with you, and I think that also, even sometimes I read that as part of the job description, and that's a great thing to include there... So that you know how much effort you're going to have to put into that. You both agreed on what the steps are that are the same everywhere... What are your thoughts of automating one of the steps? For example, I can already see some companies using tools like LeetCode, and so on. + +**Emma Draper:** Yeah, I've had mixed feedback on the different tools. If you're considering them, definitely do research on it. I know I've used some that seemed kind of helpful, and some that have been dropped, because you just don't get very good signal from them... In a perfect world, I would like to never use them, or have to use them, because ultimately, I think you get the best signal from being in front of a person and having stuff specific to what you're trying to hire. I do understand their place... If you are really rapidly trying to scale and hire a lot of people, and you don't have a lot of existing engineers that can spend their entire time interviewing, so I can see the place for it... But ultimately, I think it's not the best tool, and I think particularly if you're trying to hire more senior and experienced people, you're never going to get the signal from that. You might even turn away candidates with stuff like that... So use it wisely, sparingly, know what it's good for and what it's not... Have you used them much, Emma? + +**Emma Draper:** \[12:07\] I haven't. I don't have a lot of exposure to -- I mean, in terms of how you can automate the hiring process, I've used a centralized ATS, which I think is really helpful, like Greenhouse. So what candidates are applying... You can then move them through, which I think is great... But in terms of how you administer interviews and using automation there, I don't have a lot of experience. And I am not opposed, but I do think that I would proceed with caution to what Jonas mentioned, for the main reason of - it's really important to the culture of the team and how that's set from a micro-culture perspective of the roles that you're hiring for that the team gets a lot of exposure. And we're talking about six interviews total, where you get 40 minutes to get a gauge of if that individual is going to be a great add to the team. So minimizing that and reducing it by any amount is like -- um, I don't know if it's worth the steps it might remove, I guess... + +**Natalie Pistunovich:** So you both agree that this is not -- like, use with caution is the thing that you would both say? And is this your answer for automating any of the steps? Or are you in particular opposed to automating the first interview, or the technical? Or is there one that makes more sense or less sense to automate? + +**Jonas:** Yeah, I've only seen these used in the very first step. And that's the only place I would. And again, it's because you're really just trying to -- if you get a ton of applicants and have a limited pool of people that can actually do that initial screen, then it can be a helpful tool just to help quick-filter basic coding ability, and that's where I think it can be effective, just to help that process. But yeah, I think to Emma's point, after that you have relatively such little time to actually assess the person and get a sense for it... And if you're starting to automate those later stages, I don't think you're gonna get to a signal; you're probably not creating a good candidate experience either... So I don't see you getting a lot of benefit on either side. + +**Emma Draper:** I also can think of it -- again, I haven't interviewed and it be automated, I haven't gone through that process, but I almost would say it's like when you call and the service is just putting you through these prompts, and you're like "I just want to speak to a person..." From an interviewee perspective, it really is nice to meet with people who have been at the company, who can speak to what the culture is like today, what challenges and areas for improvement they've identified... You don't get those from a system that's prompting you with questions, so I think I'd be like "Zero! How do I get to a person?!" \[laughs\] + +**Angelica Hill:** I've had that... I've had really terrible experiences with automated interviews. Not in the software engineering space, but in my prior life in Academia. I had a whole interview that was an hour and a half long, and it was just a robot asking me a question and then getting me to record a video of me... And they were like deep questions, they were like "What does diversity mean to you?" And they were like, "Okay, ready to record?" And then one of them would start the interview, it would auto-start the recording, and you have 30 seconds to give your answer. And if you stuttered, it was like "Next question..." So I would plus-one that. + +So we've talked through the high-level at different stages, but I'd love to hear a little bit about the difference of interview process depending on the level, i.e. if someone coming in as associate, a staff, a mid-level, a senior - what are the key differences at each level? + +**Jonas:** I've seen different ways that's approached. At some companies the interview process is still very much the same, but the rubrics, and they cater the questions a bit different for leveling. So that's one way I've seen it done. Others, there is an actual different inner flow, where you might have maybe a bit less for like an associate, because you don't want to put them through a system design maybe, because they haven't done a lot of system design probably... Whereas, maybe if you're interviewing staff and principal, you would do that; you might do an additional one, that's where you might have them actually meet with someone like a VP, because you want to really assess their strategic thinking... Whereas with the other way I've seen it is you kind of use the same format, just the questions are catered a little differently for the level, and then you have a rubric... So those are the two different ways I've seen it done. Is it similar for you, Emma, or anything else? + +**Emma Draper:** \[16:19\] No, I think you touched on the main key things that I've seen, which is it'll likely be a reduced and shortened process for junior engineers or associate-level engineers... So you'll just have, again, perhaps the same interview structure, but there will be interviews that you're not going to see, like system design, and maybe you'll lean heavier on the code review and debugging, so that there's a key indicator of what collaboration with the individual and mentorship opportunities would look like. + +And then I think the other primary thing that I'll just echo is evaluation criteria I think is really important. So how you're evaluating that individual is going to be different, and the things that you're looking for at a staff level or a principal level, going back to impact, is going to be different from what my expectations as a hiring manager are for an associate or software engineer. So those are great additives, but I wouldn't expect an associate-level engineer coming out of college to say, "Oh, I'm in all of these women in tech groups, and I host Go Time." Like, that's amazing, but that wouldn't be my expectation; like, you just went through school, right? So... + +**Angelica Hill:** So do you feel like it would be a fair statement - slash I'm looking for a reaction from you here - to say that you can give a little bit more leeway to hire someone and they can learn on the job when they're more junior, but maybe you have less of a tolerance for like knowledge gaps as we get higher up i.e. a senior software engineer? ...you might maybe not give them as much leeway to learn on the job, and more lean upon them demonstrating having worked in a system, having done something at a prior job. I would love to get just a reaction, a thought around that. + +**Emma Draper:** I look for ability to learn, and that self-starter initiative in any level. I think that that just speaks highly of what type of learner the engineers are... And that when they get on the job, they're gonna face -- even if they have tons of experience, there's gonna be a whole new problem set that they likely haven't faced before, so I definitely want to see how they deal with ambiguity at any level. And I think that it's more about, in my mind, hire staff and senior-level engineers - I think it's more about what kind of experience you're bringing to the table. So how you've dealt with systems at scale, right? That is not something likely junior-level engineers have exposure to, and so I'm looking for you to be a resource for the team to be able to lean on and learn from... So I think that that's probably the main difference. But Jonas, I'm definitely curious what your experience is. + +**Jonas:** Yeah, I would agree that -- yeah, ability to learn I think is pretty important in any role, because of the nature of tech... But yeah, really as you start to go more senior, it's where that ability to really lead and provide that expertise from experience to lead and guide a team - that becomes really important. And it's really where I think the depth of knowledge starts to become more important; like, you should have gained a lot more depth in that area, depending on the role, and that's what I would tend to be expecting more. And then add ability to really speak to different trade-offs, and still be able to admit that you don't know everything... I mean, obviously, we're always like -- even at the most seniors, like you know that, but like, you're able to really understand where your limits are even, and speak to that, but then also show a lot of depth in other areas. I think that can show a lot of maturity and that's really great. + +**Natalie Pistunovich:** Definitely. I very agree with you both also that you should probably not apply to a job where you mean 100% of requirements or you'll just be bored. + +**Jonas:** Yeah. + +**Natalie Pistunovich:** So it makes a lot of sense, what you all say... + +**Break:** \[20:09\] + +**Natalie Pistunovich:** What are your thoughts on interviewing at a startup, versus at the corporate? + +**Jonas:** I'm curious -- I have not worked at a ton of startups. I have worked for very small organizations, and so I will say, you might be interviewing with very few technical people, because they may not have a lot. I interviewed where I was the first technical person. That was a weird interview, because no one knew how to assess me... \[laughs\] + +**Natalie Pistunovich:** "How do you self-evaluate yourself? On a scale of 1 to 10..." + +**Emma Draper:** It's like grading... It's like, "How well did I do on this exam?" \[laughter\] + +**Jonas:** "I was great!" So my sense is probably startups, you may have some of that; it's probably not always gonna be a structure, because it's a small company and you might be one of the first engineers getting hired; you'll probably interview with like the CTO, or the CEO, which in an enterprise if you're interviewing with the CTO, you're either really high up, or like something's broken in their process, and that's probably a concern. But I don't know, have you had more startup experience, Emma? + +**Emma Draper:** So I worked at a small startup, and that interview process, when I was going through it I did interview with the CEO. So speaking of just when you have really small organizations - the pool of people able to conduct an interview is much smaller. And that interview was intense. I mean, I think it was 20 minutes, and it was just rapid-fire, whatever questions kind of popcorned up, I was expected to answer, and to be really -- I think that's where brief and concise answers... It's a good muscle to exercise, like, how are you able to answer the question pointfully, without rambling. So I think at a startup, that was more -- I think it's applied at any level of interviewing, but I definitely noticed it when I was on a call with the CEO, and just... What you say matters. Their time is really limited, and so you don't wanna mess it up. + +**Angelica Hill:** \[24:01\] And then as people are thinking through how much time it's going to take to interview - how long does a typical interview process take? If you're working at your current job, and you're saying, "Oh, I kind of want to find myself a new job..." what are the expectations around the length of that process before you may or may not (hopefully you will) get that job offer? + +**Emma Draper:** I think that there's a difference between how long the job search takes, and then the interview process. So I would delineate between those two things. If you think that you've hit a ceiling at your existing company or in your existing role and you want to start exploring, I would give yourself a lot of time to find that right fit, right? Because there's this really unique balance of finding people that you're inspired by, as well as the mission that you're aligned with... And that doesn't happen overnight. And oftentimes, it's applying to a lot of companies and hearing back from a really small margin and fraction of those. And then, in my experience, once the interview process has kicked off, it can take anywhere from - at the small startups it was a week, or a week and a half, two weeks; it was really accelerated... Versus at The New York Times when conversations started, and it probably took in total four months, or five months. So it's just about what that role's opening is, and what their timeline looks as a company to fill that position. + +**Jonas:** Yeah, I think that's right. It can definitely range on the type of company and what their goals are for the hiring, and their process. I mean, I think in terms of the overall time you spend interviewing, I think it's about six to eight hours of actual interviews; that feels about the norm. But yeah, there can definitely be gaps, like between where those different gates are... And because -- and especially at an enterprise New York Times or something there might be, "We need to do so many candidates before we can move to another gate." Like, some of those checks and balances that are just not even in your control. I think sometimes as a candidate you can feel like "Oh, no...!" but having then been on the other side, it's also like some of it is just there are certain processes in place, and they have to go through the wait sometimes between your final interview and an offer... There can be so much happening in the background sometimes. So that's just something to consider, especially I think at larger organizations; startups, or smaller, they can move a little faster. So yeah, I think be ready for that; sometimes it might take longer than you hope. + +**Emma Draper:** Yeah. And I would say lean on your recruiter, or the hiring team that has brought you on, and just ask. I was really lucky to have an over-communicative recruiter, that was just telling me "Nothing's gonna happen for about a month. Don't have anything for you, again." That's also fine. But I think it's nice to know where you're at, and that the company is still thinking of you and considering you, right? So I would just say you can ask. Ask questions. + +**Jonas:** Yeah. Plus one on that. Make your recruiter your friend. Talk to them all the time. + +**Natalie Pistunovich:** So in Germany, where I am based, the average developer has a three months' notice in the contract, and the average manager would have even -- well, not average manager, but many managers would have six months of a notice. So I wonder if the answer would be different if it's here... It would be interesting to run -- maybe we can make this into an unpopular opinion so it will be a poll, or something... I can imagine if this is like a four-month interview, and then another six months of waiting... It's like, you're gonna get somebody next year. So I would expect them to be faster, but my experience is mostly with startups, so it's always fast. But it is interesting... + +**Jonas:** Yeah. + +**Natalie Pistunovich:** So zooming in on that duration of the process... How long should a coding assignment take if you do it at home, versus if you do it on-site, on Zoom, or whatever the equivalent is? Please don't say eight hours. \[laughs\] + +**Jonas:** No, I was gonna say -- I mean, I think there's like what happens in reality versus my advice... I prefer to keep them roughly the same. I've definitely seen some where it's like this take-home exercise "Yeah, it's been four hours", and that's kind of absurd to me, because you're interviewing at a ton of places, and you would never ask someone to sit in an office for four hours... I mean, hopefully you wouldn't; maybe someone would. But I just don't think that's realistic. But I've interviewed and had like, "Here's a take-home. It's four hours", and I was just like "Oh, my God..." It was not a great experience. + +\[28:27\] I think take-homes can be a nice option, but you should treat it the same as you would an in-person one, and try to keep it around an hour, maybe 90 minutes, depending what it is... But I just think after that point, you're just exhausting people. I don't know if you're gonna get much better signal... Unless you actually expect your engineers to sit at their computers working on one thing for four hours... + +**Natalie Pistunovich:** When you do this specifically, do you say like "Start a timer, and 90 minutes - whatever, you're done. Even if you didn't finish, just submit what you had." Or do you say like "It should take 90 minutes, and this is what you should do", and then it ends up in people sometimes taking longer, sometimes less? Which of the two approaches do you prefer? + +**Jonas:** I feel a little torn on this honestly, I feel like I'd rather just give them the recommended time, because the idea of a timer feels stressful to me... when I've done it with a with a timer-- + +**Natalie Pistunovich:** Not a real timer, but in the sense of "Don't do over 90 minutes" is what I meant. + +**Jonas:** Yeah, yeah... I think I would rather just say "Yeah, please don't spend more than this." Because there are other tools that then do that, but they enforce it. So you start the exercise and it's forced, and that can feel stressful. I want to minimize stress as much as possible. So yeah, ideally, when I have seen it, it's more just like "Here is an exercise. Please spend no more than 90 minutes on it. It shouldn't require more than that." + +**Natalie Pistunovich:** And then the guideline is basically if it's been 90 minutes and you didn't finish, just submit what you have. You don't say "I expect this to take 90 minutes, and I expect you to do A, B, C." + +**Jonas:** No... Yeah, I was gonna say, it's usually pretty open. + +**Natalie Pistunovich:** Emma, what are your thoughts? + +**Angelica Hill:** I have a follow-up that I really need to say... If you're a software engineer and you're interviewing for a company, and they're like "Do this thing. Create this app. It should take no longer than 90 minutes", it's like, I'm gonna put my hands up and say, "I will spend all night to make it the perfect app I've ever built in my entire life, to submit..." So what I wanted to bring up was the reality of like, even if you say - and maybe this is just me - like "Hey, build this basic search engine for searching cat memes", I'm gonna build that and have the most beautiful UI, the most optimized... Like, how do you get around that? And the fact that maybe if I follow the rules, and I'm wonderful, and I only spend 90 minutes, and I provide you with what you wanted, maybe I just I didn't manage to finish it, but I did follow the guidelines, versus someone else who has submitted and they spent all night on this beautifully perfected... Like, do you take that into consideration and go, "Hm, this is really good. Angelica didn't do this in 90 minutes..." How do you mitigate that? + +**Jonas:** So when I've used these - and I'll say I've only done take-homes to do interviews a couple times - is one is really how you structure it. So I would not be like "Build an app", because to your point, there's almost no end, and you can go so far with that... Usually, it's something that's already pre-built, and then it's like three pieces that I want you to fix, or add to, or something... So it's kind of targeted; I'm not going to leave it so open-ended that it makes it really easy for you to want to keep working on it forever... Because I know what you mean; I would do the same. I think so many people -- you want to be perfect. So if it's too open-ended, then you leave it like that... So the idea is to have a very specific set of problems that people can touch in, and hopefully that -- so basically even if you do them perfectly... Like, there's nothing further to do, if that makes sense. There's no extra credit, even if you really wanted it. But I don't know if you have any other -- have you used them much, Emma, or any other approaches there? + +**Emma Draper:** \[32:12\] So I tend to lean in favor of a technical screen in-person as opposed to a take-home. If I have either been sent a take-home to fill out myself, or if I'm sending that to candidates, in my personal experience, if I see it and it's going to take -- I mean, I've received one that was just an absurd number of questions... And my reaction is, "Okay, I'm going to politely bow out", because like you said, this isn't your full-time job. Like, you have a full-time job and you're also exploring other companies... So I think being really mindful of the fact that this also should be a good hiring experience for the candidates is huge, and crucial, and I think ultimately determines the success of being able to fill the roles well and quickly. + +When I have crafted or helped craft, what those take-homes look like - it is very similar to what Jonas is saying. The first question is something like "Remove the duplicates in this array." And then the follow-up questions would be, "Can you do that using recursion?" or "Can you do that and inherit a class?" or something like that. So it builds upon itself. And it should be, in my mind, less than five questions. It shouldn't take very, very long. But I also don't believe that that -- in my experience, it hasn't been time-boxed by the recruiters... Like, "I'm going to hand you this take-home, and you need to get back to me by Tuesday afternoon." It's typically "Here's the take home. When you finish it, we'll schedule the rest of the interviews." + +**Angelica Hill:** You talked a lot about the many, many different parts that make up the full interview process, and the fact that it could be spread over many months, it could be one day of intense interviews... I've certainly experienced and done that... So how do you prepare for all this? How do you get yourself ready to do all these many different interviews? + +**Jonas:** Yeah. I mean, ultimately, practicing. And I think I even saw a study that basically said people are more successful the more they practice... Which isn't surprising, but it just shows you that, unfortunately, I think for the most part, interviewing is a skill, and you have to practice to develop it. There's a whole other conversation we could have on how to actually assess juniors maybe that's like another day. So I think for the world we're in, and what you're gonna face, you have to practice and just like set aside time every day, every week... Understand your schedule and plan around that, and know how much prep you think you're going to need to be comfortable... Because we can dig into different tips there, but the main thing is try to have a learning plan and prioritize it. + +I feel like I definitely didn't do that well earlier on, and so then you just go in like a wormhole of answering like coder challenges, and then suddenly you're like "Oh, but that actually isn't what I need to practice." You know, you can just get really stuck in one thing. So try to actually step back and be like "Okay, these are the types of companies, these are the types of roles, so I should like focus on this thing with API design", or whatever you know your weaknesses are where you need to really practice and have a plan for it... Because otherwise, you could spend a lot of time studying stuff that won't be relevant, and that really is rough. + +**Emma Draper:** If I were going to tips and tricks, I like to create a document and a list of questions that I would like to practice, to Jonas's point. A kind of learning plan of what I would like to cover in preparation for -- and again, it's more broad. It's not specific to a particular role, but the roles that I'm applying to at large... And then thinking through what attributes do I believe are going to be assessed. And then I think it's really important to practice speaking out loud against those problems. Because I think it's one thing for an engineer and for anyone to be able to solve a problem, and that's wonderful and great, but I also think it's crucial for obvious reasons that you are able to talk through in an interview, how you came to that conclusion, how you're thinking about the problem space, and the delivery of those answers needs to be understood and digestible. + +\[36:27\] So I think that in preparation for the technical portions of your interview, I think you should familiarize yourself with pretending or doing some sort of role-play, and just say "This is what collaborating meet with me would look like and mean, and so this is how I'm thinking about doing checks against "Is this what I'm supposed to be solving for?" and questions like that. Because I don't think you can do that if you're just writing down the answer to a coding problem on a piece of paper. You won't be doing that, especially in a virtual setting. + +**Angelica Hill:** And prepping those stories, prepping your examples. + +**Emma Draper:** Yes. + +**Angelica Hill:** And what is your fun, little story that demonstrates you've done this thing, whether made up or real? + +**Jonas:** \[laughs\] Oh, and just read your resume, and then remind yourself what's in there... Because sometimes you write it, and then you just send it out there... And then practice talking through those examples; because to your point, it's crazy how you just forget everything you've ever done when someone asks you about it suddenly. Right? Like you have to kind of -- like I was practicing... + +**Natalie Pistunovich:** "Tell me about yourself." + +**Jonas:** Yeah, right? I even did a practice with a friend recently, who was getting into an interview again, because it's just --especially if you haven't interviewed in a while, right? You just need to wake up that part of your brain that knows how to talk about yourself in a nice, concise, very clear and direct way... So yeah, practice that, for sure. + +**Angelica Hill:** So when you're asked to talk about like a challenge you faced, or tell us about a time you like failed, how honest should you be about that? Like, "Hey, I once took down the production DB..." True story. \[laughter\] Like, how honest should you be with those stories? Or how do you frame it as a win? + +**Jonas:** Yeah, the weird thing about those - and it's helpful because I now have interviewed but think about what the interviewer is trying to learn from it. Generally, those questions are trying to get at your ability to learn, your ability to have empathy, your ability to recognize mistakes... To me, it's not so much about how honest, it's more about how well do you craft that story, so that it comes off like "I brought down a production database", right? You can say that, because I think most people accept that engineers make mistakes... But it's more about like how you say it, right? Can you speak to the fact of like "This was the mistake I realized. Here's what we did to improve, and I got this from it." That's really more what you're trying to get at. + +Again, I think you want to be careful about treading into weird waters of, I don't know, maybe really weird interpersonal conflicts or something, because that could just start sounding a bit like "Ooh, I don't know..." For the most part, you want to just think about what the takeaway from that story is, and that's really what counts, and make sure that comes through. + +**Emma Draper:** Yeah. I agree with everything Jonas is saying. I think that as with anything, it's not that you haven't made mistakes; that would be absurd, if your interview -- like, that would be a red flag for me as an interviewee, if the interviewer that I am running through this process with is looking for perfection. I think that that's an unrealistic expectation. So I think there needs to be this mindfulness of the interviewer... The interviewer's job is not to trip candidates up when they ask a question, it's to better understand what the key takeaways and the lessons learned are, especially in a question like "What is a difficult experience that you've encountered in your career, whether that's a difficult situation with a colleague, or a system, a production-level issue that you've released, and how you dealt with it?" I think that it's "How do you deal with adversity", right? That amounts to, you know, are you the type of person that doesn't tell anyone when there's a production-level outage, and you've introduced a bug fix? Or are you the first one to say, "Something's gone awry, and I need some help. Who is my network of people that I can reach out to to help?" + +**Natalie Pistunovich:** \[40:24\] I also have a genuine follow-up question that came up to me, very similar to what Angelica was asking... My question is how would you craft your answer to "Tell me about yourself"? Because many interviews begin with this; like, two minutes where you have to put everything into context, and also tell your life story, also put the important things... What should be there, what should not be there? How long should it take? How do you do this? + +**Jonas:** Yeah, I would definitely recommend not too long. Like, keep it at a minute. I tend to avoid that question, to your point, because I think it does end up being a little too weird and open-ended... And because I have seen sometimes you ask it and someone just starts talking, and they're going for like a really -- and you know, in these interviews you only have so much time. And so sometimes he would like starts to tell -- + +**Natalie Pistunovich:** That was me at my very first job interview... \[laughter\] + +**Jonas:** Sometimes it can be interesting! + +**Natalie Pistunovich:** "I studied in high school..." + +**Jonas:** Yeah... You give like your whole life story, and it's like, it's nice, but it's also like "Oh, my God, we have 45 minutes, and I have ten other questions to ask you." So yeah, I usually think it's good to just talk a little bit about -- like, try to find a way to summarize your career journey through a couple key points... Like, "Yeah, I've spent most of my experience working in backend engineering, I've recently transitioned to management, because I'm really interested in how to empower people, and I'm looking for new roles because I'm looking for a new challenge in mentorship." Like, hit a couple key points that highlight where you've grown, where your key experience has been, and then what you're looking for... And I think that usually gives enough of an open, where the person could follow up on any of that if they want to, or they can just be like "Great!" and then move on to the next thing if they want... But they at least have a little bit of context about you. + +It's amazing how many people sometimes don't read resumes... So I do think it's a nice chance to just highlight a couple things from your resume, in case they didn't read it... Because at least you're calling that out; because otherwise it actually may not come up in your interview at all. + +**Emma Draper:** The thing I'm thinking through is - you know how you can watch a preview for a movie, right? That's how I think about that question, if you do get asked it, is "What are the highlights? What would you say about yourself that either is on your resume, and you just want to bold and underline, and say "Hey, this is really great, and this speaks to who I am, and my type of character, and how I approach leadership, or mentorship, or great things that I've done as a software engineer", and then allowing space and room for the interviewers to either probe further... But at the end of the day, there's a list of questions that they're trying to get through. "Tell me about yourself" is not the main one, so... Yeah, I would just leave that open for the interviewers. + +**Natalie Pistunovich:** Okay. And then the other side of this question, as an interviewer this time, what are some things that you're looking for in a candidate, and you want to make sure that they happened in the interview? + +**Jonas:** For something like the technical side, the coding challenges or system design, some things I'm looking for is, one, are they able to dig in and ask more questions, and get more clarification, and try to really understand the problem? But it doesn't always happen, and most people just jump in, and they end up missing big things, because they don't take the time to do that understanding... So that's usually really important to me. Because again, I don't expect you to know everything, but I do want to see how you try to understand the problem. So I'm usually looking for that. + +And then yeah, that ability to really speak to why they're doing certain things... Because that's the biggest -- I think in a technical thing, you want to make sure they understand the trade-offs of different algorithms they use, or different libraries, or approaches, and not just that they have always done it one way and memorized that one way to do it... Which happens, right? Sometimes we fall into that, and it's like "I don't know, that's just the way I've always said it." But ideally, it's someone that has a bit of experience, that can speak to trade-offs, why they would do this, why not, or where the limitations might be... + +\[44:17\] I also think though being okay to admit where -- "Oh, but this I'm not 100% sure, and so I would do this to understand it further." That's okay too, again, because I don't expect you to know everything, so it's okay... You don't need to pretend. It's also okay at a certain point to be like "This piece I don't know very well, so I would look into it", or something. + +And then I think just more on the team and communication, it's really all about just "Will you be a good teammate? Are you going to work effectively in this team? Are you going to be a good mentor, or partner, depending on that maturity? Will you be able to learn? Will you be able to add to this culture in some way, or extend, or bring us something? And will you work well with different partners or peers, depending on the role, depending on that need?" Yeah, what else? What am I missing, Emma? I feel like I'm jumping around all the interviews... What else is there? + +**Emma Draper:** I think you definitely touched on the things that I feel are most important. I think it's a red flag if a candidate should likely have said, "I'm not sure, but I would be happy finding out, and this is what my process for further exploring the solution would look like." I think interviewers can definitely tell when you're not being genuine, and I think it's crucial for a candidate to be honest and humble. + +I think it's crucial that they ask clarifying questions before beginning to explain how they're going to solution their problem. As engineers, we love to build things, but you don't want to jump in and start solutioning prior to really understanding what problems you're solving for. And I think that goes across the board. So yeah, I'd caution that candidate should start with explaining the logic behind the solution that they're thinking through, and then allow the team the opportunity to collaborate and say, "Can I validate my working assumptions with you?" + +And I would also say, refreshing yourself on what you have on your resume and not exaggerating the scope of your role... Because that's just an easy red flag for me, if I see someone who doesn't actually have the experience that they're speaking to... It's obvious. + +**Break:** \[46:25\] + +**Angelica Hill:** So in the spirit of being clear and concise, which is very important in an interview, I would love to hear your three green flags and your three red flags. Jonas. + +**Natalie Pistunovich:** Sorry, just to confirm, I would like to validate... As Emma said. So a red flag is obviously like a deal-breaker; a green flag - is it like immediate throw at them a job offer, or just like a checklist that this definitely has to happen? What do you mean with a green flag? + +**Angelica Hill:** Like something that is an indication that they could be the right person for the job. So it's something that if they say something, if they demonstrate a certain skill... It might even be like they have a friendly, personable demeanor in the interview. That could be a green flag, and it just makes you as an interviewer more likely to give them a yes. + +**Jonas:** Well, I'll just repeat myself with that, like, asking clarifying questions, digging in, showing that effort to want to understand really well... I think candidates that seem to really almost be like a step ahead, so they're a little future thinking with each of their solutions, to where it's like I almost don't even have to ask the question of like "Oh, but what would happen in this error scenario?" because they're already starting to think... You know, they're kind of like, "Alright, so this is how I do this, but then in this --" They can almost take you through, because they're thinking through in very clear steps of how the scope or complexity would expand. I think that logical ability to think really clearly and express it is great; when I see candidates do that well, I'm like "Yeah, let's do it." + +And just generally anyone that shows just a really strong empathy for others, and like someone that just wants to be part of a good team, and is okay learning from failure... You know, low ego, and just generally wants to be a strong teammate is really, really important. I don't know... That's the thing, we can always learn a lot of stuff. Like, you can teach someone Go, but can you teach them how to be a good collaborator? Yeah, but that's harder. There's a lot less textbooks out there... \[laughs\] + +**Emma Draper:** \[50:22\] Definitely. Do you want to do your red flags? Or do you want me to go on green lights, and give you a pause? + +**Jonas:** Yeah, do your green, and then let me think of my red, and then yeah... \[laughs\] + +**Emma Draper:** Good. So I think being able to pull from being concise, but being able to grab the interviewer's attention with the examples that you're giving... I really appreciate it when candidates pause before answering; it shows a thoughtfulness to understanding the question before -- again, it's that, like, jump to solution before you really understand. I think that it's totally appropriate and okay, and kind of preferred if you say, "I'm gonna take a minute to make sure, just to answer your question thoughtfully." I think that that's great. + +I really appreciate it and love when a candidate comes into the interview and has researched anything they can about the company, about the mission, about what language has probed for the recruiter to say "What language does this team program? What problems are they solving for?" Like, you can ask... I think it speaks volumes when a candidate has done that initial research and legwork to just show interest in, and initiative, when taking the interview. + +And then the last one I would say is - perhaps a little bit antiquated, but I really love it when there's like a thank you. And that could be in any way, but if I get an email from a candidate that says thank you, or I get a LinkedIn message... And I've done this myself after all interviews. I just think recognizing that people have taken time to jump on a call with you... It just shows that connection, right? Like, "Hey, I really enjoyed..." and what you perhaps enjoyed from the interview; what you took away. I think that that is not done enough, and it's like a green light. Alright, red flags... \[laughter\] + +**Angelica Hill:** Back over to you. What should you just not do? + +**Jonas:** Yeah. Red flags... I mean, hopefully I don't have to say this, but obviously, any sign of like sexism, racism, ableism, ageism, any of that - just immediately no. I'll just make sure I say that, because sadly, I've encountered it interviewing. I think, opposite of a green flag a bit, just really not asking questions or communicating, particularly in technical... It's hard for me to even understand what's happening if I don't get anything... And even if it is -- because I do this a bit; when I've had to do more like the whiteboarding ones, I'll say, "Hey, I'm gonna work through this a bit silently, so I can think, because that's how I think... And then I'm going to give you a summary afterwards." Like, even letting someone know up front if you have a different style, right? Because I do understand that people have different styles of learning and communicating... But saying that upfront can save you time and make the interviewer better. So not doing that. + +And then I think the other one that I've seen sometimes is an attitude that sort of implies the user or the customer is dumb. There's a lack of empathy there, I think, or a lack of creativity in terms of as engineers, how we can make our things actually better and easier to use... That's something I've just seen sometimes, and they're just like "Oh, those users..." And like, I get it, I definitely can understand, but I think especially in an interview setting, try to really show empathy for users, and trying to make the best product you can for them. + +**Angelica Hill:** What do they not do, Emma? + +**Emma Draper:** I think, to Jonas' point -- I mean, first I'll bold and underline, don't come into an interview -- just please don't be racist, or sexist... That's just -- + +**Angelica Hill:** Or ask the interviewer how old they are... + +**Emma Draper:** \[53:52\] I don't wanna speak to you... It's just -- I don't want to speak to you. Sorry, but there's no room for that. I think it's a red flag to me if, when pulling from your answer bank, a candidate showcases the quality of needing to be right, versus needing to be understood. I think that that speaks to their character, of how do they show up and up-level their team. I think if you are always needing to be and there are subtleties of your answer that say, "I enter into a room and my way is the right way, and that is the end of it", those candidates I just won't pass forward. + +And trying to think through what else... I think not being able to bring strong examples to the table of how you have dealt with moments in your career where you disagree, either with leadership, or with your boss, or how you challenge biases in the workplace... I think candidates need to be able to show how they do that and how they provide actionable and valuable feedback... Because I think that it comes back to - I as a manager won't do everything right, and so I definitely am looking for members of my team that hold me accountable, but that also do so in a way that's compassionate, like Jonas said. So I think it's a red flag if, again, somebody is not able to demonstrate those qualities. + +**Angelica Hill:** So as is the case with it seems every Go Time episode me and Natalie do, we're going to have to have a follow-up, because we're coming to the end of our time... But not before we do Unpopular Opinion. + +**Jingle:** \[55:33\] to \[55:50\] + +**Angelica Hill:** So, Emma, what is your unpopular opinion? + +**Emma Draper:** Do we just give one? + +**Angelica Hill:** Yes. + +**Emma Draper:** Oh, okay. + +**Angelica Hill:** First time on Go Time. So, unpopular opinion - we're gonna ask you for an opinion that you believe may or may not be unpopular... We're then going to tweet about it, and we're going to have a Twitter poll where people will vote on whether they agree or disagree with your opinion. If your opinion is too popular, you're gonna have to come back on Go Time and give a better unpopular opinion. + +**Emma Draper:** Okay, I'm gonna hope that this doesn't go viral... \[laughs\] I think that mustard is better than ketchup. + +**Angelica Hill:** Okay... + +**Emma Draper:** Do I need to...? + +**Angelica Hill:** Is there data to support that claim? + +**Natalie Pistunovich:** Emma is right. + +**Emma Draper:** I thought it was an opinion... \[laughter\] + +**Jonas:** Universally, like with any food... + +**Emma Draper:** Yup. + +**Jonas:** Like, you'd rather put your pretzel in mustard than ketchup? + +**Emma Draper:** Yup. + +**Jonas:** No a pretzel makes sense. But hot dogs. So hot dog...? Burgers...? + +**Emma Draper:** Yup. Teeny-tiny little, like, pigs in a blanket? 100%, all the time, and mustard. + +**Angelica Hill:** Do you have a preference? Is it Dijon? Or any mustard? + +**Emma Draper:** I'm not gonna be particular. I just like mustard better. + +**Angelica Hill:** Solid. Okay, we will see. + +**Emma Draper:** You wouldn't ask me what ketchup I prefer... Like, it's just ketchup. + +**Angelica Hill:** I wouldn't. I would ask what kind of ketchup. + +**Natalie Pistunovich:** The red one. + +**Angelica Hill:** The red one. \[laughter\] + +**Jonas:** I would say clearly you don't like ketchup, because this can get contentious, so whether you're like a Heinz person, or... You know, there is-- + +**Angelica Hill:** Inside the ketchup battle. \[laughter\] + +**Natalie Pistunovich:** In some countries Heinz is not a ketchup, because like percents of tomatoes... Same as Pringles. Not exactly chips and all those... + +**Jonas:** \[laughs\] True. + +**Angelica Hill:** We've unlocked Pandora's box of condiments... + +**Natalie Pistunovich:** Somebody didn't read the docs. \[laughter\] + +**Angelica Hill:** Jonas. Unpopular opinion. + +**Jonas:** Yeah... So think I want to do a bit of a tech one, because this came up recently as I was redoing my setup... But I think you don't really need dual monitors. I think one good monitor is all you need. And I don't know, I feel like people get really into dual monitors, but I just think ultimately, it creates too much distraction... Humans are actually not good at multitasking, so one monitor... A good size. Like, I get -- like, a tiny little laptop is too much... But get a good one monitor and you'll be able to do everything you need. That's probably my unpopular opinion. Let's see... + +**Emma Draper:** I think that might be popular. + +**Jonas:** I don't know, I feel like people get so into -- like, you see people's work-from-home displays, and they've got this... Well, super-widescreen I guess it's technically one, but I might count that as two. But like everyone's got dual, or like the vertical and the horizontal, and... I just think you don't need it. You only need one. + +**Emma Draper:** I feel like the hacker movies have done us a disservice, where everybody is like this... It's not how it works. \[laughter\] You just need one screen. + +**Angelica Hill:** Well, we'll see if everyone agrees with you. And on that unpopular/possibly popular opinion note, it's been a pleasure talking to you all today... We'll have to get you back on very, very soon, side note being I'm loving looking at these lovely four ladies faces on Go Time... Side note... Couldn't stop the episode without saying that... But let's say goodbye to our lovely listeners. diff --git a/Functional programming with generics_transcript.txt b/Functional programming with generics_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..df482974dce175a5ec8634e1f29bd4e53d378658 --- /dev/null +++ b/Functional programming with generics_transcript.txt @@ -0,0 +1,220 @@ +**Johnny Boursiquot:** Welcome, welcome, friends... And those who want to be friends. You want to be friends with me, right Aaron? + +**Aaron Schlesinger:** Of course. Why do you think I'm here? + +**Johnny Boursiquot:** Awesome. Awesome. So today we're gonna talk about functional programming with generics. I'm your host, Johnny Boursiquot. If you haven't heard from me for a while, that's because I've been heads down, trying to ship some stuff that's coming out later this year, and I'm pretty excited about it, but I can't spill the beans on that yet. You're just gonna have to stay tuned... And if you want, you can go to golandjohnny.com. \[laughs\] + +**Aaron Schlesinger:** Nice... + +**Johnny Boursiquot:** Since I own that, I made sure to snap that up. Last time, Jon Calhoun mentioned it on a show... So joining me today is - you know him, he's been on the show before, and you see his name everywhere, especially as it relates to things like GopherCon, and getting us ready for the big show coming up next month, actually... + +**Aaron Schlesinger:** Coming up! + +**Johnny Boursiquot:** Yeah, yeah. Please join me in welcoming Aaron Schlesinger. + +**Aaron Schlesinger:** Hi, everybody. Thanks for having me, Johnny. It's wonderful to be here again. + +**Johnny Boursiquot:** Yes, yes. And it's great to have you. So we did a show on functional programming a little while back... This was episode 87, right? This was a while; this was like May something of -- let me see, let me double-check... 2019. Back in 2019. + +**Aaron Schlesinger:** Before days... + +**Johnny Boursiquot:** So a lot has happened in the GO world since then, right? Most notably, the introduction of generics in 1.18. So it was suggested, actually... There was a listener who went back and listened to the old show, and basically said, "Hey, I've found an episode on functional programming from back in 2019. However, it was a time before generics appeared in Go. It'd be nice to hear what can be done with functional programming in Go with generics." So this was from Steve Nicholas. So Steve, shout-out to you; we indeed got this show put together on your suggestion. + +So the others - you can follow in Steve's footsteps if you want to ask for certain shows, and you can basically suggest them. We often look through the list and see what's interesting. You tell us what you want to hear, and we put on shows like this as a result of that. + +So you and I, along with Mat - we did that episode, and we learned a lot during that time; we learned how to use what we had back then, mainly things like the empty interface, and things like even generators, and things like that... And even using reflection, right? Because those were some of the things that really enabled or facilitated the creation of functional programming style libraries, and things like that, or approaches to Go. But those were also some of the gripes when trying to do functional programming in Go. Too much use of the empty interface, too much use of reflection, on and on. + +\[06:15\] And even thinking that "Well, Go is really for imperative programming", right? The verbose style - not really suitable for functional programming. But I think we made a pretty good case for it, listening back to the show... But now that we have generics, I wanted to bring you back so we can talk about what's easier now to do with Go. Like, what was hard to do then, that is easier to do now? + +**Aaron Schlesinger:** Yeah. I think maybe we can do like a little bit of a recap of what FP is... That's actually a really good introduction into -- or I should maybe say segue into why generics can help, and kind of answer your question to "Where did generics unlock new areas of functional programming awesomeness in Go?", if you will. That's a technical term, awesomeness, just for our listeners... + +**Johnny Boursiquot:** \[laughs\] + +**Aaron Schlesinger:** So I'll give it a crack to start off the recap of functional programming... Really, you can go from a math perspective and say everything in FP is based on the function, and from a theoretical perspective, a function takes inputs and returns outputs. From a math perspective, you can chart a graph from a function, for example, just to map the inputs to the outputs. That's what it is, it's a mapping. + +In real life, of course, our functions don't just spit something back out. They also do some IO, or maybe talk to a file, or use a timer, or send down a channel, or whatever else. And so that's really important, actually, to remember that all of this stuff that we're going to talk about in functional programming comes from the theoretical, but when we apply it, it looks like a design pattern. It's not like adopting a framework, or if you're in school, like doing an assignment. There's not just a right answer and a bunch of wrong answers; or a right answer, and then everything else is just wrong. There's some gradients here. And we kind of saw that with the -- there have been talks on FP before, including one that I did way back, in 2017... There's been talks at GopherCon UK, there's another talk from GopherCon in 2020... And obviously, like you said, Johnny, there's been previous Go Times, there's been blog post... They all attack FP from a different perspective. And it was all possible to do some FP things before generics. We were able to take some of the theory and apply it even before generics. Obviously, now generics has unlocked more stuff, and now we can unlock more things. + +So FP, the theory is things are based on the function. We can transform functions by putting two together. So if the output of one gets passed into the input of the second one, now you essentially have a transitive property where you can input to the first one and get an output from the second one, and it all looks like one function. Composition, right? You can curry functions, so you can think of it as partially applying a function, meaning if a function takes in three parameters, pass a lot of the parameters, and now you have a function that takes two parameters, and that first parameter that you already passed in is already baked into it. + +Then you can take those tools in Go to what I kind of think of as the next higher level. So moving into the programming world, not just the math world - because all of that stuff that I just mentioned is basically just the math world... In the programming world you can have sequences of things, like a list, or an array, or in Go a slice. So what if you could apply a function on every item in the slice, without having to write a for loop? That's called, in Go and in a lot of other languages, it's called a map, right? You just apply the function on each element in the list, and then you get a new list out the other end, that has the outputs of that function in the same indices as the respective inputs. + +\[10:26\] So now starting to bend the mind a little bit, what if you had a function that took in an element of a list, but then output a list itself? Now you can do something called a flat map, where you take in an element, so you apply the function on each element - the output is a list, right? So now you're taking a bunch of lists and combining them together to make a much bigger list. So whereas before you were just dealing with input one element, output another element, now you're inputting one element, outputting N elements for each element in the original list, and you end up with a much bigger list. + +Now, you can go even further here, and you can start doing things like filtering... So if you have a function that takes in an element in a list and returns a boolean, you can use that to decide, "Can I make a new list, and can I decide which elements from the original list are gonna end up in the new list?" + +You can also do things like zipping, which is using a function to determine the ordering of a final list given two, or three, or four, or five, or six, or ten initial lists. So you can determine how they're interspersed together. And I'm gonna stop there, but there are a lot of other ways that we can apply this very basic, but powerful concept of a pure function - just takes in an input, returns an output. Doesn't do IO, or anything else. So this concept of a pure function... And applying that into the programming world of maps, and lists, and other things, other data structures, like trees, and so forth. + +Forgive the overloading of the term list, but this list of things we could do - you know, it goes on... And I could spend another hour and a half talking about it. Of course, I'm not going to do that. But the reason that I went into all that is to kind of start to talk about what could we do before, and what can we do now, with the addition of generics. + +And really, thinking about generics - this is another branch of math, right? We're talking about types here, statically defined types. And not only that, but now we're able to vary the type of the input of a function, or the output of a function - we're able to vary that type based on a parameter. It's called type parameterization. So if we can do that, now we can write one function in Go, but we can have an infinite number of actual implementations of it, because the type can change. + +So if you've got a function that takes in a type parameter, or a parameter of type t, where t is a type parameter - well, I can invent infinity types, right? So if I can invent infinity types, and that function can take an infinity different types of parameters, which means there are infinitely different implementations of that function. And you know, the Go compiler is going to figure out how to define actual implementations that take in those actual types. I don't have to worry about that anymore, which means I don't have to generate code, or anything like that... But now that we have this power - and this is an immense amount of power - now we can start to have things like... Say, take the map example - we can start to have things like a single function called map, that can operate on a list of any type. Instead of before, we had to have a function called map, that only could operate on a list of one type, and we had to repeat the implementation of map over and over and over again, for all our different lists. Now the compiler just does that for us, like magic; thank you, Go team. You all gave us this opportunity to save tons of generated code, and that's amazing. + +\[14:11\] So that's kind of where we're at now. The implications of this go beyond map, but fundamentally, now we're at a point where instead of using raw interfaces, or you know, the any type now, instead of doing reflection, now we can add compile-time proof that certain functions will work for some definition of work, right? We can prove via the type system that certain types of functions will work. Like map, for example. We don't have to wonder whether our reflection code or our type assertion on an empty interface does the right thing and figures out the right type of the list, and then applies that to the function. Now we know that if our code compiles, that you've passed in the right function, that takes in the right type for the type of elements in the list, and you will get back another list with the output type of that function in it. + +**Johnny Boursiquot:** So if I'm hearing you correctly, it seems like now -- so we're no longer arguing about some of the benefits of FP, right? So you've done a pretty good recap of why you'd want to introduce FP, or at least know of its usefulness in certain contexts. It's about now how easy is it to actually implement in an elegant way in Go, and that's what a generic gives us, right? Because like you said, there were FP libraries, and talks, and blog posts, and everything; it was possible before. We just either had to write a lot of boilerplate ourselves, or use go generate to produce a pile of variants of a particular function to support the different types, as many types as we wanted to support... Or we had to do some voodoo reflection to get to some type information. So now it's about the elegance that we get to leverage with type parameterization. Say that three times fast... \[laughter\] + +**Aaron Schlesinger:** That's crazy, yeah... + +**Johnny Boursiquot:** Okay, so if I'm understanding correctly, then some of the libraries pre-generics - is it fair to say that you are going to see a lot of -- at least for those who practice FP and are interested in practicing FP, and doing FP in Go, which is something we'll touch on as a separate thread in our conversation... So we can expect that a lot of those FP libraries are gonna start adopting generics, because it just makes the implementation that much easier. + +**Aaron Schlesinger:** Yeah. I think, as someone who has written library to do a bunch of FP things - yeah, I saw not only did it make the implementation easier in the form of less code and more guarantees at compile-time about types, but also, it extends those guarantees to the caller, so the user of the library, too. So like I said, I no longer have to take in an interface, or any type, and return in any type from, or a list of any types, return a list of any types from map. And since now I can return the type parameter T, and then a new type parameter for the return type of the list - well, now those guarantees are extended to my user, the user of a library. And now they can also say, "Oh, well, my program compiles, therefore my types are right, too." + +So this is like a domino effect of benefits that if a user is using someone's FP library, or even rolled their own, now they know that "Well, my types are right." They don't have to remember -- it's one less thing they have to keep in their head, that "Oh, well, the input - it should be a slice of ints, and return should be a slice of strings", but all a compiler knew about in the past was just an empty interface. Or maybe a little better, a slice of empty interfaces, or something like that. It is simplicity, but it is benefit, too. + +**Johnny Boursiquot:** \[18:04\] So I was looking at this, how Go talks about this instantiation of a generic function... So for those who haven't gotten into generics yet in Go, pretty much when you create a generic function, what you're basically saying is that you're going to support a number of types based on the type-set information you provide, and basically, what happens behind the scenes when you compile your code is that Go is gonna generate an instance of your function for each of the types you said you support, right? So it's kind of like a natural code generation that we have, that provides the type safety, the type checks that the compiler is able to provide. Now the language itself is actually producing that for you behind the scenes, and you don't have to actually write your own generators. So effectively, you're doing the exact same thing you were doing before, except now it's basically baked in and supported into the language. + +So this is an efficiency game, in terms of you the programmer being able to write these libraries, being able to support multiple types of given functions, and things... This is an efficiency game really; not really a game-changer for functional programming in Go per se, just a more efficient way of writing the functions, basically. + +**Aaron Schlesinger:** Yeah, that is a really great point, is that everything we can do now, technically, you could do before, because we had the empty interface. The empty interface is the set of all types, all possible types in this context. So you could have a function that took in an empty interface, and that is a function that can take in anything you can come up with. Same thing it can return - if you have the empty interface return, then it can be a set of anything. + +So your task before was write the code to define a subset of all things that your function wants to deal with. Now, there is a massive issue again, because now you can constrain a type parameter and have the compiler compute for you whether a given type, for a type parameter that has constraints on it, is legal. So do you want it to be integral, or do you want it to be comparable? Do you want it to be stringable, or anything else under the sun that we can come up with? You write less than a line of code, and maybe you write a new constraint somewhere and then you apply it on your type parameter, and the compiler does what I consider a very advanced and very useful computation for you across your entire codebase. And if you're writing a library across, everyone's entire codebase that uses it, which is pretty amazing. + +**Johnny Boursiquot:** That is amazing. + +**Break:** \[20:41\] to \[22:39\] + +**Johnny Boursiquot:** So the Go team has been very vocal and deliberate about advising pretty much everybody to not go gangbusters which generics on all the things, right? Not use it as a hammer with everything being a nail; just don't spread it everywhere. It's a very cautious approach, until some of the best practices start to emerge, and we can see the best use cases for generics in Go. Do you think that applies even more so for those who want to practice functional programming in Go? Or do you think, "Ah, actually, it makes our lives a lot easier? Let's just go full throttle?" + +**Aaron Schlesinger:** Well, that is good advice for all software, I think. So yes, the answer to your question is yes. There is no telling right now whether map is enough. Or no one wants to use it. And if people do want to use map, there's no telling whether the parallel version of that is better, or complementary to the regular serial version of map, for example, right? So the parallel version, for listeners, would be instead of applying the function on every element of the list, one after the other, it's applying them in parallel with X number of goroutines, or at least concurrently with X number of goroutines, and then return the result. There's no telling right now... You can say this is correct, you can say this as possible, you can guess this might be useful for this case, or this case, or this type of software, or that type of software... But there's no substitute for usage in the wild, in the community, if you will. There's just no substitute. + +\[24:22\] So these things should be tried... Maybe we should go a little crazy, as they say, as the Go team has said, a little overboard, to see what is possible... But we should not say that this is the result, this is the solution to all problems in FP. We should get these things out there, play around with them, experiment with them, see what works at a larger scale, and then say "That is the solution to problems X, Y and Z." This is why we all say we test in production, right? Because there's no substitute for real -world use cases. In that case, it's data; in our case, it's programmers experimenting with things, and seeing what actually works, in real-world code bases, but it's the same effect. + +**Johnny Boursiquot:** So piggybacking off of that, the thread that we paused and now I want us to resume - you see what I did there...? \[laughs\] + +**Aaron Schlesinger:** I do, I love it. + +**Johnny Boursiquot:** The idea of -- basically, the when to use generics, and in our case here, when to use functional programming... I think that's a fair question to ask, regardless of generics. So one of the strong gripes that I found out there, that I've heard even in talking to some folks, is that, look, Go is imperative, through and through. We're kind of side-bending, and kind of pushing it in a direction it wasn't meant to, even though some of its features, some of its capabilities are well-suited for functional programming; you know, functions being first-class citizens etc, etc. A lot of the things you look for in functional programming, you can do in Go. But just because you can doesn't mean you should, right? That old adage. So is functional programming still in the realm of experimentation, and just people who are curious, they can play around with it, but really, at work or whatever, they don't really use it? Where are we in terms of -- I guess it's an adoption question, right? Like, are people interested or increasingly more so interested in using FP, now even with the bells and whistles that are enabled through generics? Is FP attractive enough for people who are traditionally doing this imperative style, especially in the world of Go? Should they be looking, or really, should we try to bring FP into our production code? + +**Aaron Schlesinger:** Well, it's already here. Anyone who uses context is using a part of FP, right? Because context has a -- with cancel and with timeout, those return a function. This is how pervasive FP is. That is technically a higher-order function, because with cancel is a function that's returning a function that you call to close, to free the resources; there's an internal goroutine running, with a timer, and it frees that resource. That is FP. + +And I think, like all technologies that are used by more than a trivial number of people in the world, a corner of this technology is used a lot, right? SQL is as another good example here. You can do many, many things with SQL. You can turn Postgres SQL, or PostgreSQL... I think that's how you say it, right? + +**Johnny Boursiquot:** Yeah, yeah... I'll allow it... \[laughs\] + +**Aaron Schlesinger:** \[27:53\] Yeah? Okay. It's got the Johnny stamp of approval... So you can turn PostgreSQL into a time-series database, if you want to, for example, right? By using very advanced features of SQL. But not many people do that, because obviously, a lot of times it's not the best tool for the job. But also, most people prefer, including myself, to stick with a smaller set of foundational features of the technology. And same thing with FP. There is a smaller set than even all the stuff that I said today, there's a smaller set of functionality that most people prefer to use, and my hypothesis for why that is is because it makes more sense, and it fits into more workloads. + +So if you need to capture some state of a function, and expose one operation on that state to the caller, using a higher-order function makes a lot of sense. It's a lot easier than doing a bunch of boilerplate, and building a whole struct, and storing the state in the struct, and having a bunch of methods on it. And that - it just fits right into that use case, and that use case is applicable and exists on a ton of workloads, and so we use it. So FP, strictly speaking - it is everywhere. + +Now, I'm going to modify your second question a little bit, accordingly... You said "Should we be trying to get FP into more workloads?" Well, given that it's already in a lot of places, I'm going to modify it and say, "Should we get more features of FP into more workloads?" Will you allow that? \[laughter\] + +**Johnny Boursiquot:** Yeah, let's do it. Yeah, tell me more. + +**Aaron Schlesinger:** Alright, so now we've got the seal on that, alright... So the way that I think about it is trying to model it as the imperative/declarative difference, I guess. There's a fundamental difference between imperative and declarative. You can see it with SQL versus Go, right? Because SQL, you say, "Dear database, this is the result that I want" and then the database has a ton of usually very clever implementation to figure out how to get whatever the way that it represents the data on disk into the result that you want. So you can filter, you can join, you can group, you can order etc. + +Now, in the FP world, let's just take this map example I keep using. So with map, the intention is to take a function, apply it onto every element of the list, the slice, and then return a new slice with the results. You might return the same slice with the new results. That's also sometimes legal as well. But let's just say, for the sake of logic, we're returning a new slice, with new results. Well, we're iterating through a list, so that's a for loop. That's just what it is, right? We can write a for loop, and that's the imperative way. It's telling the machine to go through the list one by one, call this function every time, put the result into a new list. Or, the other way is someone else did that for you, and you just call map. And that's the kind of declarative way. Because when you call map, you're effectively saying -- like, with SQL, you're saying "This is the data that I want back." With map, you're saying, "I want a new list back, with the results of the old list applied to the function that I gave to map." And that's pretty much it. + +So when we think about this question of "Should we get more features of FP into codebases that are mainly Go codebases?" the way that I think about it is "Would this codebase (or this part of a codebase) benefit in some way from more declarativity?" I think I may have made up that word, declarativity... + +**Johnny Boursiquot:** \[laughs\] It works. + +**Aaron Schlesinger:** \[31:53\] Yeah. I think people know it. We did benefit in some way. And there are a lot of possible ways that declarativity can help. It can reduce lines of code, it can add structure, it can add readability... It can also reduce readability if you're not careful, so that's a trade-off... It can fix bugs; like, if you have an error in one of your for loops - this happens a lot with parallel and concurrent code, and thus also code that deals with channel sends a lot, too. I've seen that a lot as well. So it can reduce bugs, because you're just getting someone else's code that did it the right way. + +And so if you think about it in this way - and this is where I start, with any codebase. I always start "Okay, well, is there code in here that we can reduce by using the declarative features of FP?" Whether it's maps, or reduces, or filters, or zips, and so on, and so forth. + +And as a quick anecdote to close that thought out, I have seen filter as one of the most valuable next tools to take and put into codebases, because tons and tons of code has for loops that reduce down into selectively taking things out of the list and putting them somewhere else. And that's "filter". + +**Johnny Boursiquot:** Awesome. Awesome. Yeah, I can definitely see that. I think, at least for me, it's a mindset shift. I don't have a reason not to use FP, or not to bring some of the ideas... Or I should say, not to think FP-first in terms of "How do I approach, how do I solve this problem?" I'm so used to the way I'm used to doing it... It's more of maybe education, and maybe seeing more people talk about it, present about it, and having shows like this, where people knowledgeable like yourself come on and advocate for it... I think it's more of an education issue, more than the merits of it... Because all these things you're talking about here - these are things that make my program safer, they make my programs easier to understand... And obviously, as with everything in programming, we sprinkle things where they make sense... And if your team at work is not using FP style for things, and perhaps they want to or don't want to, so "Don't force it" kind of thing... But I think every once in a while, even using your example like filter, for example - I see the use case for it, now that you mention it, I see that everywhere. It could definitely be used in a lot more places. And who knows, maybe that triggers somebody to say "Hey, what's that? I've never seen this particular approach" And then boom, that's a brown bag lunch right there; you can educate some co-workers around the merits of FP. + +So for me, I think it's an education thing more so than capabilities or features... Which kind of leads me to my next question. Is there anything missing in Go to nudge us even more towards the -- I don't want to say traditional FP, because I don't think we want to make go an FP language per se... But is there missing features in Go that would allow us to take even more advantage of FP concepts? + +**Aaron Schlesinger:** Well, generics are young, and FP using generics is even younger. So really, the best I can answer that is hypothesize based on very, very unscientific things like gut feel, and things I've seen, and anecdotal evidence. I think the biggest thing that could unlock just building these things a little easier would be type parameters on methods of a struct or a type. You can't do that now. So you can't have -- let's say, you have a custom list type that's a list of type t; you can't have a method called map on that list that takes a type parameter u, and then returns a new list of type u. + +\[36:05\] You can still build map; I've done it, and I know others who have done it, I've seen it all over the place. You can still build it, it has the same effect, and it works the same. The reason that I'm hypothesizing that type parameters on methods would make things easier is because for many folks who are trained using Java or C++, or other what I'll call pure object-oriented languages, although whether they're pure or not is another podcast... + +**Johnny Boursiquot:** "Pure" in air quotes, for those not watching. + +**Aaron Schlesinger:** Air quotes, yes. Thank you for that call-out. Definitely in air quotes. So folks who are trained with the classical OOP type of programming, OOP design patterns and methodology, having a method on a struct feels a lot like a method on a class in Java. It's not strictly the same, but for folks who are used to thinking of methods, if you have a method - again, air quotes; if you have a "method" on your list struct in Go, well, it makes more sense to do it that way, rather than to do it the more kind of pure, again, air quotes... "Pure" FP way, which would be a totally separate function that takes in a certain type, and returns a new type. Again, for some of the things that I've talked about today, that's not going to limit or reduce or eliminate correctness at all... But there are some more advanced things that could be unlocked by having something called higher-kinded types in Go. Maybe this is my unpopular opinion right now, is just to say higher-kinded types make -- and I was just talking to my manager about this two hours ago... Higher-kinded types add a ton of complexity to a compiler and a type checker inside of the compiler. But they do offer basically something called a type constructor, which is a type that depends on another type, right? And if you can do that, I'll leave it as an exercise to the reader, or listener or watcher, what kinds of things could you do with that. It'd take me a while to get into it, so I'm not going to go there now... But if you had that, now, assuming it was correct in the compiler, you can start to do some very advanced things. But the only thing that I'll say is, whenever I've seen that feature in a language, or not in the language, and then introduced into the language, at that point it starts to become what you call a functional language... Because getting to that level of sophistication in a type system often means that there are a bunch of people already there who have been demanding it, and those people are the majority, and those people write very heavily function-oriented code. + +So really, I'm gonna just limit it to the type parameters on methods thing, because that thing is not very controversial, but higher-kinded types or higher order types are a very interesting and cool thing that I really do wish was in Go, because it would enable some very slick FP things. + +**Johnny Boursiquot:** Yeah, that doesn't sound controversial to me. I can see that. + +**Aaron Schlesinger:** Some might. Maybe some people will. + +**Johnny Boursiquot:** \[laughs\] + +**Break:** \[39:35\] to \[41:12\] + +**Johnny Boursiquot:** Okay, so before we transition into unpopular opinions, maybe I want one of those controversial ideas... \[laughs\] + +**Aaron Schlesinger:** Okay... + +**Johnny Boursiquot:** Alright, let's get it going. + +**Aaron Schlesinger:** I'm not gonna give one that talks about higher-kinded types, because it's been a while, first of all, so I don't think I would do it justice, or maybe even get some parts incorrect... Also, it would take me a while. But I want to mention one pattern from FP that I think is really great, and it's not that hard to do yourself in your code. I also don't think it's too hard to imagine how it would work in a codebase; it's called the lens. It maybe sounds like an intimidating thing to some people. It certainly did to me. The lens is essentially a tupple of two functions. One function is a getter for some data, and then the other function is the setter for some probably the same type of data. But you can have separate things, too. You can have a getter for one piece of data, and a setter for another piece of data in the same tupple. So usually, it's the getter first in the tupple, and the setter second in the tupple. + +Now, the first function takes no parameters and returns the result, the data; it might take one parameter if the data is multi-dimensional, like a map, and you want it to specify a key... And that's up to you; that's part of the design pattern. And then the setter potentially takes in the dimension of the data, again, like the key of the map... And then it also takes in the data that you want to set into that place. So it might be the value of a map, if we're going to use the map, like the example of a map, right? And then it returns nothing. These things can also alternatively return an error if this is doing something over the network or something that might fail. Traditionally, they're not. +So if you could kind of imagine how these functions would be created, they would probably be closures. They would close over your data, they would take some parameter... Let's take the get case, for example - it would take the parameter for the thing to get, and then they would return the value of it. Same thing with the set - it would close over the data and then do its thing. + +It's a fancy term called a lens, but it's a fairly basic, foundational thing. And that is a common theme in FP, actually. So I call this the lens, and it is in Wikipedia, and it's in the textbook, and everything... But it just boils down to functions, like most things in FP do. They just boil down to one or more functions, maybe combined together in some creative way, or maybe one function applied across a data structure in some other way. + +\[43:59\] So I would challenge everyone listening to just give it some thought. Can that reduce code somewhere in your codebase? If you have a bunch of getters, can you get rid of those getters and just replace them with a closure, the lens? That's a very interesting one to me, because it is exceptionally simple... I'll put it a different way - the return on investment is very high, right? It's very simple, but it tends to be applicable and useful in a lot of places in the codebase. + +**Johnny Boursiquot:** Very cool. I'd love to see that in action, like in practical use. The way you explain it, it sounds great. That sounds like something I'd be up for. + +**Aaron Schlesinger:** There you go, yeah. + +**Johnny Boursiquot:** Perhaps we need a blog post from you when you're done with school, or something... \[laughs\] If you don't mind me asking, what did you go back to school for? + +**Aaron Schlesinger:** Well, I went back to school for computer science. I got an undergraduate Computer Science degree in 2008. I'm not that old, but I feel older than I am, I guess... You can do the math at home, see and figure out how old I'm likely to be... And so then I've been in industry ever since, focused on industry things... And they all tend to be very practical, they all tend to be single-goal-oriented... And all of the learning that I've done has been very focused on a goal, rather than broader. So I went back to get a masters, to get that breadth again... Because I've kind of missed that. + +I kept using the SQL and the database examples because right now I'm taking a course in databases, so that -- + +**Johnny Boursiquot:** \[laughs\] There you go. + +**Aaron Schlesinger:** ...easy extension there. But the thing that I'm planning to specialize in now is formal methods, because I love this stuff... Programming language theory, and things like that - I really do enjoy it a lot. And the formal methods specialization happens to have quite a bit of language theory stuff in it, so I look forward to that. + +**Johnny Boursiquot:** Yeah, it sounds like you're in your bread and butter kind of space right now... That's nice. Nice, nice. + +**Aaron Schlesinger:** Yeah. + +**Johnny Boursiquot:** Now that you mention it, interesting enough -- I think we tend to go to school to get a specialization and to narrow down on a very particular set of skills... Even though when we go in, we might not notice what we're doing, but we're going to learn how to do X, and then we come out and we get a job to do X. It's very rarely do you go back just for the sake of breadth, right? Just to basically go find out what you don't know about it, and then you sort of see how you can apply it to your already existing set of knowledge that you acquired in industry. So yeah, I commend you for that. It's very cool. + +**Aaron Schlesinger:** Well, I'm very, very lucky and privileged to be able to do that at this point. Because I think probably all of us remember in undergrad we probably just went so we could get to whatever our next step was. Most of us it probably was a job; maybe some of us was grad school, but most of us was a job. It certainly was me; I was laser focused on getting done, so I could get paid. But these days is different, right? I have a job, and I'm able to continue working in this job while I do this part-time program... So I fully acknowledge and I try to remind myself daily that I'm very lucky to be able to do something like this, just because I want to, rather than having to do something. + +**Johnny Boursiquot:** Pretty cool, pretty cool. Well, Aaron, unless you have another more unpopular opinion for me -- although I don't think anything you said here is unpopular. Maybe it's just the way you talk and put it, but I don't think any of it was unpopular... + +**Aaron Schlesinger:** Diplomacy, yeah... Well, I do have an unpopular opinion. + +**Johnny Boursiquot:** Okay... + +**Aaron Schlesinger:** This one will be unpopular. + +**Johnny Boursiquot:** Okay, wait, then I've gotta play the song if that's the case. + +**Aaron Schlesinger:** Okay, let's do it. + +**Jingle:** \[47:55\] to \[48:11\] + +**Johnny Boursiquot:** Go on, lay it on us. + +**Aaron Schlesinger:** Alright. Well, I've been doing a bunch of Rust lately, some very low-level stuff dealing with virtual machines... And I believe now that I've used Rust long enough... I believe the type system in Rust is more complete, and leads to more concise and easier to write and read programs in Go. + +**Johnny Boursiquot:** Okay then... + +**Aaron Schlesinger:** Yeah, so the Go type system has some work to do. + +**Johnny Boursiquot:** Okay, okay... + +**Aaron Schlesinger:** I don't mean to fan the flames of the Go/Rust thing, but I'm sure I just have. + +**Johnny Boursiquot:** \[laughs\] Too late, buddy... + +**Aaron Schlesinger:** I've just done that... The flame war... + +**Johnny Boursiquot:** It's been zero days now since the flame wars... \[laughter\] + +**Aaron Schlesinger:** I know. I know... I talked a lot about type system things today, and it's not that Rust lets you do all those things. It certainly doesn't. But the type system in Rust has more features than the type system in Go. It certainly allows for some programs that take forever to compile and are hard to read, so that's the pro and con thing here... But from my experiences so far, it lets you write more expressive code, with fewer lines than the Go equivalent. + +**Johnny Boursiquot:** Okay. Well, we shall put that to the test... The audience - they will tell us whether they agree with you or disagree with you, and come find you in the night, or something... + +**Aaron Schlesinger:** Yeah, you all know my Twitter, so you can send me hatemail... I still love Go, don't get me wrong. I've been writing Go for something like ten years, or something... And there's no language I can build something more quickly in... And I'm just getting started with Rust, so this my initial opinion, though... + +**Johnny Boursiquot:** Hey, you can love more than one language, that's cool. As long as it's Go and Go, that's cool. \[laughter\] + +**Aaron Schlesinger:** Love anything you want, as long as it's Go. \[laughter\] I love that. + +**Johnny Boursiquot:** Awesome, awesome. So thank you so much, Aaron, for coming back on the show and talking about functional programming. Hopefully, we hit a few nails on the head for folks that have been curious about functional programming, and how it's done, how it's been done, and what generics enable for us moving forward. This was a fun discussion. Thanks again, Aaron, for coming on the show. + +**Aaron Schlesinger:** Thank you so much for having me, Johnny. diff --git "a/Functional programming with generics\357\274\237_transcript.txt" "b/Functional programming with generics\357\274\237_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..16c42d59ab610d30b4e0a5615dd819696e3a92e0 --- /dev/null +++ "b/Functional programming with generics\357\274\237_transcript.txt" @@ -0,0 +1,734 @@ +[0.00 --> 15.26] As someone who has written a library to do a bunch of FP things, not only did it make the implementation easier in the form of less code and more guarantees at compile time about types, it extends those guarantees to the caller. +[15.26 --> 36.52] So I no longer have to take in an interface or a list of any types and return a list of any types from map. And since now I can return the type parameter T and then a new type parameter for the return type of the list, well now those guarantees are extended to the user of the library. +[36.52 --> 54.36] And now they can also say, oh, well, my program compiles, therefore my types are right too. So this is like a domino effect of benefits that if a user is using someone's FP library or even rolled their own, now they know that, well, my types are right. +[54.36 --> 66.40] This episode is brought to you by Sourcegraph. Sourcegraph is universal code search to let you move fast, even in big code bases. +[66.90 --> 73.66] Here's CTO and co-founder, Byung Lu, explaining how Sourcegraph helps you to get into that ideal state of flow in coding. +[73.66 --> 89.14] The ideal state of software development is really being in that state of flow. It's that state where all the relevant context and information that you need to build whatever feature or bug that you're focused on building or fixing at the moment, that's all readily available. +[89.30 --> 94.68] Now the question is, how do you get into that state where, you know, you don't know anything about the code necessarily that you're going to modify? +[94.68 --> 104.14] That's where Sourcegraph comes in. And so what you do with Sourcegraph is you jump into Sourcegraph. It provides a single portal into that universe of code. +[104.30 --> 110.98] You search for the string literal, the pattern, whatever it is you're looking for. You dive right into the specific part of code that you want to understand. +[111.26 --> 124.10] And then you have all these code navigation capabilities, jump to definition, find references that work across repository boundaries that work without having to clone the code to your local machine and set up and mess around with editor config and all that. +[124.10 --> 129.56] Everything is just designed to be seamless and to aid in that task of, you know, code spelunking or source diving. +[129.88 --> 137.74] And once you've acquired that understanding, then you can hop back in your editor, dive right back into that flow state of, hey, all the information I need is readily accessible. +[137.96 --> 142.44] Let me just focus on writing the code that influenced the feature or fixes the bug that I'm working on. +[142.44 --> 152.44] All right, learn more at Sourcegraph.com and also check out their bi-monthly virtual series called DevToolTime covering all things DevTools at Sourcegraph.com slash DevToolTime. +[154.10 --> 169.42] Let's do it. +[169.98 --> 171.04] It's Go Time. +[171.62 --> 176.42] Welcome to Go Time, your source for diverse discussions from all around the Go community. +[176.90 --> 179.54] Check out our back catalog at GoTime.fm. +[179.54 --> 186.74] There you'll find the most popular episodes, our favorites, and a request form so you can let us know what you want to hear about on the pod. +[187.10 --> 191.88] Special thanks to our partners at Fastly for shipping our shows super fast to wherever you listen. +[192.12 --> 193.62] Check them out at Fastly.com. +[194.02 --> 195.70] And to our friends at Fly.io. +[196.06 --> 198.04] Host your app servers close to your users. +[198.30 --> 199.16] No ops required. +[199.50 --> 201.26] Learn more at Fly.io. +[201.52 --> 202.28] Okay, here we go. +[203.56 --> 206.22] Welcome, welcome, friends. +[206.22 --> 209.16] And those who want to be friends. +[209.78 --> 211.02] You want to be friends with me, right, Aaron? +[211.46 --> 211.90] Of course. +[212.20 --> 213.10] Why do you think I'm here? +[213.60 --> 214.48] Awesome, awesome. +[215.08 --> 220.44] So today we're going to talk about functional programming with generics. +[220.90 --> 222.88] So I am your host, Johnny Borsico. +[223.60 --> 230.42] If you haven't heard from me for a while, that's because I've been heads down trying to ship some stuff that's coming out later this year. +[231.10 --> 233.66] And I'm pretty excited about it, but I can't spill the beans on that yet. +[233.66 --> 235.18] You're just going to have to stay tuned. +[235.76 --> 238.54] And if you want, you can go to GolangJohnny.com. +[240.38 --> 247.92] Since I own that, I made sure to snap that up the last time John Calhoun mentioned it on our show. +[248.38 --> 252.12] So joining me today is, you know him, he's been on the show before. +[252.52 --> 253.92] And you see his name everywhere. +[254.12 --> 259.60] And especially as it relates to things like GopherCon and getting us ready for the big show coming up next month, actually. +[259.82 --> 260.26] Coming up. +[260.26 --> 260.98] Yeah, yeah. +[261.10 --> 264.14] Please join me in welcoming Aaron Schlesinger. +[264.70 --> 265.30] Hi, everybody. +[265.56 --> 266.56] Thanks for having me, Johnny. +[266.76 --> 268.08] It's wonderful to be here again. +[268.46 --> 269.06] Yes, yes. +[269.08 --> 270.40] And it's great to have you. +[270.64 --> 276.06] So we did a show on functional programming a little while back. +[276.34 --> 278.72] This was episode 87, right? +[279.24 --> 279.90] This was a while back. +[279.94 --> 284.82] This was like May something of, let me see, let me double check. +[285.40 --> 287.14] 2019, back in 2019. +[287.54 --> 288.34] Four days. +[288.34 --> 291.66] So a lot's happened in the Go world since then, right? +[291.98 --> 295.80] Most notably, the introduction of generics in 118, right? +[296.42 --> 307.32] So it was suggested, actually, that there was a listener who basically went back and listened to the old show and basically said, hey, I found an episode on functional programming from back in 2019. +[307.60 --> 310.12] However, it was a time before generics appeared in Go. +[310.48 --> 314.88] It'd be nice to hear what can be done with functional programming and Go with generics. +[314.88 --> 316.20] So this one's from Steve Nicholson. +[316.26 --> 317.36] So, Steve, shout out to you. +[317.72 --> 320.60] We indeed got this show put together on your suggestion. +[320.82 --> 327.46] So the others you can follow in Steve's footsteps if you want to ask for certain shows. +[327.64 --> 328.92] And you can basically suggest them. +[329.24 --> 331.72] And we often look through the list and see what's interesting. +[331.72 --> 333.28] And you tell us what you want to hear. +[333.46 --> 335.70] And we put on shows like this as a result of that. +[335.98 --> 340.60] So you and I, along with Matt, were basically, we did that episode. +[340.94 --> 342.72] And we learned a lot during that time. +[342.82 --> 349.88] We learned how to use what we had back then, mainly things like the empty interface and things like even generators, right? +[349.88 --> 352.86] And things like that and even sort of using reflection, right? +[353.06 --> 363.84] Because those were some of the things that really sort of enabled or facilitated the creation of functional programming and sort of style libraries and things like that or approaches to Go. +[364.22 --> 369.38] But those were also some of the gripes, right, with trying to do functional programming in Go, right? +[369.48 --> 371.02] Too much use of the empty interface. +[371.56 --> 375.04] Too much use of, you know, sort of reflection on and on, right? +[375.04 --> 380.66] And even the thinking that, well, Go is really for imperative programming, right? +[380.78 --> 382.04] The verbose style, right? +[382.06 --> 383.72] Not really suitable for functional programming. +[383.84 --> 386.36] But I think we made a pretty good case for it, right? +[386.44 --> 387.84] And listening back to the show. +[388.14 --> 395.58] But now that we have generics, right, I wanted to bring you back so we can talk about what's easier now to do with Go. +[395.68 --> 398.34] Like what was hard to do then that is easier to do now? +[398.72 --> 398.92] Yeah. +[398.92 --> 404.56] I think maybe we can do like a little bit of a recap of what FP is. +[404.56 --> 415.80] That's actually a really good introduction into, or I should maybe say segue into why generics can help and kind of answer your question too. +[416.30 --> 423.24] Where did generics unlock new areas of functional programming awesomeness in Go, if you will? +[424.24 --> 427.48] That's a technical term, awesomeness, just for all the listeners. +[428.22 --> 433.54] So I'll give it a crack to start off, you know, the recap of functional programming. +[433.54 --> 440.46] Really, you can go from a math perspective and say, you know, everything in FP is based on the function. +[441.06 --> 446.32] And from a theoretical perspective, a function takes inputs and returns outputs. +[446.58 --> 454.28] From a math perspective, right, you can chart a graph from a function, for example, just to map the inputs to the outputs. +[454.62 --> 455.22] That's what it is. +[455.24 --> 455.76] It's a mapping. +[455.76 --> 460.14] In real life, of course, our functions don't just spit something back out. +[460.28 --> 467.58] They also do some I.O. or maybe talk to a file or use a timer or send on a channel or whatever else. +[468.16 --> 477.84] And so that's really important, actually, to remember that all of this stuff that we're going to talk about in functional programming comes from the theoretical. +[477.84 --> 482.28] But when we apply it, it looks like a design pattern. +[482.52 --> 487.58] It's not like adopting a framework or if you're in school, like doing an assignment. +[487.76 --> 494.10] There's not just a right answer and a bunch of wrong answers or a right answer that everything else is just wrong. +[494.70 --> 496.18] There are some gradients here. +[496.18 --> 504.12] And we kind of saw that with the there have been talks on FP before, including one that I did way back in 2017. +[505.02 --> 507.40] There's been talks at GopherCon UK. +[507.72 --> 510.58] There's another talk from GopherCon in 2020. +[511.38 --> 515.40] And obviously, like you said, Johnny, there's there's been previous go times. +[515.44 --> 516.52] There's been blog posts. +[516.88 --> 521.90] They all attack FP from a different perspective. +[521.90 --> 522.58] Right. +[523.16 --> 527.98] And it was all possible to do some FP things before generics. +[528.32 --> 528.46] Right. +[528.52 --> 532.24] So we were able to take some of the theory and apply it even before generics. +[532.62 --> 537.16] Obviously, now generics has unlocked more stuff and now we can unlock more things. +[538.14 --> 541.76] So FP, the theory is things are based on the function. +[541.96 --> 545.30] We can transform functions by putting two together. +[545.58 --> 545.74] Right. +[545.74 --> 556.16] So if the output of one gets passed into the input of a second one, now you essentially have a transitive property where you can input to the first one and get an output from the second one. +[556.16 --> 558.10] And it all looks like one function. +[558.74 --> 559.26] Composition. +[559.56 --> 559.66] Right. +[559.98 --> 561.28] You can curry functions. +[561.28 --> 570.76] So you can think of it as partially applying a function, meaning if a function takes in three parameters, pass one of the parameters. +[570.76 --> 573.82] And now you have a function that takes two parameters. +[574.14 --> 578.28] And that first parameter that you already passed in is already baked into it. +[578.94 --> 579.08] Right. +[579.68 --> 585.84] Then you can take those tools and go to what I kind of think of as the next higher level. +[585.84 --> 586.64] Right. +[586.72 --> 594.42] So moving into the programming world, not just the math world, because all that stuff I just mentioned is basically just the math world. +[594.70 --> 601.08] In the programming world, you can have sequences of things like a list or an array or in go a slice. +[601.78 --> 609.02] So what if you could apply a function on every item in the slice without having to write a for loop? +[609.66 --> 612.50] That's called in go and a lot of other languages. +[612.74 --> 613.48] It's called a map. +[613.88 --> 613.96] Right. +[613.96 --> 618.10] You just apply the function on each element in the list. +[618.16 --> 626.06] And then you get a new list out the other end that has the outputs of that function in the same indices as the respective inputs. +[626.46 --> 630.14] So now starting to bend the mind a little bit. +[630.70 --> 635.98] What if you had a function that took in an element of a list, but then output a list itself? +[636.80 --> 636.96] Right. +[637.02 --> 641.30] Now you can do something called a flat map where you take in an element. +[641.30 --> 644.02] So you apply the function on each element. +[644.46 --> 645.60] The output is a list. +[645.60 --> 645.88] Right. +[645.88 --> 651.22] So now you're taking a bunch of lists and combining them together to make a much bigger list. +[651.94 --> 652.06] Right. +[652.10 --> 656.14] So whereas before you were just dealing with input one element, output another element. +[656.52 --> 663.48] Now you're inputting one element, outputting n elements for each element in the original list, and you end up with a much bigger list. +[663.48 --> 669.34] Now you can go even further here and you can start doing things like filtering. +[669.34 --> 677.10] So if you have a function that takes in an element in the list and returns a Boolean, you can use that to decide, can I make a new list? +[677.24 --> 681.46] And can I decide which elements from the original list are going to end up in the new list? +[681.82 --> 693.66] You can also do things like zipping, which is using a function to determine the ordering of a final list given two or three or four or five or six or ten initial lists. +[693.66 --> 694.26] Right. +[694.30 --> 696.60] So you can determine how they're interspersed together. +[697.48 --> 706.84] And I'm going to stop there, but there are a lot of other ways that we can sort of apply this very basic but powerful concept of a pure function. +[707.16 --> 709.04] Just takes in an input and returns an output. +[709.42 --> 711.26] Doesn't do IO or anything else. +[711.74 --> 722.46] So this concept of a pure function and applying that into the programming world of maps and lists and other things, other data structures like trees and so forth. +[722.46 --> 728.98] So forgive the overloading of the term list, but this list of things we can do, you know, it goes on. +[729.22 --> 729.34] Right. +[729.40 --> 731.80] And I could spend another hour and a half talking about it. +[731.84 --> 733.42] Of course, I'm not going to do that. +[733.82 --> 744.30] But the reason that I went into all that is to kind of start to talk about what could we do before and what can we do now with generics, with the addition of generics. +[744.60 --> 748.90] And really, thinking about generics, this is another branch of math, right? +[748.90 --> 752.10] We're talking about types here, statically defined types. +[752.30 --> 761.24] And not only that, but now we're able to vary the type of the input of a function or the output of a function. +[761.68 --> 764.58] We're able to vary that type based on a parameter. +[765.10 --> 766.52] It's called type parameterization. +[766.52 --> 777.60] So if we can do that, now we can write one function in Go, but we can have an infinite number of actual implementations of it because the type can change. +[777.60 --> 784.26] So if you've got a function that takes in a type parameter or a parameter of type T, where T is a type parameter. +[784.96 --> 788.94] Well, I can invent infinity types, right? +[789.38 --> 799.66] So if I can invent infinity types, then that function can take in infinity different types of parameters, which means there are infinity different implementations of that function. +[800.36 --> 806.10] And, you know, the Go compiler is going to figure out how to define actual implementations that take in those actual types. +[806.10 --> 810.82] I don't have to worry about that anymore, which means I don't have to generate code or anything like that. +[811.24 --> 815.06] But now that we have this power and this is immense amount of power. +[815.88 --> 821.62] Now we can start to have things like, say, take the map example. +[822.24 --> 829.78] We can start to have things like a single function called map that can operate on a list of any type. +[829.78 --> 836.62] Instead of before, we had to have a function called map that only could operate on a list of one type. +[837.08 --> 842.48] And we had to repeat the implementation of map over and over and over again for all our different lists. +[843.02 --> 845.50] Now the compiler just does that for us like magic. +[845.78 --> 846.66] Thank you, Go team. +[846.88 --> 850.82] You all gave us this opportunity to save tons of generated code. +[850.92 --> 851.58] And that's amazing. +[852.24 --> 853.84] So that's kind of where we're at now. +[853.84 --> 856.16] The implications of this go beyond map. +[856.86 --> 874.78] But fundamentally, now we're at a point where instead of using interface, raw interfaces, or, you know, the any type now, instead of doing reflection, now we can add compile time proof that certain functions will work for some definition of work. +[875.10 --> 875.20] Right. +[875.62 --> 880.20] We can prove via the type system that certain types of functions will work. +[880.70 --> 881.98] Like map, for example. +[881.98 --> 893.88] We don't have to wonder whether our reflection code or our type assertion on an empty interface does the right thing and figures out the right type of the list and then applies that to the function. +[894.28 --> 908.12] Now we know that if our code compiles, that you've passed in the right function that takes in the right type for the type of elements in the list and you will get back another list with the output type of that function. +[908.12 --> 916.44] So if I'm hearing you correctly, it seems like now, so we're no longer arguing about the benefits of FP, right? +[916.52 --> 923.98] So we've done a pretty good recap of why you'd want to introduce FP or at least know of its usefulness, right? +[924.10 --> 925.32] In certain contexts. +[925.76 --> 931.04] It's about now how easy is it to actually implement in an elegant way and go. +[931.04 --> 932.70] And that's what generic gives us, right? +[932.74 --> 937.72] Because like you said, there were FP libraries and talks and blog posts and everything. +[937.94 --> 939.38] It was possible before. +[940.00 --> 951.38] We just either had to write a lot of boilerplate ourselves or use, you know, go generate to produce, you know, a pile, a variance of a particular function to support the different types as many types as we wanted to support. +[951.88 --> 952.12] Right. +[952.12 --> 957.12] Or we had to do, you know, some voodoo with reflection to get to some type information. +[957.30 --> 957.42] Right. +[957.74 --> 965.16] So now it's about sort of the elegance, right, that we get to leverage, right, with type parameterization. +[965.46 --> 966.66] Let's say that three times last. +[967.26 --> 967.50] Yeah. +[968.90 --> 969.30] Okay. +[969.40 --> 981.62] So if I'm understanding correctly, then some of the libraries pre-generics, is it fair to say that you were going to see a lot of, at least for those who practice FP and are interested in practicing FP and doing FP in Go, +[981.62 --> 985.58] which is something we'll touch on as a separate sort of a thread, right, in our conversation. +[986.18 --> 993.18] So we can expect that a lot of those FP libraries are going to start adopting generics because it just makes the implementation that much easier. +[993.60 --> 993.88] Yeah. +[993.88 --> 1009.22] I think as someone who has written a library to do a bunch of FP things, yeah, I saw not only did it make the implementation easier in the form of less code and more guarantees at compile time about types, +[1009.22 --> 1015.86] but also it extends those guarantees to the caller, so the user of the library too. +[1016.46 --> 1028.16] So, you know, like I said, I no longer have to take in an interface or any type and return an any type from, or a list of any types and return a list of any types from map. +[1028.16 --> 1042.82] And since now I can return the type parameter T and then a new type parameter for the return type of the list, well, now those guarantees are extended to my user, the user of the library. +[1043.42 --> 1048.82] And now they can also say, oh, well, my program compiles, therefore my types are right too. +[1048.82 --> 1061.84] So this is like a domino effect of benefits that if a user is using someone's FP library or even rolled their own, now they know that, well, my types are right. +[1062.18 --> 1065.84] They don't have to remember it's one less thing they have to keep in their head, right? +[1065.98 --> 1072.16] That, oh, well, the input should be a slice of ints and return should be a slice of strings. +[1072.16 --> 1081.14] But all a compiler knew about in the past was it's just an empty interface or maybe a little better, a slice of empty interfaces or something like that. +[1081.56 --> 1083.70] It is simplicity, but it is benefit too. +[1084.28 --> 1093.18] So I was looking at sort of the, how Go sort of provides this, or how Go talks about sort of this instantiation, right, of a generic function. +[1093.18 --> 1104.14] So for those who haven't sort of gotten into generic setting Go, pretty much when you create a generic function, what you're basically saying is that you're going to support a number of types, right, based on the typeset information you provide. +[1104.60 --> 1113.06] And basically what happens behind the scenes when you compile your code is that Go is going to generate an instance, right, of your function for each of the types you say you support, right? +[1113.18 --> 1121.96] So it's kind of like an actual code generation that basically that we have that provides a type safety, the type checks, right, that the compiler is able to provide. +[1121.96 --> 1128.14] But now the language itself is actually producing that for you behind the scenes and you don't have to actually write your own generators, right? +[1128.34 --> 1133.58] So effectively you're doing the exact same thing you were doing before, except now it's basically baked and supported to the language, right? +[1133.80 --> 1142.54] So this is an efficiency game, right, in terms of you, the programmer, being able to write these libraries, being able to support multiple types of given functions and things. +[1142.72 --> 1147.80] This is an efficiency game, really, not really a game changer for functional programming and Go per se. +[1147.80 --> 1152.38] I think just a more efficient way of writing the functions, basically. +[1152.80 --> 1162.06] Yeah, that is a really great point is that everything we can do now, technically you could do before because we had the empty interface, right? +[1162.38 --> 1167.50] The empty interface is the set of all types, all possible types in this context. +[1167.50 --> 1174.36] So you could have a function that take in an empty interface and that is a function that can take in anything you can come up with, right? +[1174.50 --> 1179.98] Same thing, it can return if you have the empty interface return and it can be the set of anything. +[1179.98 --> 1188.28] So your task before was write the code to define a subset of all things that your function wants to deal with. +[1188.70 --> 1191.84] Now there is a massive efficiency gain, right? +[1191.94 --> 1203.58] Because now you can constrain a type parameter and have the compiler compute for you whether a given type for a type parameter that has constraints on it is legal. +[1203.58 --> 1213.58] So do you want it to be integral or do you want it to be comparable or do you want it to be stringable or anything else under the sun that we can come up with? +[1214.08 --> 1217.26] You write less than a line of code, right? +[1217.42 --> 1222.26] And maybe you write a new constraint somewhere and then you apply it on your type parameter. +[1222.94 --> 1232.02] And the compiler does what I consider a very advanced and very useful computation for you across your entire code base. +[1232.02 --> 1237.84] And if you're writing a library, across everyone's entire code base that uses it, which is pretty amazing. +[1238.00 --> 1238.38] It's awful. +[1238.76 --> 1239.00] Yeah. +[1239.20 --> 1239.92] That is amazing. +[1258.62 --> 1261.40] This episode is brought to you by Honeycomb. +[1261.40 --> 1263.88] Find your most perplexing application issues. +[1264.18 --> 1271.10] Honeycomb is a fast analysis tool that reveals the truth about every aspect of your application in production. +[1271.58 --> 1275.56] Find out how users experience your code in complex and unpredictable environments. +[1275.86 --> 1280.76] Find patterns and outliers across billions of rows of data and definitively solve your problems. +[1281.20 --> 1282.68] And we use Honeycomb here at Change. +[1282.72 --> 1286.52] Well, that's why we welcome the opportunity to add them as one of our infrastructure partners. +[1286.52 --> 1294.36] In particular, we use Honeycomb to track down CDN issues recently, which we talked about at length on the Kaizen edition of the Ship It podcast. +[1294.62 --> 1295.30] So check that out. +[1295.52 --> 1296.02] Here's the thing. +[1296.26 --> 1299.52] Teams who don't use Honeycomb are forced to find the needle in the haystack. +[1299.52 --> 1302.80] They scroll through endless dashboards playing whack-a-mole. +[1303.00 --> 1306.06] They deal with alert floods, trying to guess which one matters. +[1306.44 --> 1311.64] And they go from tool to tool to tool playing sleuth, trying to figure out how all the puzzle pieces fit together. +[1312.00 --> 1318.32] It's this context switching and tool sprawl that are slowly killing teams' effectiveness and ultimately hindering their business. +[1318.32 --> 1325.48] With Honeycomb, you get a fast, unified, and clear understanding of the one thing driving your business. +[1325.74 --> 1326.16] Production. +[1326.66 --> 1329.16] With Honeycomb, you guess less and you know more. +[1329.56 --> 1334.74] Join the swarm and try Honeycomb free today at honeycomb.io slash changelog. +[1334.90 --> 1338.36] Again, honeycomb.io slash changelog. +[1348.32 --> 1358.32] How do you think you're going to be able to do this? +[1358.32 --> 1368.64] So the Go team has been very vocal and deliberate about advising pretty much everybody to not go gangbusters, which is on other things. +[1369.30 --> 1372.02] Not use it as a hammer with everything being a nail. +[1372.24 --> 1373.68] Just don't spread it everywhere. +[1373.68 --> 1377.20] Where it's a very sort of cautious approach, right? +[1377.24 --> 1383.82] Until some of the best practices start to emerge and still can see the sort of best use cases for generics in Go. +[1384.14 --> 1390.34] Do you think that applies even more so for those who want to practice functional programming in Go? +[1390.46 --> 1394.80] Or do you think, actually, it makes our lives a lot easier. +[1394.92 --> 1395.78] Let's just go full level. +[1396.54 --> 1399.34] Well, that is good advice for all software. +[1399.84 --> 1402.24] So yes, the answer to your question is yes. +[1402.72 --> 1402.84] Right. +[1402.84 --> 1411.54] There is no telling right now whether map is enough, right? +[1411.72 --> 1413.20] Or no one wants to use it. +[1413.56 --> 1423.02] And if people do want to use map, there's no telling whether the parallel version of that is better or complementary to the regular serial version of map, for example. +[1423.02 --> 1423.52] Right. +[1423.52 --> 1438.68] So the parallel version for listeners would be instead of applying the function on every element of the list one after the other, it's applying them in parallel with X number of Go routines or at least concurrently with X number of Go routines and then returning the result. +[1439.18 --> 1439.32] Right. +[1439.32 --> 1442.86] There's no telling right now. +[1443.06 --> 1445.14] You can say this is correct. +[1445.14 --> 1446.74] You can say this is possible. +[1447.10 --> 1454.00] You can guess this might be useful for this case or this case or this type of software or that type of software. +[1454.00 --> 1459.30] But there's no substitute for usage in the wild in the community, if you will. +[1459.72 --> 1461.60] There's just no substitute. +[1461.60 --> 1463.74] So these things should be tried. +[1464.32 --> 1473.14] Maybe, you know, we should go a little crazy, as they say, as the Go team has said, a little overboard to see what is possible. +[1473.14 --> 1478.08] But we should not say that this is the result. +[1478.32 --> 1481.00] This is the solution to all problems in FP. +[1481.48 --> 1492.64] We should get these things out there, play around with them, experiment with them, see what works at a larger scale, and then say that is the solution to problems X, Y, and Z. +[1493.12 --> 1496.02] You know, this is why we all say we test in production, right? +[1496.24 --> 1501.04] Because there's no substitute for real world use cases. +[1501.04 --> 1501.40] Yeah. +[1501.40 --> 1503.46] In that case, it's data. +[1503.72 --> 1510.52] And in our case, it's programmers experimenting with things and seeing what actually works in real world code bases. +[1510.72 --> 1511.68] But it's the same effect. +[1512.62 --> 1517.38] So picking back off of that, the thread that we paused and now I want us to resume. +[1517.92 --> 1518.68] You see what I did there? +[1519.36 --> 1519.96] I do. +[1520.20 --> 1520.70] I do. +[1520.88 --> 1521.44] I love it. +[1522.66 --> 1528.32] The idea of basically the when to use generics, right? +[1528.32 --> 1531.24] And in our case here, when to use functional programming. +[1531.24 --> 1534.96] I think that's a fair question to ask whether or not, you know, regardless of generics, right? +[1535.02 --> 1545.62] So one of the sort of the strong gripes that I found out there, right, that I've heard even in talking to some folks is that, look, Go is imperative, right? +[1545.62 --> 1546.58] Through and through, right? +[1546.70 --> 1553.68] We're kind of side bending and kind of pushing it in a direction it wasn't meant to, right? +[1553.68 --> 1558.62] Even though some of its features and its capabilities are well suited, right, for functional programming. +[1558.80 --> 1562.52] You know, functions being first class citizens and et cetera, et cetera. +[1562.68 --> 1566.46] A lot of the sort of the things you look for in functional programming, you can do and go. +[1566.46 --> 1569.04] But just because you can doesn't mean you should, right? +[1569.08 --> 1570.00] That whole adage, right? +[1570.24 --> 1583.20] So is functional programming still in the realm of sort of experimentation and just people who are curious and they can play around with it, but really like at work or whatever, right? +[1583.26 --> 1585.52] You know, you don't they don't really use it, right? +[1585.52 --> 1589.40] Like, where are we in terms of really, I guess it's an adoption question, right? +[1589.48 --> 1597.46] Like, are people interested or increasingly more so interested in using FP now, even with the bells and whistles that are enabled right through generics? +[1598.04 --> 1605.04] Is FP attractive enough for people who are traditionally doing this imperative style, especially in the world of Go? +[1605.68 --> 1610.54] Should they be looking or really should we try to bring FP into our production code? +[1611.00 --> 1611.86] Well, it's already here. +[1611.86 --> 1618.40] Anyone who uses context is using a part of FP, right? +[1618.48 --> 1622.64] Because context has a with cancel and with timeout. +[1622.76 --> 1625.34] Those return a function, right? +[1625.50 --> 1627.80] This is how pervasive FP is. +[1627.92 --> 1632.96] That is technically a higher order function because it's a function returning. +[1633.16 --> 1639.32] With cancel is a function and that's returning a function that you call to close to free the resources. +[1639.32 --> 1644.82] There's an internal Go routine running, right, with a timer and, you know, and freeze that resource, right? +[1645.26 --> 1645.94] That is FP. +[1646.08 --> 1659.52] And I think like all technologies that are used by more than a trivial number of people in the world, a corner of this technology is used a lot, right? +[1660.10 --> 1662.08] SQL is another good example here. +[1662.36 --> 1664.20] You can do many, many things with SQL. +[1664.20 --> 1669.64] You can turn Postgres SQL or PostgresQL, is that how you say it, right? +[1670.04 --> 1670.56] Yeah, yeah. +[1670.60 --> 1671.12] I'll allow it. +[1671.74 --> 1672.14] Okay. +[1672.48 --> 1672.76] Yeah. +[1672.84 --> 1673.10] Okay. +[1673.14 --> 1674.72] It's got the Johnny stamp of approval. +[1675.38 --> 1680.46] You can turn PostgresQL into a time series database if you want to, for example, right? +[1680.46 --> 1683.48] By using very advanced features of SQL. +[1683.76 --> 1689.50] But not many people do that because obviously it's not the, a lot of times it's not the best tool for the job. +[1689.62 --> 1700.04] But also most people prefer, including myself, to stick with a smaller set of foundational features of the technology. +[1700.64 --> 1703.16] And same thing with FP, right? +[1703.16 --> 1706.86] There is a smaller set than even all the stuff that I said today. +[1707.40 --> 1712.54] There's a smaller set of functionality that most people prefer to use. +[1712.64 --> 1721.42] And my hypothesis for why that is, is because it makes more sense and it fits into more workloads, right? +[1721.42 --> 1734.60] And so if you need to capture some state of a function and expose one operation on that state to the caller, using a higher order function makes a lot of sense. +[1734.94 --> 1741.88] It's a lot easier than doing a bunch of boilerplate and building a whole struct and storing the state in the struct and having a bunch of methods on it. +[1741.88 --> 1750.92] And that, it just fits right into that use case and it applies, that use case is applicable and exists on a ton of workloads. +[1751.08 --> 1752.90] And so we use it, right? +[1752.98 --> 1756.28] So FP, strictly speaking, is, is everywhere. +[1756.76 --> 1760.46] Now I'm going to modify your second question a little bit accordingly. +[1760.90 --> 1764.44] You said, should we be trying to get FP into more workloads? +[1764.44 --> 1773.68] Well, given that it's already a lot of places, I'm going to modify it and say, should we get more features of FP into more workloads? +[1774.30 --> 1775.42] Will you allow that? +[1776.86 --> 1777.98] Yeah, yeah, let's do it. +[1778.50 --> 1779.58] Yeah, tell me more. +[1779.92 --> 1781.70] All right, so now we've got the seal on that. +[1781.94 --> 1782.18] All right. +[1782.92 --> 1789.16] And so the way that I think about it is try to model it as the imperative declarative difference, I guess. +[1789.38 --> 1791.84] There's a fundamental difference between imperative and declarative. +[1791.84 --> 1795.72] You can see it with SQL versus Go, right? +[1795.82 --> 1801.22] Because SQL, you say, dear database, this is the result that I want. +[1801.30 --> 1813.38] And then the database has a ton of usually very clever implementation to figure out how to get whatever the way that it represents the data on disk into the result that you want. +[1813.78 --> 1814.00] Right? +[1814.06 --> 1818.28] So you can filter, you can join, you can group, you can order, et cetera, et cetera, et cetera. +[1818.28 --> 1824.36] Now, in the FP world, let's just take this map example I keep using. +[1825.10 --> 1825.20] Right? +[1825.30 --> 1835.72] So with map, the intention is to take a function, apply it onto every element of the list, the slice, and then return a new slice with the results. +[1836.18 --> 1839.28] You might return the same slice with the new results. +[1839.42 --> 1841.74] That's also sometimes legal as well. +[1841.74 --> 1846.76] But let's just say for sake of logic, we're returning a new slice with new results. +[1847.26 --> 1849.52] Well, we're iterating through a list. +[1849.58 --> 1850.50] So that's a for loop. +[1850.86 --> 1852.00] That's just what it is. +[1852.14 --> 1852.26] Right? +[1852.64 --> 1853.86] We can write a for loop. +[1854.08 --> 1855.22] And that's the imperative way. +[1855.28 --> 1862.48] It's telling the machine to go through the list one by one, call this function every time, put the result into a new list. +[1862.48 --> 1868.10] Or the other way is someone else did that for you, and you just call map. +[1868.62 --> 1870.52] And that's the kind of declarative way. +[1870.98 --> 1877.06] Because when you call map, you're effectively saying, like with SQL, you're saying this is the data that I want back. +[1877.52 --> 1884.66] With map, you're saying I want a new list back with the results of the old list applied at the function that I gave to map. +[1885.14 --> 1885.38] Right? +[1885.44 --> 1886.48] And that's pretty much it. +[1886.48 --> 1887.38] Right? +[1887.48 --> 1907.98] So when we think about this question of should we get more features of FP into language, into code bases that are mainly go code bases, the way that I think about it is would this code base or this part of a code base, would it benefit in some way from more declarativity? +[1908.52 --> 1911.42] I think I may have made up that word declarativity. +[1913.50 --> 1914.28] But it works. +[1914.28 --> 1915.92] I think people know what I mean. +[1915.92 --> 1916.10] Right. +[1916.28 --> 1917.82] Would it benefit in some way? +[1918.50 --> 1922.44] And there are a lot of possible ways that declarativity can help. +[1923.00 --> 1925.00] It can reduce lines of code. +[1925.66 --> 1927.18] It can add structure. +[1927.92 --> 1929.30] It can add readability. +[1929.74 --> 1932.26] It can also reduce readability if you're not careful. +[1932.46 --> 1933.48] So that's a tradeoff. +[1933.86 --> 1935.42] It can fix bugs. +[1935.42 --> 1938.60] Like if you have an error in one of your for loops. +[1938.60 --> 1942.80] This happens a lot with parallel and concurrent code. +[1943.12 --> 1947.96] And thus also code that deals with channel sends a lot too. +[1948.06 --> 1949.06] I've seen that a lot as well. +[1949.34 --> 1954.30] So it can reduce bugs because you're just getting someone else's code that did it the right way. +[1954.30 --> 1955.12] Right. +[1955.12 --> 1960.86] So if you think about it in this way, and this is where I start with any code base. +[1960.86 --> 1970.22] I always start, okay, well, is there code in here that we can reduce by using the declarative features of FP? +[1970.22 --> 1974.72] Whether it's maps or reduces or filters or zips and so on and so forth. +[1975.40 --> 1988.56] And as a quick anecdote to close that thought out, I have seen filter as one of the most valuable tools to next tools to take and put into code bases. +[1988.56 --> 1997.70] Because tons and tons of code has for loops that reduce down into selectively taking things out of a list and putting them somewhere else. +[1997.90 --> 1997.94] Right. +[1998.10 --> 1998.68] And that's filter. +[1999.14 --> 1999.26] Right. +[1999.88 --> 2000.26] Awesome. +[2000.48 --> 2000.74] Awesome. +[2001.32 --> 2002.28] Yeah, I can definitely see that. +[2002.36 --> 2005.52] I think at least for me, it's a mindset shift. +[2005.66 --> 2005.80] Right. +[2005.86 --> 2009.58] So like I don't have a reason not to use FP. +[2009.72 --> 2009.86] Right. +[2009.90 --> 2015.72] Or not to bring some of the ideas or not really, or I should say not to think FP first. +[2015.72 --> 2016.02] Right. +[2016.04 --> 2018.64] In terms of how do I approach, how do I solve this problem? +[2018.82 --> 2019.02] Right. +[2019.04 --> 2023.60] Like I'm so used to the way I'm used to doing it. +[2023.62 --> 2023.78] Right. +[2023.78 --> 2031.76] So it's more of a sort of a, I think maybe education and maybe sort of a seeing more people talk about it, present about it. +[2031.96 --> 2036.88] And having, you know, shows like this where people knowledgeable like yourself come on and sort of advocate for it. +[2037.08 --> 2043.10] I think it's more of a sort of a education issue more than the merits of it. +[2043.10 --> 2046.74] Because all these things you're talking about here, these are things that make my program safer. +[2047.02 --> 2050.12] They make, you know, my programs easier to understand. +[2050.44 --> 2050.84] Right. +[2050.88 --> 2056.48] And obviously, like as with everything in programming, we sprinkle things where they make sense. +[2056.48 --> 2057.34] Yeah. +[2057.34 --> 2066.70] And if your team at work is not using sort of FP style for things, and perhaps they want to or don't want to, so don't force it kind of thing. +[2066.82 --> 2074.46] But I think every once in a while, even using your example, like Filter, for example, that I see the use case for it, not that you mentioned it. +[2074.48 --> 2075.52] I see that everywhere. +[2075.96 --> 2079.24] It could definitely be used in a lot more places. +[2079.90 --> 2082.92] And maybe, who knows, maybe that sort of triggers somebody to say, hey, what's that? +[2082.92 --> 2084.94] Like, I've never seen this particular approach, right? +[2084.98 --> 2087.14] And then, boom, that's a brown bag launch right there. +[2087.14 --> 2090.48] You can educate some coworkers around the merits of FP. +[2090.64 --> 2098.20] So for me, I think it's an education thing more so than capabilities or features, which kind of leads me to my next question. +[2098.30 --> 2110.50] Is there anything missing in Go to nudge us even more towards sort of the, I don't want to say traditional FP, because I don't think we want to make an FP language per se. +[2110.68 --> 2116.62] But is there missing features in Go that would allow us to even take even more advantage of FP concepts? +[2117.14 --> 2120.26] Well, it's so generics are young. +[2120.48 --> 2123.88] And FP using generics is even younger. +[2124.40 --> 2134.36] So really, the best I can do to answer that is hypothesize based on very, very unscientific things like gut feel and things I've seen and anecdotal evidence, right? +[2134.36 --> 2146.22] I think the biggest thing that could unlock just making the building these things a little easier would be type parameters on methods of a struct, right? +[2146.22 --> 2148.30] Or a type, right? +[2148.64 --> 2149.98] So you can't do that now. +[2150.10 --> 2156.52] So you can't have, let's say you have a custom list type that's a list of type T. +[2156.52 --> 2165.58] You can't have a method called map on that list that takes a type parameter U and then returns a new list of type U. +[2166.28 --> 2168.04] You can still build map. +[2168.14 --> 2169.94] I have done it and I know others who have done it. +[2169.98 --> 2171.10] I've seen it all over the place. +[2171.30 --> 2172.26] You can still build it. +[2172.54 --> 2173.56] It has the same effect. +[2173.56 --> 2175.56] And it works the same. +[2176.26 --> 2194.66] The reason that I'm hypothesizing that type parameters on methods would make things easier is because for many folks who are trained using Java or C++ or other pure, what I'll call pure object-oriented languages, although whether they're pure or not is another podcast. +[2195.04 --> 2198.24] Pure in air quotes for those not watching. +[2198.62 --> 2199.14] Air quotes. +[2199.30 --> 2199.54] Yeah. +[2199.70 --> 2200.60] Thank you for that call. +[2201.10 --> 2202.32] Definitely in air quotes. +[2202.32 --> 2217.62] Folks who are trained with the classical OOP type of programming, the OOP design patterns and methodology, having a method on a struct feels a lot like a method on a class in Java. +[2217.92 --> 2219.92] It's not strictly the same. +[2219.92 --> 2247.62] But for folks who are used to thinking of methods, if you have a method, quote unquote, again, air quotes, if you have a method, air quotes on your list struct in Go, well, it makes more sense to do it that way rather than to do it the more kind of pure, again, air quotes, pure FP way, which would be a totally separate function that takes in a certain type and returns a new type. +[2247.62 --> 2277.60] Again, for some of the more advanced thing. +[2277.60 --> 2279.02] It's not two hours ago. +[2279.02 --> 2286.10] You know, higher kind of types add a ton of complexity to a compiler and a type checker inside of the compiler. +[2286.72 --> 2294.42] But, you know, they do offer basically something called a type constructor, which is a type that depends on another type. +[2294.62 --> 2294.80] Right. +[2294.80 --> 2300.08] And if you can do that, I'll leave it as an exercise to the reader or listener or watcher. +[2300.78 --> 2302.94] What kinds of things could you do with that? +[2303.22 --> 2304.52] Take me a while to get into it. +[2304.68 --> 2306.04] So I'm not going to go there now. +[2306.04 --> 2316.22] But if you had that right now, assuming it was correct in the compiler, you can start to do some very advanced things. +[2317.08 --> 2330.50] But the only thing that I'll say is whenever I've seen that feature in a language or not in the language and then introduced into the language, at that point, it starts to become what you called sort of a functional language. +[2330.50 --> 2331.10] Right. +[2331.10 --> 2331.20] Right. +[2331.70 --> 2342.08] Because getting to that level of sophistication in the type system often means that there are a bunch of people already there who have been demanding it. +[2342.44 --> 2344.84] And those people are the majority. +[2345.16 --> 2349.00] And those people write very heavily function oriented code. +[2349.00 --> 2354.44] So really, I'm going to just limit it to the type parameters on methods thing because that thing is not very controversial. +[2355.26 --> 2367.12] But higher kinded types or high order types are a very interesting and cool thing that I really do wish was in Go because it would enable some very slick FP things. +[2367.56 --> 2368.98] Yeah, that is not controversial to me. +[2369.10 --> 2369.88] I could see that. +[2370.12 --> 2370.28] Okay. +[2370.56 --> 2371.04] Some might. +[2371.14 --> 2372.32] Maybe some people will think of it. +[2372.32 --> 2372.64] Okay. +[2379.00 --> 2384.56] We are going to ship it. +[2384.56 --> 2387.08] Three, two, one. +[2387.54 --> 2394.80] I'm Karhara Zhu, host of Ship It, a show with weekly episodes about getting your best ideas into the world and seeing what happens. +[2394.80 --> 2401.30] We talk about code, ops, infrastructure, and the people that make it happen like charity majors from Honeycomb. +[2401.72 --> 2404.56] We act like great engineers make great teams. +[2404.78 --> 2406.22] And it's exactly the opposite. +[2406.22 --> 2409.84] In fact, it is great teams that make great engineers. +[2410.42 --> 2413.74] And they finally win the founders of continuous delivery. +[2414.12 --> 2416.90] Start off assuming that we're wrong rather than assuming that we're right. +[2417.16 --> 2418.16] Test our ideas. +[2418.28 --> 2419.78] Try and falsify our ideas. +[2419.92 --> 2421.92] Those are better ways of doing work. +[2421.98 --> 2424.20] And it doesn't really matter what work it is that you're doing. +[2424.34 --> 2426.00] That stuff just works better. +[2426.00 --> 2437.62] We even experiment on our own open source podcasting platform so that you can see how we implement specific tools and services within changelog.com, what works and what fails. +[2437.62 --> 2441.86] It's like there's a brand new hammer and we grab hold of it and everyone gathers around. +[2441.94 --> 2445.74] We put our hand out and we strike it right on our thumb. +[2445.96 --> 2450.10] And then everybody knows that hammer really hurts when you strike it on your thumb. +[2450.18 --> 2451.46] I'm glad those guys did it. +[2451.54 --> 2452.90] I've learned something instead. +[2453.04 --> 2453.16] Yeah. +[2453.32 --> 2457.92] I think that's a very interesting perspective, but I don't see that way. +[2458.06 --> 2458.32] Okay. +[2458.42 --> 2461.52] It's an amazing analogy, but I'm not sure that applies here. +[2461.52 --> 2464.16] Listen to an episode that seems interesting or helpful. +[2464.32 --> 2465.98] And if you like it, subscribe today. +[2466.08 --> 2467.22] We'd love to have you with us. +[2471.70 --> 2472.10] Okay. +[2472.20 --> 2476.72] So before we transition into unpopular opinions, maybe I want one of those controversial ideas. +[2478.48 --> 2478.88] Okay. +[2479.64 --> 2480.12] All right. +[2480.46 --> 2481.14] Let's get it going. +[2481.72 --> 2490.88] I'm not going to give one that talks about higher kind of types because it's been a while, first of all, so I don't think I would do it justice or maybe even get some parts incorrect. +[2490.88 --> 2493.50] But also it would take me a while. +[2493.94 --> 2499.38] But I want to mention one pattern from FP that I think is really great. +[2499.92 --> 2503.16] And it's not that hard to do yourself in your code. +[2503.52 --> 2507.44] I also don't think it's too hard to imagine how it would work in a code base. +[2507.96 --> 2509.12] It's called the lens. +[2509.84 --> 2513.48] So it maybe sounds like an intimidating thing to some people. +[2513.64 --> 2514.72] It certainly did to me. +[2515.38 --> 2518.14] The lens is essentially a tuple of two functions. +[2518.14 --> 2520.92] One function is a getter for some data. +[2521.44 --> 2526.22] And then the other function is the setter for some probably the same type of data. +[2526.72 --> 2528.26] But you can have separate things too. +[2528.34 --> 2533.24] You can have a getter for one piece of data and a setter for another piece of data in the same tuple. +[2533.80 --> 2537.82] So usually it's the getter first in the tuple and then the setter second in the tuple. +[2537.82 --> 2543.12] Now the first function takes no parameters and returns the result, the data. +[2543.70 --> 2550.48] It might take one parameter if the data is multidimensional like a map and you want it to specify a key. +[2551.14 --> 2551.82] And that's up to you. +[2551.90 --> 2553.10] That's part of the design pattern. +[2554.04 --> 2559.00] And then the setter potentially takes in the dimension of the data, again like the key of the map. +[2559.00 --> 2563.80] And then it also takes in the data that you want to set into that place. +[2564.72 --> 2570.86] So it might be the value of a map if we're going to use the map, like the example of a map. +[2571.50 --> 2573.00] And then it returns nothing. +[2573.46 --> 2579.74] These things can alternatively return an error if this is like doing something over the network or something that might fail. +[2580.14 --> 2581.24] But traditionally they're not. +[2581.24 --> 2590.00] All right, so if you could kind of imagine how these functions would be created, they would probably be closures. +[2590.58 --> 2592.40] They would close over your data. +[2593.04 --> 2595.18] They would take some parameter in the get. +[2595.34 --> 2596.78] Let's take the get case, for example. +[2597.28 --> 2601.90] They would take the parameter for the thing to get and then they would return the value of it. +[2602.26 --> 2603.16] Same thing with the set. +[2603.24 --> 2605.34] It would close over the data and then do its thing. +[2605.34 --> 2613.00] It's a fancy term called a lens, but it's a fairly basic foundational thing. +[2613.80 --> 2618.30] And that is a common theme in FP actually. +[2618.98 --> 2623.70] So I called this the lens and it is in Wikipedia and it's in the textbook and everything. +[2624.14 --> 2628.12] But it just boils down to functions like most things in FP do. +[2628.66 --> 2633.32] They just boil down to one or more functions, maybe combined together in some creative way. +[2633.32 --> 2637.88] Or maybe one function applied across the data structure in some other way. +[2638.24 --> 2643.52] So I would challenge everyone listening to just give it some thought. +[2643.80 --> 2647.48] Can that reduce code somewhere in your code base? +[2648.12 --> 2656.10] If you have a bunch of getters, can you get rid of those getters and just replace them with a closure of the lens? +[2656.52 --> 2661.66] That's a very interesting one to me because it is exceptionally simple. +[2661.66 --> 2663.10] I'll put it a different way. +[2663.26 --> 2665.72] The return on investment is very high. +[2666.28 --> 2671.02] It's very simple, but it tends to be applicable and useful in a lot of places in the code base. +[2671.48 --> 2671.84] Very cool. +[2672.34 --> 2675.18] I'd love to see that in action, like in practical use. +[2675.52 --> 2677.20] The way you explain it, it sounds great. +[2677.54 --> 2679.68] I'm like, yeah, that sounds like something I'd be up for. +[2679.98 --> 2680.36] There you go. +[2680.54 --> 2680.66] Yeah. +[2680.96 --> 2681.16] Yeah. +[2681.16 --> 2685.16] Perhaps we need sort of a blog post from you when you're done with school or something. +[2685.94 --> 2686.34] Yeah. +[2686.34 --> 2686.74] Yeah. +[2687.94 --> 2690.56] So if you remember me asking, what did you go back to school for? +[2690.96 --> 2693.22] Well, I went back to school for computer science. +[2693.44 --> 2697.56] I got an undergraduate computer science degree in 2008. +[2698.16 --> 2701.24] You know, I'm not that old, but I feel older than I am, I guess. +[2701.52 --> 2705.26] You can do the math at home, see and figure out how old I'm likely to be. +[2705.94 --> 2710.98] And so then I've been in industry ever since, focused on industry things, right? +[2710.98 --> 2713.06] And they all tend to be very practical. +[2713.76 --> 2717.06] They all tend to be, you know, single goal oriented. +[2717.98 --> 2724.02] And all of the learning that I've done has been very focused on a goal rather than broader. +[2724.28 --> 2731.04] So I went back to get a master's to get that breadth again, because I've kind of missed that. +[2731.64 --> 2737.18] Right now, I kept using the SQL and the database examples, because right now I'm taking a course in databases. +[2737.18 --> 2741.28] So that's an easy extension there. +[2741.96 --> 2750.86] But the thing that I'm planning to specialize in now is formal methods, because I love this stuff, like programming language theory and things like that. +[2751.12 --> 2753.16] I really do enjoy it a lot. +[2753.96 --> 2760.20] And the formal method specialization happens to have quite a bit of language theory stuff in it. +[2760.48 --> 2762.06] So look forward to that. +[2762.06 --> 2765.20] Yeah, it sounds like you're in your bread and butter kind of space right now. +[2765.38 --> 2767.32] That's nice, nice, nice. +[2767.58 --> 2767.74] Yeah. +[2767.98 --> 2768.20] Yeah. +[2768.30 --> 2776.18] And now that you mention it, interestingly enough, I think we tend to go to school to get a specialization, to narrow it down on a very particular set of skills. +[2776.66 --> 2781.90] Even though when we go in, we might not know that's what we're doing, but we're going to learn how to do X, right? +[2781.90 --> 2784.78] And then we come out and we get a job to do X, right? +[2784.78 --> 2789.20] It's very rarely do you go back just for the sake of breath, right? +[2789.26 --> 2793.16] Just to get, basically to go find out what you don't know about, right? +[2793.24 --> 2799.02] And then you sort of see how you can sort of apply it to your already existing set of knowledge that you require in the industry. +[2799.36 --> 2800.30] So yeah, I commend you for that. +[2800.44 --> 2801.28] That's very cool. +[2801.56 --> 2806.44] Well, I'm very, very lucky and privileged to be able to do that at this point, right? +[2806.44 --> 2815.28] Because I think probably all of us remember in undergrad, we probably just went so we could get to whatever our next step was. +[2815.70 --> 2817.46] Most of us probably was a job, right? +[2817.58 --> 2820.96] Maybe some of us was grad school, but most of us was a job. +[2821.04 --> 2822.56] It certainly was me, right? +[2822.56 --> 2826.72] I was laser focused on getting done so I could get paid. +[2827.98 --> 2829.66] But these days it's different, right? +[2829.66 --> 2836.32] I have a job and, you know, I'm able to continue working in this job while I do this part-time program. +[2837.06 --> 2849.36] And so I fully acknowledge and I try to remind myself daily that I'm very lucky to be able to do something like this just because I want to rather than having to do something. +[2849.66 --> 2850.02] Pretty cool. +[2850.16 --> 2850.46] Pretty cool. +[2850.96 --> 2858.28] Well, Aaron, unless you have another like more like popular opinion for me, although I don't think anything you said here is unpopular, right? +[2858.28 --> 2862.88] I think maybe it's just the way you talk and put it, but I don't think any of it was unpopular. +[2863.54 --> 2864.32] Diplomacy, yeah. +[2864.94 --> 2865.52] Yeah, yeah. +[2865.52 --> 2867.40] Well, I do have an unpopular opinion. +[2867.66 --> 2867.92] Okay. +[2868.24 --> 2869.68] This one will be unpopular. +[2870.02 --> 2870.52] Okay, wait. +[2870.64 --> 2872.18] Then I got to play the song if that's the case. +[2872.44 --> 2873.12] Okay, okay. +[2873.14 --> 2873.76] Let's do it. +[2878.14 --> 2878.78] Unpopular opinion. +[2878.98 --> 2879.96] You want to. +[2879.96 --> 2881.64] I actually think you should probably leave. +[2882.14 --> 2886.58] Unpopular opinion. +[2887.14 --> 2887.22] Unpopular opinion. +[2887.90 --> 2888.42] Boom. +[2892.76 --> 2893.46] Lay it on us. +[2893.46 --> 2894.18] All right. +[2894.18 --> 2894.20] All right. +[2894.58 --> 2896.88] Well, I've been doing a bunch of Rust lately. +[2897.04 --> 2900.96] Some like very low level stuff dealing with virtual machines. +[2901.82 --> 2917.60] And I believe now that I've used Rust long enough, I believe the type system in Rust is more complete and leads to more concise and easier to write and read programs than Go. +[2917.60 --> 2918.84] Okay, then. +[2919.32 --> 2919.62] Yeah. +[2919.62 --> 2923.46] So the Go type system has some work to do. +[2923.88 --> 2924.26] Okay. +[2924.64 --> 2925.06] Okay. +[2925.32 --> 2925.60] Yeah. +[2925.98 --> 2930.34] I don't mean to fan the flames of the Go Rust thing, but I'm sure I just have. +[2930.56 --> 2931.90] Oh, too late, buddy. +[2931.90 --> 2932.70] I'm just done it. +[2932.70 --> 2932.74] I'm just done it. +[2933.78 --> 2934.62] The flame war. +[2935.58 --> 2938.56] It's been zero days now since the flame wars. +[2939.96 --> 2941.30] I know, I know. +[2941.30 --> 2946.56] I talked a lot about type system things today, right? +[2946.72 --> 2949.70] And it's not that Rust lets you do all those things. +[2949.88 --> 2950.84] It certainly doesn't. +[2951.40 --> 2958.96] But the type system in Rust has more features than the type system in Go. +[2959.40 --> 2963.70] It certainly allows for some programs that take forever to compile and are hard to read. +[2963.70 --> 2966.14] So that's the pro and con thing here. +[2966.48 --> 2976.22] But from my experiences so far, it lets you write more expressive code with fewer lines than the Go equivalent. +[2976.84 --> 2977.14] Okay. +[2977.68 --> 2979.98] Well, we shall put that to the test. +[2980.30 --> 2980.56] Yeah. +[2980.82 --> 2981.68] The audience. +[2982.20 --> 2986.74] They will tell us whether they agree with you or disagree with you and come find you in the night or something. +[2987.06 --> 2987.32] Yeah. +[2987.46 --> 2987.66] Yeah. +[2987.72 --> 2989.70] You all know my Twitter. +[2990.00 --> 2991.92] So you can send me hate mail. +[2991.92 --> 2993.98] I still love Go. +[2994.22 --> 2995.10] Don't get me wrong. +[2995.30 --> 2998.70] I've been writing Go for something like 10 years or something. +[2999.48 --> 3003.58] And there's no language I can build something more quickly in. +[3004.10 --> 3005.46] I'm just getting started with Rust. +[3005.56 --> 3006.92] So that's my initial opinion, though. +[3007.88 --> 3009.42] Hey, you can love more than one language. +[3009.62 --> 3009.98] That's cool. +[3010.18 --> 3010.50] Oh, yeah. +[3010.62 --> 3011.60] As long as it's Go and Go. +[3011.74 --> 3012.04] That's cool. +[3014.54 --> 3016.46] Love anything you want as long as it's Go. +[3018.30 --> 3019.28] I love that. +[3020.08 --> 3020.48] Awesome. +[3020.66 --> 3020.98] Awesome. +[3020.98 --> 3025.68] So thank you so much, Aaron, for coming back on the show and talking about functional programming. +[3026.38 --> 3037.00] Hopefully we hit a few nails on the head for folks that have been curious about functional programming and how it's done, how it's been done, and what Generics enabled for us moving forward. +[3037.56 --> 3038.80] This was a fun discussion. +[3039.04 --> 3040.36] Thanks again, Aaron, for coming on the show. +[3040.68 --> 3042.18] Thank you so much for having me, Jimmy. +[3042.18 --> 3047.80] All right. +[3047.80 --> 3049.40] That is Go time for this week. +[3049.64 --> 3050.36] Thanks for listening. +[3050.36 --> 3056.42] If you're FP curious and want some more like this, we had Eric Normand on JS Party Whileback talking about it. +[3056.58 --> 3061.84] Eric is one of the best people I've heard explain FP principles and why they are worth putting to use in your code base. +[3061.98 --> 3062.48] Take a listen. +[3062.88 --> 3065.82] Have you ever seen the original show, the Get Smart? +[3065.82 --> 3069.88] Like the intro where he has like 30 different doors he walks through. +[3070.10 --> 3070.44] Yeah. +[3070.66 --> 3073.98] And then he holds up his shoe to his ear and talks into it. +[3074.00 --> 3074.02] Yeah. +[3074.20 --> 3075.22] And he has a phone. +[3075.38 --> 3076.72] He has a cell phone in his shoe. +[3077.42 --> 3077.58] Yeah. +[3077.68 --> 3080.36] But you go through all these doors, right? +[3080.42 --> 3080.58] Right. +[3080.58 --> 3084.00] And you get deeper and deeper into the sanctum of functional programming. +[3084.00 --> 3091.86] Well, that first door is just recognizing the difference between what I call actions, calculations and data. +[3092.14 --> 3092.50] Okay. +[3093.20 --> 3096.48] Calculations are often known as pure functions. +[3096.68 --> 3104.72] They're the stuff you can do in your language that always gives you the same answer no matter how many times you run them or when you run them. +[3105.38 --> 3105.66] Okay. +[3105.70 --> 3108.08] So this is like addition, right? +[3108.12 --> 3110.10] Addition is always going to two plus two. +[3110.18 --> 3110.84] It's always four. +[3111.28 --> 3113.00] It doesn't matter how many times you run that. +[3113.00 --> 3118.04] But then there's actions that do depend on when you run them or how many times you run them. +[3118.48 --> 3131.70] So reading from a mutable variable, if you read after someone has written to it, you're going to get a different answer than reading before the other, you know, other part of the code writes to it. +[3131.94 --> 3135.60] Likewise, sending an email or writing something to disk. +[3135.60 --> 3143.02] These are all actions because, you know, sending the email zero times is different from sending it one time or 10 times. +[3143.24 --> 3157.58] And so making this distinction between actions that depend on time because they're hard to deal with and calculations is like the first gateway into functional programming. +[3157.58 --> 3159.12] Oh, and data is easy. +[3159.12 --> 3162.72] Data is just, you know, the stuff that doesn't do anything. +[3162.84 --> 3163.52] It's just inert. +[3163.80 --> 3169.34] You know, the strings and numbers and hash maps and lists and stuff like that. +[3169.34 --> 3172.30] That's episode 163 of JS Party. +[3172.30 --> 3178.60] If you want to hear more, find it at jsparty.fm slash 163 or search for it in your podcast app. +[3178.72 --> 3182.46] The title of the episode is JS is an Occasionally Functional Language. +[3182.96 --> 3185.86] Thanks once again to Fastly and Fly for partnering with us. +[3186.10 --> 3187.40] Please check out what they're up to. +[3187.56 --> 3188.74] They support everything we do. +[3189.00 --> 3191.10] And of course, thank you to our Beat Freakin' residents. +[3191.36 --> 3195.32] Breakmaster Cylinder, our beats are dope because BMC makes dope beats. +[3195.64 --> 3196.52] It's as simple as that. +[3196.52 --> 3200.22] That's all I have for you, but we'll talk to you next time on Go Time. diff --git a/Go and PHP sitting in a tree..._transcript.txt b/Go and PHP sitting in a tree..._transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..ccdb6d7429138b8873d5d1d0071108a3eb1789b8 --- /dev/null +++ b/Go and PHP sitting in a tree..._transcript.txt @@ -0,0 +1,776 @@ +[0.00 --> 3.76] But if you also know Golang, you can do even crazier stuff. +[3.98 --> 8.30] Like you can take, for example, machine learning model from Keras, compress it, put it into +[8.30 --> 12.44] the Golang SDK, and then invoke from PHP, and ta-dum, you now have machine learning +[12.44 --> 14.68] embedded directly into PHP via Golang Bridge. +[15.12 --> 19.14] So you can do these crazy things if you know multiple languages, specifically Golang. +[19.28 --> 22.02] Golang is beautiful because it's quite easy to learn it, right? +[22.04 --> 26.18] It doesn't have many of these legacy overhead and nuances you have to remember. +[26.18 --> 31.00] So it's not required, but if you do, well, you're essentially a superhero in PHP world. +[35.08 --> 38.70] This episode is brought to you by our friends at Sourcegraph. +[38.80 --> 41.38] They recently launched a new feature called Code Insights. +[41.70 --> 45.12] Now you can track what really matters to you and your team in your code base, transform +[45.12 --> 49.38] your code into a querable database to create customizable visual dashboards in seconds. +[49.78 --> 52.66] Here's how engineering teams are using Code Insights. +[52.66 --> 56.88] They can track migrations, adoption, and deprecation across the code base. +[57.14 --> 60.36] They can detect and track versions of languages or packages. +[60.78 --> 64.18] They can ensure the removal of security vulnerabilities like Log4j. +[64.46 --> 69.74] They can understand code by team, track code smells and health, and visualize configurations +[69.74 --> 70.72] and services. +[71.30 --> 73.88] Here's what the engineering manager at Prezi has to say about this new feature. +[74.30 --> 74.40] Quote, +[74.40 --> 99.24] The next step is to see how other teams are using this awesome feature. +[99.24 --> 104.38] Head to about.sourcegraph.com slash code dash insights. +[104.62 --> 106.12] This link will be in the show notes. +[106.24 --> 110.98] Again, about.sourcegraph.com slash code dash insights. +[125.62 --> 126.60] Let's do it. +[127.16 --> 128.22] It's go time. +[128.22 --> 130.52] Welcome to go time. +[130.78 --> 133.74] Your source for diverse discussions from around the go community. +[134.14 --> 136.60] We record live on Tuesdays at 3 p.m. +[136.68 --> 137.00] U.S. +[137.02 --> 137.68] Eastern time. +[137.96 --> 141.72] Subscribe at youtube.com slash changelog so you don't miss it. +[141.92 --> 144.62] And don't forget to follow the show on Twitter at go time FM. +[144.94 --> 149.44] Special thanks to our partners at Fastly for shipping our shows super fast to wherever you +[149.44 --> 149.76] listen. +[150.02 --> 151.94] Check them out at fastly.com. +[152.18 --> 153.38] Okay, here we go. +[158.22 --> 163.18] Welcome to this episode of go time everyone. +[163.18 --> 165.12] Glad to have you back. +[165.62 --> 170.64] So, go and PHP sitting in a tree. +[171.36 --> 175.70] I don't know there's, you know, kissing going on, but we're going to figure out why go and +[175.70 --> 179.32] PHP are together on this particular episode. +[179.32 --> 183.24] Because last time I touched PHP was many years ago. +[183.24 --> 187.66] And although it has it had its charms, it did have some quirks as well. +[187.66 --> 190.12] And we may touch on those during this episode. +[190.12 --> 197.68] But joining me today are two folks who actually work with go and PHP on a regular basis. +[197.68 --> 199.64] And that got us curious. +[199.90 --> 202.30] Actually, someone actually suggested this episode. +[202.54 --> 203.56] I believe. +[203.68 --> 204.86] Let me, let me look up there. +[204.86 --> 210.34] Oh, I will find who it was and give them a shout out for suggesting this episode. +[210.34 --> 219.74] Shout out to listener Seb for requesting this episode and putting Roadrunner and PHP on our +[219.74 --> 220.04] radar. +[220.56 --> 221.08] Thanks, Seb. +[223.84 --> 227.64] And they actually recommended Roadrunner, which is a project we'll definitely be touching on +[227.64 --> 231.56] by name as an example of go and PHP working well together. +[231.84 --> 239.92] So, joining me today are two guests, Mr. Valerie, and I'm going to butcher your last name, Piacchansky +[239.92 --> 241.14] Did I get that right? +[241.88 --> 242.38] Close enough. +[244.18 --> 247.00] And then also Anton Titov. +[247.10 --> 248.10] Hopefully I got that right. +[248.20 --> 248.88] Yeah, that's correct. +[249.18 --> 249.72] Nice to meet you. +[249.94 --> 250.16] Okay. +[250.80 --> 256.18] So, Valerie is a software developer at Spiral Scout, the team that works on Roadrunner, +[256.26 --> 258.12] the particular project that we'll be touching on. +[258.34 --> 263.34] He enjoys working on algorithms, writing his own operating system for learning purposes in +[263.34 --> 267.48] C++, and helping folks get into programming, which is pretty awesome. +[267.48 --> 272.56] He also streams on Twitch, mostly open source Go stuff. +[273.20 --> 278.66] Anton is actually the CTO and co-founder at Spiral Scout, also on the same team that actively +[278.66 --> 280.70] works behind Roadrunner. +[281.24 --> 282.86] He loves software and hardware. +[283.12 --> 287.12] We were just talking before the show about some of the hardware stuff he's working on +[287.12 --> 293.58] with Go, rather than going the traditional route of the embedded stuff with C or Python +[293.58 --> 293.92] and whatnot. +[294.32 --> 295.60] I got a chance to take a peek at that. +[295.70 --> 296.64] Very interesting stuff. +[296.64 --> 300.30] He also enjoys some DIY robotics and machine learning. +[300.88 --> 302.26] Welcome to you both, gents. +[302.78 --> 303.32] Yeah, thank you. +[303.74 --> 306.82] I was saying I do not try to pronounce my surname. +[307.82 --> 310.16] So it's like rather complicated. +[311.22 --> 312.68] But you're almost correct. +[313.62 --> 314.38] It's all good. +[314.44 --> 314.94] It's all good. +[315.04 --> 315.36] Yeah, yeah. +[315.68 --> 317.42] So Go and PHP. +[317.80 --> 320.76] Let's start with what brought you to PHP. +[320.98 --> 324.10] Like why are you working on PHP this day and age? +[324.10 --> 326.78] Well, not that there's anything wrong with working on PHP this day and age. +[326.78 --> 330.48] I'm saying like, you know, out of all the languages one could pick, PHP has been around +[330.48 --> 332.10] a long, long, long time. +[332.28 --> 334.06] Like back in the infancy of the web. +[334.16 --> 339.40] It was like at some point, PHP was like the it language to actually write, you know, dynamic +[339.40 --> 340.18] websites and whatnot. +[340.40 --> 343.78] Competing with the like of ASP, you know, classic, as we refer to it now. +[343.78 --> 348.68] So in the likes of ColdFusion, I mean, these things are like granddaddies or grandmommies +[348.68 --> 349.92] of the early web. +[350.08 --> 351.06] What led you to PHP? +[351.48 --> 353.90] Well, as you mentioned, it's a very old language. +[354.16 --> 358.40] I was young and naive and I wanted to have my own forum or my own CMS board. +[358.58 --> 364.02] And if you're trying to build a forum back in like 05, the only option would be for you +[364.02 --> 364.52] is PHP. +[364.92 --> 369.86] It's like I remember trying to download the source code of like the website and trying +[369.86 --> 371.08] to figure out why it doesn't work. +[371.08 --> 375.10] Well, apparently you had to install a bunch of instruments to also make it work. +[375.26 --> 378.84] But it's just been the beginning of kind of this long, long journey. +[379.46 --> 380.58] And I'm still sticking to that. +[380.88 --> 383.38] So it's a beautiful language these days and it changed a lot. +[383.70 --> 384.14] Yeah, yeah. +[384.24 --> 384.72] Very nice. +[385.38 --> 389.08] So surprisingly, I'm not a PHP developer. +[389.34 --> 394.60] So I'm a good developer and working on a whole part of the ecosystem. +[394.96 --> 396.86] We're trying to connect into PHP parts. +[396.86 --> 402.12] So Anton for me, it's like a light at the end of the tunnel connected to me from the +[402.12 --> 402.80] PHP side. +[403.28 --> 403.34] Okay. +[403.48 --> 407.32] So then you've got the PHP as part of your background. +[407.32 --> 409.68] And then here comes a long go. +[409.82 --> 414.32] Like when did you get into it and what led you to actually combining those two things? +[414.88 --> 417.96] I can probably talk from kind of combine part. +[418.10 --> 421.44] I mean, the goal been around when I was kind of started studying it. +[421.48 --> 424.48] It was around for a few years and I only hear like a good things about it. +[424.48 --> 430.52] It's like performance, fast, concurrent, all these kind of terms which now, well, we all +[430.52 --> 431.34] know about Golang. +[431.76 --> 434.72] Well, I mean, I actually just tried to just play with that. +[434.98 --> 437.26] And I played, I made a couple interesting applications. +[437.64 --> 443.32] And since PHP was the main kind of production stack, I was just trying to see how I can kind +[443.32 --> 444.68] of use it within this practice. +[444.82 --> 449.14] Because all the examples on Golang was kind of like small and easy, right? +[449.14 --> 454.16] And on PHP, we have frameworks, like 10 layers of abstractions, OREMs, and et cetera. +[454.26 --> 456.50] So it's kind of very different worlds. +[457.18 --> 458.48] So, I mean, it was just curious. +[458.64 --> 462.56] It was kind of very curious language for someone who didn't work in this, like this type of +[462.56 --> 463.82] language for a very long time. +[464.56 --> 467.28] And, well, that has been an idea. +[467.82 --> 469.72] Can we actually make them work together? +[469.84 --> 472.20] Can we get benefits of PHP and benefits on Go? +[472.20 --> 477.12] And can improve the developer experience or our own experience? +[477.58 --> 485.60] I came to Spiral Scout, like, in 2018 as just a regular developer on one of the projects, +[485.86 --> 486.90] like Golang developer. +[487.46 --> 490.38] Previously, I was programming in .NET. +[490.50 --> 495.32] So I was heavily involved into .NET ecosystem, like C Sharp. +[495.32 --> 501.28] I've got some, like, I guess, 40, 70, 483 exam paths. +[501.44 --> 504.28] It's like C Sharp, something like C Sharp Professional. +[504.62 --> 511.36] So, and I saw a guy who worked on a very famous taxi company based in New York. +[512.00 --> 519.82] So he rewriting old Ruby on Rails system into some interesting language. +[519.82 --> 526.22] So it was, like, 2015, maybe in early days of the Golang. +[526.52 --> 528.82] So I, like, what is language? +[529.04 --> 533.04] So could you please explain me about what is... +[533.04 --> 537.24] So I was really impressed about first web server written in Golang. +[537.34 --> 544.92] So it was so little lines of code, and it brings you to, like, a web server that can +[544.92 --> 547.78] respond to, like, just hello world, but it just works. +[547.78 --> 553.76] After that, I came into, like, SpiralScal was involved into, like, internal projects. +[554.08 --> 559.58] And one time, I started to write a test in the error genre. +[560.02 --> 561.36] So it was, like, a mess. +[561.78 --> 567.16] Because entering from the PHP expertise, like, forgot to turn errors from the functions. +[568.08 --> 571.02] Like, there are no errors if you don't return them. +[571.02 --> 577.28] Of course, I fixed it, and it was, like, a lot of errors in the tests. +[577.94 --> 584.46] And I sent this PR to Anton, and Anton was, like, oh, my God, why you did this? +[585.26 --> 586.74] And we, like, okay. +[586.98 --> 592.40] We started fixing it, and we eventually, and we finally fixed all these, like, errors. +[592.40 --> 598.24] And after that, I was, like, involved into the Golang part of the road genre. +[598.50 --> 603.76] And we started working with Anton together to improve the quality, too. +[604.72 --> 610.16] Well, he pretty much rewrited most of the parts of it, but that was for the good. +[611.72 --> 612.68] Yeah, might be. +[612.68 --> 620.40] Obviously, you saw a need, and you decided to fill in and started meeting, doing some of the things that you saw that needed to be done. +[620.98 --> 621.66] And that's always awesome. +[621.78 --> 625.74] That's how, you know, the majority of open source contributions happen, right? +[625.84 --> 630.52] You know, you find something that you enjoy working on, and you contribute code to it, and that's a beautiful thing. +[630.52 --> 643.52] Obviously, when I think about my experience with PHP and all sort of the hoops we had to jump through to sort of, to, quote, unquote, make it scale and sort of using today's terminology and whatnot. +[644.04 --> 655.34] When I sort of read the description of the project, and I'm like, okay, this is a load balancer rolled into some sort of application server rolled into some sort of, I mean, it's trying to do a lot of things. +[655.34 --> 665.96] So why don't we start with what the difference is between plain-jane PHP application server, like the last one I was used to was like Zen or something like that. +[666.12 --> 667.20] It was a very long time ago. +[667.44 --> 668.46] Yeah, exactly. +[668.90 --> 670.14] I'm dating myself here. +[670.46 --> 680.48] But what's the difference between sort of those, I guess, for lack of a better terminology, those traditional application servers that are designed to run PHP versus this new approach? +[680.48 --> 688.48] To answer this question, it's actually important to understand how PHP actually become like this type of language and bottleneck, which is hard to scale. +[688.72 --> 694.86] So, like, imagine every time you write that golden-link application, which say, let's say, doing some endpoint on HTTP. +[695.32 --> 705.20] Every time you've been getting an HTTP request, imagine that you have to bootload your application from the disk, start it, answer this request, and then kill this application. +[705.32 --> 707.46] And do it over and over and over for every request. +[707.88 --> 709.54] This sounds super expensive, right? +[709.54 --> 714.62] Well, that's how PHP has been working for 26 plus years. +[715.00 --> 720.16] And it's quite amazing that you have the tag, which quite literally restarts on every request. +[720.64 --> 724.70] And it still kind of powers, like, the pretty much majority of backend on internet. +[725.20 --> 726.98] Well, I mean, public backend, let's say. +[727.10 --> 727.20] Right. +[727.54 --> 731.06] So the idea was actually quite simple. +[731.14 --> 732.40] Let's just remove this overhead. +[732.40 --> 739.90] I mean, when I started working with Roadrunner, I started working with actually a protocol just making communication between two languages. +[740.02 --> 741.74] And the first example was quite simple. +[741.88 --> 748.02] Okay, we have, let's say, a function in Golland to do, let's say, some heavy math, which on PHP might be not optimal. +[748.02 --> 752.90] And I have this, like, highly OP-strict code in PHP. +[752.90 --> 759.82] And by the way, like, modern PHP is all about, like, OP, strict types, annotations, attributes. +[760.06 --> 765.94] Like, it's all very, very similar to Java these days, except it's free, open, and very easy to learn, let's say. +[765.94 --> 774.12] So then I just tried to make a call from PHP to Golland using this, like, internal socket or Unix socket RPC call. +[774.42 --> 775.50] And it did work. +[775.98 --> 778.74] And then I tried to make a very stupid experiment, actually. +[778.82 --> 783.34] I tried the native PHP library for RabbitMQ to push message. +[783.88 --> 787.04] And I used the Golland library for RabbitMQ to push message. +[787.34 --> 790.56] But with additional this RPC overhead from PHP. +[790.56 --> 802.06] And we ran some tests, and we found out that the PHP to Golland bridge to RabbitMQ works, like, not, like, margin much, but, like, few percent farther than a native solution. +[802.36 --> 804.08] And it was, like, that's weird. +[804.76 --> 806.32] This shouldn't be happening, right? +[806.78 --> 812.30] And this kind of led to the idea that, like, PHP is, like, a very beautiful language to model business processes. +[812.64 --> 817.94] Not, like, high-scale I.O. operations, like traffic management or, well, ingresses. +[818.46 --> 819.40] It's single-threaded. +[819.40 --> 824.32] And it's very dummy, like, in terms of, like, you can go left, you can go right in some cases. +[824.66 --> 828.18] You still can shoot yourself to the foot, but these days it's much harder. +[828.56 --> 834.26] But it's very good to have good libraries to explain, like, permission models, document mapping, data mapping. +[834.44 --> 843.02] If you'll see how to work with mapping JSON in Golland and PHP, you'll definitely see a major difference in favor of PHP because it's, well, dynamic language. +[843.02 --> 855.58] And the goal, like, on the other side is beautiful to manage, like, all of these long-porting connections, sockets, like, retries, restarts, delays, all of this fun stuff with PHP just by the definition by model can't. +[855.58 --> 861.76] So then we just tried to create the method which has been invoking code from PHP and worker pools. +[861.76 --> 867.96] So you have, like, hot processes of PHP, which are already in memory, like, let's say, one per your CPU core. +[868.18 --> 872.24] And then you just ask one of them, just do this payload, do this work for me. +[872.40 --> 873.74] You don't kill your application. +[873.90 --> 874.88] You don't restart it. +[874.96 --> 876.04] Like, you have no overhead. +[876.04 --> 881.48] And when we did this code, well, it was working, like, 11 times faster than native approach. +[881.66 --> 884.96] So we created HTTP layer at top called Roadrunner. +[885.36 --> 888.14] And it's been, well, kind of with us since then. +[888.18 --> 892.94] And we haven't written a single application without this model probably, like, since 2019. +[893.82 --> 893.96] Wow. +[894.50 --> 895.40] Yeah, pretty amazing. +[895.90 --> 899.04] So who is this for? +[899.68 --> 903.80] Is it the Go developer who has to work with PHP or the PHP developer who has to work with Go? +[903.80 --> 906.68] Or who are you targeting with this approach? +[906.98 --> 907.78] I guess both. +[908.44 --> 909.04] Both of them. +[909.62 --> 911.78] Well, it's actually a very good question. +[911.94 --> 914.20] I mean, the main auditor is obviously PHP people. +[914.38 --> 926.24] Because what the main idea of Roadrunner is, like, you can take these complex aspects of, like, queue load balancing, HTTPS traffic, temporal gRPC, and you're going to make them boring for these developers. +[926.66 --> 927.56] But just out of the box. +[927.64 --> 928.62] You want gRPC, sure. +[928.78 --> 929.72] Just plug and play. +[929.80 --> 931.58] You want a temporal, sure. +[931.70 --> 932.82] Like, it's already here. +[932.88 --> 933.60] Just make it work. +[933.60 --> 934.64] You don't need to install anything. +[935.04 --> 937.78] It basically manages the complex stuff for you. +[937.78 --> 945.34] But at the same time, it's kind of for the Go link engineers who typically work in pair or on the same team as PHP engineers. +[945.84 --> 953.10] Because this is application server, like, it's very easy to intercept and modify the requests and calls which you do with PHP. +[953.10 --> 957.28] So, like, you can add your own validations, like, author indications. +[957.28 --> 964.16] And all obviously going to work much faster and possibly, like, with much deeper integration with modern, like, cloud native tools. +[964.26 --> 969.06] You have metrics, you know, readiness, healthy endpoints. +[969.06 --> 974.10] Like, all of the stuff you need to make application, basically, like, native to the current environments. +[974.10 --> 980.64] But obviously, the first target auditorium, well, it's just engineers and companies who are just trying to write scalable code. +[981.16 --> 985.08] But at the same time, don't necessarily want to hire, like, 10 Rust engineers. +[985.42 --> 994.40] Like, it's more like a balance between price for the engineer and how fast you can find them and the performance and quality of the software you create. +[994.40 --> 998.16] Okay, so this is as much a technical decision as it is a business one. +[998.32 --> 1005.40] Well, in a long term, yes, because how many startups you've seen which come to the point we're going to scrap out PHP and move to something else? +[1005.52 --> 1006.28] It's been a bunch. +[1006.70 --> 1006.80] Right. +[1007.16 --> 1009.20] Facebook invented their own language, you know. +[1009.30 --> 1009.52] Right. +[1009.88 --> 1010.26] Hack. +[1010.66 --> 1015.30] And we've contacted their, there are some Russian competitor who did the same thing, you know. +[1015.30 --> 1017.64] So it's kind of, like, become so expensive. +[1017.64 --> 1021.64] So you even have to jump in and make your own compiler for this stuff. +[1021.64 --> 1028.66] And we can just move this line when you have to move from one tag to another, ideally up to infinity. +[1028.96 --> 1032.06] Just if you need something fast, do it in other language. +[1032.20 --> 1035.18] I mean, it's all about microservices these days and C-Bit applications. +[1035.46 --> 1037.44] So, like, you're no longer stuck to one language. +[1038.02 --> 1040.64] But at the same time, well, you're a startup. +[1040.86 --> 1048.02] You're trying to integrate with a few providers, and you need to create 12, 15 API endpoints. +[1048.70 --> 1050.64] Like, who are you going to be using for that? +[1050.64 --> 1053.62] Do you really want to hire senior engineers who's going to be doing that? +[1053.68 --> 1060.20] Or you can just use senior PHP engineers, which is much higher, like, availability since it's so old language. +[1060.58 --> 1064.04] We're going to do the same thing, which is going to work the same on the same performance. +[1064.26 --> 1067.50] Well, it's just going to be done easy because you can source people easier. +[1068.34 --> 1068.74] Yeah. +[1068.74 --> 1085.92] And at the same time, if you wanted to do some pretty hard work or some low-level stuff, you can easily write a simple plugin and plug it into Roadrunner, compile it, and solve your needs with that. +[1085.92 --> 1089.38] So, we also write our own plugin system. +[1089.60 --> 1094.58] Because, so, initially, we wanted to use, like, Golang native plugin system. +[1094.68 --> 1097.14] You know, it works only on Linux at the moment. +[1097.14 --> 1103.82] So, we waited a little for the Windows support, but I guess it doesn't seem to happen. +[1104.16 --> 1107.16] So, the Windows support for Golang native plugins. +[1107.16 --> 1121.86] But we wrote our own plugin system called Endure, which is suitable for, like, plugging all of the pieces of Golang, combining into one part, like, initializing it, starting to serve it, stopping. +[1122.18 --> 1125.26] So, building a tree with all of those plugins. +[1125.26 --> 1141.84] So, you can, for example, if you write once a configuration parser, for example, if you write, like, a parser from the YAML, for example, you don't need to write or copy or create some SDK to bring this part into every plugin. +[1141.84 --> 1145.00] So, you can just request this init function. +[1145.72 --> 1155.80] And Roadrunner will take care about finding this dependency, initializing it, topologically sorting the graph, and providing this initialized dependency for you. +[1156.04 --> 1162.48] You only need to just, okay, configuration, please give me the gRPC section or give me some other section. +[1162.48 --> 1178.52] So, if something goes wrong, Roadrunner will take care about this, of course, and, like, provide you, like, nice but unreadable for PHP users message about, like, going to some panic or some errors, something like this. +[1178.90 --> 1187.46] And for the PHP, all complexity for the PHP user is to properly define what do you need in the configuration. +[1187.46 --> 1197.50] Like, you have a configuration, you need HTTP, so you just enable HTTP section, put your configuration, and Roadrunner will remove all other plugins from the tree. +[1198.02 --> 1199.86] Like, it won't even start. +[1200.50 --> 1204.02] It just runs your section for your needs, like HTTP. +[1204.56 --> 1205.30] Or gRPC. +[1205.46 --> 1211.46] Or you can write your own plugin, put your section in the configuration, build with our tool called Velux. +[1211.46 --> 1218.30] Velux is a tool which helps you to build Roadrunner with your own custom plugins based on GitHub. +[1219.38 --> 1232.20] So, it's all of this heavy, complex stuff was moved to the GoLang part, and the open, nice things moved to, like, PHP. +[1232.20 --> 1240.82] That's actually quite an exchange of knowledge, because this first container actually came, like, as architectural pattern in most of the PHP applications. +[1241.40 --> 1246.56] Because if you've seen .NET, Java, PHP applications, you have a ton of classes, interfaces that inherit each other. +[1246.94 --> 1249.78] They, like, use class declarations and etc. +[1249.90 --> 1251.98] So, like, you can't work without container. +[1252.20 --> 1254.80] And it being dependencies, you need to manage all of this stuff. +[1254.80 --> 1260.80] So, like, essentially, we use this idea inside of Roadrunner, but obviously, it's the GoLang favor tool. +[1260.92 --> 1263.46] Initiate it, lays it all correctly in the correct order. +[1264.10 --> 1267.30] And answering the question, like, how, like, it's a large tool already. +[1267.70 --> 1273.04] Well, it's kind of not, because it's just a container with a bunch of, like, CLI tools and instruments. +[1273.52 --> 1276.80] And the rest is just kind of, like, independent projects and plugins. +[1277.08 --> 1282.80] So, we can add them without kind of influencing on each other or worrying to break the tool. +[1282.80 --> 1284.76] Do you want Roadrunner with HTTP layer? Sure. +[1284.92 --> 1287.48] If you don't, well, disable the plugin and build it. +[1287.58 --> 1290.98] It's going to be exactly the same thing, just less memory to manage. +[1303.92 --> 1307.74] This episode is brought to you by our friends at FireHydrant. +[1308.04 --> 1310.80] FireHydrant is the reliability platform for every developer. +[1310.80 --> 1315.02] For incidents, they impact everyone, not just SREs. +[1315.26 --> 1322.90] They give teams the tools to maintain service catalogs, respond to incidents, communicate through status pages, and learn with retrospectives. +[1323.26 --> 1328.68] What would normally be manual, error-prone tasks across the entire spectrum of responding to an incident, +[1328.94 --> 1332.18] they can all be automated in every way with FireHydrant. +[1332.18 --> 1337.52] They have incident tooling to manage incidents of any type with any severity with consistency. +[1338.04 --> 1341.18] Declare and mitigate incidents all from inside Slack. +[1341.56 --> 1347.90] Service catalogs allow service owners to improve operational maturity and document all your deploys in your service catalog. +[1348.48 --> 1355.86] Incident analytics allow you to extract meaningful insights about your reliability over any facet of your incident or the people who respond to them. +[1355.86 --> 1360.18] And at the heart of it all, incident runbooks, they let you create custom automation rules, +[1360.40 --> 1365.64] convert manual tasks into automated, reliable, repeatable sequences that run when you want. +[1366.02 --> 1370.02] You can create Slack channels, Jira tickets, Zoom bridges instantly after declaring an incident. +[1370.50 --> 1373.08] Now your processes can be consistent and automatic. +[1373.08 --> 1375.22] The next step is to try it free. +[1375.36 --> 1379.74] Small teams, up to 10 people, can get started for free with all FireHydrant features included. +[1380.06 --> 1381.48] No credit card is required. +[1381.92 --> 1384.10] Get started at FireHydrant.io. +[1384.42 --> 1386.38] Again, FireHydrant.io. +[1403.08 --> 1406.02] Let's dive in a little bit into the weeds, if we will. +[1406.36 --> 1408.66] Are you shipping a binary? +[1409.04 --> 1410.88] Are you interpreting PHP? +[1411.40 --> 1415.92] If I'm a developer, what does my experience look like? +[1416.06 --> 1420.66] So from a PHP perspective, we're trying to do the less invasive work possible. +[1420.86 --> 1424.08] You literally don't need to do anything to make it work from a PHP side. +[1424.44 --> 1425.22] There's no extensions. +[1425.58 --> 1428.20] There's no special CLIs, interpreters, nothing. +[1428.20 --> 1435.84] What we actually do, we do the 20, 30 years old approach, which still drives all of the applications. +[1436.38 --> 1437.62] We manage the worker pool. +[1437.90 --> 1443.38] So what WordRunner does, it actually uses the default PHP interpreter, which is a binary, +[1443.94 --> 1449.08] invokes it with your application, and then keeps it in memory in a pre-warmed state. +[1449.08 --> 1457.48] And when the request payload comes, which can come for HTTP endpoint, task queue, temporal workflow, +[1458.04 --> 1463.14] gRPC, whatever, you name it, you just send this payload to PHP and wait for it to complete. +[1463.46 --> 1468.76] But over the lifetime, the only main difference for the engineer, which is, well, for some engineers, +[1468.86 --> 1474.38] it's quite hard, is to realize that your application leaves longer than just a single request. +[1474.38 --> 1481.24] And you can't just have a global variable counter, which is going to plus plus and expect it's going to be zero on the next request. +[1481.84 --> 1482.80] That's the only difference. +[1482.88 --> 1489.22] But besides that, it's just the same exact PHP, same exact extensions, configurations as you typically use. +[1489.54 --> 1491.56] It's just managed in a bit different flavor. +[1491.90 --> 1495.98] But surprisingly, WordRunner knows nothing about the PHP. +[1496.66 --> 1500.04] So WordRunner is not binded especially to a PHP. +[1500.04 --> 1504.80] It just runs some command you specify in your configuration. +[1505.02 --> 1506.64] So you can do it in Python. +[1507.22 --> 1513.16] You can even run Golang inside the Golang, like Golang inside the Golang workers. +[1513.50 --> 1517.48] So the main purpose for the WordRunner is to manage the process. +[1518.06 --> 1519.52] Who will be in this process? +[1519.80 --> 1522.40] So for the WordRunner, it doesn't care about it. +[1522.54 --> 1524.30] It cares about the protocol. +[1524.74 --> 1526.84] So protocol is language agnostic. +[1526.84 --> 1532.60] So I saw a project, some guy wrote this, implemented this protocol in Python. +[1533.04 --> 1536.02] Anton showed me like some time ago. +[1536.22 --> 1539.08] So it's like Python running inside the WordRunner. +[1539.36 --> 1544.42] So because it's the same, pretty the same model, like the one-threaded model in the Python. +[1544.60 --> 1548.16] So some guy wrote this in and yeah, so it works. +[1548.62 --> 1551.30] This sounds like a process manager, if you will. +[1551.30 --> 1561.74] So if you wanted to, you could have Python, obviously, as you just mentioned, Ruby, obviously, PHP natively, and whatever else other sort of interpreted languages you want. +[1561.88 --> 1567.00] Like even like pre-compiled things, it sounds like you can just have in there and then basically just invoke it. +[1567.26 --> 1571.42] In theory, you can take 20 years old Perl file and run it in the WordRunner. +[1571.42 --> 1574.96] But if you want to do it, obviously, that's a good question. +[1575.34 --> 1580.36] But yes, I mean, I think the only mention, main mention of PHP is actually the title of WordRunner. +[1580.46 --> 1582.02] It's a PHP application management server. +[1582.24 --> 1586.42] But besides that, there's nothing which actually ties it specifically to single language. +[1586.78 --> 1592.24] Except that the single language has a largest SDK code base to communicate with all the features from WordRunner. +[1592.50 --> 1594.58] But they're just a nuance, quite frankly. +[1594.58 --> 1607.08] Okay, so if I'm used to working with single process PHP style application, I think you touched on this earlier, where it's like I'm restarting the world traditionally when I'm dealing with PHP. +[1607.32 --> 1611.46] One request just restarts the world and basically it's like everything is like a new, right? +[1611.46 --> 1625.22] In this world where instead there's a worker pool, there's a process management happening, do you find that developers have to sort of have a mindset shift to basically to think, okay, there's not just one process here, there's multiple. +[1625.54 --> 1629.58] So does that change sort of the nature of how they program? +[1629.70 --> 1634.60] Do they switch from programming, I guess, a single process to now having multiple processes to contend with? +[1634.94 --> 1638.06] Well, I mean, when PHP runs on scale, you still have multiple processes. +[1638.06 --> 1644.06] They just create it on demand, but you still have like 500 PHP processes running the request. +[1644.42 --> 1648.26] But there is definitely the kind of conception shift in people. +[1649.06 --> 1651.38] It's been much harder earlier, like even a year ago. +[1651.56 --> 1657.68] But now, like the main reason why it's easy these days is because very minimal amount of people write on pure PHP. +[1658.26 --> 1661.52] Unlike Golang, where most of the time you write on pure Golang because you can. +[1661.52 --> 1669.48] Like you may be like using some small HTTP overlay framework, but mostly you're going to be using SPL functions, let's say, to do most of your business stuff. +[1670.04 --> 1671.26] In PHP, it's completely different. +[1671.58 --> 1677.78] You have Symfony or Laravel or our infrastructure framework, Spiral, and they're all managed for you. +[1677.78 --> 1692.60] So the beautiful thing is that all of these frameworks over the time, well, we created our own specifically for this purpose, but all other frameworks, they actually do upgrades and patches, which solves all of these kind of thinking nuances for the engineer. +[1692.98 --> 1696.40] They reboot some parts of the services, like clean up the caches and et cetera. +[1696.40 --> 1705.16] So if you're using modern framework and if you're using Roadrunner, you most likely won't even notice a difference in like 90% of your activities. +[1705.44 --> 1710.86] It's still going to be some nuances, but most of them are already known and have already been solved. +[1711.08 --> 1716.60] If you're using framework which specifically built for Roadrunner like ours, nothing is different for you. +[1716.66 --> 1723.10] Just write code and it's all been managed and like status managed in memory and all completely reset it. +[1723.10 --> 1729.60] So would a PHP developer ever even need to know what's under the hood, what's running their processes? +[1729.92 --> 1734.00] I assume they'll never really come in contact, unless they want to, like come in contact with any go whatsoever. +[1734.26 --> 1737.80] Well, it's like owning a car and being able to drive a car, right? +[1737.90 --> 1744.30] If you have Roadrunner on PHP and default PHP SDK, you can do many, many like wonderful things. +[1744.30 --> 1753.52] You can make PHP respond in like 50 microseconds, like run on the all realms and all the queries, use your PC and like use and all of this stuff. +[1753.78 --> 1757.92] But if you also know Golang, you can do even crazier stuff. +[1758.12 --> 1765.20] Like you can take, for example, machine learning model from Keras, compress it, put it into the Golang SDK and then invoke from PHP. +[1765.38 --> 1769.46] And to the end, you now have machine learning compared to PHP via Golang Bridge. +[1769.46 --> 1774.06] So like you can do these crazy things if you know multiple languages, specifically Golang. +[1774.26 --> 1778.50] And Golang is beautiful because it's so, I mean, it's quite easy to learn it, right? +[1778.56 --> 1783.46] It doesn't have many of these legacy overhead and nuances you have to remember. +[1783.90 --> 1788.98] So it's not required, but if you do, well, you essentially are a superhero in PHP world. +[1789.88 --> 1791.94] Yeah, and also your Hardware project, Anthony. +[1792.20 --> 1794.48] Well, yeah, that's another example how you can use it. +[1794.52 --> 1797.46] It's also written here in Roadrunner and PHP combination. +[1797.46 --> 1801.78] I can show it when it's going to be the right time up to you, Johnny. +[1802.62 --> 1809.88] Well, you know, given that most of our users are going to be listening, not viewing this, we may not be able to show them much. +[1810.00 --> 1811.32] I'll try to walk it through, yeah. +[1811.52 --> 1817.40] I do want to sort of understand, obviously this is an open source project and folks are contributing to it. +[1817.76 --> 1826.02] I do want to understand sort of what are some of the hardest challenges that you encountered, like while sort of coming up with this model. +[1826.02 --> 1833.90] Obviously, you're running a process manager, you have to worry about inter-process communication, you have to worry about sort of how to keep things in memory efficiently and all that stuff. +[1834.02 --> 1839.20] You know, I'm curious as to sort of what are some of those biggest challenges and perhaps that you're still facing. +[1839.20 --> 1844.50] I think, I mean, about the current challenges, what I can speak, I don't think we're facing like a lot of them now. +[1844.64 --> 1849.82] But when we started working on this tool, it's been a number of interesting things to solve. +[1849.90 --> 1854.42] Well, number one, we had to create a protocol to communicate within two different languages. +[1854.42 --> 1858.28] And I had to work over pipes, over Unix sockets and TCP sockets. +[1858.28 --> 1867.64] So it's been like, okay, how are we going to create a low-level IPC protocol if you don't want to jump to like shared memory or all of these things? +[1868.24 --> 1874.82] This being not like a hard conceptual problem because you can always like, default protocol is like net string. +[1874.94 --> 1881.12] You have the lengths of the message with like in a fixed head size and then you have the payload board. +[1881.12 --> 1884.66] So it's quite easy to just read the package between languages. +[1885.08 --> 1893.68] Then obviously it was a problem with doing process manager because PHP sometimes tends not to start if you don't send it right parameters or it might crash. +[1893.84 --> 1897.28] If again, you send an invalid payload, I mean, doesn't do it anymore. +[1897.40 --> 1902.22] It's been like a demo builds and obviously raise conditions on Goal inside. +[1902.70 --> 1906.20] Oh, that was horrible because you can't just write a process manager. +[1906.36 --> 1908.16] No one just writes a process manager. +[1909.36 --> 1910.50] Sorry for the reference. +[1910.50 --> 1912.36] You also need to collect the stats. +[1912.52 --> 1914.64] You also need to watch for the process to restart. +[1914.86 --> 1917.80] You need to collect the SDR, right? +[1917.84 --> 1922.48] You need to understand how many times you invoked it, when it was started. +[1923.02 --> 1925.84] It thinks like, okay, let's check how long this process exists. +[1926.42 --> 1930.14] And you call, let's say, time now and you immediately kill your performance, right? +[1930.16 --> 1934.10] Because it's doing a syscall, well, back in the day, something like that. +[1934.38 --> 1939.40] Or, for example, you have a request coming, but at the same time, the PHP process runs out of the memory. +[1939.40 --> 1943.08] So what's going to happen is going to fail. +[1943.54 --> 1944.56] How are you going to start? +[1944.56 --> 1952.28] It's going to be, it's been so many like little integrational hell and edge cases for this type of work. +[1952.44 --> 1957.92] But like, eventually, like once we jumped from this part, what we managed is another part of the hell. +[1958.30 --> 1963.10] When you have a single server, which has HTTP endpoint, everything is quite easy, right? +[1963.14 --> 1968.38] You have request, pack it into the binary form, send to PHP, and well, Bob's your uncle. +[1968.38 --> 1975.76] But when you're doing HTTP, and then you also want to manage the queue process manager, like cooperate with RabbitMQ. +[1976.30 --> 1983.42] But what happens, what if you have HTTP, which runs a PHP worker, which sends data to queue? +[1983.42 --> 1994.88] So now you have two plugins, which not only have to work, they also have to be created in a correct order, wait for each of them to properly connect, and only then make it work. +[1995.06 --> 2000.08] So like, that was a part of like scratching the head a lot, because it was so hard to solve. +[2000.48 --> 2007.22] With all these dependencies and plugins and hidden dependencies, because like PHP worker can theoretically do anything which it wants. +[2007.22 --> 2017.24] It's like, it can call to queue, and it can invoke HTTP endpoint through the roadrunner to itself, which is, well, I'm not even sure what's going to happen in this case. +[2017.88 --> 2020.02] But that's why we created the container. +[2020.22 --> 2023.46] That's what we contributed to eventually Endure, which solves all these problems. +[2023.72 --> 2029.38] And now, like, there is barely no integrational hell between like plugins and Roadrunner. +[2029.50 --> 2033.10] They all like have interfaces, very easy to connect them together. +[2033.38 --> 2036.54] It's basically become a framework for the application server. +[2036.54 --> 2050.56] So if you have this deployed, is it recommended basically that you have basically a single tenant kind of situation whereby you don't want sort of processes from multiple parties that are not really associated with the same, say, the same company, right? +[2050.56 --> 2062.54] You don't want to run this as some sort of, you know, open to all multi-tenant sort of system, because you could have one process sort of peeking in into what's happening with another process within this sort of, within this world. +[2062.54 --> 2066.58] Or is it isolation between these things running all in the same system? +[2066.88 --> 2067.96] Well, it could be. +[2068.30 --> 2071.36] We can isolate it by many different ways. +[2071.46 --> 2075.38] We can isolate them by running with different permission models in PHP. +[2075.50 --> 2081.26] You can literally forbid most of the functions, which, well, some engineers and hackers will still be able to bypass. +[2081.88 --> 2085.30] You can run them in different user groups and user in different memory spaces. +[2085.30 --> 2091.28] I mean, right now we don't run it in like shared fashion, like old-fashioned shared hostings. +[2091.50 --> 2094.52] It's mostly suitable for the most classic approach now. +[2094.62 --> 2095.38] You have a container. +[2095.86 --> 2102.10] Within this container, you have your application, your APIs, or like other service functions of this application. +[2102.46 --> 2105.30] And it's fully self-contained inside this container. +[2105.30 --> 2110.80] If you work on a multi-tenant model, you don't run 10 different PHP scripts with different users. +[2110.98 --> 2116.50] You solve the multi-tenancy on the, well, application design level within your application domain. +[2117.12 --> 2117.90] I hope it makes sense. +[2119.00 --> 2121.14] But I've been thinking about this problem for a while. +[2121.72 --> 2125.54] Basically, in short, if you want to do the multi-tenancy, you're doing it in your application code. +[2125.78 --> 2132.40] Roadrunner is designed to actually work the best in modern, like, Docker environment or container-based environment. +[2132.40 --> 2135.84] It's a single application per application instance, let's say. +[2136.08 --> 2139.14] Single instance of Roadrunner per or single instance of application, sorry. +[2139.34 --> 2139.78] Gotcha. +[2140.20 --> 2140.40] Yeah. +[2140.56 --> 2146.72] And I guess, as for me, it was a lot of challenges in, like, to write all these things. +[2147.10 --> 2154.40] Because basically, Roadrunner, as you can see in, like, Roadrunner repository, is just a CLI interface. +[2154.40 --> 2158.44] So it's just a Roadrunner server, Roadrunner workers command. +[2158.80 --> 2163.62] But everything is hidden under the main, I guess, three parts. +[2164.20 --> 2167.14] Those parts are, it's Azure, as Anton mentioned. +[2167.42 --> 2169.72] It's like, it was a surprise for me. +[2169.84 --> 2172.70] Because, you know, in GoLang, we don't have any containers. +[2172.70 --> 2184.60] So we, because we just don't need them to get something dependency or, so I heard about some project in Uber, like Uber FX or Google Wire, I guess. +[2185.04 --> 2187.70] But I don't think it's much popular solutions. +[2188.72 --> 2192.22] So in my project, I never, I worked with a container. +[2192.22 --> 2198.42] But BHP is everything about containers and everything about dependency injection and so on. +[2198.44 --> 2205.44] So we have to write such algorithmic container, which, like, mutate based on the configuration. +[2206.04 --> 2209.04] So it's not like, please give me some dependency. +[2209.66 --> 2212.60] It's about, I have a configuration. +[2213.18 --> 2217.36] I have a set of, I guess, at the moment we have 20 plugins or more, I guess. +[2217.50 --> 2219.96] I don't know how much exactly. +[2219.96 --> 2227.26] But you provide a bunch of plugins, like provide a configuration and say, okay, now build this. +[2227.76 --> 2231.44] Build this properly, managing the connection, as Anton says. +[2231.84 --> 2239.08] Like, if you provide, for example, initialize, like, a logger and logger needs a configuration. +[2239.56 --> 2246.84] So you have to properly topologically sort all the things to provide first, to initialize first, like, configuration. +[2246.84 --> 2250.22] Then provide this pointer to logger. +[2250.86 --> 2258.52] So the GoLang race flag won't help us here because it's so distributed all over the plugins. +[2258.52 --> 2264.08] So you have to manage and see every, like, race condition by yourself. +[2264.40 --> 2269.16] So you have to be very careful with writing all the things. +[2269.50 --> 2273.02] But we hide all this complexity inside the Endure. +[2273.36 --> 2278.74] And you have to be sure that provided dependency is, like, concurrent free. +[2278.96 --> 2279.12] Yeah. +[2279.12 --> 2282.48] So you can't, like, use it from the different threads. +[2283.18 --> 2285.64] But the second part is also Gorich. +[2286.08 --> 2287.10] It's a protocol. +[2287.58 --> 2290.02] It's in Gorich version 1 and version 2. +[2290.14 --> 2293.12] It was, like Anton said, it's a very basic protocol. +[2293.50 --> 2297.24] Like, I guess, 14 bytes of, correct me if I wrong, Anton. +[2298.02 --> 2299.68] 14 bytes or 18 bytes. +[2299.68 --> 2303.18] So the first person had 17 bytes heater for whatever reason. +[2303.38 --> 2304.66] 17 bytes, yeah. +[2305.08 --> 2311.66] It's, like, a few flags, like, payloads in Big Indian, Little Indian, and payload. +[2312.26 --> 2317.02] But imagine the situation if you pass your payload over the pipes, for example. +[2317.46 --> 2326.26] You don't have any mechanism to CRC or to check if this payload is correct, passing it from the one side of the wire to another. +[2326.26 --> 2334.14] But Roadrunner and Gorich version 3 has a protocol based on IP protocol. +[2334.28 --> 2347.38] So I've wrote, recently, a few protocols, like TCP IP, IP protocol, and combined them all together to have a proper protocol for communicating with PHP parts. +[2347.52 --> 2349.56] So we have, like, a synchronic support. +[2349.86 --> 2352.88] We have a header length, like in IP protocol. +[2353.04 --> 2354.94] We have variable length options. +[2354.94 --> 2361.94] So we have all this funny stuff to extend it and not to break it from version to version. +[2362.84 --> 2368.74] And I guess the third part is decay, which contains all these worker pools. +[2369.08 --> 2375.56] And this is complicating stuff because, you know, you have to, for example, imagine you have a HTTP plugin enabled. +[2375.92 --> 2378.48] The user want to get a statistic about the workers. +[2378.48 --> 2382.08] So the one approach is to stop the world. +[2382.70 --> 2390.76] Okay, stop all the HTTP requests and get all the statistics about, like, get all the pointers to workers. +[2391.22 --> 2395.02] Get the stat, print it with some format and show to the user. +[2395.02 --> 2406.50] And another approach is to have some shared place where you can safely get at any time, like, without logs, even, this pointer and to provide to user. +[2406.72 --> 2412.32] But when the restart happens or some issue happens, it will log only in this case. +[2413.20 --> 2415.34] So it's like statistics in the Roadrunner. +[2415.34 --> 2417.02] It's basically free for users. +[2417.02 --> 2419.80] It's not interrupt the actual request. +[2420.26 --> 2422.64] I mean, not HTTP, not jobs. +[2422.94 --> 2424.76] So it's none of them. +[2425.02 --> 2428.36] It's quite funny how it actually, how we jump into this edge cases. +[2428.54 --> 2434.92] Because most of our users are actually, like, already mature PHP applications and, like, large, large startups. +[2435.20 --> 2437.38] And, like, they don't play with, like, 10 requests. +[2437.50 --> 2439.80] Like, okay, I mean, we just bump a few millions a day. +[2439.80 --> 2441.12] And we see this bug. +[2441.22 --> 2442.08] So what is bug about? +[2442.20 --> 2442.90] Oh, you know what? +[2443.10 --> 2447.98] Over time, it's, like, things which you could never imagine while developing the thing. +[2448.26 --> 2452.34] But when they catch them, thankfully, the user reports them. +[2452.84 --> 2456.46] It's just easy to, like, it's kind of very easy to see how it works on scale. +[2456.74 --> 2456.92] Yeah. +[2457.34 --> 2462.42] And basically, imagine a situation when you have a completely fresh scheduler in Golink. +[2462.60 --> 2465.18] So I guess you saw this ticket recently. +[2465.64 --> 2469.00] It's about to write a completely fresh scheduler. +[2469.00 --> 2480.40] So we started working with that, I guess, a year before to provide a completely fresh scheduler to schedule jobs inside the Roadrunner. +[2480.62 --> 2484.36] So we don't finish it yet, but we're continuously working on it. +[2484.88 --> 2490.12] So to provide, like, a binary HIPs algorithm to sort by priorities all of these jobs. +[2490.12 --> 2500.58] Like, for example, if you have an urgent jobs to execute, you have to set the priority one, and it will be sorted and scheduled properly. +[2500.58 --> 2512.02] So there are a lot of such things, very complicated things, which you should work inside the Roadrunner and hide all of this complexity for users just to specify a few values in the configuration. +[2512.26 --> 2514.94] Just, okay, I don't need a gRPC, so it's removed. +[2515.08 --> 2517.96] But all of this complexity hidden under the hood. +[2517.96 --> 2522.32] That's why it's hard to create very nice user-friendly APIs. +[2522.74 --> 2527.36] You know, there's a lot going on under the hood, but, you know, like, you're making it easy for people to actually use. +[2527.44 --> 2527.98] That's the beauty. +[2528.28 --> 2531.32] What can be easier to make just a Hello World endpoint? +[2531.56 --> 2532.36] Because it's so true. +[2532.36 --> 2537.66] Yeah, well, the loan balancer, Kubernetes cluster, control plane, and a few other things. +[2538.06 --> 2538.70] That's easy. +[2539.54 --> 2539.90] I know. +[2540.08 --> 2543.66] Man, things have changed over the last decade or two. +[2544.22 --> 2554.44] So I do want to switch gears a little bit here, sort of to understand, like, if someone wanted to contribute, because it sounds like there's still a lot of awesome ideas sort of at play and coming to the project. +[2554.44 --> 2560.46] If someone wanted to contribute to that, like, what would basically be sort of a safe expectation of them? +[2560.54 --> 2562.68] Should they know how to work in PHP? +[2562.98 --> 2564.96] Should they know how to do Go? +[2565.42 --> 2567.18] Is there work for both sides of the fence? +[2567.32 --> 2569.08] Like, what do you need to contribute to this project? +[2569.50 --> 2571.48] Basically, I don't know PHP at all. +[2572.00 --> 2578.10] I started Googling, like, okay, Anthony, I need some test script. +[2578.64 --> 2581.04] Could you please write it for me? +[2581.04 --> 2586.18] Or I need to, like, a for loop to write or some variable. +[2586.38 --> 2589.60] So I need to Google how to write a for loop in PHP. +[2589.78 --> 2590.06] For loop. +[2590.20 --> 2594.76] And this is great, because I don't need to involve in the, like, PHP part. +[2595.00 --> 2599.78] I can concentrate only to improve the GoLang user experience. +[2599.78 --> 2606.02] And, like, to contribute to the roadrunner, it depends on who are you as a developer. +[2606.18 --> 2608.94] You are a PHP developer or you are, like, GoLang developer. +[2608.94 --> 2612.32] If you're, like, a GoLang developer, it's very easy. +[2612.42 --> 2622.84] You just need to go into the rootrunner issues, find an issue marked, like, help needed or easy to resolve, like, some entry-level issues. +[2622.84 --> 2634.70] Or if you want to, like, contribute to PHP part, I guess it's a spiral framework, which you can, like, also have a lot of tickets to improve our PHP part. +[2634.80 --> 2638.68] And you don't need to know both sides of this at the same time. +[2638.76 --> 2640.96] You only need to know, like, PHP or GoL. +[2641.12 --> 2645.42] Or if you want to contribute both, usually, yeah, you need to know both. +[2645.68 --> 2646.58] But, yeah, it's... +[2646.58 --> 2648.58] Sounds like there's room for either side. +[2648.58 --> 2648.98] Yeah. +[2649.36 --> 2651.08] We're pretty friendly to contribution. +[2651.36 --> 2654.38] Like, on PHP, you can help us to improve SDKs and GoLang. +[2654.82 --> 2661.64] You can go as deep into the weeds as, like, SSL wish, like, dump little algos to run some crazy stuff. +[2661.72 --> 2668.00] Even Python developers can write a protocol version 3 and, like, or Ruby on Rails developer. +[2668.24 --> 2668.52] Yeah. +[2668.78 --> 2673.36] If you want, we will definitely accept the contributions to make it work for other languages. +[2673.36 --> 2688.90] This episode is brought to you by LaunchDarkly. +[2689.34 --> 2691.34] Fundamentally change how you deliver software. +[2691.80 --> 2692.68] Innovate faster. +[2693.04 --> 2693.64] Deploy fearlessly. +[2694.18 --> 2698.46] And take control of your software so you can ship value to customers faster and get feedback sooner. +[2698.92 --> 2701.96] LaunchDarkly is built for developers but empowers the entire organization. +[2701.96 --> 2705.28] Get started for free and get a demo at LaunchDarkly.com. +[2705.64 --> 2707.44] Again, LaunchDarkly.com. +[2707.80 --> 2713.90] And by our friends at FlatFile, the leading data onboarding platform for teams who don't want to build yet another CSV uploader. +[2714.30 --> 2718.32] FlatFile's powerful, out-of-the-box solution takes the import burden off of your shoulders, +[2718.60 --> 2722.04] freeing you to solve bigger business problems and build products that people love. +[2722.30 --> 2726.34] Get to usable data faster so you can focus on what matters most to you and your business. +[2726.34 --> 2728.36] It is incredibly fast to set up. +[2728.40 --> 2732.16] Just write a few lines of code and get up and running in hours, not days or weeks. +[2732.50 --> 2733.64] It is framework agnostic. +[2733.78 --> 2739.30] Use the SDK to integrate FlatFile into any JavaScript application with support for all major frameworks. +[2739.68 --> 2741.64] Learn more and get started at FlatFile.com. +[2741.64 --> 2743.94] Again, FlatFile.com. +[2758.02 --> 2767.86] All right. +[2767.86 --> 2770.34] So, let's start with you, Anton. +[2770.60 --> 2771.30] What did you bring? +[2771.64 --> 2774.00] You only need 64 kilobytes of RAM. +[2774.32 --> 2774.74] For what? +[2775.20 --> 2775.88] Just for everything. +[2776.02 --> 2776.20] Ever? +[2776.74 --> 2777.60] Well... +[2777.60 --> 2779.04] Okay. +[2779.72 --> 2780.36] Prove it. +[2784.86 --> 2793.66] For the most of the stuff, I mean, I would just say, like, people have to, like, try to work on kind of more memory-efficient applications. +[2793.66 --> 2797.72] Because when you work with hardware, 64 kilobytes of RAM is a ton. +[2798.06 --> 2802.92] You know, you can make robot moves, blink eyes, go on the stairs, and do some other stuff. +[2803.48 --> 2806.56] And what can you do with 64 kilobytes of JavaScript application? +[2806.74 --> 2807.26] Tell me, please. +[2808.70 --> 2809.44] Maybe nothing. +[2809.84 --> 2810.42] Not much. +[2811.86 --> 2814.54] I have 64 gigs and even can't run a stream. +[2814.54 --> 2820.54] I mean, I will say that over the years, we've gotten more... +[2822.14 --> 2823.52] Well, let me put it nicely. +[2823.66 --> 2828.58] We don't worry very much about sort of CPU and memory and disk, right? +[2828.66 --> 2833.20] Things that used to be expensive, you know, like, you know, even like 20, 30 years ago, right? +[2833.36 --> 2834.44] Not so much now. +[2834.74 --> 2835.78] We take these things for bringing it down. +[2836.04 --> 2836.24] Yeah. +[2836.24 --> 2842.00] Yeah, but, like, if you know how to pack it down to this level, you can create much larger scalable applications. +[2842.22 --> 2848.42] Because when you create, let's say, the traffic filtering software or, like, VPN cores, right? +[2848.54 --> 2853.22] The things which actually, like, well, real IP, let's say, not just API endpoints. +[2853.78 --> 2855.70] That's where you have to optimize it. +[2855.76 --> 2861.34] And, like, knowing these basics and knowing that, yes, 64 kilobytes sounds like a small amount. +[2861.34 --> 2867.28] And it's a joke, which Bill Gates said back in the day, if someone don't remember, like, obviously he's wrong. +[2867.48 --> 2883.54] But if you realize how actually huge this amount, like, 64 kilobytes of stack on, like, some hardware chip can let you, well, to stream a ton of traffic, you know, and build something like a Netflix, build something like Starlink. +[2883.94 --> 2887.72] Because all these things which are doing great, great things, they have to be optimized. +[2887.72 --> 2891.96] You can't put, you know, 10 CPU server in space. +[2892.14 --> 2893.26] It's still going to work. +[2894.50 --> 2895.94] All right, all right, all right. +[2896.08 --> 2896.96] Valerie, what did you bring? +[2897.20 --> 2901.02] So my popular opinion is open source is a hard work. +[2901.54 --> 2905.26] In my opinion, it's much harder than some enterprise development. +[2905.82 --> 2908.90] Because I was involved in, like, in different enterprise projects. +[2909.24 --> 2912.70] And the flow is pretty much defined. +[2912.70 --> 2922.76] Like, if customer support has some ticket, it, like, can process it, send to specialists, like, quality assurance. +[2922.98 --> 2927.38] It, like, can test it, write, like, test cases it brings to you. +[2927.82 --> 2932.56] You can, like, see this ticket, fix the problem, run the tests, and so on. +[2932.56 --> 2941.70] But in open source, a lot of people think that they should not bother themselves to write a proper description of the issue. +[2942.04 --> 2944.90] It's like, I have a problem, please fix it. +[2945.16 --> 2950.08] Or one of my favorite issues is, like, nuts and question mark. +[2951.46 --> 2952.84] What does it mean? +[2953.40 --> 2955.12] So a lot of people... +[2955.12 --> 2957.84] Are they offering nuts or are they asking you if you want not? +[2957.84 --> 2969.72] But when I say, please describe what do you want, like, you want to not support or something else and close this ticket, the guy asked me, why are you so rude? +[2970.04 --> 2971.50] Why are you closing my ticket? +[2971.80 --> 2975.90] So, and there are a lot of such things in the open source development. +[2976.08 --> 2978.08] So you should handle a lot of things. +[2978.08 --> 2990.68] You should have various types of virtual machines to run on Ubuntu previous versions, Ubuntu current versions, Debian, FreeBSD, macOS, and Windows different versions. +[2990.94 --> 2999.12] So if someone send you a ticket, I guess some people think this is a joke, like open source development is like a joke for us. +[2999.12 --> 3000.14] So I don't know. +[3000.14 --> 3008.74] So it's like, write three, two or three, like lines of the description or do not provide like test cases. +[3009.16 --> 3013.68] So it's very complicated to figure out what do you really want. +[3014.70 --> 3019.68] So one guy asks me, like, the docker doesn't work on my machine. +[3020.34 --> 3021.36] Like, what? +[3021.68 --> 3023.82] So I have to fix the docker on the machine. +[3023.82 --> 3031.68] So, but you see, in CI, everything like brings from the scratch, like a docker installs from the scratch. +[3031.80 --> 3033.56] So you see everything works inside the docker. +[3033.96 --> 3035.46] It's something inside your machine. +[3035.64 --> 3038.54] But the guy said, no, it's a problem in your code. +[3038.82 --> 3040.22] So it's because I can't get around. +[3040.38 --> 3047.78] So it was so long and so boring, like communications with a lot of people to prove that this is not a joke. +[3048.20 --> 3050.66] It's a hard work, really hard work. +[3050.94 --> 3051.14] Yeah. +[3051.30 --> 3051.62] Yeah. +[3051.62 --> 3054.88] So, and actually all the court to write is public. +[3055.08 --> 3058.22] So like, because people are going to see it and going to blame you. +[3058.32 --> 3061.60] I mean, they're going to blame you anyway, but now they're going to have a reason. +[3062.22 --> 3062.60] Right. +[3062.72 --> 3063.16] Exactly. +[3063.58 --> 3063.90] Okay. +[3064.02 --> 3069.94] You have a squirt, like N squirt algorithm or you can N factorial algorithm. +[3070.22 --> 3073.92] So you shouldn't do it. +[3074.96 --> 3075.44] Awesome. +[3075.62 --> 3076.08] Awesome. +[3076.08 --> 3089.04] Thank you guys so much for coming on the show and talking about really what piqued my curiosity when it showed up on my desk, my virtual desk, you know, PHP and go and how these two things sort of complement each other. +[3089.18 --> 3090.50] It's been awesome having y'all. +[3091.12 --> 3094.28] And I will now play our outro song. +[3094.28 --> 3097.44] And we will try to wrap this up. +[3097.60 --> 3098.58] Right on time too. +[3106.10 --> 3107.64] That is our show for this week. +[3108.04 --> 3110.92] Thanks again to Seb for requesting this episode. +[3110.92 --> 3111.88] We hope you enjoyed it. +[3112.10 --> 3113.34] Yes, we take requests. +[3113.68 --> 3119.14] Head to gotime.fm slash request and let us know what you want to hear about on the pod. +[3119.66 --> 3123.14] Special thanks again to Fastly for delivering our shows all around the world. +[3123.40 --> 3127.22] To Breakmaster Cylinder for hooking us up with all the excellent beats you hear on the show. +[3127.46 --> 3128.42] And to you for listening. +[3128.88 --> 3130.46] We appreciate you spending time with us. +[3130.92 --> 3135.06] If you have a couple more minutes, enjoy this clip from the changelog 486. +[3135.06 --> 3140.12] We invited Frank Kruger on to discuss his practical guide to solving hard problems. +[3140.38 --> 3141.96] And he shared so much wisdom with us. +[3142.36 --> 3142.76] Listen in. +[3143.14 --> 3147.64] And I was reading Wikipedia page after Wikipedia page, modern treatment after modern treatment. +[3147.86 --> 3150.88] What I was trying to do was synthesize these V nodes. +[3151.00 --> 3153.38] It's a complicated thing of data management. +[3153.38 --> 3159.60] And I couldn't understand any of the algorithms until I opened the Dragon Book and saw in the 1970s, +[3159.60 --> 3165.36] their pseudocode implementation of the algorithm, which threw away all the details, +[3165.86 --> 3169.18] ignored all these modern advances that aren't actually advancements. +[3169.24 --> 3170.28] You don't actually need them. +[3170.74 --> 3175.36] And written out in this very clear style and all capital letters. +[3175.52 --> 3178.36] I don't even know what language they were pretending to be in that book. +[3178.56 --> 3183.40] But just finally getting it from this old, old resource and realizing, +[3183.62 --> 3186.62] oh my God, in the 1970s, there's chapter five, section four, +[3186.62 --> 3188.64] and they describe exactly the problem I'm having. +[3188.88 --> 3191.58] And they, oh my God, even better, have a solution to it. +[3191.82 --> 3191.92] Wow. +[3192.00 --> 3195.76] And then you can transcribe that solution from their crazy, +[3195.86 --> 3198.82] whatever language that was, into whatever you want to be using. +[3199.22 --> 3201.30] And you learn a lot during that process. +[3201.44 --> 3203.90] That felt so good to me when I finally found that. +[3204.56 --> 3207.76] It's like coming across hidden treasure somewhere. +[3207.96 --> 3209.38] You're like, look at this. +[3209.52 --> 3210.12] Look what I found. +[3210.40 --> 3211.38] I knew they were smart. +[3211.46 --> 3211.88] That's crazy. +[3212.54 --> 3214.34] You want to tell somebody at that moment, but nobody, +[3214.34 --> 3216.14] not that they don't care, they just can't care. +[3216.14 --> 3218.40] It's like, they just can't care. +[3218.52 --> 3219.10] They can't care. +[3219.20 --> 3220.88] It's like, I have no idea what you're talking about, Frank. +[3220.98 --> 3221.34] Okay. +[3221.60 --> 3223.12] But congratulations on something with the problem. +[3223.56 --> 3225.24] Well, there's a little street cred too. +[3225.48 --> 3230.32] Like just knowing about the book shows that you're semi-interested in compiler technology. +[3230.56 --> 3232.06] Actually having a use for the book. +[3232.22 --> 3234.24] I feel like I became a computer scientist that day. +[3234.30 --> 3236.96] I actually applied something from the Dragon book. +[3237.12 --> 3237.34] Yeah. +[3237.42 --> 3239.90] It was a real high point in my career, to be thoroughly honest. +[3239.94 --> 3242.74] And that's where you're standing on the shoulder of giants. +[3243.18 --> 3245.08] It's like you graduated from Hogwarts that day. +[3245.08 --> 3245.86] You became a wizard. +[3246.14 --> 3247.00] You became a real wizard. +[3247.08 --> 3248.28] By copying a wizard's spell. +[3248.48 --> 3248.68] But yeah. +[3249.44 --> 3251.80] But I realized the wizard's spell worked. +[3252.06 --> 3252.20] Yeah. +[3252.24 --> 3253.78] I was very Harry or Hermione there. +[3256.02 --> 3262.14] Continue listening and subscribe to the changelog at changelog.fm slash 486. +[3262.80 --> 3266.16] Hey, are you ready for our next installment in the maintenance series? +[3266.66 --> 3272.22] Chris assembled an awesome panel to discuss what to do when Go projects get big and messy. +[3272.22 --> 3277.54] Johnny joined him, as did Ian Lopshire and Sam Boyer for an excellent conversation. +[3278.00 --> 3280.96] That's something to look forward to next time on Go Time. +[3281.24 --> 3281.74] Go. +[3281.74 --> 3282.86] Go Zen 3 fuckin 3 at 9! +[3282.90 --> 3284.78] Stop talking here! +[3285.22 --> 3287.20] Ciao! +[3287.44 --> 3287.54] Bye! +[3287.68 --> 3288.22] No! +[3288.70 --> 3288.72] No! +[3288.72 --> 3289.04] No! +[3289.04 --> 3289.08] No! +[3289.10 --> 3289.28] No! +[3289.28 --> 3289.32] Oh! +[3289.32 --> 3289.38] No! +[3289.38 --> 3290.22] I know you have enough. +[3290.56 --> 3293.42] No! +[3293.42 --> 3299.24] Game on. diff --git a/Go and PHP sitting in a tree_transcript.txt b/Go and PHP sitting in a tree_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..902e589690d45c8727878ef16912854365ff45d3 --- /dev/null +++ b/Go and PHP sitting in a tree_transcript.txt @@ -0,0 +1,253 @@ +**Johnny Boursiquot:** Welcome to this episode of Go Time, everyone, and glad to have you back. So Go and PHP, sitting in a tree... I don't know if there's kissing going on, but we're gonna figure out why Go and PHP are together on this particular episode... Because the last time I touched PHP was many years ago, and although it had its charms, it did have some quirks as well; we may touch on those during this episode... + +Joining me today are two folks who actually work with Go and PHP on a regular basis, and that got us curious. Someone actually suggested this episode, I believe -- let me look up there. I will find who it was and give him a shout-out for suggesting this episode. + +**Jerod Santo:** Shout-out to listener Seb for requesting this episode and putting RoadRunner and PHP on our radar. Thanks, Seb. + +**Johnny Boursiquot:** And they actually recommended RoadRunner, which is a project we'll definitely be touching on, by name, as an example of Go and PHP working well together. So joining me today are two guests - Mr. Valery, and now I'm gonna butcher your last name... Piashchynski? Did I get that right? Close enough... And then also Anton Titov. Hopefully, I got that right. + +**Anton Titov:** Yeah, that's correct. + +**Valery Piashchynski:** Nice to meet you! + +**Johnny Boursiquot:** \[04:10\] Okay. So Valery is a software developer at Spiral Scout, the team that works on RoadRunner, this particular project that we'll be touching on. He enjoys working on the algorithms, writing his own operating system for learning purposes in C++, and helping folks get into programming, which is pretty awesome. He also streams on Twitch, mostly open source Go stuff. + +Anton is actually a CTO and co-founder at Spiral Scout, also on the same team that actively works behind RoadRunner. He loves software and hardware. We were just talking before the show about some of the hardware stuff he's working on with Go, rather than going the traditional route of the embedded stuff with C, or Python, and whatnot. I got a chance to take a peak and that. Very interesting stuff. He also enjoys some DIY robotics and machine learning. Welcome to you both, gents. + +**Valery Piashchynski:** Yeah, thank you. I was saying I do not try to pronounce my surname... \[laughter\] It's rather complicated. But you were almost correct. + +**Johnny Boursiquot:** It's all good, it's all good. + +**Valery Piashchynski:** Yeah, yeah. + +**Johnny Boursiquot:** So Go and PHP. Let's start with what brought you to PHP. Why are you working on PHP this day and age? Not that there's anything wrong with working on PHP this day and age... I'm saying, out of all the languages that one could pick, PHP has been around a long, long time. Back in the infancy of the web. At some point, PHP was the it language to actually write dynamic websites and whatnot, competing with ASP classic, as we refer to it now, and the likes of ColdFusion... I mean, these things are grand-daddies or grand-mommies of the early web. What led you to PHP? + +**Anton Titov:** Well, as you mentioned, it's a very old language. I was young and naive, and I wanted to have my own forum, or my own CMS board. And if you were trying to build a forum back in '05, the only option for you would be PHP. I remember trying to download the source code of the website and trying to figure why it doesn't work... Well, apparently you had to install a bunch of instruments to also make it work... But it was just the beginning of kind of this long, long journey. And I'm still sticking to that. So it's a beautiful language these days, and it changed a lot. + +**Valery Piashchynski:** Yeah, yeah. Very nice. \[laughs\] So... Surprisingly, I'm not a PHP developer. I'm a Go developer and working on the Go part of an ecosystem. We're trying to connect it into the PHP parts, so Anton for me is like the light at the end of the tunnel, connected to me from the PHP side. + +**Johnny Boursiquot:** Okay. So then you've got the PHP as part of your background, and then here comes along Go. When did you get into it and what led you to actually combining those two things? + +**Anton Titov:** I can probably talk from the kind of combining part... I mean, when I started studying Go, it had been around for a few years, and I only heard good things about it - performant, fast, concurrent, all this kind of terms which now... well... we all know about Golang. Well, I mean, I actually just tried to just play with that. I made a couple of interesting applications, and since PHP was the main production stack, I was just trying to see, can I use it within this practice? Because all the examples in Golang were kind of like small and easy... And in PHP we have frameworks, like ten layers of abstractions, ORMs etc. So it's a very different worlds. + +So it was a very curious language for someone who had been working on this type of a language for a very long time. And that has been an idea - can we actually make them work together? Can we get the benefits of PHP and the benefits of Go and can we improve the developer experience/our own experience? + +**Valery Piashchynski:** \[07:58\] I came to Spiral Scout in 2018 as just a regular developer on one of the projects; a Golang developer. Previously I was programming in .NET, so I was heavily involved into the .NET ecosystem, like C\#. I've got some (I guess) \[unintelligible 00:08:17.23\] something like a C\# professional. So I saw a guy who worked on a very famous taxi company based in New York... \[laughs\] He was rewriting all the Ruby on Rails system into some interesting language. It was like 2015 maybe, in the early days of Golang. So I was like "What is this language? Could you please explain me?" + +I was really impressed about the first web server written in Golang. It was so little lines of code, and it brings you to a web server that can respond with just Hello World, but it just works. + +After that, I came into Spiral Scout, I was involved into internal projects, and one time I started to rewrite a test in RoadRunner. It was a mess, because Anton with his php expertise forgot to return errors from the functions. \[laughs\] Like, there are no errors if you don't return them. + +Of course, I fixed it, and it was a lot of errors in the tests... And I sent this PR to an intern, and the intern was like "Oh my God, why did you do this?" \[laughs\] And we're like, "Okay..." We started fixing it, and we finally fixed all of these errors. After that, I was involved into the Golang part of the RoadRunner, and we started working with the intern together, to improve the quality. + +**Anton Titov:** Well, he pretty much rewrote most of the parts of it, but... \[laughter\] That was for the good. + +**Valery Piashchynski:** Yeah, maybe. + +**Johnny Boursiquot:** Obviously, you saw a need and you decided to fill in, and started doing some of the things that you saw that needed to be done... And that's always awesome. That's ho the majority of open source contributions happen - you find something that you enjoy working on, and you contribute code to it, and that's a beautiful thing. + +Obviously, when I think about my experience with PHP, and all the hoops we had to jump through to "make it scale", and using today's terminology and whatnot - when I read the description of the project, I'm like, "Okay, this is a load balancer rolled into some sort of application server, rolled into..." I mean, it's trying to do a lot of things, so why don't we started with what the difference is between plain Jane PHP application server -- like, the last one I was used to was like Zen, or something like that... + +**Anton Titov:** It was a very long time ago, yeah. + +**Johnny Boursiquot:** Exactly. I'm dating myself here. But what's the difference between those -- I guess, for a lack of better terminology, those traditional application servers that are designed to run PHP, versus this new approach? + +**Anton Titov:** To answer this question, it's actually important to understand how PHP actually became this type of language, and both a language that's hard to scale. Imagine every time you write a Golang application - let's say doing some endpoint in HTTP - every time you get an HTTP request, immediately you have to boot load the application from the disk, start it, answer this request, and then kill this application. And do it over and over and over, for every request. This sounds super-expensive, right? + +**Johnny Boursiquot:** Mm-hm. + +**Anton Titov:** Well, that's how PHP has been working for 26+ years. It's quite amazing that you have the tech which quite literally restarts on every request, and it still kind of powers pretty much the majority of backends on the internet. Well, public backends, let's say. + +\[12:07\] So the idea was actually quite simple - let's just remove this overhead. I mean, when I started working with RoadRunner, I started working -- I was actually a protocologist, making communication between the languages. And the first example was quite simple. Let's say we have a function in Golang to do some heavy math, which on PHP maybe is not optima, and they have like this highly OOP, strict code in PHP. And by the way, modern PHP is all about OOP, strict type, annotations, attributes... It's all very similar to Java these days, except it's free, open, and very easy to learn, let's say. + +So then I just tried to make a call from PHP to Golang using this internal socket, or UNIX socket rpc call and it did work. And then I tried to make a very stupid experiment, actually. I tried the native PHP library for RabbitMQ to push message, and I used the Golang library of RabbitMQ to push message, but with additional RPC overhead from PHP. And we ran some tests and we found out that the PHP to Golang bridge through RabbitMQ works - not like margin much but a few percents faster than a native solution. And it was like, "That's weird." This shouldn't be happening, right? + +And this kind of led to the idea that -- PHP is a very beautiful language to model business processes. Not like highest-scale IO operations, like traffic management or ingresses. It's single-threaded, it's very dummy, in terms of like it can go left, in can go right in some cases... You still can shoot yourself in the foot, but these days it's much harder. But it's very good to have good libraries to explain permission models, document mapping, data mapping. If you see how to work with mapping JSON in Golang and PHP, you'll definitely see a major difference in favor of PHP, because it's a dynamic language. + +And Golang, on the other side, is beautiful to manage all of these long polling connections, sockets, retries, restarts, delays... All of this fun stuff which PHP, just by definition, by model, can't. So then we tried to create the method, invoking code from PHP's worker pools. So you have like hot processes of PHP, which are in memory, let's say one per CPU core, and then just ask one of them "Do this payload. Do this work for me." You don't kill the application, you don't restart it. You have no overhead. And when we did this code - well, it was working like eleven times faster than the native approach... So we created an HTTP layer on top, called it RoadRunner, and it's been \[unintelligible 00:14:44.25\] and we haven't written a single application without this model probably since like 2019. + +**Johnny Boursiquot:** Wow. + +**Valery Piashchynski:** Yeah, a pretty amazing speech. \[laughs\] + +**Johnny Boursiquot:** So who is this for? Is it the Go developer who has to work with PHP, or the PHP developer who has to work with Go, or who are you targeting with this approach. + +**Valery Piashchynski:** I guess both of them. + +**Anton Titov:** Well, it's actually a very good question. The main audience is obviously PHP people, because the main idea of RoadRunner is you can take these complex aspects of queue load balancing, HTTPS traffic, temporal gRPC, and you're going to make them boring for these developers... But just out of the box. You want gRPC? Sure, just plug and play. You want a temporal? Sure, it's already here; just make it work. You don't need to install anything. It basically manages the complex stuff for you. But at the same time, it's kind of for the Golang engineers who typically work in pair or on the same team as PHP engineers... Because this is an application server, it's very easy to intercept and modify the requests and calls which you do with PHP. So you can add your own validations, authentications, and all obviously are going to work much faster, and possibly with much deeper integration, with more than cloud-native tools. + +\[16:04\] You have metrics, readiness, healthy endpoints... All the stuff you need to make application, basically like native current environments. But obviously, the first target audience is just engineers and companies who are just trying to write scalable code, but at the same time don't necessarily want to hire ten Rust engineers. It's more like a balance between price for the engineer, and how fast you can find them, and the promise and quality of the software they create. + +**Johnny Boursiquot:** Okay, so this is as much a technical decision as it is a business one. + +**Anton Titov:** Well, in the long-term yes, because how many startups you've seen which come from the point "We're going to scrap out PHP and move to something else"? It's been a bunch. Facebook invented their own language. VKontakte, their Russian competitor, did the same thing. So it's kind of like become so expensive that you even have to jump in and make your own compiler for this stuff. And we can just move this line when you have to move from one tech to another, ideally up to infinity. If you need something fast, do it in other languages... I mean, it's all done in microservices these days, in distributed applications... So you're no longer stuck to one language. But at the same time - well, you're a startup, you're trying to integrate with a few providers, and you need to create 12-15 API endpoints. Who are you going to be using for that? Do you really want to hire senior engineers who are gonna be doing that, or you can just use senior PHP engineers, which are much higher availability since it's an old language. But you're going to do the same thing, it's going to work the same, same performance, it's just going to be done easier because you can source people easier. + +**Valery Piashchynski:** Yeah. And at the same time, if you wanted to do some pretty hard work or some low-level stuff, you can easily write a simple plugin, and plug it into RoadRunner, compile it, and solve your needs with that. So we also wrote our own plugin system, because -- so initially, we wanted to use the Golang native plugin system. It works only on Linux at the moment, so we waited a little for the Windows support, but I guess it doesn't seem to happen, the Windows support for Golang-native plugins. But we wrote our own plugin system called Endure, which is suitable for plugging all of these pieces of Golang, combining into one part, initializing it, starting to serve it, stopping... So building a tree with all those plugins. + +For example, if you write once a configuration parser -- if you write a parser from the Yaml, for example, you don't need to write or copy or create some SDK to bring this part into every plugin. You can just request this init function, and RoadRunner will take care about finding this dependency, initializing it, topologically sorting the graph, and providing this initialized dependency for you. You only need to just – "Okay, configuration, please give me the gRPC section", or "Give me some other section." + +If something goes wrong, RoadRunner will take care of this of course, and provide you nice - but unreadable for PHP users - message about some Golang panic, or some error, something like this. And for the PHP -- all the complexity for the PHP is to properly define what do you need in the configuration. You have a configuration, you need HTTP, so you just enable an HTTP section, put your configuration, and RoadRunner will remove all other plugins from the tree. It won't even start. It just runs your section for your needs, like HTTP, or gRPC. Or you can write your own plugin, put your section in the configuration built with our tool called Velox. This is a tool which helps you to build RoadRunner with your own custom plugins based on GitHub. So all of this heavy, complex stuff was moved to the Golang part, and the open, nice things moved to PHP. + +**Anton Titov:** \[20:32\] That's actually quite an exchange of knowledge, because this first container actually came as an architectural pattern in most of the PHP applications... Because if you've seen .NET, Java or PHP applications, you have a ton of classes, interfaces that inherit each other, they use class declarations etc. So you can't work without a container. And with dependencies - you need to manage all of this stuff. So essentially, we used this idea inside of RoadRunner, but obviously \[unintelligible 00:20:58.17\] initially lays it all correctly, in the correct order, and answering the question, like "It's a large tool already..." Well, it's kind of not, because it's just a container with a bunch of CLI tools and instruments, and the rest is just kind of independent projects and plugins. So we can add them without influencing each other, or worrying to break the tool. Do you want RoadRunner with http layer? Sure. If you don't - well, disable the plugin and build it. It's going to be exactly the same thing, just less memory to manage. + +**Break:** \[21:33\] + +**Johnny Boursiquot:** Let's dive in a little bit, into the weeds, if you will... Are you shipping like a binary? Are you interpreting PHP? If I'm a developer, what does my experience look like? + +**Anton Titov:** From a PHP perspective, we're trying to do the least invasive work possible. You literally don't need to do anything to make it work from the PHP side. There's no extensions, there's no special CLIs, interpreters, nothing. What we actually do, we do the 20-30 years old approach, which still drives old applications - we manage the worker pool. + +\[23:57\] So what RoadRunner does - it actually uses the default PHP interpreter, which is a binary, invokes it with the application, and then keeps its memory in like a pre-warmed state. And when the request payload comes, which can come for HTTP endpoints, task queue, Temporal Workflow, gRPC, whatever, you name it - you just send this payload to PHP and wait for it to complete. But over the lifetime, the only main difference for the engineer - which is for some engineers quite hard - is to realize that the application lives longer than just a single request. And you can't just have a global variable counter which you are going to ++ and expect it's going to be zero on the next request. That's the only difference. But besides that, it's the same exact PHP, same exact extensions and configurations as you typically use, it's just managed in a bit different flavor. + +**Valery Piashchynski:** But surprisingly, RoadRunner knows nothing about the PHP. So RoadRunner is not binded especially to PHP; it just runs some command you specify in your configuration, so you can do it Python... You can ever run Golang inside the Golang workers. So the main purpose for RoadRunner is to manage the process. Who will be in this process? For RoadRunner - it doesn't care about this. It cares about the protocol. The protocol is language-agnostic. + +A project, some guy implemented this protocol in Python. Anton showed me some time ago. So it's Python running inside RoadRunner, because it's pretty much the same model as a one-threaded model in Python. Some guy wrote this, and it works. + +**Johnny Boursiquot:** This sounds like a process manager, if you will... So if you wanted to, you could have Python, obviously, as you've just mentioned, Ruby, obviously PHP natively, and whatever else other interpreted languages you want. Even precompiled things, it sounds like you can just have in there, and then basically just invoke it. + +**Anton Titov:** In theory, you can take a 20-year old Perl file and run it on RoadRunner... \[laughter\] But if you want to do it - obviously, that's a good question. But yes, I mean -- I think the main mention of PHP is actually in the title of RoadRunner - it's a PHP application management server. But besides that, there's nothing which actually ties it specifically to a single language... Except that this single language has the largest SDK codebase to communicate with all the features from RoadRunner. But that's just a nuance, quite frankly. + +**Johnny Boursiquot:** Okay. So if I'm used to working with single-process PHP-style applications - I think you touched on this earlier - where it's like I'm restarting the world, traditionally, when I'm dealing with PHP; one request just restarts the world, and everything is like anew... In this world where instead there's a worker pool, there's process management happening - do you find that developers have to have sort of a mindset shift to basically think "Okay, there's not just one process here. There's multiple." Does that change the nature of how they program? Do they switch from programming (I guess) a single process to now having multiple processes to contend with? + +**Anton Titov:** Well, I mean, when PHP runs at scale, you still have multiple processes. They're just created on demand. But you still have 500 PHP processes running the request. But there is definitely the conception shift in people -- it was really much harder even a year ago, but now... The main reason why it's easy these days is because a very minimal amount of people write on pure PHP. Unlike Golang, where most of the time you're writing pure Golang. You may be using some small HTTP overlay framework, but you're mostly going to be using SPL functions, let's say, to do most of your business stuff. + +\[27:50\] In PHP it's completely different. You have Symfony, Laravel, or our infrastructure framework, Spiral, and they all manage it for you. So the beautiful thing is all of these frameworks, all the time -- well, we created our own, specific for this purpose, but all other frameworks, they actually do upgrades and patches which solves all of this kind of thinking nuances for the engineer. They reboot some part of the services to clean up the caches etc. So if you're using a modern framework and if you're using RoadRunner, you most likely won't even notice a difference in 90% of your activities. It's still going to be some nuances, but most of them already known and have already been solved. + +If you're using a framework which is specifically built for RoadRunner like ours, nothing is different for you. Just write code, and it's all being managed; state is managed in-memory, and all completely resetted. + +**Johnny Boursiquot:** So would a PHP developer ever even need to know what's under the hood, what their process is? I assume they'll never really - unless they want to - come in contact with any Go whatsoever. + +**Anton Titov:** Well, it's like owning a car and being able to drive a car. If you have RoadRunner and PHP, and a default PHP SDK, you can do many wonderful things. You can PHP respond in like 50 microseconds, run \[unintelligible 00:29:07.27\] and all the queries, use gRPC, and use all of this stuff. But if you also know Golang, you can do even crazier stuff. You can take, for example, a machine learning model from Keras, compress it, put it into the Golang SDK, and then invoke from PHP, and you now have machine learning embedded directly into PHP with Golang bridge. So you can do these crazy things if you know multiple languages, specifically Golang. + +Golang is beautiful, because -- I mean, it's quite easy to learn it, right? It doesn't have many of these legacy overhead nuances you have to remember. So it's not required, but if you do, you're essentially a superhero in PHP world. + +**Valery Piashchynski:** Yeah. And also \[unintelligible 00:29:49.23\] + +**Anton Titov:** Well, yeah, that's another example how you can use it. + +**Valery Piashchynski:** It's also written here, in RoadRunner and PHP combination. + +**Anton Titov:** I can show it when it's going to be the right time. Up to you, Johnny. + +**Johnny Boursiquot:** Well, given that most of our users are going to be listening, not viewing this, we may not be able to show them much. + +**Anton Titov:** I'll try to walk them through, yeah. + +**Johnny Boursiquot:** I do want to sort of understand -- obviously, this is an open source project, and folks are contributing to it... I do want to understand what are some of the hardest challenges that you encountered while coming up with this model. Obviously, running a process manager - you have to worry about inter-process communication, you have to worry about how to keep things in memory efficiently, and all that stuff... I'm curious as to what are some of those biggest challenges that perhaps you're still facing. + +**Anton Titov:** About the current challenges Valery can speak. I don't think there's a lot of them now, but... When we started working on this though, it's been a number of interesting things to solve. Number one, we had to create a protocol to communicate between two different languages, and it had to work over pipes, over Unix sockets and TCP sockets. So it's been like, "Okay, how are you going to create a low-level IPC protocol if you don't want to jump to shared memory, or all these things?" + +This has been not like a hard conceptual problem, because you can always... Default protocols like a netstrings - you have the lengths of the message, a fixed header size and then you have the payload body. So it's quite easy, you just read the package between languages. + +Then obviously it was a problem with the process manager, because PHP sometimes tends not to start if you don't send the right parameters, or it may crash if, again, you send an invalid payload -- I mean, it doesn't do it anymore \[unintelligible 00:31:36.09\] And obviously, race conditions on the Golang side - that was horrible, because you can't just write a process manager; no one just writes a process manager... \[laughter\] For the reference. It also needs to collect the stats, you also need to watch for the process to restart, you need to collect the \[unintelligible 00:31:54.16\] error, you need to understand how many times you invoked it, when it was started... Things like, "Okay, let's check how long this process exists", and you call, say, time now, and you immediately kill your performance, because it's doing a syscall. Well, back in the day, something like that. + +\[32:14\] Or for example you have a request coming, but at the same time the PHP process runs out of the memory. So what's going to happen? It's going to fail. How are you going to restart it? There's been so many little integrational hell and edge cases for this type of work... But eventually, once we jumped from this port, what we managed is not a part of the hell. When we have a single server which has an HTTP endpoint, everything is quite easy. You have request packets into the binary form, sent to PHP, and Bob's your uncle. But when you're doing HTTP and then you also want to manage the queue process manager, like cooperate with RabbitMQ. But what happens - what if you have HTTP, which runs a PHP worker, which sends data to queue? So now you have two plugins which not only have to work, they also have to be created in a correct order, wait for each of them to properly connect, and only then make it work. + +So that was a part of scratching the head a lot, because it was so hard to solve... With all of these dependencies, and plugins, and... Hidden dependencies, because a PHP worker can theoretically do anything it wants; it can call to queue, it can invoke an HTTP endpoint through RoadRunner onto itself, which -- well, I'm not even sure what's going to happen in this case. But that's why we created the container, that's why we created eventually Endure, which solves all these problems, and now there is barely no integrational hell between plugins and RoadRunner. They're all interfaces, very easy to connect them together... It's basically become a framework for the application server. + +**Johnny Boursiquot:** So if you have this deployed, is it recommended that you have a single tenant kind of situation, whereby you don't want processes from multiple parties that are not really associated with, say, the same company? You don't wanna run this as some sort of an open-to-all, multi-tenant system, because you could have one process peaking into what's happening with another process within this world? Or is it isolation between these things running all in the same system? + +**Anton Titov:** Well, it could be. We can isolate it by many different ways. We can isolate them by running with different permission models in PHP. You can literally forbid most of the functions, which - well, some engineers and hackers will still be able to bypass. You can run them in different user groups and in different memory spaces. + +Right now, we don't run it in a shared fashion, old-fashioned shared hosting. It's mostly suitable for the most classic approach now. You have a container, within this container you have the application, your APIs to other service functions of this application. And it's fully self-contained inside this container. If you work on a multi-tenant model, you don't run ten different PHP scripts with different users. You solve the multi-tenancy on the application design level, using the application domain. I hope it makes sense... + +But I've been thinking about this problem for a while. Basically, in short, if you wanna do multi-tenancy, do it in the application code. RoadRunner is developed/designed to actually work the best in modern Docker environments or container-based environments. It's a single application, pair application instance, let's say. A single instance of application, sorry. + +**Johnny Boursiquot:** Gotcha. + +**Valery Piashchynski:** Yeah. As for me, it was a lot of challenges to write all of these things... Because basically, RoadRunner, as you can see in the RoadRunner repository, is just an interface. It's just RoadRunner server, RoadRunner workers command, but everything is hidden under the main three parts. Those parts are - it's Endure, as Anton mentioned... It was a surprise for me, because in Golang we don't have containers, because we just don't need them to get to some dependency, or... So I heard about some project in Uber, like Uber FX, or Google Wire, I guess, but I don't think it's much popular solutions. + +\[36:28\] So in my projects I've never worked with a container, but PHP is everything about containers and everything about dependency injection, and so on, so we have to write such algorithmic containers which mutate based on the configuration. So it's not like "Please give me some dependency." It's about I have a configuration, I have a set of -- I guess at the moment we have 20 plugins, or more, I guess... I don't know how much exactly. But we provided a bunch of plugin, we provided the configuration, and said "Okay, now build this. Build this properly." Managing the connection as Anton says, if you for example initialize a logger, and the logger needs a configuration, so you have to properly topologically sort all the things to initialize first integration, then provide this pointer to a logger... So the Golang race flag won't help us here, because it's so distributed all over the plugins, so you have to manage every race condition by yourself. So you have to be very careful with writing all the things. But we hide all this complexity inside Endure, and you have to be sure the provided dependency is concurrent-free. So you can't use it from the different threads. + +But the second part is also Goridge. It's a protocol; in Goridge version 1 and version 2 it was, like Anton said, it's a very crazy protocol... I guess 14 bytes - correct me if I'm wrong, Anton... 14 bytes or 18 bytes. + +**Anton Titov:** The first version had a 17 bytes header, for whatever reason. + +**Valery Piashchynski:** 17 bytes, yeah. It's like a few flags, payloads, and \[unintelligible 00:38:27.12\] But imagine the situation, if you pass your payload over the pipes, for example, you don't have any mechanism to CRC or to check "Is this payload correct?", passing it from one side of the wire to another. But RoadRunner and Goridge version 3 has a protocol based on -- an IP protocol. So I wrote recently a few protocols, a TCP/IP protocol, and combined them all together to have a proper protocol for communicating with PHP parts. So we have asynchronous support, we have headers like in ip protocol, we have variable length options. So we have all this funny stuff to extend it, and not to break it from version to version. + +And I guess the third part is the key, which contains all these worker pools... And this is complicating stuff, because you have to -- for example, imagine you have an HTTP plugin enabled. The user wants to get a statistic about the workers. So one approach is to stop the world, stop all the HTTP requests, and get all the statistics, get all the pointers to workers, get status, pretty it with some format, and show it to the user. + +\[39:56\] Another approach is to have some shared place where you can safely get at any time, without logs even, this pointer, and to provide to users. But when the restart happens or some issue happens, it will look only in this case. + +So it's like statistics in RoadRunner is basically free for use. It doesn't interrupt the actual request. Not HTTP, not jobs... None of them. + +**Anton Titov:** Yeah. It's quite funny how we jumped into these edge cases, because most of our users are already mature PHP applications, and large, large startups... And they don't play with ten requests. Like, "Okay, we just bump a few millions a day, and we see this bug. So what's this bug about?" Oh, you know what - over time, things which you could never imagine while developing the thing, but whenever you catch them... Thankfully, users report them, and it's easy to see how it works on scale. + +**Valery Piashchynski:** Yeah. And basically, imagine a situation when you have a completely first scheduler in Golang. I guess you so this ticket recently, about writing completely the first scheduler, bring this to Golang. So we started working with that (I guess) a year before, to provide a completely free scheduler to schedule jobs inside RoadRunner. We didn't finish it yet, but we're continuously working on it, to provide a binary heaps algorithm to sort by priorities all of these jobs. For example, if you have an urgent job to execute, you have to set that priority 1, and it will be sorted and scheduled properly... There are a lot of such things, very complicated things which you should work inside RoadRunner and hide all of this complexity for users just to specify a few values in the configuration... Just "Okay, I don't need a gRPC server. It's removed", but all of this complexity hidden under the hood. + +**Johnny Boursiquot:** That's why it's hard to create very nice, user-friendly APIs. There's a lot going on under the hood, but you're making it easy for people to actually use. That's the beauty. + +**Anton Titov:** Well, it can be easier to make just a Hello World endpoint. + +**Valery Piashchynski:** Yes, it's so true. + +**Johnny Boursiquot:** \[laughs\] + +**Anton Titov:** Yeah. Well, we need a load balancer, a Kubernetes cluster, control plane, the other things. That's easy. \[laughter\] + +**Johnny Boursiquot:** I know. Man, things have changed over the last decade or two... + +**Anton Titov:** Yeah. + +**Johnny Boursiquot:** So I do wanna switch gears a little bit here, to understand - like, if someone wanted to contribute, because it sounds like there's still a lot of awesome ideas at play, and coming to the project... If someone wanted to contribute to that, what would basically be a safe expectation of them? Should they know how to work in PHP, should they know how to do Go? Is there work for both sides of the fence? What do you need to contribute to this project? + +**Valery Piashchynski:** Basically, I don't know PHP at all. I started googling, like, "Okay, Anton, I need some tests script. Could you please write it for me?" \[laughter\] Or I need a for loop to read some variables, so I need to google how to write a for loop in PHP... And this is great, because I don't need to be involved in the PHP part. I can concentrate only to improve in the Golang user experience. And to contribute to RoadRunner, it depends on who are you as a developer. Are you a PHP developer, or are you a Golang developer? If you are a Golang developer, it's very easy - you just need to go in the RoadRunner issues, find an issue marked "Help needed" or "Easy to resolve", like some entry-level issues... Or if you want to contribute to the PHP part, I guess it's the Spiral framework, which you can also have a lot of tickets to improve our PHP part. And you don't need to know both sides of this at the same time. You only need to know PHP or Go. Or if you want to contribute both, usually you need to know both, but... Yeah. + +**Johnny Boursiquot:** It sounds like there's room for either side. + +**Valery Piashchynski:** Yeah. + +**Anton Titov:** \[44:09\] We're pretty friendly to contributions. In PHP you can help us to improve SDKs, in Golang you can go as deep into the weeds as SSL, which dumps algos to run some crazy stuff. + +**Valery Piashchynski:** Even Python developers can write a protocol version 3, or all Ruby on Rails developers... + +**Anton Titov:** Yeah. If you want to, we'll definitely accept the contributions to make it work for other languages. + +**Break:** \[44:34\] + +**Jingle:** \[45:54\] to \[46:09\] + +**Johnny Boursiquot:** Alright, so let's start with you, Anton. What did you bring? + +**Anton Titov:** You only need 64 kilobytes of RAM. + +**Johnny Boursiquot:** For what. + +**Anton Titov:** Just for everything. + +**Johnny Boursiquot:** Ever? \[laughter\] + +**Valery Piashchynski:** Okay... Prove it. + +**Anton Titov:** For most of this stuff. I would just say, people have to try to work on more memory efficient applications. Because when you work with hardware, 64 kilobytes of memory is a ton. You could make robots move, bring eyes, go on stairs, and do some other stuff. And what can you do with 64 kilobytes of a JavaScript application, tell me, please? \[laughs\] Maybe nothing. + +**Johnny Boursiquot:** Not much. \[laughs\] + +**Valery Piashchynski:** I have 64 gigs and I even can't run a stream. \[laughs\] + +**Johnny Boursiquot:** I mean, I will say that over the years we've gotten more -- well, let me put it nicely. We don't worry very much about CPU and memory and disk... Things that used to be expensive 20-30 years ago. Not so much now. We take these things for granted now. + +**Valery Piashchynski:** Yeah. + +**Anton Titov:** Yeah, but if you know how to pack it down to this level, you can create a lot of scalable applications. Because when you create traffic filtering software, or vpn cores the things which actually -- well, a real IP, let's say, not just API endpoints... That's where you have to optimize it. And knowing these basics and knowing that "Yes, 64 kilobytes sounds like a small amount", and it's a joke which Bill Gates said back in the day, if someone doesn't remember; obviously, he's wrong, but... If you realize how actually huge is this amount... 64 kilobytes of stack on some hardware chip can let you stream a ton of traffic, and build something like a Netflix, build something like Starlink, because all these things which are doing great, great things - they have to be optimized. You can't pool ten CPU servers in space; this is not going to work. + +**Johnny Boursiquot:** \[48:13\] \[laughs\] Alright, alright, alright... Valery, what did you bring? + +**Valery Piashchynski:** So my unpopular opinion is open source is hard work. In my opinion it's much harder than some enterprise development... Because I was involved in different enterprise projects, and the flow is pretty much defined. If customer support has some ticket, it can process it, send it to a specialist, like quality assurance, it can test it, write test cases, it brings to you, you can see this ticket, fix the problem, run the tests, and so on. + +But in open source, a lot of people think that they should not bother themselves to write a proper description of an issue. It's like, "I have a problem. Please fix it." Or one of my favorite issues is like "Nuts?" Well, what does it mean?! + +**Johnny Boursiquot:** Are they offering nuts, or are they asking you if you want nuts? I mean... \[laughs\] + +**Valery Piashchynski:** But when I said, "Please describe what do you want? You want support, or something else?" and closed this ticket, the guys asked me "Why are you so rude? Why did you close my ticket?" And there are a lot of such things in the open source development, so you should handle a lot of things. You have various types of virtual machines to run on Ubuntu, previous versions. Ubuntu, current versions. Debian. FreeBSD. macOS. Windows, different versions. + +So if someone sends you a ticket -- I guess some people think that this is a joke, that open source development is a joke for us, or I don't know. So it's like, they write two or three lines of description, or do not provide test cases... So it's very complicated to figure out what do you really want. + +One guy asks me "The Docker doesn't work on my machine." I'm like, "What?!" So I have to fix the Docker on the machine... \[laughter\] You see, in CI, everything brings from the scratch. Docker installs from scratch. Everything works inside Docker. It's something inside your machine. But the guy said "No, it's a problem in your code, because it doesn't run." So it was so long and so boring communications with a lot of people to prove that this is not a joke, it's a hard work, really hard work. + +**Anton Titov:** Yeah. And actually, all of the code you write is public... Because people are going to see it and they're going to blame you. I mean, they're going to blame you anyway, but now they're going to have a reason. + +**Johnny Boursiquot:** Right, exactly. + +**Valery Piashchynski:** Okay, you have a sqrt like N sqrt algorithm or N factorial algorithm. So you shouldn't do it. \[laughs\] + +**Johnny Boursiquot:** Awesome, awesome. Thank you guys so much for coming on the show and talking about really what piqued my curiosity when it showed up on my desk, my virtual desk... You know, PHP in Go, and how these two things complement each other. It's been awesome having you all... And I will now play our outro song... And we will try to wrap this up. Right on time, too! diff --git a/Go beyond work_transcript.txt b/Go beyond work_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..72f3ff6f6a60e074bd415cccd57aedf44c9f64a8 --- /dev/null +++ b/Go beyond work_transcript.txt @@ -0,0 +1,317 @@ +**Natalie Pistunovich:** Hi, everyone. Another year of GopherCon. It's so exciting to see everybody online again. I hope you had a wonderful first date, and I hope you had lots of fun in the workshops yesterday for whoever attended. My co-host for today is Angelica. Hi! + +**Angelica Hill:** Hi! + +**Natalie Pistunovich:** Joining from the overseas of New York. + +**Angelica Hill:** Yup. But soon to be flying to London next week. Finally. + +**Natalie Pistunovich:** Oh. Exciting time to fly. + +**Angelica Hill:** Yeah. + +**Natalie Pistunovich:** And we are joined by three wonderful gophers who have amazing side-projects that are written in Go. We have Sebastian Spaink, who is a software engineer at InfluxDB, and also -- you know what? We'll tell about your wonderful project right after the intro. We'll keep it interesting. For those who have not attended the talk, or your talk has not been aired just yet, because there's a few more days of GopherCon ahead of us... Sebastian, where are you joining us from? + +**Sebastian Spaink:** I'm in the Microsoft Studios, here. So I'll be doing my talk here tomorrow. + +**Natalie Pistunovich:** Oh, wow. + +**Angelica Hill:** Because everyone knows where the Microsoft studios are. + +**Sebastian Spaink:** \[03:58\] Enter Redmond, Washington. + +**Natalie Pistunovich:** It's on the internet. \[laughter\] + +**Sebastian Spaink:** Yeah. + +**Angelica Hill:** Helpful. + +**Natalie Pistunovich:** Washington. Cool. Daniela Petruzalek, you are joining us from my side of the overseas, from Europe. But not from the E.U. + +**Daniela Petruzalek:** Yeah, so I'm currently based in the U.K. I'm based in Bournemouth. That is the South coast of England, basically. But yes, Europe. Team Europe, yay! + +**Natalie Pistunovich:** Yay! And you're a product owner at JPMC, and you're a Google Developer expert for Go. + +**Daniela Petruzalek:** Yes, yes. As people say. \[laughs\] + +**Natalie Pistunovich:** And we have Linus Lee joining us. Linus, you're an independent software engineer, and you're working on creative tools. Welcome. + +**Linus Lee:** Yes, I am. Thank you. I'm also in the Microsoft Studio. I'm across the hallway from Sebastian, but it doesn't look like it, through the magic of studios. + +**Natalie Pistunovich:** Yeah, it looks completely different in your background, but the sound quality is amazing. So it's great of you all to join. We are here to talk about your fun projects. + +**Linus Lee:** Let's do it. + +**Natalie Pistunovich:** Right, Angelica? + +**Angelica Hill:** We are, we are. But before we dive into your fun projects, do you wanna hear a little bit about when and why you started using Go? Where did it all start for you? Maybe, Daniela, you could dive in? + +**Daniela Petruzalek:** Sure, sure. + +**Angelica Hill:** How did Go come into your world? + +**Daniela Petruzalek:** It's a fun thing - I used to like to tell this story, that I came for the community, and stayed for the community... Because I started with Go - it was a few years ago... + +**Natalie Pistunovich:** We see what you did there... + +**Daniela Petruzalek:** Yes. \[laughter\] + +**Angelica Hill:** Good one! + +**Daniela Petruzalek:** Nice catch! A few years ago... Anyway. Basically, back in the days I used to work at a company in Brazil, and they were heavy users of Go. I was getting engaged with the community in other languages, like PyLadies and other groups, and actually they asked me to help organizing a women's group, basically a Women Who Go chapter... And I didn't know what Go was about, to be honest. I was just helping to organize the community. Then I got the opportunity to go to GopherCon Denver, and then the magic happened. I was simply in love with the language. Everything I did after that, I was trying to use Go in some way or another. + +**Natalie Pistunovich:** So attending the same GopherCon that you're attending now. This is how it all started. Wow. + +**Daniela Petruzalek:** Yes, definitely. So it's a very emotional thing to be back here and contributing to this amazing community... Yeah, it has a special place in my heart. + +**Angelica Hill:** Well, for all of us. How about you, Linus? How did Go come into your life? + +**Linus Lee:** Yeah, so I actually started out in the web world. I learned to program doing JavaScript and HTML web stuff, and then Go was sort of my way to figure out how to do lower-level (I guess) backend programming. I had dabbled into things like Python and Ruby before, but I wanted something where I could have a little bit of a lower-level control over what my data structures look like, and things like that, without having to learn -- I guess the only viable option back then was C and C++, which were great languages on their own, but more complex than I could fit into my brain at the time, and Go was sort of this nice middle ground, where there were things like a good garbage collector, and it was dynamic enough where I could wrap my head around it, but still let me write programs that I wanted to, that were fast enough and small enough, and things like that. So that's the way that I got into it, and I've since stuck with it for all the things that it offers. + +**Angelica Hill:** Awesome. How about you, Sebastian? How did you come to us? + +**Sebastian Spaink:** Yeah - well, it was love at first sight. The moment I saw that Gopher mascot, I was like, "Let's do it. This is it. This is the language for me." + +**Angelica Hill:** Me too! I had no idea what it was, and I was like, "I just... That Gopher." + +**Sebastian Spaink:** Right. It was all I needed to convince me. Before, I was working with Python and C++, and it was kind of the same thing, where it was just like -- it felt a bit much doing the C++ and C stuff, and then Python was good, but it felt a bit messy, and I really liked Go's opinionated way of doing things... And then I started using it. I've been working on it professionally now, but before it was just projects stuff. + +**Angelica Hill:** And no one really wants a Python plushy. A Gopher is so much nicer. \[laughter\] + +**Sebastian Spaink:** Right. + +**Linus Lee:** Before I came to the conference I was talking to one of my co-workers and he asked me, "You're going to this GopherCon thing... What is GopherCon about?" I was like, "You know, it's for all the people who love gophers. We just talk about gophers for four days." + +**Natalie Pistunovich:** \[08:09\] Not showing off or anything, but here's the one from Singapore. + +**Angelica Hill:** I was gonna ask. That's what I was just about to ask; I was being like, "I've not seen that purple, and I want it." + +**Natalie Pistunovich:** It has a tale from the famous count in Singapore. + +**Angelica Hill:** It's gorgeous. Okay, awesome. So we've heard about your start... + +**Natalie Pistunovich:** Yeah, are you all still working in Go now? And how long, actually, have you been doing this? + +**Linus Lee:** I could take a stab at it... So I've been using Go in one form or another for at least 2-3 years, I think, between when I wrote my first Go project... I've written Go professionally a little bit; I've squeezed into some projects here and there that aren't what you'd call production products, but that are infrastructure, or testing services, or things like that, just because it's a nice thing that you can build. + +One of the nice things about Go is that we value stability a lot, so if you write a thing and it's not maintained daily for a few months, it doesn't break over time... But most of my usage of Go has been on side projects, on things like the thing that we're gonna talk about today, but also other things, like little chat clients that I've written... Learning exercises. + +**Daniela Petruzalek:** Yeah, I can go next. Unfortunately, I'm almost like -- I'm trying to think of the best way to say this but I barely wrote any kind of Go code in production. I spent about (I don't know) one year back when I was working at GoCardless, but that's about it. And most of my "career" in terms of Go development is basically side projects and things that I like to do, and tutorials, and things, as you already know a little bit about... But yeah, I haven't been doing a lot of Go, unfortunately. I was trying to find my way into working with Go, but for some reason it didn't happen. + +**Natalie Pistunovich:** That's super-interesting. That is unexpected, but that's cool. Sebastian, how about you? + +**Sebastian Spaink:** Yeah, I guess I've had the good fortune of being able to work with Go now professionally for the past year, InfluxData, working on the Telegraf project. It's been fantastic. Because before that it was also just kind of dabbling with it, and just whenever I could, for projects. Technically, for a half year, I "professionally" worked with Go, but just because I was just doing it on a project that nobody else wanted to work on... But who knows what that ended up with. But now I'm doing it professionally with Telegraf, and it's been fantastic. An open source project written in Go... It's everything I could dream of as far as a career. + +**Break:** \[10:29\] + +**Natalie Pistunovich:** Speaking of career, Sebastian, would you like to tell us about your fun project in Go? + +**Sebastian Spaink:** Awesome! Did you 3D print that? + +**Natalie Pistunovich:** For everybody who is listening in but not watching, I have a little 3D printed gopher that I received from a colleague in a previous company... And he 3D printed it. I don't think he planned to print gophers, I don't think his team works in Go, but he is a fan of the mascot, so that was a nice gift. + +**Sebastian Spaink:** Yeah, that's awesome. I should have brought some myself, that I 3D printed... But unfortunately I've got nothing to show off here; I guess you'll have to wait for my talk tomorrow. I've got some screenshots of them. So during quarantine one of the craft projects I started was 3D printing, and the first thing I did was I created a thermo-detonator prop from Star Wars, and then used TinyGo to make the LEDlight shine and play some audio... So that was pretty fun. And then I kind of started looking at other ways I could involve Go with 3D printing. + +So they've got this whole thing called OctoPrint that helps monitor and remotely-control your 3D printer, and I was kind of looking at ways to extend that using Go. But yeah, it's been fun. It's not like I'm writing the firmware for the 3D printer or anything, as far as Go was concerned; it was kind of like building around it. And also, 3D-modelling gophers to 3D print as well. So that's the main thing I was doing there. + +**Natalie Pistunovich:** Yeah, that's not an obvious choice, to choose Go for 3D printing, for the modeling. + +**Sebastian Spaink:** Yeah, which is actually pretty cool. I wasn't aware that you could do it either, because I was mainly using Blender and common open source projects for 3D modeling that use user interfaces... But there is a package out there that you can create 3D models using Go, and yeah, it was pretty cool. It's just like primitive shapes kind of glued together... I'll show a screenshot of that tomorrow, but you end up with a very round-looking gopher, because it's just a sphere of other primitive objects glued on it... But yeah, it's all written in Go. So a gopher written in Go is possible. And the open source package I used was called -- I forgot the user's name, but sdfx, I believe is, if you look that up... Yeah. Model your own gopher with it. + +**Linus Lee:** And you're already talking about spheres and just building shapes out of spheres and it reminds me of what I built when I was looking and building ray tracers. And the sphere is just the simplest shape to write a ray tracer for, so you start just building scenes our of only spheres, and everything -- your ground is just a gigantic sphere with an infinite radius, and you're building raindrops on a giant sphere. What does it actually look like to build a model with Go? Are you outputting it to a file? Are you talking to an API for the printer? + +**Sebastian Spaink:** Yeah, so what we output is an STL file, and it's like the standard format file that you're gonna do in 3D printing. It's a file format that describes the geometry like a bunch of triangles... It's like magic, but they use something called signed distance functions to define the primitive shapes in the code, and then output that to the STL file, which you can then slice into G-Code for the 3D printer to print out. + +**Angelica Hill:** So how many times did you have to reprogram it if you actually got a beautiful gopher? Because I would assume you had a few interesting-looking shapes, that may have resembled a gopher, but may not have been a gopher. + +**Sebastian Spaink:** Yeah, making its eyes and its pupils look like an actual gopher shape - that's tricky. But I think I got something that looks like a gopher. You've got your iconic mouth and teeth, and then you're like "Bam! It's a gopher. Done." It's a sphere with eyes, and a square, but that's all you need. The beauty of the gopher transcends all shapes. + +**Daniela Petruzalek:** I must say, I really admire anyone that can do 3D modeling. This is such a thing that my brain can't process. That's why when I go to my hobbies, I only do 2D things, because my brain just can't process a third dimension. I really admire that. Looking forward to seeing your talk, how it works. + +**Sebastian Spaink:** \[15:57\] Yeah. I did cheat and use Blender to make it sexier-looking than the spheres, but... Yeah, it's definitely my first thing I've ever 3D modeled before. Usually, I'm also just working on the 2D space; it's definitely more comfortable. + +**Angelica Hill:** And did you use Go because it was the best language to do this, or because you write Go anyway, or did you think about using other languages, try it out, before settling on Go? + +**Sebastian Spaink:** That's a good question. No, I chose it because I wanted to write in the language. Because technically, Python is what OctoPrint, that software I mentioned, is written in, and that's the ecosystem... There's a lot of Go community work around it, but I was like, "No. I'm gonna against the stream, do it in Go, regardless if it's the best choice or not. I'll find out at the end." + +It's pretty nice to work -- you know, just all the benefits of using a Go binary that you can send around easily, and don't have to worry about Python versions, and setting virtual environments... + +**Natalie Pistunovich:** Interesting. You can cross-compile to all the different 3D printers. + +**Sebastian Spaink:** Right. + +**Linus Lee:** Is there anything that using Go allows you to do that you wouldn't be able to do with Blender? I guess you could programmatically generate a bunch -- like, if you want a grade of spheres you can programmatically generate it... But yeah, what have you experimented with it that would be hard to do in Blender or another piece of software? + +**Sebastian Spaink:** I am pretty much a novice as far as 3D modeling is concerned, so the reason why I used Blender to finalize the gopher -- because I couldn't really figure out how to make a nice hourglass shape for the gopher... You know, that perfect body, so it wasn't just to spheres attached to each other... \[laughs\] But I guess the benefit of using Go for it is the fact that you can now just send that piece of program to somebody and they can adjust it. It's easily shareable. + +I mean, right, you could send your blender files, share those as well, but I felt like the learning curve was also a little bit easier; the fact that it is just Go, and just defining shapes, while using the user interface for Blender is tricky. There's definitely a steep learning curve there. Have you used it before? Have you done 3D modeling? + +**Linus Lee:** I have personally not. I've looked at it... I had some friends -- my roommate in college was a mechanical engineering major, and he dabbled in it, he would make some stuff for me occasionally, and I always thought it was cool. But again, because of that learning curve, when you look at someone's computer when they're doing it and you're like "I'll never be able to do that." That looks like a professional power tool kind of interface, and that scares you off a little bit. But if it's just a Go program, maybe I'll give it a shot. + +**Sebastian Spaink:** Yeah, I definitely recommend it. + +**Angelica Hill:** I feel like I need to go buy myself a 3D printer now... + +**Sebastian Spaink:** I also recommend that. + +**Angelica Hill:** ...and give it a go. \[laughter\] + +**Sebastian Spaink:** Yeah, and there are surprisingly consumer-friendly as far as the price and the setup is concerned. You'll need space, I suppose, is the biggest thing. And patience. + +**Angelica Hill:** I'll just put it right in the middle of my lounge. It will be the centerpiece... And I'll just continuously churn out gophers. + +**Sebastian Spaink:** That's the ideal setup. That's how it works on my house. + +**Angelica Hill:** Yeah. \[laughter\] + +**Linus Lee:** A thing that you alluded to earlier, from another question, was the debug cycle, which I imagine is a little slower than just running the code. I guess you can run it and look at the model inside Blender, or something like that... But do you ever have like a hardware debug cycle, where you print it and you're like, "Oh, this looks bad", and then you have to rewrite the program? + +**Sebastian Spaink:** I kind of missed that question. Yeah, that's right... I don't know if there's a better way of doing it than just building it and running it and being like "Well, it's a couple millimeters to the left", and then scoot it over. A big advantage of using Blender, if you use their interface, you can just scoot it over, and you can see where it's supposed to land... But in the world space, I was definitely just guessing the coordinates. I'm definitely gonna mention that in my talk, where it's like, "These numbers are made up." This was just me hoping for the best, and they ended up in the right spot... So yeah. + +**Angelica Hill:** Trial and error. + +**Sebastian Spaink:** Right, yeah. + +**Daniela Petruzalek:** I had an interesting experience with that when I was writing some games. Also, you do a lot of guessing in terms of positioning things... So in the end, I ended up using Steve Francia's -- I don't remember now if it's Viper or Cobra; that library that helps load the configuration dynamically... So basically, every time I had something I need to set up a different thing on the screen, I used the dynamic configuration and it was reloading JSON files on the fly... And you just couldn't see anything, because the renderer was just updating the screen every single frame. I was just kind of automating my trial and error with dynamic configuration. I don't know if you thought about something... I'm not sure how fast is rendering, but in my case it was just one frame, so that was pretty fast. + +**Sebastian Spaink:** \[20:18\] Yeah, it was pretty quick in the sense that there was like -- I was using the Slicer program to get it to the 3D printed format, and I just... You can reload it in there. It felt pretty hacky, but -- yeah, I'm sorry; what were you gonna say? + +**Angelica Hill:** I was gonna say "Let's hear more." Let's go from 3D to 2D. Daniela, I would love to hear a little bit about -- you talked about some of your in-the-weeds, how you got it working, how you tested it... But what is your project, and how did it come to be? I would say, and I've said this to the group before, I have a very, very special place in my heart for this project, because when I was but a young gopher... Not even. When I was but a young newbie, dweeb, looking at the computer science world for a language that fit, I went to Gotham Go and saw you do your talk on this side project. And it got me so excited I went home and spent the entire weekend building it. And that was one of the main things; I was like, "This is my language. This is so fun. I'm ready. Let me be a gopher." + +**Daniela Petruzalek:** Right. I love this attitude, and I'm really happy that it had such an influence because from the beginning, since people don't know my project yet... Basically, I wrote this tutorial which I call Pac-Man From Scratch, or Pac Go for short, because it's not a man, it's a gopher, maybe... And the idea is to rebuild the classical game of Pac-Man only using Go and the standard library; no external fancy packages. And in order to abstract the drawing function - because this tends to be the most complex thing in games, how you render it to a screen, and I didn't want to expose people to this level of complexity... We ended up using the terminal as your screen. And Cisgo supports Unicode. You can actually bring emojis in the terminal. So the entire board of the game is rendered using emojis, and using some clever tricks with ancy escape sequences. You can put colors, and draw walls, and things like that... + +There are some things that Go makes really easy to do, especially -- for instance, if you want to separate your input handling from the main program, you can use a goroutine, and channels... So things like this in other languages - I tried to do this before in C++ and others, and you need to learn threads, you need to learn synchronization, all these more advanced concepts, but Go makes it really easy for a beginner to have a look, "Oh, yeah." Maybe you don't even fully understand what a goroutine is or a channel is, but it's enough for you to get started and do something. + +The whole idea of this project was to give the community something in terms of a tutorial, a starting point, a bit more interesting than building APIs. Because I know, APIs -- we ended up doing this for work. Why not having something more fun, that has visual feedback, so that you can see things moving on the screen? I don't know, for me it's magical, but maybe that's because I also -- the only reason why I'm in this industry now is because I love games, and I always wanted to work with games, and that was kind of like my "Oh, yeah! I've never worked with games before, but this is my opportunity. I'll make something with games and Go." That's how it started. + +**Angelica Hill:** Yeah. And I think the beauty of it is exactly what you've just said, that people who are learning Go, whether from another language, or completely it's their first language, you go through that phase you're like going through all the Go tutorials, and you're trying to do this little app, and you're trying to print Hello World... But then you hit this ceiling where you're like, "What do I do now? I know the basics. I wanna build something that feels like an accomplishment, that feels exciting, that feels like "Oh! This is something I wanna show my friends, show my family, and be like "Look, I can code now!" + +\[24:02\] And this is fun, it's colorful; as you say, it gives you that visual feedback, and it gives you, as a beginner, that feeling of accomplishment, and like "I did this. I've made it through that tutorial, and I accomplished this thing." + +I think the beauty of how you framed it is that you've really paid attention to making sure it is truly step by step, and it really does guide everyone through why you made the decision, how you did it, you provided the code so that you can look through... I'm waiting for your next game to come out. What's going on...?! \[laughs\] + +**Daniela Petruzalek:** Yeah, I've been a bit busy with life, and things... Definitely, there are more things coming your way, yeah. Stay tuned. I did actually a talk at GopherCon UK about my next project, which now is a card game. I'm rebuilding an old '80s card game that -- I can't even say it was popular where I came from. It was just the game we had when we were a child and we played that game. + +I couldn't find an international version of it. I only had a scan of the cards that I used to reference the game, but basically, I'm rebuilding it. But now I'm trying to be a bit more professional, if I can say that. Definitely, I can say that. Nothing close to being a professionally, but... I'm using this library called Ebiten, which is a proper game development library within Go by Hajime Hoshi. This guys is amazing. Thanks for writing Ebiten. + +Yeah, so this work in progress - it's not very pretty now, but I have a functioning prototype, and the code is open source, but I didn't advertise it that much, because the game is ugly now. It's not very beginner-friendly. It was just kind of like a proof of concept, trying different things, learning everything as it goes... As it goes, I did that again. \[laughter\] + +**Angelica Hill:** It's hard not to say it. It's very easy to call into the pattern. So as you've been iterating on this project, as you've been building, and as you've been delving more into the world of games and programming for games, is Go a good language for these kinds of projects? Or was it that you already liked Go, and it was fun? Or in fact is Go interesting, good to play with, but there are other languages that may serve the gaming function better? + +**Daniela Petruzalek:** I think Go is a very good language, with some asterisks, some side-notes maybe... Because the traditional language for writing games if you are old-school like me is C++, and C, and things like that. I tend to see Go as an improved version of those languages, because it abstracts so much. You don't need to worry so much about memory management... You still need to worry a little bit about memory, but not as much as you had to in the past with these other types of languages. Of course, they also evolved, but I really like how Go is pretty straightforward, it has a simple syntax, very consistent across all different features of the language... It gives almost that feeling of a scripted language. It's very simple to do simple things, and I think that's the build of it. It also becomes very productive. + +The problems come -- because Go is traditionally a systems language, it's not very popular in the gaming industry, you don't have a lot of library support, and things like that. The basics you have - you have SDL bindings, you have Ebiten and a few other gaming libraries, but it's not widespread, so you don't have a large community that can help you support it. But essentially, everything that you need to build a game, you have there. I love the language, and I think that's almost like a -- Sebastian said that as well, maybe Go was not the best option... But that was the language I wanted to work with, so that's kind of why I decided. + +But overall, I think that you can do good games with Go, and even Ebiten - they are just recently publishing their first game on Nintendo Switch... So you can see, it can be used for real games, like real commercial games. + +**Break:** \[28:05\] + +**Natalie Pistunovich:** Daniela, I'm so very curious, as somebody who knows nothing about game development - when you say "gaming libraries", what kind of tools do they give you? What do you need in a gaming library? + +**Daniela Petruzalek:** A lot of things, actually... But I would say the most basic ones are how do you handle inputs? When you're writing a game, you don't really want to care if the player has a keyboard, a gamepad, or anything else. You don't want to worry about specific drivers for different types of inputs. You want a game library to give you this for free. Once you know "key was pressed", you do your mappings into the thing. + +The second thing is drawing to the screen, another good example. You don't really care about the low-level part of drawing to the screen, you just want to say "plot my pixel there", and that's it. + +**Natalie Pistunovich:** So like relative 10% down from the beginning of the screen, kind of... + +**Daniela Petruzalek:** Yeah. And I'm talking only about 3D; and 3D gets way more complex, but... Scaling operations, rotation... The kind of transformations you can do in an image. And also, sound is a huge thing. I don't want to write a music player. I should get this for free from my gaming library... Things like that. Maybe sometimes you will have some kind of artificial intelligence processor that you can just write an AI as in a script, and the engine will load that script and control your enemies, and things like that. So this is kind of support... So you can focus really on the content creation. Because what makes or breaks your game is really not necessarily the code, it's the art, it's the sound, it's your assets. Actually, it's all about your assets. + +**Linus Lee:** That resonates with me a lot, actually. I'm not a big gamer, but occasionally I've come across little Steam games that I thought were really beautiful and well done. At times, I've had the thought of "Oh, maybe I should try to build a game of my own", and the thing that I always bump against -- so I can write the logic, I can probably learn to use one of the game engines and try to make something that's interactive and do the things that I wanna do, but the things that make a video game really compelling is all the art, and all the sounds, and the design... That requires making 3D models, which as we discussed, is difficult, and doing art, and having -- even just building a storyline behind it, and all the character design, and things like that... That's a whole other art that's outside of software. + +**Daniela Petruzalek:** Yeah. There's a huge iceberg of game development that's about -- the presentation is just a thin layer, but you have assets, in terms of art, sound effects and music. It's also -- it either kills your game, or it will give amazing results. + +But also, the whole part about game design, of how your game mechanics work, like what are the winning scenarios, what are your objectives, how can you make that thing rewarding for the player, so they will be engaged and really enjoy playing your game? How can you balance your game? This is a decent amount of challenge, plus reward... So these are the hard parts, and also, my brain is kind of -- I'm really bad at this, and I'm trying to get better, but also partnering with the right people can definitely help a lot. + +**Natalie Pistunovich:** I have a great idea for your game, Daniela, to make it pretty rewarding... You can use Sebastian's work and 3D print the cards. \[laughs\] + +**Daniela Petruzalek:** I think that's a great idea. + +**Natalie Pistunovich:** \[32:10\] You know, you always feel paper cards in the hand, and once you'll feel more chunkier, reminding of like a phone, it will be very interesting. You'll have to pay more attention when you do that. + +**Sebastian Spaink:** Revolutionary. + +**Natalie Pistunovich:** \[laughs\] + +**Angelica Hill:** Talking about something that's very challenging, engaging and rewarding... Linus. + +**Linus Lee:** Yes. + +**Angelica Hill:** Let's hear a little bit about your awesome project. + +**Linus Lee:** Yeah, so my project is called Oak. It's a toy programming language, is what that means. It's a programming language that you can use to write things like web servers, and little apps, and command line interfaces, and things like that... But it's a toy language, so it's not something that you're gonna use at work, or to build a production service, for lots of reasons. But it's dynamically-typed scripting language, it looks a little bit like JavaScript, it works a little bit like Lua if you've used it... But it's a Go program. So it's a language that's written in Go. And what that means is that when you run the Oak CLI to start up a repl, or to run a file, that thing that is running the file is a Go program. But the program's source code that it's running is written in a language that I made up, called Oak... If that makes any sense. It's a little bit sort of recursive. + +So part of it was designing the actual syntax and semantics of what the language is gonna be able to do, how it works, what are the types, what are the values, what are the things you can do, how do you define functions, and things like that... And the other part is actually implementing, in Oak's case, the interpreter, that actually takes the syntax, the program you've written, and does the thing that the program is supposed to do... And also, incidentally, other things, like telling you errors, and concurrency, and lots of other things that you might encounter as the interpreter. + +**Natalie Pistunovich:** How did you come up with the idea to do this in Go? Did you say "I want to learn Go, and I want to learn Go by creating a new language"? + +**Linus Lee:** It was a mix, actually. So before Oak, there was Ink, which is the conceptual predecessor to Oak. They're very similar languages, but Ink was the first toy language that I made. I made it after having gone through a tutorial on how to make a basic Lisp interpreter, and I wanted to make something that was a little more usable, and had the syntax that I wanted... Because everyone, always, when they first learn programming, they're like "I really like this language, but I wish this keyword said something different", or the symbol was a different symbol... And it was my chance to play God with my own language. + +At that point, I was mostly proficient with only the really dynamic languages, like JavaScript and Python, so I needed to learn a compiled language that was lower-level, to be able to build an interpreter that was actually usable. + +Go is not the typical language that you would use for such a task. You usually use like C++, or more increasingly common is Rust, or other languages like that. But as I've noted before, it was harder for me to wrap my head around those at the time, than it was for me to learn Go. Go was low-level enough for me to be able to build an interpreter. Ink was my first learning Go project, and simultaneously I was also trying to learn how to build a programming language. + +**Sebastian Spaink:** That's super-cool. Has anybody made anything real cool out of your programming language Ink? + +**Linus Lee:** I have... And then other people have used some of the stuff that I've built. Actually, one thing that's relevant for the Go community is - if you guys are familiar with the Go By Example website, where it sort of walks you through Go, and gives you different examples of how different parts of Go works... A guy named Andrew -- man, his name is escaping me now. But Andrew-something - we've talked on the phone before, but he made an Ink By Example site. And just as Go By Example is written in Go, and Go compiles the website, Ink By Example is written in Ink, so it compiles itself... And Ink has its own syntax highlighter, and code formatter, and things like that. + +So the website is inspired by Go By Example, but it's a website for Ink that someone else made, that isn't me, which was super-cool to see that there's another user for the language. And then on top of Ink, I've made a bunch of my own personal apps, like Twitter clients, and notes apps, and things like that, that I use day to day. + +**Sebastian Spaink:** That's very cool. + +**Daniela Petruzalek:** You mentioned Lua in your opening statement... I was just wondering, because Lua is very common in the game development community, for AI scripts. + +**Linus Lee:** Right. + +**Sebastian Spaink:** \[36:05\] Would you see your language could be used for the same type of use, like writing scripts for AIs, or things like that? + +**Linus Lee:** Yeah, that's a great question. So a lot of the domain of tool languages is actually fr embedding in other programs, and Lua is a great example of that. I think there's some specific things about Lua that make it really great for embedding in a game, specifically. The things that come to mind are -- Lua is a really small language, it's written in C, it's pretty readable, so it's fairly easy to embed in terms of just like the technical work involved, to embed it into another program. + +The other thing that makes Lua unique is that it has features of the language and of the interpreter that make it easy for you to fit it into another larger program. So Lua has -- I forget what it's called, but it has like a type that is supposed to be like an opaque type, that just wraps something else from the host program... So if you have a game, and you have like a C struct that represents a character, or an object, or something, you can just wrap that and expose it to Lua with some APIs. It's also fairly easy to define foreign function interfaces for Lua code so people could call C functions... And I think that's what makes Lua really great for game development. + +Oak, in theory, has those things, minus the C FFI, but because it's embedded inside Go, anything that includes Oak has to also include the Go runtime, which makes it a little heavier, and is more opinionated. + +**Natalie Pistunovich:** How can the community contribute to these languages, to Ink and to Oak? Are you looking for documentation, code contributions...? + +**Linus Lee:** Yes, for sure. So one of the reasons I made Oak was that I made Ink, and then I used it for a couple of years to build stuff... And it was usable, because I made stuff with it; by definition, it was usable. But it was not very good. The language was kind of tough to use, the interpreter sucks a little bit because of the particular ways... + +When you learn to build an interpreter in school or in a course, there are specific things that you spend a lot of time on, like parsing and compiling... And there's things that you don't spend as much time on, that it actually turns out are really important and hard to get right... And one of those things is error handling. If you're trying to get other people to use your language, most of the time they're gonna be spending initially is going to be writing incorrect programs and getting errors. To do error reporting correctly, you have to gather a lot of information about the runtime of a program, like where it's happening, what's causing it, and be able to give users really helpful errors. And none of the tutorials cover that properly. Mistakes like that accumulated over time, and eventually I went and wrote Oak, and fixed a lot of those. Ink isn't as usable as Oak. If you wanna check things out, I'd check out Oak. I'm still sort of tinkering away at it as a solo project, but if want to try you're welcome to try it. And then if you have any problems you find with it, or bugs, filing those would be super-helpful. + +If you have opinions on how Oak should work differently - I have my own opinions, and the point of Oak is to manifest them... But if you have your own opinions, it's a Go project; it's not that complicated, you can probably read through it, and come to my talk to figure out how it works. And it's only a few thousand lines, so you could fork it, make it into your own, change all the keywords and all the syntax, and make your own language out of it, too. + +**Angelica Hill:** That's awesome. Tempting. How about Daniela - if people wanna help, wanna get involved, wanna live their best PacGo life, how do they do that? + +**Daniela Petruzalek:** So I have the GitHub repo for PacGo is github.com/danicat/pacgo. We already had actually a lot of contribution from the community. I think I've built it to step eight, and I think we now we have step ten, or eleven, or something like that. So people definitely were inspired and started creating new content for it. It's a bit abandoned now, from my perspective. I wish I could have supported it better. So things like Go modules, testing, are some topics that were not covered in the traditional workshop... But I think if people are willing to contribute, that would be amazing. A chapter on Go modules, a chapter on testing, things like that. That would be great. + +**Angelica Hill:** Sebastian, how do we contribute? How do we help? How do we get involved? + +**Sebastian Spaink:** Yeah, I guess I'm mostly thinking of using Oak to write a Pac-Man game now... \[laughter\] But yeah, I guess I've got my gopher model uploaded, so you're welcome to download and 3D print that... That helps me, just seeing more gophers out in the world... But then part of the thing I worked on was a plugin for the Telegraf project that I'm working on now as a full-time maintainer, to interact with OctoPrint. And that's just an open source repository, you're welcome to do whatever you want with it; it's written in Go. Maybe extend it, add more features to it. + +**Angelica Hill:** Awesome. + +**Natalie Pistunovich:** Cool. Well, that's inspiring, and also gives us all ideas on what can we do in our free time. Okay, time for Unpopular Opinions. + +**Jingle:** \[40:27\] to \[40:43\] + +**Natalie Pistunovich:** Okay, so you know the rules... You have the stage to tell us what is your unpopular opinion. It does not have to be about your project, about tech, about anything at all. It can be about any theme. And we will then go and ask on Twitter to see if your unpopular opinion is really unpopular, or yes-popular. Linus, would you like to start? + +**Linus Lee:** Sure, I'll take it. My unpopular opinion is that -- it's actually probably a more popular opinion, but the way that I'm phrasing it I think is important... I think it's an accident of history that we don't expect every computer user to be able to program. I've been watching a lot of... And this is gonna reveal my age now, but I've been watching what I consider to be older TV programming about computers - computer programs that were coming out in the mid '80s, early '90s - and it's kind of astonishing how all those programs sort of expect the user to buy the program and then use macros or scripts to program it to fit their use case. Things like databases, or word processors... There's tons and tons of support for just -- you buy the thing, and then to make it useful, you program it. And to a certain extent, all these consumers of software expect it to program a little bit and know how computers work and customize it... And it's a little sad, I think, that we don't expect that of users anymore, and we expect them to just take the product and use it. + +I wish there was more software that expected and taught people to think in terms of programming and try to customize their software to work the way that they want it to. + +**Natalie Pistunovich:** Okay. Sebastian. + +**Sebastian Spaink:** Yeah. I guess my unpopular opinion is I think that if your open source project has a hand-drawn logo on it, then it's a good open source project, and it can be trusted. So you can just it by the cover. + +**Natalie Pistunovich:** Cool. Hey, Daniela... + +**Daniela Petruzalek:** Okay, I think my unpopular opinion will make me the most unpopular person in GopherCon this year... Sorry, guys, but I don't think Go really needs generics. \[laughter\] + +**Natalie Pistunovich:** What a mic drop to the end of the show. Thank you, Daniela. We knew you'd say that, maybe... That's why we asked you to go last. But yeah, that's a brave choice. + +To wrap this up, thanks everyone for your interesting insights and sharing about your fun projects, and for talking about that at GopherCon. Thanks everybody who tuned in and enjoyed the talks, and see your around on Discord. + +**Linus Lee:** Thanks for having us. + +**Sebastian Spaink:** Yeah, thanks. + +**Angelica Hill:** Thank you to our awesome host, Natalie. Awesome as always. diff --git a/Go code organization best practices_transcript.txt b/Go code organization best practices_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..358bd2c6ce4dea05913b70b50b5dbfaa973d3fc6 --- /dev/null +++ b/Go code organization best practices_transcript.txt @@ -0,0 +1,253 @@ +**Kris Brandow:** Welcome, gophers and wonderful listeners out there. I'm your host, Kris Brandow. Today on Go Time, we are going to be taking up one of our wonderful audience members' suggestions for a topic. So Thomas Eckert sent in an inquiry for us to talk about utility functions and code that we have in Go, and how we can organize that, where we should put it, and in general, what should we do with that code. + +Today I am joined by my wonderful co-host, Johnny Boursiquot. How are you doing today, Johnny? + +**Johnny Boursiquot:** I've been better, but I'm getting better. + +**Kris Brandow:** It's always good to hear. We appreciate you being here. And we're also joined by a now reoccurring guest, Ian Lopshire. How are you today, Ian? + +**Ian Lopshire:** I'm doing great. + +**Kris Brandow:** Awesome. + +**Ian Lopshire:** Yeah, happy to be here again. + +**Kris Brandow:** Fantastic. So I'm sure that we're going to get straight into the meta, since both Johnny and I are in this, or on this podcast today. But I guess we can start with something kind of high-level. So I guess I can read the actual suggestion, and then I have something I want to prompt us to start with. + +Basically, Thomas wrote in and asked, "There are a lot of utility functions that I end up writing myself in Go; these are the kinds of functions that get asked about on Stack Overflow, and met with the response of, "This is trivial to implement, do it yourself." For example, contains slice, string of slice, value slice, bool, which is just a regular contains function for an item in a slice, to determine if a slice of strings contains a particular value. "Where should I be putting these functions? I usually end up with a drunk drawer package of functions, like my own standard library, or I define them one-off within a file or a function. How do I organize the code that isn't in my business logic, but also isn't in the standard library? Additionally, how might this change given generics being introduced to the language? Could we get a generics contains function? Should we?" + +\[04:19\] So the question I want to start off - basically, have you guys run into this before? Have you had the struggle of "I have this code that's somewhat generic, somewhat reused; where should I put it? I guess, Johnny, do you want to start off? + +**Johnny Boursiquot:** Yeah, I can start off. I think we've had various conversations that touch on this topic on the podcast before. Oftentimes, I think those of us that are sort of experienced with the language sort of typically settle on this rule of three, kind of thing. + +For me personally, I don't consider something reusable, unless I've actually reused it; meaning that I don't try to sort of preempt the process of prematurely trying to make something reusable, even if I have a hunch that it is going to be reusable. + +So to piggyback off of the suggester's explanation of utility functions and things like that, like contains, or whatever it is - these things on their face seems like, "Yes, that should be a reusable thing." But for me, until I've seen something at least three times, I don't have enough data to understand all the edge cases, to understand the variety of ways a particular piece of functionality or even domain within my application, within my ecosystem, how that's going to be reused. So I don't consider anything reusable, unless I've actually reused it a few times. + +**Kris Brandow:** Ian, what are your thoughts on that? + +**Ian Lopshire:** Yes, I mean, I definitely agree with that kind of rule three idea. I would much rather copy and paste code multiple times, than try to make an abstract version that is bad. It's so much easier to add an abstraction later than it is to take one away. So I’d much rather just copy and paste. + +**Kris Brandow:** Okay. Yeah, I definitely agree with that, too. This is something we're gonna get into a little bit later, but I think sometimes people try and overdry their code, where it's like, "Oh, no, I've written this once. I have to figure out how to never write these exact same lines ever again, and put them somewhere useful." So yes, I definitely agree with you on that rule of three there. + +So I guess like on that as well, let's say that you do figure out that you have something that you've used three, four or five times, the code is very similar... What other heuristics do you use to figure out like, "Okay, should I just be copy-pasting this around?" Kind of as the Go Proverb says, "A little copy is better than a little dependency", or at what point do you decide, "Okay, I need to like put this somewhere that is reusable"? And I guess they’re obviously reusable within the same codebase, but then maybe microservices reusable between codebases. How do you start to figure out where those lines are? + +**Ian Lopshire:** I have a fun heuristic for this... If I can think of a good package name for it, that makes sense, I think that means it's reusable enough that I should probably break it into a package. But if it's going to be like this package that's four words long, I'm just not going to. I'm going to just put it where I need it. + +**Johnny Boursiquot:** Mine is more of a sort of, why do I think I need to make this reusable? So there's a reuse - I think you hinted at this, Kris... There's reuse basically within the same code base. There's also reuse within multiple projects. So I could extract away a package and make that reusable. But interestingly enough, I don't know whether maybe it's the problem domains I work in these days... This could be considered a dirty little secret, but I try not to sort of create packages, even if I happen to reuse them within the same code base. Very rarely am I sort of abstracting a package and making it sort of reusable across projects. So I tend to... If I have enough data to make something generic - not generic in the sense of the feature, but generic in terms of you reusability - I'll try to make that happen within the context of the particular application or service that I'm working in... Because I don't want to create a standalone package that other teams are going to depend on, and now I'm on the hook for keeping that thing up to date. I'm on the hook for managing and maintaining that thing. + +\[08:28\] So I'd rather keep reusable chunks of stuff in my own project, because I know the moment I make it reusable by other teams, the responsibility scope for that code has just grown for me. And honestly, perhaps this is an unpopular opinion, or perhaps I'm just getting to that point in my career where I just am not looking for more work to do... \[laughs\] I'm just keeping it constrained within my projects. If somebody happens to know of my product, whatever I'm working on, and they come, "Oh, can I reuse that?" "Yeah, sure. Just literally copy the code. But you know, I'm not making this into a separate package for you; just copy the code." + +**Ian Lopshire:** I'll mirror that. I'm a fan of the internal folder; like, I'm not even allowing you to use this. + +**Kris Brandow:** I think in this case, using the internal folder to be like, "No, no. I know you might want to use this thing, but please don't. Please don't touch this thing." But I've also seen the other extreme, where it's just like, "Okay, we're just going to assume that by default, and everything's going to go under the internal package. And that's just going to be like the top-level thing in our Go codebase, and maybe there's a couple things on the outside." Do the two of you kind of agree with that, or do you think it should be just like, "Okay, no, I've made an explicit decision. This is something that perhaps someone might want to use, so I'm going to mark it as like, "No, no, no, don't depend on this, because I don't want to maintain it", or are you on the side of like, "No, it's okay to just kind of put everything under Internal and pull things out as necessary"? + +**Ian Lopshire:** I don't have a habit of doing that. What I do have a habit of is just not exporting anything I don't have to, so they're not going to be able to get to it anyways. But I'm known to put kind of very general packages and internal just because they look appetizing to use externally. + +**Johnny Boursiquot:** Perhaps my other dirty little secret - I don't use Internal a ton. Even then, I've heard some folks basically approach projects with basically saying, "You start with Internal, and then if you need to, take things out." I don't take that approach. I tend to sort of reason about the stuff I'm working on in terms of domain-driven design, and what is it that I'm working on, and then I worry about sort of shuffling files around, and then I worry about packaging. + +I will literally have everything into a single -- sort of at the root level of the project before I even start to create folders, creating packages and things. Because I don't know what I don't know yet. So I have everything in the same top-level package, and if I know I'm creating executables, one thing I absolutely do do is I create a CMD folder, followed by another sub-folder to represent the name of the binary that I'm going to build. But at the start, that is literally the extent of my organization. + +Back in the very early days, I started following this pattern of putting things in a pkg folder, and all that jazz... Honestly, I think that's an anti-pattern at this point. Like, you don't need a pkg folder. If you happen to work on a project you've inherited that has one - fine, so be it. But you don't need these things. + +So I really rely on sort of a complete understanding of the problem I'm solving. I'll even go as far as to actually have in my cmd slash whatever name folder I need, I'll even have my main.go in there, I'll have other Go files in there. Literally, I'll have everything that I need to make this binary -- especially at the start of a project, I'll have everything I need to make this binary build to a proof of concept, whatever it is. Literally, I'm not thinking about packaging or extracting or moving anything. I just have everything in that folder, then if another folder underneath my cmd now needs some of that same functionality, then I'm like, "Okay, well, I have multiple binaries that are using these constructs and things; maybe now it's time to start extracting some things away from the first folder and move that up a level, underneath the main repository, and then basically start referencing these things from those sub-folders and from those executables." + +\[12:23\] It's all incremental. Even if I've built this kind of project before, I know, I kind of have a sense of where the seams are, I know what I'm going to do. I have the discipline or I've developed the discipline to not jump the gun on that. I really need to start seeing the edges of the service, and sometimes I will be completely happy with a bunch of stuff being in the executables folder, because I don't yet have a reason to move these things out. + +Again, basically, you don't know what you don't know yet. And sometimes, you build these projects, and they might be proof of concepts, and they might be things that don't end up going into production... Why am I going to expend all this brain energy trying to make all these reusable things and abstractions and things, and I'm writing way more code than I need to, trying to avert cyclical dependencies, and trying to do all the -- that's just too much work. Again, I think I'm getting to my grumpy old man stage of Go development, but I'm just not trying to do more than I need to. \[laughs\] + +**Kris Brandow:** Right. So the way you follow things is like start with just one package, maybe it is even just like the main package, and then as you have reasons to start pulling things out, that's when you start pulling things out... Which I would also say is a very good way to avoid having people import your code; because if it's in main, they can't. + +**Johnny Boursiquot:** \[laughs\] You can't import main. Right. + +**Kris Brandow:** Like, "Stay off my lawn, you can't touch any of my stuff." So I wonder, does that change when you go from building an application that's going to run as like a microservice or something, when you shift from that to writing a library that is going to be consumed by other people? Do you still take the same single package approach, or do you start at that point trying to like predefine where things might be laid out? + +**Break**: \[14:12\] + +**Johnny Boursiquot:** I pull a page from Ian's book, and basically saying, "Hey, I don't export anything I don't want you to use." So even then I might not have an internal folder inside, even if it's basically a package that’s designed to be used as a library... Because the moment you make something a library - again, going back to the scope of responsibility; it's not just the code you're now putting out there; you've basically taking on the responsibility, if you're a good maintainer, of maintaining this thing, and keeping up to date, and who knows, there might be security things you need to do for that package, depending on what the problem domain you're in... You have taken on the responsibility of keeping this thing up to date for the users of your package. + +So I will never expose anything prematurely. Everything will be unexported until I have a reason to export it. I used to have a bad habit of actually exporting interfaces out of my libraries... That says, "Oh, yeah, like you can use this public interface, this export interface, so you can know what to pass in, when you call in some functionality into my package, whatever it is. Here's some nice interfaces to help you out with that." I don't do that anymore. + +Right now, I rarely, if ever, export an interface from a library. Because you can actually, in your own code, you can create single-method interfaces, whatever you need. You can create local to your code interfaces for whatever you need. I don't need to give you exported interfaces for you to know how to use my library. As long as I include a document, the exported functions, as long as that makes sense, the naming of things; that's still very much important, whether it's a library or an executable. I make sure that the naming and documentation are up to snuff. I never expose more than I need to. And even the unexploited stuff, I will document them as if I'm going to export them, because who knows, I may end up exporting them. + +There's a way using GoDoc where you can actually show unexported documentation for functions, and stuff. So when I need to actually, on the command line, see what unexported functionality documentation is, I can actually do that. So things are documented as if I'm going to make them available for people to use. But I don't do that, unless I absolutely need to. So I keep my packages surface area for interaction, very small. Only doing just what it needs, never more than that. + +**Kris Brandow:** Yeah. Ian, what do you think about that? + +**Ian Lopshire:** I mean, it's hard to disagree with anything that was just said there. I think that pattern of just starting with one bit, like one package, one thing, even if you're just adding a feature - like, I like to start in one place, and do that proof of concept; and it should be that iterative approach, where you're getting it working, and then we can go back and organize if we need to. But avoiding that premature organization and abstraction, and all of that, I think, is probably the most important part of the job. One of the most important parts of the job. + +**Kris Brandow:** Yes, it sounds like a key thing to building maintainable software as well. + +**Johnny Boursiquot:** Wouldn't you know... + +**Kris Brandow:** I mean, I don't think I declared this as one of our maintenance series episodes; our mini, but now max series... I wonder if you can just follow like Apple’s naming instances, where’s it’s like Pro, and then we can have Max, and then we can have Ultra when the series gets longer and longer. So I guess this is the Pro of maintenance series now, soon to become Max, when we level it up some more. + +But yeah, from what you're saying Johnny, and what you're saying, Ian, it sounds like these are also like good strategies to kind of build more maintainable software for us in the future as well. Because once you do export something, you can never really take that back, at the end of the day. It's kind of there forever... Which, kind of going back to what we said in the beginning of the show, sounds -- you know, copying something around, "Okay, there's two copies of this thing now." Now, you don't have this kind of piece of code there that two things are vying for control of at the end of the day, which I think is super-important. + +But let's say that we have figured out that like, okay, here is some code that really is generic, at the end of the day; it really is just like, okay, I’m just like copying it and pasting it all over the place, and I'm changing... Like, the changes I make, I can just run a Copy and Replace, or a Find and Replace in my code editor, just changing some names. But aside from that, the code is truly copy-pastable. + +\[20:21\] At what point, or I guess in what way do you then take that code and make it so that it's not so much copying and pasting? Or do you just say, "Well, this code is small enough, and I don't want to make it into a library I have to support" so I do just continue copying and pasting it forever. Or do you do something else to kind of make it less arduous over time? + +**Ian Lopshire:** I think a good example here is the aws.string, where we’re taking a string and returning a pointer to it. I have that sprinkled throughout the codebase. It might be a struct that I'm trying to do... And before generics, I think the place where those were just where you needed them. But I think that's like one of the prime candidates for like a generic package, where we can just take a type and return a pointer to it. + +**Kris Brandow:** I'm kind of surprised there isn't a package that just does that, that someone maintains out there. Or maybe there is, I just don't know. I mean, I think it'd be hard to come up with a name for it, but... + +**Ian Lopshire:** There's one just called Pointer, ptr. + +**Kris Brandow:** Oh. Well, that's a good name. So following your rule, Ian, that makes sense. + +**Ian Lopshire:** Exactly. + +**Johnny Boursiquot:** You see, for me, I consider things like that to be extraneous. \[laughs\] Again, I'm going to be that old man pointing at the cloud. I'm like, "Okay, if I need a pointer to something, I will use the language proper to make a pointer to something." Yeah, it might be a little more verbose, it might not be a nice one-liner from a package, whatever it is... But again -- I'm not saying it's a bad thing. It's just style, or desire for bringing in a third-party -- I mean, I'm very shy when it comes to bringing in third-party dependencies into my codebase. Very, very shy. Because again, that whole responsibility model right for that package - I'm more likely to look favorably upon a package that I'm evaluating to bring into my codebase if it is actively maintained, if there's activity on the codebase. And obviously, certainly for very specific kinds of functionality, where I'm not going to reinvent wheels. So again, basically, not doing more work than I need to, kind of thing. + +So if I'm really going to be aggressively using a package to do things, in a specialized kind of work - yes, I will find a suitable third-party library out there. And of those, I will evaluate them to see which one is documented well enough, which one is maintained actively, which one is referenced by other things, like the pkg.go.dev now shows you - or maybe has always shown you - how many packages import whatever package you're looking at. So you can see how used, how well used that particular package is. And that way, you know folks have worked out the bugs and issues and whatnot. If it's been imported, and it's heavily used, then you know a lot of people have run into the potential issues you might have run into. All these things factor into the selection process for bringing in a third-party dependency. + +To me, it's a matter of -- like, I've been around the block long enough to understand the cost of abstractions, to understand the cost of bringing in third parties to take away perhaps some labor you might have had to do on your own, to perhaps write things the long way, or the more verbose way. So everything is a trade-off. Everything. So it depends on what costs are you willing to bear today, and also in the future, as your software, especially if it goes into production and you end up having to maintain it for the long-term - are you willing to bear the costs of all these third-party dependencies, of all the abstractions that you're introducing, even the ones you introduce into your own codebase? + +I've seen very large projects that basically have some abstractions that no longer hold water today. Perhaps back then, when they were introduced, they made total sense. But today, you have little notes, little side notes on these things, "Well, I don't use that anymore," or "Deprecated. Use this thing instead." Every time I see this, I'm like, "Yep, I know exactly how this one came to be. "There was some sort of abstraction going on, we thought we were going to do this, but business changed and now we're doing this instead." So yeah, I think it's all the lessons learned really, at this point. + +**Kris Brandow:** \[24:23\] And part of me wonders if this is like a shift from the pendulum swinging... Like, I know when I first started my career, the kind of ethos for everything - I think partially because it was so difficult to get dependencies - it was like, "Build stuff yourself." And I think I joined the industry right as we were swinging away from what was deemed kind of negatively as a not-invented-here syndrome, toward the proudly found elsewhere, which I think is the positive spin... Obviously, trying to get toward one side of those things. + +I kind of wonder if what you're saying, Johnny, is the kind of evolution that's happened from us swinging a little too far to the other side of "probably found elsewhere", where we have these problems where doing this type of evaluation on external dependencies to do it properly takes a lot of time and effort and energy. And it seems like the rate at which we keep pulling in dependencies, we're not really doing that at the end of the day. It's like, there's a famous Node.js bugs where someone deletes a package, and then the entire ecosystem just breaks for a long period of time. And even more recently, the different types of people sneaking malware into their own packages to get people to stop using them, or because they're mad about something... Just all sorts of problems that exist because we depend on code that isn't our own for a little bit too much stuff. + +So I wonder if either of you -- do you kind of agree with that, that we have swung a little bit too far? Do you think it's something else that is kind of like, "Okay, well maybe we should be rethinking our dependencies"? Or is it just like my perspective, like... I've seen this change over my career, and it's just like, "Okay, well, if you'd started ten years earlier, you would have seen the same change over the course of those first ten years" or whatever? + +**Ian Lopshire:** I think we're kind of lucky with Go, because we have such a good standard library. A lot of what you would have to like kind of import from a third party before, you just don't need to. So I think in Go, we might have veered towards the kind of just rewrite it and don't import a dependency, but I think that's okay, because we have such a good standard library. Obviously, there are things not in the library, but... Yeah, I don't know where I was going with that one. + +**Johnny Boursiquot:** \[laughs\] I can add some flavor. I'm not sure if it's so much as sort of the pendulum swinging back and forth... And it very well could be, depending on where you're sitting. But I think generally speaking, whether it be with Go, Ruby, or Rust, or whatever it is, there's a reason why you have things like best practice. You need bad practice, or you need not-so-good practice, in order to have best practice. So that means it takes a whole lot of people doing a whole lot of things, sometimes smart, sometimes not so smart, in order to see enough patterns of usage and start to recognize over time - and that's the critical element here, time - to understand, "Well, if I engineer my piece of software this way, I am least likely to run into these kinds of problems that we've seen occur again, and again and again and again." Hence, best practice. + +So you need the time element. That's sort of a key ingredient in understanding sort of that pendulum, if you will, to understand where are we now, given this context, and how is it likely to change when things swing back around. And really, as you mature as an engineer, I think that the best thing you can do for yourself is expose yourself to enough patterns that you can recognize which one is applicable where, and when. There's never going to be the perfect solution; such a thing doesn't exist. It's only what are the trade-offs you're making today, given the information you know now, with the hope of having maintainable, well-engineered software into the future. You’re just hedging, it's all you're doing. Only time can tell you. And even if the pendulum seems to be swinging, "Oh, this is the new hotness, organize your code this way" or, "This is the new hotness..." + +\[28:22\] I mean, I remember when -- again, going back to that whole pkg thing - I remember when that was the new hotness. I mean, everybody was putting pkg into every darn Go package. Heck, I did it, because I was just going along with the flow. I'm like, "Oh, yes, some of the most well-respected people in the community are using pkg." I'm like, "Okay, I'm going to start doing that, too." And I’d literally cargo-culted the thing into my own Go code. I don't even know why I was doing it. I was like, "Oh, yeah, so and so does it. They must know something I don't." But only over time do you realize, "Okay, why am I doing this?" You start asking yourself, "Why am I doing this?" Because I read a blog post somewhere, where somebody had a specific set of concerns and constraints, and they said this was the right way, and now all of a sudden I'm using their solution to solve a problem I do not have? What kind of sense does that make? + +So when I see blog posts around package structure and things like that - and there are a few well-written ones - I always try to understand "What problem are you solving?" Some of the best blog posts on the matter - I think Jon has written one as well, that we'll put in the show notes. I've also seen some well-written ones from Ben Johnson of BoltDB fame. I've also seen some from Bill Kennedy... People who know their stuff. But not all of their approaches and solutions works in all of my use cases. + +So you could say, "Well, Ben said do this. So all my projects now are going to look like this." Well, no. No. What problem is Ben trying to solve in his blog post? What problem is John trying to solve in his blog? So you have to take context into account. And you have to use time as your friend, as an engineer, to learn to recognize patterns and say, "You know what? You know what would work best here? I'm not a big fan of MVC, but you know what? Organizing things as if it was sort of a Rails project, by having models, views, controllers, folders, instead of arranging things as a sort of domain-driven way - this model may be better in this particular case, because I'm building a CRUD app, than this model here, which is more of a business vertical, very specific service that I need to build, or a third-party package I need to build right for other teams to use." + +So again, the context is going to be what drives you towards one or the other, but you're not going to know that you have options, unless you educate yourself, unless you add time to mix to learn to recognize patterns. + +**Kris Brandow:** Yeah, no, I think that's spot on. I wonder too, I wonder if this is like part of our problem with best practices. The way that a lot of people conceptualize them right now is that it's kind of like -- sometimes it's just kind of like, "Oh, no, just do this thing." Because as you said, like "Oh, these important people are saying that we should do them" or "These important people are saying that this is the way that we should make these things happen." + +I feel like Go is kind of unique in this way too, where - when I look at things like the idioms we have, or even things like Go Proverbs, I've never seen them as things where it's just like, "No, shut up and just accept that's true." I feel like they kind of call out to you to and they’re like, "No, no, no... Sit down and actually think about this thing a little bit more thoroughly, and make sure you actually understand what it's saying." Because even at face value, it forces you into that kind of thought process. I wonder if that's a hallmark of what a good best practice is; this thing that's not simple on its face. I think "Don't Repeat Yourself" is one of those things that is like sort of a best practice, but it doesn't have that same type of impetus for you to kind of dig a little bit beneath that to make sure you understand it a bit. I feel like it's so catchy that we just kind of repeat it all of the time, without understanding -- maybe its genesis at the end of the day. Do you agree with that, Johnny? + +**Johnny Boursiquot:** \[32:16\] We have these sort of sayings, these best practices that we've accumulated over years of practice, the software engineering practice. "Don't Repeat Yourself" is a good one; like, DRY, DRY, DRY, DRY all the things, all the time. Man, I violate that rule every single day I write Go code. I'm like, "Yeah. I mean, yeah, don't repeat. Yeah, sure. Yeah. I mean, yeah." In most circumstances, generally, that's -- yeah, you don't want to repeat yourself. Well, you have to understand the intent behind that saying. The intent is, "Well don't repeat the same functionality, because now when you need to modify it here, you also need modify it there, there, there and there." But if I never have a there, there, there, there, and there, and only have here, why do I care about DRY? Or if I only be repeat it a couple of times, or three times. Again, that rule of three - once I hit it three times, I might think, "Okay, is it worth refactoring?" Okay, maybe the fourth time it comes up - yes, maybe. Yeah, maybe I'll create some abstractions and sort of reuse it within that codebase. + +But we can't go into projects with sort of this Bible of sayings, be it Go Proverbs, or the Gang of Four Book, or domain-driven design... We can't go into our projects with these seminal works, and expect "Well, I'm just going to throw these books at the problem." Honestly, especially at the early days of a project, I'm never concerned about design. Again, maybe dirty little secret of mine, but I'm never concerned about design. I'm never concerned about reusability. I'm never concerned about abstractions. It's too early. It's too early. + +The only thing I know is -- you know, a manager comes to me and says, "Well, we need a service to take this data from point A to point B." That's all the requirements I have. Now I'm digging through docs, a lot of projects, asking other team members, "Hey, have you seen this thing? Do you know what this thing is?" I'm literally researching what the heck it is I've just been asked to do. I have zero idea what it is I'm about to build, so what makes me think I know how to design this solution? + +I mean, don't get me wrong, I'm not going to approach it as a novice; I'm not going to be shipping something that is not of high quality. The code will be high quality, but the code will not be concerned about design and reusability and architecture and all that stuff. That's the stuff that I do when I know I'm going to put it out in the world. If it's a package that's going to be reused by teams, then I sit down -- after I've solved the problem, then I sit down, I worry about design and architecture and rearranging and putting things in the right places. If I'm worried about sort of performance, then I'm going to take my proof concept and then okay, I'm going to run that through pprof and saying, "Okay, well, where do I have unnecessary Go routines spinning up and doing things? Where do I have waste? Where do I have some allocations that I could maybe reduce?" Only then do I have these concerns, but we can go into projects with these - granted, great ideas that have come over time... It will be an analysis paralysis for days before you even try to solve a single problem. + +It's maddening to think that we can just throw all these best practices at every problem every time, and then we wonder why we can't move fast; because we just stuck designing. + +**Kris Brandow:** Or do you wind up with factory, factory, factory, factory. \[laughter\] "I just took the whole Gang of Four book and started applying it all over my codebase, and now my code is correct. These are all the best practices and the patterns we should be following." Ian, do you have anything you want to add to that? + +**Ian Lopshire:** I just want to kind of extract something from what Johnny was saying. I think what's underlying everything he said is thinking what problem am I solving here? You get your proof of concept, and what's my next step? Is it making this available to another team? I'm going to make my decisions based on that. If it's performance, I'll make my decisions based on that. So I think no matter what your best practice or proverb is, the underlying "why" is what you have to think about. + +**Break:** \[36:26\] + +**Kris Brandow:** Part of what you were saying, Johnny, reminds me of this thing I've been thinking about for the past week. I don't know why I started thinking about it, but it just kind of like popped into my head. Our industry is very good at taking those early prototypes and just kind of shoving them into production. Right and just doing "Okay, we're going to we're going to make an MVP, and then it's going to prod, and we're going to rewrite that code later." And then later never actually comes until it's like so terrible that we are like, "Okay, we've got to throw out this whole project and then do a greenfield," even though greenfield is really a brownfield, but "We've got to just rebuild the entire thing from scratch." + +And the thing I was kind of thinking about was like, if you're going to go build an actual product at the end of the day, whether it's a car, or a widget, or whatever, there's the prototype version of it you build, the proof of concept that's like, "Okay, we can do this." But then you have to go and kind of have like a production process, a manufacturing process, where a lot of those elements of that thing might change; the overall idea of it stays the same. You're still producing a car, or a widget, or something. But the way that that thing gets produced is significantly altered in order to make it so that you can mass-produce it, so you can put it on a production line, and turn those out so people can buy them. And it feels like we don't really have an equivalent of that in software engineering. It feels like we're missing that step of, "Okay, we wrote this thing, we built this thing as the prototype. Now we know our idea works. Okay, put that over there, we can reference that. Now we're going to go rebuild the new thing." + +I kind of wonder, do you guys feel that yeah, we are missing that kind of step, or do you think like, "Okay, we don't really need that sort of step", because we're not making a million cars at the end of the day, we're making like a set of features, so that's a completely different thing, and that doesn't really fit the same analogy... What do you guys think about that kind of analogy I just made there? + +**Johnny Boursiquot:** So the car industry - I'm not from a car industry, so I'm just looking at it as an outsider here; but from what I've noticed, the car industry - they'll have these shows, where they have those sort of beautiful, futuristic-looking designs of automobiles... And you've look at the thing, you’re like, "Man, this thing is like -- why don't they actually build these things and put them on the road?" But these concepts, these concept vehicles - their purpose is not necessarily to sort of provide a proof of concept, even though there's concept in the name, but it's not to provide a proof of concept for a practical version of the thing. The part of the reason you have these very futuristic-looking, not really going to go to production sort of looking vehicles, is that they're part of marketing. They're beautiful to look at, they're catchy, they demand attention... You look at it and it's like, "Wow, I would love to sit in this thing and ride around with this all-glass vehicle", or whatever the main appeal is. But they know these things aren't going to go to production, one, because there are lots of constraints, including government requirements that certain vehicles have a certain mileage or safety things associated with it... There's a bunch of constraints that are going to force this beautiful, futuristic-looking vehicle to being something lesser than the concept itself. But that doesn't mean there's no need for the concept. But its purpose is not to give you something that you can sort of take to production. Its purpose is to get people excited about something. + +So if we take that exact same analogy and bring it into our world, as software engineers, the purpose of a proof of concept is to have something that proves an idea is possible. It is feasible. The problem we have - and this is something I've noticed and been a part of, unfortunately, many, many times over my career, is that the proof of concept, somebody will see it, and be it service or frontend, or whatever it is, the business will see it or somebody will see it and be like, "Oh, you know what? That's good enough. Let's put that in production." + +And all of a sudden, you're an engineer, you put both hands over your head, you're like, "Nooo...! This is not what I wanted. This is not what this is meant to do." And you argue, and you fight, and you say, "Hey, really, we shouldn't do this. This is not ready for production." You make all the arguments and all the cases, and then at the end of the day, the thing that you didn't think was going to be production-worthy finds its way in some version of production. So now we get shy, we become shy about the possibility of that happening again. If it's happened to you more than a couple of times, your proof of concepts start looking less like proof of concepts, and they start looking more like, "Possibly, maybe, might go to production" versions of a proof of concept. You spend the time with on design, you spend the time on packaging, you spend the time on folderization, and making sure if a developer stumbles on your code, you're not going to be embarrassed. You spend your time on tweaking and massaging this thing, when you could spend the time actually proving out a concept, that an idea is possible. + +\[44:27\] So I think it's an unfortunate place that we're in with proof of concepts, but it's one that I think we've been driven to this state where we as developers are shy, we are afraid that our proof of concepts that are throwaway, what’s meant to be throwaway code will actually end up being thrown into production instead. + +**Kris Brandow:** Right. And we also have the engineers, they go through and they make that proof of concept and business asks, "How long till we can put it?" And they're like, "Oh, it's like 80% done." Okay, that might actually be technically true, but that last 20% is the most difficult part of it. So yes, it's not really that. + +I think there's a little bit of both playing in there as well. Some people are very like, "Okay, no, we're only going to do this once, and we've got to do it the right way." And you have a whole bunch of other people that are like, "I don't know, we just need to need to build it and go." Ian, I'd be curious to hear your thoughts on the analogy, on what Johnny said, and all of that as well. + +**Ian Lopshire:** Yeah. So only analogy side, I think the equivalent of those production cars, those sample cars you see at the big shows are like the dribble UI mockups. I think that's where those exist. But as far as proof of concepts - I'm at the point where I kind of struggle knowing when a proof of concept is even necessary. When you're just building a new feature or something, a lot of times you know it's possible. So what do you think of that? Do you think proof of concepts are always necessary, or do you think it's— + +**Johnny Boursiquot:** If you've never done a thing before... For example, a few years ago I was asked to look into this whole serverless thing. "Is this going to work for our workloads?" and things like that. And I was given a particular problem to solve, and then it says, "Well, can we do it without EC2 Instances?" I was like, "Well, I don't know, it depends. How long is the workload?" Because at the time, I think there was a five-minute time limit, two Lambdas or 15 minutes, whatever that was, whatever the initial limit was. And I was like, "Well, if your workload is going to take longer than that time, then no, you shouldn't use serverless; you shouldn't use Lambdas and things." And they’re like, "Well, what if we break it apart? What if we do smaller things, a bunch of small things", like the best practice at the time? What if we make a bunch of different Lambdas, or invoke the same thing, put it in the pipeline and invoke the same Lambda multiple times, and you can scale it infinitely, according to the AWS marketing? Just do that instead. + +I’m like, "Well, now we need to orchestrate", and at the time AWS had Step Functions, which I think pretty closely followed Lambdas. Or at least the use of Step Functions for Lambdas basically became a thing. And then I’m like, "Well, orchestration - yeah, rather than to have one Lambda call another, and then one function waiting on another and timing out the whole thing, and now you can use step functions, and you can have ASL, Amazon State Language, whatever it is, you can write some of that YAML or whatever, and basically orchestrate all your things" - these were things that we had never done before. So in that case, yeah, I'm definitely doing a proof of concept, because I don't know what I don't know; I need to understand the characteristics of the platform, what it offers. I need to understand what all the knobs are right for the service that I'm supposed to be using. So there's a lot there that I need to do. + +So this is kind of an example that goes that's sort of at perhaps the infrastructure level - to some code level, obviously, because the programming language that you use is going to have an impact on your serverless stuff, no doubt. So if I'm working with a Java project, because of the JVM, and I know that needs to be booted up, whatever it is, I'm going to take that into consideration. If you're telling me to build something with Java and Lambda, as opposed to if you tell me to build something with Go and Lambda, then I'm not going to suffer the cost of a JVM. So again, all these trade-offs that you're looking at. + +\[48:13\] So again, the more you don't know about the kind of problem you solving, the more I think you need a proof of concept. And that goes -- this was an example at that sort of infra level. But if somebody says, "Hey, decode this data stream." Like, I've never seen that before. It's not like anything I've ever seen, so I might need to do a proof of concept to sort of understand what I'm doing. So I think it's going to depend. To your point, Ian, I think it's just something that we don't naturally gravitate to at first, because - maybe it's hubris, I don't know; maybe we think we can just start working and we'll figure it out, and whatever we come up with, "Ah, ship it to production." + +**Kris Brandow:** I guess I would say that I think -- if I'm thinking about this from a writer's perspective, I would say you should always do prototyping and proof of concept. And I think that way because it's like - if I think about how I generally tend to write code versus how I would ideally write something... If I was going to write something, I’d sit down and I'd do some drafting, or some basic sketching of an idea... But I'm not going to try and edit as I write. I'm not going to try and copy it as I write, or even try and make the thing fit together properly, because that's going to slow down the whole process overall, and it might even prevent me from getting to the goal I want, because I've taken up so much time doing it. + +So I'm just going to rough out the idea, quick and dirty, get all of the hearts of it on page, and then start the editing process and refining it and trying to get it all shiny to send out to production or publication. And I feel like that is something we should do with code, and I feel like this is something that I would love to see more engineers doing, is like, "Okay, I'm just going to like rough out the idea, quick and dirty" like maybe not even have that many tests or anything... Just like, "Does this thing work for basically everything, even the smallest of features that we want to do?" And then have the solid idea of what we want to create, and then go back, and then rewrite it and redraft it, and put it into the structure and format that we want. + +Because I think not doing those types of prototypes means that we're kind of cutting down on the number of iterations or the space that we can think of an idea. For me, if I go to write something, and it's like, "Okay, well, I know this will have to go to production", and I spend a bunch of time having to write tests to make sure this small part of it works, it kind of kills my creativity at the end of the day. I'm kind of like, "Oh, I've gotta think about all of this stuff", and it kind of knocks me into a different stream of thought than the stream of thought I was using when I was prototyping this feature, and I was kind of figuring out how to put this thing together. + +So that to me seems like one of the failures we've had as an industry, is that we don't encourage people to prototype always. And I think the reason we don't do it is because -- I think at one level, we kind of hold prototyping up on this pedestal, and it's like this fun thing that you get to do. It's what you do when you go play with new technologies. It's not something you do when you're just going to implement some boring business feature. I think it's that mixed with a little bit of hubris, too. It's like, "Well, this thing is so simple." Of course, we can just write it correct the first time. + +There's so few lines of code that I've written right the first time that I actually came back a week later and been like, "No, that was the right thing." Actually, I don't think I've ever written code where it's like "I wrote it once and that was it", and then I came back and looked at it and didn't think it was like... Every time it's like, "No, no, this could be so much better." + +I think that's like a practice that maybe we should start introducing into the way we do things, of just, you know, ease the business into it... And be like, "No, no, we have to actually write the code rough and dirty first, get all the small, high-level parts correct, and then we'll go in productionize the code. And then we'll go and add all the testing and then monitoring and all of the logging and all of this other stuff that has to go around it to make it spick and span." + +And obviously documentation, too. The amount of code I see in prod that doesn't have a line of documentation, or they add the function name or the type name and then, dot-dot-dot to get around the winter, or whatever thing they could do. I think that part of that is because we don't sit down and we're not like, "Oh, no, we have to actually prototype this first, and then there's this other series of steps we take to make it productionized. + +**Ian Lopshire:** \[52:18\] Yeah, I almost don't think of it as like a proof of concept. I think of it kind of as levels of drafting. Like you said, I guess the first draft is like that proof of concept. I'm going to do it quick and dirty, and make sure it works. And then we go back and make sure we have observability, and all of this. Yeah, I almost think about it more as drafts. + +**Kris Brandow:** Yeah. So I get to move on to the next segment, but Thomas, if you're listening, I hope that we answered your question. And if we got on a meta rant and didn't, I guess we'll just have to do another episode and try to answer your question. But now it is time for unpopular opinions. + +**Jingle:** \[52:54\] to \[53:10\] + +**Kris Brandow:** Alright. Ian, since you are a guest, do you have an unpopular opinion? + +**Ian Lopshire:** I feel like I never do, but... I don't know. I cannot think of a good one. + +**Kris Brandow:** It doesn't have to be good. It just has to be unpopular. + +**Ian Lopshire:** Give me a minute, maybe come back. + +**Kris Brandow:** Okay. I'm sure Johnny has a few. So do you want to share one with us, Johnny? + +**Johnny Boursiquot:** Sure. I mean, I think I've shared quite a few during this episode... I mean, yeah, if I could maybe summarize some of what I guess I hold most dear these days, especially around sort of code organization, is don't organize your code in the beginning. You don't know what you're organizing yet. Just get the thing done, solve the problem, and then say, "Okay, well, what am I trying to achieve in now organizing this code? Am I organizing for readability and maintainability? Am I organizing because it's going to be used by another team? Or by maybe the open source community? What am I organizing for?" + +Obviously, I'm a big fan of domain-driven design. I always keep that in the back of my mind. I don't tend to use sort of frameworky things, generating folders for me, and scaffolding. Again, the type of work I do - I have very little used scaffolding and things, and frameworks, and whatnot. + +But if these are the kinds of problems you're solving, if you're building CRUD apps for a living and you want to use frameworks to generate things, and put files and generate things for you, put them in a folder, and for things to wire up properly you need to put them in a Models folder, and or whatever - I mean, if you if that's your jam, that's your jam. I'm not kicking it, by any means. + +But yeah, beyond those use cases - man, solve the problem first, and then worry about organizing it, then worry about abstractions, then worry about what's being used and reused and things, and don't be shy to copy from across projects even. Just don't be afraid; you can just copy a file here and drop it in your project. Just get things done. Just get it done, and then worry about organizing it. + +**Kris Brandow:** So if you were to write a blog post on how you should organize your Go code, it would just say, "Don't." + +**Johnny Boursiquot:** Don't. \[laughs\] + +**Kris Brandow:** Like, by the time you need to organize your codebase, you won't need a blog post to tell you how to do it. + +**Johnny Boursiquot:** Exactly. + +**Kris Brandow:** We just publish that on the Changelog blog right now... \[laughter\] No, I mean, I definitely agree with that. I think people do rush way too fast to try and sit down and organize their codebases. I personally always start with one file, one package, just put everything in it, and then once it starts making sense to split things up, I start splitting them up. I think that's the nice thing about Go, is that you can have a file that is maybe even 1,000 lines long, and it's still very readable. You can still find things in it. I think the language is— + +**Johnny Boursiquot:** Whoa, whoa, whoa! This is where you and I fork. \[laughs\] A thousand lines long file? No, you and I are going to have words... \[laughter\] + +**Kris Brandow:** \[56:08\] Disagreements! We'll have a whole episode about this. Whether it stays at 1,000 lines is debatable, but I think files are -- I'm more okay with people splitting things up into files. I think the thing I like to avoid is people that make 40 files with 20 lines of code in them. There's a middle ground here. If I do write a file that gets to 1,000 lines, then I'm kind of like, "Okay, this is a little long in the tooth. Let's, let's break this up into pieces. Let's see how we can format this better." I'd definitely do that. But yes, I definitely think one package to start with, and then go from there. So, Ian, did you come up with an unpopular opinion? + +**Ian Lopshire:** I think I did. This might be a repeat, though. I don't know. + +**Kris Brandow:** Okay. + +**Ian Lopshire:** My unpopular opinion right now is that monoliths are probably the way to go for most companies right now. I think all these companies going microservices are doing one of those things where "Microservices are best practice, so we're just going to mindlessly do it", and not thinking "Why am I doing this?" + +**Johnny Boursiquot:** Wait, we haven't learned from all the case studies of people who went micro and went back to, I guess, macro, or monolith, or something? A lot of times people think that, "Yeah, I'm going to adopt microservices, because it's cool and trendy and whatnot." And they're realizing that when you go microservices, you're not making just a coding decision, you’re also make it an operational decision. + +So your developers, while they may have input into the decision to microservice or not to microservice, I think the operational folks, the SREs, the people who need to keep this thing running and orchestrate all the things and deploy all these things across infrastructure, make your thing into a distributed system that actually cohesively works. They need a say into the "to microservice or not to microservice" decision, because everybody is affected by that. So it's not just the developer writing code and choosing to put a network boundary in between their projects. Everybody needs to go in with both eyes open. + +**Kris Brandow:** See, we've just got to rebrand monoliths into macroservices, and that will be the new hotness, and that's how we can get people over macro. It's like, "Obviously, macroservices are better, because they're bigger, right? Macro. Why would you do something micro when you can do a macro? It's innovation. It's bigger. We want to do bigger." \[laughter\] + +Alright... So yeah, we'll put the polls up on Twitter and see if your unpopular opinions are actually unpopular. I am curious to see if they are. And as the rule goes, if they're not, you have to come back and try again, to be unpopular. + +**Johnny Boursiquot:** Justify. + +**Kris Brandow:** And perhaps one day you will be as unpopular as some of my opinions have been, to once again, warrant us some negative reviews on Apple Podcasts. If you liked the podcast, go and give us a nice review on Apple Podcasts, please. But yeah, with that, I want to thank you, Ian, for joining us for this lovely panel discussion. + +**Ian Lopshire:** Yeah. Thanks for having me. + +**Kris Brandow:** Yeah, of course. And thank you, Johnny, for co-hosting, as always. Hopefully, we didn't get too meta for our listeners out there. + +**Johnny Boursiquot:** Just the right amount. + +**Kris Brandow:** Just the right amount. diff --git "a/Go for beginners \342\200\232\303\264\302\252\303\224\342\210\217\303\250_transcript.txt" "b/Go for beginners \342\200\232\303\264\302\252\303\224\342\210\217\303\250_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..04f7dad43defc3a628caf907c9e80b5ef5a1c2f5 --- /dev/null +++ "b/Go for beginners \342\200\232\303\264\302\252\303\224\342\210\217\303\250_transcript.txt" @@ -0,0 +1,365 @@ +**Carmen Andoh:** Welcome back, everyone, to Go Time. I am Carmen Andoh and Mat is out this week, so I will be your host for the journey. Joining me today is our recurring panelist, Ashley McNamara. Say hello, Ashley. + +**Ashley Willis:** Hello, everyone. + +**Carmen Andoh:** And we have a returning guest, Jon Calhoun, who is the content founder for Gophercises.com, but also UseGolang.com, TestWithGo.com, ErrorsInGo.com... So I'm excited to have you on, Jon. Say hi. + +**Jon Calhoun:** Hey, everybody. + +**Carmen Andoh:** And now we have a new guest, Dave Valentine. He's the instructor of Udemy's A Gentle Introduction to Golang for Beginners. Welcome to our show! + +**David Valentine:** Thank you so much! It's a pleasure to be here for the first time. + +**Carmen Andoh:** Speaking of beginnings, when we asked Dave to show up on the show and give a profile pic, he submitted one of him as like a toddler, with TRS-80 Model 1, circa 1980... I think that's probably when you began? + +**David Valentine:** Absolutely! And Carmen, I have to say, I was chuckling when you were like "What computer is that?!" \[laughter\] But yes, yes, I got a start in the microcomputer generation world, so I started with TRS-80, and I moved over to a PC Junior, and then continued on through microcomputers and programming way back at that time... So I've had an early start on things. + +**Carmen Andoh:** Wow...! + +**Ashley Willis:** That is hard! That is a hardcore nerd. + +**Carmen Andoh:** \[06:02\] Yeah... Total nostalgia ahoy, I'm sure, for many people. Certainly you can share what you started with when you were beginning, your first computer, in the \#GoTimeFM channel on GopherSlack. + +Okay, so let's talk a little bit about today's theme... This is Golang For Beginners. I had originally envisioned this episode to be meant to engage both non-Go-users, that listen to maybe sister podcasts on Changelog, or any Go-curious programmers out there, as well as encourage those have just started with Go and want to level up beyond the basics. I'm hoping that maybe we can start and just see where the conversation takes us. We have three wonderful experts online today, and see what their perspectives and their experiences and how they've shared content to help gophers worldwide. + +I guess the first thing I wanted to start with is to ask - what is a beginner? What types of beginners are there? Anyone? + +**Ashley Willis:** Good question. There are beginners that know one programming language and they want to learn another; there are people that are career-transitioning, that don't know any programming... There's all kinds of beginners. + +**Jon Calhoun:** I was gonna say, there's even the curious kids who don't know what they're doing and like to break their parents' technology, which is probably some of us... \[laughter\] + +**Carmen Andoh:** Yeah. + +**Ashley Willis:** Those are my favorites. + +**David Valentine:** Absolutely. I've experienced that with one of the courses that I've published... One of the kids was literally saying "I don't have system privilege to install Go on my computer. What do I do? How do I get around that kind of a thing?" + +**Carmen Andoh:** Whoa...! \[laughs\] + +**David Valentine:** ...because his dad had bought him the course. I've actually in my program originally created a course that was in advance of the one I currently have in the marketplace, which I've since retired, that was kind of trying to get the entire beginners in Go space, where it's for people that are experienced with programming, and it's for people that are new to programming for the first time... But what I found was I was trying to appeal to too many different people, so the course content wasn't that effective. + +Since then, I've actually specialized in making something very clear it is for absolute beginners, because I actually walk them through some of the fundamentals of computer science in learning Go, and giving them a foundation if they've never seen anything before. So I've actually learned to specialize. + +**Carmen Andoh:** And is this the course that I mentioned before, or is this a different course? + +**David Valentine:** It is exactly the course you've just mentioned, yes. That one's been in the marketplace since January of this year. It's a reboot specifically - and I've taken the other one out of the marketplace - intended for people that have never seen programming languages before, and wanna get started with the exciting computer language Go. + +**Carmen Andoh:** That's great. And just a reminder to everyone - what Dave is referring to is Udemy (Udemy.com) and it's called A Gentle Introduction to Golang for Beginners. Is that what you mean by "Gentle", the idea that you're a complete beginner, not just a beginner to Go? + +**David Valentine:** Exactly right, that is just it. What tends to happen with beginners is that they end up making assumptions about what the computer is doing underneath, and they end up almost with a heuristic knowledge, or almost some voodoo... "Why is it like that?!" Well, I prefer to try to explain tip to tail everything that's happening; maybe it's that deep computer background from that TRS-80 timeframe... But I get into memory, and what computer code actually looks like, and the role of a compiler, and I try to expose all of the pieces, so at the end the light bulb goes on, and somebody understands -- especially when it comes to understanding the concept of what's at a memory location, versus what a memory location is itself. That is the type of journey that I bring people through, specifically though in relation to the computer language Go. + +**Ashley Willis:** \[10:09\] That is the kind of course that I value specifically as somebody who does not have a CS degree. I think I get into the weeds often, because what happens - I am my own worst enemy - is I try to learn something and I'm like "No, no, I need to know the thing before that, and then the thing before that, and then the thing before that", until I'm like "Have I learned anything at all?" \[laughter\] So I like to get into the weeds a little bit, but if it's self-guided, sometimes I'm too in the weeds. + +**Jon Calhoun:** I think this is one of the reasons why at times I feel like the younger you are, the easier it is to learn this stuff... Because you're more willing to just take things for granted, I think. Whereas the older you get, you're like -- I just think that it's easy, as an adult, to think "I need to figure out what all these things are doing and understand and at all." As a kid, you're just like "I'm willing to just trust you and just go with it, and let it go." + +**Carmen Andoh:** Interesting. So you think that's maybe the key, just sort of trusting whoever has shaped the content is gonna structure the content, that's how you're gonna learn, or...? + +**Jon Calhoun:** I think sometimes. One example I can give is I talked to somebody who had started a company, and when they started the company, they couldn't find a technical founder, so they basically just went and learned programming on their own... And I was talking to them about how they learned to program, and the one thing I found interesting was that they basically forced themselves to go through 3 or 4 complete web development courses... And they said the first time there was a lot of things that were mentioned that they just didn't quite get, but by the time they went through their third course, some of those things -- they had enough understanding and foundation that those more intricate details started to make sense... And I think if you get too focused on them early, you just go down too many rabbit holes and you don't really get anywhere. But if you're willing to just push yourself through it and realize that "I'll come back to this later" or "We'll go through the material again later", then that can really help. + +**Carmen Andoh:** That's a good point. Now, you've mentioned a little bit about kids and whatnot - I also wanted to ask a little bit what everyone's thoughts were about industry trends and educational backgrounds... Because I'm seeing that we are increasingly having people within our industry that are not formally trained with the university CS degree, but maybe are self-taught, or they went through a bootcamp. I know Ashley that was your path, right? + +**Ashley Willis:** Yes. + +**Carmen Andoh:** So is there anything in terms of difference between how we teach people that are not CS degree learners? + +**Ashley Willis:** Bootcamps are predatory. Sorry, all bootcamp grads... I value you; you don't know that you're predatory... I did all of the OpenCourseWare for all of the major universities. That material was way more valuable. But it's hard when you don't know where to start; you believe that you need a classroom setting. + +Some of us are not in a place where we can go get our CS degrees. That is a place of privilege... So some of us have to learn. With all of these online courses out there, there are many more opportunities to self-pace and self-teach... And not all of them are great, so it takes some time to get through them. But for me, I wanna say that the most valuable learning that I did when I was learning to program was the OpenCourseWare. + +**Carmen Andoh:** Jon and Dave, can you tell us a little bit about your personal journeys, both as beginners, but just beginners to Go, how you leveled up there? + +**David Valentine:** Absolutely. I'll jump in there first. Go is -- gosh, my eighth, ninth, tenth (I'm not even sure anymore) computer language, and obviously having a background in computer science and having done a lot of things with it... Python is another one of the language that I do a lot with, because I also teach and introduce people to machine learning and artificial intelligence concepts and courses. + +\[14:02\] With Go, to be honest, it sort of came about for me from market research, because I became curious about what I should make my next great course on. And then I found this amazing computer language that really is a next-generation computer language... Because almost any other language that we may even consider new was really developed in the last century, right? And I love the Go story, the legend being that people were waiting for a C++ program to finish compiling and said "Hey, if we were to develop a language now, what could/should/would that look like?" And you've got some brilliant engineers that ended up putting it together... So I became absolutely fascinated by it. + +I realized that - here was a language that back in the day when I learned C for the first time, very much smelled like, tasted like, acted like C, exposed some of those fundamental computer bits, but had grown into being so much more than a system-based language. So I literally fell in love with it. + +Then because I had intended to develop a course on it -- what I really try to do in all the courses that I teach is I try to develop a roadmap so that my students aren't depending on me. With the internet there is this amazing amount of material out there, but it's not curated in a meaningful way... And that is maybe the secret sauce that I try to bring to the table. Because people have the itch, they have that "How do I get up to speed and make meaningful use of my time in order to achieve a learning understanding with Go?", or whatever their thing is. And so in the course I have what I call an emergency Golang parachute, which is learning resources, right out of the gate, saying "If you get stuck, here's all the really other cool stuff that's out there that you should know about, so that you can take advantage of that in order to supplement your learning." And then I finish off the course with "Here's where to go from here to continue your learning journey", and then take them through that piece where they can effectively then use those examples and resources. + +It's that very beginning. That beginning is so hard for people that don't have any experience, so giving them that ignition, even that permission to break things and experiment with things, and to think about things as they're coding and developing their exercises and so on - it gets them that little bit of traction with their wheels and gets them started. And the more students I can get started, the happier I am. + +**Carmen Andoh:** That's great! I think that's also a really good companion or complementary to what Jon does... Which is - okay, you have an ignition, you know the basics and the foundations; the next thing you've gotta do is... \[laughs\] Gophercises! Right, Jon? Tell us a little bit about that. + +**Jon Calhoun:** Yeah, so... Basically, whenever I learned, what I generally found was that it didn't matter if I was coding the prettiest code, or anything; it generally came down to if I coded a lot, I learned a lot. And if I didn't build things, I sort of just stopped learning. And I think a lot of times I've talked to tons and tons of people who are like "I'm trying to build things, but I can't come up with a good project." And what's even worse is if they come up with a project they like -- I had a friend who did this all the time... He'd say "I have this cool project" and he'd present it to me, and I'd be like "Well, it's gonna be really hard for you to do that project at your current skill level because there's like six other factors that you don't know much about, and I know from experience that they're going to be very hard." They'll want to get data from it and they'll assume there's an API to get some data, and it's like "Nah, that doesn't really exist. You're probably gonna have to scrape web pages, and that's gonna be terribly hard." + +\[17:53\] So Gophercises was kind of me saying -- you know, if I was starting over and I just wanted some random exercises to build, that weren't completely boring, but would challenge me, would teach me to read the docs, to look at different APIs in the standard library, to do stuff like that, what would they look like? And I basically just sat down and picked out 20 projects that -- I sort of picked them intentionally to try different stuff. I didn't wanna keep using the same thing; I didn't wanna build like 20 web applications... But the idea was, you know, if you go through all of these and you code them all and you actually give it an honest shot of trying to solve each problem on your own, you will learn a ton in the process. Even if you don't complete them fully, or you don't understand everything, or your code is pretty ugly - it doesn't matter; you'll still learn a lot doing it. + +**Ashley Willis:** I think that that's so valuable as well, because - I tweeted about this recently... When I was learning to code, I was already pretty well established in tech on Twitter; so I'm like "How should I start?" "Just build something!" "Build what?" "Something." "What? Where do I start? How do I start? What do I build? I don't know." "Well, you know what - figure it out; that's how you learn, just figure it out." + +**Jon Calhoun:** And that's really frustrating. + +**Ashley Willis:** So frustrating... Like, build a to-do app. Tell me something, tell me where to start. I feel like really experienced programmers - that's their go-to advice. "Well, just go build something." + +**David Valentine:** Actually, the other advice that I'd add to that though... I mean, if you've got a project, build it. But don't just build it just once, build it three times. \[laughs\] + +**Carmen Andoh:** Huh... + +**David Valentine:** Because the first time you build it, you're gonna commit all the sins and you're gonna build the wrong thing. It's not gonna meet the needs that you're looking to address with it, it's not gonna work well. You're gonna think "Ugh, that was horrific..." + +The second time it will probably function and achieve what you want it to achieve in terms of the end results of the things that you're building, but you're gonna think "Ugh, that is some ugly, awful, evil code." There are monsters working inside that that you hate. And the third time, you're in a position where you can actually put together some really elegant approaches to develop some beautiful code. + +So if I were to add any advice with that, once you do find that something, whatever that is, if that's reproducing someone's example or finding anything that inspires you to build something, build it three times. + +**Ashley Willis:** Wow... I love that advice. + +**Carmen Andoh:** Me too. I heard once that if you want to become a great writer, don't read 100 books; just read one book 100 times. I think this is kind of in the same vein, Dave, in that this gives you the chance to revisit a thing at different stages - which is reality - in terms of maintaining a piece of software as it ages, as you age and increase in your skillset etc. That's really great. I love that and I'm gonna try that. + +**Jon Calhoun:** Just to add something to that - I know some people get bored doing that, or I should say some people seem like they do; even if you don't build the exact same thing, I think building similar things probably would go in line with what Dave is saying... Because I'll see people take a course, and they'll just build whatever is in the course, and they'll be done. And what I really like to encourage is to go back through the course and build something similar, but not quite the same, so it forces you to go out on your own and sort of do what he's saying, to try different stuff, to do a little bit different... But you're still building the same basic building blocks; you're building a web application, or you're building a CLI or whatever, but it might do something slightly different, that forces you to really think about what you're doing and consider stuff. + +**Carmen Andoh:** Yes, John. And I think that learning is repetition, right? If you learn anything, you're gonna need to repeat it in order to make it go to long-term memory. But I think that that's also a great approach. + +**Break:** \[21:36\] + +**Carmen Andoh:** Well, I wanna segue into instead of maybe talking more about how we learn and approaches to learning, let's talk about learning in Go. We kind of touched on that, about the starting point and foundations and computer science fundamentals in your Udemy course, Dave, but... How do you teach beginners Go? And we can talk about this either from other languages, in terms of their experience, or from scratch. What do you think is unique to learning in Go, versus just programming in general? + +**David Valentine:** Well, I specifically at this point specialize in starting from scratch, so I have my course that is designed really to bring someone up from scratch. One thing I did find is that originally some people were finding some trouble with the technical aspects of it; they were having some trouble with GOPATH and GOROOT, and some of those things... And originally, I had created Windows installation videos and saying "Hey, if you want, you can go ahead and do the same thing on Mac, you can do the same thing on Linux...", and when I rebooted the course and specialized more in helping beginners, I said "To heck with that noise", and I literally have approached it to give them videos, and follow-through instructions on Windows 10, on macOS, as well as Ubuntu Linux, so that they at least have something that looks, tastes and smells similar to whatever platform they have, because there were some people that were experiencing problems. And my intent with it is to reduce the amount of friction at all possible. + +Once they achieve that though, what I find is that people are really functional within it, they get an understanding... One of the beautiful things that is unique about Go as well is that if people want, they can get into the Go code itself; it's there, it's in your work station. If you're interested in how println() works, and things of that nature, you can sort of dive into it, and start to see these other elements and how they come together in the compiler, and provide the end user experience of the compilation and the code that gets executed. + +\[25:48\] So I think it's really similar to other programming languages in the sense that the fundamentals are the same, but it has great syntax, it's really thoughtful, it's really built for concurrency, and I think it's a really effective language for beginners, because some of those obscure elements in earlier languages aren't there, it's very quick to compile... So I'm really an evangelist when it comes to it. But I did find that some of the students that always experience that have that little bit of friction at first, and now if I can get them over that, then they're off to the races. It's usually a very early problem early on, or they're off to the races and then asking more advanced items out of the gate. So there's this pendulum that swings in their experience, I find, where they're like "Oh, that was awesome! What's next?" + +**Ashley Willis:** In the beginning -- well, you know, until recently (Thanks, Go team!), GOPATH was a nightmare. Once you got through that hurdle, after you're ready to throw your computer, and then you build something, you're feeling better... But I think that the hurdle of setting up your GOPATH was such a nightmare that people were just like "No... No. If this is just setup, how's it gonna be?" + +**David Valentine:** Oh, absolutely. And getting to be set out of the default was the most brilliant thing that could happen. + +**Jon Calhoun:** I think that's one of the reasons why you see PHP being so popular still to this day. And I still think one of the big reasons it was was because if you knew very little, you could find a free PHP server and just upload a file, or you FTP in, or something, and connect to it and change a file, and you would see results. You didn't have to understand or install anything, and you could get stuff working. And I think every new language should strive for as much simplicity as possible. + +That's one of the reasons why I love the Go Playground, because I'm glad they thought about -- like, we need some way for somebody to quickly and easily just write some code; yeah, it's limited to the standard library, and some stuff like that, but it's still a great learning tool for somebody who, like you said, couldn't install it for some reason, or they just wanna see it before they actually spend that time... Because I think if we don't focus on that sum, there's languages like JavaScript where you can have interactive tutorials really easily in JavaScript, and somebody can just bring up Chrome and open up Chrome DevTools and write some JavaScript... So the barrier to entry there is so small that I think it's important for other languages to keep that in mind. + +**Carmen Andoh:** Yeah, this is a good thing that I kind of was wondering... Like, how much is setup a part of learning Go versus other languages? And you mentioned PHP and Java, and if it's different. What are some of the setup gotchas in addition, or you think are gonna be the gotchas, or the learning gotchas? And maybe we can ask for audience participation on the \#GoTimeFm Slack channel. + +**Ashley Willis:** Well, some of these showoffs are saying that GOPATH made perfect sense to them, so... Good job, guys! Good job. + +**Jon Calhoun:** I think the problem with it was it either made sense to you, or it made no sense to you. There was no middle ground. + +**Ashley Willis:** Yes. + +**David Valentine:** And it was also very platform-specific as well, right? Because at that point when you're hooking into environment variables, or whatever your platform is, you can get stuck in the details of your particular platform. + +**Jon Calhoun:** Yeah. Especially because a lot of beginners are coming from Windows, and setting environment variables has always seemed easier to me in Mac or Linux, whereas in Windows, trying to get somebody to go set that stuff up correctly was sometimes a pain. + +**Ashley Willis:** Super-pain. Me and Steve Francia (Hi, Steve!) used to teach workshops, and the first thing that we would do is "Raise your hand if you use a Windows machine. Great. You guys are now friends, because you're going to need to help each other during setup, because we cannot help you." + +**Carmen Andoh:** Yeah, taught them, and I have that material and I thought it was amazing... A CLI workshop, and I think it was for OSCON last year, or the year before. The slide deck is something like 300-some slides, and setup, Ashley, was like the first third of that. So I just wonder - is that the hump that we just want to help beginners get over, and then they'll be able to get really productive soon after that? + +**Ashley Willis:** It really was the first half of the workshop. Setup was difficult for a lot of people. There was a lot of going around and helping people get their machines set up. Once we got into building the app, things seemed to flow much easier... So yeah, setup was definitely a big issue. + +**Jon Calhoun:** \[30:11\] I think setup is also the biggest quitting point, too. + +**Carmen Andoh:** Yeah. + +**Jon Calhoun:** At some point during setup, if it stops working, you quit. But if you're writing code and you've got most things running, I think you're less likely to quit at that point. + +**Carmen Andoh:** And is this unique to Go? I wanna give a nod to any beginner out there, whether it's somebody who's already learned something else in terms of the programming language, or someone who's a complete beginner - is setup a quitting point for all languages, or is it a little bit more painful in Go? + +**Jon Calhoun:** I think it's all. I think every language has some barrier to entry, and once you get it, you're good, but up until that point it can be frustrating. + +**David Valentine:** I do think it also depends whether you're talking about an interpreted language or a compiled language as well, right? Because it is a little hard to compare Go to an interpreted language, the like of PHP perhaps, in the way that it's structured and some of those elements that it actually achieves underneath. + +One other element I think comes into play is once you have Go set up, what do you hook in after that? Is it an IDE? Do you have plans for an IDE? Because there's some really outstanding things you can do afterwards, and yeah, I show - just like everyone else does - how do you make a Hello, world program on the command line. But then if you can actually trace, and set up breakpoints, and have an IDE experience after that that will help guide the student, then they can trace their way through the code, which again, will help them understand what's actually happening underneath the covers. + +So there's that initial setup, but then also -- ask anyone, they have their favorite, whether it's Atom, Sublime, Visual Studio Code... Everyone has their favorite tools to then add to whatever language they're programming in at that point. + +**Carmen Andoh:** And then what do you think about in terms of content or setup, in terms of audience competence? We have a comment here from Cory LaNou in the \#GoTimeFM Slack, who said that go install was the easiest that he's ever used... But again, it was geared towards me as an audience, so we're saying somebody who's an inexperienced beginner. So for the people who have sort of looked through content, created content, in this panel, do you feel that there are audience for whom there are content gaps for beginners? + +**Ashley Willis:** Yeah, I do think... When people ask me what language they should start with, I usually say Python, because there's so much information out there, and I felt like setup was easier, for some reason. For Go, I feel like there's a lot of highly technical things out there, which is great, and then there are some beginner courses that may or may not work; I'm excited to try your guys' (courses). So I just felt like beginner materials were super-lacking. + +For me, when I learn a new programming language - I am not ashamed by this, I don't care if you laugh - I buy children's books. I love kids' programming books. They are the greatest. There needs to be one for Go, yesterday. + +**Carmen Andoh:** Oh my god, yes! \[laughs\] Well, you know, there was this trend - it's probably about eight years old now; do you remember ELI5, Explain It Like I'm 5? + +**Ashley Willis:** Yes. + +**Carmen Andoh:** Those were great. We had someone in the chat earlier saying "Well, I need to learn the thing to get to the thing, and learn the thing before that thing", and it's just a recursive -- but I don't know this all the way down, so I'm hoping your Gentle Beginners' Course tackles exactly that. + +**David Valentine:** Well, I appreciate that, and I'll tell you, you're almost throwing the gauntlet down, because I have twin eleven-year-olds, and while they have tasted a little bit of Python and a lot of Scratch, I'm thinking I should be maybe the first to write a Golang children's book at this rate, so... We'll see. I'll give it some thought. + +**Ashley Willis:** Please do. + +**Carmen Andoh:** Yes. + +**Jon Calhoun:** \[33:56\] I hope you have a good artist, because I feel like that'd be my biggest limitation; if I was drawing it, it would like a five-year-old drew the whole thing. + +**David Valentine:** Oh, I am a coder art scheme. I'd be full up on the red, green and blue colors. It would look horrific, so... Yeah, I need -- any artists, shout it out and I could probably use some folks to collaborate with on this, no doubt. + +**Ashley Willis:** I'm here, I'm taking note... \[laughs\] + +**Carmen Andoh:** Well, there may or may not be a pre-eminent artist for the Golang community speaking to us at this moment, right now, maybe... I don't know, what you think...? \[laughs\] Yeah, Ashley's gophers are -- how many are there in your automated-- + +**Ashley Willis:** Hundreds. + +**Carmen Andoh:** Billions, in gopherize.me. + +**Ashley Willis:** Oh, in gopherize.me there are billions. I don't know how many combinations there are right now. I think I might have added some things... But billions. + +**Carmen Andoh:** And I really like whoever did your artwork on your Gopercises, Jon. I think it looks like Ashley-esque, but it could be someone else. + +**Ashley Willis:** Yeah, it's not me. + +**David Valentine:** I think that was Marcus Olsen, I wanna say... + +**Ashley Willis:** Yes, it was. Yes! + +**Jon Calhoun:** Because Ashley is who introduced me to him... But I've actually gotten into the habit of -- so Ashley is usually really swamped, so I've gotten into the habit of trying to find different artists for all of the courses, because I like them to have a slightly unique feel. The testing course was Egon Elbre, and I have an algorithms course that I'm working on that is another Gopher in the community... I think it's Adrian -- I'm drawing a blank on his last name though. And then I have another artist who's working on another course that's way down the pipeline. + +So I'm trying to use different artists for all of them, because I think it's cool to have different people drawing gophers, and I know that Ashley is just overwhelmed with stuff, so I don't wanna keep bugging her... + +**Carmen Andoh:** But it just goes to show that I think when you are making content, whether that is for beginners, or for anyone - we are not a black and white or grey world, right? We are a world of color, and we like visuals, and we like fun, and we learn better when we feel we're at play. + +I remember Richard Feynman talking about how he battled burnout and won a Nobel prize for his Feynam techniques; we're kind of veering into physics, but still learning for beginners... And he always said "Just keep a sense of play about you", and Ashley, that's what I love about the gophers that you bring to the table and to the community. + +Anyway, Jerod Santo, who's also part of the Changelog crew, said that they'd do an ELI5 (explain it like I'm 5) on the sister podcast, JS Party. That would be a great segment for Go Time. So if any of you are down to come back and do a repeat, but for 4-5 year olds, by our inner 5-year-olds, I'd love to do that episode. + +**David Valentine:** That'd be awesome. + +**Ashley Willis:** Yeah, same. I'm here for it. + +**Break:** \[36:32\] + +**Carmen Andoh:** I want to talk a little bit about learning mediums. Each of you has created or taught content in this new world of online, so I would love to talk about your perspectives about pros and cons about each of these mediums. Jon, we're talking about your Gophercises with code, accompanied by videos; Dave, this would be your online course, workshops that you've given at events and conferences, Ashley... So what are some of the pros and cons about each of these mediums for beginners? + +**Jon Calhoun:** I guess I can start with some of them... I like videos because you can show mistakes. I think it's important, especially for beginners, to see that even experienced developers make mistakes, but also to see how you get to derived code... Because I think there's a lot of times where we'll just show them the final code, and as developers, we know that there might be three refactors that got us there... And I think having a video makes it possible to do that, versus if you're doing a book or something like that, it's much harder, because to actually show them "Well, I went to this line and changed this one thing, then I went to this other file and changed this one thing" - it becomes a lot. + +So I think videos are really good for that interactive -- or not quite interactive, but something like that... But I've also started to find that books are probably more accessible; something that I hate about videos is that you almost need to find translators for a couple different languages, or you need to get somebody to come in and actually write all the transcripts up, because anything automated just doesn't do a good enough job. + +So basically that's something I've been struggling with lately, trying to figure out the right approach to that... Because I think that making videos more accessible is something that needs to happen in the future. + +**David Valentine:** \[41:56\] One of the things that I think is important - because I recognize that my courses have reached 160-something different countries - is having good closed captioning. And I have to admit right now my Golang course doesn't have it yet, but to second Jon's thoughts around closed-captioning -- because some people, English may not be their native language; they speak several different languages, so having something there is extremely useful for them... But the challenge is that the automated closed-captioning just doesn't work. I don't know how many times I've said "Udemy", and it gets translated to "you and me", on the Udemy platform itself... Or something absolutely crazy. And I'm thinking, if there's any word that Udemy would have right in their closed-captioning, it would be Udemy, right? So that is a huge technical challenge and hurdle. + +The other thing - because I did have a background in teaching outside of this, where I've done some workshops in-person in advance of this type of experience, as well as having tutored one-on-one - is you lose that interaction when you have an online video, and it is extremely difficult to iterate, and make changes. There's substitutes for it, but I like to see eyes, and even just "Hello!" and putting content out there into the world - it's a very different platform when you're doing a static video than when you're recording. + +I think ideally what I'd almost prefer to do as I grow and continue to make new courses is to teach in a live event, record those, so at least I've got somebody else I'm talking to. And if you see that deer in headlights look, you get that sense of "Oh, okay, I've fallen off track here. I need some more explanation here." It's a very different approach. + +Mind you, you get unlimited redo's when you're recording video, right? So it's like "Oh, that was -- I need more coffee", or something. So there's pros and cons of each, but everyone has its own flavor and piece to it, right? + +**Ashley Willis:** I feel like every way that we do this is valuable. We all have different learning styles. For me - it's hard for me to consume the content on video. I'm like "Okay, now I have to pause the video, and do this step. Play the video again." And then I go back, "Did I do that step right?" So for me, I like step-by-step instructions, so the workshops are really helpful... But the cons of a workshop is that a lot of people are afraid to ask questions. So they will sit there and act like they know what's going on when they don't know what's going on, so we have to constantly go around and be like "Do you really understand? You can ask questions, it's fine." + +So video, written tutorials, workshops - they're all valuable in their own way; we all learn differently. + +**Jon Calhoun:** Just to chime in one additional -- one of the reasons I have never focused on workshops, conferences, that sort of stuff as much... And I think it's easy to forget when you live in a city, but there's a lot of people who live nowhere near a city, or they don't have the resources to pay for a workshop, or something... To give you an example, I live two hours away from the nearest city, and that's Pittsburgh, which is not exactly a massive city. So just knowing that there's a lot of people out there who cannot go to that type of thing, I definitely think that there's a good -- like you said, we need almost everything, because some people are gonna do better in the workshops, and some people meet more people and they'll actually collaborate with them afterwards, other people will do online courses... + +I think one of the big things I've just noticed is that people find other people to collaborate with and to learn with - that's very huge. + +**Carmen Andoh:** So collaboration, in-person, gauging for deer in the headlights, course correcting - that's typically been the traditional way of learning, right? But it doesn't scale. It doesn't scale the teacher, and it's not accessible to rural and suburban learners. + +**Jon Calhoun:** I think it can scale, it just scales differently. One way I've seen it scale is I've seen people who do online courses, and they'll essentially have -- they call them classes, or whatever... But essentially, a bunch of people sign up and they say "Alright, we're gonna start in December, and every week you're supposed to go through so much content", and then you're all in a Slack channel and you discuss that content, and you're expected to keep up with the classroom on that content. It's less acceptable in the sense that you can't just start whenever you want, but I think there are ways to get that simulated classroom environment. + +\[46:12\] I think as course creators, it's things we have to think about... Like, is there a way that we can make that possible? I think tools like Slack have made it much better, too. I've gotten into this habit of -- all the courses I have, I provide support for, but I've gotten into this habit of "Any of the paid of courses, you've got access to a Slack", and I basically say "Ask your question there." Even if you email it to me and I can answer in five seconds, I say "Ask the question there", because this is the best way to foster this growth between different people taking the course - you each try to answer the questions and help each other out. That will establish that learning, it'll help you get used to helping each other... + +Over time, it's gotten to this point where I can sometimes check the Slack, and somebody asked a question and somebody answered it better than I could have, before I even got to it. + +**Carmen Andoh:** That's great. That's always great, the feedback loops, and in-person for me is when I can get unblocked. I know that the other exercise site that I've tried is Exercism.io, for not just Go, but a variety of languages... And I was a mentor for Go. So when I signed on to be a mentor, they were like "Listen, if you agree to walk through the code exercise, we all as a community try to get the next person in the queue", so that their feedback is as quick as possible, so there is that online mentorship, and trying to scale that differently, as you said. + +**Ashley Willis:** I love Exercism.io. Thanks Katrina and team. Also, I feel like I need to mention it, because it's great and we haven't talked about it yet, but JustForFunc - I love it. + +**Carmen Andoh:** Yeah. And I just wanted to ask Jon, before I talk about other possible resources for gophers... Can you tell us a little bit more about where learners could go for Gophercises? Are they gonna get that content for asking questions within the course, as they sign on? + +**Jon Calhoun:** Gophercises - there's a channel in the GopherSlack, \#gophercises, and there's some people there; I don't know if that one's actually that active... That one's a little bit harder because it's not paid, so supporting it is a little bit trickier. People email me, and I do try to answer where I can, in the Slack, if they ask questions... I try to check there every so often and respond. So all those are options. Because I think that one right now has something like 20,000 students, and it can be a little bit trickier sometimes, whenever I've got a bunch of other things I have to do. But I do try to help, and I do try to answer email, and that sort of stuff, so those are all viable options. + +**Carmen Andoh:** Great. Well, we're about ten minutes out... Let's go beyond Hello, world or Go 101, and Go 201. We have a great foundational course, and we have exercises - where else can Go beginners go next, once they get there? + +**David Valentine:** I'll jump in there. Once you've got your head around some basic Go, you've done lots of different examples and you have the basic language across, I think at that point jumping into effective Go in terms of learning how to write good, clean, idiomatic Go code is your very next step. Then, as Ashley was mentioning when we were talking about building examples, another item I think that is fantastic, which is similar to Awesome Python, is Awesome Go. If you go to awesome-go.com, you'll end up with a list of curated Go frameworks, libraries and software, and anything that floats your boat - whether you're interested in audio and music, or bot building, or any kind of different piece, there are some really outstanding resources available there. + +And then finally, the third thing I'd lead off with in that respect would be to go to Go users groups, because almost any geography that you're in, either there's online forums or in-person groups (especially the in-person groups), you'll find that there are Go user groups. I live in a fairly small city in Canada called Winnipeg, which is smack central in the prairies, and yet we have a thriving Go user group locally. So unless you're rural, you may be able to find and crash a Go user group (even if you're traveling) and to connect with Go users in-person which I think it's amazing. + +**Ashley Willis:** \[50:12\] Plus one to Go user groups. + +**Carmen Andoh:** Yeah, we're actually trying to build that up and strengthen a unified Go user group called The Go Developer Network. I joined Google a couple months ago and I have another colleague that joined the team recently - his name is Van Riper - and he wants to complement GoBridge's efforts, as well as Women Who Go's efforts into getting this network of groups to be able to maybe give them content, help them, maybe something like a live broadcast, give them a chance to go at their own pace... But I really like the idea of community and the importance of in-person learning, and trying to balance that off with the people for whom maybe online courses are not working. + +**Jon Calhoun:** I think it's a really good thing that that's something you focus on. We talked about all these things with learning languages, but we didn't really talk about the community, or the fact that some coding communities really weren't that inclusive or inviting. I think the fact that Go has stressed that from the start is one of the huge things for learning it, especially if you happen to be in a minority group, or something... That can be very helpful, that you can actually feel okay being yourself and not having to pretend you're somebody else as you're learning. + +**Ashley Willis:** Absolutely. That's why I moved from Python to Go. The Python community is fine, but... So I've found that within the Go community there's no such thing as a dumb question, and I'm really good at asking dumb questions. People are just super, super-helpful; they will go out of their way to help you. I've never been part of a community like this. + +**Carmen Andoh:** Agreed. And it's the reason I chose Go and stuck with Go, and I just really am happy for that... So yeah, thank you Jon for bringing that up, because it is, I feel, of paramount importance in terms of learning/mastering the language, and then staying and keeping in community and contributing to the language. + +One thing you mentioned, Dave, was Awesome Go. There's also libhunt.com, which I think works for all languages, and then they have go.libhunt.com... Which is a similar thing. You can go see third-party packages, and compare them to other packages, depending on what you wanna do. They release the newest/greatest, and I kind of like that, to keep up with my things. + +I know that Mark Bates, who is a panelist on this show, along with Cory LaNou, they made Gopher Guides, and I think this is along the same lines of trying to find a curated path based on a certain thing that you wanna learn, whether that's a data structure, or a type, or an algorithm... So I'm gonna give a little shout-out to them for that. + +Any other resources for people to go, or go-to's (pun intended)? We've been very short on Go puns this episode. Shame on us. + +**Jon Calhoun:** I think one that everybody says and it's very hard, but if you can find an open source project you like, it can be daunting at first, because you'll be like "How do I get started here? There's so much there..." But you mentioned Mark Bates - I think one of the things that he's done a great job with Buffalo is that he can help people get started; help with documentation, or help with something that's relatively easy, as you get familiar with stuff. + +But that one comes to mind because there were one or two people that were very early students of one of the courses I made, and they later were telling me that they were actually contributing more heavily to those projects, and it was because they really helped them grow as a developer, and learn more about them, and get to the point where they could actually fix PRs. + +It's not gonna happen overnight. I don't think you can ever, as a beginner, jump into a project and make meaningful code changes, or big code changes. People who run open source projects have to resist this urge to fix a one-line bug; they sort of have to set it aside for a beginner to tackle. But if you can find the right projects, and maybe ask around in the Go community for suggestions, that's a great way to grow, where you just gradually do small changes and work your way up to handling more and more of the codebase. + +**Ashley Willis:** \[53:56\] I agree. That is advice I give often. And what I would like to see in the future is people prioritizing PRs, like "Here's what's great for a beginner. Here's what's great for people that are more advanced." There's a site called Up-For-Grabs. It's not .com, I can't remember what it is (up-for-grabs.net), but it does just that, where it's like "These are great bug fixes that you can do as a beginner. So it will rank things. I would love to see people do that within open source in general. + +**Carmen Andoh:** Yeah. I love when GitHub tags "Great first issue", and then you can just sort by the tags, and they optimize their project not for getting things done necessarily, but for -- well, yes, that's certainly important, but also for being inclusive and trying to onboard new members into their project or their ecosystem and whatnot. + +**Jon Calhoun:** Another one is FirstTimersOnly, where basically they limit specific issues, and you have to be a first-time contributor to that project to do it. + +**Carmen Andoh:** Oh, that's nice. + +**Ashley Willis:** I love it. + +**Carmen Andoh:** And we would be remiss if we didn't mention \#golang-newbies in the GopherSlack. When I first began, I just loved being able to ask, as Ashley said, all the questions. And I was sort of fearless, because I had people that I had already met in person who said "Just ask the question. When you do, if you do that, then it will make people more comfortable in asking what they are afraid of, to ask questions." So yes, \#golang-newbies in GopherSlack. + +**Jon Calhoun:** It's funny, because that reminds me of -- before GoTime got rebooted, I was talking to Mat Ryer and he had said that one of his goals was to say "I don't know" in the podcast, at least once or twice... And his goal for that was basically just so beginners realize that it's okay to admit you don't know and to ask questions. I think that's important, like you said, just to really reinforce that it's okay to ask questions, and that nobody's gonna think you're a bad developer because of it. + +**Carmen Andoh:** Yes. + +**Ashley Willis:** And there's so much that Mat doesn't know... \[laughter\] + +**Carmen Andoh:** We had to throw Mat some shade... But I think our episode is quasi-complete, because we threw MAt some shade we've given Steve a hello, we mentioned Mark and his Gopher Guides... Anything else we're missing before we go? + +**David Valentine:** Just to add on the question piece, one thing I notice with the courses that I teach - so often someone asks a question, and then next thing you know there's a crowd of me-too's. And I'm like "Where were you a minute ago? Why did so-and-so have to be the first one to ask?" And then inevitably, when the answer is made, you get a handful of a dozen or so thank-you's. Make it interactive. Especially when you're trying to bring that about, I really find that people are unnecessarily shy, for whatever reason. + +**Ashley Willis:** It's not even unnecessarily shy... I think that people's fear of embarrassment rules them. All psychological. So there has to be somebody in the room who just doesn't embarrass by not knowing something. + +**Jon Calhoun:** I think also how you answer can make a huge impact... Because if you make it sound like it was something obvious, or - another example is if somebody says "Well, I'm not sure, but I wanna try to help you", and if you just jump in and just disregard everything they said, then they're not gonna try that in the future. So I think there's a lot of thinking about how we answer questions and how we present stuff to people, and just trying to be as gentle or friendly as possible is very important. + +**David Valentine:** Absolutely. + +**Carmen Andoh:** Very good. + +**Ashley Willis:** Definitely encourage those people to ask more questions; it might open it up for other people to ask more questions. + +**Carmen Andoh:** Great! Well, we're kind of coming up on the top of the hour. There's a couple things I wanted to mention... First of all, there's one last segment that I'm really interested in learning more about, and I made a survey; I am not a survey designer, but I really am curious - how many of you here on this panel have had to learn a new programming language while on the job? + +**Ashley Willis:** I have. + +**Carmen Andoh:** Yeah... And is learning while on the job, whether it's because you have to or you're exploring for new tooling, or because it's a change in the technical stack - how does learning or beginning from that headspace any different from learning maybe on your own time? + +**Ashley Willis:** \[58:10\] I feel like it's different, because usually if it's for work, you know why you need to learn it. So you kind of have some base knowledge. You're like "I have to learn it because I need to build this feature, or we're building this tool, and that's why I have to learn it." So you can google more specifically... + +**Carmen Andoh:** Okay. Jon, Dave? + +**Jon Calhoun:** I've had to learn on the job, and like Ashley said, it's specifics, but also your focus is on being productive, not academic learning, so that forces you to not let yourself get distracted as much, I guess is the best way to put it. If you need to build a WebSocket server or something, you're gonna focus more on "I just need to get this done. I can come back and fine-tune it later, or I can come back and look at those things I'm not quite sure about later." Whereas when you're learning it just for leisure, there's no really time restrictions or anything, so you have a little bit more ability to get distracted. + +**David Valentine:** Well, I think that getting distracted is the enjoyment that you have when you decide to explore something in your own personal time, for your own personal benefit. You can go down those rabbit holes, you can explore... It's almost like a buffet of knowledge. And I love reading, and learning, and I listen to a lot of podcasts because I wanna experience those tangential things that I may not otherwise have. But I will say, even though programming isn't a day-to-day part of my job, I think anyone in IT or anyone in the world in general can literally have what happened to me this recent Friday, where the CIO called the boss and said "I want Dave full-time on this other project. Whatever he's doing right now goes away." The next thing you know is "Okay... So let's figure out how we're gonna approach this", and what I need to do there, and what his objectives are, and how do I fulfill those. + +So it's very pointed, because you're trying to get to a particular objective... But the pleasure of just learning something for kicks and giggles - there's a joy with that as well, and it's different than when you're trying to fulfill what the CIO asked on a Friday morning. + +**Jon Calhoun:** \[01:00:13.09\] Yeah. + +**Ashley Willis:** Yes. + +**Jon Calhoun:** I guess I should say, those distractions can be very good things, but I just think that's one of the big differences - sometimes they're good, sometimes they're bad, and I think that having them there... It just is one of the big differences for me, as I just felt like in the work environment I never had that freedom to just really look at the stuff, explore things... You can't go compare three different HTTP routers; it's like "I just need to pick one and use it." + +**Ashley Willis:** Exactly. You definitely get less in the weeds when it's for work. + +**Carmen Andoh:** Very cool. I wanna hear more about this... I am gonna put a link in the chat - it's bit.ly/onjobcodesurvey. I wanna share it on my Twitter, and maybe everyone else can share it... I just wanna know, because it's kind of my one "How is that different?", and are there any gaps out there in the ocean of content that you think that need to be most filled at this point, as we close out? + +**David Valentine:** Children's book. Dang it. \[laughs\] + +**Carmen Andoh:** Alright, I think that's what's happening... Children's book. It is happening. + +**Ashley Willis:** Please do it. Brian Ketelsen and I were supposed to do it, but we are just far too busy. I really need somebody to do this. Just selfishly, please do it. + +**Carmen Andoh:** Amazing. + +**Jon Calhoun:** I almost feel like there needs to be resources teaching people how to ignore all the bad advice they're gonna get. Like Ashley had said, people keep telling you to build something, or if you get on the Go Reddit, they're gonna tell you never to use third-party libraries... There's a lot of things like that, and I think when you're learning, especially as a beginner, it's really easy to get suckered into this "I need to use all these things." It's almost like they need a practical guide to just ignoring all that other stuff that doesn't matter, and just focusing on what does matter. + +**Carmen Andoh:** Yes...! + +**Ashley Willis:** Yes, also please write that. + +**Carmen Andoh:** Yes, I have that as a survey question, which is like "What is this letter to your past self? What do you wish that you didn't have to unnecessarily go through?" So yeah, totally. I think that's great. I wish we had more time to discuss that, but that is just really some of the best -- if we could get rid of some of the headaches that we ourselves had to endure... But maybe that's all part of the journey. + +So I'm gonna close out... I have a closing out section here, so let me just gather my wits and then I'll read it, and then we'll say bye. + +Whatever kind of beginner you are, we hope this episode has given you a good starting point as you start to learn and improve your Go. The learning gotchas, how learning Go might be different from your current programming language, where to ask questions and get help from the community, and where to continue on your learning journey. Whatever you do, embrace failure and don't stop trying. Engage with others. Use your resources - GopherSlack, \#golang-newbies, attend meetups (online or in-person). Like the famous quote from Benjamin Franklin, "Tell me and I forget. Teach me and I may remember. Involve me, and I learn." + +Thanks everyone for today's episode of GoTime FM. Thank you, Ashley, thank you Jon, thank you Dave for joining me. We'll see you next week. diff --git "a/Go for beginners \342\231\273\357\270\217_transcript.txt" "b/Go for beginners \342\231\273\357\270\217_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..f652a384f27632a43840fe0f9b6977a4c9fc77db --- /dev/null +++ "b/Go for beginners \342\231\273\357\270\217_transcript.txt" @@ -0,0 +1,1982 @@ +[0.00 --> 2.56] Hello friends, Jared here, GoTimes producer. +[2.96 --> 5.16] This week we're doing another classic +[5.16 --> 6.44] from the back catalog. +[6.96 --> 9.26] One common request we get from listeners +[9.26 --> 12.18] is for more beginner level conversations. +[12.72 --> 14.90] Well, that's exactly what this episode is. +[15.14 --> 17.72] It was originally recorded on May 7th, 2019, +[18.12 --> 21.26] back when John Calhoun was merely a guest panelist. +[21.56 --> 23.40] We'll be back with some freshens next week. +[23.48 --> 25.50] In the pipeline, we have a debugging episode, +[25.74 --> 27.28] what's new in Go 1.19, +[27.28 --> 30.98] Gophers Say, live from GopherCon EU and more. +[31.50 --> 32.70] Okay, go for beginners. +[33.16 --> 33.58] Enjoy. +[38.08 --> 40.84] This episode is brought to you by Sourcegraph. +[40.94 --> 42.92] With the launch of their Code Insights product, +[43.06 --> 45.92] teams can now track what really matters in their code base. +[46.24 --> 48.34] Code Insights instantly transforms your code base +[48.34 --> 49.96] into a queryable database +[49.96 --> 52.54] to create visual dashboards in seconds. +[52.92 --> 53.94] And I'm here with Joel Cortler, +[54.16 --> 55.98] the product manager of Code Insights for Sourcegraph. +[55.98 --> 58.82] Joel, the way teams can use Code Insights +[58.82 --> 60.56] seems to pretty much be limitless, +[60.74 --> 63.68] but a particular problem every engineering team has +[63.68 --> 66.14] is tracking versions of languages or packages. +[66.64 --> 69.34] How big of a deal is it actually to track versions for teams? +[69.80 --> 71.54] Yeah, it's a big deal for a couple of reasons. +[71.72 --> 73.54] The first is, of course, just compatibility. +[73.78 --> 75.44] You don't want things to break when you're testing locally +[75.44 --> 77.66] or to break on your CI systems or test systems. +[78.04 --> 80.40] You need to have some sort of level of version unification +[80.40 --> 81.40] and minimum version support, +[81.56 --> 84.30] and all of that needs to be compatible forward. +[84.30 --> 87.02] But the other thing we learned was that for a lot of customers, +[87.14 --> 88.44] especially, you know, engineering organizations +[88.44 --> 89.50] that are pretty established, +[89.94 --> 91.40] they have older versions of things +[91.40 --> 93.32] or even older versions of, like, SaaS tools +[93.32 --> 94.08] they don't use anymore +[94.08 --> 95.58] that they haven't fully removed +[95.58 --> 96.96] because they're, like, not sure if it's still in use +[96.96 --> 98.46] or they, you know, lost focus on that. +[98.76 --> 100.58] And they're spinning up old virtual machines +[100.58 --> 101.58] that they're still paying for. +[101.68 --> 103.26] They're using, you know, old SaaS subscriptions +[103.26 --> 104.08] they're afraid to cancel +[104.08 --> 105.70] because they're not sure if anyone's actually using it. +[105.80 --> 107.16] And so getting off of those versions +[107.16 --> 108.78] not just, like, saves you the headaches +[108.78 --> 109.94] and the risks and the vulnerabilities +[109.94 --> 111.22] of being on old versions, +[111.22 --> 113.36] but also literally the money of, you know, +[113.42 --> 115.06] older systems running more slowly +[115.06 --> 115.90] or the build times +[115.90 --> 117.36] or, you know, virtual machines +[117.36 --> 119.08] and SaaS tools that you're no longer using. +[119.40 --> 120.56] Before you had this ability, +[120.66 --> 121.28] we talked to teams +[121.28 --> 122.96] there were basically three ways you could do this. +[123.24 --> 124.50] You could slack a million people +[124.50 --> 126.68] and ask for just, like, an update point in time. +[126.94 --> 128.46] You could have sort of one human +[128.46 --> 129.12] and one spreadsheet +[129.12 --> 130.68] where, like, it's somebody's job +[130.68 --> 132.02] every Friday or every two weeks +[132.02 --> 133.64] to just, like, search all the code +[133.64 --> 134.56] and find all the versions +[134.56 --> 135.94] and write it down in a Google sheet. +[136.20 --> 137.66] Or there were a couple of companies +[137.66 --> 139.58] I came across with in-house systems +[139.58 --> 140.48] that were sort of complicated. +[140.68 --> 142.44] You had to know, you know, maybe Kotlin, +[142.62 --> 143.60] but you didn't know Kotlin. +[143.66 --> 144.54] But if you wanted to use this system, +[144.58 --> 145.30] you had to learn Kotlin +[145.30 --> 146.64] and you'd have to sort of build +[146.64 --> 147.66] the whole world from scratch +[147.66 --> 149.52] and run basically a tool like this +[149.52 --> 150.88] with a pretty steep learning curve. +[151.18 --> 152.52] And now for all three of those, +[152.56 --> 153.26] you could replace it +[153.26 --> 155.00] with a single line, Sourcegraph Search, +[155.12 --> 156.08] which is basically just +[156.08 --> 156.82] the name of the thing +[156.82 --> 157.34] you're trying to track +[157.34 --> 158.04] and the version string +[158.04 --> 158.80] in the right format. +[159.08 --> 160.00] And then we have templates +[160.00 --> 160.76] that'll help you get started +[160.76 --> 161.60] if you're not sure +[161.60 --> 162.32] what that format is. +[162.48 --> 163.54] And then it'll automatically track +[163.54 --> 164.68] all the different versions for you, +[164.86 --> 165.58] both historically. +[165.76 --> 166.74] So even if you start using it today, +[166.82 --> 167.90] you can see your historical patterns. +[167.90 --> 169.38] And then, of course, going forward. +[169.96 --> 170.64] Very cool. Thank you, Joel. +[170.72 --> 172.12] So right now there is +[172.12 --> 174.42] a treasure trove of insights +[174.42 --> 175.18] just waiting for you. +[175.42 --> 178.04] Living inside your code base right now, +[178.32 --> 180.10] teams are tracking migrations, +[180.44 --> 182.02] adoption, deprecations. +[182.36 --> 183.50] They're detecting and tracking +[183.50 --> 185.20] versions of languages and packages. +[185.20 --> 187.18] They're removing or ensuring +[187.18 --> 189.18] the removal of security vulnerabilities. +[189.56 --> 191.14] They understand their code by team. +[191.22 --> 192.90] They can track their code smells and health +[192.90 --> 193.90] and they can visualize configurations +[194.52 --> 196.60] and services and so much more +[196.60 --> 197.50] with code insights. +[197.90 --> 200.18] A good next step is to go to +[200.18 --> 202.18] about.sourcegraph.com +[202.18 --> 204.10] slash code dash insights. +[204.38 --> 205.88] See how other teams are using +[205.88 --> 206.90] this awesome feature. +[207.22 --> 209.84] Again, about.sourcegraph.com +[209.84 --> 211.98] slash code dash insights. +[212.24 --> 213.98] This link is in the show notes. +[213.98 --> 230.04] Let's do it. +[230.26 --> 231.70] It's go time. +[232.26 --> 233.94] Welcome to go time, +[234.10 --> 235.78] your source for diverse discussions +[235.78 --> 237.26] from around the go community. +[237.26 --> 240.20] We record live on Tuesdays at 3 p.m. +[240.20 --> 240.62] U.S. +[240.68 --> 241.08] Eastern. +[241.46 --> 243.36] Subscribe at youtube.com +[243.36 --> 245.24] slash changelog to be notified +[245.24 --> 247.14] so you can be part of the action. +[247.48 --> 249.30] And don't forget to follow the show on Twitter. +[249.46 --> 250.72] We're at go time FM. +[250.92 --> 252.84] Special thanks to our partners at Fastly. +[253.04 --> 255.24] Everything we ship here at Changelog is fast +[255.24 --> 257.80] because Fastly serves it up super fast +[257.80 --> 258.74] everywhere on Earth. +[258.98 --> 261.04] Check them out at fastly.com. +[261.44 --> 262.72] Okay, here we go. +[262.72 --> 266.04] Welcome back, everyone, to go time. +[266.58 --> 269.52] I am Carmen Ondo and Matt is out this week, +[269.60 --> 272.06] so I will be your host for the journey. +[272.94 --> 275.36] Joining me today is our recurring panelist, +[275.54 --> 276.58] Ashley McNamara. +[277.16 --> 278.30] Say hello, Ashley. +[278.76 --> 279.82] Hello, everyone. +[280.46 --> 283.26] We have a returning guest, John Calhoun, +[283.26 --> 288.40] who is the content founder for gophersizes.com, +[288.48 --> 292.26] but also use goling.com, testwithgo.com, +[292.72 --> 293.94] errorsandgo.com. +[294.64 --> 296.40] And so I'm excited to have you on, John. +[296.50 --> 296.98] Say hi. +[297.24 --> 297.68] Hi, everybody. +[298.08 --> 300.32] And now we have a new panel, +[300.44 --> 301.04] or new guest, +[301.36 --> 302.80] and it's Dave Valentine, +[303.28 --> 305.52] and he's the instructor of Udemy's +[305.52 --> 308.42] A Gentle Introduction to GoLang for Beginners. +[309.12 --> 310.20] Welcome to our show. +[310.60 --> 311.54] Thank you so much. +[311.58 --> 313.28] It's a pleasure to be here for the first time. +[313.68 --> 314.58] Speaking of beginnings, +[314.80 --> 318.40] when we asked Dave to show up on the show +[318.40 --> 319.70] and give a profile pic, +[319.70 --> 323.04] he submitted one of him as like a toddler +[323.04 --> 327.20] with a TSR-80 Model 1 circa 1980. +[327.96 --> 331.36] So I think that's probably like when you began. +[331.72 --> 332.24] Absolutely. +[332.60 --> 333.78] And Carmen, I have to say, +[333.88 --> 335.02] I was chuckling when you're like, +[335.10 --> 337.00] what computer is that? +[337.00 --> 339.72] But yes, yes, +[339.72 --> 343.02] I got a start in the microcomputer generation world. +[343.16 --> 345.26] So I started with a TSR-80 +[345.26 --> 347.98] and then moved over to a PC junior +[347.98 --> 353.26] and then continued on through microcomputers +[353.26 --> 356.42] and programming way back at that time. +[356.52 --> 358.10] So I've had an early start on things. +[358.54 --> 359.10] Wow. +[359.66 --> 360.50] That is hard. +[360.86 --> 361.94] That is hardcore nerd. +[362.22 --> 362.44] Yeah. +[362.82 --> 364.74] Total nostalgia ahoy, +[364.74 --> 366.00] I'm sure for many people. +[366.20 --> 368.04] Certainly you can share what you started +[368.04 --> 368.98] with your beginning, +[369.52 --> 372.14] your first computer in the GoTime FM channel +[372.14 --> 373.10] on GoForSlack. +[373.32 --> 373.68] Okay. +[373.76 --> 376.60] So let's talk a little bit about today's theme. +[377.16 --> 379.30] This is GoLang for beginners. +[380.14 --> 382.72] I had originally envisioned this episode +[382.72 --> 383.88] to engage, +[384.08 --> 386.66] to be meant to engage for both non-Go users +[386.66 --> 390.04] that listen to maybe sister podcasts on ChangeLab +[390.04 --> 393.28] or any Go curious programmers out there, +[394.00 --> 396.14] as well as encourage those +[396.14 --> 397.68] that have just started with Go +[397.68 --> 400.20] and want to level up beyond the basics. +[400.62 --> 403.04] So I'm hoping that maybe we can start +[403.04 --> 405.72] and just kind of see where the conversation takes us. +[405.78 --> 409.74] We have three wonderful experts online today +[409.74 --> 411.80] and see what their perspectives +[411.80 --> 412.86] and their experiences +[412.86 --> 414.18] and how they've shared content +[414.18 --> 415.80] to help Gophers worldwide. +[415.80 --> 416.24] Okay. +[416.96 --> 418.02] So I guess the first thing +[418.02 --> 420.24] I wanted to start with is ask, +[420.78 --> 422.04] what is a beginner? +[422.46 --> 424.80] Like what types of beginners are there? +[425.22 --> 425.52] Anyone? +[426.46 --> 427.64] Good question. +[428.64 --> 429.92] So there are beginners +[429.92 --> 432.30] that know one programming language +[432.30 --> 433.42] that want to learn another. +[433.76 --> 436.24] There are people that are career transitioning +[436.24 --> 438.98] that don't know any programming. +[439.36 --> 441.08] There's all kinds of beginners. +[441.68 --> 442.04] Yeah. +[442.16 --> 442.48] I was going to say, +[442.54 --> 444.04] there's even the curious kids +[444.04 --> 445.50] who don't know what they're doing +[445.50 --> 447.06] and like to break their parents' technology, +[447.42 --> 449.40] which is probably some of us. +[449.62 --> 449.90] Yeah. +[450.12 --> 451.02] Those are my favorites. +[451.48 --> 451.68] Yeah. +[452.12 --> 452.52] Absolutely. +[452.68 --> 453.54] I've experienced that +[453.54 --> 454.92] with one of the courses +[454.92 --> 456.32] that I've published +[456.32 --> 457.92] that one of the kids was, +[458.38 --> 458.90] what are we saying? +[458.98 --> 460.60] I don't have system privilege +[460.60 --> 462.64] to install Go on my computer. +[462.76 --> 463.44] What do I do? +[463.62 --> 463.86] Oh. +[464.12 --> 464.48] You know, +[464.58 --> 466.02] and how do I get around +[466.02 --> 466.94] that kind of a thing? +[467.14 --> 467.34] So, +[467.68 --> 469.18] because his dad had bought him +[469.18 --> 469.98] the course. +[470.42 --> 471.18] I've actually, +[471.58 --> 472.24] in my program, +[472.24 --> 473.60] originally created a course +[473.60 --> 475.32] that was in advance +[475.32 --> 476.28] of the one I currently have +[476.28 --> 476.78] in the marketplace, +[476.94 --> 477.80] which I've since retired, +[478.28 --> 479.70] that was kind of trying to get +[479.70 --> 481.00] the entire beginners +[481.00 --> 481.78] in Go space, +[481.88 --> 483.46] where it's for people +[483.46 --> 484.12] that are experienced +[484.12 --> 484.70] with programming +[484.70 --> 485.76] and it's for people +[485.76 --> 487.42] that are new to programming +[487.42 --> 488.32] for the first time. +[488.46 --> 489.90] But what I found was +[489.90 --> 491.04] I was trying to appeal +[491.04 --> 492.26] to too many different people +[492.26 --> 494.04] and so the course content +[494.04 --> 495.08] wasn't that effective. +[495.72 --> 496.40] And since then, +[496.42 --> 497.34] I've actually specialized +[497.34 --> 500.00] in making something very clear. +[500.00 --> 501.64] It is for absolute beginners +[501.64 --> 503.48] because I actually walked them +[503.48 --> 504.58] through some of the fundamentals +[504.58 --> 505.70] of computer science +[505.70 --> 507.60] in learning Go +[507.60 --> 509.12] and giving them a foundation +[509.12 --> 510.18] if they've never seen +[510.18 --> 510.86] anything before. +[511.06 --> 511.42] So, +[511.48 --> 512.26] I've actually learned +[512.26 --> 512.78] to specialize. +[513.38 --> 514.68] And is this the course +[514.68 --> 516.08] that I mentioned before +[516.08 --> 517.30] or is this a different course? +[517.68 --> 519.00] It is exactly the course +[519.00 --> 519.78] you just mentioned. +[519.94 --> 520.16] Yes, +[520.20 --> 521.30] that one's been in the marketplace +[521.30 --> 523.34] since January of this year +[523.34 --> 524.84] and so it's a reboot +[524.84 --> 526.28] specifically +[526.28 --> 527.78] and I've taken the other one +[527.78 --> 528.64] out of the marketplace. +[528.64 --> 531.12] It is specifically intended +[531.12 --> 532.48] for people that have never +[532.48 --> 533.86] seen programming languages before +[533.86 --> 535.44] and want to get started +[535.44 --> 536.50] with the exciting +[536.50 --> 537.62] computer language Go. +[538.24 --> 538.76] That's great +[538.76 --> 540.08] and just a reminder to everyone +[540.08 --> 541.76] what Dave is referring to +[541.76 --> 542.48] is Udemy +[542.48 --> 543.60] Udemy.com +[543.60 --> 544.68] and it's called +[544.68 --> 546.08] A Gentle Introduction +[546.08 --> 546.86] to Golang +[546.86 --> 548.20] for Beginners. +[548.76 --> 549.00] So, +[549.12 --> 549.72] is that what you mean +[549.72 --> 550.24] by gentle? +[550.50 --> 550.68] So, +[550.86 --> 551.68] the idea that you're +[551.68 --> 552.52] a complete beginner +[552.52 --> 554.02] not just a beginner +[554.02 --> 554.70] to Go. +[555.38 --> 556.48] Exactly right. +[556.56 --> 557.38] That is just it +[557.38 --> 558.18] because I did find +[558.18 --> 558.98] that people were like, +[559.14 --> 559.36] you know, +[559.88 --> 560.80] what tends to happen +[560.80 --> 561.40] with beginners +[561.40 --> 562.30] is that +[562.30 --> 564.28] they end up +[564.28 --> 565.02] making assumptions +[565.02 --> 565.90] about what the computer +[565.90 --> 566.58] is doing underneath +[566.58 --> 568.44] and they end up +[568.44 --> 569.80] almost with a heuristic +[569.80 --> 570.30] knowledge +[570.30 --> 571.80] or almost some voodoo. +[571.92 --> 572.74] Why is it like that? +[572.80 --> 573.02] Well, +[573.42 --> 574.66] I prefer to try +[574.66 --> 576.30] to explain tip to tail +[576.30 --> 577.32] everything that's happening. +[577.88 --> 578.44] Maybe it's that +[578.44 --> 579.54] deep computer background +[579.54 --> 580.70] from that TSR-80 +[580.70 --> 581.34] time frame +[581.34 --> 583.06] but I get into memory +[583.06 --> 584.80] and what computer code +[584.80 --> 585.70] actually looks like +[585.70 --> 586.20] and the role +[586.20 --> 586.88] of a compiler +[586.88 --> 589.20] and try to expose +[589.20 --> 590.08] all the pieces +[590.08 --> 591.00] so at the end +[591.00 --> 591.98] the light bulb +[591.98 --> 592.70] goes on +[592.70 --> 593.88] and somebody understands +[593.88 --> 594.86] especially when it comes +[594.86 --> 595.58] to understanding +[595.58 --> 596.56] the concept +[596.56 --> 597.68] of what's at +[597.68 --> 598.42] a memory location +[598.42 --> 599.34] versus what +[599.34 --> 600.04] a memory location +[600.04 --> 600.82] is itself. +[601.38 --> 601.48] So, +[601.58 --> 602.02] that is, +[602.08 --> 602.60] you know, +[602.62 --> 603.36] the type of thing +[603.36 --> 603.92] journey +[603.92 --> 604.52] that I bring +[604.52 --> 605.26] with people through +[605.26 --> 606.70] specifically though +[606.70 --> 607.60] in relation +[607.60 --> 608.64] to the Computer Language Go. +[608.64 --> 610.20] that is the kind +[610.20 --> 610.64] of course +[610.64 --> 611.72] that I value +[611.72 --> 612.38] specifically +[612.38 --> 612.86] as somebody +[612.86 --> 613.70] who does not +[613.70 --> 614.90] have a CS degree +[614.90 --> 616.46] I think I get +[616.46 --> 617.46] into the weeds +[617.46 --> 618.12] often +[618.12 --> 619.18] because what happens +[619.18 --> 620.52] I'm my own worst enemy +[620.52 --> 622.06] what happens is +[622.06 --> 623.16] I try to learn something +[623.16 --> 623.46] and I'm like +[623.46 --> 623.84] no, +[623.88 --> 623.94] no, +[623.96 --> 624.54] I need to know +[624.54 --> 625.52] the thing before that +[625.52 --> 626.16] and then the thing +[626.16 --> 626.88] before that +[626.88 --> 627.50] and then the thing +[627.50 --> 628.26] before that +[628.26 --> 629.34] until I'm like +[629.34 --> 630.20] have I learned +[630.20 --> 630.72] anything +[630.72 --> 631.92] at all? +[632.06 --> 632.40] So, +[632.50 --> 633.34] I like to get +[633.34 --> 634.08] into the weeds +[634.08 --> 634.96] a little bit +[634.96 --> 636.58] but if it's +[636.58 --> 637.66] self-guided +[637.66 --> 638.16] sometimes +[638.16 --> 639.14] I'm too +[639.14 --> 640.00] in the weeds. +[640.68 --> 641.50] I think this is +[641.50 --> 642.02] one of the reasons +[642.02 --> 642.76] why at times +[642.76 --> 643.20] I feel like +[643.20 --> 643.86] the younger you are +[643.86 --> 644.40] the easier it is +[644.40 --> 645.04] to learn this stuff +[645.04 --> 645.94] because you're +[645.94 --> 647.00] less likely to +[647.00 --> 647.36] try, +[647.44 --> 648.18] like you're more +[648.18 --> 648.62] willing to just +[648.62 --> 649.38] take things for granted +[649.38 --> 649.78] I think +[649.78 --> 651.04] whereas like +[651.04 --> 651.56] the older you get +[651.56 --> 651.92] you're like +[651.92 --> 652.98] I just think +[652.98 --> 653.68] that it's easy +[653.68 --> 653.98] to, +[654.34 --> 654.44] you know, +[654.46 --> 655.14] as an adult +[655.14 --> 655.50] to like +[655.50 --> 656.38] think I need +[656.38 --> 656.96] to figure out +[656.96 --> 657.34] what all these +[657.34 --> 657.88] things are doing +[657.88 --> 658.66] and understand it all +[658.66 --> 659.14] and as a kid +[659.14 --> 659.78] you're just like +[659.78 --> 660.82] I'm willing to +[660.82 --> 661.88] just trust you +[661.88 --> 662.54] and just go with it +[662.54 --> 663.44] and you know +[663.44 --> 664.04] let it go. +[664.04 --> 664.96] Interesting. +[665.30 --> 665.74] So you think +[665.74 --> 666.46] that's kind of +[666.46 --> 667.30] maybe the key +[667.30 --> 668.28] is just sort of +[668.28 --> 669.38] trusting how +[669.38 --> 670.72] whoever has shaped +[670.72 --> 671.36] the content +[671.36 --> 671.90] is going to +[671.90 --> 672.76] structure the content +[672.76 --> 673.76] that's how +[673.76 --> 674.06] you're going to +[674.06 --> 674.74] learn or? +[675.12 --> 675.98] I think sometimes +[675.98 --> 677.46] like one example +[677.46 --> 678.00] I can give +[678.00 --> 679.58] is I talked to +[679.58 --> 680.36] somebody who +[680.36 --> 681.38] had started a company +[681.38 --> 682.36] and when they +[682.36 --> 682.98] started the company +[682.98 --> 683.74] they couldn't find +[683.74 --> 684.48] a technical founder +[684.48 --> 685.38] so they basically +[685.38 --> 685.92] just went and +[685.92 --> 686.36] learned programming +[686.36 --> 686.82] on their own +[686.82 --> 687.88] and I was talking +[687.88 --> 688.40] to them about +[688.40 --> 688.96] like how they +[688.96 --> 689.66] learned to program +[689.66 --> 691.04] and the one thing +[691.04 --> 691.78] I found interesting +[691.78 --> 692.80] was that they +[692.80 --> 693.48] basically forced +[693.48 --> 694.02] themselves to +[694.02 --> 694.38] go through +[694.38 --> 695.30] like three or four +[695.30 --> 696.12] complete like web +[696.12 --> 696.80] development courses +[696.80 --> 697.70] and they said +[697.70 --> 698.32] the first time +[698.32 --> 698.76] there was a lot +[698.76 --> 699.14] of things that +[699.14 --> 699.48] were mentioned +[699.48 --> 699.96] that they just +[699.96 --> 700.90] didn't quite get +[700.90 --> 701.92] but by the time +[701.92 --> 702.32] they went through +[702.32 --> 702.84] like their third +[702.84 --> 704.04] course some of +[704.04 --> 704.40] those things +[704.40 --> 704.94] like they had +[704.94 --> 705.78] enough understanding +[705.78 --> 706.38] and foundation +[706.38 --> 707.02] that those more +[707.02 --> 707.72] intricate details +[707.72 --> 708.34] started to make +[708.34 --> 709.62] sense and I think +[709.62 --> 710.12] if you get too +[710.12 --> 710.76] focused on them +[710.76 --> 711.74] early you just +[711.74 --> 712.56] you know go down +[712.56 --> 713.12] too many rabbit +[713.12 --> 713.98] holes and you +[713.98 --> 714.32] don't really +[714.32 --> 714.84] you know get +[714.84 --> 716.08] anywhere but if +[716.08 --> 716.52] you're willing to +[716.52 --> 717.18] just push yourself +[717.18 --> 717.72] through it and +[717.72 --> 718.46] realize that I'll +[718.46 --> 719.04] come back to this +[719.04 --> 719.56] later or we'll +[719.56 --> 719.96] go through the +[719.96 --> 720.74] material again later +[720.74 --> 721.36] then that can +[721.36 --> 721.82] really help. +[721.82 --> 723.60] That's a good +[723.60 --> 723.98] point. +[724.14 --> 724.88] Now you've +[724.88 --> 725.50] mentioned a little +[725.50 --> 726.08] bit about like +[726.08 --> 726.86] kids and whatnot +[726.86 --> 728.38] I also wanted to +[728.38 --> 729.08] ask a little bit +[729.08 --> 729.76] what everyone's +[729.76 --> 730.42] thoughts were about +[730.42 --> 731.08] like industry +[731.08 --> 731.84] trends and +[731.84 --> 732.90] educational backgrounds +[732.90 --> 733.54] because I'm +[733.54 --> 734.34] seeing that we +[734.34 --> 735.72] are increasingly +[735.72 --> 736.54] having people +[736.54 --> 736.90] within our +[736.90 --> 737.50] industry that are +[737.50 --> 738.50] not formally +[738.50 --> 739.18] trained with the +[739.18 --> 739.98] university CS +[739.98 --> 741.42] degree but maybe +[741.42 --> 742.90] are self-taught or +[742.90 --> 743.76] they went to a +[743.76 --> 744.16] boot camp. +[744.28 --> 744.96] I know Ashley that +[744.96 --> 745.62] was your path +[745.62 --> 745.92] right? +[745.92 --> 746.92] So is there +[746.92 --> 747.98] anything in terms +[747.98 --> 749.08] of like difference +[749.08 --> 750.08] between how we +[750.08 --> 751.36] teach people that +[751.36 --> 753.06] are not you know +[753.06 --> 754.64] CS degree learners? +[755.34 --> 756.74] So I what so +[756.74 --> 757.50] boot camps are +[757.50 --> 758.78] predatory sorry all +[758.78 --> 760.62] boot camp grads I +[760.62 --> 761.84] value you you don't +[761.84 --> 762.28] know they are +[762.28 --> 762.72] predatory. +[763.58 --> 764.98] I did all of the +[764.98 --> 766.38] open courseware for +[766.38 --> 767.28] all of the major +[767.28 --> 768.04] universities. +[768.92 --> 770.72] I that material was +[770.72 --> 772.18] way more valuable but +[772.18 --> 773.66] it's hard when you +[773.66 --> 774.68] don't know where to +[774.68 --> 775.08] start. +[775.08 --> 776.06] You believe that you +[776.06 --> 776.80] need a classroom +[776.80 --> 777.24] setting. +[777.74 --> 779.02] Some of us are not +[779.02 --> 780.56] in a place where we +[780.56 --> 782.40] can go get our CS +[782.40 --> 782.94] degrees. +[783.10 --> 784.16] That is a place of +[784.16 --> 785.78] privilege and so +[785.78 --> 786.60] some of us have to +[786.60 --> 786.76] learn. +[786.88 --> 787.50] So with all of +[787.50 --> 789.14] these online courses +[789.14 --> 791.00] out there there are +[791.00 --> 791.92] many many more +[791.92 --> 793.14] opportunities to +[793.14 --> 794.20] self-pace and +[794.20 --> 795.84] self-teach and not +[795.84 --> 796.52] all of them are +[796.52 --> 797.72] great so it takes +[797.72 --> 799.04] some time to get +[799.04 --> 800.38] through them but +[800.38 --> 801.16] for me I want to +[801.16 --> 802.04] say that the most +[802.04 --> 804.20] valuable learning that +[804.20 --> 805.06] I did when I was +[805.06 --> 805.74] learning to program +[805.74 --> 806.64] was the open +[806.64 --> 807.10] courseware. +[808.44 --> 809.58] John and Dave can +[809.58 --> 810.56] you tell us a little +[810.56 --> 811.34] bit about your +[811.34 --> 812.70] personal journeys as +[812.70 --> 814.24] beginners both +[814.24 --> 815.06] beginners but just +[815.06 --> 816.56] beginners to go how +[816.56 --> 817.42] you leveled up +[817.42 --> 817.72] there. +[818.06 --> 818.52] Absolutely. +[818.92 --> 819.76] So I'll jump in +[819.76 --> 820.28] there first. +[820.50 --> 822.30] So go is gosh my +[822.30 --> 823.40] eighth ninth tenth I'm +[823.40 --> 824.36] not even sure anymore +[824.36 --> 827.24] computer language and +[827.24 --> 828.46] obviously having a +[828.46 --> 829.28] background in computer +[829.28 --> 830.06] science and having +[830.06 --> 831.20] done a lot of +[831.20 --> 832.54] things with with it. +[832.54 --> 834.12] Python is another +[834.12 --> 834.42] one of the +[834.42 --> 835.16] languages that I do +[835.16 --> 835.92] a lot with because +[835.92 --> 837.26] I also teach and +[837.26 --> 838.06] introduce people to +[838.06 --> 838.78] machine learning and +[838.78 --> 839.66] artificial intelligence +[839.66 --> 840.64] concepts and courses. +[841.36 --> 843.22] With Go to be honest +[843.22 --> 845.06] it sort of came about +[845.06 --> 846.08] for me from market +[846.08 --> 847.52] research because I +[847.52 --> 848.54] became curious about +[848.54 --> 849.58] what I should make my +[849.58 --> 850.64] next great course on +[850.64 --> 852.34] and then I found this +[852.34 --> 853.56] amazing computer +[853.56 --> 855.10] language that really +[855.10 --> 856.06] is a next generation +[856.06 --> 856.72] computer language +[856.72 --> 857.38] because almost any +[857.38 --> 858.12] other language that +[858.12 --> 858.98] we may even consider +[858.98 --> 860.32] new was really +[860.32 --> 861.08] developed in the last +[861.08 --> 861.46] century. +[861.46 --> 861.88] Right. +[862.00 --> 863.32] And I love the +[863.32 --> 863.94] ghost story. +[864.26 --> 865.82] The legend being that +[865.82 --> 866.68] people are waiting for +[866.68 --> 868.04] a C++ program to +[868.04 --> 868.92] finish compiling and +[868.92 --> 870.16] said hey if we were +[870.16 --> 871.02] to develop a language +[871.02 --> 873.34] now what could should +[873.34 --> 874.54] would that look like +[874.54 --> 876.02] and you've got some +[876.02 --> 876.92] brilliant engineers +[876.92 --> 878.72] that that ended up +[878.72 --> 879.66] putting it together +[879.66 --> 880.76] and so I became +[880.76 --> 882.40] absolutely fascinated by +[882.40 --> 884.30] it and realized that +[884.30 --> 885.24] here was a language +[885.24 --> 886.32] that back in the day +[886.32 --> 888.10] when I learned C for +[888.10 --> 889.52] the first time very +[889.52 --> 890.52] much smelled like +[890.52 --> 891.64] tasted like acted +[891.64 --> 893.30] like C exposed +[893.30 --> 894.04] some of those +[894.04 --> 894.88] fundamental computer +[894.88 --> 896.54] bits but had grown +[896.54 --> 897.78] into being so much +[897.78 --> 898.46] more than a system +[898.46 --> 900.12] based language and so +[900.12 --> 901.52] I literally fell in +[901.52 --> 902.30] love with it and +[902.30 --> 903.78] then because I had +[903.78 --> 904.86] intended to develop a +[904.86 --> 906.34] course on it what I +[906.34 --> 908.22] really try to do in +[908.22 --> 909.16] all the courses that I +[909.16 --> 911.06] teach is I try to +[911.06 --> 912.60] develop a road map so +[912.60 --> 914.10] that my students aren't +[914.10 --> 915.26] sort of depending on +[915.26 --> 915.42] me. +[915.52 --> 917.06] My intent like with +[917.06 --> 918.84] the internet there is +[918.84 --> 920.74] this amazing amount +[920.74 --> 921.74] of material that's +[921.74 --> 922.50] out there but it's +[922.50 --> 923.66] not curated in a +[923.66 --> 926.06] meaningful way and +[926.06 --> 927.38] that is maybe the +[927.38 --> 928.34] secret sauce that I +[928.34 --> 929.06] try to bring to the +[929.06 --> 931.12] table right it's you +[931.12 --> 931.64] know because people +[931.64 --> 932.48] have the itch they +[932.48 --> 933.66] have that how do I +[933.66 --> 935.60] get up to speed and +[935.60 --> 937.40] make meaningful use of +[937.40 --> 939.56] my time in order to +[939.56 --> 940.58] achieve a learning +[940.58 --> 941.90] understanding with go +[941.90 --> 943.24] or whatever their +[943.24 --> 945.12] thing is right and so +[945.12 --> 946.52] in the course I have +[946.52 --> 947.08] what I call an +[947.08 --> 949.08] emergency Golang +[949.08 --> 950.26] parachute which is +[950.26 --> 951.24] learning resources +[951.24 --> 953.48] right out of the gate +[953.48 --> 954.72] right saying if you get +[954.72 --> 956.50] stuck here's all the +[956.50 --> 958.18] really other cool stuff +[958.18 --> 959.90] that's out there that +[959.90 --> 961.16] you should know about +[961.16 --> 962.04] so that you can take +[962.04 --> 963.10] advantage of that in +[963.10 --> 963.82] order to supplement +[963.82 --> 965.24] your learning and then +[965.24 --> 966.38] I finish off the course +[966.38 --> 968.18] with here's where to go +[968.18 --> 969.90] from here and and to +[969.90 --> 970.98] continue your learning +[970.98 --> 973.00] journey and then take +[973.00 --> 974.52] them through that piece +[974.52 --> 975.02] where they can +[975.02 --> 977.16] effectively then use +[977.16 --> 978.26] those examples and +[978.26 --> 979.68] resources and they +[979.68 --> 980.64] have the you know +[980.64 --> 981.44] it's that very +[981.44 --> 982.62] beginning that that +[982.62 --> 984.12] beginning is so hard +[984.12 --> 985.68] for people that don't +[985.68 --> 986.78] have any experience so +[986.78 --> 987.56] giving them that +[987.56 --> 989.06] ignition even that +[989.06 --> 990.94] permission to break +[990.94 --> 992.12] things and experiment +[992.12 --> 993.10] with things and to +[993.10 --> 995.34] think about things as +[995.34 --> 996.18] they're coding and +[996.18 --> 996.92] developing their +[996.92 --> 998.24] exercises and so on +[998.24 --> 1000.44] gets them that little +[1000.44 --> 1001.60] bit of traction with +[1001.60 --> 1002.88] their wheels and gets +[1002.88 --> 1003.90] them started and the +[1003.90 --> 1004.56] more students I can +[1004.56 --> 1005.72] get started the +[1005.72 --> 1007.02] happier I am that's +[1007.02 --> 1008.24] great I think that's +[1008.24 --> 1009.48] also a really good +[1009.48 --> 1010.96] what I say companion +[1010.96 --> 1012.14] or complimentary to +[1012.14 --> 1013.86] what John does which +[1013.86 --> 1014.86] is okay you have an +[1014.86 --> 1016.94] ignition you know the +[1016.94 --> 1017.60] basics and the +[1017.60 --> 1018.90] foundations the next +[1018.90 --> 1019.52] thing you got to do +[1019.52 --> 1022.50] is go for sizes right +[1022.50 --> 1023.90] John tell us a little +[1023.90 --> 1025.70] bit about that yeah so +[1025.70 --> 1026.96] I mean basically +[1026.96 --> 1028.14] whenever I learned what +[1028.14 --> 1029.40] I generally found was +[1029.40 --> 1030.86] that it didn't matter +[1030.86 --> 1032.28] if I was coding the +[1032.28 --> 1033.04] prettiest code or +[1033.04 --> 1034.10] anything it generally +[1034.10 --> 1035.26] came down to if I +[1035.26 --> 1036.22] coded a lot I learned +[1036.22 --> 1037.48] a lot and if I didn't +[1037.48 --> 1038.74] build things I you +[1038.74 --> 1039.40] know I sort of just +[1039.40 --> 1041.20] stopped learning and +[1041.20 --> 1042.56] I think a lot of times +[1042.56 --> 1043.56] I've talked to tons and +[1043.56 --> 1044.34] tons of people who are +[1044.34 --> 1044.84] like you know I'm +[1044.84 --> 1045.66] trying to build things +[1045.66 --> 1046.40] but I can't come up +[1046.40 --> 1047.48] with a good project and +[1047.48 --> 1048.48] what's even worse is if +[1048.48 --> 1048.98] they come up with a +[1048.98 --> 1049.98] project they like like I +[1049.98 --> 1050.52] had a friend who did +[1050.52 --> 1051.74] this all the time he'd +[1051.74 --> 1052.42] say I have this cool +[1052.42 --> 1053.70] project and he presented +[1053.70 --> 1054.62] to me and I'd be like +[1054.62 --> 1055.92] well it's gonna be +[1055.92 --> 1056.90] really hard for you to do +[1056.90 --> 1057.64] that project at your +[1057.64 --> 1058.36] current skill level +[1058.36 --> 1059.98] because there's like six +[1059.98 --> 1060.80] other factors that you +[1060.80 --> 1061.74] don't know much about +[1061.74 --> 1063.44] and you know I know +[1063.44 --> 1064.24] from experience that +[1064.24 --> 1064.72] they're going to be +[1064.72 --> 1065.68] very hard you know +[1065.68 --> 1066.14] like they'll want to +[1066.14 --> 1066.86] get data from it and +[1066.86 --> 1067.36] they'll assume there's +[1067.36 --> 1068.16] an API to get some +[1068.16 --> 1069.50] data and it's like no +[1069.50 --> 1070.02] that doesn't really +[1070.02 --> 1070.66] exist you're probably +[1070.66 --> 1071.26] have to like scrape +[1071.26 --> 1072.10] web pages and that's +[1072.10 --> 1072.60] going to be terribly +[1072.60 --> 1074.12] hard so go for sizes +[1074.12 --> 1075.20] was kind of me saying +[1075.20 --> 1076.08] you know if I was +[1076.08 --> 1076.82] starting over and I +[1076.82 --> 1077.42] just wanted some +[1077.42 --> 1078.52] random exercises to +[1078.52 --> 1079.28] build that weren't +[1079.28 --> 1080.36] completely boring but +[1080.36 --> 1081.48] would challenge me +[1081.48 --> 1082.40] would teach me to +[1082.40 --> 1084.02] read the docs to look +[1084.02 --> 1085.06] at different APIs in the +[1085.06 --> 1086.40] standard library to do +[1086.40 --> 1087.74] stuff like that what +[1087.74 --> 1088.44] would they look like +[1088.44 --> 1089.74] and I basically just sat +[1089.74 --> 1090.54] down and picked out +[1090.54 --> 1092.22] 20 projects that I +[1092.22 --> 1092.88] sort of picked them +[1092.88 --> 1093.66] intentionally to try +[1093.66 --> 1094.36] different stuff I +[1094.36 --> 1094.86] didn't want to keep +[1094.86 --> 1095.88] using the same thing so +[1095.88 --> 1096.40] I didn't want to build +[1096.40 --> 1097.46] like 20 web applications +[1097.46 --> 1099.00] but the idea was like +[1099.00 --> 1099.56] you know if you go +[1099.56 --> 1100.26] through all of these +[1100.26 --> 1101.12] and you code them all +[1101.12 --> 1101.72] and you actually give +[1101.72 --> 1102.48] it an honest shot of +[1102.48 --> 1103.18] trying to solve each +[1103.18 --> 1104.14] problem on your own +[1104.14 --> 1105.74] you will learn a ton in +[1105.74 --> 1106.94] the process even if you +[1106.94 --> 1107.68] don't complete them +[1107.68 --> 1108.92] fully or you don't +[1108.92 --> 1110.16] understand everything or +[1110.16 --> 1111.00] your code's pretty ugly +[1111.00 --> 1111.76] it doesn't matter you +[1111.76 --> 1112.48] will still learn a lot +[1112.48 --> 1112.92] doing it. +[1113.44 --> 1114.12] I think that that's so +[1114.12 --> 1115.26] valuable as well because +[1115.26 --> 1116.48] I tweeted about this +[1116.48 --> 1117.56] recently when I was +[1117.56 --> 1119.94] learning to code I was +[1119.94 --> 1120.74] already pretty well +[1120.74 --> 1122.20] established in tech on +[1122.20 --> 1123.96] Twitter and so like how +[1123.96 --> 1125.42] should I start just build +[1125.42 --> 1126.78] something build what +[1126.78 --> 1128.94] something what where +[1128.94 --> 1130.28] where do I start how do +[1130.28 --> 1131.86] I start what do I build +[1131.86 --> 1133.78] I don't know well you +[1133.78 --> 1134.62] know what figure it out +[1134.62 --> 1135.72] that's how you learn just +[1135.72 --> 1136.34] figure it out. +[1136.34 --> 1137.38] and that's really +[1137.38 --> 1137.90] frustrating. +[1138.18 --> 1139.50] So frustrating because +[1139.50 --> 1141.94] you just like build a to-do +[1141.94 --> 1144.62] app like tell me something +[1144.62 --> 1145.86] tell me where to start +[1145.86 --> 1148.04] right and so I feel like +[1148.04 --> 1149.24] really experienced +[1149.24 --> 1151.14] programmers that's kind of +[1151.14 --> 1152.46] their go-to advice we'll +[1152.46 --> 1153.26] just just go build +[1153.26 --> 1153.58] something. +[1154.16 --> 1155.60] Actually the other advice +[1155.60 --> 1156.42] that I'd add to that +[1156.42 --> 1157.28] though I mean if you've +[1157.28 --> 1158.70] got a project right build +[1158.70 --> 1160.24] it right but don't build +[1160.24 --> 1161.34] it just once build it +[1161.34 --> 1164.00] three times because the +[1164.00 --> 1164.94] first time you build it +[1164.94 --> 1165.46] right you're going to +[1165.46 --> 1166.54] commit all the sins and +[1166.54 --> 1167.12] you're going to build the +[1167.12 --> 1168.44] wrong thing right it's not +[1168.44 --> 1169.96] going to meet the needs +[1169.96 --> 1170.68] that you're looking to +[1170.68 --> 1172.16] address with it right it's +[1172.16 --> 1172.94] not going to work well +[1172.94 --> 1173.56] you're going to be like oh +[1173.56 --> 1174.38] that was that was +[1174.38 --> 1175.98] horrific right the second +[1175.98 --> 1177.24] time it will probably +[1177.24 --> 1178.86] function and achieve what +[1178.86 --> 1180.66] you want it to achieve in +[1180.66 --> 1181.62] terms of the end results +[1181.62 --> 1182.38] of the thing that you're +[1182.38 --> 1183.66] building but you're going +[1183.66 --> 1185.10] to think oh that is some +[1185.10 --> 1187.50] ugly awful evil code +[1187.50 --> 1189.42] there are monsters working +[1189.42 --> 1191.14] inside that that you hate +[1191.14 --> 1193.22] and the third time then +[1193.22 --> 1194.06] you're in a position where +[1194.06 --> 1195.54] you can actually put +[1195.54 --> 1196.78] together some really +[1196.78 --> 1199.48] elegant approaches and to +[1199.48 --> 1200.60] develop some beautiful +[1200.60 --> 1202.72] code and so if I were to +[1202.72 --> 1203.74] add any advice with that +[1203.74 --> 1204.80] once you do find that +[1204.80 --> 1205.86] something whatever that +[1205.86 --> 1207.42] is if that's reproducing +[1207.42 --> 1208.50] someone's example or +[1208.50 --> 1210.52] finding anything that +[1210.52 --> 1211.70] inspires you to build +[1211.70 --> 1213.04] something build it three +[1213.04 --> 1213.48] times. +[1214.34 --> 1216.06] Wow I love that advice. +[1216.26 --> 1217.94] Me too I heard once that +[1217.94 --> 1219.98] if you want to become a +[1219.98 --> 1222.46] great writer don't read a +[1222.46 --> 1224.28] hundred books just read one +[1224.28 --> 1226.58] book a hundred times and I +[1226.58 --> 1227.64] think this is kind of in the +[1227.64 --> 1229.30] same vein Dave in that this +[1229.30 --> 1230.38] kind of gives you the chance +[1230.38 --> 1231.42] to revisit a thing in +[1231.42 --> 1232.72] different stages which is +[1232.72 --> 1234.16] reality right in terms of +[1234.16 --> 1236.00] maintaining a software a +[1236.00 --> 1237.34] piece of software as it +[1237.34 --> 1239.42] ages as you age and an +[1239.42 --> 1240.60] increase in your skill set +[1240.60 --> 1242.48] etc etc so that's really +[1242.48 --> 1244.12] great I love that I'm going +[1244.12 --> 1244.72] to try that. +[1245.30 --> 1246.16] Just to add something to +[1246.16 --> 1247.28] that I know some people get +[1247.28 --> 1249.04] bored doing that or I should +[1249.04 --> 1249.94] say some people seem like +[1249.94 --> 1251.10] they do even if you don't +[1251.10 --> 1252.34] build the exact same thing I +[1252.34 --> 1253.68] think building similar things +[1253.68 --> 1255.48] is probably would go in line +[1255.48 --> 1256.58] with what Dave's saying +[1256.58 --> 1258.36] because I'll see people take +[1258.36 --> 1259.20] like a course and they'll +[1259.20 --> 1260.00] just build whatever's in the +[1260.00 --> 1260.82] course then they'll be done +[1260.82 --> 1262.58] and what I really like to +[1262.58 --> 1264.44] encourage is to like go back +[1264.44 --> 1265.34] through the course and build +[1265.34 --> 1266.50] something similar but not +[1266.50 --> 1267.68] quite the same so it forces +[1267.68 --> 1269.22] you to go out on your own and +[1269.22 --> 1270.40] sort of you know do what he's +[1270.40 --> 1271.42] saying to try different stuff +[1271.42 --> 1272.34] to do a little bit different +[1272.34 --> 1273.76] but you're still building you +[1273.76 --> 1274.88] know the same basic building +[1274.88 --> 1276.02] blocks you know like you're +[1276.02 --> 1277.06] building a web application or +[1277.06 --> 1278.72] you're building a CLI or +[1278.72 --> 1279.88] whatever but it might do +[1279.88 --> 1280.92] something slightly different +[1280.92 --> 1281.78] that forces you to really +[1281.78 --> 1282.74] think about what you're doing +[1282.74 --> 1283.60] and consider stuff. +[1284.10 --> 1285.22] Yes John and I think that +[1285.22 --> 1286.84] learning is repetition right +[1286.84 --> 1288.48] if you learn anything right +[1288.48 --> 1289.48] you're gonna you need to +[1289.48 --> 1291.10] repeat it in order to sort of +[1291.10 --> 1292.62] make it go to long-term +[1292.62 --> 1293.92] memory but I think that +[1293.92 --> 1295.20] that's also a great approach. +[1308.72 --> 1321.00] This episode is brought to you by +[1321.00 --> 1322.76] our friends at Fire Hydrant. +[1322.96 --> 1324.28] Fire Hydrant is the reliability +[1324.28 --> 1325.82] platform for every developer. +[1326.22 --> 1328.00] Incidents, they impact everyone +[1328.00 --> 1330.00] not just SREs. +[1330.16 --> 1331.50] They give teams the tools to +[1331.50 --> 1332.98] maintain service catalogs, +[1333.18 --> 1334.26] respond to incidents, +[1334.44 --> 1335.74] communicate through status pages, +[1335.74 --> 1337.90] and learn with retrospectives. +[1338.26 --> 1339.62] What would normally be manual +[1339.62 --> 1341.52] error prone tasks across the +[1341.52 --> 1343.08] entire spectrum are responding +[1343.08 --> 1343.66] to an incident. +[1343.96 --> 1345.50] They can all be automated in +[1345.50 --> 1347.16] every way with Fire Hydrant. +[1347.36 --> 1348.86] They have incident tooling to +[1348.86 --> 1350.72] manage incidents of any type with +[1350.72 --> 1352.50] any severity with consistency. +[1353.04 --> 1354.64] Declare and mitigate incidents +[1354.64 --> 1356.18] all from inside Slack. +[1356.58 --> 1357.74] Service catalogs allow service +[1357.74 --> 1359.12] owners to improve operational +[1359.12 --> 1361.28] maturity and document all your +[1361.28 --> 1362.90] deploys in your service catalog. +[1363.50 --> 1364.84] Incident analytics allow you to +[1364.84 --> 1366.10] extract meaningful insights +[1366.10 --> 1367.68] about your reliability over any +[1367.68 --> 1369.66] facet of your incident or the +[1369.66 --> 1370.86] people who respond to them. +[1371.24 --> 1372.06] And at the heart of it all +[1372.06 --> 1373.46] incident run books, they let you +[1373.46 --> 1375.16] create custom automation rules, +[1375.40 --> 1376.76] convert manual tasks into +[1376.76 --> 1378.80] automated, reliable, repeatable +[1378.80 --> 1380.64] sequences that run when you want. +[1381.00 --> 1381.98] You can create Slack channels, +[1382.12 --> 1383.30] Jira tickets, Zoom bridges +[1383.30 --> 1384.74] instantly after declaring an +[1384.74 --> 1385.02] incident. +[1385.50 --> 1386.50] Now your processes can be +[1386.50 --> 1388.08] consistent and automatic. +[1388.52 --> 1390.22] The next step is to try it free. +[1390.36 --> 1392.08] Small teams, up to 10 people, +[1392.22 --> 1393.54] can get started for free with all +[1393.54 --> 1394.74] FireHydrant features included. +[1395.08 --> 1396.46] No credit card is required. +[1396.92 --> 1398.98] Get started at FireHydrant.io. +[1399.36 --> 1401.38] Again, FireHydrant.io. +[1401.38 --> 1419.18] Well, I want to segue in to, +[1419.82 --> 1421.82] instead of maybe talking more about +[1421.82 --> 1424.60] how we learn and approaches to +[1424.60 --> 1426.76] learning, let's talk about +[1426.76 --> 1428.02] learning and go. +[1428.02 --> 1430.36] And so we kind of touched on that +[1430.36 --> 1432.08] about the starting point in +[1432.08 --> 1433.80] foundations and computer science +[1433.80 --> 1435.02] fundamentals in your course, +[1435.34 --> 1436.46] your Udemy course, Dave. +[1436.88 --> 1440.72] But how do you teach beginners go? +[1440.98 --> 1442.42] And we can talk about this either +[1442.42 --> 1445.18] from other languages in terms of +[1445.18 --> 1447.84] their experience or from scratch. +[1448.18 --> 1450.40] And what do you think is unique to +[1450.40 --> 1452.08] learning and go versus just +[1452.08 --> 1453.20] programming in general? +[1453.90 --> 1455.70] Well, I specifically at this point +[1455.70 --> 1457.46] specialize in starting from scratch. +[1457.46 --> 1460.48] And so I have my course that is +[1460.48 --> 1462.42] designed really to bring someone up +[1462.42 --> 1463.00] from scratch. +[1463.44 --> 1465.56] And one thing I did find is that +[1465.56 --> 1466.86] originally some people were having +[1466.86 --> 1468.20] some trouble with the technical +[1468.20 --> 1468.98] aspects of it. +[1469.04 --> 1470.00] They were having some trouble with +[1470.00 --> 1471.84] GoPath and GoRoot and some of those +[1471.84 --> 1472.16] things. +[1473.00 --> 1475.48] And originally I had created +[1475.48 --> 1476.98] Windows installation videos and +[1476.98 --> 1479.70] saying, hey, if you want, you can go +[1479.70 --> 1481.88] ahead and, you know, do the same +[1481.88 --> 1482.60] thing on Mac. +[1482.66 --> 1483.84] You can do the same thing on Linux. +[1483.84 --> 1486.26] And when I rebooted the course and +[1486.26 --> 1487.94] specialized more in helping beginners, +[1487.94 --> 1489.82] I said, the heck with that noise. +[1489.82 --> 1493.24] And I literally have approached it to +[1493.24 --> 1495.62] give them videos and follow through +[1495.62 --> 1499.70] instructions on Windows 10, on Mac OS, +[1499.80 --> 1502.18] as well as Ubuntu Linux, so that they +[1502.18 --> 1503.62] at least have something that looks, +[1503.80 --> 1506.12] tastes and smells similar to whatever +[1506.12 --> 1507.08] platform they have. +[1507.10 --> 1508.60] Because there were some people that +[1508.60 --> 1509.66] were experiencing problems. +[1509.66 --> 1512.16] And my intent with it is to reduce +[1512.16 --> 1513.82] the amount of friction at all +[1513.82 --> 1514.28] possible. +[1514.96 --> 1517.32] Once they achieve that, though, what +[1517.32 --> 1519.34] I find is that people are really +[1519.34 --> 1520.42] functional within it. +[1520.50 --> 1521.54] They get an understanding. +[1522.02 --> 1523.62] One of the beautiful things that is, +[1523.70 --> 1526.36] I think, unique about Go as well is +[1526.36 --> 1528.06] that if people want, they can get into +[1528.06 --> 1529.18] the Go code itself. +[1529.34 --> 1529.96] It's there. +[1530.08 --> 1531.08] It's in your workstation. +[1531.28 --> 1533.18] If you're interested in how, you know, +[1533.26 --> 1534.68] print line works and things of that +[1534.68 --> 1536.40] nature, you can sort of dive into it +[1536.40 --> 1538.66] and start to see these other elements +[1538.66 --> 1540.76] and how they come together in the +[1540.76 --> 1543.40] compiler and provide a, you know, +[1543.42 --> 1545.36] the end user experience of the +[1545.36 --> 1547.14] compilation and the code that gets +[1547.14 --> 1547.66] executed. +[1548.08 --> 1549.56] So really, I think, you know, it's +[1549.56 --> 1550.92] similar to other programming languages +[1550.92 --> 1553.58] in the sense that the fundamentals +[1553.58 --> 1554.28] are the same. +[1554.68 --> 1556.92] But it has, you know, great syntax. +[1557.30 --> 1558.24] It's really thoughtful. +[1558.58 --> 1560.52] It's really built for concurrency. +[1561.06 --> 1563.28] And I think it's really an effective +[1563.28 --> 1565.48] language for beginners because some of +[1565.48 --> 1566.90] those obscure elements in earlier +[1566.90 --> 1568.06] languages aren't there. +[1568.56 --> 1570.06] It's very quick to compile. +[1570.66 --> 1572.14] So I'm really an evangelist when it +[1572.14 --> 1572.78] comes to it. +[1573.04 --> 1574.50] But I did find that some of the +[1574.50 --> 1575.84] students that I was experiencing had +[1575.84 --> 1577.50] that little bit of friction at first. +[1577.88 --> 1580.10] And now if I can get them over that, +[1580.22 --> 1581.68] you know, then they're off to the +[1581.68 --> 1581.90] races. +[1582.14 --> 1584.46] It's usually a very early problem early +[1584.46 --> 1586.72] on or they're off to the races and +[1586.72 --> 1589.08] then asking more advanced items out +[1589.08 --> 1589.46] of the gate. +[1589.52 --> 1591.70] So there's this pendulum that swings in +[1591.70 --> 1593.72] their experience, I find, where they're +[1593.72 --> 1594.74] like, oh, that was awesome. +[1594.82 --> 1595.40] What's next? +[1595.40 --> 1595.84] Right. +[1596.26 --> 1596.92] In the beginning. +[1597.28 --> 1597.50] Right. +[1597.58 --> 1598.96] Well, you know, until recently. +[1599.28 --> 1600.16] Thanks, Go team. +[1600.80 --> 1603.22] Go path was a nightmare. +[1604.22 --> 1606.82] So once you got through that hurdle, +[1607.20 --> 1609.06] after you're like ready to throw your +[1609.06 --> 1611.04] computer and then you build something, +[1611.26 --> 1612.34] you're feeling better. +[1612.50 --> 1614.64] But I think that the hurdle of setting +[1614.64 --> 1616.50] up your Go path was such a nightmare +[1616.50 --> 1618.30] that people were just like, no, no. +[1618.36 --> 1620.18] If this is just set up, how's it going +[1620.18 --> 1620.48] to be? +[1620.84 --> 1621.50] Oh, absolutely. +[1621.50 --> 1623.44] And getting that to be set out of the +[1623.44 --> 1625.06] default was the most brewing thing that +[1625.06 --> 1625.44] could happen. +[1625.44 --> 1627.06] So I think that's one of the reasons +[1627.06 --> 1629.48] why, like, you see PHP being so popular +[1629.48 --> 1630.40] still to this day. +[1630.64 --> 1632.40] And I still think one of the big reasons +[1632.40 --> 1634.56] it was was because if you knew very +[1634.56 --> 1636.92] little, you could find a free PHP server +[1636.92 --> 1639.48] and just upload a file or your FTP in or +[1639.48 --> 1640.64] something and connect to and change a +[1640.64 --> 1642.02] file and you would see results like you +[1642.02 --> 1643.28] didn't have to understand or install +[1643.28 --> 1644.64] anything and you could get stuff working. +[1644.64 --> 1646.54] And I think like every new language +[1646.54 --> 1648.14] should strive for as much simplicity +[1648.14 --> 1648.82] as possible. +[1649.22 --> 1650.16] Like, that's one of the reasons why I +[1650.16 --> 1652.66] love the Go Playground, because I'm +[1652.66 --> 1654.02] glad they thought about like we need +[1654.02 --> 1655.36] some way for somebody to quickly and +[1655.36 --> 1657.42] easily just write some code and you +[1657.42 --> 1658.58] like, yeah, it's limited to the standard +[1658.58 --> 1659.84] library and some stuff like that. +[1659.90 --> 1661.58] But it's still a great learning tool +[1661.58 --> 1663.00] for somebody who, you know, like you +[1663.00 --> 1664.24] said, couldn't install it for some +[1664.24 --> 1666.12] reason or they just want to see it +[1666.12 --> 1667.58] before they actually spend that time. +[1667.74 --> 1669.42] Because I think if we don't focus on +[1669.42 --> 1671.22] that some there's languages like +[1671.22 --> 1672.78] JavaScript where, you know, you can +[1672.78 --> 1674.76] have interactive tutorials really, +[1674.90 --> 1675.88] really easily in JavaScript. +[1676.62 --> 1677.74] And, you know, somebody can just +[1677.74 --> 1678.94] bring up Chrome and open up Chrome +[1678.94 --> 1680.50] DevTools and write some JavaScript. +[1680.96 --> 1682.48] So like, you know, the barrier to +[1682.48 --> 1684.44] entry there is so small that I think +[1684.44 --> 1685.54] it's important for other languages +[1685.54 --> 1686.50] to keep that in mind. +[1687.16 --> 1688.98] Yeah, this is a good thing that I +[1688.98 --> 1690.50] kind of was wondering, like how much +[1690.50 --> 1692.12] is set up a part of learning Go +[1692.12 --> 1693.02] versus other languages? +[1693.02 --> 1694.58] And you mentioned PHP and Java. +[1694.82 --> 1697.38] And if it's different, what are some +[1697.38 --> 1700.32] of the setup gotchas in addition or +[1700.32 --> 1701.42] you think are going to be the setup +[1701.42 --> 1703.72] gotchas or the learning gotchas? +[1704.16 --> 1705.54] And maybe we can ask for audience +[1705.54 --> 1707.84] participation on the GoTime FM +[1707.84 --> 1708.50] Slack channel. +[1708.98 --> 1710.26] Well, some of these show offs are +[1710.26 --> 1712.54] saying that GoPath made perfect +[1712.54 --> 1713.36] sense to them. +[1713.90 --> 1715.78] So good job, guys. +[1715.98 --> 1716.64] Good job. +[1717.06 --> 1718.70] I think the problem with it was like +[1718.70 --> 1720.16] it either made sense to you or it +[1720.16 --> 1721.06] made no sense to you. +[1721.16 --> 1722.12] There was no middle ground. +[1722.80 --> 1723.24] Yes. +[1723.58 --> 1725.04] And it was also very platform +[1725.04 --> 1726.40] specific as well, right? +[1726.40 --> 1727.28] Because at that point when you're +[1727.28 --> 1728.54] hooking into environment variables +[1728.54 --> 1730.50] or whatever your platform is, +[1730.52 --> 1731.80] you can get stuck in the details +[1731.80 --> 1733.00] of your particular platform. +[1733.66 --> 1735.14] Yeah, especially because a lot of +[1735.14 --> 1736.26] beginners are coming from Windows +[1736.26 --> 1737.72] and like setting environment +[1737.72 --> 1739.24] variables has always seemed easier +[1739.24 --> 1739.58] to me. +[1739.68 --> 1741.22] And, you know, in Mac or Linux, +[1741.40 --> 1743.04] whereas in Windows, trying to get +[1743.04 --> 1744.14] somebody to go set that stuff up +[1744.14 --> 1745.72] correctly was sometimes a pain. +[1746.18 --> 1746.50] Yeah. +[1746.66 --> 1747.46] Super pain. +[1747.72 --> 1750.38] We used to, me and Steve Francia, +[1750.74 --> 1753.24] who, hi, Steve, used to teach +[1753.24 --> 1754.04] workshops. +[1754.32 --> 1755.72] And the first thing that we would do +[1755.72 --> 1756.94] is raise your hand if you use +[1756.94 --> 1757.72] a Windows machine. +[1758.08 --> 1758.44] Great. +[1758.98 --> 1760.42] You guys are now friends +[1760.42 --> 1762.50] because you're going to need +[1762.50 --> 1763.78] to help each other during setup +[1763.78 --> 1764.96] because we cannot help you. +[1765.58 --> 1765.90] Mm-hmm. +[1766.14 --> 1767.86] Yeah, you taught, and I have that +[1767.86 --> 1768.80] material, and I thought it was +[1768.80 --> 1770.96] amazing in terms of CLI. +[1771.32 --> 1772.70] So CLI workshop, and I think it +[1772.70 --> 1774.28] was for OSCON last year or the +[1774.28 --> 1774.86] year before. +[1775.60 --> 1777.20] And it has like, the slide deck +[1777.20 --> 1778.80] is something like 300-some slides. +[1779.36 --> 1781.14] And setup, Ashley, was like the +[1781.14 --> 1782.96] first third of that, right? +[1782.96 --> 1784.96] And so I just wonder, like, is that +[1784.96 --> 1786.68] the hump that we just want to help +[1786.68 --> 1787.96] beginners get over, and then they'll +[1787.96 --> 1789.16] be able to get really productive +[1789.16 --> 1790.08] soon after that? +[1790.52 --> 1793.26] It really was the first half of +[1793.26 --> 1794.16] the workshop. +[1794.76 --> 1796.76] Setup was difficult for a lot of +[1796.76 --> 1797.06] people. +[1797.20 --> 1798.74] There was a lot of going around and +[1798.74 --> 1800.20] helping people get their machines +[1800.20 --> 1800.68] set up. +[1800.98 --> 1804.50] Once we got into building the app, +[1804.78 --> 1806.94] things seemed to flow much, much +[1806.94 --> 1807.40] easier. +[1807.82 --> 1810.32] So yeah, setup was definitely a big +[1810.32 --> 1810.64] issue. +[1810.64 --> 1812.42] I think setup's also like the +[1812.42 --> 1813.50] biggest quitting point, too. +[1813.94 --> 1814.12] Yeah. +[1814.32 --> 1815.64] Like, at some point during setup, if +[1815.64 --> 1816.76] it stops working, you quit. +[1816.92 --> 1818.88] But like, if you're writing code and +[1818.88 --> 1820.44] you've got most things running, I +[1820.44 --> 1821.68] think you're less likely to quit at +[1821.68 --> 1822.06] that point. +[1822.38 --> 1824.18] And is this unique to Go, or are we +[1824.18 --> 1826.62] just, I want to kind of give a nod to +[1826.62 --> 1828.38] any beginner out there, whether it's +[1828.38 --> 1829.48] somebody who's already learned +[1829.48 --> 1830.92] something else in terms of the +[1830.92 --> 1831.86] programming language, or someone +[1831.86 --> 1834.48] who's a complete beginner, is set up +[1834.48 --> 1836.34] a quitting point for all languages, or +[1836.34 --> 1838.14] is it a little bit more painful in Go? +[1838.44 --> 1839.02] I think it's all. +[1839.02 --> 1840.72] I think every language has some, like, +[1840.78 --> 1842.46] barrier to entry, and once you get it, +[1842.56 --> 1842.92] you're good. +[1843.04 --> 1844.62] But up until that point, it can be +[1844.62 --> 1845.02] frustrating. +[1845.70 --> 1846.94] I do think it also depends whether +[1846.94 --> 1848.04] you're talking about an interpreted +[1848.04 --> 1849.92] language or a compiled language as +[1849.92 --> 1850.50] well, right? +[1850.74 --> 1852.44] Because it is a little, you know, hard +[1852.44 --> 1853.98] to compare Go to an interpreted +[1853.98 --> 1856.20] language, like the like of PHP, +[1856.40 --> 1857.82] perhaps, in the way that it's +[1857.82 --> 1859.20] structured and some of those elements +[1859.20 --> 1860.46] that it actually achieves underneath. +[1860.90 --> 1863.10] But one other element I think comes +[1863.10 --> 1864.74] into play is, you know, what do you +[1864.74 --> 1866.56] then, once you have Go set up, what do +[1866.56 --> 1868.60] you hook in after that, right? +[1868.68 --> 1870.12] Is it an IDE? +[1870.44 --> 1871.78] Do you have plans for an IDE? +[1872.18 --> 1873.72] Because there's some really outstanding +[1873.72 --> 1874.98] things you can do afterwards. +[1875.12 --> 1877.18] And yeah, I show, just like everyone +[1877.18 --> 1878.32] else does, how do you make a Hello +[1878.32 --> 1879.98] Word program on the command line? +[1880.48 --> 1882.62] But then if you can actually trace and +[1882.62 --> 1884.48] set up breakpoints and have an IDE +[1884.48 --> 1887.46] experience after that, that will help +[1887.46 --> 1889.60] guide the student, then they can trace +[1889.60 --> 1891.32] their way through the code, which again, +[1891.32 --> 1892.72] will help them understand what's actually +[1892.72 --> 1894.06] happening underneath the covers. +[1894.06 --> 1896.08] So, you know, there's that initial setup, +[1896.22 --> 1898.40] but then also, and that's the kind of +[1898.40 --> 1900.48] thing where, ask anyone, they have +[1900.48 --> 1902.22] their favorite, is it Atom, is it +[1902.22 --> 1903.94] Sublime, is it Visual Studio Code? +[1904.44 --> 1906.30] Everyone has their favorite sort of +[1906.30 --> 1909.14] tools to then add to the language, +[1909.32 --> 1910.94] whatever language they're programming +[1910.94 --> 1911.98] on at that point, right? +[1912.30 --> 1912.50] Yeah. +[1912.64 --> 1914.34] And then what do you think about in +[1914.34 --> 1917.26] terms of content or setup, in terms +[1917.26 --> 1918.90] of audience competence? +[1919.44 --> 1921.54] We have a comment here from Corey Lanou +[1921.54 --> 1924.06] in the GoTime FM Slack, who said that +[1924.06 --> 1925.64] GoInstall was the easiest that he's +[1925.64 --> 1928.14] ever used, but again, it was geared +[1928.14 --> 1929.62] towards me as an audience. +[1930.04 --> 1932.10] And so that we're saying somebody who's +[1932.10 --> 1933.70] an experienced beginner. +[1934.18 --> 1936.50] So is there content, do you feel like +[1936.50 --> 1938.42] for the people who have sort of looked +[1938.42 --> 1940.48] through content, created content in this +[1940.48 --> 1942.76] panel, that there are audiences for whom +[1942.76 --> 1944.66] there are content gaps for beginners? +[1945.04 --> 1946.24] Yeah, I do think. +[1946.24 --> 1948.70] So when people ask me what language they +[1948.70 --> 1951.02] should start with, I usually say Python, +[1951.52 --> 1953.68] because there's so much information out +[1953.68 --> 1955.08] there, and I felt like setup was easier +[1955.08 --> 1955.84] for some reason. +[1956.30 --> 1959.22] For Go, I feel like there's a lot of highly +[1959.22 --> 1961.26] technical things out there, which is great. +[1961.62 --> 1964.02] And then there are some beginner courses +[1964.02 --> 1966.26] that may or may not work. +[1966.40 --> 1967.94] I'm excited to try your guys' +[1968.18 --> 1971.94] So I just felt like beginner materials were +[1971.94 --> 1973.20] super lacking. +[1973.20 --> 1975.78] For me, when I learn a new programming +[1975.78 --> 1978.08] language, I am not ashamed by this. +[1978.24 --> 1979.12] I don't care if you laugh. +[1979.56 --> 1980.78] I buy children's books. +[1981.06 --> 1982.90] I love kids' programming books. +[1983.10 --> 1984.10] They are the greatest. +[1984.58 --> 1986.50] There needs to be one for Go yesterday. +[1987.42 --> 1988.56] Oh my God, yes. +[1988.96 --> 1991.88] Well, you know, there was this trend that +[1991.88 --> 1993.08] was, I don't know, it's probably about +[1993.08 --> 1993.82] eight years old now. +[1993.92 --> 1995.36] Do you remember ELI 5? +[1995.50 --> 1996.74] Explain It Like I'm 5? +[1996.90 --> 1997.22] Yes. +[1997.34 --> 1999.18] And it was like a popular, yeah, those +[1999.18 --> 2001.70] were great because it was, you know, we +[2001.70 --> 2004.06] had someone in the chat earlier said, well, +[2004.10 --> 2006.16] I need to learn the thing to get to the thing +[2006.16 --> 2008.04] and learn the thing before that thing. +[2008.50 --> 2011.84] And it's just a recursive, but I don't know +[2011.84 --> 2012.88] this all the way down. +[2012.98 --> 2015.38] And so I'm hoping your gentle beginners course +[2015.38 --> 2016.94] tackles exactly that. +[2017.16 --> 2018.20] Well, I appreciate that. +[2018.24 --> 2019.96] And I'll tell you, you know, you're almost +[2019.96 --> 2021.26] throwing the gauntlet down because I have +[2021.26 --> 2022.32] twin 11-year-olds. +[2022.86 --> 2024.48] And while they have tasted a little bit of +[2024.48 --> 2027.02] Python and a lot of scratch, I'm thinking +[2027.02 --> 2029.04] I should be the, you know, maybe the first to +[2029.04 --> 2031.76] write a Golang children's book at this rate. +[2031.96 --> 2032.92] So we'll see. +[2033.00 --> 2033.74] I'll give it some thought. +[2034.42 --> 2035.66] Please do. +[2035.84 --> 2036.14] Yes. +[2036.24 --> 2037.58] I hope you have a good artist because I feel +[2037.58 --> 2038.78] like that'd be my biggest limitation. +[2038.92 --> 2040.12] If I was drawing it, it would look like a +[2040.12 --> 2041.20] five-year-old drew the whole thing. +[2041.50 --> 2044.54] Oh, I am a coder, you know, coder art scheme. +[2044.66 --> 2046.68] I'd be, you know, full up on the red, green +[2046.68 --> 2047.76] and blue colors. +[2047.88 --> 2048.86] It would look horrific. +[2049.90 --> 2053.70] So yeah, I need to, any artists shout it out +[2053.70 --> 2056.20] and I could probably use some folks to +[2056.20 --> 2056.84] collaborate with. +[2056.90 --> 2057.44] There's no doubt. +[2057.44 --> 2058.20] I'm here. +[2058.32 --> 2059.20] I'm taking note. +[2059.94 --> 2062.26] There may or may not be a preeminent artist +[2062.26 --> 2064.50] for the Golang community speaking to us +[2064.50 --> 2065.84] at this moment right now. +[2065.96 --> 2066.20] Maybe. +[2066.40 --> 2066.80] I don't know. +[2066.88 --> 2067.48] What do you think? +[2068.78 --> 2069.22] Yeah. +[2069.58 --> 2071.72] Ashley's gophers are, there's how many? +[2071.90 --> 2074.28] Like in your automated, you know, like +[2074.28 --> 2075.60] billions in the gopher eyes. +[2075.66 --> 2077.10] Oh, in gopher eyes me. +[2077.18 --> 2078.12] Yeah, there are billions. +[2078.38 --> 2081.32] I don't know how many combinations there +[2081.32 --> 2081.98] are right now. +[2082.04 --> 2083.98] I think I might have added some things, but +[2083.98 --> 2084.56] billions. +[2084.56 --> 2087.16] And I really like whoever did your artwork +[2087.16 --> 2089.34] on your gopher sizes, John. +[2089.38 --> 2091.38] I think it looks like Ashley-esque, but it could +[2091.38 --> 2092.10] be someone else. +[2092.22 --> 2092.74] Yeah, it's not me. +[2092.84 --> 2096.34] I think that one was Marcus Olson is who I want +[2096.34 --> 2096.60] to say. +[2096.76 --> 2097.88] Yes, it was. +[2098.06 --> 2098.36] Oh. +[2098.36 --> 2098.84] Yes. +[2099.26 --> 2100.50] I was going to say Ashley's who introduced me +[2100.50 --> 2102.56] to him, but I've actually gotten into the habit +[2102.56 --> 2104.64] of, so Ashley's usually really, really swamped. +[2105.04 --> 2106.92] So I've gotten into the habit of trying to find +[2106.92 --> 2109.68] different artists for all of the courses because +[2109.68 --> 2111.36] I like them to have a slightly unique feel. +[2111.86 --> 2114.52] So like the testing course was Egon Elbrey, and I +[2114.52 --> 2117.30] have an algorithms course that I'm working on that +[2117.30 --> 2118.94] is another gopher in the community. +[2119.36 --> 2120.38] I think it's Adrian. +[2120.62 --> 2122.18] I'm drawing a blank on his last name though. +[2122.36 --> 2124.18] And then I have another artist who's working on +[2124.18 --> 2125.96] another course that's way down the pipeline. +[2125.96 --> 2127.86] So like I'm trying to like use different artists +[2127.86 --> 2129.48] for all of them because I think it's cool to have +[2129.48 --> 2130.78] different people drawing gophers. +[2131.34 --> 2133.68] And I know that Ashley's just overwhelmed with +[2133.68 --> 2133.88] stuff. +[2133.98 --> 2135.82] So it's like, I don't want to like keep bugging +[2135.82 --> 2136.06] her. +[2136.32 --> 2138.96] But it just goes to show that I think when you are +[2138.96 --> 2142.08] making content, whether that is for beginners or +[2142.08 --> 2145.64] for anyone, you know, we are not a black and white +[2145.64 --> 2146.68] or gray world, right? +[2146.70 --> 2149.20] We are a world of color and we like visuals and we +[2149.20 --> 2152.22] like fun and we want, we learn better when we feel +[2152.22 --> 2153.60] we're at play, right? +[2153.60 --> 2156.30] So I remember Richard Feynman talking about how he +[2156.30 --> 2159.28] bottled burnout and won a Nobel prize for his +[2159.28 --> 2160.16] Feynman techniques. +[2160.16 --> 2161.96] We're kind of veering into physics, but still +[2161.96 --> 2162.84] learning for beginners. +[2162.84 --> 2164.98] And he always said, just keep a sense of play +[2164.98 --> 2165.40] about you. +[2165.44 --> 2168.08] And Ashley, that's what I love about the gophers +[2168.08 --> 2170.14] that you bring to the table and to the community. +[2170.46 --> 2172.96] And so anyway, Jared Santa, who's also part of the +[2172.96 --> 2175.84] changelog crew said that they do an ELI five and +[2175.84 --> 2178.66] explain it like you're five on the sister podcast +[2178.66 --> 2179.30] JS party. +[2179.58 --> 2181.54] That would be a great segment for go time. +[2181.54 --> 2183.82] So if any of you are down to come back and do a +[2183.82 --> 2186.72] repeat, but for four or five year olds by our inner +[2186.72 --> 2189.08] five year olds, I'd love to do that episode. +[2189.32 --> 2189.88] That'd be awesome. +[2190.14 --> 2190.42] Yeah. +[2190.48 --> 2190.70] Yeah. +[2190.70 --> 2190.98] Same. +[2191.04 --> 2191.80] I'm here for it. +[2197.50 --> 2200.52] This episode is brought to you by our friends at +[2200.52 --> 2203.60] Acuity, a new platform that brings fully managed Argo +[2203.60 --> 2206.96] CD and enterprise services to the cloud or on premise. +[2206.96 --> 2209.26] And I'm here with two of the co-founders from Acuity, +[2209.56 --> 2212.14] Jesse Suen and Alexander Matrusenchev. +[2212.40 --> 2214.96] So the Acuity platform is in beta right now. +[2215.18 --> 2217.50] You guys have some big ideas you're executing on around +[2217.50 --> 2220.68] Argo CD, managed Argo CD, Kubernetes native application +[2220.68 --> 2222.58] delivery and the power of GitOps. +[2222.64 --> 2224.98] Help me understand the what and the why of what you're +[2224.98 --> 2225.50] doing right now. +[2225.50 --> 2230.10] So we started Acuity because we saw what was happening in the +[2230.10 --> 2232.42] Kubernetes community, the challenges that people were +[2232.42 --> 2234.86] facing about developer experience. +[2235.26 --> 2238.98] And having run Argo CD for Intuit for a couple of years, we +[2238.98 --> 2241.90] knew it took like a small team to build this and scale it and +[2241.90 --> 2244.96] provide a performant solution for the developers. +[2245.46 --> 2248.52] And so at Acuity and the Acuity platform, what we're trying to +[2248.52 --> 2251.38] do is the first thing we're trying to do is actually provide +[2251.38 --> 2255.18] Argo CD as a fully managed solution to our users. +[2255.18 --> 2257.72] But that is just actually the start of things. +[2257.82 --> 2262.70] And we actually want to take the next steps on improving the whole +[2262.70 --> 2265.86] GitOps and developer experience and providing new tools and +[2265.86 --> 2268.66] ecosystems around Argo and the Argo project. +[2269.00 --> 2269.80] Yeah, that's right, Jesse. +[2269.98 --> 2273.96] So Argo CD is just the beginning, but every company eventually +[2273.96 --> 2277.40] needs way more tools integrated into the DevOps platform. +[2277.78 --> 2280.54] And that's what we're hoping to deliver with Acuity platform. +[2281.02 --> 2284.26] So we're hoping to provide a great user interface that enable +[2284.26 --> 2288.00] developers to achieve what they need in a matter of just a few clicks. +[2288.48 --> 2291.34] But we also want to make Argo CD enterprise ready. +[2291.82 --> 2296.80] What that means is our customers will get audits and insightful +[2296.80 --> 2300.36] analytics out of the box without configuring anything. +[2300.86 --> 2302.46] That's what we did at Intuit. +[2302.58 --> 2304.40] And we learned that it was not so easy to do. +[2304.78 --> 2307.60] And that's what we're hoping to solve for multiple organizations. +[2308.12 --> 2308.38] Very cool. +[2308.46 --> 2308.98] Thank you, Jesse. +[2309.14 --> 2309.84] Thank you, Alex. +[2309.84 --> 2313.02] Again, listeners, this is a closed beta. +[2313.30 --> 2313.96] Check it out. +[2314.06 --> 2316.76] Acuity.io slash changelog. +[2316.82 --> 2319.72] Head there and see what this platform is all about. +[2320.04 --> 2322.38] Again, Acuity.io slash changelog. +[2322.50 --> 2324.10] Links are in the show notes. +[2324.56 --> 2327.48] And by Honeycomb, find your most perplexing application issues. +[2327.78 --> 2333.28] Honeycomb is a fast analysis tool that reveals the truth about every aspect of +[2333.28 --> 2334.70] your application in production. +[2335.18 --> 2339.16] Find out how users experience your code in complex and unpredictable environments. +[2339.16 --> 2343.90] Find patterns and outliers across billions of rows of data and definitively solve your +[2343.90 --> 2344.36] problems. +[2344.82 --> 2346.28] And we use Honeycomb here at Change. +[2346.32 --> 2350.14] Well, that's why we welcome the opportunity to add them as one of our infrastructure partners. +[2350.66 --> 2354.84] In particular, we use Honeycomb to track down CDN issues recently, which we talked about +[2354.84 --> 2357.96] at length on the Kaizen edition of the Ship It podcast. +[2358.20 --> 2358.90] So check that out. +[2359.16 --> 2359.62] Here's the thing. +[2359.88 --> 2363.12] Teams who don't use Honeycomb are forced to find the needle in the haystack. +[2363.24 --> 2366.40] They scroll through endless dashboards playing whack-a-mole. +[2366.40 --> 2371.18] They deal with alert floods, trying to guess which one matters, and they go from tool to +[2371.18 --> 2375.26] tool to tool playing sleuth, trying to figure out how all the puzzle pieces fit together. +[2375.62 --> 2379.94] It's this context switching and tool sprawl that are slowly killing teams' effectiveness +[2379.94 --> 2381.92] and ultimately hindering their business. +[2382.32 --> 2388.36] With Honeycomb, you get a fast, unified, and clear understanding of the one thing driving +[2388.36 --> 2389.08] your business. +[2389.32 --> 2389.76] Production. +[2390.26 --> 2392.76] With Honeycomb, you guess less and you know more. +[2392.76 --> 2398.36] Join the swarm and try Honeycomb free today at honeycomb.io slash changelog. +[2398.48 --> 2401.98] Again, honeycomb.io slash changelog. +[2401.98 --> 2424.20] I kind of want to talk a little bit about learning mediums. +[2424.20 --> 2429.34] So each of you has created or taught content in this new world of online. +[2429.34 --> 2436.02] And so I would love to talk about your perspectives about pros and cons about each of these mediums. +[2436.08 --> 2440.66] And so, John, we're talking about your gopher sizes with code accompanied by videos. +[2441.06 --> 2442.70] Dave, this would be your online course. +[2443.22 --> 2445.78] Workshops that you've given at events and conferences, Ashley. +[2445.96 --> 2449.46] So what are some of the pros and cons about each of these mediums for beginners? +[2449.86 --> 2451.46] I guess I can start with some of them. +[2451.76 --> 2453.62] I like videos because you can show mistakes. +[2453.62 --> 2459.26] I think it's important, especially for beginners, to see that even experienced developers make mistakes, +[2459.26 --> 2462.80] but also to see how you get to derived code. +[2462.98 --> 2465.94] Because I think there's a lot of times where we'll just show them the final code. +[2466.58 --> 2470.42] And as developers, we know that there might be three refactors that got us there. +[2471.06 --> 2475.50] And I think having a video makes it possible to do that versus if you're doing a book or +[2475.50 --> 2478.94] something like that, it's much, much harder because to actually show them, well, I went to +[2478.94 --> 2480.32] this line and changed this one thing. +[2480.32 --> 2482.36] And then I went to this other file and changed this one thing. +[2482.46 --> 2483.46] Like, it becomes a lot. +[2483.92 --> 2487.28] So I think videos are really good for that sort of interactive or not quite interactive, +[2487.42 --> 2488.66] but, you know, something like that. +[2488.96 --> 2494.12] But I've also started to find that books are probably more accessible, which is something +[2494.12 --> 2498.94] that I hate about videos is that you almost need to find translators for a couple different +[2498.94 --> 2503.22] languages or you need to get somebody to come in and actually, like, write all the transcripts +[2503.22 --> 2505.60] up because anything automated just doesn't do a good enough job. +[2505.60 --> 2509.52] And so basically, that's something I've been struggling with lately is trying to figure +[2509.52 --> 2513.36] out the right approach to that, because I think that making videos more accessible is +[2513.36 --> 2515.34] something that needs to happen in the future. +[2515.88 --> 2520.24] So one of the things that I think is important, because I recognize that my courses have reached +[2520.24 --> 2525.00] 160 something different countries is having good closed captioning. +[2525.02 --> 2527.62] And I have to admit right now, my GoLine course doesn't have it yet. +[2527.62 --> 2533.32] But to second John's thoughts around closed captioning, because some people, you know, +[2533.36 --> 2537.50] they're coming off of, you know, where English may not be their native language. +[2537.62 --> 2538.96] They speak several different languages. +[2538.96 --> 2542.50] So having something there is extremely useful for them. +[2542.70 --> 2546.32] But the challenge is, is that the automated closed captioning just doesn't work. +[2546.40 --> 2551.90] I don't know how many times I've said Udemy and it gets translated to you and me on the +[2551.90 --> 2555.02] Udemy platform itself or something absolutely crazy. +[2555.02 --> 2559.40] And I'm thinking if there's any word that Udemy would have right in their closed captioning, +[2559.44 --> 2560.50] it would be Udemy, right? +[2560.76 --> 2563.74] So that is a huge technical challenge and hurdle. +[2563.98 --> 2568.84] But the other thing, you know, because I did have a background in teaching outside of this, +[2568.94 --> 2574.28] where I've done some workshops in person in advance of this type of experience, as well +[2574.28 --> 2581.30] as having tutored one on one, you lose that interaction when you have an online video. +[2581.30 --> 2587.72] And it is extremely difficult to iterate and make changes and there's substitutes for it. +[2587.80 --> 2590.20] But I like to see eyes, right? +[2590.26 --> 2596.10] And even just hello and putting content out there into the world is a very different platform +[2596.10 --> 2599.28] when you're doing a static video than when you're recording. +[2599.38 --> 2604.12] I think ideally what I'd almost prefer to do as I grow and continue to make new courses +[2604.12 --> 2607.38] is to teach in a live event, record those. +[2607.38 --> 2610.02] So at least I've got somebody else I'm talking to. +[2610.16 --> 2615.30] And if you see that deer in headlights look, you get that sense of, oh, okay, I've fallen +[2615.30 --> 2616.02] off track here. +[2616.18 --> 2620.52] I need some more explanation here because it was a very different approach. +[2621.02 --> 2624.10] Mind you, you get unlimited redos when you're recording video, right? +[2624.14 --> 2627.32] So it's, oh, that was, I needed more coffee or something, right? +[2627.32 --> 2632.76] So there's pros and cons of each, but it's a very, everyone has its own flavor and piece +[2632.76 --> 2633.24] to it, right? +[2633.24 --> 2637.48] I feel like every way that we do this is valuable. +[2637.90 --> 2640.02] We all have different learning styles. +[2640.20 --> 2643.88] For me, it's hard for me to consume the content on video. +[2644.06 --> 2648.16] I'm like, okay, now I have to pause the video and do this step, play the video again. +[2648.30 --> 2649.22] And then I go back. +[2649.26 --> 2650.56] I'm like, did I do that step right? +[2650.66 --> 2653.10] So for me, I like step-by-step instructions. +[2653.74 --> 2658.82] And so the workshops are really helpful, but the cons of a workshop is that a lot of people +[2658.82 --> 2660.64] are afraid to ask questions. +[2660.74 --> 2665.14] So they will sit there and act like they know what's going on when they don't know what's +[2665.14 --> 2665.66] going on. +[2665.66 --> 2668.94] So we have to constantly go around and be like, do you really understand? +[2669.26 --> 2670.34] You can ask questions. +[2670.50 --> 2671.18] It's fine. +[2672.00 --> 2678.34] So video, written tutorials, workshops, they're all valuable in their own way. +[2678.40 --> 2679.38] We all learn differently. +[2680.16 --> 2684.26] Just to chime in one additional, like one of the reasons I have never focused on workshops, +[2684.38 --> 2685.78] conferences, that sort of stuff as much. +[2685.78 --> 2689.02] And I think it's easy to forget when you live in a city, but if there's a lot of people +[2689.02 --> 2693.60] who live nowhere near a city or they don't have the resources to pay for a workshop or +[2693.60 --> 2697.82] something, like to give you an example, I live two hours away from the nearest city and that's +[2697.82 --> 2699.76] Pittsburgh, which is not exactly a massive city. +[2700.26 --> 2703.80] So, you know, just knowing that there's a lot of people out there who cannot go to that +[2703.80 --> 2704.34] type of thing. +[2704.44 --> 2708.04] I definitely think that there's a good, like you said, we need almost everything because +[2708.04 --> 2711.60] some people are going to do better in the workshops and some people will meet more people +[2711.60 --> 2713.16] and they'll actually collaborate with them afterwards. +[2713.16 --> 2715.62] Other people will do online courses. +[2716.08 --> 2719.76] I think one of the big things I've just noticed is that people find other people to collaborate +[2719.76 --> 2720.76] with and to learn with. +[2720.84 --> 2722.10] That's very, very huge. +[2722.98 --> 2730.04] So like collaboration in person, looking or gauging for like deer in the headlights, course +[2730.04 --> 2730.50] correcting. +[2731.08 --> 2733.48] That's typically been the traditional way of learning, right? +[2733.52 --> 2737.76] But it doesn't scale and doesn't scale the teacher and it's not accessible to, you know, +[2737.82 --> 2739.54] rural and suburban learners. +[2740.00 --> 2740.96] I think it can scale. +[2740.96 --> 2742.08] It just scales differently. +[2742.46 --> 2746.18] Like one way I've seen it scale is I've seen people who do online courses and they'll +[2746.18 --> 2750.02] essentially have, they call them classes or whatever, but essentially a bunch of people +[2750.02 --> 2755.50] sign up and they say like, all right, we're going to start in maybe December and every +[2755.50 --> 2758.58] week you're supposed to go through so much content and then you're all like in a Slack +[2758.58 --> 2761.74] channel or something and you discuss that content and you're sort of expected to keep +[2761.74 --> 2763.44] up with the classroom on that content. +[2763.70 --> 2767.86] It's less acceptable in the sense that you can't just start whenever you want. +[2767.86 --> 2771.64] But I think there are ways to sort of get that simulated classroom environment. +[2772.16 --> 2774.94] And I think that's, you know, as course creators, it's things we have to think about. +[2775.00 --> 2777.02] Like, is there a way that we can make that possible? +[2777.32 --> 2779.76] I think tools like Slack have made it much, much better too. +[2780.28 --> 2783.92] Like I've gotten into this habit of like all the courses I have, I provide support for, +[2784.34 --> 2786.40] but I've gotten into this habit of any of the paid courses. +[2786.40 --> 2790.36] If you, you got access to a Slack and I basically say, ask your question there. +[2790.36 --> 2794.16] Even if you email it to me and I can answer in five seconds, I say, ask the question there +[2794.16 --> 2797.72] because this is the best way to foster this growth between different people taking the +[2797.72 --> 2798.06] courses. +[2798.06 --> 2800.64] Like you each try to answer the questions and help each other out. +[2800.76 --> 2802.80] And that will, you know, establish that learning. +[2802.90 --> 2804.68] It'll help you like get used to helping each other. +[2804.96 --> 2808.92] And over time, it's gotten to this point where I can sometimes check the Slack and somebody +[2808.92 --> 2811.52] will have asked a question and somebody will have answered it better than I could have before +[2811.52 --> 2812.30] I even got to it. +[2812.66 --> 2813.34] Oh, that's great. +[2813.34 --> 2814.30] Yeah, that's always great. +[2814.38 --> 2817.98] The feedback loops and in-person for me is where I can get unblocked. +[2818.02 --> 2823.16] And I know that the other exercise site that I've tried is exorcism.io for not just Go, +[2823.22 --> 2824.28] but a variety of languages. +[2824.84 --> 2826.72] And I was a mentor for Go. +[2826.94 --> 2832.22] And so when I signed on to be a mentor, they were like, listen, if you agree to sort of walk +[2832.22 --> 2837.90] through the code exercise, we all kind of as a community try to get the next person in +[2837.90 --> 2840.38] the queue so that their feedback is as quick as possible. +[2840.38 --> 2845.62] So there was that like online mentorship and trying to scale that differently, as you said. +[2846.06 --> 2847.94] So I love exorcism.io. +[2848.46 --> 2850.22] Thanks, Katrina and team. +[2850.98 --> 2855.84] Also, I feel like I need to mention it because it's great and we haven't talked about it yet, +[2855.84 --> 2857.82] but just for funk, I love it. +[2858.10 --> 2858.30] Yeah. +[2858.76 --> 2863.70] And I just wanted to ask John before I kind of talk about other possible resources for +[2863.70 --> 2864.24] gophers. +[2864.24 --> 2868.56] Can you tell us a little bit more about where learners could go for gophersizes? +[2868.80 --> 2873.88] Are they going to get that content for asking questions within the course as they sign on? +[2874.38 --> 2877.48] So gophersizes, there's a channel in the gopher Slack. +[2877.62 --> 2877.80] Okay. +[2878.18 --> 2879.12] Hashtag gophersizes. +[2879.62 --> 2880.64] And there's some people there. +[2880.70 --> 2882.10] I don't know if that one's actually that active. +[2882.44 --> 2885.12] That one's a little bit harder because it's not paid. +[2885.28 --> 2887.06] So supporting, it's a little bit trickier. +[2887.32 --> 2889.88] And people email me and I do try to answer where I can. +[2889.88 --> 2893.70] You know, in the Slack, if they ask questions, I think I try to check there every so often +[2893.70 --> 2894.24] and respond. +[2894.72 --> 2898.82] So all those are options because I think that one right now has something like 20,000 students +[2898.82 --> 2902.92] and it can be a little bit trickier sometimes whenever I've got a bunch of other things +[2902.92 --> 2903.42] I have to do. +[2903.92 --> 2906.74] But I do try to help and I do try to answer email and that sort of stuff. +[2906.80 --> 2908.18] So like those are all viable options. +[2908.66 --> 2908.76] Great. +[2909.22 --> 2910.96] Well, we're about 10 minutes out. +[2911.36 --> 2915.24] Let's go beyond Hello World or go 101 and go 201. +[2915.94 --> 2917.62] So we can start to name. +[2917.62 --> 2921.26] So we have a great foundational course and we have exercises. +[2921.66 --> 2925.94] Where else can go beginners go next once they get there? +[2926.12 --> 2927.10] I'll jump in there. +[2927.18 --> 2930.82] So one of the, you know, once you've sort of got your head around some basic go, you've +[2930.82 --> 2934.24] done lots of different examples and you have the basic language across. +[2934.46 --> 2939.88] I think at that point, you know, jumping into effective go in terms of learning how to write +[2939.88 --> 2943.58] good, clean, idiomatic go code is your very next step. +[2943.68 --> 2947.30] And then as Ashley was mentioning, you know, and we were talking about building examples, +[2947.30 --> 2953.28] another item I think that is fantastic, which is similar to awesome Python is awesome go. +[2953.28 --> 2959.10] And if you go to awesome hyphen go.com, you'll end up with a list of, you know, curated go +[2959.10 --> 2960.74] frameworks, libraries and software. +[2961.30 --> 2965.32] And, you know, anything that floats your boat, you know, whether you're interested in audio +[2965.32 --> 2970.86] and music or bot building or, you know, any kind of different piece, there are some really +[2970.86 --> 2973.70] outstanding resources available there. +[2974.30 --> 2978.56] And then finally, I think, you know, the third thing I'd lead off with in that respect would be +[2978.56 --> 2980.44] to go to go user groups, right? +[2980.52 --> 2986.62] Because almost any geography that you're in, either there's online forums or in-person groups, +[2986.68 --> 2990.66] especially the in-person groups, you'll find, you know, that there are go user groups. +[2990.66 --> 2991.66] There's a go user group. +[2992.02 --> 2998.00] I live in a fairly small city in Canada called Winnipeg, which is smack central in the prairies. +[2998.50 --> 3001.00] And yet we have a thriving go user group locally. +[3001.78 --> 3006.86] And so, you know, unless you're rural, you may be able to find, you know, where you can +[3006.86 --> 3010.62] crash a go user group, even if you're traveling and to connect with go users in person, which +[3010.62 --> 3011.48] I think is amazing. +[3012.12 --> 3013.54] Plus one to go user groups. +[3013.98 --> 3018.98] Yeah, we're actually trying to build that up and strengthen a unified go user group, if +[3018.98 --> 3020.58] you will, called the Go Developer Network. +[3020.74 --> 3024.44] So I joined Google a couple of months ago, and I have another colleague that joined the +[3024.44 --> 3025.10] team recently. +[3025.24 --> 3026.12] His name is Van Riper. +[3026.88 --> 3032.82] And he wants to complement GoBridge's efforts, as well as Women Who Go's efforts, into sort +[3032.82 --> 3038.02] of getting this network of groups to be able to maybe give them content, help them maybe +[3038.02 --> 3041.78] something like a live broadcast, give them a chance to go at their own pace. +[3041.88 --> 3046.40] But I really like the idea of community and the importance of in-person learning and trying +[3046.40 --> 3050.84] to balance that off with the people for whom maybe online courses are not working. +[3051.20 --> 3054.26] I think it's a really good thing that that's something you focus on. +[3054.70 --> 3057.36] Because like we talked about all these things with learning languages, but we didn't really +[3057.36 --> 3061.54] talk about the community or the fact that like some coding communities really weren't +[3061.54 --> 3062.80] that inclusive or inviting. +[3063.40 --> 3067.64] And I think the fact that Go has stressed that from the start is one of the huge things +[3067.64 --> 3071.00] for learning it, at least as a, you know, especially if you happen to be in a minority group +[3071.00 --> 3075.52] or something that can be very helpful that you can actually feel okay being yourself and +[3075.52 --> 3077.42] not having to pretend you're somebody else's you're learning. +[3077.78 --> 3078.16] Absolutely. +[3078.42 --> 3081.72] That's why I moved from Python to Go. +[3082.30 --> 3084.46] Python community is fine. +[3084.72 --> 3090.54] But so I found that within the Go community, there is no such thing as a dumb question. +[3090.54 --> 3093.58] And I am really, really good at asking dumb questions. +[3094.22 --> 3096.64] And people are just super, super helpful. +[3096.78 --> 3098.62] They will go out of their way to help you. +[3098.62 --> 3102.80] I have never been part of a community like this. +[3103.52 --> 3103.76] Agreed. +[3104.10 --> 3106.78] And it's the reason why I chose Go and stuck with Go. +[3106.96 --> 3108.62] And I just really am happy for that. +[3108.76 --> 3111.00] So yeah, thank you, John, for bringing that up. +[3111.02 --> 3115.58] Because it is, I feel, of paramount importance in terms of learning the language, mastering +[3115.58 --> 3119.42] the language, and then staying and keeping and communicating to the language. +[3119.70 --> 3122.34] One thing you mentioned, Dave, was Awesome Go. +[3122.48 --> 3126.30] There's also Go or LibHunt, which I think works for all languages. +[3126.30 --> 3130.18] And then they have go.libhunt.com, which is a similar thing, right? +[3130.22 --> 3135.62] You can go see third-party packages and kind of compare them to other packages depending +[3135.62 --> 3136.58] on what you want to do. +[3137.08 --> 3139.06] They sort of release the newest, greatest. +[3139.42 --> 3142.36] And I kind of like that to keep up with my things. +[3142.52 --> 3146.22] I know that Mark Bates, who is a panelist on the show, along with Corey Lanoue, they made +[3146.22 --> 3147.22] Go for Guides. +[3147.22 --> 3152.42] And I think this is kind of along those same lines of trying to find a path or a curated +[3152.42 --> 3156.96] path based on a certain thing that you want to learn, whether that's a data structure or +[3156.96 --> 3158.20] a type or an algorithm. +[3158.64 --> 3160.80] So I'm going to give a little shout out to them for that. +[3161.44 --> 3165.56] Any other resources for people to go or their go-tos, pun intended? +[3165.72 --> 3168.70] We've been very short on go puns this episode. +[3168.94 --> 3169.58] Shame on us. +[3169.58 --> 3173.66] I think one that everybody always says, and it's very hard, but if you can find an open +[3173.66 --> 3177.96] source project you like, it can be daunting at first because you'll be like, how do I +[3177.96 --> 3178.56] get started here? +[3178.62 --> 3179.46] There's so much there. +[3179.70 --> 3180.98] But you mentioned Mark Bates. +[3181.08 --> 3185.36] I think one of the things that he's done a great job with Buffalo is that he can help +[3185.36 --> 3188.40] people get started and be like, just help with documentation or help with something +[3188.40 --> 3191.04] that's relatively easy as you get familiar with stuff. +[3191.54 --> 3195.70] But that one comes to mind because there were one or two people that were very, very early +[3195.70 --> 3197.68] students of one of the courses I made. +[3197.68 --> 3201.66] And they later were telling me that they were actually contributing pretty like more heavily +[3201.66 --> 3202.54] to those projects. +[3202.54 --> 3206.88] And it was because they really helped them grow as a developer and learn more about them +[3206.88 --> 3209.04] and get to the point where they could actually fix PRs. +[3209.48 --> 3211.08] And it's not going to happen overnight. +[3211.32 --> 3215.48] I don't think you can ever as a beginner jump into a project and make meaningful code changes +[3215.48 --> 3217.00] or big code changes. +[3217.26 --> 3221.98] And people who run open source projects have to resist this urge to fix a one line bug. +[3222.38 --> 3224.88] They sort of have to set it aside for a beginner to tackle. +[3224.88 --> 3230.00] But if you can find the right projects and maybe ask around in the Go community for suggestions, +[3230.00 --> 3234.96] that's a great way to grow where you just gradually do small changes and work your way up to handling +[3234.96 --> 3236.06] more and more of the code base. +[3236.52 --> 3236.84] I agree. +[3236.96 --> 3238.86] That is advice I give often. +[3239.66 --> 3245.92] And what I would like to see in the future is people prioritizing PRs. +[3246.00 --> 3248.00] Like, here's what's great for a beginner. +[3248.30 --> 3250.32] Here's what's great for people that are more advanced. +[3250.46 --> 3252.26] There's a site called Up4Grabs. +[3252.26 --> 3253.62] It's not .com. +[3253.70 --> 3255.06] I can't remember what it is. +[3255.10 --> 3260.16] But it does just that, where it's like, these are great bug fixes that you can do as a beginner. +[3260.62 --> 3261.98] And so it will rank things. +[3262.12 --> 3264.98] I would love to see people do that within open source in general. +[3265.40 --> 3265.64] Yeah. +[3265.84 --> 3269.06] I love when GitHub tags, like, great first issue. +[3269.26 --> 3271.34] And then you can just kind of sort by the tags. +[3271.34 --> 3286.26] And then they've optimized their project, not for getting things done necessarily, but for, well, yes, that's certainly important, but also for being inclusive and trying to onboard new members into their project or their ecosystem and whatnot. +[3286.72 --> 3293.30] Another one is first timers only, where basically they limit specific issues and you have to be a first time contributor to that project to do it. +[3293.48 --> 3294.28] Oh, that's nice. +[3294.46 --> 3294.78] Love it. +[3294.78 --> 3299.76] And we would be remiss if we didn't mention Golang-Newbies channel in the Gopher Slack. +[3299.94 --> 3305.20] When I first began, I just loved being able to ask, as Ashley said, all the questions. +[3305.48 --> 3310.66] And I was sort of fearless because I had people that I already had met in person that said, just ask the question. +[3310.78 --> 3316.48] And when you do, if you do that, then it will make people more comfortable in asking what they are afraid of to ask questions. +[3316.78 --> 3318.90] So, yes, Golang-Newbies and Gopher Slack. +[3318.90 --> 3328.76] It's funny because that reminds me of before GoTime got rebooted, I was talking to Matt Reier and he had said that one of his goals was to say, I don't know in the podcast at least once or twice. +[3329.30 --> 3334.12] And his goal for that was basically just so beginners realize that it's okay to admit you don't know and to ask questions. +[3334.54 --> 3342.00] And I think that's important, like you said, just to really reinforce that it's okay to ask questions and that nobody's going to think you're a bad developer or something because of it. +[3342.10 --> 3342.38] Yes. +[3342.74 --> 3345.12] No, and there's so much that Matt doesn't know. +[3345.12 --> 3349.30] We had to throw Matt some shade. +[3349.50 --> 3355.04] So, I think our issue, our episode is quasi-complete because we have thrown Matt some shade. +[3355.14 --> 3356.30] We've given Steve a hello. +[3357.12 --> 3359.94] We mentioned Mark and his Gopher guides. +[3361.12 --> 3362.98] Anything else we're missing before we go? +[3363.22 --> 3374.40] Well, just to add on the question piece, one thing I notice, you know, with the courses that I teach, so often someone asks a question and the next thing you know, there's a crowd of Me Too's. +[3374.40 --> 3376.42] And I'm like, where were you a minute ago? +[3376.70 --> 3381.32] You know, like, you know, why did, you know, so-and-so have to be the first one to ask? +[3381.46 --> 3388.70] And then inevitably when the answer, you know, is made, you get a handful of a dozen or so thank yous. +[3388.74 --> 3394.26] And you're like, you know, make it interactive, especially when you're trying to bring, you know, that about. +[3394.58 --> 3399.70] I really find that people are unnecessarily shy for whatever reason. +[3400.32 --> 3402.22] It's not even unnecessarily shy. +[3402.22 --> 3408.26] I think that people's fear of embarrassment rules them, right? +[3408.40 --> 3409.38] It's all psychological. +[3409.62 --> 3414.72] So there has to be somebody in the room who just doesn't embarrass by not knowing something. +[3415.10 --> 3415.18] Yeah. +[3415.28 --> 3425.22] I think it's also how you answer can make a huge impact because if you make it sound like it was something obvious or another example is like if somebody says like, well, I'm not sure, but I want to try to help you. +[3425.22 --> 3429.48] And if you just jump in and just disregard everything they said, then they're not going to try that in the future. +[3429.72 --> 3439.42] So like I think there's a lot of thinking about how we answer questions and how we present stuff to people and just trying to be as, I don't know, gentle or friendly as possible is very important. +[3439.98 --> 3440.42] Absolutely. +[3440.72 --> 3441.18] Very good. +[3441.52 --> 3448.10] Definitely encourage those people to ask more questions might open it up for other people to ask questions. +[3448.10 --> 3448.54] Great. +[3449.58 --> 3452.26] Well, we're kind of coming up on the top of the hour. +[3452.40 --> 3454.88] There's a couple of things I wanted to mention. +[3455.34 --> 3459.86] First of all, there's one last segment that I'm really interested in learning more about. +[3459.98 --> 3465.10] And I made a survey because I'm just going to like I am not a survey designer, but I really am curious. +[3465.56 --> 3470.94] How many of you here on this panel have had to learn a new programming language while on the job? +[3470.94 --> 3471.68] I have. +[3471.96 --> 3472.16] Yeah. +[3472.38 --> 3482.68] And is learning while on the job, whether it's because you have to or you're exploring for, you know, new tooling or because it's, you know, a change in the technical stack? +[3482.78 --> 3489.22] Like how does learning or beginning from that headspace any different from learning maybe on your own time? +[3489.86 --> 3495.76] I feel like it's different because usually if it's for work, you know why you need to learn it. +[3495.96 --> 3496.16] Okay. +[3496.68 --> 3500.10] So you kind of have some base knowledge, right? +[3500.10 --> 3506.06] So you're like, I have to learn it because I need to build this feature or we're building this tool and that's why I have to learn it. +[3506.12 --> 3508.14] So you can Google more specifically. +[3509.28 --> 3509.64] Okay. +[3510.34 --> 3511.20] John, Dave? +[3511.34 --> 3512.48] I've had to learn on the job. +[3512.88 --> 3520.02] And I think like Ashley said, it's specifics, but it's also like your focus is on being productive, not academic learning. +[3520.32 --> 3525.66] So that forces you to not let yourself get distracted as much, I guess, is the best way to put it. +[3525.66 --> 3535.00] You know, so like if you need to build a web server, whatever, a WebSocket server or something, you're going to focus more on, I just need to get this done and I can come back and fine tune it later. +[3535.00 --> 3537.48] Or I can come back and look at those things I'm not quite sure about later. +[3537.48 --> 3543.70] Whereas when you're learning it just, you know, for leisure, it's, there's no really like time restrictions or anything. +[3543.70 --> 3545.84] So you have a little bit more ability to get distracted. +[3546.14 --> 3555.22] Well, I think that getting distracted is the enjoyment that you have when you decide to explore something in your own personal time for your own personal benefit. +[3555.32 --> 3555.52] Right. +[3555.58 --> 3556.64] You can go down those rabbit holes. +[3556.78 --> 3557.46] You can explore. +[3557.72 --> 3559.78] It's almost like a buffet of knowledge. +[3560.24 --> 3562.24] And I love reading and learning. +[3562.36 --> 3567.78] I listen to a lot of podcasts because I want to experience those tangential things that I may not otherwise have. +[3567.78 --> 3580.36] But I mean, I will say, even though programming isn't a day to day part of my job, I think anyone in IT or anyone in the world in general can literally have what happened to me this recent Friday where the CIO called the boss. +[3580.36 --> 3580.70] Right. +[3580.72 --> 3583.12] And said, I want Dave full time on this other project. +[3583.62 --> 3586.08] Whatever he's doing right now goes away. +[3586.16 --> 3586.48] Right. +[3586.52 --> 3588.56] And the next thing you know, it's OK. +[3588.84 --> 3596.60] So let's figure out, you know, how we're going to approach this and and what I need to do there and what his objectives are and and how do I fulfill those? +[3596.60 --> 3600.68] So it's very pointed because you're trying to get to a particular objective. +[3600.68 --> 3607.82] But the pleasure of just learning something for kicks and giggles sometimes can be there's a joy with that as well. +[3607.86 --> 3608.18] Right. +[3608.22 --> 3613.02] And it's different than when you're trying to fulfill, you know, what the CIO asked on a Friday morning. +[3613.18 --> 3613.62] Yeah. +[3613.80 --> 3614.14] Yes. +[3614.32 --> 3617.80] I guess I should say those distractions, like you said, can be very good things. +[3617.80 --> 3622.26] But I just think that's one of the big differences is that, like, sometimes they're good, sometimes they're bad. +[3622.26 --> 3629.90] And I think that having them there, it's just is one of the big differences for me is I just felt like in the work environment, I never was. +[3630.30 --> 3633.74] I never had that freedom to just really go look at the stuff, explore things. +[3634.08 --> 3636.98] You know, you can't like go compare three different HTTP routers. +[3637.08 --> 3638.42] It's like I just need to pick one and use it. +[3638.64 --> 3639.08] Exactly. +[3639.24 --> 3643.40] No, you definitely get less in the weeds when it's for work. +[3643.66 --> 3644.22] Very cool. +[3644.36 --> 3646.08] I want to hear more about this. +[3646.08 --> 3649.40] I am going to put a link in the chat. +[3649.80 --> 3653.00] It's bit.ly slash on job code survey. +[3654.00 --> 3657.36] And I just I'm going to share it on my Twitter and maybe everyone else can share it. +[3657.42 --> 3659.82] I just want to know because that's kind of my my one. +[3660.24 --> 3661.30] How is that different? +[3661.56 --> 3669.96] And are there any gaps out there in the ocean of content that you think that need to be most filled at this point as we close out? +[3670.32 --> 3670.88] Children's book. +[3671.16 --> 3671.74] Dang it. +[3672.08 --> 3672.50] All right. +[3672.60 --> 3674.02] I think that's what is happening. +[3674.42 --> 3675.14] Children's book. +[3675.14 --> 3676.06] It is happening. +[3676.44 --> 3677.60] Please do it. +[3677.76 --> 3681.98] Brian Kettleson and I were supposed to do it, but we are just far too busy. +[3682.74 --> 3684.88] I really need somebody to do this. +[3685.12 --> 3686.78] Just selfishly, please do it. +[3687.40 --> 3687.80] Amazing. +[3688.40 --> 3693.60] I almost feel like there needs to be resources teaching people how to ignore all the bad advice they're going to get. +[3693.82 --> 3700.60] Like Ashley had said, people keep telling you to build something or like if you get on the go Reddit, they're going to tell you never to use third party libraries or like there's just right. +[3700.88 --> 3702.02] There's a lot of things like that. +[3702.02 --> 3706.88] And I think when you're learning, especially as a beginner, it's really easy to get suckered into this. +[3707.02 --> 3708.22] I need to use all these things. +[3708.86 --> 3710.42] And I don't know. +[3710.50 --> 3716.90] It's almost like they need a guide to like a practical guide to just ignoring all that other stuff that doesn't matter and just focusing on what does matter. +[3716.90 --> 3717.26] Yes. +[3717.44 --> 3717.74] Yes. +[3717.78 --> 3718.82] Also, please write that. +[3719.50 --> 3719.86] Yes. +[3719.96 --> 3724.30] I have that as a survey question, which is like, what is this like letter to your past self? +[3724.42 --> 3728.76] Like what do you wish that you didn't have to unnecessarily go through? +[3728.88 --> 3729.58] So, yeah, totally. +[3730.08 --> 3730.66] I think that's great. +[3730.72 --> 3733.46] I wonder if we, I wish we had more time to discuss that. +[3733.52 --> 3735.44] But that is just really some of the best. +[3735.50 --> 3739.56] If we can sort of get rid of some of the headaches that we ourselves had to endure. +[3739.56 --> 3741.88] Eh, maybe that's all part of the journey. +[3742.98 --> 3744.94] So let's, I'm going to close out. +[3745.10 --> 3751.26] Whatever kind of beginner you are, we hope this episode has given you a good starting point as you start to learn and improve your go. +[3751.68 --> 3761.22] The learning gotchas, how learning go might be different from your current programming language, where to ask questions and get help from the community, and where to continue on your learning journey. +[3761.82 --> 3765.12] Whatever you do, embrace failure and don't stop trying. +[3765.64 --> 3766.54] Engage with others. +[3767.16 --> 3768.78] Use your resources. +[3768.78 --> 3770.22] Go for Slack. +[3770.54 --> 3771.30] Go Lang Newbies. +[3771.80 --> 3773.90] Attend meetups online or in person. +[3774.50 --> 3778.58] Like the famous quote from Benjamin Franklin, tell me and I forget. +[3779.22 --> 3780.60] Teach me and I may remember. +[3781.14 --> 3782.36] Involve me and I learn. +[3782.72 --> 3785.58] Thanks everyone for today's episode of Go Time FM. +[3786.12 --> 3787.12] Thank you, Ashley. +[3787.38 --> 3788.18] Thank you, John. +[3788.32 --> 3789.92] Thank you, Dave, for joining me. +[3790.20 --> 3791.18] We'll see you next week. +[3795.14 --> 3797.70] Thank you for listening to this episode of Go Time. +[3797.70 --> 3802.98] If you liked this oldie but goodie, it's from our list of recommended episodes. +[3803.38 --> 3806.42] Find the rest at gotime.fm slash recommended. +[3806.98 --> 3809.50] And of course, subscribe now if you haven't already. +[3809.80 --> 3813.52] We are in all the podcast apps or on the web at gotime.fm. +[3813.52 --> 3817.34] And if you enjoy the show, please tell a friend to give us a listen. +[3817.54 --> 3820.60] It's the best way you can help Go Time grow and thrive. +[3821.00 --> 3829.48] Thanks again to our partners at Fastly for CD-ing for us, to the mysterious Breakmaster Cylinder for these fresh beats, and to you for listening. +[3829.84 --> 3830.66] We appreciate you. +[3830.66 --> 3836.16] Next time on Go Time, Natalie talks debugging Go with some new friends. +[3836.52 --> 3839.50] We'll have that episode ready for your ear holes next week. +[3839.50 --> 3854.28] Game on. diff --git a/Go in medicine & biology_transcript.txt b/Go in medicine & biology_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..d4b8572220621cfe4347ffb33d8de4df4eb5b697 --- /dev/null +++ b/Go in medicine & biology_transcript.txt @@ -0,0 +1,335 @@ +**Natalie Pistunovich:** Hey, Ian, how are you doing? + +**Ian Lopshire:** I'm doing great. How about you? + +**Natalie Pistunovich:** I am doing well. I'm excited about our episode today. We're gonna talk about Go in biology and medicine. And our guest is Tim Stiles. Hi. + +**Tim Stiles:** Hey, how's it going? + +**Natalie Pistunovich:** Good. We are very excited that you're here to talk with us about something that I personally did not do enough since high school. + +**Tim Stiles:** Hah! This is great. I've taught a lot of people this; this is wonderful. I don't have any slides, I don't have any whiteboards, but I'll try my best. + +**Natalie Pistunovich:** \[laughs\] So Tim, you are doing biology with Go. How did that happen? Why? Tell us everything. + +**Tim Stiles:** So I guess I'll start with the short story, and then you can start asking me "How did you even become a software engineer in biology?" But starting with Go was -- I've been writing biology software for a while, and I'd come up on this new project where essentially it was a version control system for cell lines. So think of it as like Git for like a cell when you're programming a cell to do a new thing, and you sequence the thing, and the cell may or may not do it, and you try to track all the changes you've made across these cell lines as you engineer them. And I actually went into the original Git code to learn this. And I'd written Git servers before, so for me, it was familiar territory. But the thing was that Python just - it's a just-in-time compile language. It's not really built for command line tools. And I was "Well, what is?" And it came down to Go, which I'd played around a little bit before in undergrad a few years before, and Rust, and pretty much my decision was "Well, Go's about as fast, and much easier to use, so I'm gonna go with that." + +So that's sort of how I started with that in biology. And then it spun into something more, because at that time I was getting tired of biotech software... Because it's a hard topic; you have to know how to write software, and you need to know how to do biology, and the intersection is very, very small. And people will tell you this all across the industry, it's like the thing where like -- I was in a meetup last night in biotech software, and everyone's "Why do the biologists not know how to code? Why do the programmers not know how to do biology?" I'm "Yo, I'm right here, guys. Don't forget about me..." Because maybe one of a handful can do this. + +And so what happened was I was getting ready to leave for tech, because I was burnt out on this lack of tooling... And a friend of mine who was working for a professor at Stanford, he called me on his drive down from Stanford to LA where his parents lived, because he was quitting his job at Stanford... He was only 20 at the time; he was quitting his job at Stanford, and he was just going back to live with his parents, so he started with this idea for a startup. And he's "Hey, Tim, before I leave my job at Stanford, I need to take all of these genetic parts that people have designed, and I have to take them from the JSON we have internally and turn them back into GenBank, which is this weird, esoteric sequence data format that we've been using since 1970, and it's like THE sequence data format. This is what the government ships." The NIH and various agencies in Europe and Japan have this consortium of groups that - they have their own, I guess you could call it like a GitHub for DNA, or DNA parts, or sequences; it's not quite the same, it's a little bit more esoteric and bureaucratic... But it essentially functions in a similar way. + +\[06:06\] And so there's this really esoteric, designed in 1978, pre even XML format, that's all whitespace-based, and he's like, "Hey, I can't convert this JSON to GenBank, and GenBank to JSON and back. I can't do it." I was like, "You've got to be kidding me. This is the world's most common data interchange format, JSON, and the world's most common DNA sequence format, GenBank, and you can't find a reliable way between the two?" And he's like, "No. I can't." + +**Natalie Pistunovich:** Wait, in any programming language, or in Go? + +**Tim Stiles:** He was trying to do in Python, and the thing that was happening is that he kept getting this fatal error, on several sequences, where instead of going \[unintelligible 00:06:40.29\] like, you know, letting you just to handle the error yourself, it would just kill the whole run... And so the run would take forever, and you'd be like six hours in, and it'd be like "I don't like this specific sequence. No." Or "I don't like this specific metadata." No, I would kill the whole run. And at that time -- he's a much better programmer now. He's a great programmer. He spent a lot of time developing this with me; he's been more of the biologist side, but now he's learned a lot of software engineering working with me... And it's turned into this real project where... Like, I announced I made GenBank parser, and people were into it. I was like, "Hey, I'm done with this." It took me three weeks of going through various poor forms of documentation. I found one European website that I think is official from the government, that had bad SSL certificates... + +So it gave me warnings when I tried to click on it in Firefox... Like, "Don't go to this bug-infested site from the government." I was like, "Okay..." I'm like, "I got to it." And that's the only specification I could find for the file format. It took me three weeks to write this parser. It's not like I haven't written a parser before. I've written plenty of web scrapers in my lifetime. It's not like it's new to me. But it became this real thing where - I tell people often, for biotech it's a big deal if you put it in JSON, because they have been handling data formats since before the public; think Usenet era, 1978. We don't even know who made the most -- maybe we can find out who came up with JSON, but GenBank, I've tried to figure out who came up with GenBank. I can't. All I know is there's some council of elders, in 1978 got together in the deserts of New Mexico and decided on this format. And there's no record of who these people were. And we've all been living with the consequences ever since. + +**Natalie Pistunovich:** This is really mind-blowing. I had no idea... Wow. + +**Tim Stiles:** It starts with the complexity of this -- I guess you call this legacy data, or legacy code or software... Because bioinformatics has existed since before the web. We've definitely evolved a lot since the internet came about. But scientists had been using the internet before everyone else by at least like 15 years maybe... So there's a lot to parse there. So the most recent file format that I've seen GenBank, the NIH-backed, the United States NIH-backed database use something called ASN1, which is like the precursor to XML. It's like what XML was to JSON, but for XML. It's super-weird. And someone asked me, "Have you ever written a parser for this?" I'm like, "I didn't even know this was a thing. This is amazing. But also, no, I never want to do this." + +And so a lot of biotech software is limited by the fact that there's these data formats that the government uses, or some repository uses, but everyone else is just sort of like - most people doing this work are scientists, they're not software engineers. And there's a big push right now for scientists to learn a little bit of software engineering, DevOps practices... I'll make a post, like "This is what unit testing is. Have you ever heard of example tests? They're really great in Go", and they'll be like "You saved my life. I can actually write my code." It's wild. + +So there's this whole discussion right now in the field of like "How do we write better software?" and I've found myself in the center of it, because I wanted to write this to make cool stuff. Like, it does have medical applications, but I'm thinking of stuff like flying seaweed, and \[unintelligible 00:09:44.01\] trees and all this other goofy stuff that I think of in my spare time, where I'm like "Someday I'm gonna need the software to do that." But it just didn't exist. And the tools that you mostly find are usually either old C or Python packages, or there are companies that rewrap it in a nice GUI and sell it to the sciences for a good price... And you know, they do that, but it's not really a programmatic approach; you can't do much with it other than drag and drop. + +\[10:10\] So for example, DNA synthesis, which is what we use to make new DNA for certain genes that we're trying to engineer, there's only like two companies that have an API for that, out of like the dozens that exist. Most of the time, if you want to send something to them, you have to do like a drag and drop into their GUI. And I've talked to a lot of vendors about this, and it's just like -- yeah, there's only two or three I know of that have an API. One of them you have to send your DNA in Excel format. They don't take JSON. It's Excel format. I know, wild. But yeah, we can get off this diatribe for now, but... Yeah, it feels a little wild. + +**Ian Lopshire:** So that's the tech side of things. How'd you get into the biology side of things? I think before the show you said you had a degree in computer science, but how did the biology come in? + +**Tim Stiles:** That's true. So this is really funny... So back when I was a student, I first started as a mechanical engineering student, and then I switched to as a design student, and then I got really into robotics. I was playing around with 3D printers, and I was doing a lot of computer vision stuff... And that was like my jam for most of undergrad, was doing computer vision stuff. And at the time, the NSF here in the United States was giving out grants to undergraduates, this "Do computational bioinformatics biology research." And so I was like, "Well, I like money, and I want to get my undergrad paid for", so I found a couple of professors to work with, I tried a few things, and I kind of stuck to computer vision until I got this job that bridged out from undergrad, where I was working at Harvard Med as a research assistant doing computer vision for endothelial cell morphology, which is a fancy term for the cells that line your blood vessels, how they move and shape and morph as they respond to hypoxic cells. So if you have a cell that doesn't have enough oxygen, it releases something called vascular endothelial growth factor, VEGF, which is a chemical that goes through all of the nice little fluids and tissues in your arm, or your body... And when it hits the endothelial cell, the endothelial cell goes, "Hey, that cell's in trouble", and it coordinates all the other nearby endothelial cells to shift and morph and change the rigidity of their cell membranes to sort of push their way towards this hypoxic cell to deliver oxygen. And the reason why this is important for study - I mean, obviously, it's just cool, first off... But the reason why a lot of people like to fund it is this is a dynamic that's very central to tumor growth in cancer research... Because what tumors do and what cancerous cells do is they put out this VEGF factor to an extreme; they're just saying they're constantly hypoxic, even though they're not. They're saying they're constantly hypoxic, and recruiting all these endothelial cells to feed it more and more blood, essentially, to deliver oxygen to it, and let it grow rapidly. + +So usually, the line of research there is "Can we find a way to mess with that to stymie tumor growth?" And the answer is, "Yeah, a little bit." But cancer is a very complex disease, with lots of different ways of presenting itself, and it's not the end-all there. + +So I got started in that, but at the same time, my spouse, Ren - we were dating them, but now we're married - she was working at a lab at Harvard Med under George Church, who was a famous synthetic biologist. He's like in Wired Magazine... He's a really nice guy. He says yes to everything. So if you have a synthetic biology startup, he'll be on your board. I don't know if he'll have enough time for you, but he will. + +And so she was working in his lab, doing self-aid programming, which is this concept... Have you ever heard of stem cells? Do you know what stem cells are? + +**Natalie Pistunovich:** Mm-hm. + +**Tim Stiles:** So stem cells - the idea is that they're -- a self-aid program is the idea that you can turn a stem cell into another kind of cell, a blood cell, or a tissue cell, an eye cell... You know, a different cell. And you do this by introducing various chemicals to it, and there's different stages. So you're programming of some various enzyme, or hormone, or some odd factor, and it becomes a different cell. And the big thing that people have been trying to do is figure out how to take cells that have gone from stem cell to fibroblast cell, or skin cell, and then turn it back into a stem cell. Because if you can figure out how to do that, and culture those, it has a lot of applications for, say, tissue engineering, and replacing degenerated tissues, and things like that. It's still highly -- I wouldn't say speculative, but it's new. It's like, we've been doing it for a while, maybe there's a few therapies for it, but I'm actually not a medical guy; I am a little squeamish, which is hilarious to a lot of people, but it's true. If I'm not familiar with it, I may get a little nauseous talking about it. + +\[14:26\] But how I got in biology is that eventually doing this, walking home with my wife, we were in the same area, so I'd walk 50 minutes to her office, and I was on the way home, and we'd walk home together, and she'd be explaining her work to me... And she was mentioning this thing called plasmids, which we're gonna talk about a little bit more later, but essentially, they're these little circular tokens of DNA, these little -- consider them like tiny functions, that contains like a gene, and maybe some things to promote the expression of this gene, and a couple other things, like maybe some sort of resistance to an antibiotic, which is useful in lab plating... It's called plate selection. I can talk about that, too... But essentially, it contains like two or three genes, it's like 10,000 base pairs, which is tiny compared to all of your genome. But they occur in nature and bacteria as a way of getting around the fact that they reproduce asexually, and they need some genetic variation... But we've figured out how to use them in other places to do sort of like small genetic testing. So they're like tokens, but she would call them circular. And I was like, "Wait, when you project it, isn't the helix kind of circular, when you look at it from a certain projection?" She's like, "No, no, no, no, no, no, no, no." + +And so I was having a hard time -- there was some problem with staining, where I couldn't get quite the information I wanted out of the data I had. There was something going on there. It's been so long since I looked at this, but essentially, I was like, "I need to be able to explain what colors I need, and how it works for staining." And I can't rely just on the wet lab biologist to kind of bridge this gap for me; it's not going to work that way. + +And so eventually, she got tired of explaining things to me, and very gently suggested I go to this garage lab in Somerville, Massachusetts called BOSS Lab, which has been around forever in some various way, shape, or form... And I showed up there, and I was like, "Hey, I need to learn biology for my work", and they're like "Yeah, we were gonna make a class, but we haven't done that yet. So if you teach the class, we'll teach you how to teach the class." And so I spent, I'd say, like six weeks in this lab, nights and weekends, working on what's called the heat shock transformation protocol to put these plasmid tokens in the bacteria, and kind of learn the basics of biology that way... And it was kind of like a world-class education in like a sparse environment, because the people I was messaging for help were PhD's from Institute Pasteur, and UC Berkeley, and UW Madison... And there's a software engineer who's the VP of like engineering at DataRobot, and they'd all be giving me advice over Slack, like, "Oh this over here, this over here..." But I was also kind of on my own, because none of them had made this protocol before. And from there, I kind of just got too into it, kept getting contracts, and kept doing things, and eventually, I was kind of struggling and I wanted to get out of the field, and I just stumbled on something that worked. And now I make my living as a consultant, doing something that works, which is really great. + +**Natalie Pistunovich:** You have also built Poly, which is a Go package for DNA... + +**Tim Stiles:** Yes, so for engineering DNA, which is very specifically -- it's very different from what a lot of other people do with DNA. A lot of times they're looking for diagnostics, they're looking for variants, they're trying to assemble it into a whole genome for reference... What my package does is it tries to engineer DNA to be used for certain experiments and certain designs. And so what my package is mostly focused on is going from designing DNA as a concept into something that you can put into cells... Which is surprisingly unique. You'd think there'd be more of this, and there are; there's a lot of companies that write this internally for themselves. Imagine like they're writing their own React framework, and then not sharing it... And that's what's going on in almost every company that does this. They burn like millions of dollars doing this, and they don't share back, in most cases. And so I wrote this as a thing where I was thinking, "I want to be doing this for the next 10 years, I want a really stable library, I want it to be really well tested, and I want it to have the features that I need to create cool stuff." + +\[18:10\] And so I started from sort of that principle, and it started with just parsers, which every bioinformatician will tell you that's like the first thing they have to do, is they have to write parsers. It's like the bane of our existence, is taking government data and putting it into JSON. I wish I was kidding. But after you get past that, then you get into stuff like, "Well, how do you manage having a hash ID for a circular sequence?" Because the string is circular; there's no way of telling where it's going to start officially. So it turns out there's this old algorithm called \[unintelligible 00:18:39.27\] that given a certain sequence, a circular, it will always bring it to the deterministic point. And we use that to create hash IDs for these plasmids, because otherwise it would take an inordinate amount of time to figure this out. Because then you're using like pairwise alignment algorithms... And these are not linear algorithms; they're kind of gnarly, and not my specialty. People actually specialize in like string alignment algorithms for this specific stuff. \[unintelligible 00:19:01.16\] a professor at Harvard Med specializes in this, and his software runs the world quietly. He does not get enough credit for how much bioinformatics relies on his stuff. The practitioners know, but the people downstream have no idea that this one Harvard professor is keeping all of our -- you know, if you've ever used 23andMe, Ancestry or anything else, or any genomic testing ever, your DNA has probably gone through one of his algorithms. He's very prolific. And unlike myself -- you know, I still have a small user community. I have some very hardcore users that have their own startups, that are raising money for it, and I have a handful of people that give me feedback, and I've even had large companies go, "Hey, we're thinking about using this", but I still feel like a small player in comparison to people who have come before me. Again, I wrote it because I want to engineer DNA, and I think it's super-cool, and it was hard to do that with the tools available. + +**Ian Lopshire:** So I want to get into the nitty-gritty details of it, but first... + +**Tim Stiles:** Yeah. + +**Ian Lopshire:** ...can we talk about like high-level what it does? What kind of projects could you do with it? Are there projects you've used for it? ...just to kind of get some context here. + +**Tim Stiles:** Okay. The thing is, it's a very vast library, despite how much I've underplayed it so far... So some real things are - I know that a couple of consultants have actually engineered microbes as a potential therapeutic, and I've done that as a consulting gig for a couple of YC companies that I probably shouldn't name, so I will not. There's another company that's trying to automate DNA synthesis and cloning, which is actually a big(ish) business. There's plenty of companies trying to do it, and he's like a main contributor. They just closed a $3.5 million dollars -- or are about to. I don't know, actually, etc. But if you're interested in like writing Go and doing synthetic biology, check out \[unintelligible 00:20:42.28\] in San Francisco. They're really great. I keep my surfboard there. I can recommend them wholeheartedly. + +But there are things that people could be doing with it, that they aren't yet. One is designing primer tests, like the PCR tests, like the ones you've seen for COVID. There's very explicit tooling to be able to do that, for designing primers, and looking at -- I'm currently working on a project where I'm essentially taking all this GenBank data that I told you about, the antiquated GitHub of DNA, and I'm putting it into a graph database of all these little fun extras, and part of that is now that I could probably, hopefully, be able to design primer tests for almost every viral string of DNA we have... Which is really cool. + +So there's a lot of medical applications, there's a lot of biomedical applications, which most people when they talk to me are very interested in, because it's innately human. Everyone's had a medical condition, or knows someone who has a medical condition. And one of the examples that I should have started with - I don't know why I didn't - is... If you're a diabetic, or you know anyone diabetic, they are taking insulin that was produced in yeast. Not the classic way, which was originally derived from pigs. But in the 1970s, I believe it was Eli Lilly engineered yeast to produce human-compatible insulin. + +\[21:58\] And this is how synthetic biology works, at least at the early stages, for a lot of companies, is they engineer yeast, or E. coli; not the dangerous kind, not the Chipotle \[unintelligible 00:22:04.13\] kind, but like lab-safe, like difference between like a \[unintelligible 00:22:07.18\] and like a domestic cow sort of difference... And they engineer it to produce some molecule, some chemical, some drug, some valuable -esque thing, and then they bring it to a contract brewer in Wisconsin, and they go, "Hey, can you take this strain and brew a ton of it, so we can extract the thing of value?" And so that's how insulin has been produced since the '70s, is this way. And it's actually part of the reason why I think it's so criminal how expensive it is, because it's actually really cheap to produce. + +But the thing is that this is a technology that's been around for longer than I've been alive. It's been around for 40 or 50 years, but it's sort of become this new thing in the public eye of, "Oh, I can do it, too." It's not just the realm of Monsanto, and Eli Lilly, and pharmaceutical companies; it's now getting to the realm of people like me, who have created this library that companies use. And I wrote this in the comfort of my living room. I haven't had access to a lab in too long. I want to build my own lab again. But it's at the point where now anyone with a computer can contribute in some way, even if they don't quite understand biology. + +**Ian Lopshire:** Sorry, I'm processing for a second there... + +**Tim Stiles:** Yeah, yeah, sorry. + +**Natalie Pistunovich:** Yeah. A lot of inflammation. Super-interesting. + +**Break:** \[23:20\] + +**Ian Lopshire:** So that kind of made sense... \[laughter\] To my limited mind. So what does it do? Are you like running simulations? Is it statistical modeling? Is it like how DNA reproduce? + +**Tim Stiles:** Not statistical modeling, it's more construction. + +**Ian Lopshire:** Okay. + +**Tim Stiles:** So think of it this way... If you need to engineer a DNA construct, like these little plasmid tokens I've told you about, these little, essentially swappable functions we'd like to put into different cell types, you have to construct that. They exist in nature's bacteria, but for our purposes, we need to make sure that it's human-compatible, that it will express the gene, and all this other stuff. And some people, they take essentially templates where they just swap out the gene for a different thing, and it'll express in the organism choice... But my software makes it easy to design thousands of these at a time. And traditionally, the way that most people do this is they design it one at a time, with like a CAD-like GUI software. + +Say you want to run a whole DNA synthesis operation - my software does that end-to-end. You can design the software, you can design the construct of interest, you can simulate how you're going to construct it, you're going to simulate these special ways of printing DNA... So DNA synthesis is a little funny, in that there's some sequences we can't print, but we can print analog sequences that are similar enough it doesn't matter... But we have to know that. And so my software, for example, fixes that little quirk; it optimizes to be expressed in the cell, because different strings of DNA will be expressed differently, in different organisms. Like, they'll express in both, but at different rates, depending on like usages of what we call codons, which... I should have just started with a lecture on the Central Dogma, but you can look it up at home. It's called "The Central Dogma of Biology", if you're listening now. It's like the core tenet of the biology you need to understand this. + +\[26:15\] So essentially, my focus was just on making it so you could engineer a DNA string to do what you want, and to express faithfully, and you wouldn't have problems when you send it to a DNA synthesis vendor, for them to go like, "Oh, it's not the right sequence." And so you could do it at scale, instead of like one at a time, via one of these many CAD-like tools; you could engineer thousands at a time, which is what a lot of scientists do. I'm not sure how they do it; I think they're mostly also writing Python, which is not fun for them, because they're writing a lot of this from scratch... But that's like the method of production here, is you either use CAD software and you write one or two, or maybe you can get more out of it, or you write a script that generates thousands of primers for, say, PCR tests that you're trying to develop, or thousands of variants of a plasmid, these little tokens I've been talking about, that will express something for an experiment. Or you have to do it one at a time. + +And so my software is really the only software out there that's fast and stable; it can do this reliably with high unit test coverage. There are some - I would call them predecessors, that are really impressive. And there's also really impressive GUI tools that you'll do this one at a time sort of thing too, that are open source. + +My project is like the first that makes it really possible to do this at a scale where you're not looking at bad compute bills... I mean, this is the thing... Like, synthetic biology shouldn't have high compute bills. It's not like we're looking at whole genomes all the time, unless you're working on human-specific stuff, which not everyone in synthetic biology is. A lot of people are, obviously, because that's where a lot of the money is, but the strings you're working with are on the order of like 10,000 - kilobase pairs is what we call it, instead of kilobytes, which is \[unintelligible 00:27:47.22\] to like 4 million, I think is what E. coli can be, and like 10 million maybe for yeast... I forget. Some biologist out there, if they're listening, is probably screaming at me for getting this number wrong. But it also varies between these strains. I'm running out of words, so I think I'll let you guide the topic to something... + +**Ian Lopshire:** Yeah... Could I repeat that back to you and just see if -- + +**Tim Stiles:** Yeah, of course. Yeah, this is a biology lesson for everybody. I wish I was talking more about Go and my practices there... I should have done that too, but yeah... + +**Ian Lopshire:** We'll do Go right after this, I promise. + +**Tim Stiles:** Okay. + +**Ian Lopshire:** So basically, someone has engineered this like plasmid that they want to be able to produce, right? ...to do whatever X thing, make a new protein, or whatever. + +**Tim Stiles:** Yup. They've literally just whiteboarded it. It's a concept on like a napkin. + +**Ian Lopshire:** Okay. Yeah. And so they take your software - or this is one thing it can do at least, is run simulations to try to find actual DNA that we can produce... Like, synthesize in the real world. But we don't have to touch the real world, we can just simulate it, and yours can be like, "Yeah, we can make this." + +**Tim Stiles:** Yes, exactly. + +**Ian Lopshire:** Okay, that's cool. Okay. + +**Tim Stiles:** It's really valuable, because each one of those plasmids I've been talking about, the price is always going down for DNA synthesis, but right now, it's not uncommon to hear of like a $500 to $700 bill per plasmid. It's not uncommon. I have a customer who I'm working with right now, and they make DNA libraries that every time they send out an order, it's $20,000 to $40,000. So for them, it's worth it if they have a little extra confidence that instead of getting their grad student or whoever they -- a lot of scientists don't know how to hire programmers; we can get into that dynamic too a little bit. It's a really fascinating one. But essentially, instead of having to worry about \[unintelligible 00:29:24.19\] and maybe blowing $30,000 out of his research budget, he just comes to me and says, "Hey, can you write a pipeline for this out of the software that you wrote?" And I go, "Yeah, I can do that." And it's cheaper than what it would cost him hiring his own people, and it's just a nice deal all around, and he gets the feel like, "Oh, this is secure. There's testing behind this." + +Because a lot of scientists - there's a lot of prayer. I mean, if you ask a scientists what they do before they send out a sample to be sequenced, or go through \[unintelligible 00:29:56.25\] there's this false dichotomy between religion and science, but a lot of them, even if they aren't religious, do a little prayer, like "I hope, I hope this works..." + +\[30:10\] Because scientists are engineers; it's a difference in practice. Scientists are trying to discover something new, and they're doing it like the startuppy, like "Gotta do as fast as possible, plant the flag, get the research, get the data, boom!" And engineers, we can be a little slower, we can be a little bit more pragmatic, we can make a little bit more stable sort of maintenance-conscious decisions. And I think that's sort of what my role is with scientists, is I make these maintenance-conscious decisions where they can go, "Oh, that's a good library. I can really trust it, I can really believe in it. I know that Tim isn't trying to run as fast as possible towards this goal; he's trying to build for the long-term." + +**Natalie Pistunovich:** That is really interesting. We talked about what it is, what led to that, what it is doing now... Tell us about the future of bioengineering as you see it. + +**Tim Stiles:** So this is a conversation I come into a lot. And actually, I have friends who - we are like polar opposites. It's one of those things where we both agree where the future's going, but we disagree on how the future's gonna get there, and what the practice of it is going to look like. So for me, the practice of it in my mind is, you know, software engineers who have a decent enough biological experience, it's easy to learn, hard to master, all the theory behind it. It's a lot of little tiny variables, which is - computer scientists and engineers, we know; we have lots of tiny little variables all over our code. We're aware of this. We're gonna have to keep track of lots of little tiny details. But there's this whole breadth of research, and every once in a while I get dragged into a field that's adjacent, that I know nothing -- like, I had to learn about immunology recently. Immunology is like its own thing, where it's heavily related to like genetic engineering practice, especially with these new immunotherapies we're doing for cancer, and stuff. But I had to go and learn about immunology as like a biological basis in like human medicine. + +But the thing is that we're gonna get to a point where hopefully there's enough software engineers with biology experience that can talk to these people that have this experience in more pointed, niche parts of the field, like immunology, or plant physiology, or these different biological topics, where they can work together and some will say, "Hey, I need to go probe this." And they'll talk to the engineer, and the engineer will get it enough that they can work with them. And that's what I would like to see, is two scientists with like 20 software engineers, writing some awesome code, doing stuff. + +And then my friend - he's entirely the opposite. He thinks we should have a few software engineers writing - not GUIs, but kind of like that; you know, report generators maybe. And there's like 20 scientist analyzing them. We have entirely different outlooks on how it's going to be. So what I'm betting on is that in the future we're going to have a lot of software engineers coming from languages like Go, and Rust, coming into biology and sort of speeding things up. + +One of the things that's fun is that every time I rewrite an algorithm from Python into Go, it's like 25 times faster by default. We have the world's fastest DNA synthesis-fixing function by several orders of magnitude, which is kind of nuts. Like, we weren't aiming for that. That's just kind of how it happened. + +So there's a lot of work to be done there, and I think the real impetus for most software engineers is the biology is intimidating. And I think the one thing if there's a software engineer listening to this right now who's interested in biology - don't be so intimidated by the biology. Seek to learn it and be humble when you ask questions to scientists, because they've gotten a lot of physicists and software engineers over the years coming into the field and acting cocky because they see this code that they've been working with, and they're like, "Oh, these guys don't know how to code. They don't know nothing." It's like, "No, they know something. It's just they're not software engineers." And you've really got to listen to them. You've got to be humble; you can't come in with this attitude that you're going to solve all their problems, because you're not. + +My favorite is -- do you guys know the three-body problem? ...like, this classic problem of physics that's been around forever... + +**Ian Lopshire:** Yeah. + +**Tim Stiles:** ...where essentially we don't have the math to model three bodies orbiting each other, like all related to each other; we just don't have the math for it. And so that's what the protein folding problem is. That's why everyone gets excited when there's a new -- I mean, people here probably have heard of AlphaFold here, maybe; maybe you haven't. But every time AlphaFold comes up with something new, it gets really exciting, because they're fighting the three-body problem with huge gobs of data and models... But I've had mathematicians and software engineers go \[unintelligible 00:34:06.15\] "Oh, that? That's a simple heat model" and spend like 45 minutes trying to explain to five biologists why they're wrong, and it's easy. I'm like, "No, we all know it's -- dude, we took freshman physics, too. We know what the three-body problem is. Like, stop." + +\[34:20\] That's the one piece of advice, if there's a software engineer who wants to get into like genetic engineering or DNA synthesis or anything out there, it's just - remember to be humble, because you'll learn something new literally every day, and you'll be like "Wow, this should be some basic stuff." But that's kind of how it goes; this broad layer of basic stuff, and then you get into the really fine details. + +**Ian Lopshire:** So you mentioned Go, and Rust, and the future there... Do you think Go has like a prominent future there? + +**Tim Stiles:** Yes, I do. + +**Ian Lopshire:** Why? + +**Tim Stiles:** Usability is the biggest thing there. And also just the DevOps tools are amazing. There's nothing that makes my life easier than writing in Go. First off, almost every tool, every DevOps tool is written in Go now. Almost every single one. For my project I regularly use -- Gitpod's a real favorite of mine personally, but... + +**Natalie Pistunovich:** That's a German thing. + +**Tim Stiles:** Yeah, no, they're great. + +**Natalie Pistunovich:** Yeah. + +**Tim Stiles:** I really love them. Oh, by the way, anyone listening out there, if you haven't been to Gitpod's Discord, go. They're super-friendly. Just show them your work, with whatever you're doing. They're super-excited, supportive and ask a lot of questions. They're really, really nice. + +But when I got started, Go is like test-first; all the things you need... Like example tests are actually a great thing, where Python or a lot of other languages -- like, maybe there's a library for it, but in Go it's just standard. Whenever I write an example test, it runs every time I've run my tests, and it's an example that never does doc rot. The problem with doc rot, with a lot of scientific software is big, because people will write these huge things, the documentation, then they'll change the API like two versions down, and then they won't change the documentation. + +And so I think the fact that Go has all these opinions, which if you're like a -- I guess you would call it like a gray beard or white beard software engineer... Like, maybe you have different opinions, but if you're just starting out and you don't really know where you're going DevOps-wise, Go's defaults are beautiful; they work. And yeah, you may struggle with generics, which Go has obviously been working on, which I think scientists will have a little bit hard time going into typed languages... They don't have that much of a hard time. But the thing that when I was originally considering writing Poly, I was looking at what could be compiled to a binary, because I thought it was gonna be more of a command line tool. Eventually, I was like "Wait, this doesn't make any sense as a command line tool." But originally, I thought it was gonna be more of a command line tool... So I was like "What compiles to a binary, what's easy to learn, what's fast, and what has like a good DevOps ecosystem?" And it pretty much came down to like Rust was a little faster, but Go won on everything else. And that's kind of what I needed. + +I don't know if you've ever done like string manipulation in Rust, but it's hard. It's not something you want to teach the people coming from Python. And maybe there's a good reason for that, but I did not personally enjoy it. So when I was testing those two, it came down to Go winning on almost everything. And I think it's also just a great step from what would be considered these scripting languages, where the syntax is easy enough, the concepts are still there. I mean, yeah, you can get into concurrency and all this other stuff, but coming from R and Python, you have to learn that anyways. You're not learning multi-threaded stuff, and what most scientists use for languages, which are -- usually, most scientists are using Python, R, MATLAB and Giulia, which - actually, Giulia does have multi-threading, a lot of modern features. But Go just has a bigger breadth, and a bigger DevOps community, and a better community of people that use it, which actually makes it really easy to find odd functions that I need. + +One of my contributors convinced me that we need to use \[unintelligible 00:37:33.08\] is the default, and so Go's crypto library doesn't have that yet, so some guy wrote a working implementation and we just used that. Lua didn't have that, so my friend who wanted to implement it in Lua had to go and write BLAKE3 three by hand. He's like, "I don't understand this. It's all a bunch of math and single-little letter variables." I was like, "So now do you realize why I make you write all your variables as whole words?" He's like -- + +**Ian Lopshire:** Grumble-grumble... + +**Tim Stiles:** \[37:59\] That's something that -- some of my contributors are like "Why do you make us do whole words?" I'm like, "You'll thank me in six months." And Go has all these great tools for detecting like data race conditions... And that's something I learned recently. I'm still learning various things... There's a tool where you can find unnecessary conversions, and there's code coverage tools that are just native, and it's all easy to throw into like a GitHub Action, or any other CI/CD thing you're doing. And so for me, it's this thing where there's all these wheels that are built-in, then this community, people that really care about how usable and shippable their code is... Which makes it just so much easier to integrate their stuff into mine. + +So a lot of scientists, they heavily rely on Docker and Jupyter Notebooks. Like super, super, super heavily. And usually, like, "What do you do to containerize your software?" I was like "Well, Docker is also written in Go, so I just do what they do, which is ship binaries if it's an application. \[unintelligible 00:38:49.23\] or CI/CD thing that puts out every architecture and operating system that it possibly can. And I think the only way -- like, if you find a system that can't run that, you're advanced enough that you can figure out how to package it yourself. Like that's it. If I'm doing -- was it Fedora, Ubuntu, Arch Linux -- like, you have enough. + +And so I just think as a programming language, it's just really powerful and very suitable for people that are trying to build something stable, that may not have had that experiences, or -- you know, people in biotech have this insecurity of feeling like they're not real software engineers. That's not necessarily true, it's just sort of they're working from this corpus of work that is, you know, as I've said earlier in the show, since the '70s. There's a lot of legacy there that we're still dealing with. And part of this project is kind of jumping over the legacy and going to something faster and easier. + +The stuff I've written has mostly been for DNA engineering, but next I want to get into metabolic pathway engineering, and protein engineering. That's sort of the whole reason I've gotten into this, is I want to make -- proteins, for those listening, you've probably heard this in the bodybuilder terms, of "You need enough protein to build muscle mass." But proteins, believe it or not, \[unintelligible 00:40:07.29\] They're actually nanobots. Believe it or not, you are the gray goo. There's all of these proteins that are coded in your DNA, that are expressed all the time, that do little functions, from moving DNA from one part of the cell to the other, to helping the cells split, that are super-specialized, and they even -- like, have you ever seen like a bunch of magnets when you throw them on a table, and they kind of just \[unintelligible 00:40:25.15\] all together? That's kind of how protein folding works, in a little bit of a way. They're all like little molecular magnets folding together and doing this thing. So what I'd like to do is I'd like to make the nanobots. There's a lot of really interesting applications for energy for, for plant lifecycles, for almost anything. I mean, they're nanobots; what can you do with these? + +And so I'm really excited about that, but I'm also excited about the macro stuff; the stuff that we haven't touched on. Traditionally, with genetic engineering, the moneymakers are first pharma, then ag tech, and then other. And I'm really, really, really interested in the other part. I think that's the coolest part. Like, can we make plants as infrastructure? You may have seen the Glowing Plants project, which is -- + +**Natalie Pistunovich:** Like mushrooms. + +**Tim Stiles:** Yeah, like the mushroom bricks that you may have seen. Why can't we just have the mushrooms grow to house size? Is there some sort of fundamental physical limit there, or have we just not tried hard enough yet? I saw a great clip on Twitter where someone was talking to a group of kids, like "So are you working on --" I think the kid asked "Can we make elephants fly?" And they're like, "I don't really know." He's like, "What, so you're just gonna give up?" And it's like "Yeah, why don't we try?" I'm a little more squeamish about working with whole organisms, like big -- the kind of organisms that vegans would not want to work with. I get -- oh my God, mice work in biology is scarring. If anyone's out there thinking about working in biology lab, just don't do... Like, I hate to say this, because mouse work's important, and it's a big part of science... Europe actually tried to ban mouse work at one point; there was talk about it, and like a bunch of Nobel Prize winners were like "I'm really sorry, guys, but we can't -- like, that would be a huge cost to society. We can't do it yet." + +\[42:08\] But with plant work, for the most part, there's a lot more to be done there with like working in morphology, and like expressing different genes... You can actually grow plants that produce medicine, which people have been working on... My personal favorite is people really want to build infrastructure, like they want to have -- you may have seen on Instagram these trees, where people have guided routes across rivers for generations, and it becomes a tree bridge... And it's like, wouldn't it be cool if we could just plant a tree and it was just like "Yeah, I want it to go over this little thing and make a bridge." That'd be great. That'd be super-cool. And that's like the sort of future utopia that a lot of people might feel they're looking towards, is like the "other" stuff. + +But we also know that pharmaceuticals is super-important, and it's also a very meaningful thing, because we've all had families in hospitals before, so it's something that we all respect. And you know, if you're squeamish like me, I try to avoid it, but also, it's important. And the ag tech stuff - Monsanto ruined the whole -- so the reason why we call ourselves synthetic biologists instead of genetic engineers... Monsanto - their lawyers and PR team just really messed up. I don't blame the scientists at all for what's going on, but their PR and legal team - they really messed up. And that's why we call ourselves synthetic biologists now, for everyone at home... Because we didn't want to have the same flack, because we weren't doing the same stuff. + +So yeah, I think that's the future of biotech, maybe... There's a lot more to it. But you'll see like a lot more -- oh, specialized cancer therapy. Specialized gene therapy is another thing that is probably going to be in the near future. We're already starting to have it... There's a lot of immuno -- essentially, what we call CAR T-cell programming, or reprogramming, where we take a sample of a patient... There's probably someone in immunology and CAR T-cell therapy listening to this that's like "You're so wrong, Tim. Don't get this wrong." Essentially, what we do is you take the patient's cells, we take their CAR T-cells, we reengineer them to target the cancer cells that we've also sequenced, like we do this in the lab, then we reintroduce these CAR T-cell to attack the cancerous cells, and not the healthy cells. Because one of the issues with cancer is essentially your immune system doesn't know how to attack, how to specify, like, things in the body. I mean, there are some cancers where it figures it out, but the problem with cancer is that this is your own cells; they're very close to what your regular healthy cells would be doing. So if your immune system would be like "Hey, that's bad", it's kind of hard for it to pick up. And there's a couple of reasons how like the adaptive immune system works, and why that has to be... But if we can take these people's CAR T-cells, these immune cells out, and engineer them to attack cancer cells without attacking the -- that's the crux, that's the dangerous part, is if you don't engineer it right, it attacks the healthy cells, which is not good. But the idea being there's already therapies that don't involve chemotherapy, I'm pretty sure. Again, I'm not a medical expert, this is not medical advice. Please talk to your oncologist about any cancer-related needs. + +Actually, that's one of the big things about being a biologist, is you find yourself in this position where you're explaining biology and they're like "Hey, should I do that, like health-wise?" Like, "No, no, no. Go talk to a doctor. They have liability insurance, they've been trained in patient-related matters, they know what drugs do... I'm just the guy that does the biology, that gives them the tools to do that." That's the difference here. + +I've gone to the doctor, I'm like, "My stomach hurts. Is there something wrong with my microbiome?" and they're like "No, goofy, you have acid reflux disease. Take this \[unintelligible 00:45:14.01\] and stop eating spicy food after eight." I was like "Okay, cool." And they were right. And that's the thing - again, when you're being a biologist, you just have to be humble. There's always going to be another expert in some specific niche that you're working in. + +**Natalie Pistunovich:** Will I be one of those experts? + +**Tim Stiles:** You can absolutely be one of those experts. Actually, that's been a joke that I've had for -- + +**Natalie Pistunovich:** Can AI - like, can artificial intelligence...? + +**Tim Stiles:** Oh, AI. I thought you said "I, personally." I said YOU can be -- + +**Natalie Pistunovich:** But I didn't even dream of going that direction... \[laughs\] + +**Tim Stiles:** No, you can do it. I'm not kidding. I had one coworker that's like "How long would it take me to be a master in alignment software?" I was like "Master? A long time. Second-best, maybe third-best? It depends on how fast you read, but it could be this year." It really depends on what your focus is and what you're doing. We all have our unique specialties. + +\[46:01\] But AI is this weird thing where - you know, machine learning is so obvious; everyone in tech knows it's obviously overhyped, right? It's like this thing where now we're having all these general models, which are awesome... I was writing basic genetic programming with generative models for a weird little Pokémon a decade ago, but now we have full artistic renderings of -- for example, my Twitter banner is a flying seaweed monster. I had to tweak it a bunch, and be like"Flying Spaghetti Monster, but with seaweed, and some other stuff", and eventually, I got it to do the thing. But we're at that point where we have these generative models that - it's been a long time since I've been in deep learning, so not only has this field changed a lot, but the terminology they use is changing a lot. Every five years they have to reinvent every term for linear algebra. I don't know why they do this, but they do. It's a weird thing. + +I had a similar problem when I was first learning it like five years ago. I think it was "What's a linear filter?" Isn't this just a kernel? And then the PhD \[unintelligible 00:46:53.03\] "Yeah, we just do that... I'm sorry." And that's a similar thing. Now there's all these different terms, I'm like, "Oh, is this basically a convolutional neural net?" They're like "Yeah, but it's different..." And I'm like, "Uhm, okay..." + +So with machine learning and AI I think there's a lot of opportunity to get around what have been these traditional problems in modeling, where we don't have the math for it; we don't have -- like I said with the three-body problem, we don't have the math. No one's \[unintelligible 00:47:16.13\] model three bodies orbiting each other. It's just, we haven't done it yet. But David Baker, who is a professor at the University of Washington - his lab came up with this game called Fold It, where essentially they'd get all these molecular biology students in college, and maybe high school even, to play this game where -- humans are actually pretty good at figuring out from like a basic string and some mechanical principles, like a 3D model of this DNA laid flat, how to fold it like origami into a shape that's what we expect at the end. So they'd have the flat sequence, and then the expected sequence they'd get experimentally, and they'd have these students essentially be mechanical turks for fun, to train this model to figure out how proteins fold... Which is super-useful. + +So in this case, there's some real roadblocks, where machine learning is absolutely vital, and you do need machine learning for it. But there's also a lot of people that do machine learning models, and then you're like "Yeah, you didn't need machine learning for that." It really depends. + +But the folding stuff and the molecular dynamics stuff, stuff that is molecular interactions - I see a lot of potential there. There's a lot of work that can be done there, especially with -- so there's something called ligand binding, which is like a fancy term for "We take a small molecule, we'll make it attach to some protein of interest, or something, disrupt some function, that helps cure some disease." It's all downstream... Like, if you tried to do like a stack trace of like how medicine works, instead of a bug failure, it's pretty brutal to be able to do that. We're still at the point in life where we don't really have that. The closest we get is my software, which is - yeah, that's not good enough, I'm gonna be honest. Go's stack trace cannot heal you. Sorry, that's not a good endorsement of the language, but we'll get there someday. + +But that sort of space, of figuring out what molecules interact with each other and how they fit, and how they puzzle-piece together - that's some really good use cases for machine learning, where you're just \[unintelligible 00:49:06.15\] structures against each other, you're doing - I call it sub sampling; or people call it fingerprinting, where you decide, "Oh, this is the only place where we really care about seeing \[unintelligible 00:49:15.10\] because that's the therapeutic effect. And so there's a lot of research there in that, and that's super-valuable, and I don't think there's going to be many other routes to do it. Of course, there's these classical deterministic ways, which are good shorthands, and they're probably faster, and you don't require like a big, beefy supercomputer to pull it off. But I think that's the place where machine learning is really the -- that's the real strong suit there. + +There are things with -- people have figured out how to do DNA synthesis fixing, like I do, but like with machine learning, and it's like... I don't know, mine works like 99% of the time, or something, and there's no machine learning involved... And I haven't really looked at the other people's models, mostly because I don't even think they're on Hugging Face. I mean, I've gotta give the Hugging Face guys super props. There's also a machine learning library in Go that directly connects to Hugging Face. I wish I remembered the name of it right now, because I want to give their developer a shout-out. You should have him on the show, actually. + +**Natalie Pistunovich:** We'll add them in the show notes, for sure. + +**Tim Stiles:** \[50:09\] Oh, yeah. And so - yeah, machine learning has a place, but I feel like you still need the software engineers and biologists to understand each other's fields; you can't just kind of play it off on the hopes that this mushing of matrices will get the product you want, because you still have to figure out how to set up the mushing of matrices; you still have to have some concept of the data you're putting in. And the data cleaning part, as many people listening in probably know, regardless of field, is a lot. It's like 90% of the job \[unintelligible 00:50:35.23\] engineering to make sure you can shove it into a matrix, or shove it into some acceptable format for this to work. There's so much of that still, and I think it's still super-critical, and we haven't figured out a way to generate data, or manage data in a way that works, to put it in these models correctly. I think we're getting better at it, obviously, but we're still working on it. + +**Break:** \[50:55\] + +**Ian Lopshire:** From what I understand, a lot of the innovation in machine learning in the last five years has been around that piece you were talking about; like, we're making sure the data is usable. + +**Tim Stiles:** So this is a really interesting thing about biotech that's held true, is that biotech is ten years behind on any trend when it comes to software. So in five years I'll believe it, but right now we're still in the weeds, my man. We're still there; we're still trying our best. I mean, I'm working on a project, again, where we're essentially making a big graph database server thing, just to have all these protein and DNA sequences and metabolic pathways in one place, so you can query them and sort of engineer around it... And I'm hoping that's the innovation that's coming in five years for us as a field, is that we have this same amount of data cleaning that the rest - I would call it the rest of the field that doesn't rely on heavy domain expertise, someone who has a PhD in this stuff to kind of reason about what you're working with - I hope we can catch up, because that I feel like that's just the lag. + +One of the things that I would say for a lot of people that are listening is I have a -- egotistically, I call this Tim's Inverse Law of Software Quality. The more important the thing, the worse the code. And there's some organizational psychology theory around this, but essentially, the code behind Instagram is infinitely better engineered than the code behind a NASA rocket ship. Or your Volkswagen car. It's just how it is. And there's a couple of reasons why, but essentially, if you can find some super-niche thing, like the operating system for a car, or... Maybe get the clearance to work with nuclear reactors; there's a lot of work to be done there. Nuclear reactors still run on Fortran and COBOL man; they need some help. Bank software still runs on COBOL. I have a friend whose parents came out of retirement because they couldn't find any people to program bank software in COBOL. + +And so there's a lot to be done in writing sort of these niche libraries that are open source, where people can see it... Because the thing for me is that if I didn't make Poly open source, I wouldn't have nearly as much business. People wouldn't be able to believe in me. I don't have a PhD. The proof in this field is you either have to have a PhD or a corpus of work that is obviously there. + +\[54:04\] And so I guess what I'd like to see in the future for Go developers is writing scientific-esque or heavy industry software that's critical, in Go, or any other language that's appropriate, and it's well-documented in open source. And maybe with the nuclear reactor stuff - you probably couldn't make that totally open source; the US government may have something to say about it. But I think there's some real value there, because then you start having clients come to you and go, "Hey, can you adapt this for our needs, our specific car?" And that's how I've made my career. People would be like "Hey, it's really close to what we need. Can you adapt this for my needs?" And I think that's a real path to making a living as an open source developer. Not as a startup, where you're GitLab and you're exiting for 7 billion... Or was it 11? 7 billion? More than GitHub, hilariously enough... But it's a path to making a decent consulting business and a living as an open source -- mostly open source developer. + +Some clients, they're like "Hey, I just need this, and I need to query for it." I'm like, "Cool. I'll write this pipeline, and I'll keep most of it for myself, and make it open source, and you get this one query that I promise never to tell anyone what this query was, and it's your data forever, or as long as you're still a company, or something." And that's totally reasonable to a lot of companies, because a lot of biotech software - people are starting to learn that they don't really have a business-owning software. People think that owning software is like owning land, and that it always appreciates... But no, you have to maintain it. Like, you can't just clean it up at the end and call it a day. It's like a constant -- it's kind of like surfing, it looks easy. But if you've ever surfed before... Like, I've just started, because I live in California, and it's what everyone does here, apparently... Like, you think, "Oh, the waves are gentle." Like, you get your -- I'm trying not to swear on this podcast... I grew up in New England, so that's all we do, is just speak in swears... But you just get your butt kicked by the ocean by standing still. And that's kind of what it's like to own software; the waves keep coming in, and you keep getting your butt kicked, and you're just hoping that you can figure out how to not have your butt continually kicked by these changing tides. + +**Ian Lopshire:** No, that makes sense. So we'll do one last quick question before we do the unpopular opinion... But this is all open source, and you said that we need more software engineers in biotech. How can the community get involved? + +**Tim Stiles:** That's an amazing question. So there's a bunch of issues on Poly right now, but what I truly need from engineers or software engineers right now is like - I don't know how to deploy large, scalable things. I write really efficient DNA manipulation algorithms. That's my jam. That's what I'm really good at. But now I'm getting this data problem where we're doing this system where we have at the least five terabytes of legacy data that I've figured out how to write a really fast parser for, but now we have to store it in a database... + +I'm using SurrealDB right now; I contributed to their Go client... Like a minor one, but still one. And it's been so far a great experience with them, and I think they have like a real product going on there, that's really nice, and a proper graph database... So if there's someone out there that knows how to deploy things at scale, that are reliable as a DevOps \[unintelligible 00:57:00.17\] that'd be great. + +The other thing is, I'm actually writing tutorials for Poly using Gitpod and an awesome plugin that has like 10 stars on GitHub, where essentially it lets you just put in the configs, like breakpoints for code, for the debugger... So the whole tutorial idea is that you open up this Gitpod instance and it brings you to the first tutorial, and it's just the debuggers there, and you hit Run, and you just go through each checkpoint or breakpoint and go, "Oh, this is what it looks like here. This is what it looks like here", and just keep going through that. Because I really love Go's -- like the Godoc's example of things, and that you can run them... But since a lot of this stuff is parsers, and parser-based, you can't really get everything you want out of that... So that's why I kind of made the tutorials for that. But I've only made the first one. There's like five others that I'd like to have, that I've put in like little "Oh, please, someone help me. Write it here." + +So if you wanna learn biology, like I did with this wet lab, where I was like "Hey, I want to learn biology", and they're like "We'll teach you, but you've got to make the class..." Like, if you want to do a similar thing, but on the internet, that would be super-helpful. + +\[58:02\] And if anyone's listening and you're into like learning how to do the wet lab biology yourself, there are plenty of garage labs all over the world, except for Germany. I'm so sorry, Natalie; Germany has a law against this. I wish I was kidding... But the rest of the world has these. I know, it's super-weird, super-random, but every time I've talked to someone in the community that's German, they're like "I can't even \[unintelligible 00:58:16.26\] my house. This is ridiculous." + +**Natalie Pistunovich:** Never tried... + +**Tim Stiles:** You can try... Just don't tell anyone, apparently. + +**Natalie Pistunovich:** \[laughs\] + +**Tim Stiles:** But there's plenty of these garages; they're called community labs, or biohacker spaces. They're all over the world, and they're used to software engineers just showing up and being like "Hey, what's all this biology stuff about?" And they'll take you in. They'll be nice. They'll show you the ropes. But if you want to get involved in the software, tutorials and deployment are like the big things that I really truly need help with. + +**Natalie Pistunovich:** And it's a good time to remind that at least today, when we're recording the episode, it's still October, so it's still time for Hacktoberfest. + +**Tim Stiles:** It's Hacktoberfest...! + +**Natalie Pistunovich:** You also get something for that contribution. + +**Tim Stiles:** I added a tag in there, you're right. Did you get T-shirt? What do you get for that? I've never gotten the swag. + +**Natalie Pistunovich:** I don't know what's this year's swag. + +**Tim Stiles:** I have a disappointingly small amount of tech swag... Like, I'm gonna be honest, I don't have enough. And every time I get like a Google hoodie or something, I just give it to my dad. He thinks it's hilarious. + +**Natalie Pistunovich:** So... Time for Unpopular Opinions! + +**Jingle:** \[59:13\] + +**Tim Stiles:** I like the harmony there... Okay, unpopular opinion. This is probably not unpopular to people listening, but maybe it is... But open source has always been sustainable, and I see no business difference between that and closed source software, personally. + +**Natalie Pistunovich:** Now's the time to drop that mic. \[laughs\] + +**Tim Stiles:** Really? Okay... Yeah, so that's like my hot take, is I don't see a difference between them business-wise. And I know a lot of people do, and this is weird... I don't know what -- there's something weird about open source in people's heads. When they hear I'm an open source developer, they're "Oh, you're a fan of Richard Stallman, and you don't have proper hygiene etc. You don't believe in money." And it's like, maybe like two of those things are true, or maybe one. But the point being is that there's this weird little stereotype of open source can't make money, and it's this weird thing where most people, and I think most engineers, especially when they're coming from - in my realm, of Academia and biotech software, they don't understand that there has to be a sales team, there has to be a person out there selling the software, getting the customers... And as a purveyor of the service, you are the expert; you are the person that -- they're not just paying for your code, they're paying for your expertise that's embodied in the code. That's in your ability to deploy it for them in meaningful and useful ways. + +So for me, looking at like a -- like, if you look at MongoDB, obviously, GitLab, and... What's the other one that I've already forgotten? Red Hat, obviously - they sell services; they're the experts in these things. They help people deploy it in a way that's scalable, that probably saves them money compared to them doing it themselves, and they keep it mostly open source or open core. And the point being there is that they have sales teams; they have people going out, finding customers, and doing all this other stuff. But I meet a lot of engineers who have the belief of "If you build it, they will come." And for consumer-level stuff - which is not what biotech is; biotech is very business-to-business. I mean, there are some consumer biotech things as end products, and there are some people that are consumers, like students, or people that buy the home kits, or things like that... But it's a very business-to-business thing, where like my code, being open source -- people ask me, "Aren't you afraid of someone's gonna copy your code?" It's like, "No. I'm the guy who wrote it." Sure, you could use my code, but why would you pay someone twice as much to do half as much with it as I would do, if you can just hire me? + +I think that's something that a lot of people don't recognize with open source stuff, is that it's a very viable -- it has almost no effect on their actual business model. That's what's always been surprising to me. Unless your business model is to not develop something novel... Like a lot of SaaS'es, obviously, different projects, just \[unintelligible 01:02:08.19\] open source packages put together in some sort of microservice framework that blah-blah-blah... And obviously, sometimes it's novel stuff out of that, and sometimes it's just re-skinning of other people's open source stuff, but with like a fancier GUI, or something... + +**Natalie Pistunovich:** Yeah. + +**Tim Stiles:** But if you're writing something truly novel, that no one else has done before, and you're the expert, you really have no fear of anyone copying you, because... You have to market yourself, that's the big thing. You can't be the quiet guy in the corner. You can't be like "Oh, I wrote this thing", and then you let some super marketing guy just come in and be like "Ah, this is my thing now. I made it" etc. You can't let them do that; you have to have some hold there. But I don't see much of a difference between a company whose product is open source and whose it's not; it's still the same thing for the most part. But you do get the benefit as an open source company, as people can believe that your software is good, because they can literally see it. The proof is in the pudding there. They can look at it. + +And then with a lot of biotech software, they're like "Just believe me that it's good. Please, give me $100,000 to start, before we even show you the code. Just believe me." And so it's a big difference there. And then you get into -- you know, there's obvious code quality, then you have people that are submitting pull requests and little bug fixes, little documentation fixes... It's not like you have a real engineering force behind that, but it does add up. And it does give you also some marketing power. Now people are like "Hey, have you heard about --" Word of mouth can happen there a little bit. I show up to meetups here in the Bay Area, and people are like, "Oh, yeah, you're the guy who wrote Polymerase." I was like, "Nice to meet you." + +So there is some real value to it, especially if you're not the traditional purveyor of truth, if you're not the PhD who's gone to Harvard, who's the fancy guys, fancy guys, grad students... It's a great way to show that you're the best, and you've got the right stuff. + +So I feel like there's gonna be more companies, hopefully, like GitLab, Red Hat, MongoDB... And I'm hoping companies like Gitpod, and NextFlow, sort of get to this IPO stage too, where people can see the engineer, and they can see the service, they can see the value in it, and they don't care that it's open source, because they know that it would cost them 60k to have a dev do it internally, or it could cost them 5k to 10k to hire the guys who wrote it. And I think that's just sort of - I hate to use VC terminology. I've been spoiled by the city... That's "the moat". You are the moat. You're the creator. You're the person who made this. You are the artist who brought it to life, and there's some real value behind that. + +**Natalie Pistunovich:** I agree with you very much. + +**Tim Stiles:** Really? Okay. So it's not controversial, and I've ruined the segment. + +**Ian Lopshire:** I think it might be controversial... + +**Natalie Pistunovich:** Yeah. So there will be a vote... As with every unpopular opinion, we will put this on our Twitter and we'll see how many people agree or disagree with you. But to summarize it, we said that your unpopular opinion is that open source always has been sustainable, and... Yeah, we'll see how that goes. I have to say, this must be the episode with the most show notes we will have, at least for me... + +**Tim Stiles:** Oh God, I'm so sorry... \[laughter\] + +**Natalie Pistunovich:** No, that's amazing. You taught us so much, and I want to say thank you very much for that. + +**Tim Stiles:** I wish I had a whiteboard. I'm gonna have to go and make a new PowerPoint presentation... + +**Natalie Pistunovich:** I learned one million things today. + +**Tim Stiles:** Okay. Alright. If anyone is out there listening to this, or you, Natalie, or Ian wanna ask more questions about biology, I'm always available. I may get tired of it eventually... + +**Natalie Pistunovich:** And your Twitter is mentioned here, for sure. + +**Tim Stiles:** Oh, yeah. This is the Twitter. If you jump in the Discord, there's gonna be hopefully some nerd that'll be like "I know that bit of biology." I love my Discord for that. Someone can be like "Yeah, I can explain that." I'm like, "Great. I had no idea." + +**Natalie Pistunovich:** Perfect. We'll add the Discord to the show notes as well. And Ian, thank you so much for joining, and co-learning here, live. Tim, thank you for teaching us so many things. This was really fascinating. + +**Ian Lopshire:** It was. + +**Tim Stiles:** Thank you for having me. I hope I did alright... + +**Natalie Pistunovich:** It was amazing. + +**Ian Lopshire:** Amazing, yeah. + +**Tim Stiles:** Alright. + +**Natalie Pistunovich:** Thanks, everyone who joined. + +**Tim Stiles:** Thank you very much. Talk to you later. Bye! diff --git "a/Go tooling \342\200\232\303\264\302\252\303\224\342\210\217\303\250_transcript.txt" "b/Go tooling \342\200\232\303\264\302\252\303\224\342\210\217\303\250_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..165a42e183360b587db450118280972359725448 --- /dev/null +++ "b/Go tooling \342\200\232\303\264\302\252\303\224\342\210\217\303\250_transcript.txt" @@ -0,0 +1,309 @@ +**Mat Ryer:** Hello, and welcome to Go Time. I'm Mat Ryer, and on today's episode we're talking about tooling. All those great tools that help us be successful, help us do our job. We use the Go tools all the time, every day. We use them for building, for running code, for testing, we use them for formatting our code, for linting and vetting, and many more things, too. + +I think this show will be useful to anybody new to Go that wants to get a sense of the tooling that we all use. I'm sure there will also be some golden nuggets for the seasoned gophers also... And I'm so confident because of who's joining me. I'm joined today by - in no particular order - Jaana Dogan. Hello, Jaana. + +**Jaana Dogan:** Hello, hey! + +**Mat Ryer:** Welcome back to Go Time! How have you been? + +**Jaana Dogan:** Yeah, it's been a while. I've been traveling, I guess. + +**Mat Ryer:** Yeah, yeah. Where did you go? + +**Jaana Dogan:** I was in Marbella, Spain. The last time we talked I was just going for a conference, and then I never came back to the show. I'm so sorry... + +**Mat Ryer:** That's alright. I can understand. This year you're off on travels to exotic places for work... It's a tough life. + +**Jaana Dogan:** Such a tough life, yeah. + +**Mat Ryer:** And you told me earlier that everything you do at work is completely confidential. Do you wanna just break all the rules and tell us anyway, or...? + +**Jaana Dogan:** Well, kind of... I'm actually about to switch to a new job. I mean, not a new job, but sort of like a new role. Currently, I'm still exploring what I'm supposed to do, and... It's confidential, not because it's supposed to be super-confidential, but I am not sure about what I will be focusing on, so... I think I will need a week, or something. Just don't get oppressive now. It's not about you, it's just... I'm still exploring. + +**Mat Ryer:** Yeah, absolutely. Obviously, I do take it very personally, but I'll pretend that I don't. Well, also joining us on today's show, it's only Johnny Boursiquot. Hello, Johnny! + +**Johnny Boursiquot:** Hello there. + +**Mat Ryer:** And speaking of new gigs, you've just started yours, haven't you? + +**Johnny Boursiquot:** Yeah, recently. It's been a couple weeks. I'm still on-boarding, as they say. But yeah, it's still exciting, I'm still looking forward to contributing and learning... New gigs are always exciting that way; there's that honeymoon period where everything is new, and you're learning; you're learning about the systems, and the people, and all that good stuff... And then at some point I'm gonna cross that threshold, I'm like "Ugh, what is going on...? I need to start fixing things." But so far everything is going well. + +**Mat Ryer:** \[06:01\] Oh, good. I'm glad to hear it. Yes, it is exciting. It's scary and exciting all at the same time, new jobs. But yeah, I wish you all the best. If you don't mind, we'll keep asking you about it on the show, because I'm very interested; I think it's useful for other people as well to hear about things that we get up to in our professional lives... So if you don't mind, I'll keep bugging you about that. + +**Johnny Boursiquot:** Sure thing. + +**Mat Ryer:** Let's jump straight in. We're gonna be talking about Go tools today. I asked on Twitter earlier which of the Go tools are people's favorites, or which ones do they like the most... I'll just kick off - mine probably has to be `go fmt` (go format). For those that don't know, it formats all the Go code so it looks the same, and all the rules are baked into the tool. So you don't get to choose tabs versus spaces, you don't get to choose where the braces go, you don't really get to choose a great deal about the actual format of your code... Which again, I think to some people when they're used to having tools that allow them to configure all this, they feel like that's a deficiency in Go... But it turns out to be one of Go's super-powers, in my opinion. Because what happens is all Go code starts to look the same, and starts to look familiar... And I've done it where I've been to a project and found that the code just looks like I wrote it, and I definitely didn't. I think that's awesome. + +If you think about pull requests, with white space... Sometimes pull requests having loads of white space makes it really difficult to see what the crux of the change is. Well, with `go fmt` we don't have that problem, because it's all formatted nicely. + +Anyone else? How do you feel about `go fmt`? How do you pronounce it, by the way? Let's just get that one out of the way. + +**Jaana Dogan:** `go fmt`, right? + +**Mat Ryer:** Okay, good. + +**Jaana Dogan:** I mean, that's what I know. \[laughs\] + +**Mat Ryer:** Yeah, that's what I'm told. + +**Johnny Boursiquot:** It's always awkward when I'm teaching something and I say "the fmt package", for example... People kind of look at me sideways; I'm like "Yeah, I know. I know. Just go with it." If you say "fmt" or "format", God forbid, gophers are gonna look at you a little weird, like "We did just go with it." + +**Jaana Dogan:** Yeah, it takes a while for people to parse it initially, and then they learn it and they take it and they don't question it... I'm trying to keep it consistent by saying "`go fmt`". + +**Mat Ryer:** Yeah, same. I agree. I've done that, I don't think naturally, but I heard about it, and yeah, I do it for consistency, too. + +It's funny, because sometimes people will say "golang", because when we use Google and we search, or when we use hashtags, we tend to write "golang", but we never say "golang." So it's a little pro tip for anyone that's new to the Go community - when you're talking about the language, just call it Go. Don't say "golang." Same with "fmt". + +**Johnny Boursiquot:** With regards to `go fmt`, the reason -- well, let me step back for a little bit. When I first came across `go fmt`, I was taken aback, honestly, because I wasn't used to basically tooling, formatting my code to look in a standardized way. I come from programming languages where everybody has their little pet peeves, their little quirks about "I like my braces to be lined up together." Another person might be like "I like my braces to end at the declaration, and then for the closing brace to be at the end", or whatever. + +\[10:04\] People would have these back-and-forths around styling, what's more readable versus what's not as readable... And obviously, it was all sort of subjective. Everybody has their own preferences, their own quirks and what they're used to and what they're not used to... But `go fmt` sort of threw all of that out the window when I first came across it. + +I'll be honest, for the first month or so, I was like "I don't like everything about what it does. I'm happy with 90% of it, but I don't like everything about it." But then as time went on, I really began to love the tool and what it does. The beauty of it - I think you touched on that - is that every Go code started looking like I expected it to. Basically, that cognitive load, that aspect of looking at code and reviewing code - that just went out the window. I didn't have to worry about "Okay, is this person's code gonna look different formatted than this other one?" Basically, I could just focus on the actual code and what it was doing, as opposed to trying to figure out, parse in my head, "Okay, this person's quirks are that way, and that person's quirks are that way", kind of thing. So it was valuable in that way. + +**Jaana Dogan:** Yeah. There's actually something from Robert Griesemer that he used to say... He's the person who's maintaining `go fmt`, and all the rules and so on. He says that he doesn't agree with all the styling; he doesn't necessarily agree with `go fmt`, but some tool is enforcing it, so there's no question. + +I work for a very large company and I witnessed -- it took like four years to just tweak one little style guideline change on the Java style guideline. And can you imagine - there's all these hundreds of people with strong opinions about style just wasting four years debating on minor style issues... I like the fact that there's this `go fmt`, this canonical place, and there's no debate; there's one source of truth type of thing, and everybody has to agree with it, even if the formatting is not always what you desire. + +**Mat Ryer:** Yeah. Do you think they would be able to retrospectively fit that into the toolchain? Say that there wasn't `go fmt` originally, and it just came out now... Do you think the community and everyone would rally around it in the same way, or do you think there's something to be said for the fact that this was there from the very beginning? + +**Jaana Dogan:** I think it's necessary that initially you create some initial culture around -- you know, just relying on a tool... Because I think it creates this community with enough people supporting an idea, and understanding why it's valuable. If you try to inject this type of tool at a later time, the community is already fragmented and there's a lot of excuses to prefer a personal style, because you already (for example) invested in one particular style all across a company, and there's no way to just fix things at a later time. So it's really good that they came up with that tool initially... At least that's my opinion. + +**Mat Ryer:** Yeah, I agree with you. There's a few examples where the foresight or the insight from the team in the original design -- I think we really benefit from some of those decisions, and we'll talk about more of them as well. I think the fact that another one of the tools, go test - that was there as well, from the very beginning. So testing as a concept was a first-class concern in Go. And that of course makes sense, because at the time it was being designed that was how we were building software. We were writing tests a lot, it was an important part of software engineering... But the fact that they make these decisions early just sets a precedent... And yeah, from there it pays dividends every day. + +**Jaana Dogan:** \[14:04\] Yeah, I think Go is doing a good job in terms of identifying 80% of what is essential in software engineering, and I think tooling is also representing those priorities. + +**Mat Ryer:** Thinking beyond `go fmt` then, if we look at Golint and also Govet, does anyone wanna have a stab at describing the difference between those two, or describing what they actually do? \[pause\] Cool. Well... \[laughter\] Yeah, so Golint - I like it. It essentially looks at your code and does some static analysis, and can catch common mistakes and can give you warnings about them. And sometimes they're not mistakes, but they're just best practices. You can run the lint tools on your code and see if it's got any recommendations for things that you might change. + +One example is if you have something in a package that's exported, if it starts with a capital letter, then you should have a comment on that, really. That's the accepted practice. Now, the Go spec doesn't say that, so it's not a compile error if you don't have a comment there... But the Golint tool will catch it and say "For maximum quality, for the best quality, you should consider putting a comment here." + +And there are a few rules around how we write comments as well, where we repeat the name as the first word in the comment. So there's a few little things like that that are encoded in the linter, right? + +**Jaana Dogan:** Yeah. Well, we need to mention first, I think there's a difference between that and lint. That is reporting more of like suspicious stuff, and some patterns that might be a misuse of an API that may corrupt some memory, or whatever. Think about the typical example of printf - if you pass the wrong type of arguments, that is going to complain about it. + +So lint is more about style errors. For example Godoc, a public API, is going to complain about that type of problems. So that became a part of the test, but not all the things that is reported as a part of that is genuine. There could be false positives, as far as I know. And it also applies to lint as well. These are not part of the compiler because there's some reports that are not accurate, or something... But they generally generate genuine enough reports, and they're really useful. + +**Mat Ryer:** Yeah, you're right. If you use printf or wrapf, if you use one of those f methods and then you don't put the correct number of verbs/arguments in, catching things like that is extremely useful, because it's quite hard at a glance to just see those kinds of mistakes. So yeah, I think people should switch on those tools for their codebase, at least run them for their codebase and see what kinds of things it is actually saying... Because you might find you agree with them. + +The comment one is a good example. It's quite dogmatic. It just says "Okay, it's exported, so it needs a comment." Now if that function says "new thing", then it's obvious that's making a new thing, and your comment is probably gonna say "new thing makes a new thing." So we have a little bit of redundancy, but I think generally speaking, if you do follow the lint tools, I find that the code (again) starts to look more familiar, and you get all the benefits of `go fmt`. + +**Johnny Boursiquot:** \[18:22\] One of the things that I typically do, and partly the reason why for me, off the top of my head, differentiating between the linting and the vetting was I was like "Hm, I guess, I've never really talked about the difference that much, because they're part of my toolchain." In my day-to-day I use a lot of VS Code and Vim as my editors of choice, and they have plugins and extensions built-in; that's part of my workflow. So every time I hit Save, these tools are running, and I'm getting different markers at different spots, from different tools. + +There's another popular open source project out there that I think is called gometalinter, which includes a bunch of those tools, as well. You can configure it, turn some off and others on, and whatnot. But these tools together, they give you a set of outputs that you can basically go through and figure out "Oh yeah, I use the wrong verb here. I'm supposed to use an integer and I'm using a string instead." Things that the linter and vet would find for you if you ran them individually; but because they're part of my toolchain, basically I just look at the view at the bottom of my editor and get a list of things that I go and fix. + +I almost don't care, I should say, which tools give me what, unless I really need to work with a specific tool... But it's part of my workflow, they're just a part of my editor, and everytime I hit Save, formatting gets done, Goimports does its thing, whatever I'm referencing in my code that is not imported - it brings that in automatically. All these things happen. The tooling makes it easy to just focus on writing the code and not worry so much about having to run individual tools, one-at-a-time kind of thing. + +**Jaana Dogan:** Yeah, it's a good point actually making it a part of the everything experience; it's really useful. Especially vet is reporting a lot of useful stuff, like "Hey, this is unreachable", or you're passing for example unmarshal a non-pointer, and stuff like that... It's so hard sometimes when you're typing, and when you're just like coding, but tools are sort of helping you to do the right thing as you are programming. + +**Mat Ryer:** Yeah, and I extend that to running tests as well. I tend to write unit tests, which run very quickly, and then you can run those everytime you save the package, usually. If they start getting too slow, then of course you have to have a different strategy, but certainly in the beginning, if it's unit tests, they just run very quickly... And the build time in Go is still phenomenal. We always kind of forget about it, until you have to go and build a different codebase; then you appreciate it again. + +And Johnny, by the way, the gometalinter now apparently is called golangci-lint. So if you want to install that into VS Code, it's golangci-lint. That's the new name of that. + +**Johnny Boursiquot:** Interesting. + +**Mat Ryer:** But you're right, it's the metalinter - it runs a range of other linters, and kind of gives you that one view of it. And they integrate brilliantly into the IDEs as well. So that's the other things, like you say... You can run it on Save, but even if you don't, you can still usually integrate it into the IDE in some way that just makes it part of your routine... Because you know, anytime you can get that live feedback from the code, that's valuable, because usually, as you're working, you learn, too. And that's a great way to learn things, as you're writing code, to see a linter saying "This is unreachable now" or "That things's over there now." And if it's tests too, then "Oh, these tests are broken over here that you didn't expect." You just get that feedback from the code which is so useful when you're working. + +**Johnny Boursiquot:** \[22:25\] And you shouldn't have to wait until -- if you have continuous integration (CI), you should; but you shouldn't have to wait until that code reaches that remote server, where all these tools are run, for you to get that feedback. It's much easier and much faster. Like you're saying, that feedback loop is much tighter when it's part of your tooling. So there's some things you can do locally to make sure your code is fmted, it's vetted, it's linted, and all that good stuff. + +Then when it goes up for review, for a PR, whatever CI tool you're using - Travis, Circle, whatever; there's dozens of them these days - they give it a blessing, and now people can just focus on what does the code do. They don't have to tell you "Hey, you forgot to run `go fmt` on it", or something. You take advantage of these tools locally. They're very good tools, so I wholeheartedly encourage folks to make them part of your development workflow. + +**Jaana Dogan:** Yeah, one of the best parts is they are really fast also. It's part of the editing experience, because they're fast. I'm coming from a background where I used a lot of Java tools, and it's not like it's a smooth experience. We used to have similar static tools, but it was not as smooth as all these Go tools. Nobody is making it optional because it doesn't really make the editing experience more challenging, because they are fast and they are useful. + +**Break:** \[23:55\] + +**Mat Ryer:** We mentioned go test. That's another tool that we use a lot. Anyone that's not used it - if you write test codes in your Go programs, you do that usually by naming the file with \_test.go at the end, and then you run go test. It will look through all those test files and it will actually run all the test code for you, and that's really how if you do TDD, you know that your code is fulfilling its promises, it's doing what you said it was gonna do. + +There's also another little feature in the test tool which I think gets overlooked a little bit... It's the race detector. When you're writing concurrent code, it's possible for you to break the rules and try and read and write from the same data at the same time; if you try and do something like that, that's illegal and it will crash the program. But of course, it's very difficult to see that sometimes if you've written the concurrent code, and certainly difficult to write tests for it, because sometimes it might just not happen, just because of the way that things get scheduled... But there is a race flag which you can pass into the go test; it's a bit slower, but it does some additional checks, and you can catch those potential deadlocks early... Which is kind of cool. + +**Jaana Dogan:** Yeah, the tooling is also part of the standard tooling. It's not just a test, but it's a really good edition \[unintelligible 00:27:00.05\] is also a part of the tests, because we all have this workflow of not merging things if the tests are not passing... So you would ideally wanna enable the race detector as a part of your CI. + +And it's amazing, but there's one thing I think we should mention - your tests should cover concrete cases, so that the detector can detect them. If you don't represent those by concrete situations, the detector won't be able to detect them... But it's amazing, because it's just so on point, and it's easy, and it's a part of the standard tools, so you don't have to figure out all these additional, extra tools in order to get the benefit. + +**Mat Ryer:** Yeah. Now, it's worth saying that the race detector - if it reports that there's a violation, then that is a violation, but it doesn't necessarily catch everything... Isn't that true? + +**Jaana Dogan:** \[28:01\] Yeah. + +**Mat Ryer:** Okay. But it's still -- to be honest, I've never seen a race condition get through after testing it with... + +**Jaana Dogan:** Because you are actually good, in terms of like you care about your tests, so you represent all the cases, probably... I've seen a lot of times people are just not creating those situations where concurrency is a problem. They have all these super micro-tests, so they don't really capture it. And I think it's really important to tell that your tests should represent those cases, so the race detector can detect them. + +**Mat Ryer:** Yeah, that's a really good point. Well, with TDD you tend to get good coverage. Code coverage, by the way, is also another part of the tooling that we just get for free, which is awesome. But yeah, I never try and shoot for 100% code coverage or anything, but naturally, it's quite high with TDD, and I suppose naturally you'll also cover a lot of those cases that you talked about, as well. + +I like go run, as well. You don't tend to have much magic in Go, but go run is probably the magic tool, because it actually secretly does a build, and then executes... It does a few steps behind the scenes. But it's great if you're just learning to code, or you just wanna write a little script quickly and just execute a program. You can use go run and you pass in the name of the file(s), and it just runs it. It builds it to a temporary directory, and I think it gets deleted afterwards; I'm not sure. But I think that also is such a nice thing, to be able to just quickly see results and see feedback from what you're doing, and go run is another example of that. + +**Jaana Dogan:** Yeah. I think people use go run for their first Hello, world program. It sometimes becomes complicated... They have this habit of using go run -- I think before GOPATH was a little bit more difficult to rely on; go run was able to work outside of GOPATH, so the behavior of go build and go run was not quite the same. People have been advocating to always rely on go build or install, rather than go run... But I think it's just really nice for a Hello, world, or if you have a script type of thing that you just go run. It's really useful. + +**Johnny Boursiquot:** You were right, Mat, when you were saying -- from my understanding, it does the same thing as go build, the difference being that once the programs run, it just discards that temporary artifact. At least that's the high-level of what I think it does. + +One thing that's worth mentioning also is you can run it with -race as well. That way, if there's any sort of race conditions in the code, when the program fails, if it panics, then you'd actually get some information around where that race condition occurred, as well. + +**Mat Ryer:** I didn't know that. That's brilliant. + +**Jaana Dogan:** Yeah, I think race is supported in test, build, run... Generally, all across the tools. + +**Mat Ryer:** Hm. But it adds overhead, doesn't it? And it slows down your program, and things. It's not something you would just always switch on. + +**Jaana Dogan:** Yeah, that's why I think it's useful to make it an optional thing for tests. But apart from that, you don't wanna have the race detector always on. + +**Mat Ryer:** Yeah. + +**Johnny Boursiquot:** I've had mixed results, depending on the size of the codebase, obviously... But these days I'm working on small codebases. I've been working a lot with microservices, that kind of thing, so these codebases tend to be somewhat small, relatively speaking... So by default, I use the make command. \[unintelligible 00:32:10.06\] run the test with a -race flag. + +\[32:18\] I haven't noticed significant slowdown in that, but again, obviously, it might vary, depending on the size of your project and how many things you've got going on. + +**Jaana Dogan:** There was a benchmark about this... I think it was kind of like memory usage is like five times larger if you have the race detector on. And I think execution time-wise - again, there were some reports, but it's really depending on the use case, let's just say... It's kind of like adding some overhead, which could be 2x to 20x, or something, if I can remember the numbers correctly. There's a really good blog post actually, or an article on the golang.org about the race detector, and there must be some numbers over there. + +**Mat Ryer:** Yeah, cool. Okay. I was thinking as well about go get. Go get is another one of the tools which I think -- obviously, things have changed a lot, especially in the module space... But I've gotta say, when I was first using Go, to just be able to install packages by saying "go get" and then the package name... And then for that package name also to be the import path and to be the URL of where that package lives - I've found that to be such an elegant thing, that it was very easy to install things. This is in a GOPATH world where everything just gets put into one place, but go get just really made that very easy. How do you feel about go get versus the new module tools? Because working with modules is a little bit more complicated. + +**Johnny Boursiquot:** I'll punt the modules to JBD and let her tackle that, but I can tell you when using go get, especially when I'm teaching, being able to say "Look, we're gonna import this package. Before we can actually import this package and use it in our code, we need to go get it." So I'd literally say "Okay, go get", and then basically I'd find the name of the package \[unintelligible 00:34:26.24\] github.com, or wherever the public repository is... And then basically I'd get this blank stare from the students; they'd be like "Okay, what just happened?" + +Then it dawned on me that if I literally copy that path, go into the browser and paste it in the URL bar and navigate to that repository, immediately they were like "Oh, okay. I see what this is. You are literally pulling this code, that lives at this very path; you're putting it on the command line. You are pulling it down." Now I can actually see and read that code in my browser and see what it is I'm actually pulling down. So the whole thing about pulling down the package, it goes in your GOPATH - none of that stuff makes sense for them, but the moment they can actually go into a browser and put that very path in, it sort of clicked. Now they understood the value of go get, and it didn't really matter much where it was being put in the GOPATH. The fact that they knew how to get it, they knew where to go and see whatever was being pulled was almost magical for them. + +**Mat Ryer:** Funny, because it's no magic, and the fact that it's so obvious, i.e. "That's the URL. Go look at it. You know what a URL is." I think that's great, and the little story you've just told makes total sense. If I use some npm stuff for a project, I install a few things and I look in that Node modules folder - there's 16 million folders in there. \[laughter\] And I don't know where they've come from... It's kind of hidden. It is magic. Whereas that thing of being very simple and clear, even if you sacrifice some features for that - I think that's such a positive dividend that it keeps paying again and again later. + +**Jaana Dogan:** \[36:35\] I think we need to make an episode on Go mods, but I agree that go get is a really good initial experience. One thing I like about it - if you're go getting a main package, it installs it, it puts it in your GOPATH /bin directory... So it's just a good way to distribute tools, as well. Before Go, I was just publishing binaries, and making sure that I have the right version all across. Versioning still is a problem with go get, but I think it's an okayish sacrifice. + +**Mat Ryer:** Yeah. What I'm gonna do is just keep moving on to different Go tools, because I'm already learning things about these as well... The other one with go build which I love is the fact that we can do cross-compilation. Now, this has been around from (I think) the beginning. + +**Jaana Dogan:** Yeah. + +**Mat Ryer:** Essentially, for those that don't know, you can choose the target architecture, the target machine to build your Go code for. That's very useful if you're using Docker, because on a Mac you can do the build for Docker, and then you've got the Linux binary that you can then put into the Docker image. Or you can, of course, put the code into Docker and build it in there, in that environment. How's your experience with cross-compilation so far? + +**Jaana Dogan:** I think it was magic. When I first saw -- they were typing GOOS (it's pronounced goose) and Windows and go build, and you'd get a Windows binary... It was like "Whoa!" It was fascinating. I usually generate binaries for Linux, so I kept working on my Mac without any worry, or anything. It was so awesome. + +**Mat Ryer:** Yeah. Have you used it, Johnny? + +**Johnny Boursiquot:** Absolutely. One of my first jobs using Go full-time, my first responsibility was to have a multi-platform build process. I relied on GOOS and GOARCH quite a bit. And for those of you who don't know what GOARCH is - that's the companion to GOOS, for Go architecture. Using GOOS and GOARCH were sort of bread and butter for having that work done, and being able to push out binaries for all kinds of different platforms. + +There are a ton of them that Go supports out of the box. For ARM processors... The sheer combinations you can have - I've lost track of all the different variations you can push out. It really was a godsend. There's no way I would have been able to get that job done without these things being in there. + +**Jaana Dogan:** I think it's also awesome that -- I was doing a lot of development for ARM, for Raspberry Pi, for example... The processor on a typical Raspberry Pi is going to be not comparable to my laptop, so I would just build things on my laptop because it's going to be faster, and then I'd push it to the Raspberry Pi, because it was just so easy to do cross-compilation. It's maybe ten times faster, or something. + +**Mat Ryer:** \[40:04\] Wow. So how does that actually work? Because obviously, the compiler is doing a few steps, and then it ultimately creates a binary that's made up from the machine code. Is it just that the machine code is generated differently, depending on the architecture? + +**Jaana Dogan:** Yeah, you know - they know what to generate for each architecture, so they just basically take the input, they know what to map it, and then they generate the output based on the operating system and the architecture. + +**Mat Ryer:** That must have been possible because of the way that they built the tool system. Do you think it was deliberate that they wanted to be able to build it to any target architecture, or do you feel like they just realized they could after, because they'd just built it and designed it in a simple way? + +**Johnny Boursiquot:** I don't think you stumble on something like this by accident. If I had to guess, I'd say this was by design... Considering that the creators of the language -- basically, they were building for Google, so I imagine that at some point they needed to be able to run binaries on different platforms, for different CPU architectures; 32-bit versus 64-bit, all that good stuff. So I imagine this must have been part of the plan, part of the design. This seems way too complicated and way too powerful a feature to just come across, to have fallen out of the build system. + +**Jaana Dogan:** I think we simplified the process, but there's this intermediate assembly... The compiler first translates everything to that intermediate assembly, and from that point on they are being compiled to the architecture's specific instructions. So the internals of the compiler is this two-step thing... And this is a really typical way that compilers work. They're just taking it and converting everything into an intermediate language, and then from that intermediate language you can just basically target whatever architecture you wanna target. + +**Mat Ryer:** And of course, you can have build tags, as well. Does anyone wanna describe build tags? + +**Jaana Dogan:** Yeah, build tags are providing conditional compilation, and you can create different rules. For example, you can have constraints to say "Only use this file for Linux builds." Or you can say "I just want only ARM builds to have this file included in the build." There are many different rules provided by the toolchain. Goversion is one of them, arbitrary custom build tags is one of them... So it gives you this possibility to switch to different implementations depending on the Go version, depending on the operating system or architecture, or some custom build tags. + +**Mat Ryer:** Yeah, I've used those successfully when it comes to testing. Sometimes if there are long-running tests, or if there are integration tests that require a different dependency to be running or something, I use a build tag in our test files. That's quite an easy way to choose a subset of things to run. And it's just a special comment that goes at the top of the file, isn't it? + +**Jaana Dogan:** Yeah, I think it's on the top of the file, there's a particular place... But that's it. And it's really readable. I think my only complaint about these rules, about the build constraints, is that it's just really hard sometimes to just have multiple rules represented. It becomes really hard to parse. If you wanna have more complex rules, like "Hey, include this file on Linux, Darwin and blah-blah, but not on this particular thing; on top of that, not for this custom build type." I think expressing those more complicated/complex type of constraints is a little bit hard... But otherwise, I think it's just pretty straightforward, and I use build tags all the time. + +**Break:** \[44:18\] + +**Mat Ryer:** Okay, I want to also mention a couple of tools from the community, as well. Because remember, we are using Go tools all the time, but we can write tools as well, and some people have contributed. I think goimports was a Brad Fitzpatrick project; that was his own idea, that he just did on his own... It essentially wraps `go fmt`, so you get all the formatting, but it also resolves imports for you. And you can do these things too with your own tools. + +Some of the tooling as well doesn't have to be Go tooling running on our machine. Matt Holt has a great JSON-to-GO service. If you google "JSON-to-GO", you basically paste in a JSON blob, and then it generates the Go structures for that JSON blob. Extremely useful, especially if you're gonna consume an API and you need all of the data, so you just don't wanna sit and type out all the field names. So that's a very useful one, and that's a hosted website, so you can go to that. + +Are there any other community tools that we like? + +**Johnny Boursiquot:** I personally like the Go Report Card website, which -- well, I guess it's less of a local tool, but something that can basically evaluate how close to the idioms of the Go community your code is being kept at. I think it might even incorporate some of the tools we've mentioned before - the linter, the vetting... And it includes some other things, like cyclomatic complexity analysis, and there's a bunch of other nice adds in there. Based on these things, it gives your repository a grade, I think on a scale of A through F, or something like that. + +I find that useful, especially when I'm evaluating a repository, a third-party package, to determine whether I'm gonna use it or not. If it has a score, I will look at that. If it's anything other than A, then I'm gonna take a closer look; I'm gonna be a little bit more hesitant with bringing it in, because I'm like "Okay, what best practices, what idioms are you not following?" So I'll take a look at that. + +\[48:03\] Sometimes I may just see what's happening and maybe replicate it locally, without having to bring in the package, if I don't like the score, so to speak. So it's sort of another data point, so to speak, to help you evaluate the quality of the repository. But yeah, it's one of the things I like to see, as well. + +**Mat Ryer:** The same for GoDoc. GoDoc is a tool you can run locally, but we have also the godoc.org hosted service, which lets us view documentation for any open source project. I think that's also nice. It's a nice way to provide that capability, because it makes sense; yo want to share just a link. + +The nice thing for GoDoc - it's just godoc.org/pkg/importpath. So again, you're still referring to that import path, and we see it. + +**Jaana Dogan:** I personally use a lot of tools from Dominik Honnef. He has this go-tools repo, Staticcheck tool, which contains a lot of style check, a lot of linting type of features that Golint doesn't support. There are some cases sometimes - like, there's a controversial style topic; it's not possible to merge it into the official tool, so people would just go and put it in the Go Static tool. So it's a really useful tool to take a look at. In terms of static tools like that, I just rely on Staticcheck more than Golint. + +**Mat Ryer:** Yeah. And Fatih Arslan - he made a service which I think is called Fixmie. It's kind of a GitHub integration, as I understand it, and it analyzes -- it does a bit like the Go Report Card, but it actually creates PRs with changes in it. It's sort of proactive, like you've got another member on your team -- like the pedant, who just cares about all the style rules, and all that... That's a project that I think is worth checking out. It's Fixmie. Have a look at that one, too. It's a similar kind of idea to the Go Report Card, but more tightly integrated into GitHub. + +Has anyone here written any tooling, static analysis or otherwise? + +**Jaana Dogan:** I only wrote some tools to generate some stuff from an interface -- well, these are also some static tools... One common case is generating implementations of interfaces, and there's a lot of boilerplate, so I wrote a tool that takes the interface and creates the concrete implementation, and then you just go and fill the implementation, fill the methods. + +**Mat Ryer:** And did you use the AST stuff in the parser, and things, to build that? + +**Jaana Dogan:** Yeah, I used whatever there was in the standard library. It was not that hard. It was not that good-looking either, but it was possible to get it done in 100 lines, or something. + +**Mat Ryer:** Yeah. Okay, cool. Well, I think we should also spend some time talking about some of the performance tools as well, that we just get for free. There's some great talks on YouTube... It's quite an interesting subject, and it's talked about quite a lot, and from different angles. Jaana, perhaps you could tell us a little bit about -- did I see you do a talk about the performance tools? + +**Jaana Dogan:** \[51:56\] It might be possible, because I worked on some of the dynamic tools when I was working on Go, so it was part of my full-time job... And I generally have been working in this area for a while... So it's possible that you have seen me giving a talk, but I can't remember, because I'm giving too many talks nowadays. \[laughs\] + +**Mat Ryer:** I thought it was all confidential what you work on. + +**Jaana Dogan:** The confidential stuff is different than this. + +**Mat Ryer:** Oh. What's that? + +**Jaana Dogan:** It's none of my performance tools, it's more about computing products... + +**Mat Ryer:** Yeah. + +**Jaana Dogan:** We'll figure it out in a couple of weeks. \[laughter\] + +**Mat Ryer:** Yeah. I'm just trying to be like one of those hard-hitting journalists that tries to get out the information that you don't wanna say. \[laughter\] But I'm just too polite. You just say "I'm not gonna talk about it", and I go "Oh, okay. Bye..." + +**Jaana Dogan:** Well, the problem is I really don't know. Like, I know generally what I'm going to be working on, but I don't know the specifics, and I'm a really precise person, I think. I don't wanna give any impressions that I'm going to work on something that I'm not going to, because people will get upset. + +**Mat Ryer:** Absolutely fair enough. + +**Jaana Dogan:** Just joking, yeah. + +**Mat Ryer:** So Jaana, could you tell us about some of these tools and what they're for, for anyone that doesn't know about them? + +**Jaana Dogan:** Yeah. Generally speaking, let's go beyond the performance tools - there are a lot of dynamic tools in Go, and they are part of the standard tooling. Some of them are related to performance, some of them are more like debugging type of stuff... + +We can talk about, for example, performance initially... When Go first came around, it came around with some of those dynamic tools, because we went to the SRE team and the SRE team at Google is just really specific about what they wanna put in production. They wanna have enough visibility into things, and some of these were related to performance. + +They wanna be able to get the profiles, they wanna get some runtime traces... Because they specifically wanna be able to understand when there is something going wrong, and they wanna be able to pinpoint to those. So pprof support was baked into Go since the early times, because of that requirement, for example. It provides you some profiles, you can also add your custom profiles, which is a useful topic... But it provides the CPU profile, memory profile, goroutines, thread profile and mutex contention profile. It was really crucial to have a language mature enough to put in production, because basically -- most people think \[unintelligible 00:54:46.15\] is about development time, but it's also important in production time. + +On top of pprof support, there's good benchmarking support baked into go test. So benchmarking is a first-class citizen in Go, which is not really quite the same situation in other languages. I think it kind of creates this culture where you care about benchmarking stuff. I don't know what is your opinion on this, but I've seen lots of different communities have different opinions about benchmarking just because of the tooling, or it's really easy to write benchmarks or not. What do you think about it? + +**Mat Ryer:** Well, I've seen it used perfectly, and I've also seen it used incorrectly. I've seen an example where the benchmark, just because of a slight issue with the way it was written, it was reporting completely incorrect results. But if it's used in the right way... It depends on what you're testing, I suppose. If you're gonna be testing something and you're making HTTP requests, for example, there's so much variation anyway in HTTP you're not really gonna be getting any meaningful information. But if you've got two little algorithms and you wanna know which one's better at certain tasks, then yeah, it's great. + +\[56:14\] I agree with you, Jaana - I love the fact that it's baked straight into the language, and you just have to write a function that starts with "func benchmark name", take in the special variable, and as long as your get the for loop inside it in the right place, and also think about setup and teardown work and where that's happening, then yeah, it's a great way to really just find out which is better... Because sometimes it can be really surprising. In fact, I think it would make a great talk (or presentation), if someone out there wants to do it, of like "Here's some code. Which one's the fastest?" and have people guess. Sometimes I find the results to be very surprising. + +**Jaana Dogan:** Yeah, I think benchmarking in general is a discipline that takes a lot of time to learn, and what are the other factors that actually impact the performance. So I agree with you, I've seen a lot of wrong benchmarks, and people are super-strongly opinionated that it's actually an optimization, but it's actually one specific thing that improves performance maybe for one specific case, or something... And I think you need to have a really good understanding of the runtime and everything around the language in order to write good benchmarks, as well as interpret the results correctly. So it's a really tough game... That's true. + +**Johnny Boursiquot:** When it comes to benchmarking and performance optimization, I try very hard not to jump to that right away. I'll try to solve a problem first, and then try to optimize... Basically, prevent premature optimization. And these tools, because they're part of the standard toolchain, they make it very easy to just start using them right then there, just basically start leveraging them right away. + +There was a time - maybe we're still in that time - where it seems like there was a new HTTP \[unintelligible 00:58:13.12\] or router coming out every couple of weeks, and they were all like "Oh, benchmark. Compared to these other things, this one has zero allocation, and is 0.05% faster than the other one." I kind of found it silly a little bit, because of all that going on, and I was like "Okay, we're kind of missing the point here a little bit." But yeah, having that tool I think is great, and like you, I don't think I've seen that capability built in, part of the language, from the start. + +I'm very careful with that, because it's too easy to create a culture within an engineering team of "Okay, if I can even ship this thing, I have to make sure it's super-optimized." We're kind of putting the cart before the horse a little bit there; it's too easy to do that, so I tend to shy away from that stuff. I bring it in when I need to. + +**Jaana Dogan:** I completely agree. I think optimizations in development time is kind of like fabricated problems. You realize what needs to be optimized in production, right? For example, what we do is continuous profiling, which is we keep collecting some profiles from the production binaries, and we sort of like have an understanding of like "Within this project, what are some of the hot calls, and what is some of the stuff that is in the critical path? And what critical paths are more often being called? What happens if I just optimize this function? Or what is the actual cost of this particular function if you think about the whole system, and depending on the usage, and whatever?" + +\[01:00:09.25\] So I think it just really makes more sense to start thinking about these cases in production, and by looking at the data you just go back to the development environment and try to optimize those things, and keep using these tools. + +One of the nice things about Go profiling - the actual pprof - is it's a really low overhead type of profiling thing, and you can enable it in production. So you can just keep getting profiles from production, without impacting the critical paths so crazily. There's overhead, but there's some strategies - if you have multiple replicas of a web server, for example, you can enable production maybe for like one minute, or five minutes on one replica... Depending on how much latency you will experience, it's sometimes doable. And that's why we do it. That's what we do. Just try to optimize based on the usage, and what is the critical usage, and what are some of the hot paths; identifying those hot paths is also very important before jumping into any optimizations. + +**Johnny Boursiquot:** Right. Having a problem before you solve it. + +**Jaana Dogan:** Yeah. + +**Mat Ryer:** Jaana, when you say you do continuous profiling, when you deploy services, do you have pprof already enabled in there, and you just switch it on? + +**Jaana Dogan:** Yeah, so all the pprof tools -- pprof can be tweakable. You can turn it on dynamically, and you can turn it off. So what we do is basically turn it on for several minutes, collect the data, just parse it, store it, and then we aggregate all that data, and we have this daily, weekly etc. report... And you can take a look at "Oh, this service, particularly this handler, is often used, and all these particular functions are accounting for the most CPU time, or memory, or whatever", and you can just go and dig and optimize those particular places. + +I wish that Go had some tools around maybe supporting this type of more continuous profiling features. It's possible to write a tool that aggregates multiple pprofiles. It could be possible to write a library that automatically just turns once-in-a-while reports to some central service, and then turns it off, and so on. I think we can do much better in this field; it's just kind of up to the user right now to plan and design and do this type of thing... But that's basically what we do. I wrote on this topic for a while. + +Some companies are aware of these methodologies and some companies are not. It would be so nice if the community was producing more best practices, as well as more tooling around this. + +**Mat Ryer:** \[01:03:29.23\] Well, there we go; there's the call gone out. Anyone who's looking for a new open source project or something to hack on - what a great problem. Could you build something that samples running Go code periodically, at some schedule, and collects the results up? It would be extremely useful, and really fun probably, as well. + +**Jaana Dogan:** Yeah. It's a lot of fun once you start to see, for example, a large company aggregating all the profiling data. For example, you can actually improve your bill on your cloud provider; you can say that "Lots of the calls are dependent on this one function, and if you optimize it, we can actually cut the billing by 10%", or something. It's actually pretty useful once you start to do this systematically, everywhere. + +**Mat Ryer:** \[01:02:49.14\] Well, I love the message of "Wait till you've got something running, and then look at optimizing it." I think in some cases you can shortcut it, but generally speaking - yeah, that advice is sound. The idea of being able to profile running production systems to understand them better I think is a great goal to have... And what a great use of the tools that we have as part of our ecosystem. + +Well, on that bombshell -- I mean, I think that's it. I think we've reached the end of the hour, and so the end of this episode. Thank you very much, Johnny and Jaana. It's been awesome. Have you liked it? + +**Jaana Dogan:** I can talk about this topic for hours, and I think this was awesome... But we should keep talking about tools, I think. \[laughs\] + +**Mat Ryer:** Yeah, absolutely. Well, there's lots more to discuss, and I might even see if we can bring in some people from the community that have built some of the tools that we're using today. + +One other little bit of info that I think is quite interesting - the only actual contribution I personally made to the Go project was to remove something from Golint. One time Golint got a bit easier to satisfy thanks to me... \[laughter\] You're welcome. + +**Jaana Dogan:** Yay! \[applauds\] + +**Mat Ryer:** I delete code. I love it. Well, yeah, that's it. Thank you so much, and we'll see you next time on Go Time! diff --git "a/Go tooling \342\231\273\357\270\217_transcript.txt" "b/Go tooling \342\231\273\357\270\217_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..4be52ebd87ba508e2a75f6d32ee2eec1c8ec51dc --- /dev/null +++ "b/Go tooling \342\231\273\357\270\217_transcript.txt" @@ -0,0 +1,872 @@ +[0.00 --> 11.56] Ah, the dog days of summer. Everyone's taking time off, chilling out, maxing, relaxing all cool, maybe shooting some b-ball outside of their school. +[12.14 --> 17.64] So today we are rebroadcasting a classic episode of Go Time all about tooling. +[18.20 --> 24.26] We took you to the future a few episodes ago. Now we're headed back in time to 2019. +[24.26 --> 30.34] This was originally episode 90 and it aired almost three years ago to the day. +[30.84 --> 35.48] It was a simpler time then. Generics hadn't landed. COVID, nope. +[36.00 --> 40.12] Social distancing was only something nerds knew about. This is the way. +[40.50 --> 44.86] We didn't even have an unpopular opinions segment back then. +[45.28 --> 50.14] So some of this conversation will be quaint considering all that's changed in the world since then. +[50.14 --> 54.22] But most of it is still highly relevant. We think you'll enjoy it. +[54.44 --> 61.64] And we'll be back with some freshens next week when Natalie and Ian welcome Ronna Steinberg to discuss OOP and Go. +[66.72 --> 70.40] This episode is brought to you by our friends at Sourcegraph. +[70.48 --> 73.10] They recently launched a new feature called Code Insights. +[73.38 --> 76.34] Now you can track what really matters to you and your team in your code base. +[76.34 --> 81.10] Transform your code into a querable database to create customizable visual dashboards in seconds. +[81.46 --> 84.34] Here's how engineering teams are using Code Insights. +[84.66 --> 88.56] They can track migrations, adoption, and deprecation across the code base. +[88.80 --> 92.02] They can detect and track versions of languages or packages. +[92.46 --> 95.34] They can ensure the removal of security vulnerabilities like Log4j. +[96.14 --> 102.40] They can understand code by team, track code smells and health, and visualize configurations and services. +[103.00 --> 105.58] Here's what the engineering manager at Prezi has to say about this new feature. +[105.58 --> 113.52] Quote, as we've grown, so has a need to better track and communicate our progress and our goals across the engineering team and the broader company. +[113.90 --> 118.98] With Code Insights, our data and migration tracking is accurate across our entire code base. +[119.28 --> 125.90] And our engineers and our managers can shift out of manual spreadsheets and spend more time working on code. +[126.28 --> 126.64] End quote. +[126.64 --> 130.94] The next step is to see how other teams are using this awesome feature. +[131.28 --> 136.06] Head to about.sourcegraph.com slash code dash insights. +[136.32 --> 137.80] This link will be in the show notes. +[137.92 --> 142.50] Again, about.sourcegraph.com slash code dash insights. +[142.50 --> 158.26] Let's do it. +[158.82 --> 159.90] It's go time. +[160.50 --> 165.70] Welcome to Go Time, your source for diverse discussions from around the Go community. +[165.70 --> 169.74] We take requests, just like all the best wedding DJs. +[170.08 --> 176.12] Head to gotime.fm slash request to let us know what you want to hear about on the pod. +[176.70 --> 182.18] Special thanks to Fastly for ensuring Go Time reaches your ears super fast wherever you listen. +[182.52 --> 184.58] Check them out at fastly.com. +[184.96 --> 186.20] Okay, here we go. +[191.98 --> 194.44] Hello and welcome to Go Time. +[194.44 --> 195.18] I'm Matt Ryer. +[195.38 --> 198.12] On today's episode, we're talking about tooling. +[198.62 --> 202.94] All those great tools that help us be successful, help us do our job. +[203.50 --> 205.98] And we use them, Go tools, all the time. +[206.34 --> 207.14] Every day. +[207.26 --> 210.14] We use them for building, for running code, for testing. +[210.92 --> 215.00] We use them for formatting our code, for linting and vetting. +[215.54 --> 218.56] And many, many, many more things too. +[218.56 --> 227.44] And I think this show will be useful to anybody new to Go that wants to get a sense of the tooling around that we all use. +[228.16 --> 234.40] And I'm sure there will also be some golden nuggets for the seasoned gophers also. +[234.98 --> 238.94] And I'm so confident because of who's joining me. +[238.94 --> 243.64] I'm joined today by, in no particular order, Jana Dogen. +[244.04 --> 244.56] Hello, Jana. +[244.86 --> 245.18] Hello. +[245.78 --> 246.16] Hey. +[246.44 --> 248.28] Welcome back to Go Time. +[248.42 --> 249.00] How have you been? +[249.94 --> 250.94] Yeah, it's been a while. +[251.06 --> 252.30] I've been traveling, I guess. +[253.02 --> 253.72] Yeah, yeah, yeah. +[253.78 --> 254.82] You went to... +[254.82 --> 255.46] Where did you go? +[255.46 --> 257.56] I was in Marbella, Spain, right? +[257.72 --> 264.28] Like, the last time we talked, like, I was just going for, like, a conference and then I never came back to the show. +[264.34 --> 265.10] I'm so sorry. +[266.44 --> 267.14] That's all right. +[267.38 --> 272.78] I can understand if you're off and travels to exotic places for work. +[273.76 --> 274.62] It's a tough life. +[275.72 --> 276.44] Such a, yeah. +[276.62 --> 281.52] And you told me earlier that everything you do at work is completely confidential. +[281.70 --> 283.70] Do you want to just break all the rules and tell us anyway? +[283.70 --> 288.24] Well, kind of, like, I mean, I'm actually about to switch to a new job. +[288.40 --> 291.16] Like, I mean, not a new job, but sort of like a new role. +[291.62 --> 296.66] And currently, I'm still exploring what I'm doing, what I'm supposed to do. +[296.86 --> 304.84] And it's confidential, not because it's supposed to be super confidential, but I am not sure about, like, what I will be focusing on. +[304.96 --> 307.76] So I think I will need, like, a week or something. +[308.22 --> 308.92] That's exciting. +[308.92 --> 309.84] Just don't get it personal. +[310.12 --> 311.06] It's not about you. +[311.06 --> 313.46] It's just, you know, I'm still exploring. +[314.20 --> 315.10] Yeah, no, absolutely. +[315.44 --> 315.90] I do. +[316.02 --> 319.16] Obviously, I do take it very personally, but I'll pretend that I don't. +[320.04 --> 324.52] Well, also joining us on today's show, it's only Johnny Borsico. +[324.84 --> 325.38] Hello, Johnny. +[325.98 --> 326.52] Hello there. +[327.22 --> 330.96] And speaking of new gigs, you've just started yours, haven't you? +[331.34 --> 331.78] Yeah, yeah. +[331.78 --> 332.28] Recently. +[332.28 --> 332.92] Yeah, recently. +[333.06 --> 334.60] It's been a couple of weeks. +[334.94 --> 336.64] Still onboarding, as they say. +[337.30 --> 338.94] But, yeah, still exciting. +[339.14 --> 342.12] Still looking forward to contributing and learning. +[342.94 --> 346.12] You know, new gigs are always exciting that way, right? +[346.16 --> 353.00] There's that honeymoon period where everything is new and you're learning and you're learning about systems and people and all that good stuff. +[353.76 --> 356.00] And then at some point, I'm sure I'm going to cross that threshold. +[356.14 --> 357.78] I'm like, ah, what is going on? +[357.78 --> 359.16] I need to start fixing things. +[359.44 --> 361.06] But so far, everything is going well. +[362.10 --> 362.72] Oh, good. +[362.78 --> 363.46] I'm glad to hear it. +[363.70 --> 364.62] Yes, it is exciting. +[364.78 --> 366.86] It's scary and exciting all at the same time. +[366.96 --> 367.48] New jobs. +[368.28 --> 368.82] But, yeah. +[368.96 --> 369.78] No, I wish you all the best. +[369.86 --> 373.94] Well, if you don't mind, we'll keep asking you about it on the show because I'm very interested. +[373.94 --> 380.78] I think it's useful for other people as well to hear about things that we get up to in our professional lives. +[380.96 --> 383.82] So, if you don't mind, I'll keep bugging you about that. +[383.98 --> 384.36] Sure thing. +[386.06 --> 386.76] So, yeah. +[387.12 --> 388.16] Let's jump straight in. +[388.22 --> 390.18] We're going to be talking about GoTools today. +[390.88 --> 398.02] And I asked on Twitter earlier which of the GoTools are people's favorites or which ones do they like the most. +[398.02 --> 400.86] Mine, I'll just kick off. +[401.22 --> 407.88] Mine probably has to be GoFumpt or GoFmt or GoFormat, however you say it. +[409.20 --> 414.16] You see, for those that don't know, it formats all the Go code so it looks the same. +[415.00 --> 419.26] And all the rules are baked into the tool. +[419.42 --> 421.90] So, you don't get to choose tabs versus spaces. +[422.08 --> 424.32] You don't get to choose where the braces go. +[424.32 --> 430.24] So, you don't really get to choose a great deal about the actual format of your code. +[430.58 --> 437.10] Which, again, I think to some people when they're used to having tools that allow them to configure all this, +[437.18 --> 439.18] they feel like that's a deficiency in Go. +[439.48 --> 442.26] But it turns out to be one of Go's superpowers in my opinion. +[442.68 --> 447.98] Because what happens is all Go code starts to look the same and starts to look familiar. +[447.98 --> 456.50] And I've done it where I've been to a project and found that the code just looks like I wrote it and I definitely didn't. +[456.96 --> 458.92] And I think that's awesome. +[459.02 --> 462.96] If you think about pull requests, you know, with white space. +[463.08 --> 469.96] Sometimes pull requests and having loads of white space makes it really difficult to really see what the crux of the change is. +[470.62 --> 474.56] Well, with GoFumpt, we don't have that problem because it's all formatted nicely. +[474.56 --> 478.96] Anyone else? How do you feel about GoFumpt? +[479.00 --> 481.82] How do you pronounce it, by the way? Let's just get that one out of the way. +[483.18 --> 483.88] GoFumpt, right? +[484.42 --> 485.22] Okay, good. +[485.82 --> 487.56] I mean, that's what I know. +[488.46 --> 490.30] Yeah, that's what I'm told. +[490.78 --> 496.84] It's always awkward when I'm teaching or something and I say the Fumpt package, for example, +[497.02 --> 500.64] which is, you know, people kind of look at me sideways. +[500.78 --> 503.14] I'm like, yeah, I know, I know. Just go with it, right? +[503.14 --> 508.12] Because if you say instead, if you say FMT or format, God forbid, like, you know, +[508.28 --> 509.84] gophers are going to look at you a little weird. +[510.02 --> 511.10] Just go with it. +[511.62 --> 518.46] Yeah, it takes a while for people to, I think, parse it initially and then they learn it and like they take it and like they don't question it. +[519.18 --> 522.34] So I'm trying to, you know, keep it consistent by saying GoFund. +[522.94 --> 523.50] Yeah, same. +[523.76 --> 525.26] I mean, I agree. +[525.26 --> 527.46] Like, I wouldn't have done that. +[527.58 --> 529.92] I don't think naturally, but I heard about it. +[530.20 --> 532.38] And yeah, I do it for consistency too. +[532.38 --> 541.44] So it's funny because like sometimes people will say Golang because when we use Google and when we search or when we use hashtags, +[541.68 --> 545.38] we tend to write Golang, but we never say Golang. +[545.86 --> 550.12] So it's a little pro tip for anyone that's new to the Go community. +[550.12 --> 552.94] When you're talking about the language, just call it Go. +[553.24 --> 554.12] Don't say Golang. +[555.34 --> 556.18] Same with Phumpt. +[557.04 --> 557.34] Yeah. +[558.86 --> 564.40] So with regards to the Phumpt, well, not Phumpt, but GoPhumpt, I should say. +[565.80 --> 570.28] The reason, well, let me step back a little bit. +[570.28 --> 583.88] When I first came across Golang, I was taken aback, honestly, because I wasn't used to basically tooling, sort of formatting my code to look like a standardized sort of any sort of way. +[584.06 --> 584.16] Right. +[584.20 --> 593.36] So, you know, I come from programming languages where everybody has their little pet peeves, little quirks about, you know, I like my braces, you know, to be lined up together. +[593.36 --> 604.00] And another person would be like, I like my braces to me to end at the declaration and then for the closing bracket or brace to be at the end or whatever. +[604.16 --> 611.90] And so it's like people would have sort of these back and forths around sort of styling, you know, what's more readable versus what's not as readable. +[612.44 --> 614.42] And obviously, it was all sort of subjective, right? +[614.44 --> 617.46] Everybody has their own preferences, their own quirks and what they're used to and what they're not used to. +[618.06 --> 622.70] But GoFund sort of threw all of that out at the window when I first came across it. +[622.70 --> 630.12] And I'll be honest, I mean, for the first month or so, I was like, I don't like everything about what it does. +[630.30 --> 634.48] You know, I'm happy with like 90% of it, but I don't like everything about it. +[634.84 --> 648.94] But then over as time went on, I really began to love the tool and what it does because the beauty of it, I think you touched on that, is that every basic Go code started looking like I expected it to, right? +[648.94 --> 655.22] So basically, that cognitive load, that aspect of looking at code and reviewing code, that just went out the window. +[655.30 --> 659.82] I didn't have to worry about, okay, is this person's code going to look differently formatted than this other one? +[660.22 --> 669.42] Basically, I could just focus on the actual code and what it was doing as opposed to, you know, sort of trying to figure out parts in my head, okay, this person's quirks are that way and that person's quirks are that way kind of thing. +[669.42 --> 670.76] So it was valuable in that way. +[671.46 --> 671.68] Yeah. +[671.84 --> 678.20] There's actually something from Robert Grismer that he used to say, he is the person who is maintaining GoFund and like all the rules and so on. +[678.42 --> 681.46] He says that he doesn't agree with like all the styling. +[682.54 --> 689.26] You know, I mean, he doesn't necessarily agree with GoFund, but it's really good that like somebody, some tool is enforcing it. +[689.26 --> 690.52] So there's no question. +[691.52 --> 705.20] I mean, I work for a very large company and I witnessed, it took like four years to just tweak one little side guideline change on the Java style guideline. +[705.20 --> 715.16] And can you imagine like, you know, there's all these like hundreds of people with like strong opinions about style, just like wasting four years debating on minor style issues. +[715.34 --> 721.42] I like the fact that it is like GoFund, there's like this canonical place and there's no debate. +[721.74 --> 726.58] There's like one source of truth type of thing and everybody has to agree with it. +[726.94 --> 731.60] Even if, you know, the formatting is not always what you would desire. +[731.60 --> 737.22] Yeah. Do you think they would be able to retrospectively fit that into the tool chain? +[737.34 --> 742.10] Say that there wasn't GoFund originally and it just came out now. +[742.28 --> 746.68] Do you think the community and everyone would rally around it in the same way? +[747.04 --> 751.26] Or do you think there's something to be said for the fact that this was there from the very beginning? +[751.94 --> 759.70] I think it's necessary that like initially you create some like, you know, initial culture around, you know, just relying on a tool. +[759.70 --> 767.90] Because I think it creates like enough people, you know, it creates this community with enough people supporting the idea and understanding why it's valuable. +[768.28 --> 773.18] If you try to like inject this type of tools at a later time, the community is already fragmented. +[773.32 --> 782.66] And there's a lot of excuses to, you know, prepare a personal style because you already, for example, invested in one particular style all across a company. +[782.66 --> 786.78] And like, there's no way to, you know, just kind of like fix things at a later time. +[786.78 --> 790.60] So it's really good that they, you know, came up with a tool initially. +[790.86 --> 791.86] At least that's my opinion. +[792.46 --> 793.40] Yeah, I agree with you. +[793.48 --> 800.88] I mean, there's a few examples where the foresight or the insight from the team in the original design, +[800.88 --> 804.58] I think we really benefit from some of those decisions. +[805.16 --> 806.86] And we'll talk about more of them as well. +[806.86 --> 812.18] I think the fact that another one of the tools, GoTest, that was there as well from the very beginning. +[812.72 --> 817.62] So testing as a concept was part of, it was a first class concern in Go. +[818.04 --> 824.96] And that was, of course, makes sense because we, at the time it was being designed, you know, +[824.98 --> 827.14] that was kind of how we were building software. +[827.26 --> 828.92] Now we were writing tests a lot. +[829.02 --> 831.90] It was an important part of software engineering. +[831.90 --> 839.54] But the fact that they make these decisions, I think, early just sets a precedent. +[840.18 --> 844.14] And yeah, from there, I think it pays dividends every day. +[844.60 --> 847.30] Yeah, I think Go is doing a good job in terms of like, you know, +[847.40 --> 851.72] identifying 80% of what is essential in software engineering. +[851.92 --> 855.40] And I think, you know, tooling is kind of also representing those priorities. +[855.40 --> 859.60] Yeah, so extensive thinking beyond GoFund then. +[859.84 --> 863.58] If we look at GoLint and also GoVet, +[864.16 --> 868.18] does anyone want to have a stab at describing the difference between those two +[868.18 --> 870.24] or describing what they actually do? +[871.74 --> 872.08] Cool. +[872.54 --> 876.52] Well, yeah, so GoLint. +[876.96 --> 879.54] GoLint is, I like it. +[879.54 --> 887.52] It essentially looks at your code and does some static analysis and can catch common mistakes +[887.52 --> 889.38] and kind of give you warnings about them. +[890.02 --> 896.02] And usually, sometimes they're not mistakes, but they're just best practices. +[896.68 --> 902.30] And you can run the Lint tools on your code and see if it's got any recommendations +[902.30 --> 903.94] for things that you might change. +[903.94 --> 908.68] So one example is, if you have something in a package that's exported, +[908.94 --> 913.92] if it starts with a capital letter, then you should have a comment on that, really. +[914.04 --> 915.78] That's the sort of accepted practice. +[916.28 --> 918.48] Now, the Go spec doesn't say that. +[918.66 --> 922.74] So, of course, nothing, you know, it's not a compile error if you don't have a comment there. +[923.08 --> 928.10] But GoLint tool will catch it and say, you know, for maximum kind of quality, +[928.30 --> 931.94] for the best quality, you should consider putting a comment here. +[931.94 --> 934.88] And there are a few rules around how we write comments as well, +[934.94 --> 938.82] where we repeat the name as the first word in the comment. +[939.18 --> 943.84] And so there's a few little things like that that are encoded in the Lint, right? +[944.42 --> 944.60] Yeah. +[945.12 --> 948.24] There's, you know, initial actually, like we need to mention first, +[948.30 --> 950.38] I think there's a difference between that and Lint. +[950.74 --> 953.66] That is, you know, reporting more of like suspicious stuff. +[953.66 --> 963.16] And, you know, like some patterns that might be just, you know, might be just, +[963.76 --> 969.38] I mean, a misuse of the, of an API that it may actually, you know, +[969.42 --> 971.80] just kind of like corrupt some memory or whatever. +[972.30 --> 975.30] Like think about like the typical example of printf. +[975.70 --> 978.88] If, you know, you pass the wrong type of arguments, +[979.00 --> 982.20] Beth is going to complain about it. +[982.20 --> 986.70] So both Lint is more about like, I think style errors, +[986.96 --> 989.72] more of like, if you don't, for example, +[990.08 --> 991.80] Godok, a public API, +[992.16 --> 995.24] it's going to complain about that type of problems. +[995.96 --> 997.84] So that became a part of the test, +[997.96 --> 1003.30] but like not, I think all the things that is reported as a part of Beth is genuine. +[1004.24 --> 1008.34] So you can like, there could be like false positives as far as I know. +[1008.34 --> 1011.84] And it also applies to Lint as well. +[1011.84 --> 1015.64] So these are not like a part of the compiler because, you know, +[1016.00 --> 1021.96] there's like some reports that is not accurate or something. +[1022.86 --> 1026.22] But it's just generally like, you know, you need to follow. +[1026.22 --> 1032.08] So they generally generate like genuine enough reports and they're really useful. +[1032.74 --> 1033.18] Yeah, you're right. +[1033.26 --> 1037.40] When it catches, if you use like printf or wrapf, +[1037.44 --> 1039.08] if you use one of those f methods, +[1039.08 --> 1044.50] and then you don't put the correct number of verbs or whatever the arguments in, +[1044.90 --> 1047.26] catching things like that is extremely useful +[1047.26 --> 1051.38] because it's quite hard at a glance to just see those kinds of mistakes. +[1052.16 --> 1057.36] So yeah, I think people should switch on those tools for their code base, +[1057.42 --> 1058.94] at least run them for their code base, +[1059.06 --> 1062.00] and see what kinds of things it is actually saying, +[1062.12 --> 1064.38] because you might find you agree with them. +[1064.78 --> 1066.22] The comment one's a good example. +[1066.52 --> 1069.34] I mean, it's quite dogmatic. +[1069.50 --> 1072.00] It just says, okay, it's exported, so it needs a comment. +[1072.00 --> 1076.38] Now, if that method is something like, +[1076.52 --> 1078.90] or if it's a function that says new thing, +[1079.38 --> 1082.12] then it's obvious that's making a new thing. +[1082.24 --> 1085.48] And your comment's probably going to say, new thing makes a new thing. +[1085.88 --> 1088.56] So we have a little bit of redundancy. +[1088.92 --> 1093.24] But I think generally speaking, if you do follow the Lint tools, +[1093.52 --> 1097.96] then I find that, you know, the code, again, it starts to look more familiar +[1097.96 --> 1101.76] and you get all the other benefits of GoFund. +[1102.00 --> 1105.08] One of the things that I typically do, +[1105.24 --> 1109.80] which is probably the reason why, for me, +[1110.28 --> 1111.68] like off the top of my head, +[1111.74 --> 1115.42] sort of differentiating between the linting and the vetting was sort of, +[1115.76 --> 1118.62] I was like, hmm, I guess I've never really thought about the difference that much +[1118.62 --> 1120.72] because they're part of my tool chain. +[1120.94 --> 1127.90] So like on my day-to-day, I use VS Code and Vim as sort of my editors of choice. +[1127.90 --> 1131.62] And basically they have the plugins, you know, +[1131.76 --> 1134.60] and the extensions sort of built in as part of my workflows. +[1134.76 --> 1138.78] Every time I hit save, right, these tools are running, right? +[1139.12 --> 1142.06] And I'm getting different, basically, +[1142.42 --> 1146.18] markers at different spots from different tools, right? +[1146.18 --> 1149.06] So there's another popular open source project out there, +[1149.26 --> 1150.68] I think it's called the GoMetLinter, +[1151.32 --> 1153.94] which includes a bunch of those kinds of tools as well. +[1154.02 --> 1157.24] You can configure, you can turn some off and others on and whatnot. +[1157.60 --> 1159.14] But these tools together, +[1159.22 --> 1163.00] they give you sort of a set of outputs that you can basically go through +[1163.00 --> 1164.98] and figure out, oh, yeah, I missed, you know, +[1164.98 --> 1165.96] I used the wrong verb here. +[1166.02 --> 1167.60] I'm supposed to use an integer. +[1167.60 --> 1169.12] I'm using a string instead, right? +[1169.32 --> 1174.44] Things that the linter and vet would sort of find for you +[1174.44 --> 1176.16] and if you run them sort of individually. +[1176.70 --> 1178.32] But because they're part of my tool chain, +[1178.82 --> 1182.18] basically I just look at the view at the bottom of my editor +[1182.18 --> 1184.40] and get a list of things that I go and fix. +[1184.54 --> 1188.54] So I've sort of almost basically, I don't care, I should say, +[1188.68 --> 1193.36] which tools give me what unless I really need to work with a specific tool. +[1193.82 --> 1195.42] But I kind of, you know, it's part of my workflow. +[1195.54 --> 1196.44] It's just part of my editor. +[1196.44 --> 1198.90] Every time I hit save, formatting gets done. +[1199.62 --> 1200.72] Go import says this thing. +[1200.88 --> 1203.52] Whatever I've referenced in my code that is not imported, +[1203.74 --> 1205.02] it brings that in automatically. +[1205.66 --> 1206.88] All these things sort of happen. +[1207.20 --> 1211.32] The tooling kind of makes it easy to sort of just focus on writing the code +[1211.32 --> 1215.08] and not worry so much about having to run individual tools one at a time. +[1215.86 --> 1219.30] Yeah, it's a good point that actually making it that part of, you know, +[1219.34 --> 1221.26] the editing experience is really useful. +[1221.82 --> 1225.44] Like a special vet is reporting a lot of like, you know, useful stuff like, +[1225.44 --> 1229.72] okay, this is unreachable or, you know, you're passing the wrong, you know, +[1229.72 --> 1235.36] you're passing, for example, unmartial and non-pointer and like stuff like that. +[1235.42 --> 1241.40] Like it's so hard sometimes by just when you're typing and when you're just like coding, +[1241.40 --> 1245.34] but like tool is really helping you to do the right thing as you are, you know, programming. +[1245.34 --> 1248.60] Yeah, and I extend that to running tests as well. +[1248.76 --> 1255.68] I tend to write unit tests which run very quickly and then you can run those every time you save the package. +[1256.12 --> 1262.40] Usually, you know, if they start getting too slow, then of course you have to have a different strategy. +[1262.54 --> 1269.32] But certainly in the beginning, if it's unit tests that just run very quickly and the build time in Go is still phenomenal. +[1269.32 --> 1275.32] We always kind of forget about it until you have to go and build a different code base. +[1275.52 --> 1277.38] Then you appreciate it again. +[1278.24 --> 1283.08] And Johnny, by the way, yeah, the MetaLint now apparently is called Golang CI-Lint. +[1283.40 --> 1288.18] So if you want to install that into VS Code, it's Golang CI-Lint. +[1288.90 --> 1291.14] That's the new name of that. +[1291.60 --> 1292.48] Yeah, but you're right. +[1292.54 --> 1293.42] It's the MetaLint. +[1293.42 --> 1298.78] It runs a range of other Linters and kind of gives you that one view of it. +[1299.02 --> 1301.32] And they integrate brilliantly into the IDEs as well. +[1301.40 --> 1305.08] So that's the other thing, like you say, you can run it on save, but even if you don't, +[1305.18 --> 1311.84] you can still usually integrate it into the IDE in some way that just makes it part of your routine. +[1312.26 --> 1318.72] Because, you know, anytime you can get that live feedback from the code, that's valuable. +[1318.72 --> 1323.96] You know, because usually as you're working, you learn too. +[1324.32 --> 1329.50] And that's a great way to learn things as you're writing code and to see a linter saying, +[1329.62 --> 1333.34] oh, you know, this is unreachable now or that thing's over there now. +[1334.68 --> 1339.70] You know, and if it's tests too, then, oh, these tests are broken over here that you didn't expect. +[1340.46 --> 1344.96] And you just get that feedback from the code, which is so useful when you're working. +[1344.96 --> 1352.84] And you shouldn't have to wait until basically, you know, if you have CI, continuous integration, you should. +[1353.14 --> 1359.68] You shouldn't have to wait until the code reaches, you know, that remote server where all these tools are run for you to get that feedback. +[1359.94 --> 1361.94] It's much easier, much faster, right? +[1362.00 --> 1365.98] Like you're saying, that feedback loop is much tighter when it's part of your tooling. +[1366.10 --> 1369.74] So there's some things you can do locally, right, to make sure your code is funpted, +[1369.96 --> 1373.46] make sure it's vetted, it's linted, and all that good stuff. +[1373.46 --> 1380.22] And then when it goes up for review for PR, you know, Circle, you know, whatever CI2 you're using, Travis, Circle, whatever, +[1380.70 --> 1381.66] there's dozens of them. +[1381.80 --> 1388.00] So you can sort of, you know, they give it a blessing, and then now people can just focus on what does the code do, right? +[1388.02 --> 1392.42] They don't have to tell you, hey, you forgot to, you know, run GoFundit or something, right? +[1392.48 --> 1394.04] You take advantage of these tools locally. +[1394.56 --> 1395.46] They're very good tools. +[1395.62 --> 1399.90] So I wholeheartedly encourage folks to sort of make them part of your development workflow. +[1399.90 --> 1404.18] Yeah, one of the, I think, best parts is like they are really fast also. +[1404.44 --> 1407.82] You know, it's a part of the editing experience because they are fast. +[1409.00 --> 1417.36] And, you know, it's just, I'm coming from like background where, you know, I use a lot of Java tools. +[1417.36 --> 1421.50] And it's, you know, not like it is, I think, smooth the experience. +[1421.72 --> 1426.22] We used to have like similar static tools, but it was not as smooth as all these Go tools. +[1426.80 --> 1433.20] So nobody is making it optional because it doesn't really, you know, make the editing experience more challenging +[1433.20 --> 1435.26] because they are fast and they are useful. +[1435.26 --> 1449.52] This episode is brought to you by Square. +[1449.78 --> 1454.98] Millions of businesses depend on Square partners to build custom solutions using Square products and APIs. +[1454.98 --> 1461.84] When you become a Square solutions partner, you get to leverage the entire Square platform to build robust e-commerce websites, +[1462.22 --> 1465.34] smart payment integrations, and custom solutions for Square sellers. +[1465.84 --> 1468.58] You don't just get access to SDKs and APIs. +[1469.04 --> 1476.30] You get access to the exact SDKs and the exact APIs that Square uses to build the Square platform and all their applications. +[1476.96 --> 1479.04] This is a partnership that helps you grow. +[1479.18 --> 1483.48] Square has partner managers to help you develop your strategy, close deals, and gain customers. +[1483.48 --> 1490.52] There are literally millions of Square sellers who need custom solutions so they can innovate for their customers and build their businesses. +[1491.10 --> 1492.74] You get incentives and profit sharing. +[1493.14 --> 1497.52] You can earn a 25% SaaS revenue share, seller referrals, product bounties, and more. +[1497.84 --> 1500.00] You get alpha access to APIs and new products. +[1500.12 --> 1502.46] You get product, marketing, tech, and sales support. +[1502.68 --> 1504.46] And you're also able to get Square certified. +[1504.66 --> 1507.64] You can get training on all things Square so you can deliver for Square sellers. +[1508.10 --> 1512.32] The next step is to head to changelog.com slash Square and click become a solutions partner. +[1512.32 --> 1514.84] Again, changelog.com slash Square. +[1516.98 --> 1527.62] We mentioned GoTest. +[1527.82 --> 1529.54] That's another tool that we use a lot. +[1530.04 --> 1534.58] Anyone that's not used it, if you write test codes in your Go programs, +[1534.58 --> 1539.48] and you do that usually by naming the file with underscore test dot go at the end. +[1540.24 --> 1541.68] And then you run GoTest. +[1541.76 --> 1547.14] It will look through all those test files and it will actually run all the test code for you. +[1547.20 --> 1553.06] And that's really how, you know, you can, if you do TDD, you know that your code is fulfilling its promises. +[1553.24 --> 1555.40] It's doing what you said it was going to do. +[1555.40 --> 1562.02] There's also another little feature in the test tool, which I think gets overlooked a little bit. +[1562.24 --> 1563.90] And it's the race detector. +[1564.96 --> 1572.08] So when you're writing concurrent code, it's possible for you to break the rules +[1572.08 --> 1576.02] and sort of try and read and write from the same data at the same time. +[1576.02 --> 1581.06] And if you try and do something like that, that's illegal and it will crash the program. +[1581.40 --> 1588.30] But of course, it's very difficult to see that sometimes when you, if you've written the concurrent code +[1588.30 --> 1593.38] and certainly difficult to write tests for it, because sometimes it might just not happen +[1593.38 --> 1595.14] just because of the way that things get scheduled. +[1595.14 --> 1603.20] But there is a race flag, which you can pass into GoTest, which will, it's a bit slower, +[1603.32 --> 1613.14] but it does some additional checks and you can catch those potential deadlocks early, which is kind of cool. +[1613.60 --> 1619.32] Yeah. And the tooling is also a part of, you know, the standard tooling. +[1619.32 --> 1625.66] It's not just a test, but it's a really good addition that like T-SAN, you know, detector is also a part of the tests +[1625.66 --> 1631.16] because we all have this workflow of not merging things if the tests are not passing. +[1631.52 --> 1637.10] So you would, you know, ideally want to enable the race detector as a part of your CI. +[1638.00 --> 1645.16] And it's amazing, but there's like one thing I think we should mention that your tests should cover concrete cases, +[1645.36 --> 1648.94] concrete cases so the, you know, detector can detect them. +[1648.94 --> 1654.82] If you don't represent those like concrete, you know, situations, the detector won't be able to detect them. +[1654.96 --> 1662.06] But it's amazing because it's, it's just like so on point and it's easy and it's a part of the standard tools. +[1662.22 --> 1668.54] So you don't have to like, you know, figure out like all these additional extra tools or whatever in order to get the benefits. +[1669.34 --> 1675.76] Yeah. Now it's worth saying that the race detector will, if it reports that there's a violation, +[1675.76 --> 1681.24] then that is a violation, but it doesn't necessarily catch everything. Isn't that true? +[1682.36 --> 1682.64] Yeah. +[1682.92 --> 1683.08] Yeah. +[1683.60 --> 1694.40] Okay. So, but still, I mean, you know, it's still, to be honest, I've, I've never seen a race condition get through after testing it with. +[1694.40 --> 1699.36] Because you are actually good in terms of what you care about your tests. +[1699.36 --> 1701.10] So you represent all the cases. +[1702.10 --> 1710.04] I've seen a lot of times people are just like, you know, not creating those like situations where concurrency is a problem. +[1710.96 --> 1713.76] They have this, all these like super micro tests. +[1714.14 --> 1716.50] So they don't really, you know, capture it. +[1716.50 --> 1720.64] And I think it's really important to tell that like your tests should represent those cases. +[1720.64 --> 1722.48] So the race detector can detect them. +[1723.18 --> 1725.60] Hmm. Yeah. Okay. That's, that's a really good point. +[1726.24 --> 1732.86] Well, with TDD, you tend to get good coverage, even though I don't, covering by the way, +[1733.02 --> 1739.34] code coverage is also another part of the tooling that we just get for free, which is, is awesome. +[1739.34 --> 1747.20] But yeah, I, I never try and shoot for a hundred percent code coverage or anything, but naturally it's quite high with TDD. +[1747.42 --> 1752.44] And I suppose naturally you'll also cover a lot of those cases that you talked about as well. +[1752.98 --> 1754.98] I like Go Run as well. +[1755.32 --> 1767.50] Go Run is like a, it, you don't tend to have much magic in Go, but Go Run is probably the magic tool because it actually secretly does a build and then executes, +[1767.50 --> 1777.86] you know, it does a few steps behind the scenes, but it's great if you're just learning to code or you just want to write a little script quickly and just execute a program. +[1778.04 --> 1785.32] You can use Go Run and you pass in the name of the file or files and, and it just runs it. +[1785.38 --> 1788.46] I mean, it builds it to a temporary directory and I think it gets deleted afterwards. +[1788.68 --> 1789.54] Well, I'm not sure. +[1789.54 --> 1799.70] Um, but yeah, I think that also is, it's such a nice thing to be able to just quickly see results and see feedback from what you're doing. +[1799.90 --> 1802.10] And Go Run is another example of that. +[1802.38 --> 1802.90] Yeah. +[1802.92 --> 1806.90] I think, uh, people use Go Run for their like first Hello World program. +[1807.42 --> 1809.76] Um, it sometimes becomes like complicated. +[1809.76 --> 1812.96] So then they, you know, have this habit of like using Go Run. +[1812.96 --> 1822.06] Uh, Go Run, I think before Go Pad was a little bit more difficult to rely on because it was some sort of like, you know, it was able to work outside of Go Pad. +[1822.46 --> 1826.68] So, um, the behavior of Go Build and Go Run was not quite the same. +[1826.68 --> 1834.70] So, you know, it's just kind of like people have been advocating to always rely on Go Build or install rather than Go Run. +[1834.90 --> 1840.66] But I think like, it's just really nice for a Hello World or if you have a script type of thing that you just go run. +[1840.66 --> 1842.00] Um, it's still useful. +[1842.74 --> 1843.10] Right. +[1843.20 --> 1851.92] The, the Go Run, I think, yeah, like you were right, um, Matt, when, when you're saying basically it, from my understanding is that this, it does the same thing as, as Go Build. +[1852.06 --> 1856.68] It's just the difference being that, okay, once a program is run, it just discards that temporary artifact. +[1857.34 --> 1861.18] Um, at least that's, that's, that's the high level of what I think it does. +[1861.52 --> 1866.56] Um, one thing is worth mentioning is also you can run it with a, I believe it is run, you can run with a dash race as well. +[1867.04 --> 1870.36] Um, that way, you know, if there's any sort of a race conditions in the code, +[1870.36 --> 1879.44] it'll actually, um, when, if, when the program fails, if it panics, um, then you'll actually get some information around, um, where that sort of a race condition occurred as well. +[1880.58 --> 1880.98] Hmm. +[1881.56 --> 1882.54] I didn't know that. +[1882.94 --> 1883.60] That's brilliant. +[1883.60 --> 1883.88] Yeah. +[1883.92 --> 1893.10] I think race is, race is supported in, um, like test, build, um, run, like general all across the tools. +[1893.50 --> 1893.70] Hmm. +[1893.70 --> 1893.82] Hmm. +[1894.42 --> 1897.18] But you, but it adds overhead, doesn't it? +[1897.20 --> 1898.80] And slows down your program and things. +[1898.84 --> 1901.18] It's not something you would just always switch on. +[1901.66 --> 1902.04] Yeah. +[1902.04 --> 1906.88] That's why I think it's useful to just, you know, make it, uh, an optional thing for tests. +[1907.10 --> 1912.50] Um, but you know, apart from that, like you don't want to have the race detector always on. +[1913.28 --> 1913.68] Yeah. +[1913.68 --> 1917.74] I've, I've had mixed results depending on the size of the code base, obviously. +[1917.98 --> 1922.76] Um, the, but these days I work on a lot of somewhat small, um, code bases. +[1922.90 --> 1924.64] I work a lot with my services, that kind of thing. +[1924.70 --> 1928.14] So these, these code bases tend to be somewhat small, relatively speaking. +[1928.14 --> 1932.72] Um, so I, I, by default, whenever I, my, the default make command, right. +[1932.80 --> 1933.46] For, I use make. +[1933.58 --> 1939.42] So when the default, whenever I run make for the default is basically to just run it with the dash race flag, run the test with dash race flag. +[1939.42 --> 1947.74] Um, I, I haven't noticed, um, significant slowdown in that, but again, you know, obviously your mileage is going to vary depending on the size of your project and how many pages we got going on. +[1948.16 --> 1960.18] There was a benchmark about this and like, um, I think it was kind of like memory usage is again, like five times, you know, larger if you, um, have the like race detector on. +[1960.30 --> 1966.64] And I think execution time wise, like, again, like there was like some reports, but it's really dependent on the use case, as you say. +[1966.64 --> 1973.52] So it's kind of like adding some overhead, which could be, I think, um, two to 20 X or something. +[1973.52 --> 1976.22] If I can, you know, remember the numbers correctly. +[1976.46 --> 1983.46] Uh, there's a really good blog post actually, or an article on the, uh, go lang.org about the race detector. +[1983.76 --> 1986.24] And, uh, there must be like some numbers over there. +[1987.46 --> 1988.26] Yeah, cool. +[1988.38 --> 1988.64] Okay. +[1988.94 --> 1996.30] Well, so I was thinking as well about, um, go get, go get, go get another one of the tools, which I think, you know, +[1996.30 --> 2000.62] obviously things have changed a lot, especially in the module space. +[2000.96 --> 2009.64] Um, but I've got to say when I was first using go to just be able to install packages by saying, go get, and then the package name. +[2009.64 --> 2016.46] And then for that package name also to be the import path and to be the URL of where that package lives. +[2017.10 --> 2024.96] Um, I found that to be such an elegant thing that it was very easy to, to install things. +[2024.96 --> 2030.06] I mean, this is when we had, this is in a go path world where everything just gets put into one place. +[2030.06 --> 2034.58] Um, but go get just really made that very, uh, very easy. +[2035.26 --> 2039.88] Um, how do you feel about go get versus the new module tools? +[2039.88 --> 2043.58] Cause they're the, the working with modules is, is a little bit more complicated. +[2044.46 --> 2051.70] So, um, I'll, I'll, I'll punt the modules, um, to, uh, to JVD and let her tackle that. +[2051.70 --> 2061.88] But I can tell you that for the, when using go get, like, especially when I'm teaching, like being able to say, look, like we're gonna, we're gonna import this package before we can actually import this package and use it in our code. +[2062.10 --> 2063.36] You know, we need to go get it. +[2063.40 --> 2063.58] Right. +[2063.58 --> 2067.80] So, you know, I'd literally say, okay, go get, and then basically I'd find the name of the package. +[2067.80 --> 2074.20] If we qualify path, basically, you know, with github.com, whatever, um, or whatever the, wherever the public repository is. +[2074.70 --> 2078.62] Um, and then, you know, and then basically I'd get this blank stare from the students. +[2078.62 --> 2080.42] They'd be like, okay, what just happened? +[2080.92 --> 2092.02] You know, and then it, I, it dawned on me that, okay, if I literally copied, right, that path, go into the browser and paste it into your URL bar and navigate to that repository. +[2092.76 --> 2095.08] Immediately they were like, oh, okay. +[2095.08 --> 2098.06] I see, I see what this is, right? +[2098.08 --> 2102.32] You are literally pulling this code that lives at this very path, right? +[2102.34 --> 2104.76] You're putting on the command line, you are pulling it down. +[2104.76 --> 2111.30] Now I can actually see and read that code, you know, in my browser and see what it is I'm actually pulling down, right? +[2111.30 --> 2115.52] So the whole thing about, you know, pulling down the package, you know, it goes in your go path. +[2115.72 --> 2118.26] None of that stuff makes sense, right, for them. +[2118.32 --> 2124.24] But the moment that I can actually go into a browser and put that very path in, it sort of clicked, right? +[2124.24 --> 2133.72] They, now they understood the value of GoGet and it didn't quite, you know, it didn't matter really much where it was being put in the go path. +[2133.88 --> 2135.96] It's just the fact that they knew how to get it. +[2136.04 --> 2140.36] They knew how to go, where to go and see whatever was being pulled was, was almost magical for them. +[2141.72 --> 2143.64] Funny because it's no magic. +[2143.64 --> 2148.56] And it's almost the fact that it's so obvious, i.e. that's the URL, go and look at it. +[2148.68 --> 2149.90] You know what a URL is. +[2150.84 --> 2153.36] I think that that's great. +[2153.66 --> 2156.92] And that you, the little story you just told then makes total sense. +[2157.00 --> 2166.56] I mean, when, when I use, if I use some NPM stuff for a project, I install a few things and I look in that node modules folder. +[2166.72 --> 2168.70] There's 16 million folders in there. +[2168.70 --> 2174.48] And yeah, and I don't know where they've come from. +[2174.82 --> 2176.24] It's kind of hidden. +[2176.58 --> 2177.70] It's, it is magic. +[2178.14 --> 2194.64] Whereas it's just, you know, that thing of being very simple and clear, even if you sacrifice some features for that, I always think is, has, has such a positive kind of dividend that it keeps paying again and again later. +[2194.64 --> 2198.62] I think we need to make an episode on go month. +[2198.96 --> 2202.88] But I think go, I agree that like go get is a really good, you know, initial experience. +[2203.32 --> 2214.62] And one thing I like about is if, if you're go getting a main package, it, you know, installs it, puts in your go, go pat being directory. +[2214.62 --> 2215.06] Yeah. +[2215.06 --> 2218.62] So it's just like a good way to, you know, distribute tools as well. +[2219.08 --> 2226.44] Before I think go, I was just publishing binaries and like making sure that like I have the, you know, the right version all across. +[2226.78 --> 2232.32] Versioning still is a problem with go get, but like, I, I think it's a, it's an okay sacrifice. +[2233.44 --> 2233.96] Yeah. +[2233.96 --> 2234.40] Okay. +[2235.30 --> 2245.10] I'm just going to, what I'm going to do is just keep moving on to different go tools because I've, I've, I'm already learning things about these as well. +[2246.44 --> 2252.74] And the other one, the other one with go build, which I love is the fact that we can do cross compilation. +[2253.40 --> 2256.96] Now this, this has been around from, I think the beginning. +[2256.96 --> 2266.50] Um, essentially for those that don't know, you can choose the, the target architecture, the target machine to build your go code for. +[2266.90 --> 2272.66] That's very useful if you're using Docker, cause you can do like on a Mac, you can do the build for Docker. +[2272.90 --> 2279.24] And then you've got the doc, you've got the binary, the Linux binary, uh, that you can then put into the Docker image. +[2279.54 --> 2284.26] Or you can of course put the code into Docker and build it in there, uh, in that environment. +[2284.26 --> 2288.74] But do you, how would you, how's your experience with cross compilation so far? +[2289.40 --> 2290.42] I think it was magic. +[2290.42 --> 2299.36] Like when I first saw, um, you know, they were typing go, oh, it's go, actually it's pronounced goose, um, and windows and go build. +[2299.36 --> 2300.84] And like, you get a windows binary. +[2301.00 --> 2302.08] It was like, whoa. +[2302.38 --> 2302.74] Right. +[2302.74 --> 2308.00] Like, um, I, it was fascinating and I usually generate binaries for Linux. +[2308.00 --> 2312.72] So it was like, I kept, you know, working on my Mac without any worry or anything. +[2312.72 --> 2313.84] It was so awesome. +[2314.26 --> 2314.82] Yeah. +[2315.02 --> 2315.96] Have you used it, Johnny? +[2316.50 --> 2316.94] Absolutely. +[2317.28 --> 2321.14] Um, one of my, uh, first, uh, one of my first jobs using go full time. +[2321.38 --> 2328.68] My responsibility was the, basically to, to have, uh, sort of a multi, multi platform, um, build process. +[2328.86 --> 2332.78] Um, so basically I relied on, on goose and gorge quite a bit. +[2332.78 --> 2342.72] Um, and for those of you who don't know what gorgeous is busy, the, that's the companion to, to goose, uh, G O A R C H, um, for go architecture. +[2342.72 --> 2344.06] Um, yeah. +[2344.06 --> 2345.06] Yeah. +[2345.06 --> 2352.10] So using goose and gorge were sort of a bread and butter, um, to having that work done and basically being able to push up binaries for all kinds of different platforms. +[2352.10 --> 2358.16] And I mean, there are a ton of them that, you know, go support out of the box, um, for ARM processors. +[2358.16 --> 2362.54] And, and, and I mean, there's, there's a, there's a, there's a, the combination, a sheer combination you can have. +[2362.78 --> 2370.52] Um, I've lost, I've lost track of, of, of, of basically how all the different variations you can push out, but it's, it's really was a godsend. +[2370.52 --> 2375.14] I mean, I, there's no way I would have been able to, to sort of get that job done without, without these things being in there. +[2375.76 --> 2382.92] Uh, I think it's also awesome that like, I was doing a lot of developments, uh, for ARM and, you know, for a Raspberry Pi, for example. +[2383.32 --> 2390.26] Uh, the processor on a, you know, typical Raspberry Pi is just going to be not comparable to my laptop. +[2390.26 --> 2394.92] So I would just, you know, build things on my laptop because it's going to be faster. +[2394.92 --> 2400.08] And then I will push it to the Raspberry Pi because it's just so much easier to do cross compilation. +[2400.82 --> 2404.18] And, uh, it's just like maybe like 10 times faster or something. +[2404.88 --> 2405.32] Wow. +[2406.18 --> 2408.04] And so how does it actually work? +[2409.08 --> 2418.12] Because obviously the compiler is doing a few steps and then ultimately it then creates a binary that's made up of, um, from the machine code. +[2418.12 --> 2422.26] And is it just that the machine code is generated differently depending on the architecture? +[2422.92 --> 2423.00] Yeah. +[2423.08 --> 2426.16] You know, like, I mean, they know what to generate for each architecture. +[2426.56 --> 2431.04] So they just basically take the inputs and they know what to map it. +[2431.14 --> 2436.76] And then they generate, uh, the output based on the, you know, um, operating system and architecture. +[2438.14 --> 2444.06] And that must've been possible because of the way that they built the tool system. +[2444.06 --> 2449.10] Do you think it was deliberate that they wanted to be able to build to any target architecture? +[2449.10 --> 2455.98] Or do you feel like they just realized they could after because they'd just built it and designed it in a simple way? +[2456.76 --> 2460.70] I don't think you, you stumble on something like this, um, by accident. +[2460.70 --> 2473.70] I think I had, I mean, if I had to guess, I'd say this was by design, um, is considering that the, the creators of, of the language busy to head, they had, um, they see they were building for, for Google. +[2473.70 --> 2474.00] Right. +[2474.00 --> 2485.30] So I imagine that at some point they need to be able to run, um, binaries on different platforms with different CPU architectures and, you know, 32 bits or 64 bits and all that, and all that good stuff. +[2485.40 --> 2490.12] So I imagine this must've been sort of a part of the plan, part of the design. +[2490.34 --> 2498.96] This, this seems way too complicated and way too powerful a feature to have just come across, um, um, to have fallen out of the build system. +[2498.96 --> 2499.96] Hmm. +[2499.96 --> 2500.46] Hmm. +[2500.46 --> 2505.56] There's also like, we, uh, I think simplified the process, but there's this intermediate assembly. +[2505.56 --> 2510.30] So, uh, the compiler first translates everything to that intermediate assembly. +[2510.30 --> 2518.16] And from that point on, uh, they are being compiled to the, you know, the architecture specific, um, machine code, uh, instructions. +[2518.16 --> 2524.64] So, so, um, it's, it's actually like, you know, the internals of compilers, like this, like two step thing. +[2524.64 --> 2529.36] Uh, and this is like a really typical way the compilers work. +[2529.48 --> 2533.50] They're just, you know, taking it, converting everything into an intermediate language. +[2534.04 --> 2538.96] And then from that intermediate language, you can just basically target whatever architecture you want to target. +[2539.46 --> 2539.98] Hmm. +[2540.28 --> 2542.66] And of course you can have build tags as well. +[2543.36 --> 2546.62] Does anyone want to describe build tags? +[2547.44 --> 2547.66] Yeah. +[2547.72 --> 2551.38] Build tags are, uh, it's providing conditional completion. +[2551.38 --> 2554.66] And you can create different rules. +[2554.66 --> 2560.30] For example, you can have constraints to say, only use this file for Linux builds. +[2560.30 --> 2566.48] Or you can say, I just want only arm builds to have this file included in the build. +[2566.70 --> 2571.04] Uh, you can, there are many different rules provided by the tool chain. +[2571.16 --> 2573.70] Uh, go version is one of them. +[2574.08 --> 2576.58] Arbitrary custom build tags is one of them. +[2576.58 --> 2590.28] So it kind of gives you this like, you know, possibility to switch to different implementations depending on the go version, uh, depending on the, you know, the, uh, operating system or architecture or some custom build tags. +[2590.64 --> 2591.08] Yeah. +[2591.12 --> 2594.46] I've used those successfully when it comes to testing. +[2594.46 --> 2603.98] Sometimes if there are long running tests or if there are integration tests that require a different dependency to be running or something, I use build tag in our test files. +[2603.98 --> 2607.24] And that's quite an easy way to choose a subset of things to run. +[2607.96 --> 2611.64] Um, and it's just a special comment that goes at the top of the file, isn't it? +[2612.14 --> 2612.44] Yeah. +[2612.44 --> 2617.14] It's just like, I think it, it must be on, um, I mean, it's on the top of the file. +[2617.38 --> 2620.02] Um, there's a particular place, but that's it. +[2620.08 --> 2621.30] Um, and it's really readable. +[2621.30 --> 2630.72] I think my only complaint about this rules, uh, about the build constraints is like, it's just really hard to sometimes just, you know, have like multiple rules represented. +[2630.72 --> 2632.68] It becomes really hard to parse. +[2632.68 --> 2643.60] Like if you want to have like more complex rules, like, Hey, just include this file on Linux, um, you know, Darwin and blah, blah, but not on this particular thing. +[2643.60 --> 2652.48] And I think on top of that, like not for this custom build tag, like I think writing those expressing those, uh, complicated, more complex type of constraints is a little bit hard. +[2652.48 --> 2654.70] But otherwise I think it's just pretty straightforward. +[2655.06 --> 2656.98] And I use build tags all the time. +[2656.98 --> 2670.98] This episode is brought to you by our friends at retool. +[2671.14 --> 2677.90] Retool helps teams focus on product development and customer value, not building and maintaining internal tools. +[2677.90 --> 2687.60] It's a low code platform built specifically for developers, no more UI libraries, no more hacking together data sources, and no more worrying about access controls. +[2688.10 --> 2696.00] Start shipping internal apps to move your business forward in minutes with basically zero uptime, reliability, or maintenance burden on your team. +[2696.34 --> 2705.38] Some of the best teams out there trust retool, Brex, Coinbase, Plaid, DoorDash, LegalGenius, Amazon, Allbirds, Peloton, and so many more. +[2705.38 --> 2710.44] The developers at these teams trust retool as their platform to build their internal tools. +[2710.64 --> 2711.90] And that means you can too. +[2712.10 --> 2713.06] It's free to try. +[2713.20 --> 2718.98] So head to retool.com slash changelog again, retool.com slash changelog. +[2718.98 --> 2740.08] Okay, well, I want to also mention a couple of tools from the community as well. +[2740.08 --> 2748.10] Because remember, you know, we are, we are writing, we are using Go tools all the time, but we can write tools as well. +[2748.22 --> 2759.12] And some people have contributed, like I think Go Imports was a Brad Fitzpatrick project that was his own kind of idea that he just did on his own. +[2759.12 --> 2765.06] And it essentially wraps Go thumped, so you get all the formatting, but it also resolves imports for you. +[2765.74 --> 2768.24] And you can do these things too with your own tools. +[2768.86 --> 2774.20] And some of the tooling as well doesn't have to be Go tooling running on our machine. +[2774.54 --> 2778.36] Matt Holt has a great JSON to Go service. +[2778.36 --> 2787.00] If you Google JSON to Go, you basically paste in a JSON blob and then it generates the Go structures for that JSON blob. +[2787.72 --> 2796.16] Extremely useful, especially if you're going to consume an API and you need all of the data and you don't just, you just don't want to sit and type out all the field names. +[2796.36 --> 2797.78] So that's a very useful one. +[2797.80 --> 2798.80] And that's a hosted website. +[2798.80 --> 2800.44] So you can, you can go to that. +[2800.44 --> 2808.40] Any other community, are there any other community tools that we like? +[2809.08 --> 2826.34] I personally like the Go report card website, which, which, well, I guess it's less of a local tool, but something that can basically evaluate sort of how close to the idioms, right? +[2826.34 --> 2831.16] You have the Go community, your, your, your, your code is, is, is, is being kept at. +[2831.40 --> 2834.28] I think it might even incorporate some of the tools we've mentioned before. +[2834.38 --> 2836.20] They'll enter the vet, the vet, vetting. +[2836.46 --> 2840.24] And it includes some other things like a cyclomatic complexity analysis. +[2840.24 --> 2846.66] And there's a bunch of other nice sort of basic ads in there as well. +[2846.76 --> 2851.14] And I sort of based on these things, it gives your repository a grade, right? +[2851.34 --> 2853.82] I think on a scale of A3F or something like that. +[2854.28 --> 2855.74] So I find that, you know, useful. +[2855.74 --> 2863.58] So especially when I'm evaluating a repository, a package, third-party package to, to, to determine whether I'm going to use it or not. +[2864.18 --> 2867.32] If it has a score, I will look at that. +[2867.64 --> 2872.54] If it's, if it's anything other than A, then I'm going to take a closer look, right? +[2872.54 --> 2879.28] I'm going to be a little bit more hesitant with sort of bringing it in because I'm like, okay, well, what, what best practices, what idioms are you not following? +[2879.48 --> 2879.58] Right? +[2879.58 --> 2880.76] So I'll take a look at that. +[2880.76 --> 2890.82] And, you know, sometimes, you know, I may just, you know, sort of see what's happening and maybe replicate locally without having to bring the package if I don't like the score, so to speak. +[2890.82 --> 2895.46] So it's kind of a, it's, it's a nice, it's sort of a, another data point, right? +[2895.46 --> 2901.70] So to speak, to help you sort of evaluate the quality, I should say, of a repository. +[2902.02 --> 2904.18] But yeah, it's, it's one of the things I like to see as well. +[2904.48 --> 2905.68] The same for Godoc. +[2906.54 --> 2916.58] Godoc is a tool you can run locally, but we have also the godoc.org hosted service, which lets us view documentation for any open source project. +[2916.58 --> 2920.24] And so, yeah, I think that's, that's also nice. +[2920.38 --> 2923.92] It's a nice way to provide that capability because it makes sense. +[2924.00 --> 2925.52] You want to share just a link. +[2925.90 --> 2934.52] And the nice thing is for Godoc, it's just godoc.org slash, I think maybe pkg slash then the import path. +[2934.68 --> 2938.00] So again, you're still referring to that import path and we see it. +[2938.00 --> 2942.68] I personally use a lot of tools from Dominic Honneth. +[2942.88 --> 2960.14] Like he has this GoTools repo, static check tool, which, you know, contains a lot of like, you know, style check, a lot of like linting type of like, you know, features that GoLint doesn't support. +[2960.14 --> 2966.04] And it's, you know, there are some cases, sometimes like there's a controversial style topic. +[2966.38 --> 2969.10] It's not possible to, you know, merge it into the official tool. +[2969.22 --> 2973.82] So people would just go and like, you know, put it in the GoStatic tool. +[2974.66 --> 2980.62] So it's, it's a really useful to, you know, tool to take a look in terms of, I think, static tools like that. +[2980.70 --> 2986.28] I just rely on, you know, static check more than GoLint. +[2986.28 --> 3002.80] Hmm. Yeah. And, uh, Fatih Arslan, um, he made a service, um, which I think is called Fixing Me, which is, it's a kind of a, it's a, it's a GitHub integration as I understand it. +[3002.80 --> 3010.90] And it analyzes, it does a bit like the Go report card, but it actually creates, you know, PRs with changes in it. +[3010.90 --> 3021.90] So it actually does the fit, it's sort of proactive, like you've got another member on your team that's cares only, you know, like the pedant who just cares about all the style rules and all that. +[3022.40 --> 3025.70] And, um, and that's a project I think it's worth checking out. +[3025.76 --> 3030.58] It's called, it's, it's Fix Me, it's spelled F-I-X-M-I-E. +[3030.58 --> 3033.12] Um, so have a look for that one too. +[3033.24 --> 3038.80] It's a similar kind of idea to the Go report card, but tightly and more tightly integrated into GitHub. +[3039.42 --> 3044.40] Has anyone here written any of the, any kind of tooling, static analysis or otherwise? +[3044.76 --> 3048.50] I only wrote some tools to generate some stuff like from an interface. +[3048.50 --> 3055.56] Um, well, these are also some static tools, like, uh, one common case is generating implementations of interfaces. +[3055.56 --> 3058.30] And there's like a lot of boilerplate, uh, plate. +[3058.46 --> 3064.54] So I wrote a tool that kind of like, you know, takes the interface and creates the, you know, the concrete implementation. +[3064.54 --> 3069.22] And then you just go and like, you know, feel the implementation, feel the methods. +[3070.04 --> 3076.18] Uh, and do, did you use the AST stuff and the parser and things to build? +[3076.18 --> 3076.46] Yeah. +[3076.54 --> 3079.24] I used, uh, whatever it was in the standard library. +[3079.34 --> 3080.36] It was not that hard. +[3080.64 --> 3087.92] Uh, it was not that like, I mean, good looking either, but like it was possible to, you know, get it done in like a hundred lines or something. +[3088.64 --> 3089.08] Yeah. +[3089.30 --> 3089.56] Yeah. +[3089.66 --> 3091.80] So, okay, cool. +[3091.92 --> 3096.80] Well, I think we should also spend some time talking about some of the performance tools as well. +[3097.12 --> 3099.54] Um, that, that we just get for free. +[3099.54 --> 3101.52] There are some great talks on YouTube. +[3101.52 --> 3104.32] It's quite a, it's quite an interesting subject. +[3104.32 --> 3106.98] And it's talked about quite a lot in from different angles. +[3107.66 --> 3116.16] Um, but perhaps Yana, you could tell us a little bit about, did I see you do a talk about the performance tools? +[3116.30 --> 3122.32] It might be possible because like I worked on, uh, you know, some of the dynamic tools, uh, when I was working on Go. +[3122.32 --> 3124.82] So it was part of my full-time job. +[3124.82 --> 3129.44] Um, and I generally have been, you know, working in this area for a while. +[3129.86 --> 3134.92] So it's possible that you have seen me giving a talk, but I can't remember because I'm giving too many talks nowadays. +[3136.76 --> 3138.90] I thought it was all confidential what you work on. +[3139.36 --> 3140.10] It's not. +[3140.26 --> 3142.32] Uh, so the confidential stuff is different. +[3142.66 --> 3143.36] Oh, what's that? +[3143.54 --> 3145.24] None of my performance tools. +[3145.24 --> 3148.54] It's more about like computing, you know, products. +[3149.02 --> 3149.38] Right. +[3150.60 --> 3152.58] We'll figure out in a couple of weeks. +[3154.42 --> 3155.50] I'm just trying to, yeah. +[3155.54 --> 3161.18] I'm just trying to be like a, one of those journalists, hard hitting journalists that tries to get out the information that you don't want to say. +[3161.80 --> 3164.86] But if they're too polite, you just say, I'm not going to talk about it. +[3164.88 --> 3165.90] And I go, oh, okay. +[3166.18 --> 3166.46] Bye. +[3166.46 --> 3169.44] Well, the problem is I really don't know. +[3169.52 --> 3176.24] Like I know generally what I'm going to be working on, but I don't know the specifics and I'm a really precise person, I think. +[3177.00 --> 3177.20] Yes. +[3177.38 --> 3184.30] I don't want to like give any impressions that I'm going to work on something that like I'm not going to, because people will be upset. +[3185.08 --> 3185.32] Yeah. +[3185.78 --> 3187.40] It's absolutely fair enough. +[3187.40 --> 3187.64] Just joking. +[3187.78 --> 3188.04] Yes. +[3188.72 --> 3194.68] So, but Yada, could you tell us a bit about the, some of these tools though and what they're for, for anyone that doesn't know about them? +[3194.68 --> 3195.00] Yeah. +[3195.00 --> 3198.96] I think generally speaking, I think let's go beyond the performance tools. +[3199.34 --> 3203.54] There are a lot of like dynamic tools in Go and they are a part of the standard tooling. +[3204.12 --> 3205.96] Some of them are related to performance. +[3206.18 --> 3209.18] Some of them are more related to like debugging type of stuff. +[3210.30 --> 3216.64] The typical, you know, let's, we, we can talk about, for example, performance initially. +[3216.64 --> 3233.52] And the Go came around when it first came around, it came around with some of the dynamic tools because we went to the SRE team and SRE team is at Google is just really specific about what they want to put in production. +[3233.52 --> 3236.84] So they want to have like, you know, enough visibility into things. +[3237.06 --> 3240.52] And one of, some of these were related to, you know, performance. +[3240.98 --> 3243.60] They want to be able to, you know, get the profiles. +[3243.76 --> 3250.76] They want to get like some runtime traces because they specifically want to be able to understand when there is something going wrong. +[3250.76 --> 3253.10] Again, like they want to be able to pinpoint to those. +[3253.10 --> 3262.40] So PPROF support was baked into, you know, Go since the early times because of that requirement, for example. +[3262.88 --> 3264.10] It provides you some profiles. +[3264.92 --> 3268.08] You can also add your custom profiles, which is a useful topic. +[3268.84 --> 3277.34] But, you know, it provides the CPU profile, memory profile, you know, Go routines and thread profile and new tech contention profile. +[3277.34 --> 3291.86] And it was really crucial, you know, to have a language mature enough to put in production because that's basically most of people think that like performance is about development time. +[3291.96 --> 3294.02] But it's also important in production time. +[3294.34 --> 3301.42] On top of like PPROF support, there is like, you know, good benchmarking support baked into Go test. +[3301.42 --> 3309.08] So benchmarking is a first class citizen, which is not really, you know, quite the same situation in other languages. +[3309.74 --> 3315.74] And I think it kind of creates this culture where you care about, you know, benchmarking stuff, right? +[3315.78 --> 3328.24] Like, I don't know what is your opinion on this, but I've seen, you know, lots of different communities and different opinions about benchmarking just because of the, you know, the tooling or, you know, it's really easy to write benchmarks or not. +[3328.28 --> 3329.10] What do you think about it? +[3329.10 --> 3334.36] Well, I've seen it used perfectly and I've also seen it used incorrectly. +[3335.10 --> 3346.28] I've seen an example where the benchmark, just because of slight issue with the way it was written, it was reporting completely incorrect results. +[3346.74 --> 3355.32] So, yes, I think one, but if it's used in the right way, because like, you know, if it's, if you, depends on what you're testing, I suppose. +[3355.32 --> 3364.32] If you're going to be testing something and you're making HTTP requests, for example, there's so much variation anyway in HTTP, you're not really going to be getting any meaningful information. +[3364.32 --> 3374.10] But if you're, if you've got two little algorithms and you want to know which one's better at certain tasks and stuff, then yeah, it's, it's, it's great. +[3374.16 --> 3375.44] And I agree with you, Yannick. +[3375.44 --> 3386.92] I love the fact that it's baked straight into the language and you just have to write a function that starts with, you know, funk benchmark name, take in the special variable. +[3386.92 --> 3403.44] And as long as you get the for loop inside it in the right place and also think about setup and tear down work and where that's happening, then yeah, it's a great way to really just find out which is better because sometimes it'd be really surprising. +[3403.44 --> 3416.24] In fact, I think it would make a great talk if someone out there wants to do it or a presentation of like, here's, here's some code, which one's the fastest and have people kind of guess. +[3416.92 --> 3421.06] And sometimes I find it, the results to be very surprising. +[3421.06 --> 3421.36] Yeah. +[3421.66 --> 3422.00] Yeah. +[3422.00 --> 3431.98] I think benchmark in general is a discipline that, you know, takes a lot of time to kind of like learn and what are the, you know, the other factors that actually improve, you know, impacts the performance. +[3432.30 --> 3447.14] So I agree with you that like, I've seen a lot of like wrong benchmarks and people are like super strong opinion that it's actually an optimization, but it's actually like, like one specific thing that improves the performance, maybe like for one specific case or something. +[3447.14 --> 3447.26] Okay. +[3447.62 --> 3458.10] And I think you need to have a really good understanding of the runtime and everything around the language in order to write good benchmarks as well as like interpret the results correctly. +[3458.10 --> 3459.78] So it's, it's a really tough game. +[3460.96 --> 3461.54] That's true. +[3461.54 --> 3473.54] Would you say that if, so when it comes to benchmarking and performance optimization, like I try very, very hard not to sort of jump to that sort of right away. +[3473.54 --> 3478.66] I, you know, I'll try to solve a problem first and then, and then try to optimize. +[3478.78 --> 3478.92] Right. +[3478.96 --> 3483.98] So basically preventing premature optimization and these tools make it because they're part of the center, center tool chain. +[3484.06 --> 3489.42] They make it very easy to just, you know, start using them like right then and there, especially start, start leveraging right away. +[3489.42 --> 3500.24] Um, and there, there, there was a time, maybe we're still in that time where, you know, it seems like there was a new sort of HTTP, like MUXer router coming out every, every couple of weeks. +[3500.24 --> 3509.24] And they were all like, Oh, benchmark compared to this, these other things, you know, this one is zero allocation and it's, you know, 0.05% faster than the other one. +[3509.24 --> 3519.12] So it was, I was, I kind of found it silly a little bit, um, because of all that sort of going on and, and I was like, okay, we're kind of missing the point here a little bit. +[3519.36 --> 3522.58] Um, but yeah, I mean, it's, it's having that tool. +[3522.80 --> 3523.84] I think it's great. +[3523.92 --> 3532.14] You know, and like you, I don't think, I don't think I've seen that busy, that kind of capability sort of built in, you know, part of, part of the language from, from, from the start. +[3532.14 --> 3546.94] So the, I tend to sort of be very, I'm very careful with that because it's too easy to, to have, to create a culture within an engineering team of, of, okay, before I can even ship this thing, I have to make sure it's like super optimized. +[3547.44 --> 3552.00] And, and we're putting kind of the cart before the horse a little bit there. +[3552.26 --> 3553.80] It's too easy to do that. +[3553.86 --> 3556.34] So I tend to be, I tend to shy away from that stuff. +[3557.02 --> 3559.12] Um, you know, I sort of bring it in when I need to. +[3559.34 --> 3560.40] I completely agree. +[3560.40 --> 3565.66] I think, um, you know, optimizations in development time is kind of like fabricated problems. +[3565.66 --> 3570.06] Like, I mean, you realize what needs to be optimized in production, right? +[3570.14 --> 3580.48] Like, um, for example, what we do is, um, continuous profiling, which is we keep collecting some profiles from the, you know, the production binaries. +[3581.00 --> 3590.38] And we sort of like have an understanding of like, you know, within this project, what are some of the hot calls and what is like some of the, you know, stuff that is in the critical. +[3590.38 --> 3594.42] Like, and what critical paths are like more often being cold. +[3594.42 --> 3609.16] And like, you know, what happens if I just optimize this functional or like, what is the, you know, actual cost of this particular function in the, you know, the, the, if you think about the whole system and, you know, depending on the usage and whatever. +[3609.16 --> 3615.16] So I think it just really makes more sense to start thinking about these cases in production. +[3615.16 --> 3623.68] And like, by looking at the data, you just go back to the development environments and like, try to optimize those things and, you know, keep using these tools. +[3624.24 --> 3631.20] Uh, one of the nice things about Go profiling, the actual P-Prof is like, it's a really low overhead type of profiling thing. +[3631.20 --> 3633.12] And, uh, you can enable it in production. +[3633.12 --> 3642.82] So you can, you know, just keep, you know, getting, uh, profiles from production without impacting the critical paths so crazily, but there's a overhead. +[3643.12 --> 3644.98] Uh, but you know, there's some strategies. +[3644.98 --> 3655.70] If you have multiple replicas of a web server, for example, you can enable production maybe for like, um, one minute or five minutes, uh, on one replica. +[3656.14 --> 3665.72] And it's just like, sort of like, you know, depending on how much latency you will, you know, experience, uh, it's sometimes doable and that's what we do it. +[3665.72 --> 3674.20] Uh, that's how, you know, what we do and, um, just try to optimize based on the usage and, you know, what is the critical usage and like, what are some of the hot paths? +[3674.20 --> 3680.16] Like identifying those hot paths is also very important before jumping into any, you know, optimizations. +[3680.50 --> 3680.60] Right. +[3680.70 --> 3682.32] Having a problem before you solve it. +[3682.86 --> 3683.14] Yeah. +[3683.60 --> 3683.86] Yeah. +[3684.20 --> 3694.06] So Yana, when you say you do continuous profiling, when, when you deploy services, do you have P-Prof already enabled in there and you just switch it on or. +[3694.84 --> 3695.08] Yeah. +[3695.12 --> 3696.06] Think about like this. +[3696.06 --> 3701.08] Um, so although, you know, the Powerform P-Prof tools, uh, P-Prof is, can be tweakable. +[3701.08 --> 3705.52] Like it's dynamically turned, you can turn it on dynamically and you can turn it off. +[3705.68 --> 3714.96] So what we do is basically turn it on for like several minutes, uh, collect the data, just get the data and, you know, just parse it, store it. +[3715.04 --> 3717.38] And then we aggregate all that data. +[3717.68 --> 3721.14] And we have this like, you know, daily, weekly, whatever reports. +[3721.14 --> 3736.78] And you can take a look at like, oh, this service, particularly this handler is often used and all these like particular functions are, you know, it is, uh, accounting for the like most CPU time or memory or whatever. +[3736.78 --> 3741.78] And you can just go and like dig and like, you know, optimize those particular places. +[3743.12 --> 3753.34] Um, I, I wish that go ahead, like some tools around maybe supporting this type of like more continuous integration, uh, sort of continuous profiling, uh, features. +[3754.06 --> 3759.96] Um, you know, it's possible to write a tool kind of aggregates, you know, multiple P-profiles. +[3759.96 --> 3773.26] Maybe it could be possible to write like a library that automatically, you know, just turns it once a while reports to some like central service and then, you know, turns it off and so on. +[3773.40 --> 3777.34] Um, I think there's like some, some, some, we can do much better in this field. +[3777.34 --> 3786.62] It's just kind of like up to the user right now to plan and, you know, design and just, you know, do this type of things. +[3786.62 --> 3787.90] But like, that's basically what we do. +[3787.90 --> 3793.96] I, I gave some, I wrote, uh, on this topic, um, for a while. +[3793.96 --> 3801.20] And, um, it's just really like some companies are aware of these methodologies and some companies are not. +[3801.20 --> 3809.02] And like, it's just, it would be so nice if community was, you know, producing more best practices as well as, uh, more tooling around this. +[3809.56 --> 3810.54] Well, there we go. +[3810.60 --> 3812.10] There's the call gone out. +[3812.24 --> 3817.00] Anyone who's looking for a new open source project or something to hack on? +[3817.00 --> 3818.32] What a great problem. +[3818.68 --> 3833.76] Um, could you build something that samples running Go code and, you know, periodically at some schedule and, uh, and collects the results up, um, would be extremely useful and really fun probably as well. +[3833.76 --> 3834.26] Yeah. +[3834.26 --> 3835.54] It's like a lot of fun ones. +[3835.54 --> 3841.00] Once you start to see like, for example, a large company, you know, aggregate and all the profile and data. +[3841.00 --> 3847.82] So you can see like, Oh, the companies, you know, the, for example, you can actually improve your bill on your cloud provider. +[3847.82 --> 3852.26] Uh, you can say that like a loss of the calls are actually like dependent on this one function. +[3852.26 --> 3858.58] Um, and you know, if you optimize it, we can actually cut the billing, like by 10% or something, right? +[3858.58 --> 3863.18] Like it's actually pretty useful once you start to do this systematically everywhere. +[3863.66 --> 3864.22] Hmm. +[3864.22 --> 3870.88] Well, I love, I love the message of when wait till you've got something running and then look at optimizing it. +[3870.94 --> 3877.50] I think in some cases you can shortcut it, but generally speaking, yeah, that advice is sound. +[3877.50 --> 3889.06] And the idea of being able to profile running production systems and to understand them better, I think is, is, um, a great goal to have. +[3889.06 --> 3895.94] And what a great use of the, the tools that we, that we have as part of our ecosystem. +[3897.10 --> 3899.42] Well, on that bombshell, I mean, I think that's it. +[3899.50 --> 3902.82] I think we've, uh, we've reached the end of the hour. +[3902.82 --> 3909.16] And so the end of this episode, thank you very much, Johnny and Jana. +[3909.34 --> 3909.98] It's been awesome. +[3910.42 --> 3911.68] How have you liked it? +[3912.00 --> 3918.36] I can talk about this topic for hours and I think, you know, this was awesome, but we should keep, you know, talking about tools, I think. +[3919.30 --> 3920.06] Yeah, absolutely. +[3920.20 --> 3930.40] Well, there's a lot, there's lots more to discuss and, um, I might even see if we can bring in some people from the community that have, um, built some of the tools that we're using today. +[3930.40 --> 3935.66] One, one other little, uh, bit of info that I think is quite interesting. +[3936.12 --> 3944.02] The only actual contribution I personally made to the Go project was to remove something from GoLint. +[3944.38 --> 3949.26] So one time GoLint got a bit easier, uh, to satisfy. +[3949.58 --> 3950.32] Thanks to me. +[3950.48 --> 3951.18] You're welcome. +[3952.74 --> 3953.26] Yay. +[3953.58 --> 3954.02] Yeah. +[3954.20 --> 3955.80] I, I, I delete code. +[3956.06 --> 3956.74] I delete code. +[3956.82 --> 3957.98] I mean, I love it. +[3958.34 --> 3959.42] Well, yeah, that's it. +[3959.42 --> 3960.54] Thank you so much. +[3960.84 --> 3963.78] Um, we'll see you next time on GoTime. +[3971.42 --> 3974.68] Thanks for listening to this classic episode of GoTime. +[3975.12 --> 3976.12] Subscribe now. +[3976.20 --> 3980.16] If you're new, head to GoTime.fm for all the ways. +[3980.56 --> 3984.80] And long time listeners, do us a solid by sharing the show with your friends. +[3984.80 --> 3986.84] That's the best way people find us. +[3987.06 --> 3988.62] In fact, I will cut you a deal. +[3988.98 --> 3995.14] Email a personal recommendation to three friends and BCC go time at changelog.com. +[3995.26 --> 3997.50] I'll send you a free pack of changelog stickers. +[3997.84 --> 3999.00] That's a win, win, win. +[3999.16 --> 4000.06] With win, win, win. +[4000.30 --> 4000.92] We all win. +[4001.00 --> 4004.50] Thanks again to our partners at Fastly for having our CDN covered. +[4004.66 --> 4006.58] To BMC for these banging beats. +[4006.94 --> 4007.90] And to you for listening. +[4008.44 --> 4009.20] We appreciate you. +[4009.20 --> 4012.48] Next week, we welcome Ronna Steinberg to the show. +[4012.80 --> 4017.52] She joins Natalie and Ian to discuss object-oriented programming and Go. +[4018.12 --> 4021.56] That's what you can look forward to next time on GoTime. +[4021.56 --> 4033.84] Game on. diff --git a/Going through the news_transcript.txt b/Going through the news_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..6ce80a3f3ccb6e83d6f26408b9b1f2b3f206990b --- /dev/null +++ b/Going through the news_transcript.txt @@ -0,0 +1,343 @@ +**Natalie Pistunovich:** Good evening, morning, and afternoon, everyone, from wherever you're joining us. I'm Natalie, and I'm joined today by Ian and Kris. Hello. + +**Kris Brandow:** Hello. + +**Ian Lopshire:** How's it going? + +**Natalie Pistunovich:** I'm good. How are you all doing? + +**Ian Lopshire:** Doing well. + +**Kris Brandow:** It's a beautiful, but very warm day, the end of May. It's like 90 degrees outside, so it's quite sweltering. + +**Natalie Pistunovich:** Here, it's freezing. It's a single-digit Celsius this morning. + +**Kris Brandow:** Ooh. + +**Natalie Pistunovich:** Which is not far from snow, but far enough definitely not to happen. So today, this episode is our news episode, and you can see that because I'm holding a pen; I see all these news anchors do that. So we will be talking about what is new and interesting in the Go world, and we'll cover four interesting things from the recent news, with my interpreters for the field of growth. There goes the pen. + +Okay, so a first thing that many of us probably saw in an email or an announcement or in a tweet is that GopherCon US is moving from Florida to Chicago. And from the end of September to the beginning of October, specifically October 6th through 8th. And the reason of that being is that in Florida two bills were passed or in progress of passing; keep me honest, I'm not from the US. One is about what parents can say and not say, or schools can say in the context of gay people. I take that from the name, Don't Say Gay. And the second one is a law about abortions, that when things pass, then in Florida this will be not allowed, starting week 15. + +**Kris Brandow:** \[04:20\] Yeah. + +**Natalie Pistunovich:** So for those of us from not the US, can you say how much of my knowledge is up to date or correct? + +**Kris Brandow:** Yeah, the Don't Say Gay bill - I think it has been passed in Florida. It's one of those weirdly, vaguely-worded bills as well where it's just like, "You're not allowed to talk about gender identity in school", but clearly targeted at, "We don't want kids hearing about gayness and queerness when they're in school." So yeah, that's definitely happening. I mean, there's a lot of just-- many, many states are attempting to pass abortion bans. Many states in the South, but yeah, that's... And in their announcement, they went over both those things and said this is part of the reason why we're moving venues. + +**Natalie Pistunovich:** As a conference organizer myself, I can say that this is definitely an expensive decision. Just not doing a conference in the place where we plan to do it due to a simple reason like COVID back in 2020 - that was a huge headache, and definite financial implications to that, so I can only imagine what is it like in the US. What are your thoughts on that? + +**Ian Lopshire:** I think this is a good move. I think it's representative of our community, and I think it's kind of brave of them to do it despite the cost, and looking forward to going to Chicago. + +**Kris Brandow:** Fingers crossed as well, that by the time October rolls around that we're not in the middle of another spike with COVID, because I think that'll also be a thing that people have to kind of like wonder about now... Because it felt like we were through all of this, and they did an announcement to bring up one of the reasons they wanted to leave Florida as well is like the kind of lackluster - to put it nicely - a response that Florida had to the pandemic, and how absolutely crazy things got down there. + +So I think it's the same sort of thing of like, okay, well, even if it's in Chicago, if they're still like, "We have a giant COVID spike and a lot of our travelers are international, are they going to be able to get into the country? If they do, are they going to be able to get back into their home countries easily? So there's a lot of questions around in-person conferences. But I am excited that they are still trying to do an in-person conference at the end of the day. I'm glad it's not just like, "Okay, well, we can't do it in Florida, so we're going to do another virtual conference, like we did last year." I think it's good that they're trying to do an in-person thing. + +**Ian Lopshire:** Yeah. I do realize that in-person conferences are a huge risk. I think we just saw that with the Kubernetes Conference, and all that. But there is not a replacement for the in-person. I wouldn't know you, Kris, if I didn't go to one of these conferences. There's just not a replacement for it, so I am excited that they're trying to do it. + +**Kris Brandow:** Yeah. And hopefully in future years we won't have as many problems... With COVID, at least. + +**Natalie Pistunovich:** Yeah. Yeah. So I heard about the Google DDoS. What happened? ...for those who have not read about that. + +**Kris Brandow:** Yeah. So apparently, this is a knock-on effect of the design of modules. So for those of you that don't know how the internals of-- I hear you sigh, Ian... \[laughter\] So for those of you that don't know what came with the module system is we got this thing of module proxies, and there's the big one that everybody uses by default, that's owned and operated by Google. And essentially, what it's doing is it's going out and it's doing a full Git clone of repositories that are in the module proxy index. + +And there's this one Git hosting site called - what's it? SourceHut... And they've been getting-- about 5% of their overall traffic is just the Go module proxy hitting them and downloading the Go Git repositories that are hosted on their surface, which is a tremendous amount of traffic. I mean, at the end of the day, the thing you pay for, I think more than anything else when you're hosting Git is bandwidth and is traffic. + +\[08:02\] So basically, they're just getting DDoSed every day by this module proxy that exists. And I think this is another thing... I promise I'll get off my inclusion soapbox eventually, but this person didn't just write a blog post whining about this, they did try and actually go to the maintainers of Go and be like, "Hey, can you please fix this?", and they were basically told, "This isn't our priority. We're not going to do anything about this. Sorry." They're just stuck paying all of these costs at the end of the day for this bandwidth, because they can't just block the Go module proxy, because then any Git repositories that are hosted there just would stop working. They wouldn't get updates, or things like that. + +**Ian Lopshire:** Yeah. Google's answer to this is like, "Hey, email us and give us your domain and we'll block the periodic refresh to your servers. We'll stop doing that. But you have to submit a ticket or email us." It doesn't seem scalable. + +**Natalie Pistunovich:** Or DDoS with requests. + +**Ian Lopshire:** Exactly. And Kris, you said it was a result of the module system. I'm not sure I totally agree with that take. I think it's more a result of however Google has under the hood created this proxy. But yeah, SourceHut is seeing multiple requests for the same module, like multiple times a minute. So obviously, there is some redundant traffic happening that does not need to be, probably, at the end of the day. + +**Kris Brandow:** I think you're right there that it's not part of modules, but I feel like if modules had been designed slightly differently. Modules, at the end of the day, are designed to be agnostic of like Git, or any of this stuff. So I feel like it's like a small failing there that the proxies weren't designed in a way that didn't need to depend on Git as much as they are; or maybe my knowledge of the internal of modules is kind of old and rusty, so maybe it is designed in this way. Maybe it truly is just Google's proxy, that is the problem that needs to get fixed. + +But either way, this type of stuff, if you wanted to self-host your own modules via Git repository, this is going to hit you and you're going to have to pay the cost of this thing just coming back and just doing useless polls of your Git repository. Especially if you have large projects, it's gong to cost you a lot of money. + +In our way, it's a kind of like, if you're someone that doesn't necessarily want to put all your stuff on GitHub, or you don't want to put it on GitLab or any of these other centrally hosted services, you're kind of just stuck, I guess, dealing with this extra cost. It's like a tax that you have to pay for not wanting to be part of a centralized platform. + +**Ian Lopshire:** To me, this kind of brings up the issue like, why is the module proxy controlled by Google? Why is it close sourced? Why can't we go look and see why this traffic is happening? It seems at odds with the principles of our community. + +**Kris Brandow:** Yeah. Which is really weird too, because I remember in the early days, there was Athens, there were all these other projects that are still around, but just not nearly as active, that were there. So it does feel a little weird that it's just like this piece of infrastructure is owned and operated by Google. I mean, I think there's also plenty of other companies that would be happy to help operate this, because at the end of the day, it's not just Google that uses Go. I think, as pointed out in other articles we're going to talk about today, every cloud provider has Go at their core, so there's no reason why we couldn't have a module proxy by Microsoft, or by Amazon, or even by some of the other CDNs, like Fastly. + +To me as a gopher, it feels weird that Google has used its position to put something into the Go ecosystem that people can't really change. People can change it, but they just don't know about it. So by using Go by default, you're just opting into this module proxy, when they could have done a different way. They could have suggested that you use the Go proxy, the Google proxy, or given you some options, but they just chose to do it this way... Which kind of sucks too, because I was always one of those gophers that defended Go to such a high degree, because I was like, "Google does not control Go. Look at all of these people that are contributors to Go. The people that can actually submit, plus to a code review, the majority of them don't work at Google", although now that is also a thing that is dead and you can't do that anymore. You can still plus two, but you can't actually get the merge approved without having a Google employee sign off on it. + +\[12:19\] So there's these steps that have been happening that just make it feel like Google is trying to wrangle control over Go again, which just doesn't really feel good for an open source community project... Especially when their answer to, "Hey, there are these problems with this thing that you built that you are unintentionally forcing everybody to use" is "Well, sorry, we aren't going to put in the resources to actually fix that problem for you." I think that you can't have it both ways. It's really not cool to say, "Well, we're going to have this control over everything, but we're also not going to invest the resources into making sure that it is not harming people actively." Sorry... Okay, I promise they will get off my soapbox and stop ranting about things. + +**Break:** \[13:00\] + +**Natalie Pistunovich:** The next thing we will chat about is the article that reflects about Go, and was written by the five creators at the Association of Computer Machinery magazine. And that article was recapping who were the early users of Go, why the users chose to stay, and what were the design decisions that made Go what it is. And for anybody who's familiar with the language, you're probably not going to hear anything new, but it's still interesting to recap this. So who do you see, the early users? Why did they come? What did they like about the language? + +**Kris Brandow:** There is one thing I want to point out real quick before we jump into it - I did find it interesting that Russ Cox and Ian Lance Taylor are included here as the core people that helped create the language, because I think it's always like... You know, when you think about it, you think it's like, okay, Rob Pike, Robert Griesemer, Ken Thompson - those are the creators of Go. So I think it's nice that some of those people that did a lot of the early work... I mean, the reason we have a bootstrapped compilers is because Russ spent a lot of time figuring out how to get all of this massive C codebase into Go. So I do want to call out that it's really cool that they included these other two people that have been hugely influential on Go from the beginning. + +To answer the question you had, I think the early users of Go I think were probably a mix of people that were really interested in this idea of a programming language that has this core of simplicity. I think around the same time the early versions of Go started to become popular was the same time that Rich Hickey had his talk called Simple Made Easy. + +\[16:13\] There was a lot of thought in the software engineering ecosystem and world about how do we simplify software? How do we make it easier to build things and not have all of this complexity that comes with other languages? So I think -- at least I know for myself and a bunch of people I know, that was one of the things that drew us to Go in the early days. + +**Natalie Pistunovich:** That would be my next question - when and why did you join? + +**Kris Brandow:** Ian, do you want to take that first? + +**Ian Lopshire:** Sure. I think-- + +**Natalie Pistunovich:** I wanted to hear Ian's first. If you want to answer then also-- if you want to share in addition, who do you recognize as the early users would be interesting, too. Or why did they come-- what did they like in the language? Some of the early people. + +**Ian Lopshire:** Yeah, so I don't have much familiarity - I can't even say the word - with the people. I think I came to Go a little bit later, like Go 1.8 maybe, 1.5, something like that. The way I came to it was basically my company said, "Hey, you've got the choice between Go and C\#", and I went Go... I mean, for a lot of the same reasons - just the simplicity and the... Yeah. + +**Natalie Pistunovich:** Yeah. + +**Ian Lopshire:** Yeah. + +**Natalie Pistunovich:** Yeah, yeah. + +**Ian Lopshire:** Sorry, that was the worst answer I've done so far. + +**Natalie Pistunovich:** No, that's a perfect ramp to my next question... But first, Kris, do you want to share when and why did you join Go? + +**Kris Brandow:** Yes. I think I've shared it a number of times, actually. So the main thing is when I started my career, I was writing a lot of Drupal, PHP. It was very, very, very complex. And this one day I went to this full-stackgineering meetup where I met Sam Boyer. And Sam and I were both in the Drupal community at the time, and I was describing to him how frustrated I was with Drupal. I was like, "There's all these things that are annoying, and I don't like it", and he looked at me and he said, "I know what that is. That's complexity. You don't like complexity. You like simplicity. You should go try out Go, and you should go look at this language and see how you like it." And that's literally-- + +**Natalie Pistunovich:** You should go test it. + +**Kris Brandow:** Go test Go. That's literally how it happened. It's just like, Sam Boyer-- I just met him one night and he was literally like, "Hey, I think you'd like this other thing. I understand your frustrations." And then basically, I think a couple months later-- or not even a couple, I think a couple weeks later, I was like, "I like this language. This is going to become my professional language now. This is going to be language that I write the majority of the software that I write going forward." So it was just -- not love at first sight, but definitely something that I just was like, "Yeah. No, this is great. I absolutely love it. Let's Go." Oh, God, I'm going to say Go so many times, in different ways... Let's go...! But yeah, that's how I got into writing Go. + +**Natalie Pistunovich:** So Ian and Kris, you've both mentioned simplicity, and that's very interesting. I had a conversation last week with VP on R&D of a Berliner startup that is doing mostly B2B things. And I always like asking people if they use Go, and if they say no, whether they evaluated that, and if they say yes, then I'm always curious to hear why not Go. So his answer was that a couple of years back - five years or so - he was a developer at a company that was using Go, and he even attended one of the meetups in Berlin, and his memories and his impressions from the language at the time is that this is like a super-optimization language and, "Check out my talk of how I changed two lines of the compiler and shaved off a few more milliseconds..." And then he said, "And you know, overall, yes, SLAs were way, way far from that, so we don't need the fanciness of Go. We went for something simple, like Python." And then I said, "But wait", and he's like, "I know you're going to pitch Go to me. I know you." But this was very interesting for me, that for him-- and he sampled Go back then, and did not check on it again. And for him, this is like a fancy language, for great optimizations and actually not something with simplicity. + +**Kris Brandow:** \[19:59\] I think it comes down to how you see simplicity, because I think simplicity is one of those words that's like really, really slippery. Python, along a lot of domains along a lot of axes, is a simple language, but it's not simple in the way that Go is simple, right? Go is simple in-- if you haven't seen it, Rich Hickey's talk Simple Made Easy, very, very good talk, and I think it kind of defines it well. Simplicity doesn't necessarily mean that it's easy, or that you can get started quickly. In fact, simplicity, a lot of time means that you're going to wind up getting started slower, but it means in the long run, things remain simple; that it's harder to introduce complexity, right? It's simplicity as anti-complexity. And I think if you look at most other languages, you wind up with just lots of complexity after a few hours or days of coding. Whereas in Go, I've found that it's like the language pushes back on you a lot if you try and do things in an overly complex way. + +I think for a lot of years, there was this kind of marketing of Go that was like, "Oh, we have goroutines and channels." I remember even myself sitting down and writing tons of code with channels all over the place, and it felt awful; like, absolutely awful. And I was like, "Oh, I've made this mistake. I will not make this mistake again." Channels should be used judiciously. They should not be just sprinkled all throughout your code base. And I've learned that lesson and I haven't done that since. + +I think other languages probably would've made it so that you could keep just using channels everywhere, because like "Hey, this is like a core feature of the language. It's like a nice thing. We'll just bolt some more features onto channels." I think there's been multiple proposals to make things like error handling with channels easier. And it's just like, no, because that will violate the higher-level simplicity of Go as a whole. Channels are meant to be this simple object. If it doesn't fit your use case, you're probably using the wrong thing, and there might be something else that's a better fit for you. + +So it's simple in that way, not necessarily that it's simple to get started with, or simple to bootstrap an application with. There are certainly other languages that are easier for that. It's also quite simple to learn at the end of the day as well. I think most people can at least pick up a majority of the syntax in a weekend or in a week. It does take a lot longer to become an expert, but it's easy to pick up that early language syntax. + +**Ian Lopshire:** I saw some comments on -- I think it was on this actual article, where part of the simplicity is also the density per line. Every line is simple. Does that make sense? There's not these 40 ternary statements long, where I don't know what's returning. That's one of the big parts of the simplicity to me, that all the code is the same density. It's easy to look at each line. + +**Natalie Pistunovich:** Even the format thing being the same everywhere. + +**Ian Lopshire:** Exactly. + +**Natalie Pistunovich:** Always dive to that familiar pattern. + +**Ian Lopshire:** Exactly. And they kind of talk about that in the article, where they say part of the reason it stayed around is the environment, right? All the great tooling, all the great -- like Go format, Go... + +**Kris Brandow:** Yeah. I remember -- I don't know if it was a talk Andrew Gerrand gave, or if I was just talking to him, but I remember this comment he made about... No, I think it definitely was in one of his talks, where he was talking about how he went and looked at a bunch of Go on GitHub, and found out that all the open source Go they could see, 98% of it had been `go fmt`ed already, so it was already in that nice format. I remember that being another one of those early things that I was like, "Yes, this is nice." + +It was novel, very novel at the time for a language to ship with a formatter that was just very easy to run, and that you were expected to run, and there's no negotiation. It's like, "Oh, but I like to put brackets on the next line." "Well, too bad. You're not allowed to do that." And if you do that, someone else will just run Go fmt on the code base and will just correct it. + +\[23:50\] So that's another element of simplicity there, because then it's like if you go to any codebase you find in GitHub and you can just read it. You're not annoyed because the brackets are in different lines, or there's some other spacing issue that you don't like, that you would prefer to be a different way. No, no, this is just the way that Go looks. So I definitely agree with them that - yeah, the tooling around Go was absolutely one of the things that solidified it in the ecosystem. + +**Natalie Pistunovich:** Yeah. I'm doing these days more Python than Go, unfortunately... And one of the reasons is that, as you mentioned, Kris, that this is a quick language to just ship things fast, and then later deal with the consequences. Not necessarily very happy with that, but disagree and commit. + +I do feel that after so many years of doing only Go, the same pattern, always recognizing whatever codebase you're entering, like new company, existing company, this is just so confusing. I spend so much time that feels unnecessary or just realizing like, where am I? What is the fancy line doing? Like you said, Ian, each line is not necessarily always readable. And this is not because a bad developer wrote that, it's just because it's not necessarily a thing in many languages, not just Python. + +And I think that also the fact that Go is so consistent and always looks the same is very useful when we are in this stage, where tools like Copilot are being introduced into our world as developers. This is something that is super-easy to use with Go, because it will always do the right thing. But with different languages, like Python, it will sometimes do the right thing, sometimes it will be like a mix of, "This is too much", like too different input here, something that is a mix but doesn't really work, or just does not fit the codebase. So it's also one of the reasons I personally think that languages like Go will survive going forward, and other languages that are more a mix of things will slowly go away. + +**Ian Lopshire:** Yeah, that's a great point with the Copilot. I hadn't thought about using Copilot in a different language, and it just spits out code and you're like, "What is this?!" + +**Natalie Pistunovich:** You're from a different story. + +**Ian Lopshire:** That's just not how I do it. + +**Natalie Pistunovich:** Yeah. + +**Kris Brandow:** Yeah. + +**Break:** \[26:02\] + +**Kris Brandow:** We did get some nice commentary from Bill Kennedy in the Go Time channel. Shout-out to the Go Time channel. If you're in Gopher Slack, you should join GoTimeFM. That's where we're always hanging out when we're doing the live recordings for this. I'll just read what he said, and then we can discuss it a little. It was back when we were talking about the module proxy, and the issues with someone getting DDoSed by it. He says, "I don't see it like this. I see it as this system that provides security and durability for Go projects. They built a system which they run and manage for the community. They publish the API, and anyone can build a system such as Athens. They are not controlling anything. They are spending money and time to provide a system that must be stable and available. We should be thanking them for this. You know how hard it is to open source anything from big companies like Google." + +And I do understand that sentiment. I think I agree with Bill there, and I am happy that this exists. I think that for the community at large, this is great. Having a module proxy, making it so that you don't have to configure it is absolutely fantastic. I see how they arrived at the decision to just include it in the default Go binary that ships, and I think we should be now thankful and happy for them for building this. I do have another thought on this, but I want to give - Ian, Natalie, if you want to also respond to this a little bit... + +**Ian Lopshire:** Yeah, I do think it's really easy to forget how much monetary support these big companies do provide communities like Go, right? And I do think we should be grateful, in some regards. But in other regards, they're not doing this solely out of goodwill. + +**Kris Brandow:** Philanthropy. + +**Ian Lopshire:** Yeah, exactly. They are getting something in return, right? Having Go as a strong community provides them programmers they can hire, expands the job pool... It does a lot of different things for Google. So I also don't want to forget that it's also a benefit for them to have a strong community. We do should appreciate the support. + +**Natalie Pistunovich:** And also it's a good point of having something that is the stable thing to go to, pros and cons. + +**Kris Brandow:** Yeah. So I think my larger take on this is that I do agree, we should be thankful for them for putting out the effort to do this, but I also think that when it comes to underrepresented communities or underrepresented people, we're always like a small minority of things. So whenever something like this happens and it's a small minority of people that get affected, often it's just swept under the rug that this is a problem for some people. Like I mentioned, small entrepreneurs or people that have ethical or moral or they just feel strongly about not centrally hosting things, this is a cost for them. And as a community, by Google doing this, we've said to them, "You get to pay an extra tax for this", which is a way of saying to people, "You're not as welcome in our community." Which - if that's something we want to say because it's better for the greater good of Go, I think that's a fine thing. That's a trade-off we can make. + +\[32:07\] But I think the thing I was trying to point out here is that this is like an actual trade-off that happens, at the end of the day, and big companies have a habit of doing this sort of thing and harming underrepresented communities by doing this sort of thing. And if we just keep going on thanking them for everything they do, we don't necessarily resolve that problem at the end of the day... Which is obviously a tough thing to talk about, because once again, this is not a topic that people usually consider to be polite conversation, or things like that. But no, they are doing a good service, so it's like we should appreciate that... But at the end of the day, as someone from a plethora of marginalized backgrounds, it's tough for me when I see other people that are not thought of as marginalized, other people that don't get the spotlight as much are suffering, and they don't get necessarily the same amount of recognition; they just kind of have to put up and deal with it. + +**Natalie Pistunovich:** And it's more, responsive. Thank you, Bill, for being a member of this conversation in some way. This is interesting. + +**Ian Lopshire:** Bill does bring up something nice. It's a small team, very little runway... That probably is something we forget. Even behind these big companies are just people on teams. + +**Kris Brandow:** Yeah. And that's the same thing, too. It's like, oh, well, at the end of the day, the Go team is also a minority. They are very small. There's a couple million gophers and there's a couple dozen people on the Go team. So it's like, they need to be prioritized as well, and what they can do needs to be prioritized as well. So it's like, I think at the end of the day what I'm trying to express here is that there is no simple or easy answer to this question... Aside from maybe Google spending a bit more money on the Go team and questions about, I think, the thing I brought up in the first place, which is like, it's fine if you want to build this sort of stuff, but when it harms people, you've got to shell a little bit more money. I understand, you've shelled out a lot of money already and that's a tough thing to do, but that's part of how we actually do this thing. And that's one of the things that, at the end of the day, annoys me a lot. And once again, we're heading right into Pride months, so this is a big thing, which is when people don't want to walk the walk after they've talked the talk. Everybody wants to be like, "Yeah, we support queer rights", or "We support marginalized communities. We support diversity and inclusion", but then it gets hard, and they're like, "But this is hard, so now we don't want to do it anymore." And it's like, well, then you didn't really support the thing. That does damage to those people because they thought they had your support, and now they don't. And people might have gone to your company and they might have done something, they might have gotten themselves into a situation that they wouldn't have otherwise. So you can't actually do actual harm to people through these sorts of things. + +Thank you so much, Bill, for bringing this up. I think it's super-important to have a more balanced look. I can see how what I said was a little bit maybe too much on the other side of things... But yeah, and I also love this type of discussion that we get to have with our listeners in GoTimeFM. So if you're not in there and you're listening, go jump into that channel now. It's really great. + +**Natalie Pistunovich:** And a shout-out to Google for this. Reading what Bill wrote more as we were chatting, "Google has provided GoBridge over 200K in donations over the past three years, which is amazing and it always translated people coming to conferences or being able to do things that they would not be able to do otherwise." And the Go team has lost some resources over the past few months, and this last release burnt out a lot of the team members. Shout-out to you for being great. Thank you. It's hard to balance everything; this episode, how to manage your resources, everything. Engineering is all about trade-offs. + +**Kris Brandow:** Yeah. And we do have one more topic that we want to get to, so... + +**Natalie Pistunovich:** Which can maybe act as a potentially unpopular opinion, maybe yes, maybe not... A recent proposal by Thomas Eckert on lightweight anonymous function syntax. + +**Kris Brandow:** He pointed it out to us. I don't know if he made the actual proposal though. + +**Natalie Pistunovich:** Definitely worth-- + +**Kris Brandow:** He sent it to us. + +**Natalie Pistunovich:** ...checking again before I'm saying it wrong. Thank you for pointing this out. Maybe this can be corrected afterwards. This proposal seems back in the milieu. Let's chat about that in our remaining five minutes. + +**Ian Lopshire:** Yes. I think this proposal is getting a lot of conversation, because it's somewhat polarizing. Would you agree with that? Some people love it, some people hate it. + +**Natalie Pistunovich:** \[36:13\] That's why it's made our way to be our somewhat of an unpopular opinion... + +**Ian Lopshire:** Yeah. + +**Natalie Pistunovich:** ...for today, given it's all hosts today. + +Ian Lopshire:: Personally, I am a fan. + +**Kris Brandow:** I mean, especially the part that Robert Griesemer has done in the last 12 days - I love this. I think before I would've been kind of against it, but I think, once again, used appropriately, used judiciously, I think this could help clean up a lot of code. There's definitely some things I've done before where I'm just like "Okay, I'm writing all of this boilerplate function, definition stuff, and I have like-- the actual boilerplate makes up more of the function than the actual functiony parts of the function." + +**Ian Lopshire:** Should we take a second and explain what this proposal actually is? + +**Kris Brandow:** Yes. + +**Natalie Pistunovich:** Yes. And mention - thanks to Bill's feedback - that this is an old feature request that recently resurfaced. Yeah. + +**Kris Brandow:** Yes. So basically, Robert Griesemer, in this issue, which we'll have linked, brought up this idea of having basically a more concise way to define functions. Similar to the fat arrow syntax that is in JavaScript, he basically gave us like two different examples of how to do this... Mostly either a lot in the func keyword and having a fat arrow, or keeping the func keyword and removing parentheses around the arguments. And then also as a secondary part that's not something you have to do as part of this, but is related, is removing return, or not having to specify return, if there's a single value that's being returned from one of these more compact functions. + +**Ian Lopshire:** I am 100% against ever having implicit returns. There should never be a return without the return keyword, even in these small, anonymous functions. I would say no to this proposal all the way, just on that single point. + +**Kris Brandow:** Yeah. Maybe this will be an unpopular opinion because-- I don't know. When I look at the examples and I look at that code, I can obviously see, maybe if it was also just like, if it's multiline, you're not allowed to not specify the return... But when I look at this compact code, I'm like, "Oh, but this is like... It gets to the point so much quicker." I understand it so much easier. But I can also see how it could be like, just like you shouldn't have naked returns where you don't specify the return values because you've defined them in a function definition, I think that if you have a longer, like more than one line, or more than maybe a couple lines, you should have to specify the return. + +So I think that's one of the things-- I don't know if it's... I think -- yeah, in JavaScript, you don't have to specify the return. It'll just return the last thing, and that's always confusing to me, because I'm just like, "Where is this return value?" I'm like, "Oh, right, it just does it magically for you." + +**Ian Lopshire:** There's some talk in the issue on GitHub about changing the syntax, so you can just not specify the function parameter types or return types, just the names instead. I think that actually goes quite a ways to get there. We already allow for some of the generic stuff. You can omit generic types, and all this. So I'm not sure, it takes away too much from understanding, but I do think it's a slippery slope here, right? We've got to be careful not to make things opaque. + +**Natalie Pistunovich:** To summarize this, I want to read yet another message from Bill, which I really like, "Let's not make things easy to do. Let's focus on making things easy to understand." This circles back very nicely to our previous conversation about simplicity and how nice it is in Go. I like that about the language. It's great to keep, for sure. + +Let's say we had the tune for unpopular opinion now... We kind of discussed that, although it was quite popular, but there's all sorts of opinions out there. And this has been an interesting hour. I think it's first time we tried this format, at least from what I know, that we're kind of not going for a topic and discussing it, but actually just going over what happened recently. + +**Kris Brandow:** \[40:02\] I feel like we should have more of these nice little "Let's do some coverage of Go News and talk about things." Although, hopefully, in the future, it's not me being on a soapbox. + +**Natalie Pistunovich:** So what is a soapbox, for those who still reside outside of the US? + +**Kris Brandow:** A soapbox, quite literally, it's an actual crate they use to ship soap in, but the terminology comes from when you used to have a little town square, where someone wanted to give a speech-- + +**Natalie Pistunovich:** Hmm... Like a Hyde park thing. + +**Kris Brandow:** ...they would put a soapbox down and stand on top of it, so they could be above everybody else, so everybody could hear them in the crowd. So it's like when you get up on a soapbox, it's kind of like preaching, or basically talking to a gathering of people. + +**Natalie Pistunovich:** Lots of learnings in today's episode. Thank you. + +**Kris Brandow:** Yeah. I like the diversity of this panel too, because it's like, we've got you, who's not in the US and has never lived in the US, so you're not in tune with this stuff... And then we've got two people from the US. + +**Natalie Pistunovich:** Although I said I do speak some American, but... Nope, not enough. + +**Kris Brandow:** \[laughs\] There's a lot of American to learn. I think there's some people in America that don't speak American, so... + +**Natalie Pistunovich:** Kris and Ian and everybody who listened, thank you very much for joining. It was a great conversation, and see you next week. + +**Break:** \[41:18\] + +**Kris Brandow:** There was another thing that I wanted to point out about this, and it's a slight criticism, I guess, of the-- I guess it was the criticism of the way they were framing the conference in Florida, right? And it's around one key phrase that I think a lot of people don't understand as a dog whistle, but it is a dog whistle, and it's the phrase "family-friendly." I've seen it pop up a bunch of times in the Go community. So I want to have a little bit of a conversation around that, because it's one of those diversity inclusion things that I think is super important that we actually talk about. But for those listeners who don't know, family-friendly, in the United States, is a conservative dog whistle. And if you don't what a dog whistle is, it's basically a way of saying something that seems innocent, but is a way of basically crying out or rallying people that support you. It's a dog whistle for the conservatives that is basically anti-queer, anti-trans. It's the justification that a lot of states legislatures have used for bathroom bills and for even things like the "Don't say, gay" Bill, in Florida. So it's kind of startling to see that in a post about, "We're trying to be more inclusive. We're trying to be more welcoming of people", but then to just have that kind of anti-queer thing sticking in the middle of it; it's just kind of like, "Whoa, okay, this is sending some mixed signals here, some mixed messaging." So I don't know, I'd like to chat about that a little bit, and get maybe your perspective, Natalie, as someone who's not in the US, who's someone that probably hasn't heard this stuff before. + +**Natalie Pistunovich:** Yeah. We were just briefly chatting before the beginning of this episode, and when this came up, I told you, Kris, I was surprised to learn that. Definitely, as much as I interact with the US in my free time and in my professional time, somehow this went over my head, this term being signaling something like this. And I was also sharing with you that we were-- GopherCon Europe, we're planning to have the space and set up that is accommodating for families, accommodating for people to bring one partner, whatever gender it's from, multiple partners, kids, and so on. Probably now where we're talking about this between us in the organizers team as making this a family-friendly conference. So I learned something today, thank you. What is a better word to use to describe this type of setup? + +**Kris Brandow:** \[44:27\] I feel like either just pointing out that childcare is available, I think is one way of doing it, like "Hey, we have childcare", or just being child-friendly, or kids are welcome or something like that. Just make it actually about the thing that you're trying to include. Because you're trying to say, "Hey, if you have kids and you don't want to leave your partner at home with the kids, because that's a lot of burden on them, just bring them with you and we'll have spaces available to make it so it doesn't just feel like, "Oh, hey, you're just in this space with your kids now, and you don't get to enjoy this space." + +So I think it's like you're thinking about people that have that type of family situation and wanting to welcome them into the space. So I think just pointing out like "Kid-friendly" or "Kids are welcome" is a good stepping stone there, or a good first step. + +I'd love it if we could finally get to a place - this is along down the road - where it's just known that "Yeah, we're going to have childcare available because people have kids. So what are they supposed to do?" There's lots of single parents as well who have children. So it's like, are they just not allowed to come to conferences because they're a single parent? No, that's silly. So we should have these spaces, in general, for people. + +So yeah, I think if I read something else like, "Oh, kids are welcome", I think that would be positive. I think it's like a tough balance at the end of the day as well though, because whenever you choose to put the spotlight on something, you'll always unintentionally move the spotlight away from other things. So I can see some queer people feeling like, "Oh, this doesn't feel as inclusive to me", or still feeling like it's a dog whistle of saying "Oh, well this is a space where I can't be my whole self, because there will be children around." So there's always this weird, delicate balance that needs to be struck. But I think in general, it is good to have spaces to have childcare and have those sorts of things, because it is necessary for a large swath of the population. + +**Natalie Pistunovich:** I have multiple follow-up questions. + +**Kris Brandow:** Go for it. \[laughter\] + +**Natalie Pistunovich:** It's probably a very safe assumption to say that this word or this phrase in the announcement of GopherCon does not mean that, in my opinion. But I am not part of the queer community. What do you think? My opinion here is only a small thing, not being part of that, and this is not signaling to me. + +**Kris Brandow:** Yeah. It's definitely not intentional. I know the people that organize GopherCon. I've been a conference chair before. I do not think in any way that this was an intentional "We're trying to exclude--" I mean, it'd be stupid if it was anyway, because they're saying, "We're moving out of Florida because of this bill that's anti-queer." So I know they're not intentionally trying to do it... Which is good, but I think it's that unintentional side effect that these types of things tend to have on people that we have to start watching out for. + +At the end of the day, being inclusive is more than just trying and more of what your intentions are. It's what your impact is at the end of the day as well, so avoiding these types of things. There was a similar problem - it might still be; I haven't checked up on it recently... But with the Gopher Slack, where there was this real name policy, which is also another anti-queer, anti-trans signal that happens, because a lot of people that are trans don't go by their "real name" or whatever you would call-- whatever is a real name. + +**Natalie Pistunovich:** And then what does real name even mean? + +**Kris Brandow:** Exactly. So it's like those policies are also like dog whistles to conservatives saying, "Oh, yeah, yeah, you're welcome here. We're trying to make sure that the trans people aren't in this space, because we make it more difficult for them to exist in it." It's always a difficult thing, because I know at the end of the day, these people, their thoughts are in the right place. Their intentions are in the right place. It's just making sure that we have the impact, at the end of the day, that matches those intentions. + +\[48:07\] Just calling things like this out a little bit, or as some people calling in... This isn't meant to be like attacking GopherCon or anything obviously, or Gopher Slack, but just pointing out to people like, "Hey, you should probably avoid these types of phrases in the future, and you should probably get some more queer representation in the room when you're crafting these types of things as well." Because I don't know if there's many queer people that would've saw the words family-friendly for a conference base in the US and would not have raised their hand and said, "Hi, this is a problematic phrase. We should probably use something else." + +**Natalie Pistunovich:** Yeah. Another interesting thing to discuss is when you say-- I don't know how to phrase that... The dry meaning of family-friendly - does it mean friendly for families that don't have kids? Can you bring a partner, or if you have multiple, can you bring your partners? Is that part of what people would have in mind when saying that? And I can share that our thoughts in organizing GopherCon Europe is actually giving something like-- we have not done this yet. We're still considering. And it would be interesting to hear what people think, but what we're considering is giving sort of a discounted ticket for partners who want to travel with their partners who do want to attend, but the partner who's not interested in going will not be attending the conference talks, will have some dedicated space with WIFI so you can sit and have lunch with your partner, and then you both split on-- like you work in the same space with a person, but you only meet for coffee and lunch. + +**Kris Brandow:** Yeah. + +**Natalie Pistunovich:** Does that include as a family thing? + +**Kris Brandow:** That's amazing. I hope you guys actually do that, because that would be so awesome. But I don't think in many people's minds in the US, when you say family-friendly, that is what comes up. If I were to say-- even pushing away the conservative stuff, if I hear the term family-friendly, I'm specifically thinking about kids being included in this, right? Because I think in the US, there's this kind of thought that a family-- it's like, oh, if you just have a partner, you're like a couple; you're not really a family yet. You're a family once you have kids. And obviously, I do not personally agree with that definition of family, but I think that's the thing that gets bashed into our minds growing up in the US, it's like "Well, you're a family once you have kids, once you have grandkids, once you have this kind of bigger unit that's more than just you and a single partner." + +And I think when it comes to the polyamorous community, where there are multiple partners and all different types of defined relationships, I don't know if there's anybody in the US that would be like, "Yes, when we say family-friendly, we mean people that are poly." I just don't think that would be something that those people would believe or think. That shouldn't be included, but I think that they wouldn't see that under the umbrella of family-friendly. + +**Natalie Pistunovich:** Ian, what are your thoughts? + +**Ian Lopshire:** Yeah. I mean, I just think, especially in the United States, people have spent a lot of money over the last couple decades promoting the idea that a family is a man, a woman and kids. I think that's part of the problem with the phrase family-friendly, right? It's that it alludes to this specific ideal of what some people think life should be. And so, yeah, I don't think when you say family-friendly, it includes partners or polyamorous people; or at least I don't think most people will think of that. + +**Kris Brandow:** I don't even know if it includes single parents with kids really, because that's another thing in the US - it's just not really considered a family to a portion of the country. And I think that's why this phrase is such an awkward one. + +I have another suggestion as well, if you'd like it - just try maybe just listing out these benefits that you have in a prominent place, of saying, "Hey, we offer discounted tickets for partners, and you get these benefits." Just make it prominent, without necessarily having to wrap it in some phrasing. So I think that's where you get into the struggles of trying to wrap things in phrasing, and that's where you run into a lot of these like, "Oh, well, these words make sense when you just kind of put them together", like "Oh, we want this to be friendly toward families." But it's like, "Oh, no, there's all this underlying historical terminology that we have to go deal with as a result of that." So it's just like, "Oh, no, here's some of the benefits", call it out, like "Hey, traveling with a partner or partners? Here's some benefits that might be useful to them." Like, "Here's a discounted ticket and this is what they get for it." + +**Natalie Pistunovich:** \[52:22\] If/when you are in a relationship, would you find it interesting to traveling with a partner who is not interested in Go? + +**Kris Brandow:** I'll defer to Ian. + +**Ian Lopshire:** No, I think so... Especially if it's like in Chicago, I think a city that is popular for tourism - yeah, of course; especially if you're going to spend time before-- + +**Natalie Pistunovich:** Or any other conference also. + +**Ian Lopshire:** Yeah, exactly. But if you're spending time before and after, I think that's a great option. But also, just during the conference, I think it's a great option. I would consider it. + +**Natalie Pistunovich:** You work in different companies. Sometimes your companies are sponsoring at those conferences, including different GopherCons. How do you see companies support such a decision? Do you expect most sponsors to say, "Yes, we support that. We will bear additional costs with you. We will back out. We'll just stay with whatever we confirmed we'll do, you handle the additional costs"? What do we expect from the different sponsors to go about such a decision? Not about your employer in general, about-- I don't know whether your employer is even a sponsor. Generally, what do we expect from sponsors as people who work in companies who generally support the community? + +**Kris Brandow:** I mean, I think generally most companies are going to sponsor something that GopherCon wouldn't have a problem with this type of benefit or something. I mean, at least I hope they wouldn't. I think if they do have a problem, I think that's like a signal that, hey, maybe this is a sponsor that you might not want to have, and that's another-- once again, when you want to be inclusive, when you want to grow a diverse community, there are sacrifices that have to be made, and sometimes that's leaving money on the table from places that are not going to be supportive of the things that you value, right? So if you do value making it known that community members that have, say, multiple partners are welcome at the conference, and you have a sponsor that's like, "Well, I don't want to support that, so I'm not going to give you money for that", that's a choice you have to make. Do you actually support that, or were you just saying you support it, until it came push came to shove and you're like, "Well, I have to go leave this money on the table", so now it's like, "Okay, I actually don't support this type of thing"? + +So I think truly if you as a conference organizer - do support this sort of thing? And I think it applies broadly to the Go community as well, on things that aren't -- anytime it comes with money; even deciding to work at a company... It's like, when it comes down to money, that's where you really see what your values are at the end of the day. So it's like, are you willing to just let that money walk out the door and support this thing, or have to figure out more logistics? Because yeah, there is associated costs with-- if you really are giving someone use of a space, that's another room or another space you have to get, there's more cost for Wi-Fi, since you have more people on it. Obviously, if you're giving them food, that's like half the cost of your ticket right there, so that's a substantial cost. So it's like there is cost associated with doing this sort of thing. + +I hope that with the sponsors of things like GopherCon or GopherCon Europe, that they would not have a problem with calling out this type of arrangement. But yeah, I don't know. But yeah, if it was me personally, like if I was organizing a conference and a sponsor said, "Hey, I noticed that you have this benefit for people's partners", and having this discounted ticket for them to use the Wi-Fi and get some food, I think it's fine if you say your partner or your spouse, but if you allude to the fact they might have multiple partners, "I'm not okay with this", I'd say, "Okay, well, here's your money back. Have a nice day." I don't want to be doing business with someone that's drawing those types of lines around-- at the end of the day, what types of relationships are conference attendees allowed to be in? Why should you, as a sponsor, get to dictate that sort of thing? So yeah, I guess that's my rambly point of view on that. + +**Natalie Pistunovich:** It's particularly relevant because tomorrow begins the Pride month. + +**Kris Brandow:** \[56:04\] Yes. Or as I like to call it, queer new year. \[laughter\] We can't be contained to a month, so this is like the beginning of it, especially in the US too, because I think people have this assumption that's like, "Hey, Pride month is June, and there's Pride in June." I'm like, "No, no, no. There's Prides in June, July, August, September, October, November." I don't know if there's any in December, but there's certainly some in February. So it really is like the beginning. It's like queer new year. And then we hibernate for the spring months and then we come back and the same year again. But yeah, it's coming up. Queer new year is here. We have to prepare ourselves for the rainbow washing that will happen. I'm already mentally fortifying myself to it. For those of you who don't know, rainbow washing is essentially like whitewashing, but for queerness. + +**Natalie Pistunovich:** And for those who don't know what's whitewashing? + +**Kris Brandow:** So whitewashing is-- I guess it's easiest to describe it with an example, something that prominently happens in Hollywood, where you're casting a movie where the cast should be predominantly Asian, yet somehow, all of the main characters are white people. That's whitewashing. So it's just like, "Oh, well, we've put white actors and we're just pretending as if they are Asian, or of Asian descent, or Middle Eastern, or whatever." + +The same thing usually happens where all these companies that very much do not support queer rights and any other month of the year comes around and they're just like, "Hey, we like queer people. We're going to celebrate you now. Here's a bunch of discounts." And then there's this perfect-- I don't know if it started on TikTok or Instagram or wherever, but it's just this iconic thing that comes around every year, where it's just like this person pretending to do an ad, and it just starts with, "Hi, gay" and then it goes into all this other stuff. I hope we can find it and link it, because it's just absolutely amazing... But it perfectly describes what it feels like to be a queer person in June, because it really is just like an assault of all of these companies that are like, "I know that you support-- you are against same-sex marriage, and yet you're coming out and you're putting a float in the parade. Thanks for the money now, but how about you support us all year long?" + +**Natalie Pistunovich:** So maybe another follow -up question... For companies who do think how to celebrate queer new year, what would be a better thing to do, other than just changing your logo to something within a Pride flag? + +**Kris Brandow:** There's always the monetary support you can give to organizations that help queer people, especially queer youth, especially trans youth, people in danger, so things like donating to Trevor Project are immensely helpful. + +**Natalie Pistunovich:** We'll add links at the end of the notes for those who listen. + +**Kris Brandow:** But yeah, I think it's just like, if you're doing it so that you get recognition for it, I think you're just doing it for the wrong reasons, and I don't know if there's anything at the end of the day that will help fix that or repair that. I don't know. + +I remember I was working at one company and there was this thing that happened that was basically supporting anti-queer organizations, and a lot of employees were like, "We don't feel great about this." And one employee was like, "Hey, how about we take all of the money we get from these anti-queer organizations and we actually just donated it to the Trevor Project, or donate it to something?" And they were told no. And it's like, "Well, the company can't do that", and when asked why, it was like, "Well, we don't know how to market this." And we're like, "But that's not the reason that you would do this thing. You would do this thing because it's for the employees, for the people that are working here to know that the company does actually care and value them." So I think that's also a sort of thing - if you're going after it just to, at the end of the day, be in this like, "Hey, we're getting recognition and we're making money off of it", I think that's the wrong direction. + +**Natalie Pistunovich:** Yeah. + +**Kris Brandow:** I mean, we're going to have to retitle this episode like "Gay Time" or something, because this-- + +\[laughter\] + +**Natalie Pistunovich:** Special edition for Queer New Year. + +**Kris Brandow:** Yes, exactly. Happy Queer New Year, everybody out there. We'll put lots of rainbows around it. + +**Natalie Pistunovich:** Can you somehow-- I'm missing like one or two links, but something about generics and being gender fluid... + +**Kris Brandow:** Oh... + +**Natalie Pistunovich:** There must be something there, right? \[laughs\] It fits all sorts of things. + +**Kris Brandow:** Generics, a new feature rolled out for our generically-gendered friends out there. It's like, your gender doesn't have to be static. It can be whatever you want it to be. + +**Natalie Pistunovich:** And it's fine if your decision is different from the decision of your language of choice. + +**Kris Brandow:** Okay, we're down a rabbit hole now. \[laughter\] We're drinking tea with the Mad Hatters... diff --git a/Going with GraphQL_transcript.txt b/Going with GraphQL_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..15769e806aa15b73c4e9be520339a621cc368880 --- /dev/null +++ b/Going with GraphQL_transcript.txt @@ -0,0 +1,443 @@ +**Jon Calhoun:** Hello, and welcome to Go Time. Today we're gonna talk about GraphQL, and we are joined by two guests. The first is Mark Sandstrom. Mark is an engineer on the platform infrastructure team at Khan Academy. Mark, how are you? + +**Mark Sandstrom:** I'm great, thanks. + +**Jon Calhoun:** Well, thanks for joining us. Our other guest is Ben Kraft. He is the creator of the genclient, and he also used to work on the platform and infrastructure team at Khan Academy. Ben, how are you? + +**Ben Kraft:** Doing well, thanks. + +**Jon Calhoun:** Awesome. Well, hopefully Ben's internet sticks with us... We've had some issues before we've started, so if he cuts out, keep that in mind. Alright, so let's go ahead and just jump in... + +**Mat Ryer:** I'm here too, Jon... + +**Jon Calhoun:** Mat, you want me to introduce you? + +**Mat Ryer:** Yeah. Yeah, yeah. + +**Jon Calhoun:** Come on, now... Alright, sorry everybody. + +**Mat Ryer:** Treat me like a guest. + +**Jon Calhoun:** We're also joined by our other host, Mat Ryer. + +**Mat Ryer:** Hello. + +**Jon Calhoun:** Mat, how are you? + +**Mat Ryer:** Thank you for remembering me. I'm doing fine, thanks. Very glad to be back, it's been a while, but... + +**Jon Calhoun:** It's been a while for me, if you can't tell, based on my terrible introduction for you. + +**Mat Ryer:** Nah, it's alright, really. I know sometimes when I'm not the host I tend to mess around and be silly, and stuff, but Jon, out of respect -- well, I'm gonna obviously carry on. I'll just carry on actually doing that, probably... + +**Jon Calhoun:** Alright. We're gonna first start off by just talking about what is GraphQL. So Ben or Mark, do either of you wanna take that? + +**Mark Sandstrom:** Yeah, I'll take this. So GraphQL is a specification for building APIs. It consists of a query language for fetching data, and also a spec for implementing GraphQL servers. It can run over any transport, but it's usually run over HTTP. You send a query to the server, usually as a post request. If you squint, a GraphQL query kind of looks like a JSON document, without quotes or colons or scalar array values; pretty much what you're left with is the keys and curly braces. And that isn't a coincidence. The query document describes the JSON data to be returned by the server. + +\[04:05\] So the structure of the data that you're requesting is specified in a schema that you write... So your schema might have a type called user, and the user would have an ID, and a name, and perhaps friends... And friends could be an array of users... And if you follow that through, those friends could have friends... And a little bit later we can talk about why that's okay in GraphQL. + +**Ben Kraft:** I guess the way I like to think about GraphQL is as a solution to some of the classic difficulties of a REST API. So say you have an API and you wanna get the current user, but different pages wanna get different pieces of information about the current user... So one page might just want their ID and name, and the other page wants their friends, and their friends of friends, and some whole complicated thing. And GraphQL lets you just define one schema that can cover all of these, and then each page just says "I want this data", and then the server can send back exactly what it needs, instead of having to write a bunch of -- either call a bunch of different API calls from each page, and then you get a bunch of roundtrips, or having to write essentially an API call for each page, and then you have to kind of duplicate a bunch of things. + +GraphQL says "We're just gonna have this whole schema", and then the client says "Here's what I want", and the server just sends it back, and it's all in one API call, and it makes things super-easy. + +**Jon Calhoun:** Yeah, I know when I first looked to GraphQL, it kind of struck me as something that somebody on the client side would have dreamed up... Whereas when you look at a REST API, it looks like somebody on the server side dreamed it up, because it literally just tends to mimic exactly what's on the server side. They're like, I'm doing no data transformation, or anything. You can deal with that." + +**Ben Kraft:** Yeah, that's right. It puts a lot more work on the server, to say "Here's what the client wanted, and now I have to do all the glue to put it together." But when you're actually writing a GraphQL server, most of that is gonna be in your GraphQL library, so you still just write the application code. + +**Jon Calhoun:** Ben, you mentioned REST APIs... I know there are some out there that have been trying to sort of solve this problem. Stripe is one that comes to mind, where -- I don't know if you're familiar with their API at all, but they have a way, when you're making REST requests, that you can include... I think they call it "include", it's like an extra parameter... So it'll actually give you information about nested resources. But then it's always been interesting to me that GraphQL is something that just had that baked in from the get-go... And it seems like you can go -- like Mark, when you were talking about how you could have a user with friends, and those friends objects are probably going to be user objects of some sort... And you could literally nest that down several layers deep, and it just kind of works in GraphQL, which is kind of crazy... + +**Mark Sandstrom:** Yeah, that's right. It's considered an anti-pattern in GraphQL to have fields that are just IDs. So when returning an array of friend IDs, instead you can select into that object, and that prevents you from having to stitch together the data on the client. + +**Ben Kraft:** But yeah, I really think of GraphQL as sort of - you choose which fields you want in a REST API, and you really supercharge that and you make it the whole thing. You can nest it, you can -- that's just the whole API is structured, and that's GraphQL. + +**Mark Sandstrom:** You can also tailor fields a little bit more to the client. You have to normalize data, for example if you want, because it's okay; the client doesn't have to select a field if they don't want, so you can really include additional things that are just especially handy for the client to select directly. + +**Ben Kraft:** Yeah. I think at Khan Academy our user had like a hundred fields, because it's anything anyone needs anywhere; you just put it somewhere useful, and then most clients are only gonna select like five fields. And that's fine, they don't pay for the others. + +**Jon Calhoun:** \[07:51\] I think that's something that me personally, I wouldn't have thought about at first, because I have a lot more experience just making JSON RESTful APIs that send that data back... And if your objects aren't too big, or there's not too much there, it doesn't make much of a difference. But like you said, if you have a lot of different use cases where the information for a user - you know, you might need a lot, it is nice that you can throw it all there and not worry about the fact that that could be really slow if you had to send everything. + +So GraphQL - Mark, you mentioned that it kind of looks like JSON when you're writing these queries, and then the data you get back actually is JSON. Is that correct? + +**Mark Sandstrom:** Yeah, that's right. + +**Jon Calhoun:** Okay. So I'm assuming that means that there's both a server and a client part... So can you elaborate I guess a little bit more on that part? + +**Mark Sandstrom:** Yeah. So a GraphQL client is pretty simple. You package an operation in a JSON payload, and if you can make an HTTP request, you get a JSON response back, and clients - that's the basic thing that a client needs to do. And client libraries can provide additional facilities to make certain use cases easier. A server is a lot more complex; it takes the query that you send it, it parses it, and then resolves the data... So in the user example, it's resolving ID and name... And each field can be associated with a function on the server, and that function is only called when you ask for that field. So that's what solves the nested friends issue. + +You can't actually write a GraphQL query that's infinitely deep, so therefore you wouldn't end up in an infinite loop, and your server trying to resolve friends of friends of friends. Fields can also take arguments, and so really, each field is like a function. Some are really simple, and some are more complex. Not every field needs to correspond to a function. If you have a user model where a bunch of the fields are available, you can return those all together. + +In Go, one of the nice things - each separate resolver function is resolved (in gqlgen) in a separate goroutine. So they're resolved in parallel. + +**Jon Calhoun:** So you mentioned gqlgen, and Ben had created genclient, just so people know, I guess, when we get into this... Gqlgen is a library that helps generate the server-side stuff, and genclient is something for the client-side, correct? + +**Mark Sandstrom:** Yeah, that's right. And we've talked about JSON, we've mentioned JSON a few times... So one way to handle JSON in Go, or the way that JSON is handled in a JSON package is via struct tags. So there are some libraries that use struct tags to help you generate queries and interact with the GraphQL server, and then code generation is another popular strategy. + +**Jon Calhoun:** Mat, you had said before we started the show that you've written -- I believe you said you wrote, was it a client or a server with GraphQL? + +**Mat Ryer:** A client, yeah. + +**Jon Calhoun:** So when you were writing that, do you remember what library you were using? + +**Mat Ryer:** I wrote the thing. I wrote the client library. + +**Jon Calhoun:** Okay. + +**Mat Ryer:** I wrote the Machine Box GraphQL client library as well, yeah. + +**Jon Calhoun:** Okay. + +**Mat Ryer:** Thanks for bringing it up. + +**Jon Calhoun:** No problem. So when you were writing it, what language were you writing it in, Mat? + +**Mat Ryer:** I was writing in Go, and I was consuming a third-party GraphQL service. And at the time, there wasn't a way to do it, but you can just do it -- it's normal HTTP, as long as you get the request body that makes sense for the server... And I think you pass the query in as a parameter, but it is like a JSON-esque sort of query language. You just pass that as a string in a post request. In the client library that we wrote, it would just use the normal JSON unmarshaling kind of thing for the response. When you unmarshal it, or when you ask for a request, you give it an object or you point it to an object like you do with a JSON thing, and it just passes it through to that when it does the unmarshaling of it. + +So it's like a very simple, lightweight library, and it was very easy to put together. But that's definitely not the case on the server. I sort of almost can't really imagine what that looks like on server code, to implement the server-side GraphQL. + +**Mark Sandstrom:** \[12:14\] Yeah. I looked at your client right quick, and it looks like you specify the query as a string when you're making the query. Like you said, it populates the data... So in that case, you have the query string, and the struct, and you have to keep the two in sync with one another... So there's another client library, the shurcooL GraphQL client, and what it does is you use struct tags, and it will generate the query from the struct, and then populate the data back in the struct. + +**Mat Ryer:** Right... + +**Mark Sandstrom:** But that can get very complex, because it's not as simple as JSON, where it's just the field name. You have to start passing arguments, and there are things like fragments and interfaces. + +**Mat Ryer:** Well, that is sure cool... + +**Ben Kraft:** I guess the way that I think about it is on the client all the complexity is in the query and in the types that you're deserializing it into... And so I think Mat's client just kind of lets you deal with that, and it's really straightforward. I think shurcooL's client kind of tries to hide the query; you just write the types, which - like any kind of hiding things, is often convenient, and then sometimes a little confusing... And then we can later talk about my client, genclient - you write the query and it generates the types. + +But then on the other side, on the server, the types are actually a lot simpler, because you just have your user type, and it's got a bunch of fields, and a bunch of methods... And then the hard part is the server library itself, that glues all of that stuff together and actually has all this logic, to say "Did the user ask for an ID? Okay, then let's put the ID in. Did the user ask for a name? Let's put the name in. Did the user ask for friends? Okay, which fields of the friends did they ask for?" That's where things get really complicated. So gqlgen, which is the server library, is really quite complicated. + +**Mat Ryer:** Yeah. So it's a perfect kind of problem for code gen, I think, that particular piece. And you mentioned earlier that there were functions that basically are called optionally if somebody asks for friends of the user. There's another function that's gonna get called. So that goes some way to explain, on the Go side at least if you're writing a server in Go, how you could think about that. There's gonna be some kind of getter for the user, and another optionally away to then also load the friends of that user. And I don't know, do they tend to resolve down into joins in SQL, in a relational database? Or is that just completely up to really the implementation of the server? + +**Mark Sandstrom:** It's up to whoever is writing the server and what the database is that you're using. At Khan Academy we use a document database for a lot of things. We do have SQL here and there... + +**Mat Ryer:** So that's interesting then. + +**Mark Sandstrom:** Yeah. And you need a different strategy, because with a SQL database you can select just the fields you want, and document -- at least the document database we use, you get the entire entity from the data store. + +**Break:** \[15:11\] + +**Jon Calhoun:** So it sounds like using GraphQL on the server side is going to complicate things... So presumably it's solving some sort of problems to justify all that added complexity? + +**Mark Sandstrom:** Yeah. + +**Jon Calhoun:** So can you speak a little bit to what problems it's solving and how it was helpful at Khan Academy? + +**Mark Sandstrom:** \[16:16\] Yeah. So Ben mentioned just RUST APIs versus GraphQL APIs, where in a REST API you may be making multiple requests and stitching it together on the client. I think it really does make it easier for the client; we have a couple clients, we have a website, and we also have a mobile client... + +**Ben Kraft:** To me, one of the really nice things is that -- as someone who mostly works on the server, one of the really nice things is that you don't have to know exactly which client is going to want what, where; you don't have to -- if the frontend developers decide they need three more fields in this particular query, they just add it. As long as those fields are already in the schema, because you needed them on some other page... Like, they can decide; they can just add and remove things. So as a server engineer, you just say "How do we compute any data that we do need?" You don't have to worry, really at all, about which requests need which data. The server library just figures all that stuff out for you, because the client just asks for it. + +**Mat Ryer:** Yes. But haven't you really just moved the problem there though? At some point we have to decide what data we're gonna load... And obviously, if there's a real separation between client and server, which if you have a public API, you have customers that are not part of your organization, so that's a significant separation... But if you were building an app and you just chose GraphQL, but no one else is gonna use this API, you at some point have to decide what data you're gonna load, and you may optimize that over time as well. Basically, you're just empowering really the frontend, or the client really, in this case. Isn't it akin to just letting them run some SQL? Why don't we just let them run SQL? + +**Ben Kraft:** In some ways it is. But I think the nice thing about GraphQL over SQL is it's a much simpler language. If you wanted to try to implement a whole SQL server, that includes your business logic... Like if you build a really simple app, it may look like you're kind of just making requests to the database, and maybe you throw some auth on it, or something... But in a complicated app - it all gets more complicated. You have a bunch of business logic sitting in between. And I'm not gonna say you can't put that in a SQL server, because there are certainly people out there doing it, and I think it's a cool idea... But it's a lot harder. Whereas GraphQL is really designed so that, again, it's just calling these functions on your server. So if you have a function that knows how to get users' friends, that's really all you need to implement that field. And so GraphQL is really -- the server library is very complicated, but I think writing a server as an application developer is not that much harder than REST. You're really just sort of writing the same functions, and the entry points are just a little different. The entry point is the field, instead of the entry point being the whole API call. + +**Mat Ryer:** Yeah. + +**Ben Kraft:** And for that, you get a lot of flexibility, and that flexibility -- you know, on the other side it is still tricky for the client, in some cases, to decide "What information do I actually need?" Because you may have some huge React app, and you need to know which of your components is way down the stack; what they're gonna need. That is still a problem, but at least you get the decision closer. So it's at least written in the same language, say. Then you can use a lot more of your tooling to look into that. + +**Mat Ryer:** I'm glad you mentioned that there's better ways to do it than just doing it in SQL, because of course, this is Go Time, not Stored Procedure Time. That's a separate podcast, coming soon, starring Jon Calhoun. + +**Jon Calhoun:** Oh, boy... + +**Mat Ryer:** Jon, you're gonna do an all about SQL and stored procedures podcast. Is that right, or have I just made that up? + +**Jon Calhoun:** I hope nobody's signing me up for it without letting me know... \[laughter\] + +**Mat Ryer:** Oh... You were nearly gonna do it. + +**Jon Calhoun:** \[20:00\] Some of these problems are interesting to me, because it seems like people are approaching them from different angles. One of the more recent React frameworks was Remix, and I believe they're trying to solve some of those similar problems of not knowing what data you need further down in your React application. I don't know if either of you have used it at all, but essentially, if I recall correctly, it's supposed to be something where your entire page has like a small section that's like a panel, or a widget that needs to load some data... It tries to make it so that starts loading the data on its own API request if it needs to, without it being a waterfall, where the first one goes, and the second one goes, and it's really slow... + +So I've seen people approach this problem, like you had said, Ben, where you need to kind of know what data the entire page needs, whereas sometimes they're just like, "Well, let's just not know what the entire page needs, and let each thing load its own data." And there's pros and cons to both, I'm sure, but... You know, just seeing different approaches to similar problems. + +**Ben Kraft:** I think one of the cool things with GraphQL is that in theory - and I'm not enough of a client developer to know how much this works out in practice... But in theory, you can kind of roll that all up statically, where each of your components may need something different, and you just kind of roll that all out to the top statically, and then you can just send it to the server in one Go; you don't need to do anything clever to make that efficient. + +**Mat Ryer:** Yeah, it probably comes down to organizational things even... Like, if you've got a little team that are working on something, you don't want them to have a dependency if you can help it. One of the advantages I can see to GraphQL is that it does empower the client. So that means if there is another team that's consuming your API, they are somewhat more free to run ahead at their own speed. It doesn't have to be a conversation even. But then there's trade-offs to that, and one of the things that I was interested in is now anybody can really build any kind of query. Is it possible that they're gonna ask for something that's very inefficient, and it's gonna be quite an expensive thing to figure out? Whereas obviously if it was a RESTful service, or if you were building an endpoint dedicated to a specific task, you could really sort of ultra-optimize that. Do we lose some performance in the generalization of this? + +**Mark Sandstrom:** Yeah, so when loading data for a single-page app, you can load the data at the top level and pass it down through the tree, or you could have components own their individual queries and load a subset of the data. But to get to your question, you can construct queries that are very expensive to resolve. So we're talking about client and server separation, but as a client developer, even with GraphQL, you can't be ignorant of what the server is doing; and vice-versa, as a server you can't be ignorant of what the client is doing. + +**Mat Ryer:** So with that power that you get also comes great responsibility. That's essentially your message. + +**Mark Sandstrom:** Right. You can see the queries that the client is sending to the server, and exactly the data that they want, so it does provide an opportunity to optimize for returning that data as well. + +**Mat Ryer:** Yeah, that's a good point, because of course, even if it's inefficient, you'd be able to run a report; assuming you were instrumenting this stuff, you'd be able to find out where are the inefficiencies, and then optimize it after, without changing the interface. + +**Mark Sandstrom:** Yeah. And then another consideration is protecting a GraphQL server against denial of service. Because if anyone can construct any query, then that can get very expensive. So one strategy for that is calculating the complexity for query and not allowing queries beyond a certain complexity. + +Another strategy which we use is we have an entire list of all the queries that clients are allowed to send to our server, called the safe list. And if it's not on the safe list, our server will not execute it. + +**Mat Ryer:** Ah, so that's sort of like you've locked that down after. + +**Mark Sandstrom:** That's right. + +**Mat Ryer:** You let the clients free to build and ask for what they need, and then lock it. + +**Mark Sandstrom:** \[23:54\] That's right. Yeah, we statically gather those across the mobile client, across all of the backend services, and so we have a complete list, which allows us to do other very interesting things, since we know exactly what data is being asked for everywhere. We also use GraphQL for communication between our services. + +**Mat Ryer:** Do you really? + +**Mark Sandstrom:** So not just client-client, but service-service, yeah. + +**Mat Ryer:** Is that to have the self-similarity and the familiarity and all the dogfooding benefits? Or is it because you then also get to only select the data that you need? Do you use those features of GraphQL, too? + +**Mark Sandstrom:** It's both. We use federation to make it so the client doesn't need what services, what fields come from; backend services take advantage of that as well. So the services expose one API, and we do have access controls where we'll lock down a field so that it's just allowed to be used by other services. But then we also have access controls where it's exposed on the public API. + +**Jon Calhoun:** So when you have a list of "These are the accepted queries", how does that work with a public API where presumably people would think they can kind of generate whatever query they want? Is it just in the docs, like "You can't do certain things"? + +**Mark Sandstrom:** I think in that case you'd wanna go with a complexity approach, and have some sort of perhaps budget when the user is making requests, and if they go through their complexity budget, you start rejecting their requests. + +**Jon Calhoun:** It'd be interesting to see some of the approaches... + +**Mat Ryer:** You're also rewarding the people for keeping the queries simpler, if it's a complexity budget. That's actually quite an interesting idea, when you think about that. + +**Jon Calhoun:** I'd start to wonder if -- like, you'll see some APIs that have rate limits of like this many requests per minute, and if they start to have complexity limits instead of... You know, you can do as many requests as you want, as long as your complexity doesn't reach a certain amount. That could be very weird to convey, but... + +**Mat Ryer:** Yeah, if you're trying to build an app that figures out how many degrees of separation between you and Kevin Bacon... This was a meme a few years ago, which I'm trying to bring back. You know, then that would be expensive, complex, so you maybe can't do it... But maybe it's actually not, as well. Obviously, there's recursion going on, presumably. So if you have the case of like user and friends, and let's assume that the friends return a list of users - I guess that's the example that we're using... So then you can ask for friends of friends. And on the server, presumably that's just the same method being called twice, or the same function being called twice. Is that right? + +**Jon Calhoun:** It would be called more than twice, I assume, because for every friend it would have to call that function. + +**Mat Ryer:** Oh yeah, I meant get friends, or the friends function; the one that loads the list of friends... If you then got friends of friends... + +**Mark Sandstrom:** Yeah... + +**Mat Ryer:** Oh yeah, multiple times. You're right. + +**Mark Sandstrom:** Yeah. And exactly how you're fetching the data depends on the structure of your data. So if the friend has the ID of the friends, then if you only select friend and ID, that's gonna be much less expensive than if you select the friend name, so it actually has to fetch the friend from the database to the return the name. + +**Mat Ryer:** Right, right. + +**Mark Sandstrom:** I mean, we're talking about the server bit... And we could talk a little bit about the mechanics of what gqlgen is doing, if we wanted. I think something that's interesting to mention about gqlgen is it's a schema-first library. So you write a GraphQL schema, which is defined in the spec, and you point gqlgen at that, and from that it generates interfaces which have the various methods that you're expected to implement. + +**Mat Ryer:** So there's a really interesting detail there which I think we should probably shine a light on. You say there's a schema, and that schema describes the API. You get pretty good discoverability with that, don't you? + +**Mark Sandstrom:** You do, yeah. + +**Mat Ryer:** \[27:54\] It's an interactive web client that can connect to the end point, and gives you like IntelliSense, so you can actually look at the objects... It's kind of self-documenting, so it does feel very modern in that sense. And that's also great, if someone's consuming an API, having that at your fingertips, where you have autocomplete and features like that. That is gonna, for sure, help. And consuming services... Actually, I did consume a service that was extremely complicated, and essentially was just -- I feel like they'd basically dumped their tables through GraphQL. It wasn't much design or nuance to it, frankly. So that did make it quite difficult to consume. + +**Mark Sandstrom:** Yeah. + +**Ben Kraft:** That's definitely something people do, and something that in writing a client has -- we've seen some interesting use cases come out of folks who were using our client on GitHub... Like, "How the heck did you get this schema like that?" And it's "Oh, they're using one of those things that just kind of turns your database into an API", which I think obviously have huge benefits as well. They definitely make it a little less clean... + +**Mat Ryer:** Yeah, exactly. I think you still have to design your API. So Mark, you were saying that you take the schema and then you generate interfaces and things, which then makes... + +**Mark Sandstrom:** Yeah, that's what glqgen does. So you're expected to implement the interface type and the struct that you create that implements this interface. That's where you can inject all the dependencies into the resolvers. The resolver functions themselves are methods on the struct. And I also mentioned that when you're at the same level within the query, for the most part, the fields can be resolved in parallel, so they're all being resolved in different goroutines. + +**Jon Calhoun:** I think one thing that's worth pointing out here is that because gqlgen is schema-first and because GraphQL has a schema, that means that this is a typed query language, where you actually know what all the fields are... Whereas when some people are used to like a JSON API, there's not really a type structure that's strictly adhered to, and sometimes that can be insanely frustrating. + +**Mark Sandstrom:** Yeah. + +**Jon Calhoun:** I remember working with one API where there was like one field on it, and if it had one item to return back, it gave you back a string with the ID, but if there was multiple, it gave you an array back... There was something about the request where you could actually request multiple; so if you requested using an array with just one in it, it still just gave you a string back... And it was one of those things, it was like "Who designed this? What were you thinking?" + +**Mat Ryer:** Well, they're trying to be helpful, aren't they? They're trying to optimize. It's really clever. If there's just one, it's a string; if there's many, it's a slice of string... That feels intelligent, in some way. I mean, you don't realize that "Actually, yeah, I'd rather just have it as an array." + +I built an API once that was just all arrays. There weren't any objects. And if there was only one object, it would be an array with one object in... Because then you just could write it once. You could write an implementation that just looped over this data, and it worked for one and multiple. So you'd kind of get that for free. But yeah, I think you do need to still pay attention to your API design. + +**Mark Sandstrom:** We doubled down on making our GraphQL schema the authoritative place for our documentation... So attached to each field, you can put a doc string; attached to each type, you can put a doc string... So really, in our systems, if you wanna understand the broad data model and how things connect together, you go to our composed schema, or federated schema, and you can read through and get a pretty good sense. We have a linter that enforces that some doc strings are written, and also that they have a consistent format etc. + +**Ben Kraft:** \[31:49\] One of the nice things about GraphQL is that it's a fairly opinionated type system. I mean, you can certainly have a type that's like string or a list of string, but it's gonna encourage you to do the list of string. If you design your GraphQL schema well, that's really where you make that decision and where you say "Okay, this might be a list, so let's make it a list." + +**Jon Calhoun:** Alright, so when it comes to actually working with GraphQL - and this is a Go Time podcast... So were there any challenges that you guys experienced that felt specific to Go, or maybe specific to a typed language, versus using JavaScript or something else when using it? + +**Ben Kraft:** Yeah, so I can talk a lot about this, so please, cut me off. But I think the first challenge we had, which will contextualize the client that we wrote, is that -- you know, the great thing about GraphQL is it's a typed language, and the great thing about Go is that it's a typed language. But you've gotta make sure those types match. If you make a request and the server thinks it's one thing, or the GraphQL schema says it's one thing but the client thinks it's another thing, that's no good. And initially, this was a challenge for us, because we would have people write queries, and if you don't manage to unit-test it quite right, you put into production and it's just like "Sorry, that's the wrong type", and it's like, "We're using a typed language. We should be able to do better." + +So the client that we wrote, genclient - it uses your queries and it uses the schema, and it generates the correct types for you. So if you make a query that's incorrect, you can do that, but at compile time it's gonna say, "Sorry, we checked against the schema. There's no field called 'namee'. Maybe you meant 'name'." And it's gonna generate the right Go types for you. So if you thought name, for some reason, was an integer, it's gonna just give you a string, and it's gonna give you a Go type with a string in it. + +So that's something that I think Go is really powerful, especially using Go and using code gen, and using a really well typed language. You get the typing all the way; if you're using a library like gqlgen, where you've got good typing on the server, and you're using a library like genclient, where you've got good typing on the client, your types are checked all the way through, from the client to the server, if you do everything right... And I think that's really powerful. + +**Mat Ryer:** And they both share the same schema, right? The source of the code gen is the same data... Because there's another advantage to this. It's not just -- I mean, it is what you've said, really, but as a sort of iteration, development engineering process, having that confidence of being able to, first of all, have discussions in an API schema is very powerful. Frontend engineers often have a very different kind of perspective to backend engineers; that can happen. But this is a common place that you are speaking the same language really, and you can kind of collaborate much better there. But then not only that... Once you then generate all the types, you're faced with errors sometimes, and you have to fix them, and broken tests, and things. And it sounds bad, but actually that is exactly what you want; what's happening is the compiler sometimes is just pointing exactly to all the places that need your attention now... And that is an asset to have. + +**Ben Kraft:** Yeah. And it really helps when you're evolving your API if you wanna add a field that's pretty safe. But if you wanna remove a field or change the type of a field, you have to check that nobody is using it in a way that's gonna break. And that's the kind of thing that if you have a typed server, you have a typed client, you just change your schema and you see where you get compiler errors all the way across the stack... And ideally, that tells you everything you're gonna need to fix. + +**Mat Ryer:** Yeah, especially if you're using TypeScript in the frontend as well; you literally have types there, too. I mean... Yeah, really. + +**Ben Kraft:** Yeah. And a lot of the inspiration for gen client came from clients that already exist in TypeScript, that'll similarly generate your TypeScript types from your schema, and then just use them. + +**Mat Ryer:** It's good, ain't it? I'm usually not a fickle man... This episode has completely convinced me to now use GraphQL on the server. To be honest, I've never comprehended -- I thought it was built writing SQL, and writing joint queries, and things like this, in the backend. + +**Jon Calhoun:** \[36:03\] I think that gets confusing because -- + +**Ben Kraft:** You can do that, but... + +**Jon Calhoun:** Like, you see all these services out there that are like "This is a database with GraphQL built in", or it will turn your Postgres database into a GraphQL server... It is kind of confusing at times, because you're not really sure what it's doing or how that's working, or whether that's what you'd want... Which makes it a little bit challenging. + +**Mat Ryer:** Yeah, that can't be great. That sort of general approach can't be great. Maybe on small projects, or -- frankly, any small project you can get away with a lot. It's a great tip - if you want to write really robust software, don't make it popular. Just don't have that many users... If you can. I'm very good at that. It's actually where I shine. And actually, the whole idea of GraphQL where you can ask only for the fields you need - in some cases, that is just negligible, the difference that that really makes in reality. So it's very interesting to hear about the other benefits and things that would apply even in smaller projects. + +**Jon Calhoun:** Like you said, the fields - knowing that you don't have to request them all... I've definitely built some smaller APIs where sending the entire resource back just doesn't matter at the end of the day... So it's like, that's not a real benefit for me. It really just becomes a -- you know, trying to figure out what makes the most sense. + +**Ben Kraft:** Even with that, I think the big benefit is really -- it's not that you can skip requesting name; it's an extra 10 bytes, or whatever... It's that you can provide those linkages to friends, and friends of friends, and friends of friends of friends, and you can decide exactly how far down you want to go. I think the relationships -- and that's why it's called GraphQL; your users may be a graph. You can think of the schema, actually, as a graph. And you can decide exactly how far down that graph you wanna traverse. And that's really where it's like -- you definitely don't want to include friends of friends of friends in every API request. That would really explode. But you can get away with including it in your schema, because the client decides "Okay, I don't actually -- I really just needed the friends." + +**Break:** \[38:06\] + +**Jon Calhoun:** So I guess I'm kind of curious on both of your opinions, since you've built some stuff with this... Typically, when I'm helping somebody learn to build an API or do something like that, my general advice is to start with just a simple REST JSON API that returns some data... Mostly because if you only have one client, and you're writing that one client, and you know exactly what data it needs, you can kind of tailor it around that. Now, granted, at some point it might get so complex that you need something else, but generally speaking, I'm sort of an advocate of keeping it as simple as possible, especially when you're learning... Because you can't learn 17 things at the same time; it's just not possible. And GraphQL, from what I've seen at least - it's not that it's overly complex, but it's also not like "Pick it up in 30 seconds and move on with it." It's gonna take a little bit of learning. + +So I guess, what is your advice on when to get started with it, and then can you share a little bit about how you guys got started, resources, that sort of thing? + +**Ben Kraft:** I would say get started when you start to find REST frustrating. When you find it's like "Oh, now I have to go change all these six different APIs because I wanted to change one page" or "Now I have to figure out who's using that API, because I wanted to make a change to it, and it's used by 16 different pages, and who knows what they're doing with it." When you start to see those problems, that's when it's like, "Okay, GraphQL is gonna help you." If you're not seeing those problems yet, if you're in a small enough project that you can just build your REST API, and build your client, and keep it all in your head at once, and update everything you need to, then that's fine. I think GraphQL is gonna be unnecessary complexity for you. But it's really when you see those problems with REST. And that's really what we did at Khan Academy quite a few years ago now. We did have this huge REST API, and we did have things with 50 different fields, and it was a huge pain to go look through "Okay, I wanna remove this field because it's expensive, and only a few of the clients need it, but which ones is it?" And that's really when we started to pick up GraphQL and it really helped us a lot. + +**Mark Sandstrom:** Yeah. And I was gonna say, we also had some tooling for typing REST responses... And so you know, if you find yourself building out that kind of tooling, perhaps using something that's going to assist you that has that already built in is a good option. + +**Jon Calhoun:** Yeah, that makes a lot of sense. Earlier I mentioned the Stripe API, and if I understand correctly, they have a ton of tooling built around what they do... But at the same time, I think some of these companies had to invest stuff back when GraphQL didn't exist... So it's like, "How do we make this work knowing that's not an option?" + +**Mat Ryer:** Would you use GraphQL if you didn't have relational data in any way? Would you still use it now? + +**Mark Sandstrom:** Yeah. + +**Mat Ryer:** Cool. Simple as that. + +**Jon Calhoun:** That's all the answer you get, Mat. Just yes. + +**Mat Ryer:** Yeah. Fair play. It's just a boolean answer; I like it. Maybe you could answer in a string now, since we're talking about types... + +**Ben Kraft:** I don't know, I think most data is somehow relational. There are relationships between your data. If there's not, what is the data even? And there are certainly exceptions, but I think if you genuinely have -- just your entire data is one big table, maybe you don't need GraphQL. But I think almost everybody has some kind of relations, and you'll find value in GraphQL with this. + +**Jon Calhoun:** I guess one question I have related to that is if you're releasing a public API - GitHub is an example of one that's a GraphQL API - do you feel that releasing it as a GraphQL API makes it harder to get adoption, in the sense that people are kind of... There's a certain subset of developers who know how a REST API works and they're kind of comfortable with that, but then you throw something new in the mix and they're just like, "I'm not gonna use that API, because it is that new thing"? + +**Mat Ryer:** Yeah, I suppose there is also a bit more cognitive effort to actually deciding what data you need, and discovering that, whereas it's much easier to just say "Get this thing", but in reality, what difference does that really make? + +**Ben Kraft:** Yeah, I mean -- I think there's a risk of that for anything. I think GraphQL is easiest to use when you're working with the people who are writing the client. There are certainly public GraphQL APIs, but I think when you are kind of both sides, it's really the easiest to adopt. + +**Mark Sandstrom:** \[44:04\] Yeah. And I think writing a GraphQL query is pretty intuitive. Mat, you mentioned tooling where you can write a query, select a field, and actually that tooling will pop up, this is the type, and here are all the other fields that it has, and the documentation for those fields... So it does help with discoverability quite a bit as well, and I think that's a great place to start learning GraphQL, is just try writing some queries. It's a pretty easy entry into it. And graphql.org - really good overview if you wanna learn just more high-level how GraphQL works as a spec. Graphql.org/learn. You can run through those pages and get a pretty good, high-level overview. + +**Jon Calhoun:** Alright, we're getting sort of near the end of the episode, so is there anything else you guys would like to talk about before we move on to Unpopular Opinions? + +**Mat Ryer:** Related to GraphQL, please. + +**Mark Sandstrom:** Yeah, I know, I know... I think it was pretty good. We had it on the list to mention federation. Something I can mention right quick around that is - you had asked what advantage GraphQL was giving us at Khan Academy... We moved from a monolith written in Python to Go services, and federation was key in being able to do that safely and effectively. + +**Mat Ryer:** What do you mean by federation in this case? + +**Mark Sandstrom:** Yeah, so this is where different services own different fields, and you send your query to a gateway instead of a single backend service. It will figure out which services to contact, to connect to to get that data and stitch together and return it. So we treated Python as just any other service. And then when we started building out the Go services, we built some tooling that would send requests to both Python and Go, it would compare them, it would send the old request to the client, it would tell us if there are any differences... + +**Mat Ryer:** That's cool. + +**Mark Sandstrom:** Yeah. It worked really well. + +**Jon Calhoun:** The federation is one of those things that I've looked at because it looks cool, but I've never had a project I've been working on that's like at the scale where this makes sense... So it hasn't been something I've actually dove into. So it's interesting to see that that worked for you guys, for your particular use case of migrating. Alright, let's move on to the Unpopular Opinions. + +**Jingle:** \[46:26\] to \[46:42\] + +**Jon Calhoun:** Alright, Mark and Ben, do you have any unpopular opinions you would like to share with us? + +**Mark Sandstrom:** Sure, yeah. My unpopular opinion - so I think manually grinding through work is an underrated engineering strategy. Computers are great at automating tasks that you know how to do, and you have to know how to do something really well manually before you can effectively automate it. You see this in product development, where startups will just have staff members doing tasks instead of their API automatically doing things as they're building things out. But it also applies, I think, to other areas of engineering, where it could be that perhaps a task isn't even worth automating... I really like the approach of assisted automation, where you let the computer do what it's really good at doing, perhaps finding places in code to update, but then you just go ahead and use your ability as a person to actually make the updates. + +**Mat Ryer:** Yeah, I think that's a really interesting one, because it didn't come intuitively to me. That's something I had to learn over years of getting it wrong, basically. I definitely have an instinct to jump too soon into automating things... Because the process of solving that problem is quite nice and rewarding anyway, especially if -- because we're programmers, we're problem-solving, and things... + +\[48:09\] But you're so right. And the point you made about you should be able to manually do it first enough to know it inside out before you automate applies to lots of things I think too, and has so many benefits. I think the knowledge you get from that... + +**Jon Calhoun:** I feel like you're gonna have a hard time for that one being unpopular. I feel like you're gonna have a lot of people agreeing with you. + +**Mark Sandstrom:** Yeah? Okay... \[laughs\] + +**Jon Calhoun:** Especially developers. Because I feel like developers make that mistake a lot, of wanting to automate things. + +**Mark Sandstrom:** Yeah. It's really the manual grind... Sometimes I'll need to make updates across all of our services, and there's literally 200 things to update, and figuring out how to do that in an efficient way... Automation is not always the answer. Automating some aspect of it usually is part of the solution. But not working on a tool to do that just specific thing that you'll only be doing once. + +**Jon Calhoun:** Yeah. So I make video courses that teach programming stuff, and at first it was really tempting to build my own admin page that I'd upload a video and it would do all these things for me, like create a thumbnail, change all the stuff and do it all... And I kind of felt the urge to do that because that was like I'd have to interact with a couple of APIs, and do some other things... And what I've found was that in reality, going to the website that I host the videos on and just bulk-uploading 50 videos by dragging and dropping with my mouse, and then running FFMPEG on my own computer, just generate thumbnails and just basically naming them 01, 02, 03, and just uploading those ended up taking a little bit of time, but it was way less time than it would have cost me to actually develop all this stuff that I really didn't need, because it wasn't like I had a platform that people were all posting stuff on... + +Now, it's completely different if you host a platform, but when it's just you doing something, it's like, "Is this really worth it?" And it's really hard to say no, because that sounds so much fun. You're like, "Oh man, I'm gonna get to use this API. It's gonna be fun." + +**Mat Ryer:** It's such a good example, Jon, because that is -- even now, I'm thinking "Definitely automate that one." Even just that description of it, in the story where you're saying you learned this lesson; even then, I'm still like, "Yeah, that would be great. That's primed for automation." But Mark, would you resist that urge? + +**Mark Sandstrom:** Yes. + +**Mat Ryer:** Yeah. Boolean answer. Yes. \[laughter\] Cool. + +**Jon Calhoun:** Alright. Ben, while you're here, do you have an unpopular opinion? + +**Ben Kraft:** This may also not be that unpopular, but I think that Go really needs union types. Just like proper, disjoint union... I think generics are just making this really obvious; I think it's especially obvious because they kind of half-way added union types, except they only exist in type parameters, and it's kind of more of an untagged union. I think enums are a case of this. I think people using pointers as nullable, when you don't need a pointer, you just need something that's nullable... I think there are so many cases where it's just like "We really could just use union types." I know they wanna take the time to add it right, which I support... + +**Mat Ryer:** But you want them. + +**Ben Kraft:** ...but I think it's really getting to be time. + +**Mat Ryer:** For anyone that doesn't know, what is a union type, Ben? + +**Ben Kraft:** So a union type is when you have something like - it could be an int, or it could be a string. There are a lot of distinctions, there are slightly different ways you could think about unions, but that's the basic idea - it's something that could be an int, or it could be a string. It doesn't necessarily have to satisfy any interface. It's just, it's one of that list. + +**Jon Calhoun:** I was gonna say, if your unpopular opinion was that your ISP you have is very good... \[laughter\] + +**Ben Kraft:** Yeah, well... No, my unpopular opinion is I might need to switch to the terrible ISP whose service everyone hates. + +**Jon Calhoun:** I feel like there's lots of potential ISPs for that. + +**Mat Ryer:** I love mine. I'm one of those. I have a gigabit fiber symmetrical. It's life-changing. + +**Mark Sandstrom:** \[52:02\] Magical. + +**Ben Kraft:** Yeah. Sadly, we could not get that at this apartment. + +**Mat Ryer:** It's worth running your own fiber around the streets if you can, Ben. + +**Ben Kraft:** I wish. + +**Mat Ryer:** Yeah... No one would move it for a while, I think. If you knew someone nearby that had fiber... Just while you do the podcast. Just get a really long -- sorry, I have to stop with this now... + +**Jon Calhoun:** Can I tell you a story about this, Mat? Very related. + +**Mat Ryer:** Yeah. + +**Jon Calhoun:** So when my wife and I were building our house, I live with my parents, and their internet was terrible... And I later found out it was because their one DSL line runs underground, and apparently there's water getting in, so when it rained, it would just be a really bad connection. So apparently, it's something -- I don't know the exacts, but it was something with the union with it and everything, it's hard to get them to prioritize this job, and they can't send anybody else out to do it, so it was kind of limited on options. And somehow, at one point they got somebody to come out to run a new line, but it went into a neighbor farmer's field. So it's like just sort of dangling from a pole there, and then it ran through his field. But their internet was twice as fast, so it was great for a while. + +Then one day the farmer came through to actually plow his field, or whatever he was doing, cutting stuff, and he didn't know that line was there... So all of a sudden, the internet just went out. And then when they came to check it, basically what had happened is he went over the line and just sliced the crap out of it... And they had to go back to the old line at that point... And they never sent anybody else out to fix anything or do anything with my parents' internet. + +**Mat Ryer:** Oh, that's annoying, man... + +**Jon Calhoun:** It just shocked me, because I'm like "Who thought this was a good idea, to just run a line through some farmer's field and not tell anybody?" + +**Mat Ryer:** Yeah. Amazing. I mean, yeah, I love that story. + +**Ben Kraft:** That's gonna be me and running my fiber, you know? It's gonna work great, until somebody runs it over... + +**Mat Ryer:** Just avoid the farmers fields. Just be in the streets, I think. That'd be good. Yeah, that'd work. + +**Jon Calhoun:** I have decent internet at my place... But if you get outside of this small town that I'm in, where there's a bunch of farms and fields and stuff like that... When I was growing up, it was awful options everywhere. Like, satellite internet was sort of an option, but you always had these bandwidth caps that were so low that it was really hard to use, and everything else... And it's slowly gotten better... Whereas now they have something where they have a big tower, and if you're within direct line of sight, or if you can get a receiver within line of sight of it, it's not terrible, but otherwise you're just kind of out of luck. + +**Mat Ryer:** It sounds like you live in a dystopia. Like a Netflix special where -- well, you described this unusual village, with really modern internet, surrounded by... You know what I mean? I feel like we could sell this to Netflix. + +**Jon Calhoun:** I think the only reason the internet is modern in this small town is because there's so few people that they have -- Comcast has cable set up, and there's just not enough people for it to actually be slow. + +**Mat Ryer:** No one's using it, so it's so fast. + +**Jon Calhoun:** Pretty much. + +**Mat Ryer:** It's just ADSL. + +**Jon Calhoun:** I have an uncle who worked at Comcast... Not directly; he worked on the lines, and stuff... But he used to joke that whenever I move back to this town, that the internet got half as fast, because I'm one of the heaviest users... \[laughter\] It's what he would claim. He was just joking, but... It's interesting in different places. + +**Mat Ryer:** I knew your uncle was joking. I think I'd like him. + +**Jon Calhoun:** Maybe. + +**Mat Ryer:** Can I have his number? That'd be weird, wouldn't it? + +**Jon Calhoun:** I don't know. I mean, it'd be more weird for him, I think, if you just called him. + +**Mat Ryer:** Yeah. + +**Jon Calhoun:** He's like, "What is this long-distance call?" + +**Mat Ryer:** I'd be like "Are you Jon Calhoun's uncle?" "Yeah." "It's me, Mat!" What would he do? Should we do it now? + +**Jon Calhoun:** I don't think that's a good idea. \[laughter\] + +**Mat Ryer:** Okay, well... We need to do a -- next week we'll get ready; we'll do a new regular section, Phoning Someone's Uncle. And we'll just phone up an uncle and see what's going on. Uncles usually have the most interesting views, and some very unpopular opinions... + +**Jon Calhoun:** I was gonna say, based on where I live, I have no idea what type of unpopular opinions I might get from uncles... So that could be scary. + +**Mat Ryer:** In your dystopian Netflix village. + +**Jon Calhoun:** Pretty much. Alright, Mark, Ben - thank you for joining us. Mat, thank you for hosting with me. + +**Mat Ryer:** Pleasure. + +**Jon Calhoun:** Sorry for the poor intro... + +**Mat Ryer:** It's alright. + +**Jon Calhoun:** We'll get it right sometime. + +**Mat Ryer:** Nah, it's better like this. + +**Mark Sandstrom:** Yeah, thanks for having us. + +**Mat Ryer:** Yeah, thank you so much. + +**Ben Kraft:** Thanks for having us. diff --git a/Gophers Say! GopherCon EU Edition_transcript.txt b/Gophers Say! GopherCon EU Edition_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..fa369c949f24a2d731aba14c6945d88d07ff789d --- /dev/null +++ b/Gophers Say! GopherCon EU Edition_transcript.txt @@ -0,0 +1,487 @@ +**Mat Ryer:** Hello! Welcome to Go Time, a very special edition, live from GopherCon EU 2022 in Berlin. Say hello, everyone! \[applause\] Okay, so the purpose of the game is -- what we've done is we've asked a hundred gophers lots of questions, and the two teams here have to guess the top answers that were given. So any answer that received five or more responses will appear on the board, and you get that many points as how many people answered. So some of them will be worth a lot more, the more popular ones. + +\[04:11\] We're gonna take turns, we'll do a face-off to start with, and this is where a team captain from each team will have a guess, and the one that gets the best answer takes control of the board. We will then round robin through the team, and they'll have the opportunity to guess the answers. But be warned... You only have three lives. If you lose the three lives as a team, the other team gets the chance to steal. Stealing is essentially you just have to get one correct answer, and that's it, you can take all of the points. So stealing is key to the success in this game. + +Now, throughout the game there's no conferring, unless you're working as a team to steal. So it's important that you answer independently. We haven't got time to meet the team, but... You know, just look at them. + +**Jerod Santo:** Well, we have time now, so I will introduce them. Team Zero is GopherCon organizer and Go Time panelist Natalie Pistunovich. Ronna Steinberg, who is an engineering manager at Delivery Hero, a Google Developer expert, a Women Who Go organizer, and Go Time's Unpopular Opinion Hall of Famer. + +And Robert Burke. Robert has avidly played with Go for fun since 2009, and lately gets to do it for work, which he finds terribly exciting. His talk at GopherCon was called "Stream processing end-to-end with Go. + +Team One is comprised of V Körbes, an international speaker, Dev Rel extraordinaire, occasional writer, and Best Hair Award recipient. Tamir Bahar, an advanced technologies team lead at JFrog, who gave a talk on Generating Generators, and Konrad Reiche, a software engineer at Reddit, who works for the ranking platform team and gave a talk about the Go-based ranking system he helped create. + +And... Mat Ryer. He's also there. Back to you, Mat... + +**Mat Ryer:** Okay, so who's gonna be the team captain for Team Zero? + +**Natalie Pistunovich:** Who wants to be our captain? + +**Mat Ryer:** Natalie, okay. And for Team One? + +**V Körbes:** Hello! + +**Mat Ryer:** Okay. Okay, so Natalie, we'll start with you. I want you to have a guess. The first question is "Where is the first place you turn for Go knowledge?" + +**Natalie Pistunovich:** The internet. + +**Mat Ryer:** "The internet", okay. You say "the internet." Gophers say... \[win alert\] Yes, it's on the board, top spot. Okay, so V, I'm afraid you don't even get a chance to guess; they've taken the board, you can't beat them. + +**V Körbes:** I was gonna say it's your DMs... + +**Mat Ryer:** Well, hang on there, because you might give these ideas... \[laughter\] + +**V Körbes:** Oops! + +**Mat Ryer:** Okay, Robert, it's your turn. Have a guess. See if you can get one of these. You've got 25 points already. + +**Robert Burke:** Where is the first place you turn to for Go knowledge...? Package.go.dev? + +**Mat Ryer:** Okay, the Go Docs maybe. Gophers Say... \[win alert\] Yeah, we have. Oh, Go.dev specifically. Well done, Robert, and that's the second answer. Okay, Ronna. + +**Ronna Steinberg:** My guess is going to be Ardan Labs... I'll be specific. + +**Mat Ryer:** Ardan Labs. Gophers say... \[fail alert\] I'm afraid not now, although it is an excellent resource. Okay, one life down. Natalie, your turn. As quick as you like, Natalie... + +**Natalie Pistunovich:** Reddit. + +**Mat Ryer:** I just should say, Go.dev specifically was the answer, not the general one that I gave you. It's not a clue but what's your answer? + +**Natalie Pistunovich:** Reddit. + +**Mat Ryer:** Reddit! Gophers say... \[fail alert\] No, a car crash of an answer. + +**Natalie Pistunovich:** False. + +**Mat Ryer:** Robert, that's two lives down now... + +**Robert Burke:** Um, Gophers Slack? + +**Mat Ryer:** Gophers Slack. Gophers say... \[fail alert\] No, and that's three lives down. That means you now have a chance to steal, team one. You can confer... + +**V Körbes:** So we need to confer... We're conferring now, excuse me. + +**Mat Ryer:** \[08:13\] Confer then. Don't tell me you're conferring. Confer. \[laughter\] Put the microphone up to your conferring, please. Confer into the mic, please. + +**V Körbes:** Oh, okay, sorry. We're conferring into the mic now... \[laughter\] + +**Mat Ryer:** Don't tell them you're conferring into the microphone. Just confer into it! \[laughter\] + +**Tamir Bahar:** I told him to confer into the mic... + +**Robert Burke:** Confirmed. \[laughter\] + +**Konrad Reiche:** Okay, so we're between Twitter, Stack Overflow, and Mat's DMs... Let's go with Twitter. + +**Mat Ryer:** Okay, you say Twitter. Gophers say... \[fail alert\] No, I'm afraid not. That means Team Zero takes all of those lovely points. 46 points you get there. + +**V Körbes:** What was the answer? + +**Mat Ryer:** Well, I can just tell you four people said Stack Overflow. I did get a mention from one person, but Bill Kennedy, actually - two people said that they get their answers from Bill Kennedy himself. + +**Ronna Steinberg:** I said Ardan Labs. + +**Mat Ryer:** You did. + +**Ronna Steinberg:** I said Ardan Labs... + +**Mat Ryer:** Yeah, that's true. Oh yeah, not enough people said it to make it to the board. Yeah, let's have a look... So let's see what was in answer number four then. It was... Go Time! Oh, this is not rigged. And at number three... The docs. Okay, fair enough. Okay, great. That was round one. A round of applause as we go into round two. \[applause\] + +So now we need two different captains. Robert and Tamir, you're gonna be captains for this round. The question is we asked people to describe the Go language with just one word. What word did they say? Tamir. + +**Tamir Bahar:** I guess I'll go with "simple". + +**Mat Ryer:** Simple. Gophers say... \[win alert\] Yes, it's on the board, and it's the top answer. You've stolen straight away. Okay, so do you wanna go that way around? V. + +**V Körbes:** Cute. + +**Mat Ryer:** Cute... Okay. This is a brave one. Gophers say... \[fail alert\] No, it turns out gophers aren't quite -- + +**V Körbes:** Y'all are wrong. \[laughter\] + +**Mat Ryer:** Okay... Conrad. + +**Konrad Reiche:** Let's see... There are many words I would like to say, but I think at least "fast". + +**Mat Ryer:** Fast... Ooooh...! I don't know why I'm doing it like a ghost. Let's see... Fast. Gophers say... \[win alert\] Yes! It's at number five there. Five delicious points. Tamir, back to you. + +**Tamir Bahar:** I guess I'll go with "fun". + +**Mat Ryer:** Fun! Okay, let's see. Gophers say... \[win alert\] Yes, it's on there, and it's in position three. So we're still looking for two and four. We have simple at the top there, with 28 points, fun is in third place with nine points, fast is in fifth place, with five points. Konrad, what are the other ones, do you reckon? Oh, sorry, it's V's turn. + +**V Körbes:** I thought I was dead. + +**Mat Ryer:** Sorry. + +**V Körbes:** I thought I was dead. + +**Mat Ryer:** Yeah, keep going, you're not dead. + +**V Körbes:** Okay! Um... Elegant. + +**Mat Ryer:** Sorry? + +**V Körbes:** Elegant. + +**Mat Ryer:** Elegant. Gophers say... \[fail alert\] No... + +**V Körbes:** What's wrong with y'all? \[laughter\] + +**Mat Ryer:** Konrad. Okay, so you've lost two lives now, both V... \[laughter\] Konrad. + +**Konrad Reiche:** I can think of a negative word, but this is GopherCon, so I will go with "complete". + +**Mat Ryer:** \[12:03\] Complete. Very interesting. Gophers say... \[fail alert\] No. And that means Team Zero has a chance to steal. That's three lives down... 42 points on the board here. You'll have to confer... + +**Robert Burke:** Confer into the mic... Confer into the mic... \[laughter\] + +**Natalie Pistunovich:** Safe? + +**Robert Burke:** Safe is good... Yeah. + +**Mat Ryer:** Yeah, tell us what you're thinking. + +**Ronna Steinberg:** Efficient, I would say... Readable... + +**Robert Burke:** Efficient... We already have fast. + +**Mat Ryer:** Into the mic, please. It's a podcast, remember... + +**V Körbes:** I'd say safe. Do we wanna go with safe? Or efficient? + +**Mat Ryer:** Okay, what do you think then? I'm gonna have to press you for an answer. + +**V Körbes:** So Windows is updating on this screen... \[laughter\] + +**Konrad Reiche:** So let's go with useful? + +**Natalie Pistunovich:** Yeah. Or... + +**Ronna Steinberg:** Useful, or safe, or...? + +**Natalie Pistunovich:** Useful or safe. + +**Konrad Reiche:** Okay, as team captain for the moment, I'm gonna go with "useful". + +**Mat Ryer:** Useful. That's a great answer. Let's see if it's on the board. Gophers say... \[fail alert\] No, I'm afraid not, but that means team one take those points. And now we have - let's have a look at the score boards... Yes, a round of applause for Team One. \[applause\] Team Zero have 46 points, Team One has 42. + +**V Körbes:** We're still good... + +**Mat Ryer:** Okay, let's see what was at number four in this one. "Awesome". Seven people said that. It would have got you seven delicious points. And at number two... "Opinionated". + +**Ronna Steinberg:** Oh, wow... \[laughs\] + +**Mat Ryer:** Twelve points. A round of applause as we go into round three! \[applause\] + +**Break:** \[13:40\] + +**Mat Ryer:** Okay, this is a very interesting question, I would say, and yet, we've got lots of answers. The question was "Name an animal that reminds you of a gopher." Name an animal that reminds you of a gopher. So Ronna and Konrad are gonna be facing off now on these, to try and get this. Ronna, you go first... + +**Ronna Steinberg:** I am going to go for guinea pig. + +**Mat Ryer:** Guinea pig, okay. Gophers say... \[fail alert\] No one said that, I'm afraid. Konrad. Konrad's playing this game very seriously... \[laughter\] + +**Konrad Reiche:** It hurts me a little bit to say this, but the animal that reminds me of a gopher is... A gopher. \[laughter\] + +**Mat Ryer:** A gopher. Sounds crazy. Let's see. Gophers say... \[win alert\] Yes, and it's the top answer. \[applause\] Excellent. Tamir, your team now controls the board... + +**Tamir Bahar:** \[16:19\] So an animal that reminds me of a gopher -- that reminds you of a gopher; so not me. You. \[laughter\] I'll go with a bear... + +**Mat Ryer:** You're gonna go with a bill. A bear, okay. I'm sorry -- + +**Tamir Bahar:** Don't kill me on my accent. + +**Mat Ryer:** Sorry. No, no, no. I thought you were insulting Bill Kennedy. \[laughter\] Okay, a bear. Let's see. Gophers say... \[fail alert\] No one said bear. Okay. Never mind... V. + +**V Körbes:** Okay, I'm gonna take a -- + +**Mat Ryer:** V, would you like to have a guess? + +**V Körbes:** Okay, I'm gonna take a conservative stance this time... Let's go with hamster. + +**Mat Ryer:** Ooh, a hamster. Okay, let's have a look. Gophers say... \[win alert\] Yes, it's on there. Position three, eight points there for hamster. \[applause\] Well done. + +Okay, you still have two lives left. Konrad, it's your turn again to have a guess. See if you can clean up these board, or get one of them. + +**Konrad Reiche:** Oh, here comes an unpopular one... How about a rat? + +**Mat Ryer:** A rat, okay... This didn't go to Java programmers, this survey. Okay, let's see. Gophers say... \[fail alert\] No, I'm afraid it's not on there. You still have a life though... And Tamir, it's your turn. I can tell you that rat did get four responses. Four people said that. + +**Tamir Bahar:** So close. So very close. + +**Mat Ryer:** So very close. + +**Tamir Bahar:** I'm trying to think of animal names in English. It's tricky sometimes. + +**Mat Ryer:** Yes, bonus points for people that are doing this where English isn't their first language. \[laughter\] Tamir, what do you reckon? + +**Tamir Bahar:** Okay, I'll go with something... Let's go with a mouse. + +**Mat Ryer:** A mouse. How adorable. Gophers say... \[fail alert\] No, not adorable enough. And that's your three lives, so that means you've got a chance to steal these 23 points here. You just have to get one of them correct. + +**Natalie Pistunovich:** Let's confer! Let's confer! + +**Robert Burke:** I'm just throwing this out... This is gonna be part of my Canadian bias here... How about a beaver? Same cute, little teeth. + +**Natalie Pistunovich:** Or mole? Sometimes when I'm googling for gopher pictures I see how to get rid of moles and gophers... It's very sad. + +**Robert Burke:** Like not too many big claws. + +**Natalie Pistunovich:** Get rid of moles and gophers. + +**Robert Burke:** And very beady eyes. No big eyes. + +**Mat Ryer:** Okay, what do we think? Name an animal that reminds you of a gopher. + +**Ronna Steinberg:** Between those two, I think both make very good points. + +**Natalie Pistunovich:** Thank you. + +**Mat Ryer:** Both make good points. Very diplomatic here... Diplomatic, if slow... + +**Ronna Steinberg:** I'll take note... So what do you say to a beaver? I do actually think a beaver is a good choice. + +**Mat Ryer:** Okay. Are we gonna lock it in? We're locking in beaver. Gophers say... \[win alert\] Yes, you steal the points, and it's the second position, with 13 points. That gives you 36 points that you get to steal. Let's see what else was on this board... At number five there was a marmot. Vegemite, if you're in Australia, I think it is... And at number four was groundhog. + +**V Körbes:** I have a question... + +**Mat Ryer:** Go on, V. Sorry, everyone. Stop. V's got a question. + +**V Körbes:** Did anyone say capibara? + +**Mat Ryer:** Did anyone say what? + +**V Körbes:** Capibara. + +**Mat Ryer:** \[20:11\] No, they didn't. We had rat, prairie dog, squirrel, Noodles the cat from Reddit, and one person said Mat Ryer. \[laughter\] So... That's not great for me. Round of applause as we go into the next round. \[applause\] Yes. Now, this is Go Time, and we've got a very special round. This is the unpopular round. So now we're looking for the least popular answers. So you get more points the least popular answers there are. And there's no face-off, we're just gonna go around, starting with Natalie. Which day of the week do you get the most coding done? Which day of the week do you get the most coding done? Natalie. + +**Natalie Pistunovich:** Monday. + +**Mat Ryer:** Monday... Gophers say... \[win alert\] Yes, it's on the board, in place three. 16 people said it, so for some reason, you get 15 points. \[laughter\] Okay, Team One. Tamir. What day do you get the most coding done? What do you think our gophers said? + +**Tamir Bahar:** Let's see... Let's go with Friday. + +**Mat Ryer:** Friday is a classic... Gophers say... \[win alert\] Yes, it was on there. Well done. Nine people said it, so you get 25 points here; because this is less popular, you actually get more points. It's unpopular round. Okay, Robert... What do we think? + +**Robert Burke:** Sunday... + +**Mat Ryer:** Can you say that again in a more sinister way, please? + +**Robert Burke:** \[coughs\] Muhahaha... Sunday...! + +**Mat Ryer:** Oh, there we go, look at that... \[applause\] Okay, gophers say... \[fail alert\] No, you managed to find one that's not on there. Okay. V. + +**V Körbes:** I'm gonna go with humpday, so Wednesday... + +**Mat Ryer:** Humpday... Wednesday... Gophers say...? \[win alert\] Yes, it's not a surprise it's on there. Quite popular, so it means you just get ten points. But still good. Okay, Ronna, what do you reckon? + +**Ronna Steinberg:** It's a very difficult decision to make, but I am a decision-maker... \[laughter\] I wasn't going for laughs, but alright... I'm very upset with all of you. I'm going to go with Thursday. + +**Mat Ryer:** Thursday, oh yeah. I remember that one. Gophers say... \[win alert\] Sure thing, it's there in number four position, which gets you 20 points. Pretty good. So now we know the other remaining days, I think... I hope. We just have to now see if you can get this one. So Tamir, what do you think? + +**Tamir Bahar:** So we've had Sunday, and Monday, so I guess we'll go for Tuesday. + +**Mat Ryer:** Tuesday. Let's see. Gophers say... \[win alert\] Yes, but it was very popular. + +**Tamir Bahar:** Nooo...! + +**Mat Ryer:** So that's the most popular day, a Tuesday. It kind of makes sense, maybe. Monday - no? You think that they're wrong? Yeah. Okay. Tamir thinks you're wrong, everyone. + +**Robert Burke:** Hard question. I think you've got this. + +**Tamir Bahar:** Days of the week go differently in Israel, so you know... + +**Natalie Pistunovich:** We'll need to confer... \[laughter\] + +**Mat Ryer:** Natalie, if you can't remember which days have been... I mean, seriously, just go for Monday again. Let's see what happens. \[laughter\] + +**Natalie Pistunovich:** I'll take Saturday. + +**Mat Ryer:** Oh, Saturday. Let's see. Gophers say... \[fail alert\] No, I'm afraid not. + +**Tamir Bahar:** \[24:11\] Wait, so it's not Sunday and not Saturday? + +**Mat Ryer:** Hang on, let's check your lives here... Yeah, but bear in mind, we asked this question, and they gave as answers -- we didn't give them an enum, we don't have them in Go. \[laughter\] This is string field. + +**Tamir Bahar:** Whose turn is it? + +**Mat Ryer:** It's V Körbes' go. Come on, V, you've got this. + +**V Körbes:** Every day. + +**Mat Ryer:** Every day... Gophers say...? \[win alert\] Yes! They've literally said it, every day. \[applause\] And that was the least popular answer, which gets you 30 points. Let's check out the scores... Look at this - Team One has 112 points, but Team Zero, edging into the lead there, 117 points. A round of applause for them, please. \[applause\] + +Okay... Now the final round coming up, and this is -- you're gonna be excited. You're gonna be very excited... I hope. Or let me put it another way - please, just act surprised and excited... Okay. We'll cut this out of the edit of the podcast, but... Okay, we're coming up to our final round, and you're not gonna believe it - it's double points! \[applause\] + +**Tamir Bahar:** Wow!! + +**Mat Ryer:** Yes, so basically, whatever the points were, you times it by two. Okay, the question - "Describe GopherCon EU with just one word." So which one word would people say? Let's do the face-off to see who's gonna control the board. Natalie and V. V and Natalie, let's go, face to face. Interface to face. Go pun. V, your turn first. What do you think? What one word would people use to describe GopherCon EU? + +**V Körbes:** I'm on the fence between "feature-complete", "generic"... \[laughter\] I'll go with fun. + +**Mat Ryer:** Fun. Let's see. Gophers say... \[win alert\] Lovely. It's on there, in place two. 17 people said it, so if you put that through the algorithm, it comes out 34. It's not an algorithm, is it? You just times it by two. Okay... So Natalie, you still have a chance to take control of the board if you can get this top answer. Describe GopherCon EU with just one word. + +**Natalie Pistunovich:** I'm struggling not to toot my own horn here... But I will. Awesome. + +**Mat Ryer:** Awesome. Let's see if they said that. Gophers say... \[win alert\] Yes, they did, and it's the top answer. + +**Natalie Pistunovich:** Yay! \[applause\] + +**Mat Ryer:** That means Team Zero takes control of the board and has the opportunity now to guess these other remaining two options. We've got "awesome", we've got "fun". Robert, describe GopherCon EU with just one word. What did our gophers say? + +**Robert Burke:** In-person? \[laughter\] + +**Mat Ryer:** Is in-person one word? It's got hyphen in it? + +**Robert Burke:** Maybe... + +**Mat Ryer:** Or upper-case? Camel-case. + +**V Körbes:** Robert, are you saying in-person with a dash, or without? + +**Robert Burke:** I'm gonna say those were consolidated into a single response... + +**V Körbes:** Wow... Impressive. + +**Mat Ryer:** Okay. Is that your answer, in-person? Okay, gophers say... \[fail alert\] No, I'm afraid they did not say that. That's okay though, you've still got two lives left... Ronna, what do you think? What one word did people use? + +**Ronna Steinberg:** Well, to me, GopherCon Europe is Natalie, but I am going to go with "amazing". + +**Mat Ryer:** \[28:00\] Amazing. That's a great answer. Let's see. Gophers say... \[fail alert\] No, they didn't say amazing. But don't worry, you still have another life. Natalie, don't lose the life, essentially. + +**Natalie Pistunovich:** Educational. + +**Mat Ryer:** Ooh... Is that your answer? + +**Natalie Pistunovich:** Yes. Can we confer? No. Then this is my answer. + +**Mat Ryer:** Okay. Gophers say... \[fail alert\] No, they didn't say educational. I actually don't think anyone said that, which is an insult to all of us. Although, to be fair to them, this is happening right now, so they're on to something there... The other team now get a chance to steal... So you can confer into the microphone. + +**V Körbes:** Okay, confer into the mic. Confirm this. + +**Mat Ryer:** Nothing. Just says nothing. + +**V Körbes:** Give us a suggestion, dude. + +**Robert Burke:** English as a second language... + +**V Körbes:** I don't think they said that. + +**Robert Burke:** No, they did not. I think Berlin. + +**V Körbes:** That's a word... What do you think? + +**Tamir Bahar:** I'm thinking about social... + +**V Körbes:** Social? Social... What if we go with "party"? + +**Robert Burke:** Party? + +**Konrad Reiche:** We can... \[laughter\] + +**Robert Burke:** This has been sent around when? One, two days ago? People are not partying yet. + +**Natalie Pistunovich:** This was sent on Monday. + +**Tamir Bahar:** So tomorrow doesn't work either. + +**V Körbes:** Okay, I have a very low-hanging fruit kind of suggestion... What about the word "cool"? It's not very creative, but we get points... + +**Tamir Bahar:** So by raise of hands, who said "cool"? \[laughter\] + +**Mat Ryer:** Tamir just asked for a raise of hands... + +**V Körbes:** I saw at least one hand here, so let's go with "cool". + +**Mat Ryer:** Okay, this is to steal the board and take these 72 points... And this is essentially for the game... Is "cool" on the board? What did gophers say? \[win alert\] Yes! It is. It's in the last position. \[applause\] + +And some other answers there... There were some quite -- well, let's have a look. Let's see those in number three on this one. Number three was "exciting". It would have got you 32 points. We had some amazing answers though, but not enough of them. We also had "gopherlicious". + +**V Körbes:** Hah! + +**Mat Ryer:** We had rolicking... + +**V Körbes:** Mat, can you read the suggestions that mentioned your name? + +**Mat Ryer:** No, they've not mentioned me, these... Or they do. Um, "weird". + +**V Körbes:** Weird. That checks out. + +**Mat Ryer:** "Gosome..." Like, goawesome... It doesn't really work. "Gotastic", which was also in there... "Inclusive", which is lovely... And then someone - this is lazy; they just said "gophers". + +**V Körbes:** Technically correct... + +**Mat Ryer:** Yeah, fair enough. + +**Natalie Pistunovich:** I would use the word "conference". + +**Mat Ryer:** You would use the word "conference", would you? + +**V Körbes:** Oh, my God... + +**Natalie Pistunovich:** I did not answer, so I did not use this word. But I would... + +**Mat Ryer:** Yes... Well, let's look at the final points. Team Zero - Natalie, Robert an Ronna - you did brilliantly. You got 117 points. Very respectable, indeed. Congratulations. \[applause\] But slightly more respectable, Team One, you nailed it. You got 210 points. Well done, Tamir, V and Konrad. \[applause\] Thank you so much for joining us for this live episode of Go Time and Gophers Say. Tune in next time. Tell your friends about the podcast... If you've got any. I have. Why are you looking at me like that? + +**Break:** \[31:54\] + +**Mat Ryer:** Thank you so much, everyone... Tamir, V, Konrad, Ronna, Robert... Please go away. Thank you so much though. A round of applause, everyone. \[applause\] + +**V Körbes:** Thanks, everyone. + +**Mat Ryer:** Okay. Has anybody got one of these grey gophers? You do have one. Anybody else? Two people? Three. So these are pretty rare then, I guess... And apparently, lots of people have been asking for if these are for sale, + +**Natalie Pistunovich:** Yeah, a lot of people have been asking whether these are for sale... Next to the pink gophers, next to the T-shirts, and next to the blue plushies... And so far we've been saying no... And we have come up with an idea. We really love GoBrige. + +**Mat Ryer:** What is GoBridge? + +**Natalie Pistunovich:** Bill, can you tell us about GoBridge? If you feel like it... \[applause\] Would you like to hop on? + +**Bill Kennedy:** What? No, I could get hurt. I'm 50. GoBridge is an organization that started in 2015 to really try to help a lot of the community initiatives around education. So before the pandemic we were doing tons of workshops, really all over the planet. Natalie and I went to Kenya at one point, and we did one there. We supported people that wanted to give trainings, even locally at meetups, we support a lot of the conferences with scholarships... And there's a whole bunch of other initiatives that we're trying to do. The big one that actually might affect all of you already today is if you belong to a meetup, GoBridge partnering with GDN, we pay for the meetup fees. There's a whole, big network of Gophers now because of that, and that's stuff I can show you if you're interested. + +So we're really kind of focused on a lot of community efforts, specifically around education mostly, and we work with the Go team as much as we can to kind of help out. So that's GoBridge. + +**Natalie Pistunovich:** Thank you. What we were thinking we can do with this beautiful plushie is to make an auction. And we so happen to have a person with a British accent in the house... + +**Mat Ryer:** Hello. + +**Natalie Pistunovich:** So you can run an auction house. + +**Mat Ryer:** Yes, I can \*gibberish\* I can talk really fast like this. If you put that on half speed... That's how some people listen to the podcast, by the way, on 2x speed. It's like that. It's like me just going on and talking about everything really fast. + +**Natalie Pistunovich:** Whatever he said, yes... So we would like to bid this gopher for a donation to GoBridge. So if you would like to take this plushie from this conference, and you would like to make a donation to GoBridge, please raise your hand and tell us how much would you like to donate, and Mat will be managing this auction. Please start. + +**Mat Ryer:** Yes. So I'm gonna need to see $5 first... What currency should we use? Euros? + +**Natalie Pistunovich:** Yes. + +**Mat Ryer:** Yeah, fair enough. + +**Natalie Pistunovich:** German Marks. + +**Mat Ryer:** \[36:05\] Deutsche Mark, yeah. Okay, Euros will do. Five euros though... Do I see a bid? I do see a bid of five. Do I see 10? I see 10 down here, gentleman there with his hand up. He's got 10. Do I see 15? I see 15 in the back. Do I see 20? 20 down here, 20 euros. Do I see 25 euros anywhere? Yes, there's 25 euros, someone that's actually taken the effort to find a piece of card to hold up, which is great. + +**Natalie Pistunovich:** Your badge can be the card, yeah. + +**Mat Ryer:** Yeah. 25, okay. Do I see 30? Yes, we do, 30. What about 35? Yes, we've got it back there, the gentleman. 40. Yes, Adelina's got 40. What about 45 euros? Oh, we're getting into a bidding war, literally. Okay, what about 50 euros? Do I see 50 euros anywhere? I do see 50 euros over there. 55 euros... Maybe. Do I see 55? Yes, I see there, the gentleman with the pink... + +**Natalie Pistunovich:** This is a donation to GoBridge. + +**Mat Ryer:** A donation to GoBridge. Okay, do I see 50? Did I do 55? I'm not very good at this... 60. 60 euros, anyone? + +**Natalie Pistunovich:** Donations are tax-deductible. + +**Mat Ryer:** Okay, there we go, we've got 60 euros over there... 65 euros. Yup. We've got 65. 70? Do I see 70? Do I hear 70? I've got 70 over here... Do I have 75? 75. Looking for 75. Down here at the front, we have a new entry coming really high there. Not interested when it was cheap. Now he's like "Now I want it..." 75 euro there. Do I hear 80 euros for this little grey thing. Okay. 85, anyone? Do I hear 85? 85 euros... Adelina again, yes. 85. Okay, 90. Do I hear 90, do I see 90 euros? 90 euros here. Yes? Is that you? Okay. 90. Okay. 95? Do I see 95? Yes, Adelina is in. She wants this. + +What about -- what's after the 95 one? A hundred? Do I see - yes, in the back. New entry, a hundred. We're into three digits. Okay, let's go now 110. Do I see 110 euros? No, I'm not -- oh yes, there is 110. Okay. 120 euro donation. Do I see that anywhere? Oh, Adelina, 120. 130 anyone? Otherwise we're going for 120, unless anyone wants it for 130 euros... Okay, then going once... Going twice... Adelina gets it for 120 euros. \[applause\] Come on stage and collect your price. + +**Adelina:** Thank you. + +**Mat Ryer:** Yes, please come on upstage and bring your card. Bring your bank card. Or cash. + +**Natalie Pistunovich:** And marked bills. \[laughter\] We are in Germany, this is the land of cash. + +**Mat Ryer:** Yes. Give it up for Adelina! \[applause\] + +**Natalie Pistunovich:** We want the handshake, we want the handshake. \[laughter\] Thank you, Adelina, this is awesome. This is a great donation to GoBridge, and on behalf of everybody who got to be supported by GoBridge, thank you very much. + +**Mat Ryer:** Goodbye! diff --git a/Gophers Say! GopherCon Edition_transcript.txt b/Gophers Say! GopherCon Edition_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..b7248ef5ff88fcc3e8f6e05fa8a6232f79829ca8 --- /dev/null +++ b/Gophers Say! GopherCon Edition_transcript.txt @@ -0,0 +1,1029 @@ +**Mat Ryer:** Hello, and welcome to this Go Time GopherCon extravaganza... Yeah, extravaganza, yeah. I'm Mat Ryer, and I'm thrilled to announce that today again we're playing Gophers Say, the excellent, popular, family game show based on Family Feud, or Family Fortunes if you're in the U.K, which I am. + +Let's meet the teams. We have Team Zero, is -- well, let's meet our contestants here. We've got Julie Qiu around... Hello, Julie. + +**Julie Qiu:** Hi! + +**Mat Ryer:** Welcome to Go Time/GopherCon. We've also got Steve Francia. I can't believe it. Hi, Steve. + +**Steve Francia:** Hi, everyone. + +**Mat Ryer:** I can't believe you're here, Steve, really. Can you? + +**Steve Francia:** I mean, I'm here... + +**Mat Ryer:** Well, you won a competition, so congrats. Jon Calhoun is also here. Jon. + +**Jon Calhoun:** Hey, Mat. How are you? + +**Mat Ryer:** Hi, Jon. I'm good, mate. I'm excited. Are you gonna win? + +**Jon Calhoun:** \[03:59\] I don't know. + +**Mat Ryer:** That's a good answer. An honest answer. + +**Jon Calhoun:** If my team carries me, I'll win. But if my team doesn't carry me, I probably won't. + +**Mat Ryer:** Okay, good. So he's not gonna be pulling his own weight there. Well, Natalie Pistunovich is in town. Hello, Natalie. + +**Natalie Pistunovich:** Hi, Mat. + +**Mat Ryer:** Welcome back. Are you looking forward to the game show? + +**Natalie Pistunovich:** I'm very excited to be the one person who is aware of the rules, yes. + +**Mat Ryer:** Yeah, absolutely. + +**Natalie Pistunovich:** Because I already played it. + +**Mat Ryer:** Yes. Please remind me if I forget, as well. Katie Hockman also around. Hello, Katie. Welcome. How are you doing? + +**Katie Hockman:** I'm good. I'm very excited to win, despite Jon not believing that we can. + +**Mat Ryer:** Ah, okay. Well, you'll have to carry him. That's it. Rob Findley is also here, aren't you, Rob? + +**Rob Findley:** I am. Hi, Mat. + +**Mat Ryer:** Hey. Welcome to the Gopher Time/Go Time Mashup. Yeah, don't worry, I said it wrong on purpose. I need to tell you about this game that we're gonna play, because it's lovely. Essentially, we've asked a load of gophers some questions, and you'll have to try and guess the answers. So you're trying to match the popular answers from the survey. It's not about right or wrong, it's just about what gophers say. That's why it's called Gophers Say. + +The top answers will go on the board. Any response that had five or more answers will appear on the board, and in order to get control of the board, each team has to do a face-off, basically, where you'll just have a guess and see the highest score, and that's how you'll take control of the board. Once you've got control, you have to then guess all the other answers, and then you'll win. That's how you win, you get points. But be warned, if you get three of them wrong, you'll lose three lives, and you'll give the other team a chance to steal, and they'll be able to take your points, essentially. So don't let that happen. + +You cannot confer through the game, but while you're stealing, you can confer, and I'll remind you of that later. Do we feel like we're ready to play? I'm gonna announce the teams that we have here. + +So team zero - it's Natalie, Steve and Julie. You are our first team, team zero. + +**Natalie Pistunovich:** Woo-hoo! + +**Julie Qiu:** Woo-hoo! + +**Mat Ryer:** There you go. And you know, don't take anything by the fact that zero is in the name of your team; it's literally just zero-bound, like with the Go slice. That's why I've done that. Jon, Katie and Rob, team one. You're our second team. + +**Rob Findley:** Go team one! + +**Jon Calhoun:** Katie apparently isn't happy. \[laughter\] + +**Katie Hockman:** It's one for the first place, yeah. + +**Rob Findley:** Very happy. + +**Mat Ryer:** Yeah. Okay, good. They're definitely happy, we've been assured of that verbally. Okay, we're gonna now look at the board, we're gonna show the screen... Can we see this? I'm gonna jump into our first round here... Natalie and Jon, you're gonna go face to face; we're gonna have an interface-off, if you like puns... Natalie, what would your guess be to this question: "Describe the Go community with just one word." What did most people say when they were asked to describe the Go community with just one word? + +**Natalie Pistunovich:** Welcoming. + +**Mat Ryer:** Welcoming. Gophers say... \[win sound\] Yes, it's on the board, in number two. Nine points there. So Jon, you're gonna have a guess. If you beat Natalie, you have to get that number one spot. You'll take control of the board. + +**Jon Calhoun:** Natalie stole my answer, and I don't know if my back-up is... I guess -- is "friendly" the same as "welcoming"? I don't know how that would be categorized. + +**Mat Ryer:** Let's see if friendly is on there. Gophers say... \[win sound\] Yes, friendly... And where is it? At number three. + +**Natalie Pistunovich:** Wooooh! + +**Mat Ryer:** Natalie's team, team zero, takes control. And we're over to Steve Francia. Steve... How's it going, Steve? Are you gonna have a guess? What do you think the Go community said about this? + +**Steve Francia:** \[08:05\] I was gonna say welcoming and friendly... + +**Mat Ryer:** Okay... + +**Steve Francia:** But those have been taken. So I'm gonna say "good-looking". No, it's two words... Attractive. + +**Mat Ryer:** \[laughs\] Okay. Gophers say... \[fail sound\] Oh, no, I'm sorry. Attractive was not on the board. You lose a life. But don't worry, you've got two other lives. + +**Steve Francia:** It should have been. It should have been. + +**Mat Ryer:** Agree, yeah. Julie, would you wanna have a guess? + +**Julie Qiu:** I'm gonna say "fabulous". + +**Mat Ryer:** Fabulous, okay. Gophers say... \[fail sound\] Oh, sorry. Nobody said it was fabulous. Well, five or more people didn't say that. We'll loop back around... Natalie, what do you think? + +**Natalie Pistunovich:** Inclusive. + +**Mat Ryer:** Inclusive. Gophers say... \[fail sound\] No. Nobody said inclusive. And that's your third and final life gone, I'm afraid... So now the other team, team one, have a chance to steal. You can confer. I just need one answer from you, Jon, the team captain, after you've conferred with your team. + +**Jon Calhoun:** Alright. From my team, does anybody think the word "gopher" might be on the board? + +**Katie Hockman:** I don't know... + +**Rob Findley:** The Go community is gopher? + +**Jon Calhoun:** That, or like programmer, or like nerdy... + +**Katie Hockman:** I was gonna say diverse... + +**Rob Findley:** Yeah, I think diverse... + +**Jon Calhoun:** That's a good one. + +**Mat Ryer:** Okay, is that gonna be the answer you'd like to submit? + +**Jon Calhoun:** Are we going with diverse? + +**Mat Ryer:** Do you wanna lock it in? + +**Rob Findley:** Yeah. + +**Jon Calhoun:** Okay, we're locking diverse in. + +**Mat Ryer:** Okay, we're gonna lock diverse in. Gophers say... \[fail sound\] + +**Katie Hockman:** What?! + +**Mat Ryer:** No, I'm afraid not. You did not steal. Let's reveal the board. Number six was "opinionated". Five people said that. + +**Rob Findley:** It makes sense. + +**Mat Ryer:** And at number five, we have "fun", with six people. And in position number four, we had "helpful", seven people said that. In position three we have "friendly" with eight people, "welcoming" is in position two with nine people, and at number one it is "awesome", and that was said by 17 people. + +**Katie Hockman:** Wow... + +**Mat Ryer:** There we go. + +**Jon Calhoun:** We're getting off to a rough start. + +**Katie Hockman:** Yeah. + +**Natalie Pistunovich:** Isn't it the same as fabulous? + +**Steve Francia:** Yeah, I was gonna say, "awesome" is fabulous, right? + +**Natalie Pistunovich:** \[laughs\] + +**Mat Ryer:** Yeah. Well, there we go. Let's move on to our next round. Okay, round two... Name a place outside your house where you like to code/work. This time it's Steve versus Katie going face to face, to find out who controls the board. Steve, I'll ask you after Katie. Katie, what's your answer? + +**Katie Hockman:** Wait, so it's me first? + +**Mat Ryer:** Yeah. + +**Katie Hockman:** Okay. I had two guesses... I will say "coffee shop." + +**Mat Ryer:** "Coffee shop." Good one, I think... Gophers say... \[win sound\] Yes, of course. And it's the top answer... + +**Katie Hockman:** Yaaay! + +**Mat Ryer:** That means that Steve doesn't even get to have a guess. You've basically snapped his dream out of the clutches of his tiny -- + +**Steve Francia:** And I was gonna say coffee shop, so... + +**Mat Ryer:** There we go. + +**Steve Francia:** Good one. + +**Mat Ryer:** Yeah, good one. Okay... Right, so - now it's Rob's turn to have a guess. You've got three lives, Rob... What do you think? + +**Rob Findley:** Okay. We used to work in offices... So how about the office? + +**Mat Ryer:** Okay, let's find out. The office. Gophers say... \[win sound\] Yes, indeed. In at number three, "the office", with 21 people. By the way, 38 people said "coffee shop". Very cool. Alright, John, three lives still. Going strong. Where do you like to code/work? + +**Jon Calhoun:** \[12:02\] I'm gonna go with "park". + +**Mat Ryer:** In the park, how beautiful. Gophers say... \[win sound\] Yes, indeed. Number two. Park/garden/yard. 26 people. Okay, Katie - you've still got your three lives... + +**Katie Hockman:** I don't do this, because I watch movies instead, but I know a lot of people who do... I'm gonna say "airplane". + +**Mat Ryer:** Oh, airplane. Gophers say... \[fail sound\] + +**Katie Hockman:** Ugh... + +**Mat Ryer:** No, nobody does that outside -- they don't tend to do that, apparently. They watch movies too, I guess. You lose a life, but that's okay. You've got two more lives. Rob, what do you think? + +**Rob Findley:** Okay, how about the beach? Some people work at the beach. + +**Mat Ryer:** Oh, do they? Well, let's see. Gophers say... \[fail sound\] No. Have you met a programmer at the beach? + +**Rob Findley:** Well, yeah... + +**Mat Ryer:** Yeah, there's some cool ones, aren't there? Not in this survey, mate... + +**Rob Findley:** Not me. + +**Mat Ryer:** Not me either. Okay, Jon, back around to you, talking to cool people... What do you think? + +**Jon Calhoun:** Oh, boy... I'm gonna guess campground. + +**Mat Ryer:** Campground. Okay. Gophers say... \[fail sound\] No. I can't believe no one's going camping, setting up a tent, sorting out the electric, sort out the water, look up the local amenities... + +**Jon Calhoun:** I figured if everybody's going remote, they're gonna go camping and working. + +**Mat Ryer:** And just get some code written. They're not doing that. I can't believe it. Katie, what do you think? + +**Katie Hockman:** I always have a life left... + +**Mat Ryer:** One life left... + +**Katie Hockman:** Oh, gosh. I thought we were dead, so I wasn't even thinking... \[laughs\] Pressure's on... Um, I was gonna say -- + +**Steve Francia:** They are good, they've had three x'es. + +**Mat Ryer:** Oh, really? + +**Rob Findley:** Yeah. + +**Jon Calhoun:** Mat, are you good? + +**Mat Ryer:** Oh, okay. + +**Katie Hockman:** We did die three times, so I'm pretty sure... + +**Mat Ryer:** Yeah, that's a mistake by me then. Okay. Okay, that means -- + +**Natalie Pistunovich:** Mat started counting at zero. + +**Mat Ryer:** Yeah, thank you for getting me out of it... Okay, unfortunately that was your last life, so that means the other team get to steal. You can confer, and Natalie, I'll take your answer. What do you think? + +**Steve Francia:** So we get to talk to each other right now? + +**Mat Ryer:** yes. + +**Natalie Pistunovich:** Yes, we get to decide together. + +**Steve Francia:** What about library? + +**Julie Qiu:** I thought train. + +**Steve Francia:** I was thinking train, too. Train and library. + +**Natalie Pistunovich:** That's because you live in the U.S. \[laughs\] Maybe most of the people who took the survey live in the U.S. How about a co-working space? Is that like the office? + +**Julie Qiu:** Hm... + +**Mat Ryer:** Okay, I'm gonna have to take an answer. + +**Steve Francia:** I think four, five and six are co-working space, library and train. + +**Mat Ryer:** Okay. + +**Natalie Pistunovich:** By this order? \[laughs\] + +**Steve Francia:** Well, I don't know... Any of those sound good to me, Natalie. You choose. + +**Julie Qiu:** Yeah. + +**Mat Ryer:** Natalie...? + +**Natalie Pistunovich:** I was about Julie "What do you think?", but okay. Yeah, let's try train. + +**Mat Ryer:** Okay, "train" to steal. Gophers say... \[fail sound\] No, I'm afraid not... Well, that means the points go to team one. Let's unload this board then. Number six was "nowhere". \[laughter\] People that don't leave to do any work. That got five answers. In number five it's the pub or the bar; six people said that. I can't believe that. And at number four - let's see if this is one of yours, Steve... Number four was the library, indeed. 14 people. + +**Natalie Pistunovich:** Wow...! + +**Mat Ryer:** That's okay, don't worry though... It's going well, we've got team zero have 17 points, team one are forging ahead with 85 points. Let's go on to round three... + +**Break:** \[15:50\] + +**Mat Ryer:** Okay, round three. The question we asked our gophers is "The most useful Go keyword is what?" What is the most useful Go keyword? What did people say? We're gonna find out which team takes control of the board when Julie and Rob go head to head... Julie, this time. + +**Julie Qiu:** Oh, boy... Um, switch. + +**Mat Ryer:** Switch. Gophers say... \[fail sound\] No. + +**Julie Qiu:** Sorry. + +**Mat Ryer:** That's okay, you just lose one life, no problem. Rob, you have to have a guess. Actually, Julie, you don't lose a life. We're just still finding out who's gonna take control... So Rob, you have to get one of these to take the board. + +**Rob Findley:** Well, you can't switch without a func, so I'm gonna say func. + +**Mat Ryer:** Func indeed. Gophers say... \[win sound\] Yes, indeed, at number two. Eighteen people said func was the most useful Go keyword. And that means team one steals the control of the board, takes it. Jon Calhoun, what do you think the most useful Go keyword is? What do you think our Gophers said? + +**Jon Calhoun:** This is one of those ones where I just expect people to not name keywords as answers... + +**Mat Ryer:** \[laughs\] Why would that be, Jon? That's very unusual. + +**Jon Calhoun:** Like, if you asked me to list all the Go keywords, I couldn't list them all. I'd definitely get something wrong. + +**Mat Ryer:** But you could do some of them, right? + +**Jon Calhoun:** I'm gonna go with select... + +**Mat Ryer:** Select. Gophers say... \[win sound\] Yes, indeed. Down at number six. Seven people said select. Nice work. Okay, Katie, what do you think? Return would be very useful. Let's see if the Gophers said it. Gophers say... \[fail sound\] No...! And it honestly shocks me that that's not on there. + +**Katie Hockman:** Everybody needs that... \[laughs\] + +**Mat Ryer:** What are their programs? + +**Rob Findley:** You need to return... + +**Mat Ryer:** Well, we can't figure that out now, Rob. You've got two lives still, so... What do you think? + +**Rob Findley:** Okay. I think people will have just heard the word Go, so they'll say "go". + +**Mat Ryer:** Okay, that's how much respect you have for the audience of Gophers that have answered this... + +**Rob Findley:** It's useful. + +**Mat Ryer:** Yeah. It is useful. Let's find out if they've said it. Gophers say... \[win sound\] Yes. And look at this - top answer. 29 people said that "go" was the most useful Go keyword. Very cool, indeed. Two lives, and it's Jonny Calhoun. Go on, Jon. Do a guess. + +**Jon Calhoun:** Var? + +**Mat Ryer:** \[20:01\] Var... Good one. + +**Jon Calhoun:** I mean, how else are you gonna create those globals? + +**Mat Ryer:** Good point. + +**Jon Calhoun:** ...that everybody loves. + +**Mat Ryer:** \[laughs\] Trollin'. Gophers say... \[fail sound\] No. You lose another life. You have one more life. Katie, that life sits delicately in your hand. Are you gonna set it free, or crush it to death more? What's your guess? + +**Natalie Pistunovich:** I'm between two, so I'm gonna go with -- I think people said "if". + +**Mat Ryer:** If. Gophers say... \[fail sound\] No. These programmers are very certain. There's no uncertainty. + +**Natalie Pistunovich:** Apparently... + +**Mat Ryer:** Yeah, yeah. They know what they're doing. This is not like "Yeah, if this..." Make your mind up. + +**Jon Calhoun:** These are interesting programs. + +**Mat Ryer:** That was your last life, I'm afraid... This gives the other team, team zero, a chance to steal. Please confer amongst yourselves, and then Natalie, team captain, I'll take an answer after you've conferred with your team. + +**Natalie Pistunovich:** Alright, this time we're doing this right... + +**Mat Ryer:** Yes, please. + +**Natalie Pistunovich:** What do you think? + +**Steve Francia:** I think it's "defer". + +**Natalie Pistunovich:** I was thinking of that, too. + +**Julie Qiu:** I was thinking "panic" or "type", but I'll defer to you two... + +**Natalie Pistunovich:** \[laughs\] This was meant to be. Let's go with "defer". + +**Mat Ryer:** Oh, excellent. I don't know when to do this now... I feel like we need to wait for the end of the game show before I announce this one... I'm just kidding. + +**Natalie Pistunovich:** \[laughs\] + +**Mat Ryer:** Defer. Gophers say... \[win sound\] Yes! + +**Julie Qiu:** Yaay! + +**Mat Ryer:** You've successfully stolen! Third space there, with 11 people, "defer". And you steal the points. 65 delicious points coming your way. + +**Rob Findley:** Oh, it's so close. + +**Mat Ryer:** So team zero is now on 82, team one is on 85. Let's reveal the rest of this board then. What was the most useful Go keyword? Number seven, people said "struct", five people said that. At number six was "select", with seven people. Number five is "interface", nine people said that. At number four, "for". Clue is in the number... Ten people said so. And defer was at number three, with 11 people, func in spot two, 18 people, and go at number one there, with 29 people. + +Wow, things are heating up here... 82 and 85 it's not much in it is there? Now we've built some tension... It's time to go to round four! Pam-param-pam-pam-pow! Right... What are we doing again...? What are we doing again? Okay, the next question... The best thing about the tech industry is - what? What is the best thing about the tech industry. And it's Natalie and Jon to go head to head... Jon, what's your guess? What's the best thing about the tech industry? + +**Jon Calhoun:** The paycheck? + +**Mat Ryer:** The paycheck. Gophers say... \[win sound\] Yes, at number two. The pay/benefits. 20 people there. Interesting. Okay, Natalie, can you beat that? Can you find that number one spot? + +**Natalie Pistunovich:** Twitter? + +**Mat Ryer:** Twitter. \[laughter\] + +**Natalie Pistunovich:** It's my honest answer... + +**Mat Ryer:** Gophers say... \[fail sound\] No, I'm afraid. No one gets that. That means team one take control of the board. Katie, we're over to you... Three lovely lives. Don't ruin them. What's your guess? + +**Katie Hockman:** I think people would have said "the people". + +**Mat Ryer:** The people. That'd be nice, if they people said "the people". Let's see - did the people say "the people"? Gophers say... \[win sound\] Yes indeed, at number four. The people or the community that they're in. 15 people said that. That's very nice, isn't it? Rob, can you find a nicer one than that? Or just any? + +**Rob Findley:** \[23:53\] I don't know about that, but... How about the fact that we get to program? How about the work? + +**Mat Ryer:** Oh, the work itself, programming. Gophers say... \[win sound\] Yes, indeed. People did say that. Problem-solving, we grouped into that. 19 people, at number three, yes. + +**Rob Findley:** Generous. + +**Mat Ryer:** Very cool. Three lives still. Jonny Calhoun... + +**Jon Calhoun:** I'm gonna go with the impact. + +**Mat Ryer:** The impact. + +**Jon Calhoun:** You can impact more people. + +**Mat Ryer:** Okay, the impact - did they say that? Let's find out. Gophers say... \[win sound\] They did indeed say "impact". They said "opportunity" or "the impact to make a change", things like that. Ain't that nice? Eight people there, and it was at number six. And still, three lives... This is looking good. Katie, can you keep up the running streak of success? + +**Katie Hockman:** I don't know if I can... I feel like the thing I wanna say is probably within the category of benefits. I was gonna say "flexibility"... + +**Mat Ryer:** Hm, okay... Yeah, flexibility. Let's see if it's up there. Gophers say... \[win sound\] Well done. Yeah, work from home, or flexibility. Ten people said that. I'll tell you what - I genuinely thought this was probably gonna be the hardest question, and you've really just nailed this. Very impressive. Rob, I think we're back around to you, aren't we? + +**Rob Findley:** I'm mystified. I have no idea. I was gonna say flexibility. I don't know, how about the opportunity to grow? + +**Mat Ryer:** Okay, let's see... + +**Rob Findley:** I don't know, I'm doubtful. + +**Mat Ryer:** Let's see if more than 20 people said "opportunity to grow." Gophers say... \[fail sound\] No. And you lose a life. But it's okay, you've got two other lives, so Jon and Katie are both gonna get a guess here... Starting with you, Jonny Calhoun, down at the local saloon... What do you think? + +**Jon Calhoun:** I am equally mystified... I don't know, I'm gonna go with Go. + +**Mat Ryer:** The language, Go. + +**Jon Calhoun:** Go. + +**Mat Ryer:** The best thing. + +**Jon Calhoun:** The best thing about the tech industry is Go. + +**Mat Ryer:** Okay, let's see. Maybe Gophers did say that. Let's see, Gophers say...? \[fail sound\] No, I'm afraid not. You lose your life. Katie, the last life, to prevent the steal... + +**Katie Hockman:** Oh, no, that was gonna be my guess... \[laughs\] I don't know... + +**Mat Ryer:** You're very close with that answer, apparently... + +**Katie Hockman:** Okay... Coding? + +**Mat Ryer:** Okay, let's do it. Let's find out. Gophers say... \[fail sound\] No, I'm afraid not. So that means the other team get a chance to steal. Can you get this right, team? Please confer. + +**Natalie Pistunovich:** My idea is conferences. + +**Julie Qiu:** Me too! + +**Natalie Pistunovich:** Okay! Steve? + +**Steve Francia:** I was gonna say travel. So I think conferences is great. + +**Julie Qiu:** Yeah. + +**Natalie Pistunovich:** Cool. Okay. So that's our answer. + +**Mat Ryer:** Okay, conferences... Let's see if it's conferences. Is it conferences? Gophers say... \[fail sound\] Oh, no...! I'm afraid not. That means team one steals all those delicious points... And we'll look at the scores after. The top answer for this one was in fact the tech, the innovation. Tech, or innovation is the best thing about the tech industry. + +Okay, so let's check in with the scores. Team zero, in our slice of team here - team zero has 82 points. But team one out in the lead again, it's 157 points. You know, celebrate... + +**Rob Findley:** Yay! + +**Mat Ryer:** There we go. Okay, let's move on, shall we, to the next round. Round five. Daram-pam-pam-pow! Someone else is having a party at their own house by the sound of that... Okay, right. Here we go again. This is round five... This time the question is "The worst thing about the tech industry." This is what we're after now - what's the worst thing about the tech industry? And last time it was Jon and Natalie went head to head, so it's Steve and Katie. And Steve goes first. Steve, what's the worst thing? + +**Steve Francia:** \[28:20\] Um, bugs. + +**Mat Ryer:** Bugs. What a great answer. + +**Steve Francia:** Bugs. + +**Mat Ryer:** Bugs. Yeah. We'll submit it in your accent then, shall well. + +**Steve Francia:** Bugs. + +**Mat Ryer:** Gophers say... \[win sound\] Yes, bugs. Bad practices. Okay, yeah. There we go, we'll give you that one. Bad practices and bugs, things like that. In position four there, 13 people said that. Okay, Katie, can you beat that? There are three other answers ahead of this one. + +**Katie Hockman:** I feel like no matter what I say, I'm gonna get in trouble for it... \[laughs\] I'm gonna say this in the most diplomatic way possible... Potentially like the room for growth as far as diversity goes. + +**Mat Ryer:** Okay, so... Right, diversity being the worst thing currently. Gophers say... \[fail sound\] No, Katie. Apparently, that's all sorted. + +**Katie Hockman:** It's all done. Fixed it. + +**Mat Ryer:** Yeah, we're finished. Put that into the Done column. Okay, so that means team zero - you need it. You've taken control of the board. And Julie, it's your turn. Three lives... Have a guess. What do you think the worst thing about the tech industry is? + +**Julie Qiu:** Um, I'm gonna say the bad people. The people, but specifically the people who are bad. + +**Mat Ryer:** The baddies. Bad people. Like villains, and the like. Charlatans. + +**Julie Qiu:** Yeah. + +**Mat Ryer:** Scallywags. + +**Julie Qiu:** Exactly. + +**Mat Ryer:** Okay. The bad people. Gophers say... \[win sound\] Yes, indeed. So people are saying people or the culture. 14 there, and that was a top answer. Merry, merry festive holidays to everybody. That was a little surprise... And that surprise there means Steve Francia is gonna donate $100 to a charity. That's what that lovely piece was. Thank you so much, Steve. That's lovely. + +And Natalie - we're back around to you now. It's your turn to have a guess. What do you think the worst thing about the tech industry is? Three lives... + +**Natalie Pistunovich:** Something along the lines of agile, sprints, JIRA... Something there. + +**Mat Ryer:** Okay. So like management practices, stuff like that. + +**Natalie Pistunovich:** Yes. + +**Mat Ryer:** Okay, let's see if that's on the board. Gophers say... \[fail sound\] No, everyone loves all that. Not a problem at all. + +**Natalie Pistunovich:** Good! It's not a problem. Alright. That and diversity is all good. + +**Mat Ryer:** Steve, what do you think? Do you wanna have another guess? You've got two lives left still. + +**Steve Francia:** I think it's the hours. + +**Mat Ryer:** Oh, the hours could be the worst thing about the tech industry. It is nearly 9 PM and I'm hosting a game show... So let's see what Gophers say. \[win sound\] Yes, indeed, Steve. Number three there. The pressure and the long hours. 16 people said that. Yeah... Take a break, everyone. Not you though, Julie, because you're up to answer a question now. You've got two lives still... What do you think the worst thing about the tech industry is? We're looking for that second answer. + +**Julie Qiu:** Hm... It is very hard. Can I say like the bad impact? + +**Mat Ryer:** So the evil? Would you class it as evil? + +**Julie Qiu:** Yes, the evil. Yeah. + +**Mat Ryer:** Yeah, the negative impact of tech. Okay, let's see if that's up there. Gophers say... \[fail sound\] No, Julie. Not there. Down to one life. Natalie, you've got your thought on it, so what are you gonna do? What do you guess? What's the worst thing about the tech industry? + +**Natalie Pistunovich:** Bad code was already said, right? That was kind of the bad practices one... + +**Mat Ryer:** \[31:57\] Yeah, bad practices in existing code was at number four there. We've got number three, the pressure of long hours, and at number one is the people. + +**Natalie Pistunovich:** Legacy? + +**Mat Ryer:** Well, that would probably fall under the existing code... + +**Natalie Pistunovich:** Okay. Oh, man... I'm out of ideas. Um... + +**Mat Ryer:** Tough one, isn't it? + +**Natalie Pistunovich:** Not the people, not the long hours, not the bad practice... Nothing about management then? Did we say that? Is that management practices, that also doesn't work? + +**Mat Ryer:** Yeah. That one wasn't up there either, was it? What are we going to say? Or should we time you out? + +**Natalie Pistunovich:** Does that mean we get to keep the life? \[laughter\] + +**Mat Ryer:** No, it means the other team will get a chance to steal... + +**Natalie Pistunovich:** Right. Bad documentation? + +**Mat Ryer:** Okay, let's go for that then. Bad documentation. Gophers say... \[fail sound\] No. But what a great guess...! Okay, so now the other team have a chance to steal. Jon, I'll accept your answer after you confer with your lovely team. + +**Jon Calhoun:** So I'm assuming that if they voted the podcasts, that they would have filtered that one out, right? \[laughter\] Jerod wouldn't have let that fly. + +**Mat Ryer:** Didn't take conferences out + +**Katie Hockman:** My idea is like Twitter, Reddit, Hacker News, that kind of -- like the social media aspect. What do you guys think? + +**Jon Calhoun:** I was actually thinking Twitter, but Natalie used it for the best things about tech. But I think it's kind of like people as well - they can be good and bad. + +**Rob Findley:** They can be both. I think that's good. I was also gonna say -- + +**Jon Calhoun:** Do we go with like social media, or do we want a specific -- + +**Katie Hockman:** I would hope that they would take either, if that was actually the answer. + +**Rob Findley:** Sounds good. + +**Jon Calhoun:** Okay. So let's go with social media. + +**Mat Ryer:** Okay. Social media. Let's find out. Gophers say... \[fail sound\] No, they didn't say that, so you don't steal the points. That means that team zero keeps those lovely points... But let's find out, what was number two on this one? It was "There's too much to learn." There's too much to learn in this old tech industry of ours, and it's making us angry. That's what we've learned today. + +Okay, I don't know, I probably wouldn't have guessed that one. I thought there were some great answers there, everyone... How's it going so far? Are we having a fun quiz? + +**Natalie Pistunovich:** Yaay! + +**Rob Findley:** Wooh! + +**Steve Francia:** This is fun. + +**Jon Calhoun:** I'm realizing that I do not resonate with all the people who took the survey. \[laughter\] We've missed the top two answers several times now, I feel like. + +**Mat Ryer:** Yeah. It's surprising; you learn things. It is a surprise. Let's check in with the scores, shall we? Please? Team zero has got 160 points. Just slightly ahead of team one, who have 157. This is so tight, you won't believe it... But the next round is Double Points Round! Round six. Let's go! + +**Break:** \[34:42\] + +**Mat Ryer:** Okey-doke, this is the final round for double points. Anyone could win the amazing prize. Any one of the three of you can win this... And then you have to share it, which I'm sure will be no problem. + +The question for round six is - we asked people to describe the Go language with just one word. Describe the Go language with just a single, solitary word. What did people say? Julie and Rob, you're going head to head, starting with Rob. What do you think people said, Rob? + +**Rob Findley:** \[36:06\] I think people said "simple". + +**Mat Ryer:** Do you know? Let's see if people did say "simple". + +**Rob Findley:** I hope. + +**Mat Ryer:** Gophers say... \[win sound\] Yes, and it's the top answer. So it means you take control of the board straight away there... And we're over to Jon. You've got three lives, there are three ones to guess. The top answer is already taken. Simple. What else did people use to describe the Go language, Jon? One word. + +**Jon Calhoun:** I'm gonna go with verbose. + +**Mat Ryer:** Verbose. Okay. And just for anyone that doesn't know, verbose is when you use lots of extra words that really aren't necessary and you don't need, to try and explain something to somebody, or it could be in documentation... And essentially, you're using way too many words that's unnecessary, essentially... And the idea is -- + +**Jon Calhoun:** This is what happens when Jerod tells you we have 11 minutes... + +**Mat Ryer:** Oh, no, 11 minutes to fill? Oh, okay. I won't carry that joke on then for 11 minutes... That would have been gold. + +**Natalie Pistunovich:** About being verbose... + +**Mat Ryer:** Yup. Okay, a good bit of comedy, everyone. We're all equally guilty. Katie, what do you think? + +**Katie Hockman:** Is verbose on the board? + +**Jon Calhoun:** I was gonna say... + +**Mat Ryer:** Oh yeah, I genuinely forgot what we were doing... \[fail sound\] \[laughter\] + +**Katie Hockman:** I'm gonna assume no... Okay... + +**Mat Ryer:** No, unfortunately... It should be now, after that... So unfortunately - yeah, Jon, you lose a life. Thanks, Katie... Hopefully, we can edit that, so I don't look like an idiot. \[fail sound\] + +**Katie Hockman:** No worries... + +**Mat Ryer:** No, we can't apparently... \[laughter\] Okay, Katie... + +**Katie Hockman:** Alright, I'm gonna say "fast". + +**Mat Ryer:** Fast. The Go language with just one word - fast. Gophers say... \[win sound\] Yes, indeed. It was in at spot number three there. Nine people said it, so you get 18 points. By the way, for the "simple" answer, because 25 people said it, you've got 50 points there. So there's a lot of points on this board. And with two lives left, Rob's to guess. Robs? + +**Rob Findley:** How about "concrete"? + +**Mat Ryer:** Concrete. Oh, okay, concrete. Let's see if people said that; that'd be interesting. Gophers say... \[fail sound\] No, obviously not, Rob. But good answer. But you lose a life, I'm afraid. Jon, over to you now. + +**Jon Calhoun:** See, this is the problem... + +**Mat Ryer:** Last life, Jon... + +**Jon Calhoun:** It could be a self-fulfilling prophecy if I lose this for us... + +**Mat Ryer:** Come on, get in the minds of these surveyed Gophers. You know them by now... You got to know them over a series of five rounds of simple questions; you feel like you know them pretty well... What do you think? + +**Jon Calhoun:** Um, readable? + +**Mat Ryer:** Readable. I love that answer. Let's see. Gophers say... \[fail sound\] No, they did not. That is a surprise. And unfortunately, that was your last life, which means team zero has a chance to steal, and basically win the game if they can get either the second or the fourth answer from this board. You're allowed to confer, of course... Natalie, I'll accept your answer after you've conferred with your team. What do you think? + +**Natalie Pistunovich:** What do you say? + +**Julie Qiu:** Oh, boy... + +**Steve Francia:** I think I would say -- I'll give you a few. Efficient... + +**Julie Qiu:** That was mine! + +**Steve Francia:** Secure... + +**Julie Qiu:** That was what I have in mind. + +**Steve Francia:** I mean, it is what it says on the Go website, which I've put there. You could also say "fun". "Fun" is another one. So fun, efficient, secure. + +**Natalie Pistunovich:** Secure is a good one. + +**Julie Qiu:** I like efficient. + +**Natalie Pistunovich:** Efficient? It's interesting that some of the other questions have seven answers, and this one has only four, so it looks like everybody has the same thing in mind... Or they've just dropped out of the questionnaire. Okay, so I guess -- + +**Jon Calhoun:** I think anything below five answers gets dropped. + +**Natalie Pistunovich:** Oh, so it's actually the other way around. So many answers. + +**Mat Ryer:** Mm-hm. + +**Steve Francia:** Many people answered the same things in this one. + +**Julie Qiu:** Yeah. + +**Mat Ryer:** 25 people said "simple", 9 people said "fast". There's still two answers to get on the board, positions two and four. + +**Natalie Pistunovich:** Then how many people said "efficient"? + +**Mat Ryer:** Shall we find out? Is that your answer? + +**Natalie Pistunovich:** \[40:23\] Yes. + +**Mat Ryer:** Would you like to lock it in? + +**Natalie Pistunovich:** I would like to lock it in to number two, please. \[laughter\] + +**Mat Ryer:** Well, you don't get to choose... Let's find out. Efficient. Gophers say... \[fail sound\] You've been hanging out in Germany too long, Natalie... + +**Jon Calhoun:** They haven't been reading Steve's page. + +**Natalie Pistunovich:** I've been reading the documentation. It says there. + +**Steve Francia:** It is on the website... + +**Mat Ryer:** Yeah. Okay, we're gonna award those points, and let's find out, at number four on this board - people said, with just one word, the Go language was... Fun! Which was mentioned... + +**Rob Findley:** I knew it! + +**Mat Ryer:** Fast is at number three, nine people... At number two, pragmatic. 12 people said that. It would have earned you 24 points. + +**Katie Hockman:** At least it doesn't say secure... + +**Mat Ryer:** Yeah... \[laughter\] + +**Jon Calhoun:** Well, there are just a lot of answers with four or less then... + +**Mat Ryer:** Yeah, we'll do the data crunching and find out. It sometimes happens there's lots of answers, and it's not many people agreeing; only if you had five or more will it make it onto the board, of course... But there we go. Let's have a quick look at the scores, because these are final scores now. + +It's time for final scores now. Team zero got 160 points, which is very respectable... And I'm saying that to lessen the blow when we find out that team one thrashes you with 225 points. Please celebrate appropriately! \[applause\] + +So there we go... Any surprises? What do we think? Which was the round? Jon, when you were saying that they missed out some answers, what were you thinking of? + +**Jon Calhoun:** Oh, I just felt like half the answers people gave on that last one were ones I would have expected to at least show up a couple times... But it is hard when everybody chooses different words. + +**Mat Ryer:** Yeah, that's it... + +**Jon Calhoun:** But I did call it, that Katie and Rob were gonna have to carry me... Because I got nothing on that last round. + +**Mat Ryer:** Yeah. It was good though, ain't it? + +**Natalie Pistunovich:** That's alright. That's why you're a team. + +**Jon Calhoun:** That teamwork. Sorry, Steve. You look upset. + +**Mat Ryer:** How are you feeling, Steve, after...? + +**Steve Francia:** I'm pretty broken-hearted. I thought we were gonna win. Even at that last round, I thought we had it. I was like, "We're gonna steal it." + +**Natalie Pistunovich:** We were not efficient enough. + +**Steve Francia:** Well, we weren't fun enough, that's really what it was. + +**Natalie Pistunovich:** That's true. That is actually true. + +**Steve Francia:** Or pragmatic. + +**Jon Calhoun:** I mean, that just means you need to rebrand the go.dev website, and just put "FUN" real big. + +**Steve Francia:** I think we're gonna have to... What were the -- it was fast, simple, pragmatic and fun. I think that's what's gonna go on the website now. + +**Katie Hockman:** I think it has to. Legally obligated. + +**Mat Ryer:** That's what people think of it... But how do you feel? Do you feel like you got to know the Go community a bit more as well, in a way? + +**Steve Francia:** Yeah. + +**Mat Ryer:** Like, do you understand them? Sometimes some of the answers are unusual, aren't they? + +**Jon Calhoun:** I feel like the last time you played this -- the last time you played this, didn't you have questions that asked about editors, and people were using that? Or I think one question was like "What was the first language you learned?", so you kind of got to understand people better based on how old they were when they learned to program? This one I don't feel like there's any questions that gave that away too much. + +**Mat Ryer:** Yeah. You sort of just have to pick up the general vibes of people, I think, with this game. + +**Steve Francia:** So - interesting fact... In the first Go user survey we had a bonus question, "What was your favorite Go keyword?" And I thought doing all these surveys would help me; that's the only one it helped me with, but... "go" and "defer" were both high on the list. I remember. + +**Jon Calhoun:** I was something waiting for "else" to be on the list, even though "if" and "switch" weren't. I was just like, "If that's the case, it's just somebody trolling us." + +**Mat Ryer:** \[44:07\] Some of the answers in some of the questions -- because we've asked loads of questions, so we have loads to choose from... And some of them are -- like, you wonder if people are trolling, or joking. And sometimes the joke answers - enough people say it that it makes it onto the board as well, because... You know, it's quite funny, but... It's interesting to get into the minds of people, I think, a little bit... + +Speaking of that, we've got a couple of minutes... Steve, what have you been working on lately? + +**Steve Francia:** I've been working on a lot of stuff lately, because that's what I do, I work on a lot of different stuff... But I know -- I'll just kind of talk about some things that we're working on. A lot of work getting ready for a generics release, which is coming up in the release in February... But the betas should be -- you know, all the pre-releases we're working hard on. + +I'm also meeting with a lot of our users. A good part of how I spend my day is meeting with companies and individuals and projects who have adopted Go, and just hearing their challenges and their successes... + +**Mat Ryer:** Oh, nice. + +**Steve Francia:** It's one of my favorite parts of my job, actually. + +**Mat Ryer:** Yeah. Generics of course coming in 1.18 in February... Katie - also fuzzing is coming. First-class concern. + +**Katie Hockman:** It is. I'm so, so excited that it's happening. I'm working very hard on documentation right now. We're kind of in the bug fix, and polish, and document phase... So I think it's gonna be important for it to land with docs, or it's not gonna land successfully if people can't use it. So that's the goal that I have right now for this project. + +**Mat Ryer:** That makes sense. I love the way that it interops with the existing testing stuff; our knowledge of writing unit tests can come in handy for writing fuzz tests. I think that's a kind of really nice advantage to the fact that it's getting first-class support, rather than being a sort of external tool. + +**Katie Hockman:** Yeah, absolutely. Yeah, that was one of the main goals, and I'm glad you like that. + +**Mat Ryer:** Yeah, I do. I really do. Well, thank you so much, it's been a pleasure. If our audience wants to learn more about Rob and Julie, check out our Go Time back-catalog, and also listen to future shows, because they'll definitely be on it, hopefully. + +Thank you so much to everyone for joining this live feed, this game. A great time was had by all(most); a great time was had by most... And we'll see you next time. Bye everyone! + +**Steve Francia:** Bye. + +**Natalie Pistunovich:** Bye! + +**Break:** \[46:44\] + +**Jerod Santo:** I feel like I only really screwed up probably the tech/innovation one, which was the number one answer nobody got... I almost gave it when Jon said "Go", because that would be focused on like the tech, but it was more broad... I was struggling. + +**Katie Hockman:** What about diversity is not culture? How is diversity and culture not the same? \[laughs\] + +**Jerod Santo:** Well, yeah... I know, but you just said -- yeah, that was definitely specifically mentioned by a few people. + +**Katie Hockman:** I see. + +**Jerod Santo:** And I grouped it into like -- it was tough because there were so many... Like, "tech bros"... There were a lot of weird responses, so I kind of like put it in like the people and the culture. But when you said diversity specifically, I was kind of like "It would be too much of a mismatch with just the people." It was tough. + +**Katie Hockman:** Nah, that's fair. I'll allow it. + +**Jerod Santo:** That one was. + +**Julie Qiu:** I feel like I should have gotten the points when I said "fabulous"... + +**Jerod Santo:** Oh, those were -- that one I didn't do any grouping, because that was like the easiest one, right? Because you can actually just like "Group by" and the actual words were... So I didn't do any grouping of the words. But on the -- + +**Mat Ryer:** Maybe we should like each question and then have a regex for the answer next time. So you have to do it in your head live, but... + +**Jerod Santo:** Just do the regex live and apply it? + +**Jon Calhoun:** I would not get any answers then, because I would never compose my regex right the first time... + +**Jerod Santo:** Well, you can just have Mat write the regex, and then we can all criticize it on the air like we did last time with your team selection algorithm. + +**Mat Ryer:** Which worked, yeah. + +**Jerod Santo:** Which worked... \[laughs\] + +**Mat Ryer:** Which worked. + +**Jon Calhoun:** Yeah, that time you didn't even give us the opportunity. + +**Natalie Pistunovich:** You are all saying regex, wow. + +**Mat Ryer:** Well, what do you say? + +**Natalie Pistunovich:** Reg ex. + +**Katie Hockman:** I say reg ex... + +**Mat Ryer:** Do you? + +**Steve Francia:** Wait, what is it? + +**Natalie Pistunovich:** Reg ex, regex. + +**Katie Hockman:** I say reg ex, but I realized people say regex. I say reg ex. + +**Steve Francia:** Reg ex. Like Reg A. + +**Natalie Pistunovich:** Like regular expression. + +**Rob Findley:** Do you say gif or gif? + +**Katie Hockman:** I say gif just to mess with people. + +**Jerod Santo:** Do you say gif? + +**Katie Hockman:** I say gif, I switch. + +**Jon Calhoun:** I mean, I get reg ex in the sense that it's regular expression, but... + +**Jerod Santo:** Yeah, I understand it's illogical. + +**Jon Calhoun:** ...for some reason I don't do it. + +**Jerod Santo:** ...but regex just rolls off the tongue, whereas reg ex - you feel like you're choking. + +**Mat Ryer:** Yeah, it's hard to say. I just say regiular expressions... \[laughter\] + +**Jerod Santo:** This is funny, because... + +**Natalie Pistunovich:** It sounds like something you'd say in British. + +**Julie Qiu:** Do you say Lynux, too? + +**Jerod Santo:** No, of course not. + +**Julie Qiu:** Linux? + +**Steve Francia:** No, his name is Linus, so it's Lynux. + +**Natalie Pistunovich:** Oh... + +**Rob Findley:** Well, I thought it was based off minux + +**Jerod Santo:** No, because Linus is actually on the record of saying "My name is Linus Torvalds and I call Linux Linux." \[Linus recording 00:50:22.20\] He recorded that audio and put it out there. + +**Steve Francia:** Oh, interesting. + +**Mat Ryer:** Oh, yeah. + +**Jerod Santo:** So we have permission. + +**Steve Francia:** Yeah, that was a long time ago. + +**Jerod Santo:** It really was. + +**Steve Francia:** Who created regex? Will they give us permission? + +**Jerod Santo:** Oh, I'll give you permission to call it regex. + +**Natalie Pistunovich:** This reg ex/regex just gave my next Twitter poll, thank you. All. + +**Steve Francia:** I think the teams were super-nice and fair, and I think if we had one more round, we would have taken it. + +**Jerod Santo:** It was super-tight. This was probably the closest one we've had. + +**Rob Findley:** Yeah, it was amazing. + +**Steve Francia:** Yeah, it was close. + +**Mat Ryer:** Yeah, it was close. + +**Jon Calhoun:** Based on the way the score was going back and forth, I wouldn't have been shocked at all if you guys would have taken the next round, Steve. + +**Jerod Santo:** Well, you were one word away. "Fun" was the word, and you had it in your shortlist... + +**Natalie Pistunovich:** Yeah... + +**Jerod Santo:** So... Great game, guys. + +**Steve Francia:** That was the first thing that came to mind, and then -- stupid, I was like "I'll just look at the website, and see what the website says." I should have stuck with my guess. + +**Mat Ryer:** Does the website say "fun"? + +**Steve Francia:** No. + +**Rob Findley:** Oh, you were cheating? + +**Steve Francia:** Yeah, I was cheating. + +**Natalie Pistunovich:** Illegal! + +**Jon Calhoun:** You can't google it...! + +**Steve Francia:** Well, I didn't google it, I go.dev-ed it. + +**Jerod Santo:** \[laughs\] + +**Mat Ryer:** Is it still cheating if it doesn't work? + +**Katie Hockman:** Yes. + +**Rob Findley:** Yes. + +**Steve Francia:** Well, yes. + +**Jon Calhoun:** It's still cheating. + +**Jerod Santo:** It's still cheating. + +**Rob Findley:** But Steve did write that, so he shouldn't have had to google it. + +**Steve Francia:** I did write it. Yeah, I just wanted to double-check myself, because I actually thought it said "fun" -- at least one version of it said "fun"; another version said "simple", but those aren't on the website today. + +**Rob Findley:** I was kicking myself... I thought for sure that was concurrent. + +**Jon Calhoun:** So whoever did the study group... + +**Natalie Pistunovich:** It's on people's minds... + +**Rob Findley:** I almost said "concurrent", and I didn't. + +**Julie Qiu:** Oh...! + +**Rob Findley:** It was not there. I was surprised. + +**Katie Hockman:** I knew it was gonna be "fun". As soon as I said "fast", I was like, "I hope that can get fun" Because they said "fun" in a previous one, for a similar question. I was like, "They're gonna say it again." I thought "fun" was gonna be on the board 100%. + +**Steve Francia:** Oh, we should have said "fun"! + +**Katie Hockman:** \[52:12\] I was so sure... And then when they said it, I was like, "We're done. We're done." I knew they had it. And then they didn't, and I was like... + +**Rob Findley:** Yeah, when Steve said said "fun"... + +**Steve Francia:** So close... + +**Rob Findley:** Yeah... I was like, "Oh, man. Fun. Fun is definitely there." + +**Mat Ryer:** Well, everyone's got a time machine token that you can play at any point, and we have to rewind... + +**Katie Hockman:** Oh, dang. Okay, thanks. + +**Mat Ryer:** I forgot about that, yeah. + +**Katie Hockman:** You didn't mention that rule. + +**Steve Francia:** Oh, that'd be fun. + +**Mat Ryer:** And we'd be like \[reverse sound\] go backwards through everything, and we can get to that question again, and you can answer it differently, and we play it forward. + +**Jerod Santo:** Wow. Can we go back and erase that "verbosity" joke you did? + +**Mat Ryer:** I was trying to hold it up after that... \[laughter\] And I went too fast. + +**Jerod Santo:** I knew what you were doing immediately, and I'm like, "Oh, he's not doing this, is he?" + +**Mat Ryer:** Of course he's doing it. + +**Jon Calhoun:** You just started doing it, and I read Jerod's 11 minutes. + +**Mat Ryer:** In the edit. + +**Jerod Santo:** I guess I can go back and change it. + +**Steve Francia:** Can we make fun of whoever? Panic is not a Go keyword, is it? But someone said it. + +**Katie Hockman:** Is it not...?! + +**Natalie Pistunovich:** \[laughs\] The same complaints last time also... The previous game, same questions, same answers. + +**Jerod Santo:** Yeah, it was new last time as well. + +**Katie Hockman:** Oh... + +**Jon Calhoun:** It's like, when they do the survey, they don't actually give them a list of keywords... So on any question like that, they can just type whatever they want in. + +**Steve Francia:** No, all the words they said were keywords, but someone in our panel... + +**Jerod Santo:** Oh, someone guessed "panic"? + +**Mat Ryer:** Well, it was on -- + +**Steve Francia:** They guessed a non-keyword. + +**Mat Ryer:** It was on the last time, panic... + +**Rob Findley:** I think that's a good guess. + +**Mat Ryer:** Yeah, yeah. On keywords. + +**Jerod Santo:** Yeah, it actually was on the last time. + +**Steve Francia:** It was. + +**Mat Ryer:** People bundle that in. + +**Jerod Santo:** So they were playing 4D chess. Was that you, Natalie? She was on the last time... + +**Mat Ryer:** People bundle that in. + +**Katie Hockman:** Oh, panic is a function, not a keyword? + +**Mat Ryer:** Yeah, built-in function. + +**Steve Francia:** I don't think it's a keyword... + +**Jerod Santo:** I can't believe "return" wasn't on there. + +**Jon Calhoun:** I'm guessing that things like "new", "panic", and there's a couple others that people probably think are keywords, even though they're not... + +**Mat Ryer:** Because they're built-in functions... + +**Jon Calhoun:** I mean, some of them I wouldn't know unless I looked at the spec and was like "Okay, what actually is this?" + +**Mat Ryer:** There's loads of built-in functions as well... See, I did a talk at Gotham Go, Steve, and I remember showing the list of -- I showed the list, and you were like, "They're not them." And I was like doing my talk, trying to concentrate and be professional. And you're like "That's not them..." And I was like, "Oh..." You can see on the video when you ruined it. + +**Steve Francia:** Mat, we know you're not telling the truth though, because I've known you a long time and I've yet to see you be professional once... \[laughter\] diff --git a/Hacking with Go Part 2_transcript.txt b/Hacking with Go Part 2_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..2857b332be855fd589c3cf4e78a329a481c75f95 --- /dev/null +++ b/Hacking with Go Part 2_transcript.txt @@ -0,0 +1,263 @@ +**Natalie Pistunovich:** Hello, everyone who is joining us today, on a Wednesday of the recording. We normally record on a Tuesday, but we have a very special guest, so we need to make a very special event about that. Ian is my co-host today. Hi, Ian. + +**Ian Lopshire:** Hey. How are you doing, Natalie? + +**Natalie Pistunovich:** Good! I'm very excited to have Ivan today join us. Ivan Kwiatkowski , also known on Twitter as @JusticeRage. You are a senior security researcher at Kaspersky. + +**Ivan Kwiatkowski:** Yes. Hello, very happy to be here. Indeed. So I work in the threat intelligence field, and my daily work involves looking at malware and writing reports about it. Basically, the activity that I'm involved in is trying to figure out what the attackers are up to, what kind of tools they're using, methodologies, what types of victims they are after, and then we write stuff about it. And our customers read our reports and then it allows them to figure out whether or not this group or this group is likely to attack them or not, depending on what type of information they are after, and if so, how they may defend from those attacks by knowing more about the type of malware that they use, the type of attack vectors that they typically favor, and so on. So really, I spend my day in IDA Pro most of the time, and sometimes as well I do give out trainings, or reverse-engineering, either at universities, or for our customers as well. + +**Natalie Pistunovich:** And there's a very cool video that has two parts of you reverse-engineering a malware written about a year ago, that was written in Go, actually. + +**Ivan Kwiatkowski:** Absolutely. + +**Natalie Pistunovich:** And that was from the SolarWinds attack. + +**Ivan Kwiatkowski:** Exactly. This specific example comes from the SolarWinds incident, which I'm pretty sure that most listeners will be aware of, because it was such a high media impact case. To make a quick summary about it, what happened was a company called -- I always get those mixed up; I think the name of the company is SolarWinds, and then the product is Orion IT, but maybe the other way around, right? I do really get confused about this all the time. + +**Ian Lopshire:** I think the way the way you have it is right. + +**Ivan Kwiatkowski:** Okay, great. That wasn't really a 50/50 chance there. Anyway, this company got attacked, but it wasn't attacked for the information that it had, because it was just a software company, which in itself had little value as an intelligence target. But the thing was that it had a high number of high profile customers, and these customers were US government entities, or big companies in the field. And what the attackers did was they were able to compromise the software build chain, and they were able to insert their own code inside of the software that was then pushed to the customers. And using this, they were able to create a backdoor that would be automatically deployed at all SolarWinds customers. And then maybe two weeks or three weeks later, because this very stealthy attack had a very long sleeping time - it stayed dormant for a while, to make sure it would remain very stealthy... But after a while, then it would start connecting to the city server, and then for all the targets that were deemed interesting by the attackers, they would receive a second stage payload that would allow them to get into the network and then collect intelligence and whatnot. + +So the very first stage of the attack was just some modification of the code of the original program. This part was written in .NET. But then the second part, which is called SUNSHUTTLE, was actually written in Go language. So it was for me like the first time I was getting involved in reverse-engineering for the Go language. The learning curve was a little steep, but then again, I kind of used this as a learning experience, but also as an example in future reverse-engineering courses for other people that might be interested in learning how to reverse-engineer Go programs, but also, I think, if you are a Go enthusiast, reverse-engineering can allow you to get to know more about how the language actually works under the hood, which I think is also very interesting from a software development point of view. + +**Ian Lopshire:** So that's one famous example of Go malware. Are there other famous ones written in Go that you can think of off the top of your head? + +**Ivan Kwiatkowski:** \[07:58\] Yeah. So from the same incident, one of the companies that was breached through the SolarWinds incident was Mandiant; it now belongs to Google. And they were the ones that actually detected that there was something wrong in the network and reported it... And so kudos to them really, great job on figuring out that something was wrong. But one of the things that the attackers were very interested in was getting access to the tool sets that Mandiant was using for their own penetration testing and red teaming engagements. And it so happens that the tools that they were using were actually written in Go language, which I think is really interesting from an analyst perspective. So I think there's an interesting discussion to have about why they chose this language for their own offensive tools. There are a number of other projects on GitHub, which I can probably -- I can think of one called Stowaway on top of my head, which has been also reused and modified by some threat actors... + +**Natalie Pistunovich:** We'll add a link to that in the show notes. That sounds interesting. + +**Ivan Kwiatkowski:** Yeah, sure. It's a networking tool. It's really something that proxies the stuff in and out of a network that goes between protocols, and that kind of stuff. It's written in Go language; pretty annoying to reverse-engineer it, because it's a lot of goroutines talking to each other, very hard to figure out how it's architectured. + +And another example I can think about, that I'm not 100% sure, but I do believe that a commercial backdoor called Brute Ratel, which is a big competitor, or a new competitor maybe to Global Strike, which places enormous emphasis on evading detection, and being able to slip through EDR solutions etc. is also written in Go language, I do believe; but I would have to double-check that. So these are examples of malware families written in Go language, and I think that over time we're going to see more and more of them. + +**Ian Lopshire:** Why do you think we're going to see more and more? Is there a specific reason? You mentioned that they were hard to reverse-engineer... Is that part of it, or all of it, or...? + +**Ivan Kwiatkowski:** Yeah, there are a few reasons. The first reason I think is probably related to the ease of use for the developers. I don't mean that Go is easier to program than other languages, but the fact that it generates statically-built executables, binaries that are self-contained, that do not need any additional libraries, is kind of very comfortable for attackers. They create their backdoor, they send it to the victim, or even they deploy it at the victim one way or the other, and then it just works. You don't have to think about, "Is this DLL present on the system, or do I have to pull in additional libraries?" etc. + +So this is something that makes running programs very easy on victim machines, where you do not control the environment. A long time ago, maybe 10 years ago, it was kind of a problem, because you cannot send binaries that are two or three megabytes big to victims; if your attack vector is an infected PDF or infected Word documents, then you cannot really send over email a PDF that ends up being five megabytes big, because back in the day it would be rejected, or maybe the victim has some limit on their mailbox, or maybe they had a slow connection that is not going to be able to retrieve that binary. In Europe, or in the US, in the Western world it used to be fine, but if you think about victims that are in third world countries where the internet access is not as good... And it used to be some real issue for attackers; now that the internet connectivity is pretty much -- well, at least way better in most parts of the world, then having backdoors that are there 5, or 10, maybe 20 megabytes is really not that much of an issue anymore, I think. + +Then the second very good reason for using Go as an offensive language is going to be that reverse-engineering is difficult, which I will get back to, but also, all the standard tools that we as defenders tend to use in order to figure out quickly if a program is malicious or not tend to kind of break with Go language. The reason for this - and it ties into the discussion of why the reverse-engineering in Go is annoying for us, is that Go tends to really do its own thing. The assembly it generates really does not look like any other assembly. It's not like C, or it's not like C++ or Delphi, that kind of tend to look like distant cousins, or even brothers in some cases. Go really does things its own way, and all the automated methods or analyzing code statically, or maybe signatures you can recreate for Go language, etc. But old tools that would try to recognize specific patterns in code are not going to work, because the code generated by Go just looks like nothing you've seen before. + +So that's one reason... And then the final reason reverse-engineering is really difficult for us is because the constructs that are generated by the Go compiler tend to be very unfamiliar to us. And so the learning curve - I wouldn't say it's that steep... You mentioned, Natalie, that I had released a few videos about it - I think by the end of the videos you can have a rough idea of how to approach those programs. So it's not like an obstacle that is insurmountable; it's something that eventually you will be able to figure out. But when you've been working on similar-looking code as C for ten years, then sometimes learning something new is not something that you are easily going to do, because you have your comfort zone, and then you have to discover something different, and maybe you don't like to do this. And maybe you have ten easy malware redundancies that are waiting in the test list, and then you are going to work on those first, because it will allow you to end your day earlier next Friday, right? + +**Ian Lopshire:** So you kind of mentioned that there's assembly differences that make it hard to recognize... Are there any specific things that you've learned about Go under the hood from that? ...that differ from C, like how functions are called in the assembly, or something like that. + +**Ivan Kwiatkowski:** Yeah, absolutely. So one of the major differences -- it's not really about the assembly itself, it's about the static aspect of the executables. It's the fact that all the functions are pulled inside the final binary, and then you have this big program that's two megabytes or three megabytes big just for a print Hello, world. And now it's getting a bit better. I think IDA Pro has made significant improvements in its later versions... But maybe two to three years ago, when you were opening a Go program, you would have nothing recognized at all. Maybe you would be able to pull a few plugins here or there, or Python scripts that may or may not work... And in that case, if you were lucky, you might have been able to create signatures for the well-known functions, and maybe start from there, but it was really a huge ordeal. Now it's a bit better. + +So at least you are starting to get pretty reliably all the references to all the unknown functions. Beyond this, the Golang convention is -- well, I'm not going to it's weird, because it's as valid as any other one, it's just not the same one that we are used to seeing. The main difference is that considering that Go can return multiple return values, then you cannot have the same system as we had before... Like, for instance, in the C program the return values goes into EAX, and that's it. No difference. I mean, the EAX register of your CPU. When it comes to Go language, if you have three, four, or maybe more return values, typically one return value and also some error objects, if I'm not mistaken - then you cannot put all that into a single CPU register; it just doesn't work. And so you tend to get values that -- well, in the past, you would have all the arguments being passed through the stack, not through pushes, but direct moves from the value into the stack, directly. + +So the instruction was not pushed, which these assemblers or automated analysis tools - they just like to see push, push, push, and then call. That's something that is easy for them to recognize. But Go would just do "Move this on the stack, at this place, move this on the stack, at this place", and then when you go to another function, it knows, because the compiler knows where the stuff ends up. So it figures it out. But the IDA Pro looks at this and is like, "What the hell is this? This memory has never been initialized before. I cannot show this to you." There was an issue, and then the return values were given back exactly the same. So the program would just move back all the return values onto the stack as well, at places that it would be able to figure out later, but then when you look at IDA Pro, then it sees "Okay, values being moved on the stack", you go back into the Golang function and then you see references to the stack as well, but the offsets are going to be different, because since you are returning from a function, those things have shifted a little bit, and so the offsets do not work well anymore... And so this is like another issue that you have to face, like figuring out where your return values go. It still is, by the way, a terrible nightmare. + +\[16:18\] And finally, there is this other key difference... And this difference is the fact that usually the C compiler and other similar compilers will tend to reserve some space on the stack for specific local variables... And this tends to be very reliable; it doesn't move too much. So when you have some variable in C, it gets used in some part of the program, it's in that one place in the stack, and then that's it. And if the program needs another local variable later on, then there's just another space located for this in the stack. And the Go compiler tends to be very smart about these things. What it does is if it sees that there used to be a variable at some place on the stack and it's not used anymore, then it will feel like it's totally okay to reuse the same space to store something else later, which makes total sense. I mean, do not use more memory than you need to, right? But the Go compiler is totally right in doing this. But for me, it's really, really a problem, because what I do IDA Pro is I try to figure out where the local variables are in the stack, I name those positions by saying, "Okay, this is the error variable, this is the integer that represents an iteration count", or whatever, and I name or rename everything I can, and then eventually stuff starts to make sense, because I know what represents what on the stack, and I know what the variables are etc. But the thing is, if one position on the stack does not consistently represent a specific variable, then I cannot rename things anymore, right? There's just no way for me to do this, and the tools that we have, such as IDA I'm pretty sure Ghidra is going to function the same way; it's not going to allow me to say "Okay, up to this point, this variable should be named like this, and then from there on, then it should have another name, and then yet another" etc. + +So this is like a very, very difficult thing for us, trying to track down variables and return values, even arguments, is something extremely complex... And basically, this is the normal flow of how you analyze a program - you try to figure out what the variables are, you try to look at the functions and how they are called, what they return, and that kind of stuff. And just doing those simple things that would be the basic operations and building blocks of trying to understand what is going on in some random program are in themselves extremely complex operations due to like optimizations that were performed by the Go compiler. + +Now, the last thing I can mention is that since version probably 16.1 or something like this, or 1.16, I guess, in Go, the Golang convention actually changed, and they do things even smarter now, which is pass some arguments through the registers and not through the stack. For me, it doesn't change that much. Actually, it makes things a little bit easier, because at least I know argument one is \[unintelligible 00:18:50.27\] from memory. It might not be that one, but generally, it's going to be in a fixed register, at least for the two first arguments, and so I know where they are. That's way better. But overall, this doesn't change this bigger game of renaming things, which is not possible anymore. + +And then when it comes to the quick and easy mode, which is getting my super-expensive IDA Pro license that comes with a decompiler, then I just open a program, press F5, and hopefully I can read whatever is going on in the program - well, that just doesn't work, because the constructs that are generated by the Go compiler, especially I think when it comes to function calls, is totally alien to IDA, and every time you try to decompile code that comes from the Go language, you just end up with something that makes absolutely no sense... Because again, IDA tries to recreate pseudo C code, and pseudo C code that has just no way of representing concepts like multiple return values, or that kind of stuff. So this is a way that Go breaks everything that we hold dear in the reverse-engineering world. + +**Natalie Pistunovich:** \[19:53\] For anybody who didn't watch the video or is not familiar with how to do reverse-engineering, I can in simple words say that roughly you look at the instructions, and then you try to kind of see - the entry point is usually main, so this is probably function main, this is one thing that's been returned, and then you kind of try to follow that... Basically, this is what you do when you reverse-engineer. + +**Ivan Kwiatkowski:** Yeah. Actually, maybe I can say a few words about what reverse-engineering is for people that might not be familiar with it. The general idea is that we try to understand what a program does, even though we do not have access to the source code. But this is the typical case for malware, because we cannot call up malware authors and tell them, "Okay, please show me the code, because I don't really understand what going on in there." We don't know where they are, they don't want to be found, and they don't want to give us their code anyway. So what we have to do then is - we have no other solution but to look at the program and see what instructions the program is sending to the CPU, and then try to figure out from there, based on those instructions that are working at the CPU level, what the higher-level line of code that might have generated this type of instruction might have been. So it's not entirely a guessing game, because it's sort of a mostly exact science... But also, it's a very unnatural operation to perform, because this CPU language was really made for CPUs and machines, and for us humans, it's extremely difficult to understand. It's really not something natural for human beings to read those instructions. It doesn't make sense to us, and it really requires a lot of effort to figure out what the programmer's intent was just by looking at those instructions. So this is why, actually, we are looking for reverse engineers. I mean, not just at Kaspersky; the whole industry is looking for people that are able to do this, because it's something that most people find unpleasant, and I have to say myself, I do find it unpleasant most of the times... But at the end of the day, when I am able to figure out what was actually happening in the program, I feel very good about myself, and so this is the reason why I still do this job. But overall, this is kind of a difficult thing to do, and it's kind of painful, and it takes a lot of time to be able to figure out even the simplest programs. + +**Natalie Pistunovich:** Especially when the tooling is not even there for you. + +**Ivan Kwiatkowski:** Yes. + +**Ian Lopshire:** Just for some reference, the ratio between lines of, say, Go, to assembly - do you know what that ratio is? Just roughly... 1 to 100, 1 to 1,000? + +**Ivan Kwiatkowski:** It's a good question; it would depend on the complexity of the line. In Go I'm pretty sure that you can do function calls that are chained together in long lines. I'm not sure if it's compliant to the official Go styling code, or something like this... But if you were to do this, then you would have a -- I mean, let's take it from the other way. If you have some normal-looking Go code, like a Hello World or something like this, it would probably translate into 10 or 15 lines of assembly. So I'd say the default would be 15 lines of assembly for one line of actual Go code. But then if you get up into lines of code that are a bit more complex, that return multiple return values or function calls, then this can get a bit bigger... But this is still going to be the right ballpark. + +**Ian Lopshire:** Okay. Yeah, that gives me a good idea. + +**Natalie Pistunovich:** What does it for other languages? Is it a lot more? Is it a lot less? Is it roughly the same? + +**Ivan Kwiatkowski:** I would say it's probably going to be mostly the same. C++ tends to be very \[unintelligible 00:23:09.17\] it's very comparable to to Go. C might be a bit more direct, like the translation between C and assembly is going to be a bit more -- how would I say it in English...? The correspondence between C code and the assembly is going to be a bit more direct. That's it. But otherwise, I would say this is like a common ratio for languages. The problem is not that Go generates more assembly, the problem is that the assembly generates is not the one that we are used to seeing, and we don't like that. + +**Natalie Pistunovich:** Interesting to see if in one or two years from now it will be more supported and more pattern recognition working... + +**Ivan Kwiatkowski:** Well, that's the thing, right? It kind of depends on the attackers. If we do end up seeing more and more Go tools out there in the wild, then there's going to be pressure on the tool authors, like either IDA, Ghidra etc. to implement better detection, and better support for those languages. I'm pretty sure that since last time I tried using a decompiler on some Go program, IDA has made improvements, and it's probably not as broken as it used to be. But if we keep seeing offensive tools written in Go, then I'm pretty sure that the tools will get better. + +\[24:16\] We will still have to figure out how the Go assembly works, especially if it changes again in the future... But overall, at least the support in the last years has improved tremendously, and I think it will continue to do so also in the future, if there is a need to. And I would guess that Go is only going to become more prevalent when it comes to offensive software. + +**Natalie Pistunovich:** Because of all the reasons that you mentioned. + +**Ivan Kwiatkowski:** Yeah, exactly. + +**Natalie Pistunovich:** Some specific questions... You mentioned that -- you were kind of thinking out loud about the behavior you see in IDA Pro when you were looking at the Go code that you loaded there, or the binary of it that you loaded there... So some -- I'm gonna describe two things that you mentioned, and tell me how if you think it's good, if it's bad, how it compares other languages... This is an interesting kind of point; it can get too deep, so we'll try to keep it on a slightly high level for everybody who is kind of hearing about this and not very well familiar... So for example, you mentioned that skipping to the next instruction lands you in another place in the code of the CPU instruction. + +**Ivan Kwiatkowski:** Yeah, exactly. So this is something that was super-surprising to me, which is when I reverse-engineer programs so we can look at it statically in IDA Pro, which means you display the instructions and you read them like a book... Or there is another approach, which is not like opposite, but maybe more like a complement to it, which is to look at the program inside a debugger. The debuggers - they just work exactly the same as in the software development world; you execute the code instruction by instruction, or line by line, and you can see the state of the various variables. Except for us, we don't have the source code, so it's not lines of code, it's just assembly instructions. But we can still watch them execute one by one, and we can see the CPU registers getting updated etc. And when I was doing this with Go programs, I was very surprised to see that sometimes I would step from one instruction to the next and I would end up at a totally random place somewhere else in the program. + +And eventually, by doing some Google searches, etc. I figured out that it is actually the -- I don't know if it's the Go scheduler that is involved in there, probably it is, But there is a garbage collector that is in charge of freeing the variables that are not used anymore. And sometimes it takes priority and starts freeing stuff. And then once it's done running, it takes you back where you were in the program. And so this is something that is super-jarring for us as reverse-engineers, because we are looking at a very specific place in the program, we are frowning, looking very concentrated and focused (Because we are), looking super-serious. And then there we press F7, we step into the \[unintelligible 00:26:43.12\] and suddenly, we end up somewhere totally different, even though we didn't see any jump instruction. Suddenly, it's like, "Oh, something is going on. What's happening with my program there? ...because it's not supposed to just go somewhere else." + +Now, once I was able to figure out what was going on and understand that I just have to get out of this garbage collector function - and it will take me back exactly where I used to be, and things were fine, but initially, it was another one of Go's idiosyncrasies that felt super-alien to me. I wasn't happy about it at first. + +**Natalie Pistunovich:** So that means it's not a behavior that you see often in other languages... + +**Ivan Kwiatkowski:** Oh, no, it's something I had never seen before. I know that other languages, they do have their own garbage collectors, but when it comes to Java, we don't really have to look at the instructions, because Java is compiled to bytecode. So we just read the code disassembled or decompiled maybe, and get access to something that looks like the source code. It may be obfuscated, which means that it will be modified in a way that the variable names are not there anymore, or it has been specifically engineered to be harder to read... But in that case, or for .NET, or for Java, we just never have to worry about CPU instructions, because they are not that relevant to the language. So Go was for me a big surprise on that level, because this was the first time I had to encounter debugging your program and being taken far away somewhere, without even asking to. And it kind of happens on a regular basis, too. + +**Natalie Pistunovich:** \[28:05\] And then one more question about another behavior that was peculiar, that you pointed out... That at some point, when you had two following instructions, and they were using the same variable, you didn't see the return, but because it was right the one after or before. + +**Ivan Kwiatkowski:** I'm not sure if I remember exactly the part that you refer to... But what I noticed is - yeah, this might be one of the other ways that the compiler in Go is being very smart, which is that if you have chained function calls, it turns out I think that the way that arguments from one functions are returned on the stack happen to be the exact place where they would be considered as arguments for the next function. So you don't really see the data moving back and forth from the functions; you just have chained calls, and the compiler knows that whatever was returned happens to be at the right place for the next one, etc. + +So one of these other things that we are used to seeing, like we see a function call, we look at the input -- we look at what goes in and what goes out, basically; this helps us understand what is going on. And with Go, sometimes you just don't see that, because it's hidden from you. The complexity tends to be -- well, the complexity is still there, but all these operations are masked by the way that the stack is constructed by the Go compiler... Which, again, is a super-good thing for Go programmers, because it means that you don't have those memory movements that are taking place in the program that are actually not that useful. And every time you have a movement that involves the memory in a program, it takes a lot of time. I mean, not a lot compared to our human existence, but if you look at how a CPU works, you have the CPU that has some memory regions inside of it, which are called the registers, and then you have the RAM as well. And when you allocate memory in a C program with a malloc or calloc, it goes into the RAM. Or when you move something into the stack, it's also a region of memory that is on the inside of the RAM, the RAM stick of that computer. Every time the CPU has to talk to the RAM sticks, there has to be an electrical signal that goes from the CPU through a bus to the motherboard, and the motherboard understands it has to request the specific region of data to the RAM sticks, and you have the response that goes back the same way, converted into electrical signals. So it's pretty fast, of course, when it comes to -- it's probably in the ballpark of microseconds or milliseconds... But compared to just the CPU talking to itself, or moving stuff inside of the physical area that is the CPU, or just not moving things at all, because they are already in the right place, then you get performance increases that I think are pretty significant, especially considering the amount of function calls that you have in the program. + +**Natalie Pistunovich:** It's very interesting to hear about this from the perspective of somebody who's kind of poking this out from the outside... + +**Ian Lopshire:** No, that's this makes me want to dive more into the reverse-engineering just to learn more about the internals. + +**Break:** \[30:48\] + +**Ian Lopshire:** So let's maybe move to a bit of a higher level now. Go's community is kind of big on consistency; we have like the linters, that keep everything consistent, go format keeps everything consistent... Does that actually helped with reverse-engineering at all? ...just the only one way to do thing. Or at the level that you're doing reverse-engineering if you think it doesn't matter. + +**Ivan Kwiatkowski:** It's a good question. I have to say, I don't know that much about the linter itself. I have written a bit of C code myself. When I was trying to like look at assembly code and write Go at the same time, that would generate the same thing. So this is my extent of the experience with the language, and I really noticed something, which is that the Go language is super-strict. I have, in the past, used the expression - maybe it's going to make you laugh... I was saying that in Go if \[unintelligible 00:33:13.22\] return values, then the program is complaining. If you have unused variables, then the program complains again, right? And I was saying that to me, Go feels a bit like fascist Python; like, it doesn't let you do anything that you want, Except if it follows the rules very strictly. + +For us, it doesn't matter too much, in the sense that those checks are enforced at the compiler level, right? It's something that if the code is not compliant, then you will not get a binary at the end. So it does not add additional stuff inside the binary, And also, if there were some variable that is unused inside the program, then as reverse-engineers, we would not care, right? Because we would just consider that it's not used anymore, or probably the programmer doesn't need it, for whatever reason, and we would just move on. + +\[34:01\] So for us, it doesn't really change that much, although knowing about those guarantees kind of allows us to make more informed guesses about what is going on in the program. Like, for instance, when I do you see a function that returns multiple return values, then I am not a Go developer, but still, I am always going to assume that the last value returned is going to be the object; or the first one, I don't recall. I will have to check. But I know that since this is the normal way that people are supposed to write Go code, and since I know that the compiler is going to force people to do it, even if they don't want to, then probably I can base my hypothesis on those conventions, which is actually pretty helpful in that regard. + +**Natalie Pistunovich:** So would you say that Go is a good language to pick up for a hacker, or for a researcher in security? + +**Ivan Kwiatkowski:** Well, I'm not really in the business of helping attackers new being more efficient at varying offensive tools... But if I were to, then yes, I would guess that Go is probably a good language to pick up. Basically, anything that is away from the traditional languages is going to be more annoying for us, because we're less used to it. I think Rust is going to be a good choice as well. I haven't looked at Rust too much myself. I have a coworker that did, and also recent videos... And from what he's saying, it's like C++, but harder, which is kind of a high standard to beat. So yeah, just Go and Rust would be my advice there... Although it's not advice; please, don't. + +**Ian Lopshire:** So if those are kind of the new school ones, Go and Rust, historically, what languages has everyone used on the hacking side and on the research side? + +**Ivan Kwiatkowski:** Well, historically, everything has been used. You know Murphy's law, which says that if there is a way to misuse something, then it's going to be misused, right? And programming languages have proven time and again that law. The thing is, we are recipients of whatever the hackers are doing, right? We do not get to choose what we are going to work on. Like, hackers are going to write their tools, and they're going to choose whatever language is familiar for them, or whatever language feels comfortable, or whatever. And this is why we end up sometimes facing the most ridiculous stuff, like malware written in AutoIt; I don't know if you know about this... It's some weird scripting language that is used for UI testing, and basically allows you to simulate keystrokes and mouse clicks. Well, it turns out people write malware with this as well. Anything that has ever been available as a programming language has been, one way or the other, eventually used for malware. + +So the thing is, this is our bane as reverse-engineers, which is that we receive malware, and whatever it is, we have to work on it... Because at the end of the day, our job is to figure out what was going on in that specific incident. And so whether it's C, or C++, or it's Go, or Delphi, or Pascal, whatever... Erlang maybe... I'm pretty sure there's an Erlang malware. Whatever we receive, we have to work on, and so we cannot really afford to be picky about what languages we get interested in. We just have to be able to adapt to whatever comes, because everything will come eventually. + +**Ian Lopshire:** So you just mentioned right there, your research is on whatever hackers leave behind, let that be malware, or whatever. What other things do people leave behind? Is it just the actual binaries? Or like, are you digging into logs, and other things? + +**Ivan Kwiatkowski:** Yeah, so in a typical incident scenario, then you would have people that go into what we call forensics mode; they will collect all the logs, they will collect all the hard drives and try to figure out exactly what happened inside the network. They will collect not just machine logs with DNS logs, they will collect whatever event was generated by the Windows machines, they will collect whatever was saved by the HTTP proxy, and so on... All the NetFlow if it's available... Usually, it's not. Usually, not that much information is actually available in case of an incident. But that's someone else's problem. I'm not an incident responder, and I have enough stuff to worry about. But what I focus on is the actual malware. We do have information through the antivirus from Kaspersky that gives us information about the execution context... So we can see that, "Okay, this process launched this process", etc. So we have this type of information. But in a bigger incident context, then you would get a much clearer picture about everything that went on in the victim's network. And this whole trove of information would allow you to reconstruct the whole timeline of the incident. + +\[38:23\] So you would see that, you know, at this time, you had some suspicious request on some web frontend, and then you'd see that there was a file created at a later date on the same web server, and then you would maybe see some weird, suspicious request to the Active Directory server, with some golden ticket with meme cats or something. Well, those kinds of lateral movement methods, etc. And at the end of the day, somewhere, some attacker would have to drop some binaries to help them either persist on the victim machine, or get further into the network, or deeper... Because they will try to do whatever they can without deploying anything. Some very careful attackers will not deploy anything on disk, and they will just deploy whatever program that they need inside the memory... Which is very stealthy, but also if the machine happens to reboot, then everything that was in the memory just goes away... And so if you have no way of coming back onto the victim's machine, then all the access that you have deployed is lost. Some very stealthy attackers will decide that they would rather lose access than leave forensics traces on hard drive. Most of them, like 90%, 99% of them will feel like they would rather leave some kind of trace, knowing that most people don't look anyway, and then leave stuff for us to analyze later, if we figure out that there was an incident and someone goes there, collects everything and just sends the binaries back to us. + +**Ian Lopshire:** You said the incident response teams are the ones that collect all that data, and all of that... + +**Ivan Kwiatkowski:** Yeah, exactly. So we do have such teams at Kaspersky, but most cybersecurity companies will have either their internal incident responders... + +**Ian Lopshire:** Swoop in? + +**Ivan Kwiatkowski:** ...or a contractor that they know often, that can be called at any hour of the day or the night, and that will come and just -- exactly, swoop in with the big guns if something weird took place. Now, it doesn't mean that we do not work in direct interaction with those teams. It means that this is their job, and then we get -- we are more back-office guys, where we get escalated some stuff, and then we look into it. + +But most of the intelligence that we create doesn't actually come from incident response cases. I think it would be a good idea if we were able to gain more information from that source as well. I think it's a very valuable one. But we work mostly on the telemetry collected by our antivirus - all the samples that are suspicious or that are uploaded to the cloud for analysis. And then we can also swoop in, but much more quietly, and look at all this data and see "Okay, this looks interesting, because, we've never seen this before", or it looks like some malware that we saw 10 years ago and we haven't seen since, and it has some modifications. And then we are interested in what happened since then. But our work tends to be a bit disconnected from the actual incidence, and really more focused on looking at the big data lake that we have, and try to understand what is relevant inside of it. + +**Ian Lopshire:** That's cool. Thanks for that insight. + +**Natalie Pistunovich:** From the other side of this equation, what are some tips you can give for writing secure software for people who do Go? Or in general, if it's not specific to Go, it's also useful. + +**Ivan Kwiatkowski:** Yeah. I think one of the main appeals of Go is that you don't really need to think about security as much as with other languages. Go is a memory-safe language, unless I'm mistaken, and the compiler is never going to let you do stupid stuff, like create an array that is too small, and then write stuff that goes out of it. Like, it's just not possible. So it eliminates a whole lot of bug classes, which we call memory corruptions; it's just not going to happen. You cannot do this yourself in Go. And it means that all the old school buffer overflows that plagued all the C and C++ programs for dozens of years by now just are not going to ever happen in the Go language. It doesn't mean that the program is going to be perfectly safe from any security issues, but the issues are not going to be related to "Oh, I made a programming mistake, and if there is a bug in my program, it's going to be exploited." It's going to be more related to design issues... A memory safe language does not help you implement a secure authentication scheme, for instance; it doesn't help you write a well-thought-out network protocol. + +\[42:25\] I saw that Go really helps you with cryptography. I noticed that it's very difficult to choose algorithms that are not safe. By default, you can only -- I don't think you can choose the algorithms in Go by -- I know you can do AES, for instance, but like the cipher mode, or those kinds of stuff tends to be, unless I'm mistaken, selected by default for you, and the defaults are good... So you're not going to be making those mistakes. + +But -- oh, yeah, the IV... I was working on some code in Go that was relying on AES. I was looking at trying to figure out exactly how the IV was generated, and so on. I was seeing that nowhere the developer code, and doing some research, I noticed that it was actually Go that would, by itself, generate an IV for the encryption, this initialization vector, and then it would append it somewhere in the final encrypted buffer. And so usually, in other languages, this is something you would have to do on your own, and this is a like a big avenue for making mistakes. Like, if you choose a stupid IV, like just zeros, or if you do not select one at all, then you're going to have encryption problems. Go would not let you do this. + +So it's very obvious to me that Go was created with security in mind, not for the developers, but by the Go creators. They don't want you to shoot yourself in the foot, and they are going to make sure that there is no way for you to do it, unless you really, really want to. + +Even though you do have all those kinds of protections, cryptography can be misused. If you choose a bad key, then nobody's going to save you from that. If your protocol doesn't work, then again, you cannot be protected from it either. But I think it allows people to focus on design flaws, instead of programming flaws. And this is already a huge burden off the shoulders of developers. + +**Natalie Pistunovich:** That is a very interesting insight. + +**Ian Lopshire:** That's interesting. I see a lot of complaints outside of the GO community, just like Hacker News, about "Go is choosing your defaults for TLS, or not letting you do certain things..." But that's one I'm firmly on board with. If I don't need to think about it, I don't want to. And I don't want to make the mistake. + +**Ivan Kwiatkowski:** Would you be able to confidently select your defaults for TLS? I mean, I don't think I would feel comfortable doing this. You have to be very well-versed in cryptography to be able to make those kinds of decisions. So it's very good that Go is not making you do this, I think, in my opinion. + +**Break:** \[44:44\] to \[46:37\] + +**Natalie Pistunovich:** Another interesting -- about your interest in Go, you mentioned that you started using Go because malware was thrown at you, kind of... + +**Ivan Kwiatkowski:** Yeah, exactly. So I wouldn't say that I've started using Go; I would say that I was forced to learn Go. Not that I am unhappy about it... I'm not saying it's a bad thing. What I'm saying is that I'm not really writing Go code myself. What I did was I had assembly that was generated by the Go compiler, and I was trying to make heads or tails from it. So what I did was I looked at the assembly, I was like, "Okay, this might be the Go code that generated this assembly", and then I opened my Go IDE and I compiled my code and checked if it was the same on both ends. + +Also, when I start to learn about a language when I want to reverse-engineer it, I think it's super-useful to write some simple programs and just compile it and see how it looks at the assembly level. You know, just create a simple, stupid C function. Not C function, but some function that adds two integers, or something that will allow you to see what types of function calls the program is using, what kind of constructs the language is generating. The things that I had to face there was, again, the Go compiler being way too smart for my uses... And it tends to in-line all the function calls that are too simple. What I mean by this is, if you have a simple function that does almost nothing, and you call that function, then the Go compiler will be like, "Oh, this is not worth a function call. What I will do is I will take the code of all this function and put it inside the calling function." And when you try to look at what a function call looks like in assembly, then this is not helping. But the good thing is, I was able to find the good flags for the compiler to disable all optimizations, and things then kind of worked out for me. + +**Natalie Pistunovich:** You mentioned that IDA, which is the main tool you're using, and the other tool, are not really supporting Go. So if anybody wants to try reverse-engineer, to get into that, but also want to do that with Go, what would you recommend how to do that? + +**Ivan Kwiatkowski:** So if you're going to reverse-engineer Go programs, I still think that you don't have much choice there. So you're still going to have to use either IDA Pro or Ghidra. I want to switch to Ghidra eventually, but I haven't done so at the moment, so I cannot speak too much about its capabilities. I'm told that it's being improved at a very rapid pace, so it's probably a good choice... But when it comes to IDA - it got better. I think that a few months back, maybe a year now, you had my good friend, Juan Andrés Guerrero-Saade from SentinelOne on the podcast, and probably he told you about the various plugins that he wrote to help people reverse-engineer Go programs with IDA. I also contributed to his repository myself, with some script that I find useful.. + +But overall, even though IDA might not be perfect for the job, it's still one of the two only tools that are available for the job. So you still have to work through it, no matter what. The thing is, I find myself thinking that even though starting with reverse-engineering Go is kind of difficult, it turns out that I find myself liking reverse-engineering Go programs way more than C++ programs, that tend to be extremely complicated with virtual function tables and the very complex structures that represent classes, and so on... Because when it comes to the Go language, it turns out that it kind of feels like a scripting language in the sense that everything ends up being a call to an API function, or a call to some function that comes from the Go standard library. And so if you're able to take a debugger and look at all the arguments after you know how to do that, but if you look at all the arguments of the Go functions that are documented, by the way, and look at the return values, then actually, the meaning of the program tends to manifest itself, even though you don't really understand all the instructions that are in the middle, and you cannot track all the stuff going here and there. + +\[50:25\] So overall, my advice for people that would like to get started with Go reverse-engineering is, okay, it's going to be very different from what you're used to, but at the end of the day, I think you're going to end up liking it more than you would think, because it's going to be way easier than it looks. + +**Ian Lopshire:** How about those listeners that haven't done any reverse-engineering, that want to get started? Do you have any good resources out there? I know that you personally have made some videos. Do you want to talk about that a little bit, and anything else that would be helpful? + +**Ivan Kwiatkowski:** So yeah, the videos that I put out are just related to the Go language. If you're going to get into reverse-engineering, I would not advise you to start with Go. Not because it's going to be harder or anything, but because probably, the basics of reverse-engineering are going to be related to traditional C code, or traditional assembly code generated by C. So this is going to be like your base knowledge of reverse-engineering, and then once you are comfortable with understanding what is going on with the C language, and all the assembly that you see most places, then you can move on to other languages and see how they differ from others etc. + +But I think C is always going to be used as a reference for other languages, in the sense that when you look at assembly, first you try to understand it like you would understand C, and then if it's different, you adapt from that. But if your baseline is going to be the Go language, if the one thing you know is Go and then you try to recognize whatever you learned with Go with another language, then you're going to be into trouble, because whatever you're going to see next is not going to look like anything you saw in Go. + +So we do have a few courses at Kaspersky, people can check them out if they want. There are a few interesting online courses as well. It's something for free, which is beginners.re; it's a website, it used to be free, maybe now it's behind a paywall, I'm not sure, but it used to be this big, big reverse-engineering course written by some guy, and it was amazing. You have a book, which is called Practical Malware Analysis. It's a bit old now, but I think it's still very much up to date. It's from No Starch Press. I think for beginners it's going to be a good way to get into the field, because it explains everything that is going on, it provides links to the various tools that you might need, etc. So good resource there. + +And finally, if you want to approach this from the fun angle, I can actually recommend extremely good Steam games that allow you to get a feel for reverse-engineering. One of them is called Turing Complete. The pitch of this game is you're going to build your own computer. And so you start with -- they give you logic gates, like XOR gates on electric cables, basically, and based on this, you have to build a CPU, component by component. And then you move on, with increasing levels of abstraction. + +So it's really super-helpful to understand how a program works, or how a computer works. It allows you to get this high-level bird's eye view of how a CPU is constructed, and how it's supposed to operate. And knowing how CPUs work is then very, very helpful when you are doing reverse-engineering. + +And then you have other games, which are from a developer which is called Zachtronics. These are weird puzzle games that are really related to computing problems. One of them is called TIS-100, you have another one called EXAPUNKS, and they are dubbed "the assembly games you didn't know you wanted." And it's actually a very apt description, because these games have their own weird and limited assembly language, and you have to solve puzzles with them. You have to program some sort of small machine in order to make it do stuff, and you have to do this with assembly. And it forces you to use the language, which has the super-good design side effect of making you learn how CPUs work, or making you more comfortable with handling those weird instructions by yourself. So these would be my recommendations for people that want to get into it. + +**Ian Lopshire:** \[54:15\] Yeah, I have not thought about games. I'm gonna check those out later, actually. + +**Ivan Kwiatkowski:** And actually, if you are working from a university, or if you're a teacher somewhere, Zachtronics. I think the company may be closed doors not too long ago; I think they are done making games, or they moved on to something else, but they used to have a very extensive education program, where if you are at university and you're doing some computer science degree or something like this, you could just send them an email and they would give you access to all their games, for free, basically, and you could use them to teach, or as teaching aids. I think it's amazing of them, and also, the games are really, really fun. They are fun if you like assembly, which I think is a pretty biased statement on my end... But I do still recommend them. + +**Natalie Pistunovich:** A lot of the things you said are like a cheatsheet for reverse-engineering. Lots of useful information, and I have so many more questions about specific things about Go and reverse-engineering; we might have to do another episode about this, because we are running out of time. + +**Ivan Kwiatkowski:** Sure. Well, I can come back whenever you like. + +**Natalie Pistunovich:** We will prepare our questions, we'll ask you about things like generics... + +**Ivan Kwiatkowski:** I will have to prepare those questions as well, I guess... But no problem. \[laughter\] + +**Natalie Pistunovich:** Now, it's time for an unpopular opinion. + +**Jingle:** \[55:33\] + +**Natalie Pistunovich:** So Ivan, what is your unpopular opinion for us? + +**Ivan Kwiatkowski:** Oh, my God, I totally forgot about that. But it's okay. The good thing is I do have many unpopular opinions, so I'm going to give you things off the top of my head, and you can tell me what you want to know more about. For instance, I think that cyberspace is never going to be regulated. I think that NFTs are a scam, I think that there is no political will to limit the sale of cyber offense tools... That kind of stuff. I do have a lot of unpopular political opinions as well, but I don't think I want to inflict that onto you. You've been very nice to me. + +**Natalie Pistunovich:** What do you think about the European rule about USB-C, standardizing USBs? + +**Ivan Kwiatkowski:** Oh, I'm very, very happy about it. I know it's some pressure put on some device constructors, but I've been carrying lots of different chargers for years, and I'm super-annoyed about this... And knowing that we are going to switch to like a single USB-C for every single device makes me extremely, extremely happy. + +Another unpopular opinion I have, which you can add to the list, is that I'm not really a big fan of Apple. Like, not at all. I don't like their ecosystem. And I'm not going to get into this, but one of the things I don't like is that people have to pay 40 bucks for new chargers, and they change chargers every time they release a new product. And I'm very happy that this is going to cut off this revenue stream for them, because I think this should have never existed in the first place. + +**Ian Lopshire:** What do you think about all the walled systems, like the Google Play Store, and the Apple Store, and the Amazon store? From a security practice perspective, they say it's safer. Do you agree with that? + +**Ivan Kwiatkowski:** Yeah, this is a very good question. I do have very ambiguous feelings about them. I do believe that on the security perspective, it's kind of a good thing, in the sense that yeah, it's another one of those safeguards that prevent people from doing stupid stuff with their devices... And having to go to some friends' places, or more specifically, friends of my mom's places to debug computers, and uninstall malware, and fix the printers, then I'm very happy when there are protections that prevent them from doing that kind of stuff. But then again, they are not a perfect solution either. I think the Apple Store in terms of security is pretty good. The Google store, the Play Store has a bad track record when it comes to hosting malware. I'm not saying that they're doing a bad job; I think it's a very, very difficult job. But the fact of the matter is there are a number of apps on the Google Play Store that turned out to maybe not be total malware; some of them are, but a lot of them are just there to collect personal data, or that kind of stuff. + +\[58:23\] So I think a better way of securing those devices is not to control the app stores. Creating protections on the device level is probably where I would work. So when you look at both iOS and Android, they are doing, I think, a very good job of -- or have been doing a very good job, at least in the past years, of making sure that apps would not be able to access anything just because the user clicked Ok way back when they installed the app. So I think making sure that all those personal information cannot be pulled so easily is going to be a much better way than trying to police all the stores, and look at all those thousands of apps that are updated there every day... Which I do not think that you can realistically ensure that they are always going to be safe. + +But overall, the other issue with walled gardens, which is, okay, maybe they do provide something with security, but also I feel like they take away some agency from me as a user, right? I really like to own the devices that I use, and having some restrictions that tell me "Oh, you cannot install this app because Google says you can't", or "You cannot uninstall this app, also because Google says you can't" is something that tends to make me extremely, extremely angry. + +**Natalie Pistunovich:** So you mentioned a lot of unpopular opinions... + +**Ivan Kwiatkowski:** Yes. + +**Natalie Pistunovich:** The way that Twitter works for our podcast is that we take an unpopular opinion and then we make a vote. So there's a poll - do people agree with you or not? And then there's a Hall of Fame for unpopular opinions, and for popular unpopular opinions. So you listed several... Which one would you like us to vote on? + +**Ivan Kwiatkowski:** So if I wanted to win the contest, I guess I would go with the NFT one, because I know that this is something very divisive, and I think that a lot of the audience that you are reaching is going to be probably -- I'm not going to say that they are necessarily going to be on my side, but I think they're going to be on a side. But I think a much more interesting question that I would be actually interested in having the committee's opinion about is the one about regulation. I do believe that cyberspace is never going to be regulated, and maybe I need to say a bit more about this one, so that people can figure it out for themselves... My opinion on this is that -- we have a number of high-level discussions taking place at the UN about acceptable norms for behavior in the cyberspace, etc. And you have all these discussions between states, where they talk with each other, and they are like, "Okay, what type of offensive operations are legitimate?" Like, for instance, espionage is okay, but destructive attacks are not okay. I mean, I'm not saying this is right, I'm just saying this is probably the kind of discussions that they're having. And we may have differing opinions on what types of attacks are okay, and what types are not, or even if attacks are okay at all; it doesn't matter. + +The thing is, I do believe that -- I don't think that we will ever reach an agreement there, because, well, states do not have an incentive to regulate cyber offense. I think that they have an interest in having a way, or having some kind of framework that allows them to still conduct operations, because when they conduct operations, they know what they are winning, right? They have intelligence services that gather data, they collect it through cyber means, they take it back, and so they know that they are able to achieve certain results, because they have obtained specific information, and they can quantify that. + +On the other hand, when you look at the cost of cyber offense, which means all your companies in your country that have been breached because there are no such norms, it's something that's super-hard to quantify. You can never know that you lost some contracts overseas to sell planes, or to sell something else because of cyber means, because it's very likely that nobody knows that the breach even happened in the first place. + +So the thing is, you look at the balance of risk/reward for the decision-makers, and they see "This is what we win with cyber offense", which is a lot. And what they lose - it's painless. And also, they have no idea what it is. And so overall, I think that all those discussions that are taking place, that are saying, "Okay, we need to make a safer internet, blah, blah, blah" are actually possibly being conducted in bad faith, because there is no political will to actually stop doing this kind of stuff. This would be my unpopular opinion, especially in the diplomatic circles. + +**Natalie Pistunovich:** Alright. You will be tagged, and we will be following the results. + +**Ivan Kwiatkowski:** Okay. + +**Ian Lopshire:** I'm interested to see the results on this one. + +**Natalie Pistunovich:** Yeah. + +**Ian Lopshire:** It's an interesting way to think about it. + +**Ivan Kwiatkowski:** Yeah, I want to know what's well. + +**Natalie Pistunovich:** Cool. Thank you very much for sharing your knowledge, your thoughts and your opinions with us. This was really fascinating. We will be very happy to have you again. Thanks a lot, Ivan. + +**Ivan Kwiatkowski:** Well, thank you very much for having me. And yeah, feel free to call me up anytime, and I will be happy to be back. + +**Natalie Pistunovich:** Thanks, Ian, for joining. It was fun co-hosting together. + +**Ian Lopshire:** Yeah. Thanks to you guys. This was great. + +**Natalie Pistunovich:** Bye, everyone! diff --git a/Hacking with Go Part 3_transcript.txt b/Hacking with Go Part 3_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..181cfdf638270b964a3b4846b57a3d827b3caa67 --- /dev/null +++ b/Hacking with Go Part 3_transcript.txt @@ -0,0 +1,239 @@ +**Natalie Pistunovich:** Today, Ivan Kwiatkowski, you are joining us again, to talk more about hacking with Go and cover all the things we did not manage to cover in our last episode. + +**Ivan Kwiatkowski:** Yes, very happy to be back here. + +**Natalie Pistunovich:** So for those who did not tune in to the previous episode, could you shortly introduce yourself? + +**Ivan Kwiatkowski:** Yes, of course. So my name is Ivan Kwiatkowski. I am a French cybersecurity researcher, and I work for Kaspersky. Specifically, I work in a threat intelligence team, and my role in this team, apart from writing reports, and that kind of stuff, is really to proceed with the reverse-engineering of the malware that is provided to me by my coworkers. So basically, they do the threat hunting, they find some interesting stuff to look at, they identify implants used by the attackers, and they give them to me. And then my job from there is using the fantastic tools of the reverse-engineer, which are IDA Pro, basically... I do read the assembly code of those programs and try to figure out what they do as best as I can. And that's my life. + +**Natalie Pistunovich:** And in the previous episode, we talked a little bit about how IDA is yes or no support in Go, and that it got better over time, but still place to improve... + +**Ivan Kwiatkowski:** Yeah, absolutely. So if you tried reverse-engineering Go programs something like maybe two years ago, maybe just one year ago, then you would be in a lot of trouble, because the tools just weren't there. So it involved using some third-party plugins, some code taken from - I wouldn't say suspicious, but GitHub repositories that weren't that well maintained, or that didn't have really clear instructions... And so it was really a difficult path for the reverse-engineer that had to do it. And thankfully, over the years, the developers from IDA, this company in Belgium called Hex Rays, they've been listening to the customer complaints, I suppose, and they have made a number of improvements that allow us to support Go programs a lot more easily. And that entails having some recognition for the various functions that come from the Go standard library, better support from the Google executables as a whole... In the very latest versions, if I recall correctly, a number of the things that were actually implemented in some of the third-party plugins that Juan Andres Guerrero-Saade from SentinelOne, as well as myself, had implemented manually in Python. Those kind of features, they include them into IDA Pro, into the main line. + +So I do expect that one year from now maybe this won't even be a discussion anymore. I mean, sure, Go will still be a very kind of alien language for us to look at in terms of like the assembly code that it generates, but at the end of the day, I do suspect that the tooling problems are going to be over in the near future. So that's a good thing as far as we are concerned. + +**Natalie Pistunovich:** So is Go a better language for a security researcher to pick up, or for a hacker? + +**Ivan Kwiatkowski:** Well, when it comes to security researchers, we don't actually have to write that many programs. Most of the tools that we use are already provided to us by the community. I mentioned IDA Pro - nobody is ever going to redevelop IDA Pro. I mean, some guys did, so it wouldn't be fair saying that they wouldn't, but most people are not going to do this. And if they were going to do this, then I think that choosing either the Go language or C++ wouldn't make that big of a difference considering the scale of such a project. + +When it comes to a hacker, I think that for them the Go language is still probably a very good bet, because as far as I can tell from a very unscientific polling of my co-workers and other reverse engineers in the field, it feels like most people still really dislike having to work on the Go language; like really, really dislike it. + +In my opinion though, if I were to write malicious programs, I would use Rust, because the code generated by Rust is actually even way worse. And at the moment, I'm not exactly sure how to approach that type of code. I have to work on that, but this would be my intuition - to use Rust, because I know that I'm going to make someone's day miserable somewhere in the future. + +**Natalie Pistunovich:** And when you say "worse", you mean basically because the steps that you need to do to reverse-engineer and kind of figure out what's happening are actually more painful... + +**Ivan Kwiatkowski:** \[08:06\] Exactly. I say that in the sense that -- maybe it's just a personal thing, right? I have spent a bit of time trying to figure out how Go ticks, not at the language level, but at least at the assembly level; I would never pretend that I am a Go expert, or that I have deep insights about the inner workings of the Go language... But if you give me a binary written in Go, and I expect that eventually I will be able to tell you what it's supposed to be doing. When it comes to Rust, it's kind of uncharted territory as far as I'm concerned. As far as I know - and again, this is not something that I've checked all that much, although this is something that I probably will have to do very soon... It feels like the tooling, like IDA, for instance, the disassembler that we use on the job doesn't seem to support the Rust language as well as it supports the Go language. It doesn't recognize as many things. And at the end of the day, Rust tends to generate constructs that really look like C++, and C++ is kind of a mess to begin with. It's a very powerful language, I really love C++; if I have to write some complex program, I will write in C++, because this is the one I have the most experience with. But when it comes to reading assembly written in C++, oh my God, it's just so convoluted, and there are so many levels of indirection added at every level. So this is not something I would be happy with, and Rust being a new, more complex, or a new, more alien C++ is really not something I would be happy with. + +**Natalie Pistunovich:** How about the easy cross-compilation of Go? The fact that you write your code, and then you create one binary, and then you kind of ship it as is, and on top of that you just write one more command and then you have it for any architecture? Does this make any difference for you? + +**Ivan Kwiatkowski:** It does, in the sense that I think as a developer, it's pretty cool to have that. I also have always felt that this feature that is very often brought forward by defenders of the Go language was a bit -- I don't think it's that important, right? Not that we don't want cross-compilation or that we don't want programs that can run anywhere, but I don't feel like the Go language is especially adding something new there. I mean, when I write C++ code, it can already run everywhere, provided that I write it properly, of course. Way back when, probably like 10 years ago, when I was in school, I was learning Java, and I actually programmed a few projects in Java code. Supposedly, Java was supposed to be working on any platform as well, right? So it didn't work as well as we expected due to many reasons, but overall, it doesn't seem to me that this ability to run on any platform, and to have a code that will compile everywhere is really something that Go is actually bringing to the table. I think this is something that we already had, and that maybe Go is making easier for a lot of developers. But it's not something new, and something that would make me switch languages, by any stretch. By the way - maybe this is a question for you... I have no idea about the support for Go on exotic platforms. What about running Go code for Solaris, or for the ARM architecture? Is that something that's supported out of the box, or isn't it? Because I know for a fact that when new CPUs come out, the first thing that the manufacturers release is going to be a C compiler, or a C++ compiler. So we know that eventually, those languages are always going to work. But I have the motion - and maybe I'm wrong about this - that when it comes to the Go language, that if you have this new platform somewhere, then you will have to wait for Google to release the corresponding compiler, and that may take some time. Is that correct? + +**Natalie Pistunovich:** So as you're asking, I googled the command \[unintelligible 00:11:35.24\] which is what you run to do this, and for Solaris, it does come out of the box. What was the second one you asked for? + +**Ivan Kwiatkowski:** This one is going to be supported, I would imagine. It would be to compile for ARM or MIPS, probably less used architectures... But I would imagine that at least ARM is supported very well. + +**Natalie Pistunovich:** Yes, ARM is out of the box indeed. Yeah. + +**Ivan Kwiatkowski:** Okay. + +**Natalie Pistunovich:** ARM-64, and so on. Yeah. And MIPS - some of the variations are not, but most of them, yes. The ones that are not out of the box is MIPS-64 P32 and MIPS-64 P32LE. + +**Ivan Kwiatkowski:** \[12:15\] Okay. So yeah, overall, these are probably architectures that most people don't care about. So I don't think this is a big fault on the part of the Go language. What I will say though is that, as far as I'm concerned, C is already a multi-platform language as it is, and if other languages provide this as well, then good for them, but to me, it's not something groundbreaking. + +**Natalie Pistunovich:** Yeah, that's fair. Lots of DevOps people do love that feature, that you don't need to do much to ship everything to everyone in your favorite architecture. + +**Ivan Kwiatkowski:** Yeah. That aspect is pretty important, and also very appreciated by malware authors; it's the fact that - yeah, when you write some program in C, then you might have modules that are distributed in the form of a DLL file in Windows, or .so shared object on Linux, and so on... And then you end up with a program, an executable file, and then several object libraries that come with it. And then when you want to distribute it, you have to send this big archive that contains many files. I would agree that when it comes to the Go language, you end up with a single binary, and that's pretty useful, right? Especially in the context where you don't have control over the client, the "victim's machine"; then in that case, then of course, just having to send an executable and knowing that it is self-contained, and going to work everywhere, all the time, is going to be a big advantage. There are ways to do this in C, C++ and so on, but it will require some work... Which, agreed, is not reported on when you're using Go. So this is one point for Go, I would say. + +**Natalie Pistunovich:** Another is - tell me what you think - the concept of modules, where it has this file that says what are all the dependencies, and what version specifically is used where, in case you're using some package of an older version, and whatnot. So given that this is all kind of compiled into a module, and it's being sent out as one - does this provide any value for a hacker or for security researcher? + +**Ivan Kwiatkowski:** Well, for a security researcher or a hacker, I'm not exactly sure. I mean, as far as developers go, it prevents you from falling into this pithole of dependency hell, where -- this is stuff I actually experienced last week, while working on a Python project. I updated all my packages, and some of them were not compatible with each other, and my whole project just broke down in production... Which is always fun. So this is something I could have prevented by just fixing the version numbers, which is what you're supposed to be doing. But overall, having this mechanism is kind of a good thing. + +When it comes to security, making sure that hackers are able to compile their thing is really not something we worry about too much. What we do worry about is the fact that when you end up with a single binary that contains everything, it's kind of an issue for reverse engineers. when you compare this to a C program or C++ program that has some various DLLs, then the different files already represent some sort of separation between the code, right? The DLLs might or might not have a relevant name, but at the end of the day, you know that they're going to be split according to some form of functionality, right? This type of code is going to go in this DLL, the main intelligence of the program will go in the main executables, and so on. + +So when you have those big malware platforms that you have to work on, then having several files is actually a pretty good thing for us. When you have a big Go binary, that is 5 to 10 megabytes big, and then you have to just dive in there and try to find out where the interesting code is located - it's a good thing; where the uninteresting code (which is the library code) is; that's something that the tools can recognize pretty well. But you still have this big program that contains everything, and it's just much easier if you already have this kind of separation, where you can already focus on some specific functionality, even though you haven't been able to dig into the whole project. So in that sense, I would say that this feature is pretty useful for the offensive side, I would say. + +**Natalie Pistunovich:** \[16:22\] I wonder if there's any particular feature that is good for the defensive side... But I'll keep asking questions until we find something. Or do you have one in mind? + +**Ivan Kwiatkowski:** I do have one in mind... The best, and I think one of the strongest selling points of Go over C and C++ and all those unmanaged languages is going to be that when you write programs in Go as a developer, you know that you're never going to have any problems with memory corruption, buffer overflows, and those kinds of issues. I would be very surprised if the Go language would allow you to read outside the bounds of an array, and that kind of stuff. So all this is already taken into account for you, and it's not going to make my personal daily job as a reverse engineer easier, but what it's going to provide is that by default, people will have a much harder time shooting themselves in the foot. And this is a good thing overall, because it means that if I download an application that was written by someone else in the Go language, I don't have to worry as much about the code quality, because I know that the language is going to provide a number of guarantees, and that will make sure that at least a number of vulnerabilities are not going to affect me, ever. This is true as a whole, for the whole industry, even outside security - if programs are written in the Go language, like an FTP server, or an email server, or whatever... If such programs are written in the Go language, and we know that at least we won't have to worry about buffer overflows, and this means less weekends spent in incident response engagements because some customer didn't patch their program, or because there was some vulnerability discovered as a zero day, and that's being exploited in the wild, for an application that has some buffer overflow vulnerability, and that is just available widely on the internet. + +So overall, for defenders, less vulnerabilities and less ways for developers to make tragic, tragic mistakes is always going to be a good thing. And I think this, in fact, overshadows any advantage that the attackers are gaining over us, on the personal level, with the reverse engineers. + +**Natalie Pistunovich:** That is a very interesting point. I guess it would be interesting to see if overall Go code is more secure, however you would measure that. It's probably going to be interesting to see how to do that. + +**Ivan Kwiatkowski:** Well, having precise metrics is always going to be difficult, but if you compare CVE number quantities for projects written in the GO language and for projects within C, then I think it's very likely that you would find that programs written in memory-unmanaged languages, like C, C++ and the like, are always going to have more bugs, just because by default there are more opportunities to shoot yourself in the foot. + +So if you take the developers with equivalent skill, and for one of them some bugs are just unavailable, and for the other one you have twice as much mistakes that you can make, it feels very obvious to me that no matter if the developers have the same level of skill, then the person using the unsafe language is always going to make more mistakes. + +**Break:** \[19:35\] + +**Natalie Pistunovich:** Have you ever seen malware in COBOL? + +**Ivan Kwiatkowski:** I haven't, actually. + +**Natalie Pistunovich:** I wonder how would you evaluate that - on the more safe, or on the less safe side, how would you say all that? + +**Ivan Kwiatkowski:** I actually have no idea. COBOL is one of the languages that I know of; I know that if he wants to work in banks and be paid the big bucks, then you should definitely learn COBOL, because all the former COBOL developers - they died of old age by now, so they are kind of hard to find... Beyond this, never in my life have I heard about -- or never in maybe the last ten years have I heard about malware written in COBOL. If I were to find one, then it would be a pretty cool blog article, but it would probably be a miserable week for me, because then I would have to probably learn the language and figure out how it works. Although, to be fair, this week I actually had to reverse-engineer a program that was written in Pascal... + +**Natalie Pistunovich:** Which is the spiritual parent of Go. + +**Ivan Kwiatkowski:** Yeah, I suppose. And also spiritual parents of many other languages, because it came from the '80s, maybe the '70s, right? I remember learning Pascal, or at least the basics, when I was in school; something like 15 years ago now. + +**Natalie Pistunovich:** Same. It was my high school graduation project; it had to be in Pascal. + +**Ivan Kwiatkowski:** Yeah. There you go. I probably wrote a little bit of Pascal myself; probably never actual projects that did anything meaningful... But I did have to look at some APT malware, like real-life APT. + +**Natalie Pistunovich:** What does APT stand for? + +**Ivan Kwiatkowski:** Oh, sorry, it means Advanced Persistent Threat. So APTs are one of the categories of hackers that we track in our daily work. So you have on one end the attackers that are financially motivated, or cybercrime, ransomware groups and all the like, and on the other end of the spectrum, you have what we call the APTs, which basically are the state-sponsored actors, or the mercenary actors, all the groups that are focused on cyber espionage. Now, initially, the name APT was, I think, proposed by Mandiant - it's probably around 2010, something like this - in their first report. And back then, I think it makes sense to say, on one end, we have the low-scale cyber criminals that are doing run of the mill crimeware, and on the other end of the spectrum, you have those state-sponsored attackers that are doing very sophisticated things. I think that today, trying to separate attackers between levels of sophistication doesn't make that much sense anymore when you have extremely skilled ransomware groups that use very cutting-edge pentesting methodologies... And we do have APTs that are extremely bad, I would say; they have poor OpSec, they don't know how to use their tools, and so on... So at this stage, I think in 2022 when you hear APT, you just have to think about espionage. I think this is going to be the way to understand this. + +\[23:59\] But in any case - yeah, this week I was working on another APT case, taking place in some of the STEM countries, in the CIS, Commonwealth Independent States, I think... Anyway, and one of the malware implants that we found there was actually written in Pascal, and so it was kind of a trip down the memory lane to, one, figure out what Pascal was again, and also trying to understand what kind of assembly was generated by this Pascal compiler. It wasn't that bad, actually. Way less bad than having to discover the Go language. + +**Natalie Pistunovich:** That's interesting to know, because there is a lot of similarity between Go and Pascal, but knowing that the translation of that is different... + +**Ivan Kwiatkowski:** Oh yeah, I can tell you, even though I'm not an expert in any of the languages, even though you might have some similarities on the code level, in maybe the constructs and the way that you declare things and so on, when it comes to assembly, the languages could not be more different from one another. + +**Natalie Pistunovich:** I thought if they're conceptually similar, they might have a similar structure, but I guess not. + +**Ivan Kwiatkowski:** No. I suppose they took inspiration when it comes to how to write the code... But then when it comes to what the compiler does, then yeah, the Go compiler really does its own thing. + +**Natalie Pistunovich:** And when you say that each language has their own different thing in the assembly representation, or even when you reverse-engineer that into like a visual representation, how many different ways can you have -- can it really be like every time completely different? + +**Ivan Kwiatkowski:** It's not always completely, completely different, but there are some meaningful differences. I would say that the C language - maybe it's a misconception I have, because the C language is traditionally what you learn reverse engineering on. So the C language to me is going to be the language that is closest to the CPU. When you compile a C program, then there's going to be a kind of direct translation from your C code to the assembly language. The compiler isn't going to be too smart about things; when you do something in C, then when you write it in code, then it kind of shows in the assembly language. And of course, you can add some compiler optimizations, for speed, for space, and so on, but overall, the translation is going to be a pretty -- I wouldn't say it's easily reversible, but I think it's pretty direct. You really find your bearings from the C code to the assembly language. And I think it's actually not that much of a surprise that decompiler tools such as Ghidra, or such as the one sold with IDA Pro do take the assembly and convert back to C language, because I think this is the closest. And then when you go to languages that have, I would say, higher levels of conceptual complexity, then this is where the compiler starts doing a lot of things on its own, and this is where the code that you write ends up being super-different from the assembly that you read. When it comes to C++ and when you use an STD string, it seems like a very simple thing, right? But under the hood, the STD string class is actually a template instantiation of a very, very complex series of nested templates, and you end up with... How can I explain this? You end up with a weird structure that has first a table that contains a pointer to methods, which is something you never wrote in C; you end up with methods calling each other, nested methods that come from the template library from the C++ standard library, and so on, and things just get crazy from here. + +Taking the example of Pascal - and again, I don't write that much Pascal code, but it's very obvious to me that when I look at the assembly code and I see reference counters being incremented and decremented automatically, and all that kind of stuff, then this is something that was automatically added by the compiler. And it's, I guess, useful as far as the running a program goes. But when it comes to me understanding the program, making sure that reference counts are handled properly, and that the objects are going to be free when there are no more references to them is something that I really don't care about, and it's just cluttering my window, really. It's just code added by the compiler that has no meaning, at least when it comes to what the program is supposed to be doing. It doesn't add any intelligence to program, it's just something that gets in the way. + +\[28:06\] Go is probably one of the far extremes of this, right there with C++, because the Go compiler is really doing a lot of stuff under the hood. It really -- how can I put this...? + +**Natalie Pistunovich:** Optimizations? + +**Ivan Kwiatkowski:** Yeah, optimizations. It in-lines anything that is not worth a function call... The calling convention is its own thing... Oh, you also have a garbage collection mechanism. When you write a simple Hello World program in the Go language, it ends up being an executable that's something like one megabytes big, or something like this. I get that today storage space isn't that expensive; we don't care about one megabyte of code. Like, what's the difference between this and seven kilobytes? I think not that much when it comes to hard drives. But when you are a reverse engineer, and you have to look at one megabyte of code instead of seven kilobytes, it's actually a big deal, right? And this is the kind of thing that the Go compiler does to you, along with a number of optimizations, along with this very weird calling convention that they have etc. + +Another thing that I don't like about Go - I mean, it's a good thing, right? I don't like it as a reverse engineer; it's the goroutines. + +**Natalie Pistunovich:** Yes. This would have been my next question. Yes, please do elaborate. + +**Ivan Kwiatkowski:** Great. So it seems like a very, very easy way to create threaded programs, as far as I understand, which is great as far as developers go... But it makes it a bit too easy for malware developers to create threaded programs as well. And when it comes to understanding what a program does, we really like linear programs. We want instructions that we can look at one after the other, we want programs that we can debug very easily... And as soon as many things start happening in many threads, then oh my God, following things around becomes extremely, extremely difficult. So I would actually like for threaded programs to be more difficult to write, and to be less available for attackers, if that were an option. + +**Natalie Pistunovich:** How do threads represent themselves visually when you do reverse-engineering? + +**Ivan Kwiatkowski:** Well, they don't really represent themselves, because threads as a concept, they are a fundamentally runtime object, right? A thread is going to be a unit of execution that is going to run some code. And when you have a single thread, which is the case for a lot of programs, then you can just follow what's going on in the code linearly, and then you figure out what is going on. When you have several threads, then this is really an order of magnitude more of complexity that you have to wade through, in the sense that as a reverse engineer, you not only have to think about what is going on in the program or in the function that you are reading, but at all times, you have to think about the fact that there might be another thread running somewhere that might be doing things that are affecting what you're doing right now, or what the current function that you're reading is doing. And so you do not have the luxury of having all the information that you need in a single place. The functionality, the intelligence of the program ends up being spread over different units of execution, and you have to keep everything in your head to have any hope of understanding what is going on. So this is really a very heavy mental tax that is imposed on the reverse engineer. And of course, the more threads there are, then the more effort you have to go through to try to keep track of everything that is going on. + +One good example of this is a Go program that I mentioned in the previous podcast, it's called Stowaway. This is an open source project that is used to do various proxying operations as a pen testing tool. You can create tunnels, SOCKs proxies etc. and probably pipe them with each other. I'm not exactly sure... But what I'm sure of is that when I was reading the program's assembly, it felt very, very miserable, because it was obvious that many things were happening at the same time, which of course is going to be the case, because when you have some network program, then packets can arrive from any end of the various terminals. And you can also have many tunnels running at the same time. + +\[32:05\] So you have all these things taking place at the same time, and trying to figure out exactly what does what is extremely difficult. And if I hadn't been able to figure out that this was actually an open source project for which I was able to go find the code, then probably I might not have been able to figure out everything that the program was doing at all, because there was just too much to work through, and too much to remember, because my memory is actually quite limited, as is the memory of any human being compared to a computer, really. + +**Natalie Pistunovich:** Until we started upgrading... + +**Ivan Kwiatkowski:** Yeah, I wish... + +**Natalie Pistunovich:** All the interesting terms that you're mentioning, like APT, and also a Stowaway, which you reminded now - it will be in the show notes, for those who want to look back at that and see how can they handle threads in reverse engineering. It will be interesting to see some example of how that actually looks... Because you said that this kind of can tint the results, and whatnot... Maybe like there's data dependency between the two, but... I still try to visualize and understand -- so I'm referring back to the video that you have published on YouTube, which I will also link, of your "Reverse-engineering a Go program", and you kind of build this block diagram of the different steps, and you write what's happening there, what you guess, and so on. So affecting the results by sharing some data, or do a calculation return, and then on top of that - that all makes sense. Do commands just pop up randomly in what's happening now, and you kind of try to paint that in the relevant context once you're reverse-engineering something with multi-threads? + +**Ivan Kwiatkowski:** Actually, I think the best way to visualize it is not to try to think about the program as it is running, but imagine that instead of reading assembly code, you are reading the source code of a Go program. I think you are not going to dispute the fact that if you receive some Go project from a friend or a co-worker, and you know nothing about this project, and you have to read the source code, then this source code is going to be much easier to understand if the program is just a single thread that is doing a single thing, right? If you receive this program that, as soon as it begins, it launches three different threads that are supposed to do different things at the same time, then figuring out exactly what the program is going to be doing is going to be, I think, much more difficult. Now, imagine the same thing, but instead of receiving proper Go code, then you would receive Go code where all the variable names have been wiped, and all the variables are named A, B, and C etc. So you cannot even use the function names, or you cannot even use the variable names to try to understand what the program is supposed to do. Imagine that you have no comments inside the code as well. This is basically what reverse-engineers have to do. + +**Natalie Pistunovich:** You don't always have to imagine that... + +**Ivan Kwiatkowski:** Yes, of course. \[laughter\] Yeah, this might be real life for a lot of people out there. Shout-out to them, I guess... But this is exact what reverse-engineering is. You receive some source code; whether it is assembly or high-level code - that does a bit of a difference, because assembly is hard to read... But basically, it's going to be the same thing that you have to go through, right? You receive some code, you have to understand what it does, and the more complex this code is, the more sophisticated its operations are, then the harder time you're going to have to understand exactly what is going on in there. + +**Natalie Pistunovich:** This brings me to the next question, that generally Go best practice, let's say, or the right way to do Go is to write simple, readable code, rather than sophisticated, and like ternary operations and whatnot, and complicated things... Does this in any way help, or not? + +**Ivan Kwiatkowski:** \[35:45\] I wish, but it turns out that for the compiler, whether you write the simple way or the ternary operation, if the compiler is smart enough, then at the end of the day it's going to generate the very same assembly construct. Hopefully, the compiler, if you do this "if then... else", or if you use the ternary operator, it's going to be able to recognize that is the same thing, and in the end it's going to generate assembly that does exactly the same. So it's a good thing as far as development practices are concerned... But when you reach the assembly level, then all those helpful things and all those precautions that you have taken to make sure that other people will be able to understand what you're doing - they just get taken out by the compiler, because they are things for humans; they are not things for CPUs, and so they have no place in your compiled program. + +**Natalie Pistunovich:** Yeah, that's a great point. + +**Ivan Kwiatkowski:** Actually, one callback from one question you asked earlier - you asked how many different things can the compiler do when it comes to different languages... An additional example I can give for Go is Go programs, Go functions, they can return any number of return values, right? This is not something that most languages are doing... So when you look at the assembly language, at the end of the day it turns out to be translated in CPU code in very different ways than normal functions are supposed to work, right? When you have a function that can only return single arguments, which is the case for a lot of languages out there, like C, C++, etc. then you have a very simple convention. The convention says, "Oh, the return value will be in register EAX in assembly." This is the rule; it's very simple. When it comes to the Go language, then you don't have a single place, because you can have several arguments. And so they are returned differently through the stack., and you have to go look for them... It's just much more complex, and it's very different from a traditional language. + +And the difference between the languages are going to be small things like this when it comes to conventions, but actually, some of all of them are going to result in having a source code or an assembly code at the end of the day that is really extremely different from one another. + +**Natalie Pistunovich:** Would you say, from what you see, that the way Go handles on the compiler level the return of multiple arguments is efficient, overall? + +**Ivan Kwiatkowski:** Yeah, it really feels like it, as far as I recall. Maybe this is me making a mistake, but it feels like the return values from one function call are placed exactly where they should be on the stack, so that another function can use them as arguments immediately. And so chained calls between different functions, where you have a function calling -- when you call the function by passing an argument which is the result of another function call, it feels like an assembly that these function calls are going to be very close to each other. You won't have to move back stuff from the return values back on the stack etc. It's just already there, and I think that on that level, they're going to be pretty efficient, and go pretty fast. + +**Natalie Pistunovich:** Nice. Good to know; it's always encouraging... So Go is kind of built in a way that you don't debug this line by line, with breakpoints and so on, as you do in many other languages, but you do something and you check for errors, all the time. I will not ask you whether malware is generally written like this or not, or how good are they with their error catching... Unless you know, and then please do share. + +**Ivan Kwiatkowski:** I do know, actually, because I end up reading the code, right? So what I learned about Go by reading assembly code, and also by trying it myself to understand what was going on, is the fact that the Go language will not allow you to not catch the errors, right? If you have a function that returns two return values, and if you do not catch them, then you're going to be in trouble. The code is not going to compile. So I think you can probably create this underscore variable, that means "I don't care." + +**Natalie Pistunovich:** Yes...! + +**Ivan Kwiatkowski:** Right. But as far as I can tell now, at least for the programs that I've seen, they do catch the errors, and they check for the errors, and they handle them properly... Which I think makes sense, right? Because if the language forces you to do it, then you're going to do it. Of course, you can circumvent this by using this special variable, and not actually checking, but if the mechanism is there and if the language creates the framework where you kind of have to do it, then it kind of feels foolish, I think, to not do it, even though you can. Because if you don't want to do those things and just go back to C and play without any safety belts, and just play by your own rules... But if you're going to use the Go language, I think it makes sense to use the language as it was intended. This is in fact what I'm seeing when I look at malware code. + +**Natalie Pistunovich:** \[40:25\] That is a little bit sparking joy to know that, even \[unintelligible 00:40:27.06\] the best practices... But it's true that errors have a lot of information in them. + +**Ivan Kwiatkowski:** Is that something that, as a developer, you are not seeing? Do you see a lot of co-workers and the code you receive where the error-checking is bypassed and not used at all? + +**Natalie Pistunovich:** No, usually those things will not pass peer review... But I just don't know enough whether hackers do peer review. So that's -- that is interesting. + +**Ivan Kwiatkowski:** Well, this is what I've seen; it's going to be anecdotal at best. Of course, there are always going to be hackers out there that do things their own way, their own bad way, just like real world developers that work on other legitimate projects... And so I can only speak about the few malware programs that I have seen, and for which I can say that they looked pretty well developed as far as I was concerned, but there's bound to be somewhere out there that is going to be writing the most despicable Go code you can think of. + +**Natalie Pistunovich:** That, of course, makes sense. But still, I'm happy to hear that generally good practices are followed everywhere... But another thing that I wanted to ask about that is how does this represent in assembly, given that this is generally not a very common practice...? I guess because you don't see lots of errors, you basically don't see the representation of this? + +**Ivan Kwiatkowski:** Well, the way I see it is that when I look at the assembly code, most of the times, since I'm not a Go developer, I have to look up the functions that are being called. Sometimes the names are self-explanatory... But most of the times I have to go to the Go documentation... Which by the way, I think is extremely well done. Every time I look for functions documentation, I find it, which is always a good thing. There are languages where you try to find stuff, and you just don't; even just basic functions. But anyway... So I go look at the documentation, and then I get information about what arguments this function is supposed to receive, and what return values it's supposed to provide. And then when I do this looking up, then I get information about whether or not an error value is supposed to be returned by the function. And when that's the case, and I cannot recall of an instance where I was supposed to see an error value returned from the program, and that was not checked. + +So the way that you would see it in assembly would be like you have this function call, and then you see some random variable being taken back from the stack, and compared with value zero. So basically, if err=nil, and then you have a block, and whether or not the error is or is not nil, then you can go into that block and go into another one. But that block is here, which means that the attacker or the malware author went through the trouble of actually making sure that the error -- like, there was no error returned by the function. So this is the way that I observe it. + +**Natalie Pistunovich:** So earlier you mentioned something that you -- you compared kind of your work to getting code from a colleague, but it's all kind of no parameter names, no documentation, no function names, and so on... And that reminded me how sometimes you can use all sorts of AI tools like Codex and Copilot and whatnot to highlight that, and says, "Well, explain what it does." So did you ever have a chance to use one of those? + +**Ivan Kwiatkowski:** \[43:48\] I didn't. Now, I know that GitHub released this project. I personally have a very religious fear of such projects, just because I know that the way it works is that all the source code that I write gets uploaded into the cloud, and analyzed, and gets to feed the machine learning algorithm. And, it's kind of stupid, because all the code I write ends up being open source anyway... But I don't like that. + +**Natalie Pistunovich:** And if it's on GitHub, it goes to the same place. + +**Ivan Kwiatkowski:** Exactly. It also ends up there. So overall, there's not really a good reason for me not to do it. But I didn't try it yet. I've been told by some co-workers... I think they've used it for Python, and I've been told that it's amazing. It can pretty much guess what you're thinking, which is kind of scary. + +**Natalie Pistunovich:** When you write code, or when you reverse it? + +**Ivan Kwiatkowski:** Yeah, exactly. When you write code. I am not aware of a machine learning project that would help you reverse-engineer programs... Although I am 100% sure that this is possible. I've been playing a lot with the image generation AI, especially Midjourney. I tried the one that generates text which is called Lex... All those AIs, as far as I'm concerned, produce incredible results. If you had told me one year ago that I would be able to type some text and I would get the corresponding image generated, and that the image would actually look pretty amazing, I would really not have believed it, for real; I would have said that this is science fiction, and it's never going to happen in my lifetime. Or maybe when I'm old, and don't understand what's going on anymore. But we're there; we're there for many complex applications, such as understanding human language, and generating lots of contents... And at this stage, I would be extremely shocked if you told me that the recognizing functions that are actually generated by other computers is not something possible. This is 100% going to happen, eventually. I don't know who is going to do it. Maybe I should, actually. I don't know anything about \[unintelligible 00:45:39.20\] but this is an extremely worthy project, and I think that eventually, this is going to help us win so much time when we work on some unknown programs. we probably would have to have specialized AIs for different languages; we will need one for C, one for C++, one for Go etc. But I cannot imagine that this is not in our future, and probably in our near future, too. Hopefully, they won't sell this too expensive, because I want it. + +**Natalie Pistunovich:** So a follow-up question to that... For code generation, some languages are better than other; for example, Go is performing even better than Python and such, just because NGO has this built-in linter, and there's many things that it's not either/or, but it's definitely tabs. It's definitely curly brackets and a new line. So the AI has a more consistent dataset to be trained on... Versus Python, and many other languages that you can write in whatever way... So it just sees lots of different examples, and it might, in the best case, generate inconsistent code of, you know, one file is different from another, but even just sometimes wrong, following two different paradigms in one file. Do you think that for the reverse part of it, will this benefit kind of keep rolling, or not? For the AI perspective of it. + +**Ivan Kwiatkowski:** Yes and no. So what you're referring to is the fact that the more strict languages are, then the easier for an AI it is to understand what this meant and to generate code. And I think it makes sense, because Python, for instance, as a language, is extremely ambiguous, right? Not as ambiguous as JavaScript, or maybe as PHP, which as far as I'm concerned, they don't qualify as actual languages, because they're not rigorous at all... But when you have stuff like duck typing in a variable that you don't have to type at all - well, of course, for an AI to figure out what is going on is going to require much more effort, and it's going to probably not be possible all the time... Because I think that in many cases, a lot of stuff gets determined at runtime, which is not something that an AI is probably going to be able to do, at least not that soon. + +On the other end, when you have some assembly code, then assembly code is going to be this very strict, unique language that probably all AIs will have to work on. And I'm not exactly sure how they are going to work their way back up to either recognizing a function, or actually generating corresponding high-level code... + +\[48:08\] But the good thing that AIs will have going for them is that assembly is going to be like the exact opposite of ambiguous, right? Like, you have ambiguous, and at the exact other end of what's possible you have assembly, which is 100% precise... + +**Natalie Pistunovich:** As consistent as it gets. + +**Ivan Kwiatkowski:** Yeah, consistent, and actually done in some ways, but it's just very simple operations that can only do a single thing in a very defined way. So on that front, I think that this is actually going to be a very, very good thing for the AIs, whenever they are ready. + +**Natalie Pistunovich:** So there's already good tools out there, that just take binary and translate that into assembly; not 100%, but a very good coverage. And assembly is consistent enough, so that means that some IDA Pro plugin that uses AI must be developed as we speak, to say "Here's assembly input. Please translate that to Go code for me." + +**Ivan Kwiatkowski:** Well, it's a good question, because you would think that someone would be working on this... But when you look at the market for reverse engineering tools, it's actually quite small. You have Hex Rays, the creators of IDA Pro, and Hex Rays are kind of an old-school company. Their product is amazing, but they haven't really tried to create any form of disruption in the past 20 years. Now, they have been doing lots of improvements to their products, but I think part of it is only because they have been challenged by Ghidra, their open source competitor. It's kind of my opinion that -- but let's make this my unpopular opinion, if you want to... But it's kind of my opinion that if Ghidra had never appeared, then IDA Pro would basically have stayed kind of the same for the next decade or so, because the developers had kind of no incentive to make it significantly better, because just they had no competition there. + +The way that the compiler is working at the moment, as far as I understand, is purely through algorithmic means, and they do not use any form of machine learning. There is no AI applied to their decompilation process. Maybe they have started working on this, but as a company - and maybe I'm totally mistaken about this, I don't work there, but Hex Ray doesn't strike me as a company that could be doing groundbreaking R&D, that would maybe prepare for the next generation of decompilers. I think they would rather make incremental changes on their existing product to make it slightly better year after year. + +I think Ghidra, which has a decompiler as well, is open source, and I think this decompiler is doing pretty well, too... But I don't think it uses AI in any shape or form either, and I'm not aware of any plans to like to start working on this. Probably developing some AI product focused on reverse engineering would require some very specific AI knowledge. And my feeling is -- I don't know everyone in the security field, but I tend to... Like, my perception of this is that we have people that are extremely focused and extremely talented and skilled in the specific field of cybersecurity that they are working on... But I cannot recall anyone that I've met that was both a great researcher, or a great reverse engineer, a great pen tester, and a very qualified data scientist as well, or someone that would be able to tune a machine learning algorithm that would change our lives forever. Maybe someone is working on this somewhere, but if it's the case, then I'm not aware of it. + +I think if it were public, or if it was out there, I think I would know about it; maybe I don't. But overall, if there was this big project about to be released, I want to believe I would have heard about it. But I do still hope that some company somewhere, probably in Israel, is working on this in a secret lab, and eventually, they're going to take the market by storm, and make my job a lot easier. + +**Natalie Pistunovich:** \[51:59\] Yeah, another idea for the followers who are tuning in. So in addition to an AI that translates from assembly to Go, it would be personally interesting for me to have some AI that says "This malware is written in the style of..." And what I mean by that is that already now you can write - if you go to Codex and other tools like that, you can say, "Write a Go program that does this, and that, and this, and do it in the style of..." And if you mention a GitHub handle of somebody who is a known developer, has lots of stars, or some other big presence, and their type of code has a flavor - which is maybe less common in Go, because it's so structured - you will get their style of code. So eventually, a next step to this magical plugin would be "In whose style is this malware written?" This will be also interesting. And then all those language teachers out there will know "Oh, well, I taught that hacker." + +**Ivan Kwiatkowski:** Yeah, that's a very nice idea. And actually, there have been, maybe not rumors, but open research projects working exactly on this for probably decades in the \[unintelligible 00:53:08.15\] intelligence field. I think I recall, a CCC presentation from maybe 2010 or 2012, one of those years... And basically, people were already working on obtaining code from open source repositories - they basically downloading GitHub, and trying to extract maybe some penmanship characteristics from every single developer, and they were hoping that they would be able to take any program in the future and be able to tell you that "This might probably be the developer that created this executable." + +Now, it's been 10 years, 15 years... I'm not sure -- I haven't heard about this for a while, so maybe this either didn't work as well as they expected, or this was actually absorbed into some intelligence service somewhere... Because there are very obvious intelligence applications there, that I think would -- like, those types of services would love to have such a capability, because they would be able to identify malware authors; they wouldn't need really to have the burden of proof the way that police forces have... They would just be able to know who the guy behind some malware is, and just do their usual parallel construction stuff. + +So this is something we know they want, and this is something that we also know they spent money on, and I remember that some universities were actually working on those types of projects... For research, not for intelligence, but those worlds, they tend to communicate with each other anyway, when there are applications. + +If this came to fruition, then this did so in secret. The way that they used to be working on this was, again, algorithmic; they were trying to extract the characteristics, and they were not using those blackbox AI capabilities that we have now. Maybe this is a new avenue of research for those applications. Maybe we'll know in the future. But as far as I know, this is an existing problem that people are trying to solve, and I haven't seen any signs that they have, although it's not exactly sure that I would. + +**Natalie Pistunovich:** This episode took very interesting turns, like new ideas for tools and projects and whatnot, and remember where you heard this first. + +**Ivan Kwiatkowski:** Yeah, I was not expecting that. + +**Natalie Pistunovich:** Yeah, but that's very cool. That's really inspiring to speak with you, Ivan. Thank you for this hour of the conversation. + +**Ivan Kwiatkowski:** Of course. + +**Natalie Pistunovich:** As the previous episode, it also ended with me having lots of open questions... Please consider joining me again next year for one more episode, or ten more. \[laughs\] + +**Ivan Kwiatkowski:** Yeah, of course. Eventually - and I think it will be pretty soon - we're going to expand the whole knowledge that I have on the Go language, right? If I have to come back, then at this point I will really have to look into the language more, and maybe try to come with actual research that I can share with you... Because otherwise -- I don't want this conversation to be boring, but I'll do my best. + +**Natalie Pistunovich:** Well, next time we'll talk about generics maybe... See how that Go proposal goes. + +**Ivan Kwiatkowski:** And I'll work on that. + +**Natalie Pistunovich:** Yeah. And instead of unpopular opinion, we have provided two unicorn ideas... So you're welcome, everyone. \[laughs\] + +**Ivan Kwiatkowski:** I can pretty much give you my personal guarantee that if anyone in the audience actually implements one of those two things, that they're going to be extremely rich. So there you go. + +**Natalie Pistunovich:** That's almost as good as being on the rank of the most unpopular opinion. \[laughs\] Thanks, everyone, for joining. Thank you very much, Ivan. + +**Ivan Kwiatkowski:** Thank you. I was happy being here, it was a pleasure speaking to you, and see you next time. diff --git "a/Hacking with Go\357\274\232 Part 2_transcript.txt" "b/Hacking with Go\357\274\232 Part 2_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..cd84ea828e51a3ae788c2a2be6079c8bc3819991 --- /dev/null +++ "b/Hacking with Go\357\274\232 Part 2_transcript.txt" @@ -0,0 +1,835 @@ +[0.00 --> 4.44] I think one of the main appeals of Go is that you don't really need to think about security +[4.44 --> 6.36] as much as with other languages. +[6.62 --> 12.00] Go is a memory-safe language, unless I'm mistaken, and the compiler is never going to let you +[12.00 --> 16.68] do stupid stuff like create an array that is too small and then write stuff that goes +[16.68 --> 17.08] out of it. +[17.28 --> 18.12] It's just not possible. +[18.12 --> 22.98] So it eliminates a whole lot of bug classes, which we call memory corruptions. +[23.10 --> 24.12] It's just not going to happen. +[24.22 --> 25.82] You cannot do this to yourself in Go. +[25.82 --> 31.68] And it means that all the old school buffer overflows that plagued all the C and C++ programs +[31.68 --> 35.66] for dozens of years are not going to ever happen in the Go language. +[35.76 --> 40.76] It doesn't mean that the program is going to be perfectly safe from any security issues, +[40.76 --> 45.86] but the issues are not going to be related to, oh, I made a programming mistake and there +[45.86 --> 46.70] is a bug in my program. +[46.78 --> 47.46] It's going to be exploited. +[47.60 --> 50.38] It's going to be more related to design issues. +[50.38 --> 55.96] What's up, friends? +[56.06 --> 58.40] This episode is brought to you by Sourcegraph. +[58.84 --> 64.32] With the release of Sourcegraph 4.0 and the Starship event just a few weeks behind us, +[64.52 --> 70.12] it is super clear that Sourcegraph is becoming not just CodeSearch, but a full-on code intelligence +[70.12 --> 70.72] platform. +[71.12 --> 74.10] And I'm here with Joel Kortler, product manager of Code Insights for Sourcegraph. +[74.58 --> 79.20] Joel, this move from CodeSearch to Code Intelligence is a really big deal. +[79.20 --> 83.92] How would you explain this feature, Code Insights, if you're just talking to folks in the hallway +[83.92 --> 85.56] track of your favorite conference? +[86.18 --> 89.74] I would really start with technical because before I was a product manager, I used to be +[89.74 --> 90.40] an engineer as well. +[90.68 --> 94.42] And it's really cool and exciting just to be able to say, we're going to turn your code +[94.42 --> 95.56] base into a database. +[96.06 --> 100.12] And the structured language that you need to interact is just the ability to write a code +[100.12 --> 101.80] search, you know, literal search. +[101.94 --> 102.66] That's totally fine. +[102.80 --> 106.08] Regular expression, you know, that'll give you a few more advanced options, even a structural +[106.08 --> 106.44] search. +[106.90 --> 111.46] But the number of long tail possibilities that unlocks, truly the journey of building +[111.46 --> 115.56] this product was just saying, well, we've just unlocked, you know, an infinite number +[115.56 --> 116.18] of possibilities. +[116.62 --> 120.24] We got to figure out some immediate use cases so we can start to, you know, invest in this +[120.24 --> 121.28] product, build it and sell it. +[121.74 --> 125.52] But we're only getting started in terms of the number of uses that we're uncovering for +[125.52 --> 125.70] it. +[126.04 --> 130.04] The story I told you about discovering like version tracking turned out to be a really important +[130.04 --> 133.62] use case that wasn't even on our roadmap six months prior to discovering that as we +[133.62 --> 136.44] are already planning to launch this product until we talked to enough folks, realized +[136.44 --> 139.84] this was a problem and then found, well, oh, that's like a simple regular expression +[139.84 --> 143.34] capture group that you can just plug right in because we really built this system to +[143.34 --> 145.08] not limit the power of what we built. +[145.18 --> 148.06] We don't want to give you like three out of the box templates and you can only change +[148.06 --> 149.18] like one character or something. +[149.28 --> 152.18] It's truly like the templates are there to hold your hand and get you started. +[152.30 --> 155.76] But if you can come up with anything you want to track in your code base, you can do +[155.76 --> 156.46] that with Code Insights. +[156.70 --> 157.36] I love it. +[157.40 --> 157.86] Thank you, Joel. +[157.86 --> 163.68] So right now there is a treasure trove of insights just waiting for you living inside +[163.68 --> 164.30] your code base. +[164.46 --> 167.32] Your code base is now a queryable database. +[167.38 --> 168.34] Thanks to Sourcegraph. +[168.86 --> 172.68] This opens up a world of possibilities for your code and the intelligence you can gain from +[172.68 --> 172.92] it. +[173.18 --> 178.54] A good next step is to go to about.sourcegraph.com slash code dash insights. +[178.72 --> 180.02] The link will be in the show notes. +[180.36 --> 182.86] See how the teams are using this awesome feature. +[182.86 --> 187.94] Again, about.sourcegraph.com slash code dash insights. +[188.30 --> 190.58] Again, this link is in the show notes. +[205.74 --> 206.58] Let's do it. +[207.28 --> 208.22] It's go time. +[208.92 --> 210.36] Welcome to go time. +[210.36 --> 213.58] Your source for diverse discussions from all around the Go community. +[214.06 --> 216.70] Check out our back catalog at go time.fm. +[217.02 --> 221.66] There you'll find the most popular episodes, our favorites, and a request form so you can +[221.66 --> 223.90] let us know what you want to hear about on the pod. +[224.26 --> 228.80] Special thanks to our partners at Fastly for shipping our shows super fast to wherever you +[228.80 --> 229.04] listen. +[229.30 --> 230.78] Check them out at fastly.com. +[231.22 --> 232.86] And to our friends at fly.io. +[233.22 --> 235.20] Host your app servers close to your users. +[235.48 --> 236.32] No ops required. +[236.64 --> 238.44] Learn more at fly.io. +[238.44 --> 239.46] Okay, here we go. +[240.36 --> 245.02] Hello, everyone who is joining us today on a Wednesday of the recording. +[245.36 --> 249.18] We normally record on a Tuesday, but we have a very special guest, so we need to make a +[249.18 --> 250.76] very special event about that. +[251.38 --> 252.40] Ian is my co-host today. +[252.48 --> 252.82] Hi, Ian. +[253.14 --> 253.88] Hey, how you doing, Natalie? +[254.22 --> 254.60] Good. +[254.76 --> 257.80] I'm very excited to have Ivan today join us. +[257.92 --> 258.16] Yeah. +[258.32 --> 259.48] Ivan Kwiatkowski. +[260.02 --> 262.10] Also known on Twitter as Justice Rage. +[262.24 --> 266.10] You are a senior security researcher at Kasparsky. +[266.10 --> 266.58] Yes. +[266.76 --> 267.00] Hello. +[267.14 --> 268.16] Very happy to be here. +[268.54 --> 268.80] Indeed. +[268.96 --> 274.78] So I work in the threat intelligence field and my daily work involves looking at malware +[274.78 --> 276.36] and writing reports about it. +[276.48 --> 281.24] Basically, the activity that I'm involved in is trying to figure out what the attackers +[281.24 --> 285.04] are up to, what kind of tools they're using, methodologies, what types of victims they are +[285.04 --> 287.10] after, and then we write stuff about it. +[287.66 --> 291.76] And our customers read our reports, and then it allows them to figure out whether or not +[291.76 --> 296.58] this group or this group is likely to attack them or not, depending on what type of information +[296.58 --> 297.24] that they are after. +[297.46 --> 302.36] And if so, how they may defend from those attacks by knowing more about the type of malware +[302.36 --> 306.98] that they use, the type of attack vectors that they typically favor, and so on. +[306.98 --> 313.98] So really, I spend my day in AIDA Pro most of the time, and sometimes as well, I do give +[313.98 --> 319.18] out trainings for reverse engineering, either in universities or for our customers as well. +[319.48 --> 324.82] And there is a very cool video that has two parts of you reverse engineering and malware +[324.82 --> 328.12] written about a year ago that was written in Go, actually. +[328.40 --> 328.74] Absolutely. +[329.10 --> 330.88] And that was from the SolarWinds attack. +[331.18 --> 331.60] Exactly. +[331.60 --> 337.18] This specific example comes from the SolarWinds incident, which I'm pretty sure that most +[337.18 --> 341.90] listeners will be aware of because it was such a high media impact case. +[342.20 --> 346.64] To make a quick summary about it, what happened was a company called... +[346.64 --> 348.08] I always get dogs mixed up. +[348.16 --> 351.74] I think the name of the company is SolarWinds, and then the product is Orient IT, but maybe +[351.74 --> 352.46] the other way around. +[352.78 --> 355.06] I do really get confused about this all the time. +[355.18 --> 356.58] I think the way you have it is right. +[356.82 --> 357.34] Okay, great. +[357.50 --> 359.32] That was really a 50-50 chance there. +[359.32 --> 364.80] Anyway, this company got attacked, but it wasn't attacked for the information that it +[364.80 --> 370.52] had because it was just a software company, which in itself had little value as an intelligence +[370.52 --> 370.94] target. +[371.38 --> 377.02] But the thing was that it had a high number of high-profile customers, and these customers +[377.02 --> 382.18] were US government entities or big companies in the field. +[382.18 --> 388.86] And what the attackers did was they were able to compromise the software build chain, and +[388.86 --> 395.00] they were able to insert their own code inside of the software that was then pushed to customers. +[395.38 --> 399.82] And using this, they were able to create a backdoor that would be automatically deployed +[399.82 --> 401.42] at all SolarWinds customers. +[401.42 --> 407.84] And then, you know, maybe two weeks or three weeks later, because this very stealthy attack +[407.84 --> 412.18] had a very long sleeping time, it stayed dormant for a while to make sure it would remain very +[412.18 --> 412.60] stealthy. +[412.74 --> 416.02] But after a while, then it would start connecting to the C2 server. +[416.24 --> 420.66] And then for all the targets that were deemed interesting by the attacker, they would receive +[420.66 --> 424.18] a second stage payload that would allow them to get it to the network and then collect +[424.18 --> 425.14] intelligence and whatnot. +[425.14 --> 431.82] So the very first stage of the attack was just some modification of the code of the +[431.82 --> 432.52] original program. +[432.82 --> 434.02] This part was written in .NET. +[434.32 --> 439.24] But then the second part, which is called SunShuttle, was actually written in Go language. +[439.76 --> 444.28] So it was, for me, like the first time I was getting involved in reverse engineering for +[444.28 --> 444.96] the Go language. +[445.34 --> 446.72] The learning curve was a little steep. +[447.36 --> 451.40] But then again, I kind of used this as a learning experience, but also as an example +[451.40 --> 456.18] in future reverse engineering courses for other people that might be interested in learning +[456.18 --> 459.30] how to, well, how to reverse engineer Go programs. +[459.42 --> 463.80] But also, I think if you are a Go enthusiast, reverse engineering can allow you to get to +[463.80 --> 468.36] know more about how the language actually works under the hood, which I think is also very +[468.36 --> 470.98] interesting from a software development point of view. +[471.42 --> 473.90] So that's one like famous example of Go malware. +[474.24 --> 477.66] Are there other famous ones written in Go that you can think of off the top of your head? +[477.92 --> 478.08] Yeah. +[478.08 --> 483.38] So from the same incident, one of the companies that was breached through the Solwinds incident +[483.38 --> 485.94] was Mendiant, now belongs to Google. +[486.32 --> 489.90] And they were the ones that actually detected that there was something wrong in their network +[489.90 --> 490.60] and reported it. +[490.72 --> 494.00] And so kudos to them, really a great job on figuring out that something was wrong. +[494.64 --> 499.52] But one of the things that the attackers were very interested in was getting access to the +[499.52 --> 504.90] tool set that Mendiant was using for their own penetration testing and red teaming engagements. +[504.90 --> 509.90] And it so happens that the tools that they were using were actually written in Go language, +[510.06 --> 513.76] which I think is really interesting from an analyst perspective. +[514.04 --> 519.20] So I think there's an interesting discussion to have about why they chose this language for +[519.20 --> 520.72] their own offensive tools. +[521.20 --> 527.40] There are a number of other projects on GitHub, which I can probably think of one called Stowaway +[527.40 --> 532.02] on the top of my head, which has been also reused and modified by some threat actors. +[532.36 --> 534.32] We'll add a link to that in the show notes. +[534.44 --> 535.14] That sounds interesting. +[535.34 --> 536.06] Yeah, sure. +[536.44 --> 537.38] It's a networking tool. +[537.50 --> 542.04] It's really something that proxies and stuff in and out of a network that goes between protocols +[542.04 --> 542.82] and that kind of stuff. +[542.90 --> 543.86] It's written in Go language. +[544.02 --> 548.74] Pretty annoying to reverse engineer because it's a lot of Go routines talking to each other. +[548.74 --> 551.40] Very hard to figure out how it's architecture. +[551.40 --> 557.46] And another example I can think about is I'm not 100% sure, but I do believe that a commercial +[557.46 --> 564.72] backdoor called BrutRatel, which is a big competitor or a new competitor maybe to CobaltStrike, +[565.12 --> 572.68] which places enormous emphasis on evading detection and being able to slip through EDR solutions, +[572.82 --> 575.70] et cetera, is also written in Go language, I do believe. +[575.86 --> 577.50] But I would have to double check that. +[577.50 --> 581.74] So these are examples of malware families written in Go language. +[581.84 --> 584.30] And I think that over time, we're going to see more and more of them. +[584.74 --> 586.04] Why do you think we're going to see more and more? +[586.28 --> 587.58] Is there a specific reason? +[587.64 --> 589.42] You mentioned that they were hard to reverse engineer. +[589.56 --> 591.14] Is that part of it or all of it? +[591.36 --> 593.54] Yeah, there are a few reasons. +[593.92 --> 600.94] The first reason I think is probably related to the ease of use for the developers. +[601.06 --> 605.68] I don't mean that Go is easier to program than other languages, but the fact that it generates +[605.68 --> 610.74] statically built executables, binaries that are self-contained, that do not need any additional +[610.74 --> 613.64] libraries, is kind of very comfortable for attackers. +[614.08 --> 617.64] Like, you know, they create their backdoor, they send it to the victim, or, you know, they +[617.64 --> 620.52] deploy it at the victim one way or the other, and then it just works, right? +[620.56 --> 623.88] You don't have to think about, you know, is this DLL present on the system, or do I have +[623.88 --> 626.22] to pull in additional libraries, et cetera? +[626.22 --> 631.76] So this is something that makes running programs very easy on victim machines where you do not +[631.76 --> 632.58] control the environment. +[632.58 --> 636.94] A long time ago, like maybe 10 years ago, it was kind of a problem because you cannot +[636.94 --> 639.88] send binaries that are two or three megabytes big to victims. +[640.04 --> 644.66] You know, if your attack vector is, you know, infected PDF or infected Word documents, then, +[644.76 --> 649.14] you know, you cannot really send over email a PDF that ends up being five megabytes big +[649.14 --> 651.58] because, you know, back in the day it would be rejected. +[651.82 --> 656.46] Or maybe, you know, the victim, you know, has some limit on their mailbox, or maybe they +[656.46 --> 660.06] have a slow connection that is not going to be able to retrieve that binary. +[660.06 --> 663.78] In Europe or in the US, in the Western world, it used to be fine. +[663.88 --> 667.78] But, you know, if you think about victims that are in third world countries where the +[667.78 --> 671.50] internet access is not as good, then it used to be some real, a real issue for attackers. +[671.86 --> 676.24] Now that, you know, the internet connectivity is pretty much, well, at least way better in +[676.24 --> 680.70] most parts of the world than having backdoors that are, you know, five or 10, maybe 20 megabytes +[680.70 --> 682.80] is really not that much of an issue anymore, I think. +[682.80 --> 690.18] Then, second very good reason for using Go as an offensive language is going to be that +[690.18 --> 694.96] reverse engineering is difficult, which I will get back to, but also all these standard +[694.96 --> 700.32] tools that we, as defenders, tend to use in order to figure out quickly if a program is +[700.32 --> 703.28] malicious or is not, tend to kind of break with Go language. +[703.66 --> 708.10] The reason for this, and it ties into the discussion of why reverse engineering Go is annoying for +[708.10 --> 713.16] us is that Go tends to really do its own thing, like the assembly it generates really does not +[713.16 --> 714.78] look like any other assembly. +[714.90 --> 719.96] It's not like C or it's not like C++ or Delphi that kind of tend to like look like distant +[719.96 --> 721.78] cousins or even brothers in some cases. +[722.52 --> 729.64] Go really does things its own way and all the automated methods for analyzing code statically +[729.64 --> 735.92] or all the maybe signatures you can recreate for Go language, et cetera, but all the tools +[735.92 --> 740.80] that would try to recognize specific patterns in code are not going to work because the code +[740.80 --> 743.52] generated by Go just looks like nothing you've seen before. +[743.52 --> 745.12] So that's one reason. +[745.12 --> 747.68] And the final reason is reverse engineering. +[747.68 --> 754.40] It's really difficult for us because the constructs that are generated by the Go compiler tend to be +[754.40 --> 755.76] very unfamiliar to us. +[755.76 --> 758.56] And so the learning curve, I would not say it's that steep. +[758.56 --> 762.72] Like you mentioned, Natalie, that I had released a few videos about it. +[762.72 --> 766.56] I think, you know, by the end of the videos, you can have like a rough idea of how to approach those +[766.56 --> 766.96] programs. +[766.96 --> 771.28] So it's not that like, it's not like an obstacle that is insurmountable. +[771.28 --> 773.36] It's something that eventually you will be able to figure out. +[773.36 --> 780.80] But when you've been working on C or like similar looking code as C for 10 years, and +[780.80 --> 786.16] sometimes learning something new is not something that you are easily going to do because you are, +[786.16 --> 790.32] you have your comfort zone and then you have to like discover something different. +[790.32 --> 793.52] And maybe you don't like to do this and maybe you have, you know, 10 easy malware +[793.52 --> 795.60] return C that are waiting in the, in the test list. +[795.60 --> 799.52] And maybe you're going to work on those first because it will allow you to end your day earlier +[799.52 --> 800.08] next Friday. +[800.08 --> 800.40] Right. +[800.40 --> 805.12] So you, you kind of mentioned that there's assembly differences that make it hard to recognize. +[805.12 --> 808.88] Are there any like specific things that you've learned about Go under the hood from that, +[808.88 --> 815.20] you know, like that differ from C like how functions are called in the assembly or something like that? +[815.20 --> 816.56] Yeah, absolutely. +[816.56 --> 820.72] So one of the major differences, it's not really about the assembly itself. +[820.72 --> 825.28] It's about, you know, the static aspect of the executables is the fact that all the +[825.84 --> 828.32] functions are pulled inside the final binary. +[828.32 --> 832.72] And then you have this big program that's two megabytes or three megabytes big just for a +[832.72 --> 833.04] print. +[833.04 --> 834.08] Hello world. +[834.08 --> 836.16] And now it's getting a bit better. +[836.16 --> 841.76] I think IDA pro has made significant improvements in its later versions, but maybe two to three years +[841.76 --> 845.20] ago when you were opening a Go program, you would have nothing recognized at all. +[845.20 --> 851.04] Maybe you would be able to pull a few plugins here and there or Python scripts that may or may not +[851.04 --> 851.28] work. +[851.92 --> 856.16] And in that case, if you were lucky, you might have been able to create signatures for the +[856.16 --> 859.92] well-known functions and maybe start from there, but it was really a huge ordeal. +[859.92 --> 861.28] Now it's a bit better. +[861.28 --> 867.36] So at least you are starting to get pretty reliably all the references to all the known functions. +[867.36 --> 871.52] Beyond this, the calling convention is, well, I'm not going to say it's weird because like, +[871.52 --> 873.52] it's, I mean, it's as valid as any other one. +[873.52 --> 875.76] It's just not the same one that we are used to seeing. +[876.48 --> 881.36] The main difference is that considering that Go can return multiple return values, +[881.92 --> 886.72] then you cannot, you know, have the same system as we had to, as we had before. +[886.72 --> 890.32] Like for instance, in the C program, the return value goes into EAX and that's it, right? +[890.32 --> 891.28] No difference. +[891.28 --> 893.76] I mean, the AX register of your CPU. +[893.76 --> 897.84] When it comes to Go language, if you have three, four, or, you know, maybe more return values, +[897.84 --> 902.24] you know, typically one return value and also some error objects, if I'm not mistaken, +[903.12 --> 906.48] then you cannot put all that into a single CPU register. +[906.48 --> 907.36] It just doesn't work. +[907.36 --> 912.40] And so you tend to get values that, well, in the past, you would have all the arguments being +[912.40 --> 917.84] passed through the stack, not through pushes, but direct moves from the, you know, from the +[917.84 --> 919.68] value into the stack directly. +[919.68 --> 924.40] So the instruction was not push, which are these assemblers, they are, you know, automated analysis +[924.40 --> 924.64] tools. +[924.64 --> 926.88] They just like to see push, push, push, and then call that. +[926.88 --> 929.60] That's something that is easy for them to recognize. +[929.60 --> 933.76] But Go would just do move this on the stack at this place, move this on the stack at this place, +[933.76 --> 935.04] and then you go into another function. +[935.04 --> 937.84] It knows because the compiler knows where the stuff ends up. +[937.84 --> 938.64] So it figures it out. +[938.64 --> 941.44] But you know, the Ida Pro looks at this and is like, what the hell is this? +[941.44 --> 943.36] This memory has never been initialized before. +[943.36 --> 944.24] I cannot show this to you. +[944.88 --> 945.76] That was an issue. +[945.76 --> 948.40] And then the return values were given back exactly the same. +[948.40 --> 952.64] So the program would just move back all the return values onto the stack as well, +[952.64 --> 955.36] at places that it would be able to figure out later. +[955.36 --> 958.96] But then when you look at Ida Pro, then, you know, it sees, okay, values being moved on the +[958.96 --> 959.20] stack. +[959.20 --> 963.68] You go back into the calling function, and then you see references to the stack as well. +[963.68 --> 966.96] But you know, the offsets are going to be different because since you are returning from a function, +[966.96 --> 968.88] you know, things have shifted a little bit. +[968.88 --> 971.44] And so the offsets are not well, do not work well anymore. +[971.44 --> 975.68] And so this is like another issue that you have to face, like figuring out where your return values go. +[975.68 --> 978.08] It still is, by the way, a terrible nightmare. +[978.80 --> 983.12] And finally, there is this other key difference. +[983.12 --> 989.04] And this difference is the fact that usually the C compiler and other similar compilers will tend to +[990.00 --> 994.24] reserve some space on the stack for specific local variables. +[994.24 --> 996.48] And this tends to be very reliable. +[996.48 --> 997.84] It doesn't move too much. +[997.84 --> 1001.52] So when you have some variable in C, you know, it gets used some part of the program. +[1001.52 --> 1003.28] It's at one place on the stack, and then that's it. +[1003.28 --> 1007.68] And if the program needs another local variable later on, then there's just another space located +[1007.68 --> 1008.80] for this in the stack. +[1008.80 --> 1011.52] And the Go compiler tends to be very smart about these things. +[1012.08 --> 1016.08] And what it does is if it sees that, you know, there used to be a variable at some place on the +[1016.08 --> 1021.04] stack and it's not used anymore, then it will feel like it's totally okay to reuse the same space +[1021.04 --> 1024.16] to store something else later, which makes total sense. +[1024.16 --> 1026.24] I mean, do not use more memory than you need to, right? +[1026.24 --> 1028.72] But the Go compiler is totally right in doing this. +[1028.72 --> 1034.00] But for me, it's really, really a problem because what I do in IDA Pro is I try to figure out where +[1034.00 --> 1035.84] the local variables are in the stack. +[1035.84 --> 1039.44] I name those positions by saying, okay, this is the error variable. +[1039.44 --> 1044.72] This is the, I don't know, this is the integer that represents an iteration count or whatever. +[1044.72 --> 1047.36] And I name or rename everything I can. +[1047.36 --> 1052.48] And then eventually stuff starts to make sense because I know what represents what on the stack +[1052.48 --> 1054.88] and I know what the variables are, et cetera. +[1054.88 --> 1060.08] But the thing is, if one position on the stack does not consistently represent a specific variable, +[1060.08 --> 1062.00] then I cannot rename things anymore, right? +[1062.00 --> 1063.60] There's just no way for me to do this. +[1063.60 --> 1068.08] And the tools that we have, such as IDA and Pritcher Ghidra is going to function the same way. +[1068.08 --> 1072.88] It's not going to allow me to say, okay, up to this point, this variable should be named like this. +[1072.88 --> 1077.92] And then from there on, then it should have another name and then yet another, et cetera. +[1077.92 --> 1082.88] So this is like a very, very difficult thing for us is that, you know, trying to track down variables +[1082.88 --> 1086.64] and return values, even arguments is something extremely complex. +[1086.64 --> 1090.32] And basically this is the normal flow of how you analyze a program. +[1090.32 --> 1094.16] You try to figure out what the variables are, try to look at the functions and how they are called, +[1094.16 --> 1095.60] what they return and that kind of stuff. +[1095.60 --> 1100.08] And just, you know, doing those simple things that would be the basic operations and building blocks +[1100.08 --> 1105.12] of trying to understand what is going on in some random program are in themselves, +[1105.12 --> 1110.88] extremely complex operations due to like optimizations that were performed by the Go compiler. +[1110.88 --> 1116.00] Now, the last thing I can mention is that since version probably 16.1 or something like this, +[1116.00 --> 1120.08] or 1.16, I guess, in Go, the calling convention actually changed. +[1120.56 --> 1125.36] And they do things even smarter now, which is pass some arguments through the registers and not through +[1125.36 --> 1129.36] the stack. So for me, it doesn't change that much. Actually, it makes things a little bit easier +[1129.36 --> 1133.44] because at least, you know, I know argument one is in ECX, argument two is in EDX from memory. +[1133.44 --> 1137.52] It might not be that one, but, you know, generally it's going to be in a fixed register, +[1137.52 --> 1141.04] at least for the two first arguments. And so I know where they are, but that's way better. +[1141.60 --> 1146.00] But, you know, overall, this doesn't change this bigger game of renaming things, which is not +[1146.00 --> 1151.04] possible anymore. And then when it comes to the quick and easy mode, which is, you know, getting my +[1151.04 --> 1156.08] super expensive IDEP Pro license that comes with a decompiler, then, you know, I just open the program, +[1156.08 --> 1160.24] press F5 and hopefully you can read whatever's going on in the program. Well, that just doesn't +[1160.24 --> 1164.08] work because, you know, the constructs that are generated by the Go compiler, especially, +[1164.08 --> 1169.60] I think when it comes to function calls, is totally alien to IDA. And, you know, every time you try to +[1169.60 --> 1176.16] decompile code that comes from the Go language, you just end up with something that makes absolutely +[1176.16 --> 1183.12] no sense. Because again, IDA tries to recreate pseudo C code and pseudo C code has just no way of +[1183.12 --> 1190.00] representing concepts like multiple return values or that kind of stuff. So this is the way that Go +[1190.00 --> 1193.44] breaks everything that we hold dear in the reverse engineering world. +[1193.44 --> 1198.56] For anybody who didn't watch the video or is not familiar with how to do reverse engineering, I can, +[1199.28 --> 1204.96] in simple words, say that roughly you look at the instructions and then you try to kind of see +[1204.96 --> 1210.56] entry point is usually main. So this is probably function main. This is one thing that's being returned. +[1210.56 --> 1216.08] And then you kind of try to follow that. Basically, this is what you do when you reverse engineer. +[1216.08 --> 1221.12] Yeah. Actually, maybe I can say a few words about what reverse engineering is for people that, you +[1221.12 --> 1226.24] know, might not be familiar with it. The general idea is that we try to understand what a program +[1226.24 --> 1230.64] does, even though we do not have access to the source code. But this is the typical case for malware, +[1230.64 --> 1234.64] because, you know, we cannot call up malware authors and tell them, okay, please show me the code, +[1234.64 --> 1238.16] because I don't really understand what's going on there. They just, we don't know where they are. +[1238.16 --> 1242.64] They don't want to be found and they don't want to give us their code anyway. So what we have to do +[1242.64 --> 1248.16] then is like, we have no other solution, but to look at the program and see what instructions the +[1248.16 --> 1254.08] program is sending to the CPU and then try to figure out from there, based on those instructions +[1254.08 --> 1260.00] that are working at the CPU level, what the higher level line of code that might have generated this type +[1260.00 --> 1265.04] of instruction might have been. So it's not really, it's not entirely a guessing game because it's sort of a +[1265.04 --> 1271.20] mostly exact science, but also it's a very unnatural operation to perform because this +[1271.20 --> 1277.68] CPU language was really made for CPUs and machines. And for us humans, like it's extremely difficult to +[1277.68 --> 1282.32] understand. Like it's really, it's really not something natural for human beings to read that, +[1282.32 --> 1286.96] those instructions. It doesn't make sense to us. And it really requires a lot of effort to figure out +[1286.96 --> 1292.16] what the programmer intent was just by looking at those instructions. And so this is why actually we are +[1292.16 --> 1296.72] looking for reverse engineers. I mean, not just at Kaspersky, just the whole industry is looking for +[1296.72 --> 1301.60] people that are able to do this because it's something that is, that most people find unpleasant. +[1301.60 --> 1305.68] And I have to say myself, I do find it unpleasant most of the times, but you know, at the end of the +[1305.68 --> 1309.84] day, when I am able to figure out what was actually happening in the program, I feel very good about +[1309.84 --> 1316.32] myself. And so this is the reason why I still do this job, but overall, this is kind of a difficult +[1316.32 --> 1321.52] thing to do and it's kind of painful and takes a lot of time to be able to figure out even the +[1321.52 --> 1324.88] simplest programs. Especially when the tooling is not even there for you. +[1324.88 --> 1331.68] Yes. Just for some reference, like the ratio between lines of say, Go to assembly, +[1331.68 --> 1335.84] do you know what that ratio is? Just like a rough, it's like one to a hundred, one to a thousand, one to... +[1335.84 --> 1339.04] It's a good question. It would depend on the complexity of the line. You know, in Go, I'm +[1339.04 --> 1343.44] pretty sure that you can do function calls that are chained together in a pretty like in long lines. +[1343.44 --> 1348.24] Maybe I'm not sure if it's like the compliant to the official Go styling code or something like this, +[1348.24 --> 1353.44] but if you were to do this, then you would have a... I mean, let's take it from the other way. If you +[1353.44 --> 1358.40] have some normal looking Go code, like a hello world or something like this, it would probably translate +[1358.40 --> 1364.32] into 10 or 15 lines of assembly. So I would say the default would be 15 lines of assembly for one +[1364.32 --> 1371.12] line of actual Go code. But then if you get up into lines of codes that are a bit more complex or that +[1371.12 --> 1376.16] check or that return multiple return values or function calls, then this can get a bit bigger, +[1376.16 --> 1380.16] but this is still going to be the right ballpark. Okay. Yeah. That gives me a good idea. +[1380.16 --> 1384.32] What is it for other languages? Like, is it a lot more? Is it a lot less? Is it roughly the same? +[1384.32 --> 1391.60] I would say it's probably going to be mostly the same. C++ tends to be very, like, tends to +[1391.60 --> 1397.60] generate a lot of codes. It's very comparable to us to Go. C might be a bit more direct. Like the +[1397.60 --> 1402.72] translation between C and assembly is going to be a bit more, how would I say it in English? The +[1402.72 --> 1407.76] correspondence between C code and the assembly is going to be a bit more direct. That's it. But +[1407.76 --> 1413.36] otherwise I would say this is like a common ratio for languages. The problem is not that Go generates +[1413.36 --> 1417.60] more assembly. The problem is that the assembly generates, like, is not the one that we are used +[1417.60 --> 1422.56] to seeing. And we don't like that. Interesting to see if in one or two years from now, it will be more +[1422.56 --> 1428.64] supported and more pattern recognition working. Well, that's the thing, right? It kind of depends on the +[1428.64 --> 1434.48] attackers. Like, if we do end up seeing more and more Go tools out there in the wild, then there +[1434.48 --> 1440.96] is going to be pressure on the tool authors like IDAF, like Ghidra, etc. to implement better detection +[1440.96 --> 1445.52] and, you know, better support for those languages. I'm pretty sure that since last time I tried using +[1445.52 --> 1450.96] the compiler on some Go program, IDA has made improvements and it's probably not as broken as it +[1450.96 --> 1456.56] used to be. But if you keep seeing offensive tools running in Go, then I'm pretty sure that the tools will get better. +[1456.56 --> 1461.76] We will still have to figure out how the Go assembly works, especially if it changes again in the future. +[1461.76 --> 1466.48] But overall, at least the support in the last years has improved tremendously. +[1466.48 --> 1472.24] And I think it will continue to do so, although also in the future, if there is a need to. And I would +[1472.24 --> 1476.40] guess that Go is only going to become more prevalent when it comes to offensive software. +[1476.40 --> 1478.08] Because of all the reasons that you mentioned. +[1478.08 --> 1479.84] Yeah, exactly. +[1479.84 --> 1485.92] Some specific questions. You mentioned that you were kind of thinking out loud about the behavior you see +[1486.56 --> 1492.88] in IDA Pro when you were looking at the Go code that you loaded there, or the binary of it that you +[1492.88 --> 1498.56] loaded there. So I'm going to describe two things that you mentioned. Tell me how you think if it's +[1498.56 --> 1503.36] good, if it's bad, how it is compared to other languages, just as an interesting kind of point. +[1503.36 --> 1508.40] It can get too deep, so we'll try to keep it at a slightly higher level for everybody who's out of, +[1508.40 --> 1514.40] of, was kind of hearing about this and not very well familiar. So for example, you mentioned that +[1514.40 --> 1519.52] keeping to the next instruction lands you in another place in the code of the CPU instruction. +[1519.52 --> 1524.08] Yeah, exactly. So this is something that was super surprising to me, which is when I reverse +[1524.08 --> 1528.88] engineer a program. So you can look at it statically in IDA Pro, which means you display the instructions +[1528.88 --> 1533.92] and you read them like a book. Or there's another approach, which is not like opposite, but maybe more +[1533.92 --> 1539.04] like a complement to it, which is to look at the program inside the debugger. You know, debugger, +[1539.04 --> 1542.56] they just work exactly the same as in the software development world. Like you execute the code +[1542.56 --> 1546.56] instruction by instruction or line by line, and you can see the state of the various variables. +[1546.56 --> 1550.88] Except for us, you know, we don't have the source code, so it's not lines of code, it's just assembly +[1550.88 --> 1556.40] instructions. But we can still watch them execute one by one, and we can see the CPU registers getting +[1556.40 --> 1560.56] updated, etc. And when I was doing this with the Go programs, it turned out that I was very surprised to see that +[1560.56 --> 1566.48] sometimes I would step from one instruction to the next and I would end up at a totally random place +[1566.48 --> 1571.52] somewhere else in the program. And eventually, like by doing some Google searches, etc., I figured out +[1571.52 --> 1576.80] that it is actually the... I don't know if it's the Go scheduler that is involved in there, probably it is, +[1576.80 --> 1581.76] but there is a garbage collector that is in charge of, you know, freeing the variables that are not +[1581.76 --> 1587.12] used anymore. And sometimes it takes priority and starts, you know, freeing stuff. And then once it's +[1587.12 --> 1592.00] done running, it takes you back where you were in the program. And so this is something that is super +[1592.00 --> 1596.48] jarring for us as reverse engineers, because we are looking at a very specific place in the program. +[1596.48 --> 1601.76] We are frowning, looking very concentrated and focused, because we are looking super serious. +[1601.76 --> 1606.80] And then we press F7, F7, we step into the next instruction. Suddenly we end up somewhere totally +[1606.80 --> 1610.64] different, even though we didn't see any jump instruction. And suddenly it's like, oh, something +[1610.64 --> 1615.52] is going on. What's happening with my program there? Because it's not supposed to just go somewhere else. +[1615.52 --> 1620.24] Now, once I was able to figure out what was going on and understand that I just have to get out of +[1620.24 --> 1624.96] this garbage collector function, and it will take me back exactly where I used to be, and things were +[1624.96 --> 1631.28] fine. But initially, it was another one of Go's Indiosyncrasies that felt super alien to me, and +[1631.28 --> 1636.32] that, like, I wasn't happy about at first. So that means it's not something, behavior that you see often +[1636.32 --> 1642.00] in other languages? Oh, no, it's something I had never seen before. I know that other languages, they do have +[1642.00 --> 1645.92] their own garbage collectors. But when it comes to Java, we don't really have to look at the +[1645.92 --> 1652.40] instructions, because Java is compiled to bytecode. So we just read the code disassembled, or decompiled +[1652.40 --> 1658.08] maybe, and get access to something that looks like the source code. It may be obfuscated, which means that +[1658.08 --> 1663.52] it will be modified in a way that the variable names are not there anymore, or it has been specifically +[1663.52 --> 1668.56] engineered to be harder to read. But in that case, or for .NET or for Java, we just never have to worry about +[1668.56 --> 1673.92] CPU instructions, because they are not that relevant to the language. So Go was, for me, +[1673.92 --> 1678.08] a big surprise on that level, because this was the first time I had to encounter debugging a program +[1678.08 --> 1684.24] and being taken far away somewhere without even asking to. And you know, it kind of happens on a +[1684.24 --> 1689.04] regular basis, too. And then one more question about another behavior that was peculiar that you +[1689.04 --> 1694.24] pointed out, that at some point, when you had two following instructions, and they were using the same +[1694.24 --> 1698.32] variable, you didn't see the return, but because it was right, the one after or before. +[1698.56 --> 1703.92] I'm not sure exactly if I remember exactly the part that you referred to, but what I noticed is, +[1703.92 --> 1709.04] yeah, this might be one of the other ways that the compiler in Go is being very smart, which is that +[1709.04 --> 1714.08] if you have chained function calls, it turns out, I think that the way that arguments from one +[1714.08 --> 1719.20] functions are returned on the stack happened to be the exact place where they would be considered +[1719.20 --> 1725.12] as arguments for the next function. So you don't really see the data moving back and forth from the +[1725.12 --> 1729.04] functions. It's just, you know, you have chained calls and the compiler knows that whatever was +[1729.04 --> 1734.24] returned happens to be at the right place for the next one, etc. So one of these other things that we +[1734.24 --> 1739.92] are used to seeing, like we see a function call, we look at the input, we look at what goes in and what +[1739.92 --> 1744.48] goes out, basically. This helps us understand what is going on. And with Go, sometimes you just don't see +[1744.48 --> 1749.36] that because it's hidden from you. Like the complexity tends to be, well, the complexity is +[1749.36 --> 1755.28] still there, but you know, all these operations are masked by the way that the stack is constructed by +[1755.28 --> 1759.84] the Go compiler, which again is a super good thing for Go programmers, because it means that you don't +[1759.84 --> 1764.64] have those memory movements that are taking place in the program that are actually not that useful. +[1764.64 --> 1770.24] And every time you have a movement that involves the memory in a program, it takes a lot of time. I +[1770.24 --> 1777.76] mean, not a lot compared to our human existence. But if you look at how a CPU works, you have the +[1777.76 --> 1782.48] CPU that has some memory regions inside of it, which are called the registers. And then you have the RAM +[1782.48 --> 1787.52] as well. You know, when you allocate memory in C program with a malloc or C alloc, it goes into the RAM. +[1787.52 --> 1793.20] Or when you move something into the stack, it's also a region of memory that is on the inside the RAM, +[1793.20 --> 1797.44] you know, the RAM stick of the computer. Every time the CPU has to talk to the RAM sticks, +[1798.00 --> 1802.40] it has to be an electrical signal that goes from the CPU through a bus to the motherboard. And the +[1802.40 --> 1807.84] motherboard understands that it has to request the specific region of data to the RAM sticks. And +[1807.84 --> 1812.96] you know, you have to have the response that goes back the same way, converted into electrical signals. +[1812.96 --> 1818.72] So it's pretty fast, of course, when it comes to, it's probably in the ballpark of microseconds or +[1818.72 --> 1826.16] milliseconds. But compared to just the CPU talking to itself, or moving stuff inside of the physical +[1826.16 --> 1829.84] area that is the CPU, or just not moving things at all, because they are already in the right place, +[1829.84 --> 1834.48] then you get performance increases that I think are pretty significant, especially considering the +[1834.48 --> 1838.00] amount of function calls that you have in the program. It's very interesting to hear about this +[1838.00 --> 1841.12] from the perspective of somebody who's kind of poking this out from the outside. +[1841.12 --> 1846.40] No, this makes me want to dive more into the reverse engineering just to learn more about the internals. +[1857.68 --> 1863.52] Hey friends, this episode is brought to you by my friends and potentially your friends too at +[1863.52 --> 1868.88] Firehydrant. And I'm here with Robert Ross, founder and CEO of Firehydrant. And Robert, +[1868.88 --> 1874.56] there are several options out there for incident management, but what is it that makes Firehydrant +[1874.56 --> 1879.36] different? The reason that we think that Firehydrant is is on to something is because we're +[1879.36 --> 1885.36] meeting companies really where they are. We face the same problems that every company in the industry +[1885.36 --> 1891.28] that is building and releasing software is also facing. So where you want people to be able to sign +[1891.28 --> 1897.60] up for Firehydrant and immediately be able to kick off an incident using the best practices that we've +[1897.60 --> 1902.40] built and we've experienced and have gathered through the other amazing customers that use our tool. +[1902.40 --> 1908.00] It really is a very quick time to value. And we want people to have a long jump from where they are +[1908.00 --> 1910.88] to where they want to be in incident management. +[1910.88 --> 1916.96] I love it. Thank you, Robert. Small teams up to 10 people can get started for free with all Firehydrant +[1916.96 --> 1921.28] features included. There's no credit card required to sign up. They are making it too easy to get +[1921.28 --> 1927.04] started. So check them out at firehydrant.com. Again, firehydrant.com. +[1927.04 --> 1956.00] So let's maybe move to a bit of a higher level now. Ghost community is kind of big on consistency. +[1956.00 --> 1960.40] You know, we have like the linters that keep everything consistent. Go format keeps everything +[1960.40 --> 1965.68] consistent. Does that actually help with reverse engineering at all? You think just the the only +[1965.68 --> 1971.12] one way to do thing and or at the level that you're doing reverse engineering if you think it doesn't +[1971.12 --> 1976.32] matter? It's a good question. I have to say I don't know that much about the linter itself. I have +[1976.32 --> 1981.76] written a bit of C code myself when I was trying to like look at assembly code and write Go at the same +[1981.76 --> 1985.68] time that would generate the same thing. So this is my extent of the experience with the language and I +[1985.68 --> 1991.36] really noticed something which is that the Go language is super strict. I have in the past used +[1991.36 --> 1996.40] the expression, maybe I think to make you laugh, I was saying you know that in Go if you don't use +[1996.40 --> 2001.44] the return values that the program is complaining. If you have unused variables then the program +[2001.44 --> 2006.88] complains again, right? And I was saying that to me Go feels a bit like fascist Python. Like it doesn't +[2006.88 --> 2011.04] let you do anything that you want except you know if it follows the rules very strictly. +[2011.04 --> 2018.40] For us it doesn't matter too much in the sense that those checks are enforced at the compiler level, +[2018.40 --> 2022.80] right? It's something that if the code is not compliant then you will not get a binary at the +[2022.80 --> 2028.96] end. So it does not add additional stuff inside the binary and also if there were some variable +[2028.96 --> 2033.76] that is unused inside the program then as you reverse engineers we would not care, right? Because we +[2033.76 --> 2038.72] would just consider that you know it's not used anymore or probably the programmer doesn't need it for +[2038.72 --> 2044.32] whatever reason and we would just move on. So for us it doesn't really change that much although +[2044.32 --> 2049.68] knowing about those guarantees kind of allows us to make more informed guesses about what is going +[2049.68 --> 2054.64] on in the program. Like for instance when I do see a function that returns multiple return values +[2054.64 --> 2060.56] then I am not a Go developer but still I am always going to assume that the last value returned is going +[2060.56 --> 2065.52] to be the object or the first one I don't recall but I will have to check but I know that since this is +[2065.52 --> 2069.60] the normal way that people are supposed to write Go code and since I know that the compiler is going +[2069.60 --> 2075.52] to force people to do it even if they don't want to that probably I can base my hypothesis on those +[2075.52 --> 2081.44] conventions which is actually pretty helpful in that regard. So would you say that Go is a good language +[2081.44 --> 2089.12] to pick up for hacker or for researcher in security? Well I'm not really in the business of helping attackers +[2089.12 --> 2094.24] you know being more efficient at writing offensive tools but if I were to then I would guess that +[2094.24 --> 2100.96] Go is probably a good language to pick up. Basically anything that is away from the traditional +[2100.96 --> 2105.92] languages is going to be more annoying for us because we are less used to it. I think Rust is going to be +[2105.92 --> 2110.00] a good choice as well. I haven't looked at Rust too much myself I have a co-worker that did and +[2110.00 --> 2117.04] also released some videos and from what he's saying it's like C++ but harder which is a kind of a high +[2117.04 --> 2125.44] standard to beat. So yeah just Go and Rust would be my advice there although it's not advice please don't. +[2125.44 --> 2131.28] So if those are the kind of the new school ones right Go and Rust what are like historically what +[2131.28 --> 2137.04] languages has everyone used on the hacking side and on the research side? Well historically everything +[2137.04 --> 2142.16] has been used you know Murphy's law which says that if there is a way to misuse something that is going +[2142.16 --> 2148.64] to be misused right and programming languages have proven time and again that law. The thing is we are +[2149.20 --> 2153.76] recipients of whatever the hackers are doing right. We do not get to choose what we are going to work +[2153.76 --> 2157.92] on like hackers are going to write their their tools and they're going to choose whatever language +[2157.92 --> 2163.20] is familiar for them or whatever language feels comfortable or whatever and this is why we end up +[2163.20 --> 2169.04] sometimes facing the most ridiculous stuff like malware written in auto IT. I don't know if you know about +[2169.04 --> 2176.48] this it's like a it's some weird scripting language that is used for UI testing and basically allows +[2176.48 --> 2182.48] you to simulate keystrokes and mouse clicks. Well it turns out people write malware with this as well. +[2182.48 --> 2187.92] Anything that has ever been available as a programming language has been one way or the other eventually +[2187.92 --> 2194.32] been used for malware. So the thing is this is our bane as reverse engineers which is that we do receive +[2194.32 --> 2198.96] malware and whatever it is we have to work on it because at the end of the day our job is to +[2198.96 --> 2204.16] figure out what was going on in that specific incident. And so whether it's C or C++ or it's +[2204.16 --> 2210.64] Go or Delphi or Pascal or whatever we just have Erlang maybe no Erlang I'm pretty sure there's an Erlang +[2210.64 --> 2217.20] malware whatever we receive we have to work on and so we cannot really afford to be picky about what +[2217.20 --> 2223.36] languages we get interested in. We just have to be able to adapt to whatever comes because everything +[2223.36 --> 2228.56] will come eventually. So you just mentioned right there like you get your research is on whatever like +[2228.56 --> 2233.36] hackers leave behind let that be malware or whatever what other things do people leave behind is it just +[2233.36 --> 2240.64] the actual binaries or like are you digging into logs and other things? Yeah so in a typical incident +[2240.64 --> 2246.96] scenario then you would have people that go into what we call forensics mode they will collect all the +[2246.96 --> 2252.24] logs they will collect all the hard drives and try to figure out exactly what happened inside the network +[2252.24 --> 2257.76] they will collect not just machine logs but you know DNS logs they will collect whatever event was +[2257.76 --> 2263.44] generated by the windows machines they will collect you know whatever was saved by the HTTP proxy and so +[2263.44 --> 2268.00] on all the net flow if it's available usually it's not usually not that much information is actually +[2268.00 --> 2271.52] available in case of an incident but you know that's someone else's problem I'm not an incident +[2271.52 --> 2279.52] responder and I have enough stuff to worry about but what I focus on is the actual malware we do have +[2279.52 --> 2285.68] information through the antivirus from Kaspersky that gives us information about the execution context so we can +[2285.68 --> 2291.12] see that okay this process launched this process etc so we have this type of information but in a bigger +[2291.12 --> 2296.56] incident context then you would get a much clearer picture about everything that went on on the in the +[2296.56 --> 2301.92] victim's network and this this whole trove of information would allow you to reconstruct the whole +[2301.92 --> 2306.00] timeline of the incident so you would see that you know at this time you had some suspicious +[2306.00 --> 2311.36] request on you know some web front end and then you would see that there is a file created at a later date on the +[2311.36 --> 2316.80] same web server and then you would maybe see some weird suspicious request to the active directory server +[2316.80 --> 2322.80] with some golden ticket with a mimicat or something well those kinds of lateral movement methods etc and at the end of the +[2322.80 --> 2330.80] day somewhere some attacker would have to drop some binaries to help them either persist on the victim machine or get +[2330.80 --> 2337.36] further into the network or deeper because they will try to do whatever they can without deploying anything some very careful +[2337.36 --> 2343.52] attackers will not deploy anything on disk and they will just deploy whatever program that they need inside +[2343.52 --> 2348.08] the memory which is very stealthy but also the machine happens to reboot then you know everything +[2348.08 --> 2352.80] that was in the memory just goes away and so if you have no way of coming back onto the victim machine then +[2352.80 --> 2357.28] you know all the access that you have deployed is lost some very stealthy attackers will decide that they +[2357.28 --> 2365.12] would rather lose access and leave forensic traces on hard drive most of them like 90 99 of them will feel like they would +[2365.12 --> 2371.60] rather leave some kind of trace knowing that most people don't look anyway and then leave stuff for +[2371.60 --> 2377.60] us to analyze later if we figure out that there was an incident and you know someone goes there collects +[2377.60 --> 2382.24] everything and just sends the binaries back to us you said the incident response teams are the ones that +[2382.24 --> 2388.16] collect all that data and all of that yeah exactly so we do have such teams at Kaspersky but most +[2388.16 --> 2393.28] cyber security companies will have either their internal instance responders or swoop in you know a +[2393.28 --> 2399.76] contractor that they know of and that can be called at any hour of the day or the night and that will come and you know +[2399.76 --> 2406.80] just exactly swoop in with the big guns if something weird took place now it doesn't mean that we do not work in +[2406.80 --> 2411.76] direct interaction with those teams it means that you know this is their job and then you know we get we are more back +[2411.76 --> 2419.28] office guys where you know we get escalated some stuff and then we look into it but most of the intelligence that we create +[2419.28 --> 2424.32] doesn't actually come from incident response cases i think it would be a good idea if we were able to +[2424.96 --> 2430.40] gain more information from that source as well i think it's a very valuable one but we work mostly +[2430.40 --> 2435.60] on the telemetry collected by our antivirus you know all the samples that are suspicious or that are +[2435.60 --> 2441.68] uploaded to the cloud for analysis and then we can also swoop in but you know very much more quietly +[2441.68 --> 2446.80] and look at all this data and see okay this looks interesting because you know we've never seen this before +[2446.80 --> 2451.20] or it looks like some malware that we saw 10 years ago and we haven't seen since and you have some +[2451.20 --> 2456.40] modifications and then we are interested in what happened since then right but it's not really our +[2456.40 --> 2461.76] work tends to be a bit disconnected from the actual incidents and really more focused on looking at the +[2462.40 --> 2467.68] big data lake that we have and try to understand what is relevant inside of it that's cool thanks for +[2467.68 --> 2475.36] that that insight from the other side of this equation what are some tips you can give for writing a secure +[2475.36 --> 2481.76] software for people who do go or in general if it's not specific to go it's also useful yeah i think +[2481.76 --> 2487.44] one of the main appeals of go is that you don't really need to think about security as much as with +[2487.44 --> 2493.44] other languages go is a memory safe language unless i'm mistaken and the compiler is never going to let +[2493.44 --> 2498.56] you do stupid stuff like create an array that is too small and then you know write stuff that goes out +[2498.56 --> 2505.04] of it like it's not just not possible so it eliminates a whole lot of bug classes which we call memory +[2505.04 --> 2509.44] corruptions it's just not going to happen you cannot do this to yourself and go and it means +[2509.44 --> 2516.16] that all the old school buffer overflows that plagued all the c and c++ programs for dozens of years by +[2516.16 --> 2522.48] now just are not going to ever happen in the go language it doesn't mean that the program is going to be +[2522.48 --> 2528.48] like perfectly safe from any security issues but the issues are not going to be related to oh i made a +[2528.48 --> 2533.36] programming mistake and uh there is a bug in my program it's going to be exploited it's going to be +[2533.36 --> 2539.92] more related to design issues like the memory safe language does not help you implement a secure +[2539.92 --> 2546.08] authentication scheme for instance it doesn't help you write a well thought out network protocol i saw +[2546.08 --> 2550.96] that go really helps you with cryptography i noticed that it's very difficult to choose algorithms that are +[2550.96 --> 2556.96] not safe like by default you can only i don't think you can choose the algorithms and go by like i know you +[2556.96 --> 2562.88] can do aes for instance but like the cypher mode or the those kind of stuff tends to be unless i'm +[2562.88 --> 2567.04] mistaken you know selected by default for you and the defaults are good so you're not going to be +[2567.04 --> 2574.32] making those mistakes but oh yeah the iv something i was working on some code in go that was uh relying +[2574.32 --> 2579.52] on aes i was looking at trying to figure out exactly how the iv was generated and so on i was looking +[2579.52 --> 2585.20] saying that nowhere in the developer code and doing some research i noticed that it was actually go that +[2585.20 --> 2591.76] would by itself generate an iv for the encryption this initialization vector and then it would append +[2591.76 --> 2597.92] it to the uh somewhere in the final encrypted buffer and so usually in other languages this is something you +[2597.92 --> 2603.60] would have to do on your own and this is a like a big avenue for making mistakes like if you choose a +[2603.60 --> 2608.00] stupid iv like just zeros or if you do not select one at all then you're going to have encryption +[2608.00 --> 2613.28] problems go would not let you do this so it's very obvious to me that go was created with +[2613.28 --> 2618.48] security in mind not for the developers but by the go creators like they don't want you to shoot +[2618.48 --> 2621.84] yourself in the foot and you know they're going to make sure that there's no way for you to do it +[2621.84 --> 2627.60] unless you really really want to even though you do have all those kinds of protections cryptography +[2627.60 --> 2632.48] like can be misused like you know if you choose a bad key then you know nobody's going to save you from +[2632.48 --> 2637.92] that if your protocol doesn't work then again you cannot be protected from it either +[2638.64 --> 2645.04] i think it allows people to focus on design flaws instead of programming flaws and this is already +[2645.04 --> 2651.04] like a huge burden off the shoulders of developers that is a very interesting insight that's interesting +[2651.04 --> 2656.24] i see a lot of complaints outside of kind of the go community just you know like hacker news about +[2656.24 --> 2662.56] goes choosing your defaults for like like tls or not letting you do certain things but that's +[2662.56 --> 2666.96] one i'm firmly on board with if i don't want to if i don't need to think about it i don't want to +[2667.60 --> 2672.72] and i don't want to make the mistake would you be able to confidently select your defaults for tls i +[2672.72 --> 2677.20] mean i don't think i would feel comfortable doing this like you have to be like very well versed in +[2677.20 --> 2681.84] cryptography to be able to make those kinds of decisions so it's very good that go is not making +[2681.84 --> 2688.56] you do this i think in my opinion +[2688.56 --> 2700.56] this episode is brought to you by honeycomb find your most perplexing application issues +[2700.56 --> 2706.96] honeycomb is a fast analysis tool that reveals the truth about every aspect of your application +[2706.96 --> 2712.16] in production find out how users experience your code in complex and unpredictable environments +[2712.16 --> 2717.68] find patterns and outliers across billions of rows of data and definitively solve your problems +[2717.68 --> 2722.08] and we use honeycomb here at changel that's why we welcome the opportunity to add them as one of our +[2722.08 --> 2727.60] infrastructure partners in particular we use honeycomb to track down cdn issues recently which we talked +[2727.60 --> 2732.72] about at length on the kaizen edition of the ship it podcast so check that out here's the thing +[2732.72 --> 2737.28] teams who don't use honeycomb are forced to find the needle in the haystack they scroll through +[2737.28 --> 2741.92] endless dashboards playing whack-a-mole they deal with alert floods trying to guess which +[2741.92 --> 2747.04] one matters and they go from tool to tool to tool playing sleuth trying to figure out how all the +[2747.04 --> 2751.92] puzzle pieces fit together it's this context switching and tool sprawl that are slowly killing +[2751.92 --> 2758.24] teams effectiveness and ultimately hindering their business with honeycomb you get a fast unified and +[2758.24 --> 2764.96] clear understanding of the one thing driving your business production with honeycomb you guess less +[2764.96 --> 2770.24] and you know more join the swarm and try honeycomb free today at honeycomb.io +[2770.24 --> 2775.36] slash changelog again honeycomb.io slash changelog +[2775.36 --> 2802.56] another interesting about your interest in go you kind of mentioned that you started using go because +[2802.56 --> 2808.72] malware was thrown at you kind of yeah exactly so i wouldn't say that i've started using go i would +[2808.72 --> 2814.16] say that i was forced to learn go not that i i'm unhappy about it i'm not saying it's a bad thing +[2814.16 --> 2820.08] what i'm saying is that i'm not really writing go code myself what i did was i had assembly that was +[2820.08 --> 2830.56] generated by the go compiler and i was trying to make heads or tails from it so what i did was i looked at the assembly i was like okay this might be the go code that generated this assembly and then i opened my go id +[2830.56 --> 2838.96] i opened my go ide and i compiled my code and checked if it was the same on both ends also when i start to +[2838.96 --> 2843.92] learn about a language when i want to reverse engineer it i think it's super useful to write some simple +[2843.92 --> 2848.48] programs and just compile it and see how it looks in the assembly level you know just create a simple +[2848.48 --> 2854.48] stupid c function not c function but a sum function that you know adds to integers or you know something that +[2854.48 --> 2859.92] will allow you to see what types of function calls the program is using what kind of constructs the +[2859.92 --> 2864.48] language is generating the things that i had to face there was again the go compiler being way too +[2864.48 --> 2871.04] smart for for my uses and it tends to inline all the function calls that are too simple what i mean +[2871.04 --> 2875.76] by this is if you have a simple function that does almost nothing and you call that function then the +[2875.76 --> 2880.64] go compiler will be like oh this is not worth a function call what i will do is i will take the code of all +[2880.64 --> 2885.04] this function and put it inside the calling function and you know when you try to look at what +[2885.04 --> 2889.60] a function call looks like in assembly then this is not helping you but the good thing is i was able +[2889.60 --> 2894.80] to find the the good flags and uh for the compiler to disable opt all optimizations and things kind of +[2894.80 --> 2900.32] worked out for me you mentioned that ida and which is the main tool you're using and the other tool +[2900.32 --> 2907.44] are not really supporting go so if anybody wants to try reverse engineer to get into that but also want to +[2907.44 --> 2913.76] do that with go what would you recommend how to do that so if you're going to reverse engineer go +[2913.76 --> 2918.56] programs i still think that you don't have much choice there so you're still going to be you're +[2918.56 --> 2923.76] still going to have to use either either pro or ghidra i want to switch to ghidra eventually but i +[2923.76 --> 2929.76] haven't done so at the moment so i cannot speak too much about its capabilities i'm told that it's being +[2929.76 --> 2937.12] improved at a very rapid pace so it's probably a good choice but when it comes to to ida it got better +[2937.12 --> 2942.64] right i think that's a few months back maybe a year now you had my good friend uh juan andres +[2942.64 --> 2947.28] guerrero sad from central one on the podcast and probably told you about the various plugins that +[2947.28 --> 2953.68] he wrote to help people reverse engineer go programs with ida i also contributed to his repository with +[2953.68 --> 2960.96] some script myself that i find useful but overall even though ida might not be perfect for the job +[2960.96 --> 2966.48] it's still one of the two only tools that are available for the job so you still have to +[2966.48 --> 2972.96] you know work through it no matter what the thing is i find myself thinking that even though starting +[2972.96 --> 2979.28] with reverse engineering go is kind of difficult it turns out that i find myself liking reverse engineering +[2979.28 --> 2984.40] go programs way more than c++ programs that tend to be extremely complicated with you know virtual +[2984.40 --> 2989.52] function tables and the very complex structures that represent classes and so on because when it comes to the +[2989.52 --> 2994.80] go language turns out that it kind of feels like a scripting language in the sense that everything +[2995.44 --> 3000.48] ends up being a call to an api function or a call to some function that comes from the go standard +[3000.48 --> 3006.16] library and so if you are able to take a debugger and you know look at all the arguments after you know +[3006.16 --> 3012.32] how to do that but if you look at all the arguments of the go functions that are documented by the way and +[3012.32 --> 3018.24] look at the return values then actually the meaning of the program tends to manifest itself even though +[3018.24 --> 3022.72] you don't really understand all the instructions that are in the middle and you cannot track you +[3022.72 --> 3027.44] know all the stuff going uh here and there so overall like my advice for people that would like +[3027.44 --> 3031.68] to get started with go reverse engineering is okay it's going to be very different from what you are +[3031.68 --> 3036.72] used to but at the end of the day i think you're going to end up liking it more than you would think +[3037.52 --> 3043.04] because it's going to be way easier than it looks how about those listeners that haven't done any +[3043.04 --> 3046.72] reverse engineering that want to get started do you have any good resources out there i know that you +[3046.72 --> 3050.64] personally have made some videos you want to talk about that a little bit and anything else that +[3050.64 --> 3056.16] would be helpful so yeah the videos that i put out are just related to the goal language if you're +[3056.16 --> 3060.72] going to get into reverse engineering i would not advise you to start with go not because it's going +[3060.72 --> 3067.12] to be harder or anything but because probably the basics of reverse engineering are going to be related to +[3067.12 --> 3073.44] traditional c code or traditional assembly code generated by c and so this is going to be like your +[3073.44 --> 3078.24] base knowledge of reverse engineering and then once you are comfortable with understanding what +[3078.24 --> 3085.28] is going on with the c language and you know all the assembly that you see most places then you can +[3085.28 --> 3092.72] move on to other languages and see how they differ from others or etc but i think c is always going to +[3092.72 --> 3097.84] be used as a reference for other languages in the sense that when you look at assembly first you try to +[3097.84 --> 3102.64] understand it like you would understand c and then if it's different like you adapt from that but if your +[3102.64 --> 3107.92] baseline is going to be the go language if the like the the one thing you know is go and then you try +[3107.92 --> 3113.44] to recognize whatever you learned with go with another language then you're going to be into trouble +[3113.44 --> 3118.48] because whatever you're going to see next is not going to look like anything you saw in go so we do +[3118.48 --> 3122.80] have a few courses at kaspersky but i mean people can check them out if they want there are a few +[3122.80 --> 3130.48] interesting uh online courses as well it's something for free which is a beginners.re it's a website it used to be +[3130.48 --> 3135.28] free maybe now it's behind a paywall i'm not sure but it used to be a big big reverse engineering course +[3135.28 --> 3140.72] written by some guy and it was uh it was amazing you have a book which is called practical malware +[3140.72 --> 3147.44] analysis it's a bit old now but i think it's still very much up to date it's from no stock press i think +[3147.44 --> 3152.24] for beginners it's going to be a good way to get into the field because it explains everything that is +[3152.24 --> 3157.76] going on it provides links to the various tools that you might need etc so a good resource there and +[3157.76 --> 3163.92] finally if you want to approach this from the fun angle i can actually recommend extremely good steam +[3163.92 --> 3169.92] games that allow you to like get a feel for reverse engineering so one of them is called turing complete +[3170.48 --> 3175.52] and this one is uh like the pitch of this game is you're going to build your own computer and so you +[3175.52 --> 3181.84] start with they give you logic gates like a xor gate or the electric cables basically and based on +[3181.84 --> 3187.68] this you have to build a cpu component by component and then you move on with increasing levels of +[3187.68 --> 3194.96] abstraction so it really it's really super helpful to understand how a program works or how a computer +[3194.96 --> 3201.20] works it really it allows you to get this uh eye level bird eye view of how a cpu is constructed and +[3201.20 --> 3206.40] how it's supposed to operate and knowing how cpus work is then very very helpful when you are doing +[3206.40 --> 3212.56] reverse engineering and then you have other games which are from a developer which is called zaktronics +[3212.56 --> 3218.64] and these are like weird puzzle games that are really related to computing problems one of them +[3218.64 --> 3224.24] is called a tis 100 you have another one called hexapunks and they are dubbed the assembly games you +[3224.24 --> 3229.76] didn't know you wanted and it's actually a very apt description because these games have their own +[3229.76 --> 3233.84] weird and limited assembly language and you have to solve puzzles with them like you have to program +[3233.84 --> 3238.40] some sort of small machine in order to make it do stuff and you have to do this with assembly and +[3238.40 --> 3244.88] it forces you to use the language which uh like has the very super good desired side effect of making +[3244.88 --> 3250.56] you learn how cpus work when we are making you more comfortable with handling those uh weird instructions +[3250.56 --> 3256.00] by yourself so these would be my recommendations for people that want to get into it yeah i'd not +[3256.00 --> 3261.44] thought about games that sounds i'm gonna check those out later actually yeah actually if you are +[3261.44 --> 3266.48] working from a university or if you are a teacher somewhere zaktronics i think the company maybe +[3266.48 --> 3271.04] closed doors not too long ago i think they they are done making games or they move on to something +[3271.04 --> 3277.36] else but they used to have a very extensive education program where if you are a university and you are +[3277.36 --> 3282.16] doing a i don't know some computer science degree or something like this you could just send them an +[3282.16 --> 3287.04] email and they would give you access to all their games you know for free basically and you could use +[3287.04 --> 3292.08] them to teach or as teaching aids and i think it's uh like amazing of them and also the games are +[3292.08 --> 3297.44] really really really fun i think so they are fun if you like assembly which i think is a pretty biased +[3297.44 --> 3302.48] statement on my end but i guess still do recommend them a lot of the things you said they're like a +[3302.96 --> 3309.84] cheat sheet for reverse engineering and lots of useful information and i have so many more +[3309.84 --> 3314.80] questions about specific things about go and reverse engineering we might have to do another episode +[3314.80 --> 3320.72] about this because we are running out of time sure well i can come back whenever you like we will +[3320.72 --> 3325.20] prepare our questions we'll ask you about things like generics i will have to prepare those questions +[3325.20 --> 3341.12] as well i guess but no problem now it's time for an unpopular opinion +[3349.04 --> 3354.32] so ivan what is your unpopular opinion for us oh my god i totally forgot about that oh it's okay +[3355.52 --> 3360.72] the good thing is i do have many unpopular opinions so i'm going to like give you things on the top of +[3360.72 --> 3365.44] my head and you can tell me what you want to know more about for instance i think that cyberspace is +[3365.44 --> 3372.56] never going to be regulated i think that nfts are a scam i think that there is no political will to +[3372.56 --> 3378.24] limit the sale of cyber offense tools that kind of stuff yeah that's i do have a lot of unpopular +[3378.24 --> 3381.76] political opinions as well but i don't think i want to inflict that onto you you've been very nice to me +[3381.76 --> 3390.08] what do you think about the european rule about the usbc until standardizing usbs oh i'm very very +[3390.08 --> 3398.08] happy about it i know it's uh some pressure put on some device constructors but i've been carrying +[3398.08 --> 3403.28] lots of different chargers for years and i'm super annoyed about this and you know knowing that we are +[3403.28 --> 3409.12] going to switch to like a single usbc for every single device makes me extremely extremely happy +[3409.68 --> 3414.08] another unpopular opinion i have which you can add to the list is that i'm not really a big fan of +[3414.08 --> 3419.36] apple like not at all i don't like their ecosystem and i'm not going to get into this but one of the +[3419.36 --> 3423.20] things i don't like is that people have to pay 40 bucks for like new chargers and they change +[3423.20 --> 3427.68] chargers every time they release a new product and i'm very happy that this is going to cut off this +[3427.68 --> 3431.60] revenue stream for them because i think this should have never existed in the first place what do you +[3431.60 --> 3435.76] think about all the like the walled systems you know like the google play store and the apple store +[3435.76 --> 3440.32] and the amazon store like from a securities perspective they say it's safer do you agree +[3440.32 --> 3446.32] with that do you yeah this is a very good question i do have very ambiguous feelings about them i do +[3446.32 --> 3451.52] believe that for the security on the security perspective it's kind of a good thing in the sense +[3451.52 --> 3456.48] that yeah it's another one of those safeguards that prevent people from doing stupid stuff with their +[3456.48 --> 3463.60] devices and you know having to go to some friends places or more specifically friends of my mom's +[3463.60 --> 3468.24] places to debug computers and you know uninstall malware and fix the printers then i'm very happy +[3468.24 --> 3474.16] when you there are protections that prevent them from doing that kind of stuff then again they are not a +[3474.16 --> 3480.48] perfect solution either i think the apple store in terms of security is uh like pretty good the google +[3480.48 --> 3485.92] store the play store has a bad track record when it comes to hosting malware i'm not saying that they're +[3485.92 --> 3490.48] doing a bad job i think it's a very very difficult job but the fact of the matter is there are a +[3490.48 --> 3497.60] number of apps on the google play store that turned out to maybe not be total malware but some of them +[3497.60 --> 3504.08] are but a lot of them are just you know there to collect personal data or that kind of stuff so i think +[3504.08 --> 3510.64] a better way of securing those devices is not to control the app stores i think created protections on +[3510.64 --> 3516.96] the device level is probably where i would um work so when you look at both ios and android they are +[3516.96 --> 3521.60] doing i think a very good job of or have been doing a very good job at least in the past years of making +[3521.60 --> 3526.56] sure that apps will not be able to access anything just because the user clicked okay on a way back +[3526.56 --> 3532.24] when when they installed the app so i think making sure that all those personal information cannot be +[3532.24 --> 3539.84] pulled so easily is going to be like a much better way than you know trying to police all the stores and +[3539.84 --> 3544.24] look at all those thousands of apps that are updated there every day which i do not think +[3544.24 --> 3551.84] that you can realistically ensure that they are always going to be safe but overall the other issue +[3551.84 --> 3556.40] with world gardens which is okay maybe they do provide something with security but also i feel +[3556.40 --> 3562.32] like they take away some agency from me as a user right i really like to own the devices that i use +[3562.88 --> 3567.52] and having some restrictions that tell me oh you cannot install this app because google says you can't +[3567.52 --> 3572.40] or you can uninstall this app also because google says you can't is something that tends to make me +[3572.40 --> 3579.20] extremely extremely angry so you mentioned a lot of unpopular opinions yes the way the twitter works +[3579.20 --> 3586.32] for our podcast is that we take an unpopular opinion and then we make a vote so there's a poll do +[3586.32 --> 3590.96] people agree with you or not and then there's a hall of fame for unpopular opinions and for popular +[3590.96 --> 3598.00] unpopular opinions so you listed several which one would you like us to vote on so if i wanted to win +[3598.00 --> 3602.48] the contest i guess i would go with the nft one because i know that this is something very divisive +[3602.48 --> 3607.76] and i think that a lot of the audience that you are reaching is going to be probably i'm not going +[3607.76 --> 3611.92] to say that they are necessarily going to be on my side but i think they're going to be on the a side +[3611.92 --> 3616.72] but i think a much more interesting question that i would be actually interested in having the uh the +[3616.72 --> 3621.52] community's opinion about is the one about regulation like i do believe that cyberspace is +[3621.52 --> 3625.92] never going to be regulated and maybe i need to say a bit more about this one right so that people +[3625.92 --> 3631.12] can figure out for themselves my opinion on this is that you know we have a number of high-level +[3631.12 --> 3636.88] discussions taking place at the un about you know acceptable norms for behavior in the cyberspace etc and +[3636.88 --> 3641.68] you have all those discussions between states where they talk with each other and they are like okay +[3641.68 --> 3646.40] what type of offensive operations are legitimate like for instance espionage is okay but destructive +[3646.40 --> 3651.28] attacks are not okay i mean i'm not saying this is right i'm just saying this is probably the kind +[3651.28 --> 3655.68] of discussions that they are having and you know we may have differing opinions on what types of +[3655.68 --> 3660.40] attacks are okay and what types are not or even if attacks are okay at all it doesn't matter the thing +[3660.40 --> 3667.68] is i do believe that i don't think that we will ever reach an agreement there because well states +[3667.68 --> 3675.52] do not have an incentive to regulate cyber offense i think that they have an interest in having a way or +[3675.52 --> 3679.92] having some kind of framework that allows them to still conduct operations because when they +[3679.92 --> 3684.96] conduct operations they know what they are winning right they have intelligent services that gather +[3684.96 --> 3688.88] data they collect it through cyber means they take it back and so they know that they are able to +[3688.88 --> 3693.68] achieve certain results because they have obtained specific information and they can quantify that +[3694.32 --> 3701.28] on the other end when you look at the costs of cyber offense which means all your companies in +[3701.28 --> 3705.20] your country that have been breached because there are no no such norms it's something that's super +[3705.20 --> 3710.24] hard to quantify you can never know that you know you lost some contract overseas to sell planes or to +[3710.88 --> 3715.76] something else because the cyber means because it's very likely that nobody knows that the breach even +[3715.76 --> 3722.08] happened in the first place so the thing is you look at the balance of risk reward for the decision +[3722.08 --> 3727.68] makers and they see this is what we win with cyber offense which is a lot and what they lose is +[3727.68 --> 3732.48] actually it's painless and also they have no idea what it is and so overall i think that all those +[3732.48 --> 3737.44] discussions that are taking place that are saying okay we need to make a safer internet blah blah blah +[3737.44 --> 3742.72] are actually possibly being conducted in bad faith because there is no political will to actually stop +[3742.72 --> 3747.44] doing this kind of stuff this would be my unpopular opinion especially in the diplomatic circles +[3747.44 --> 3753.52] all right you will be tagged and we will be following the results okay no i'm interested to see the results +[3753.52 --> 3758.64] on this one yeah it's an interesting way to think about it yeah i want to know as well cool thank +[3758.64 --> 3762.56] you very much for sharing your knowledge your thoughts and your opinions with us this was +[3762.56 --> 3768.40] really fascinating we will be very happy to have you again thanks a lot ivan well thank you very much +[3768.40 --> 3773.36] for having me and uh yeah feel free to like call me up anytime i'll be happy to be back thanks +[3773.36 --> 3781.52] ian for joining it was fun co-hosting together yeah thanks you guys this was great bye everyone +[3783.76 --> 3789.92] if you enjoyed listening at the intersection of go and information security stay tuned for part three +[3789.92 --> 3795.36] it's currently scheduled to record live on november 29th and of course go back and listen to part one +[3795.36 --> 3802.00] while you're at it that was episode number 205 here's a sampler the stakes have been ratcheting up +[3802.00 --> 3806.88] it's really easy to kind of look at it that way and we don't want to make it like too dark or too +[3806.88 --> 3813.84] heady but now this is the playground of also a lot of nation states and a lot of criminals and you know +[3813.84 --> 3819.12] if you're in this in the us it's kind of like the ransomware epidemic is sort of unavoidable right like +[3819.12 --> 3824.08] you you have to talk about it every day and that's where things get less pretty right like if you're at +[3824.08 --> 3829.76] a hospital that can't help folks because all their you know tragically outdated windows xp systems are +[3829.76 --> 3834.80] in a flat network and all of them got popped at the same time that's where you go well yeah that code +[3834.80 --> 3841.36] was fun i love the idea of just having these you know kind of hacking superpowers but there's a side +[3841.36 --> 3846.40] to it that isn't quite so cute and i think we're kind of walking that line all the time right where you +[3846.40 --> 3851.92] go oh this is fascinating and you just get wrapped up in the functionality and what someone has been able +[3851.92 --> 3858.16] to accomplish and it's easy to forget like oh well there's this is actually a part of a much much +[3858.16 --> 3865.68] heavier game listen in at gotime.fm slash 205 or search hacking with go in your podcast app of +[3865.68 --> 3872.16] choice it should pop right up thanks once again to our partners at fastly and fly.io they help make +[3872.16 --> 3878.00] go time possible into the mysterious brake master cylinder go times beats are dope because bmc makes +[3878.00 --> 3884.96] dope beats it's as simple as that next time on go time angelica natalie and chris welcome tech lawyer +[3884.96 --> 3891.28] lewis via to the show to answer the age-old question who owns our code stay tuned i think +[3891.28 --> 3894.64] you're gonna dig it we'll have that episode ready for you next week +[3906.64 --> 3906.72] you diff --git "a/Hacking with Go\357\274\232 Part 3_transcript.txt" "b/Hacking with Go\357\274\232 Part 3_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..0f59ee94b698db8fcfe4c2173661e4ea17afb626 --- /dev/null +++ "b/Hacking with Go\357\274\232 Part 3_transcript.txt" @@ -0,0 +1,1856 @@ +[0.00 --> 6.52] This week I was working on another APT case and one of the malware implants that we found there was actually written in Pascal. +[6.84 --> 11.44] And so it was kind of a trip down the memory lane to one figure out what Pascal was again. +[11.82 --> 16.78] And also trying to understand what kind of assembly was generated by this Pascal compiler. +[16.94 --> 20.24] It wasn't that bad, actually. Way less bad than having to discover the Go language. +[20.24 --> 28.30] What's up, friends? This episode is brought to you by Sourcegraph. +[28.76 --> 34.20] With the release of Sourcegraph 4.0 and the Starship event just a few weeks behind us, +[34.40 --> 40.62] it's super clear that Sourcegraph is becoming not just CodeSearch, but a full-on code intelligence platform. +[41.00 --> 44.16] And I'm here with Joel Cortler, Product Manager of Code Insights for Sourcegraph. +[44.46 --> 49.08] Joel, this move from CodeSearch to code intelligence is a really big deal. +[49.08 --> 55.46] How would you explain this feature, Code Insights, if you're just talking to folks in the hallway track of your favorite conference? +[56.06 --> 60.28] I would really start with technical because before I was a product manager, I used to be an engineer as well. +[60.60 --> 65.42] And it's really cool and exciting just to be able to say, we're going to turn your code base into a database. +[65.94 --> 70.32] And the structured language that you need to interact is just the ability to write a code search. +[70.60 --> 72.54] You know, literal search, that's totally fine. +[72.70 --> 76.34] Regular expression, you know, that'll give you a few more advanced options, even a structural search. +[76.34 --> 82.88] But the number of long-tail possibilities it unlocks, truly the journey of building this product was just saying, +[83.02 --> 86.08] well, we've just unlocked, you know, an infinite number of possibilities. +[86.52 --> 91.16] We got to figure out some immediate use cases so we can start to, you know, invest in this product, build it and sell it. +[91.36 --> 95.58] But we're only getting started in terms of the number of uses that we're uncovering for it. +[95.58 --> 103.32] The story I told you about discovering, like, version tracking turned out to be a really important use case that wasn't even on our roadmap six months prior to discovering that, +[103.38 --> 111.34] as we were already planning to launch this product, until we talked to enough folks, realized this was a problem, and then found, well, oh, that's like a simple regular expression capture group that you can just plug right in. +[111.42 --> 114.96] Because we really built this system to not limit the power of what we built. +[114.96 --> 119.06] We don't want to give you, like, three out-of-the-box templates, and you can only change, like, one character or something. +[119.16 --> 122.06] It's truly, like, the templates are there to hold your hand and get you started. +[122.20 --> 126.34] But if you can come up with anything you want to track in your code base, you can do that with Code Insights. +[126.64 --> 127.74] I love it. Thank you, Joel. +[127.86 --> 132.48] So right now, there is a treasure trove of insights just waiting for you. +[132.76 --> 138.26] Living inside your code base, your code base is now a querible database thanks to Sourcegraph. +[138.26 --> 142.80] This opens up a world of possibilities for your code and the intelligence you can gain from it. +[143.06 --> 148.32] A good next step is to go to about.sourcegraph.com slash code dash insights. +[148.64 --> 149.88] The link will be in the show notes. +[150.24 --> 152.74] See how the teams are using this awesome feature. +[152.98 --> 157.82] Again, about.sourcegraph.com slash code dash insights. +[158.16 --> 160.44] Again, this link is in the show notes. +[168.26 --> 176.44] Let's do it. +[177.02 --> 178.06] It's Go Time. +[178.56 --> 179.96] Welcome to Go Time. +[180.96 --> 184.82] Your source for diverse discussions from all around the Go community. +[185.74 --> 187.72] Changelog has joined the Fediverse. +[188.56 --> 191.62] Find our Mastodon instance at changelog.social +[191.62 --> 195.06] and connect with us at gotime at changelog.social. +[195.06 --> 200.80] Thanks to our partners at Fastly for delivering Go Time super fast all around the world. +[201.00 --> 202.90] Check them out at fastly.com. +[203.22 --> 205.18] And to our friends at fly.io. +[205.50 --> 207.34] Deploy your app servers close to your users. +[207.72 --> 208.82] No ops required. +[209.16 --> 211.20] Learn more at fly.io. +[211.56 --> 213.34] Okay, here we go. +[216.34 --> 223.58] Today, Ivan Kwiatkowski, you are joining us again to talk more about hacking with Go +[223.58 --> 227.52] and cover all the things we did not manage to cover in our last episode. +[228.42 --> 229.90] Yes, very happy to be back here. +[230.78 --> 234.02] So for those who did not tune in to the previous episode, +[234.30 --> 235.88] could you shortly introduce yourself? +[236.72 --> 237.46] Yes, of course. +[237.66 --> 239.86] So my name is Ivan Kwiatkowski. +[240.12 --> 241.98] I am a French cybersecurity researcher. +[242.48 --> 244.14] And I work for Kaspersky. +[244.38 --> 247.04] Specifically, I work in a threat intelligence team. +[247.04 --> 252.18] And my role in this team, apart from writing reports and that kind of stuff, +[252.34 --> 258.22] is really to proceed with the reverse engineering of the malware that is provided to me by my coworkers. +[258.58 --> 260.04] So basically, they do the threat hunting. +[260.40 --> 263.18] They find some interesting stuff to look at. +[263.38 --> 266.90] They identify implants used by the attackers and they give them to me. +[266.90 --> 272.66] And then my job from there is using the fantastic tools of the reverse engineer, +[272.92 --> 274.24] which are Ida Pro, basically. +[274.68 --> 279.16] I do read the assembly code of those programs and try to figure out what they do as best as I can. +[279.64 --> 280.56] And that's my life. +[281.22 --> 288.30] And in the previous episode, we talked a little bit about how Ida is yes or no supporting Go, +[288.44 --> 292.96] so that it got better over time, but still a place to improve. +[293.56 --> 294.24] Yeah, absolutely. +[294.24 --> 301.16] So if you tried reverse engineering Go programs something like maybe two years ago, +[301.22 --> 306.62] maybe just one year ago, then you would be in a lot of trouble because the tools just weren't there. +[306.72 --> 313.86] So it involved using some third-party plugins, some codes taken from, I wouldn't say suspicious, +[314.14 --> 319.54] but GitHub repositories that weren't that well maintained or that didn't have really clear instructions. +[319.54 --> 324.60] And so it was really a difficult path for the reverse engineer that had to do it. +[325.12 --> 332.72] And thankfully, over the years, the developers from Ida, the company in Belgium called HexRace, +[332.92 --> 335.94] they've been listening to the customer complaints, I suppose. +[335.94 --> 344.68] And they have made a number of improvements that allow us to support Go programs a lot more easily. +[344.92 --> 352.64] And that entails having some recognition for the various functions that come from the Go standard library, +[353.08 --> 355.48] better support from the Go executables as a whole. +[355.48 --> 361.18] In the very latest versions, if I recall correctly, a number of the things that were actually implemented +[361.18 --> 365.72] in some of the third-party plugins that Juan Andres Guerrero, Satz from SentinelOne, +[365.82 --> 368.36] as well as myself, had implemented manually in Python. +[368.78 --> 374.20] Those kind of features, they included them into IdaPro, into the mainline. +[374.20 --> 379.82] So I do expect that one year from now, maybe this won't even be a discussion anymore. +[379.94 --> 384.36] I mean, sure, Go will still be a very kind of alien language for us to look at +[384.36 --> 387.00] in terms of the assembly code that it generates. +[387.40 --> 393.50] But at the end of the day, I do suspect that the tooling problems are going to be over in the near future. +[393.68 --> 395.68] So that's a good thing as far as we're concerned. +[396.58 --> 402.42] So is Go a better language for a security researcher to pick up or for a hacker? +[402.42 --> 410.28] Well, when it comes to security researchers, we don't actually have to write that many programs. +[410.60 --> 414.68] You know, most of the tools that we use are already provided to us by the community. +[415.02 --> 418.12] Like I mentioned IdaPro, nobody is ever going to redevelop IdaPro. +[418.22 --> 421.10] I mean, some guys did, so it wouldn't be fair saying that they wouldn't. +[421.18 --> 423.30] But most people are not going to do this. +[423.66 --> 428.82] And if they were going to do this, then I think that choosing either the Go language or C++ +[428.82 --> 432.72] wouldn't make that big a difference considering the scale of such a project. +[433.26 --> 440.30] When it comes to hackers, I think that for them, Go language is still probably a very good bet. +[440.46 --> 448.44] Because as far as I can tell from a very unscientific polling of my co-workers and other reverse engineers in the field, +[448.78 --> 452.50] it feels like most people still really dislike having to work on Go language. +[452.76 --> 454.00] Like really, really dislike it. +[454.00 --> 459.98] In my opinion, though, I think if I were to write malicious programs, I would use Rust. +[460.18 --> 463.38] Because the code generated by Rust is actually even way worse. +[463.82 --> 467.22] And at the moment, I'm not exactly sure how to approach that type of code. +[467.30 --> 468.24] I have to work on that. +[468.76 --> 471.72] But this would be my intuition to use Rust. +[471.84 --> 476.34] Because I know that I'm going to make someone's day miserable somewhere in the future. +[476.34 --> 482.68] And when you say worse, you mean basically because the steps that you need to do to reverse engineer +[482.68 --> 486.16] and kind of figure out what's happening are actually more painful. +[486.84 --> 487.24] Exactly. +[487.56 --> 490.62] I say that in the sense that maybe it's just a personal thing, right? +[490.68 --> 494.04] I have spent a bit of time trying to figure out how Go ticks. +[494.20 --> 496.76] Like not at the language level, but at least at the assembly level. +[497.26 --> 504.06] I would never pretend that I am a Go expert or that I have deep insights about the inner workings of the Go language. +[504.06 --> 510.78] But if you give me a binary written in Go, then I expect that eventually I will be able to tell you what it's supposed to be doing. +[511.16 --> 514.82] When it comes to Rust, it's kind of uncharted territory as far as I'm concerned. +[515.34 --> 519.40] As far as I know, again, this is not something that I've checked all that much. +[519.54 --> 522.22] Although this is something that I probably will have to do very soon. +[522.86 --> 528.56] It feels like the tooling, like IDA, for instance, the disassembler that we use on the job, +[528.56 --> 534.30] doesn't seem to support the Rust language as well as it supports the Go language. +[534.40 --> 535.98] It doesn't recognize as many things. +[536.68 --> 540.80] And at the end of the day, Rust tends to generate constructs that really look like C++. +[541.16 --> 543.98] And C++ is kind of a mess to begin with. +[544.40 --> 545.62] It's a very powerful language. +[545.76 --> 546.80] I really love C++. +[547.00 --> 549.78] If I have to write some complex program, I will write it in C++ +[549.78 --> 552.46] because this is the one I have the most experience with. +[552.88 --> 557.38] But when it comes to reading assembly written in C++, oh my god, it's just so convoluted. +[557.72 --> 560.82] And there are so many levels of indirection added at every level. +[561.04 --> 563.54] So this is not something I would be happy with. +[563.78 --> 570.46] And having Rust being a new, more complex or a new, more alien C++ is really not something I would be happy with. +[570.46 --> 573.96] How about the easy cross-compilation of Go? +[574.60 --> 579.78] The fact that you write your code and then you create one binary and then you kind of ship it as is. +[579.88 --> 584.22] And on top of that, you just write one more command and then you have it for any architecture. +[584.54 --> 586.06] Does this make any difference for you? +[586.36 --> 589.90] It does in the sense that I think as a developer, it's pretty cool to have that. +[589.90 --> 599.90] I also have always felt that this feature that is very often brought forward by defenders of the Go language was a bit... +[600.56 --> 602.96] I don't think it's that important, right? +[603.10 --> 607.28] Not that we don't want cross-compilation or that we don't want programs that can run anywhere. +[607.88 --> 613.02] But I don't feel like the Go language is especially adding something new there. +[613.16 --> 615.84] I mean, when I write C++ code, it can already run everywhere. +[616.10 --> 617.92] Provided that I write it properly, of course. +[617.92 --> 622.50] Way back when, probably 10 years ago when I was in school, I was learning Java. +[622.94 --> 625.50] And I actually programmed a few projects in Java code. +[626.10 --> 629.38] Supposedly, Java was supposed to be working on any platform as well, right? +[629.48 --> 633.00] So it didn't work as well as we expected due to many reasons. +[633.40 --> 646.10] But overall, it doesn't seem to me that this ability to run on any platform and to have a code that will compile everywhere is really something that Go is actually bringing to the table. +[646.10 --> 651.42] I think this is something that we already had and that maybe Go is making easier for a lot of developers. +[651.74 --> 656.26] But it's not something new and something that would make me switch languages by any stretch. +[656.84 --> 658.54] By the way, maybe this is a question for you, right? +[658.76 --> 659.10] How about... +[659.72 --> 662.80] I have no idea about the support for Go on exotic platforms. +[662.92 --> 667.32] What about running Go code for Solaris or for the ARM architecture? +[667.32 --> 669.80] Is that something that is supported out of the box or isn't it? +[669.96 --> 678.28] Because I know for a fact that when new CPUs come out, the first thing that the manufacturers release is going to be a C compiler or a C++ compiler. +[678.40 --> 682.04] So we know that eventually those languages are always going to work. +[682.16 --> 686.82] But I have the notion, and maybe I'm wrong about this, that when it comes to Go language, +[687.04 --> 692.08] that if you have this new platform somewhere, then you will have to wait for Google to release the corresponding compiler. +[692.08 --> 693.28] And that may take some time. +[693.72 --> 694.28] Is that correct? +[694.96 --> 700.94] So as you were asking, I googled the command of Go OS and Go Arch, which is what you run to do this. +[701.10 --> 703.64] And for Solaris, it does come out of the box. +[704.16 --> 705.40] What was the second one you asked for? +[705.78 --> 711.32] This is one that is going to be supported, I would imagine, would be to compile for ARM or MISP, +[711.48 --> 714.36] such probably less used architectures. +[714.46 --> 717.02] But I would imagine that at least ARM is supported very well. +[718.02 --> 721.06] Yes, ARM is out of the box indeed. +[721.06 --> 721.62] Yeah. +[721.90 --> 722.10] Okay. +[722.20 --> 723.54] ARM 64 and so on. +[723.66 --> 723.76] Yeah. +[724.04 --> 724.20] Right. +[724.74 --> 728.28] MIPS, some of the variations are not, but most of them, yes. +[728.56 --> 735.16] The ones that are not out of the box is MIPS 64P32 and MIPS 64P32LE. +[735.76 --> 735.98] Okay. +[736.18 --> 739.94] So yeah, overall, these are probably architectures that most people don't care about. +[740.36 --> 744.60] So I don't think this is a big fault on the part of the Go language. +[744.60 --> 751.34] What I will say, though, is that as far as I'm concerned, C is already a multi-platform language as it is. +[751.70 --> 754.66] And if other languages provide this as well, then good for them. +[754.76 --> 756.30] But to me, it's not something groundbreaking. +[757.06 --> 757.76] Yeah, that's fair. +[758.50 --> 764.90] Lots of DevOps people do love that feature that you don't need to do much to support, to ship everything to everyone. +[765.10 --> 765.54] Yeah. +[765.54 --> 766.88] In your framework architecture. +[766.88 --> 771.84] That aspect is pretty important and also very appreciated by malware authors. +[772.30 --> 782.66] It's the fact that, yeah, when you write some program in C, then you might have modules that are distributed in the form of a DLL file in Windows or .so shared object on Linux and so on. +[783.36 --> 790.00] And then you end up with a program, an executable file, and then several object libraries that come with it. +[790.00 --> 793.90] And then when you want to distribute it, you have to send this big archive that contains many files. +[794.22 --> 797.66] I would agree that when it comes to Go language, you end up with a single binary. +[798.36 --> 799.98] And that's pretty useful, right? +[800.26 --> 807.02] Especially in the context where you don't have control over the client, unquote, or victim's machine. +[807.50 --> 816.48] Then in that case, then, of course, just having to send an executable and knowing that it is self-contained and going to work everywhere all the time is going to be a big advantage. +[816.48 --> 819.92] And there are ways to do this in C, C++, and so on. +[820.54 --> 826.70] But it will require some work, which, agreed, is not required when you're using Go. +[826.82 --> 828.88] So this is one point for Go, I would say. +[829.38 --> 842.70] Another is the, tell me what you think, is the concept of modules, where it has this file that says what are all the dependencies and what version specifically is used where, in case you're using some package of an older version and whatnot. +[842.70 --> 855.34] So, given that this is all kind of compiled into a module and it's being sent out as one, does this provide any value for a hacker or for a security researcher? +[855.34 --> 860.00] Well, for a security researcher or a hacker, I'm not exactly sure. +[860.18 --> 871.40] I mean, as far as developers go, it prevents you from falling into this pit hole of dependency hell where, you know, this is stuff I actually experienced last week while working on a Python project. +[871.66 --> 876.16] I updated all my packages and some of them were not compatible with each other. +[876.22 --> 880.42] And my whole project just broke down in production, which is always fun. +[880.42 --> 886.96] So, this is something I could have prevented by just fixing the version numbers, which is what you're supposed to be doing. +[887.44 --> 890.88] But overall, like having this mechanism is kind of a good thing. +[891.18 --> 898.16] When it comes to security, making sure that hackers are able to compile their thing is really not something we worry about too much. +[898.70 --> 905.46] What we do worry about is the fact that when you end up with a single binary that contains everything, it's kind of an issue for reverse engineers. +[905.46 --> 916.02] Like when you compare this to a C program or C++ program that has some various DLLs, then the different files already represent some sort of separation between the code, right? +[916.08 --> 925.48] DLLs might or might not have a relevant name, but at the end of the day, you know that they're going to be split according to some form of functionality, right? +[925.56 --> 931.94] This type of code is going to go in this DLL, like the main intelligence of the program will go in the main executables and so on. +[931.94 --> 939.80] So, when you have those big malware platforms that you have to work on, then having several files is actually a pretty good thing for us. +[940.18 --> 952.42] When you have a big Go binary that is 5 to 10 megabytes big, and then you have to just dive in there and try to find out where the interesting code is located, it's a good thing. +[952.42 --> 958.36] You know where the interesting code, which is the library code is, that's something that the tools can recognize pretty well. +[958.74 --> 962.54] But still, you just still have this big program that contains everything. +[963.12 --> 973.92] And it's just much easier if you already have this kind of separation where you can already focus on some specific functionality, even though you haven't been able to dig into the whole project. +[973.92 --> 982.02] So, in that sense, I would say that this feature is pretty useful for the offensive side, I would say. +[982.44 --> 986.38] I wonder if there's any particular feature that is good for the defensive side. +[986.56 --> 988.86] But I'll keep asking questions until we find something. +[988.98 --> 989.98] Or do you have one in mind? +[989.98 --> 1009.56] I do have one in mind, like the best and I think one of the strongest selling points of Go over C and C++ and all those unmanaged languages is going to be that when you write programs in Go, as a developer, I mean, you know that you're never going to have any problems with memory corruption, buffer overflows and those kinds of issues. +[1009.62 --> 1016.16] I would be very surprised if the Go language would allow you to read outside the bounds of an array and that kind of stuff. +[1016.16 --> 1019.26] So, all this is already taken into account for you. +[1019.80 --> 1024.48] And it's not going to make my personal daily job as a reverse engineer easier. +[1024.70 --> 1031.94] But what it's going to provide is that, by default, people will have a much harder time shooting themselves in the foot. +[1032.48 --> 1039.28] And this is a good thing overall because it means that if I download an application that was written by someone else in Go language, +[1039.28 --> 1053.60] and I don't have to worry as much about the code quality because I know that the language is going to provide a number of guarantees and that will make sure that at least a number of vulnerabilities are not going to affect me ever. +[1053.60 --> 1059.06] This is true as a whole for the whole industry, even outside security. +[1059.74 --> 1067.14] If programs are written in Go language, like FTP server or email server or whatever, if such programs are written in Go language, +[1067.24 --> 1070.36] then we know that at least we won't have to worry about buffer overflows. +[1070.66 --> 1081.38] And this means less weekends spent in incident response engagements because some customers didn't patch their program or because there was some vulnerability discovered as a zero day +[1081.38 --> 1089.72] and that's being exploited in the wild for an application that has some buffer overflow vulnerability and that is just available widely on the Internet. +[1089.72 --> 1100.08] So overall, for defenders, less vulnerabilities and less ways for developers to make tragic, tragic mistakes is always going to be a good thing. +[1100.20 --> 1108.62] And I think this, in fact, overshadows any advantage that the attackers are gaining over us on the personal level with the reverse engineers. +[1108.80 --> 1110.48] That is a very interesting point. +[1110.48 --> 1117.76] I guess it would be interesting to see if overall Go code is more secure, however you would measure that. +[1117.98 --> 1120.06] It's probably going to be interesting to see how to do that. +[1120.62 --> 1124.92] Well, having precise metrics is always going to be difficult. +[1125.08 --> 1132.78] But if you compare CVE number quantities for projects written in Go language and for projects written in C, +[1132.78 --> 1150.02] then I think it's very likely that you would find that programs written in memory unmanaged languages like C++ and the like are always going to have more bugs just because, like, by default, there are more opportunities to shoot yourself in the foot. +[1150.02 --> 1161.58] So if you take the developers with equivalent skill and for one of them, some bugs are just unavailable and for the other one, then you have twice as much mistakes that you can make. +[1161.58 --> 1169.00] It feels very obvious to me that no matter how, like, if the developers have the same level of skill, +[1169.36 --> 1174.40] then the person using the unsafe language is always going to make more mistakes. +[1174.40 --> 1188.78] Hey, friends. +[1188.90 --> 1194.56] This episode is brought to you by my friends and potentially your friends, too, at FireHydrant. +[1194.74 --> 1198.66] And I'm here with Robert Ross, founder and CEO of FireHydrant. +[1198.66 --> 1202.62] And Robert, there are several options out there for incident management. +[1202.96 --> 1205.34] But what is it that makes FireHydrant different? +[1205.88 --> 1211.70] The reason that we think that FireHydrant is is on to something is because we're meeting companies really where they are. +[1211.96 --> 1218.94] We face the same problems that every company in the industry that is building and releasing software is also facing. +[1218.94 --> 1232.66] So we want people to be able to sign up for FireHydrant and immediately be able to kick off an incident using the best practices that we've built and we've experienced and have gathered through the other amazing customers that use our tool. +[1232.82 --> 1234.92] It really is a very quick time to value. +[1235.18 --> 1241.24] And we want people to have a long jump from where they are to where they want to be in incident management. +[1241.24 --> 1242.72] I love it. Thank you, Robert. +[1243.00 --> 1248.12] Small teams up to 10 people can get started for free with all FireHydrant features included. +[1248.24 --> 1249.98] There's no credit card required to sign up. +[1250.32 --> 1252.00] They are making it too easy to get started. +[1252.20 --> 1254.24] So check them out at FireHydrant.com. +[1255.02 --> 1257.50] Again, FireHydrant.com. +[1271.24 --> 1280.20] Have you ever seen malware in COBOL? +[1280.72 --> 1281.78] I haven't, actually. +[1282.30 --> 1285.96] I wonder how would you evaluate that and the more safer and the less safe side. +[1286.56 --> 1287.50] Now that you say all that. +[1287.86 --> 1289.16] I actually have no idea. +[1289.32 --> 1291.34] Like, COBOL is one of those languages that I know of. +[1291.70 --> 1299.98] I know that if you want to work in banks and be paid the big bucks, then you should definitely learn COBOL because all the former COBOL developers, +[1299.98 --> 1301.46] they died of old age by now. +[1301.66 --> 1303.98] So, you know, they are kind of hard to find. +[1304.88 --> 1311.68] Beyond this, never in my life have I heard about, or like never in the, maybe the last 10 years have I heard about malware written in COBOL. +[1311.92 --> 1321.66] If I were to find one, then it would be a pretty cool blog article, but it would probably be a miserable week for me because then I would have to probably learn the language and figure out how it works. +[1322.26 --> 1327.32] Although, like, to be fair, this week, I actually had to reverse engineer a program that was written in Pascal. +[1327.54 --> 1329.80] Which is the spiritual parent of Go. +[1329.98 --> 1331.24] Yeah, I suppose. +[1331.36 --> 1336.36] And also a spiritual parent of many other languages because like it came from the 80s, maybe the 70s, right? +[1336.50 --> 1343.44] I remember learning Pascal or at least the basics when I was in school, something like 15 years ago now. +[1343.94 --> 1344.22] Same. +[1344.36 --> 1349.00] It was my high school graduation project had to be in Pascal. +[1349.42 --> 1350.04] So there you go. +[1350.36 --> 1352.46] Like, I probably wrote a little bit of Pascal myself. +[1352.66 --> 1355.62] Probably never actual projects that did anything meaningful. +[1355.62 --> 1360.06] But I did have to look at some APT malware, like real life APT. +[1360.20 --> 1361.28] What does APT stand for? +[1361.68 --> 1362.14] Oh, sorry. +[1362.22 --> 1364.38] It means advanced persistent threat. +[1364.72 --> 1370.52] So APTs are one of the categories of hackers that we track in our daily work. +[1370.52 --> 1376.48] So you have on one end the attackers that are financially motivated, cybercrime, ransomware groups, and all the like. +[1377.16 --> 1389.18] And on the other end of the spectrum, you have what we call the APTs, which basically are the state-sponsored actors or the mercenary actors, all the groups that are focused on cyber espionage. +[1389.18 --> 1395.06] Now, initially, the name APT was, I think, proposed by a mendiant. +[1395.40 --> 1398.58] It's probably around 2010 or something like this with their first report. +[1398.68 --> 1406.24] And back then, I think it made sense to say, you know, on one end, we have the low-skilled cybercriminals that are doing run-of-the-mill crimeware. +[1406.46 --> 1411.38] And on the other end of the spectrum, you have those state-sponsored attackers that are doing very sophisticated things. +[1411.38 --> 1417.56] I think that today, trying to separate attackers between levels of sophistication doesn't make that much sense anymore. +[1417.68 --> 1423.72] We do have extremely skilled ransomware groups that use very cutting-edge pen-testing methodologies. +[1423.80 --> 1426.24] And we do have APTs that are extremely bad. +[1426.46 --> 1428.30] I would say they have a poor OPSEC. +[1428.52 --> 1430.20] They don't know how to use their tools and so on. +[1430.32 --> 1436.44] So at this stage, I think, in 2022, when you hear APT, you just have to think about espionage. +[1436.60 --> 1439.18] I think this is going to be the way to understand this. +[1439.18 --> 1450.52] But in any case, yeah, this week I was working on another APT case taking place in some of the STAN countries in the CIS, Commonwealth Independent States, I think. +[1450.98 --> 1455.44] Anyway, and one of the malware implants that we found there was actually written in Pascal. +[1455.74 --> 1461.12] And so it was kind of a trip down the memory lane to, one, figure out what Pascal was again. +[1461.48 --> 1466.44] And also trying to understand what kind of assembly was generated by this Pascal compiler. +[1466.44 --> 1469.94] It wasn't that bad, actually, way less bad than having to discover the Go language. +[1470.66 --> 1478.10] That's interesting to know, because there is a lot of similarity between Go and Pascal, but knowing that the translation of that is different. +[1478.76 --> 1479.04] Oh, yeah. +[1479.08 --> 1485.70] I can tell you, even though I'm not an expert in any of the language, that even though you might have some similarities on the code level, +[1485.92 --> 1488.70] and maybe the constructs and the way that you declare things and so on, +[1489.14 --> 1492.96] when it comes to assembly, like the languages could not be more different from one another. +[1492.96 --> 1498.82] I thought that if they're conceptually similar, they might have a similar structure, but I guess not. +[1499.10 --> 1503.64] No, I suppose they took inspiration when it comes to how to write the code. +[1504.06 --> 1508.88] But then when it comes to what the compiler does, then yeah, the Go compiler really does its own thing. +[1509.22 --> 1513.80] And when you say that each language has their own different thing in the assembly representation, +[1514.12 --> 1517.58] or even when you reverse engineer that into like a visual representation, +[1517.86 --> 1519.74] how many different ways can you have? +[1519.74 --> 1522.82] Is it really, can it be like every time completely different? +[1523.26 --> 1527.78] It's not always completely, completely different, but there are some meaningful differences. +[1528.14 --> 1536.30] I would say that C language, maybe it's a misconception I have because C language is traditionally what you learn reverse engineering on. +[1536.72 --> 1540.38] So C language to me is going to be the language that is closest to the CPU. +[1540.38 --> 1547.96] When you compile a C program, then there's going to be a kind of direct translation from your C code to the assembly language. +[1548.80 --> 1551.84] Like the compiler isn't going to be too smart about things. +[1552.18 --> 1557.92] When you do something in C, then when you write it in code, then it kind of shows in the assembly language. +[1558.06 --> 1563.28] Of course, you can have some compiler optimizations for speed, for space and so on. +[1563.28 --> 1570.20] But overall, the translation is going to be pretty, I wouldn't say it's easily reversible, but I think it's pretty direct. +[1570.50 --> 1573.88] Like you really find your bearings from the C code to the assembly language. +[1574.20 --> 1583.14] And I think it's actually not that much of a surprise that decompiler tools such as Ghidra or such as the one sold with Ida Pro +[1583.14 --> 1588.22] do take the assembly and convert it back to C language because I think this is the closest. +[1588.22 --> 1595.24] And then when you go to languages that have, I would say, higher levels of conceptual complexity, +[1596.06 --> 1599.88] then this is where the compiler starts doing a lot of things on its own. +[1600.26 --> 1605.08] And this is where the code that you write ends up being super different from the assembly that you read. +[1605.44 --> 1610.52] When it comes to C++ and when you use an STD string, it seems like a very simple thing, right? +[1610.52 --> 1619.74] But under the hood, the STD string class is actually a template instantiation of like a very, very complex series of nested templates. +[1620.24 --> 1623.28] And you end up with, how can I explain this? +[1623.36 --> 1630.24] You end up with a weird structure that has a first a table that contains a pointer to methods, +[1630.56 --> 1632.00] which is something you never wrote in C. +[1632.30 --> 1637.68] You end up with methods calling each other nested methods that come from the template library, +[1637.68 --> 1640.06] from the C++ standards library, and so on. +[1640.12 --> 1641.58] And things just get crazy from here. +[1642.06 --> 1646.02] Taking the example of Pascal, and again, I don't write that much Pascal code, +[1646.10 --> 1649.40] but it's very obvious to me that when I look at the assembly code +[1649.40 --> 1653.18] and I see reference counters being incremented and decremented automatically, +[1653.74 --> 1654.68] and all that kind of stuff, +[1654.74 --> 1657.40] then this is something that was automatically added by the compiler. +[1657.90 --> 1661.60] And it's, I guess, useful as far as the running the program goes. +[1661.72 --> 1664.00] But when it comes to me understanding the program, +[1664.00 --> 1667.38] making sure that reference counts are handled properly +[1667.38 --> 1671.26] and that the objects are going to be freed when there are no more references to them +[1671.26 --> 1672.74] is something that I really don't care about. +[1672.86 --> 1676.40] And it's just cluttering my window, really. +[1676.52 --> 1679.64] It's just code added by the compiler that has no meaning, +[1679.90 --> 1681.94] at least when it comes to what the program is supposed to be doing. +[1682.00 --> 1684.88] It doesn't add any intelligence to the program. +[1684.96 --> 1686.12] It's just something that gets in the way. +[1686.94 --> 1690.42] And Go is probably one of the far extremes of this, +[1690.42 --> 1696.30] right there with C++ because the Go compiler is really doing a lot of stuff under the hood. +[1697.30 --> 1700.10] Like it really, how can I put this? +[1701.04 --> 1701.48] Optimizations? +[1702.06 --> 1703.02] Yeah, optimizations. +[1703.26 --> 1707.42] It inlines anything that is not worth a function call. +[1708.00 --> 1710.54] The calling convention is its own thing. +[1711.08 --> 1715.06] It's really, oh, you also have a garbage collection mechanism. +[1715.06 --> 1718.50] When you write a simple hello world program in Go language, +[1718.62 --> 1722.38] it ends up being an executable that's something like one megabyte big or something like this. +[1722.70 --> 1726.70] Like I get that today storage space isn't that expensive. +[1726.90 --> 1728.62] Like we don't care about one megabyte of code. +[1729.10 --> 1731.46] Like what's the difference between this and seven kilobytes? +[1731.54 --> 1733.88] I think not that much when it comes to hard drives. +[1734.64 --> 1737.82] But when you are a reverse engineer and you have to look at one megabyte of code +[1737.82 --> 1741.46] instead of seven kilobytes and it's actually a big deal, right? +[1741.46 --> 1745.50] And this is the kind of thing that the Go compiler does to you, +[1745.96 --> 1747.94] along with a number of optimizations, +[1748.52 --> 1752.70] along with this very weird calling convention that they have, etc. +[1752.96 --> 1754.44] Another thing that I don't like about Go, +[1754.58 --> 1755.62] I mean, it's a good thing, right? +[1755.66 --> 1758.42] I don't like it as a reverse engineer, is the Go routines. +[1759.06 --> 1760.84] Yes, this would have been my next question. +[1761.06 --> 1762.28] Yes, please do elaborate. +[1762.96 --> 1763.40] Great. +[1763.50 --> 1768.00] So it seems like a very, very easy way to create threaded programs, +[1768.00 --> 1772.14] as far as I understand, which is great as far as developers go. +[1772.46 --> 1776.14] But it makes it a bit too easy for malware developers to create threaded programs as well. +[1776.76 --> 1779.46] And when it comes to understanding what a program does, +[1779.56 --> 1781.02] we really like linear programs. +[1781.16 --> 1784.36] We want instructions that we can look at one after the other. +[1784.54 --> 1786.50] We want programs that we can debug very easily. +[1787.14 --> 1790.48] And as soon as many things start happening in many threads, +[1790.48 --> 1794.18] then, oh my God, following things around becomes extremely, extremely difficult. +[1794.18 --> 1799.28] So I would actually like for threaded programs to be more difficult to write +[1799.28 --> 1803.10] and to be less available for attackers, if that were an option. +[1803.44 --> 1808.94] How do threads represent themselves visually when you do reverse engineering? +[1809.42 --> 1813.44] Well, they don't really represent themselves because threads as a concept, +[1813.60 --> 1816.58] they are a fundamentally runtime object, right? +[1816.86 --> 1821.26] A thread is going to be a unit of execution that is going to run some code. +[1821.26 --> 1825.00] And when you have a single thread, which is the case for a lot of programs, +[1825.20 --> 1828.20] then you can just follow what is going on in the code linearly. +[1828.38 --> 1829.82] And then you figure out what is going on. +[1829.88 --> 1833.98] When you have several threads, then this is really an order of magnitude, +[1833.98 --> 1839.90] more of complexity that you have to wade through in the sense that as a reverse engineer, +[1840.04 --> 1843.50] you don't only have to think about what is going on in the program +[1843.50 --> 1845.06] or in the function that you are reading, +[1845.18 --> 1848.32] but at all times you have to think about the fact that there might be +[1848.32 --> 1851.22] another thread running somewhere that might be doing things +[1851.22 --> 1853.60] that are affecting what you are doing right now +[1853.60 --> 1855.96] or what the current function that you are reading is doing. +[1856.12 --> 1859.84] And so you do not have the luxury of, I think, +[1859.90 --> 1861.84] all the information that you need in a single place. +[1862.14 --> 1864.62] The functionality, the intelligence of the program +[1864.62 --> 1867.58] ends up being spread over different units of execution. +[1868.44 --> 1870.68] And you have to keep everything in your head +[1870.68 --> 1873.46] to have any hope of understanding what is going on. +[1873.46 --> 1878.38] So this is really a very heavy mental tax that is imposed on the reverse engineer. +[1878.90 --> 1880.82] And of course, the more threats there are, +[1880.84 --> 1883.52] then the more effort that you have to go through +[1883.52 --> 1886.02] to try to keep track of everything that is going on. +[1886.58 --> 1889.62] But one good example of this is a Go program +[1889.62 --> 1891.22] that I mentioned in the previous podcast. +[1891.52 --> 1892.76] It's called Stow Away. +[1893.30 --> 1895.82] And this is a open source project +[1895.82 --> 1900.16] that is used to do various proxying operations +[1900.16 --> 1902.34] as a pen testing tool, right? +[1902.36 --> 1904.78] You can create tunnels, SOX proxies, et cetera, +[1904.90 --> 1906.22] and probably pipe them with each other. +[1906.28 --> 1907.02] I'm not exactly sure. +[1907.50 --> 1908.98] But what I'm sure of is that +[1908.98 --> 1910.76] when I was reading the program assembly, +[1911.34 --> 1912.86] it felt very, very miserable +[1912.86 --> 1914.76] because it was obvious that many things +[1914.76 --> 1916.06] were happening at the same time, +[1916.36 --> 1917.74] which, of course, is going to be the case +[1917.74 --> 1919.98] because when you have some network program, +[1920.10 --> 1923.04] then packets can arrive from any end of the various tunnels. +[1923.04 --> 1925.30] And you can also have many tunnels running at the same time. +[1925.64 --> 1928.88] So you have all these things taking place at the same time +[1928.88 --> 1932.44] and trying to figure out exactly what does what +[1932.44 --> 1933.66] is extremely difficult. +[1933.88 --> 1936.66] And if I hadn't been able to figure out +[1936.66 --> 1938.42] that this was actually an open source project +[1938.42 --> 1940.42] for which I was able to go find the code, +[1940.92 --> 1943.76] then probably I might not have been able to figure out +[1943.76 --> 1945.50] everything that the program was doing at all +[1945.50 --> 1947.40] because there was just too much to work through +[1947.40 --> 1949.36] and too much to remember +[1949.36 --> 1952.08] because my memory is actually quite limited, +[1952.52 --> 1954.30] as is the memory of any human being +[1954.30 --> 1955.52] compared to a computer, really. +[1955.94 --> 1957.06] Until we start upgrading. +[1957.64 --> 1958.70] Yeah, I wish. +[1958.88 --> 1961.90] All the interesting terms that you're mentioning, +[1962.10 --> 1964.14] like APT and also Stowaway, +[1964.20 --> 1965.32] which you're reminded now, +[1965.66 --> 1967.88] it will be in the show notes +[1967.88 --> 1970.26] for those who want to look back at that +[1970.26 --> 1974.16] and see how can they handle threads +[1974.16 --> 1975.34] in reverse engineering. +[1975.54 --> 1975.66] Yeah. +[1975.74 --> 1978.38] It will be interesting to see some example +[1978.38 --> 1980.06] of how that actually looks +[1980.06 --> 1982.04] because you said that this kind of can +[1982.04 --> 1984.06] tint the results and whatnot. +[1984.28 --> 1987.32] If you maybe like there's data dependency +[1987.32 --> 1988.06] between the two, +[1988.14 --> 1990.30] but I still try to visualize and understand. +[1990.82 --> 1992.40] So I'm referring back to the video +[1992.40 --> 1994.10] that you have published on YouTube, +[1994.10 --> 1995.04] which I will also link +[1995.04 --> 1997.30] of your reverse engineering AGO program. +[1997.30 --> 1999.50] And you kind of build this block diagram +[1999.50 --> 2000.64] of the different steps +[2000.64 --> 2002.18] and you write what's happening there, +[2002.26 --> 2003.34] what you guess and so on. +[2003.34 --> 2005.50] So affecting the results +[2005.50 --> 2006.60] by sharing some data +[2006.60 --> 2008.30] or do a calculation return +[2008.30 --> 2009.26] and then on top of that, +[2009.58 --> 2010.86] that all makes sense. +[2011.40 --> 2013.54] Do commands just pop up randomly +[2013.54 --> 2015.32] in what's happening now +[2015.32 --> 2017.60] and you kind of try to paint that +[2017.60 --> 2018.74] in the relevant context +[2018.74 --> 2020.18] once you're reverse engineering +[2020.18 --> 2022.20] something with multi-threads? +[2022.20 --> 2023.78] Actually, I think the best way +[2023.78 --> 2024.62] to visualize it +[2024.62 --> 2026.42] is not to try to think about the program +[2026.42 --> 2027.14] as it is running, +[2027.26 --> 2030.50] but imagine that instead of reading assembly code, +[2030.58 --> 2032.06] you are reading the source code +[2032.06 --> 2032.96] of a Go program. +[2033.42 --> 2036.10] I think you are not going to dispute the fact +[2036.10 --> 2038.90] that if you receive some Go project +[2038.90 --> 2040.58] from a friend or a coworker +[2040.58 --> 2042.30] and you know nothing about this project +[2042.30 --> 2044.26] and you have to read this source code, +[2044.68 --> 2045.86] then this source code is going to be +[2045.86 --> 2047.12] much easier to understand +[2047.12 --> 2050.46] if the program is just a single thread +[2050.46 --> 2052.02] that is doing a single thing, right? +[2052.16 --> 2053.58] If you receive this program +[2053.58 --> 2055.10] that as soon as it begins, +[2055.16 --> 2056.66] it launches three different threads +[2056.66 --> 2058.50] that are supposed to do different things +[2058.50 --> 2059.32] at the same time +[2059.32 --> 2061.34] and figuring out exactly +[2061.34 --> 2063.60] what the program is going to be doing +[2063.60 --> 2064.76] is going to be, I think, +[2064.80 --> 2065.60] much more difficult. +[2066.48 --> 2068.20] Now, imagine the same thing, +[2068.32 --> 2070.96] but instead of receiving proper Go code, +[2071.36 --> 2073.02] then you would receive Go code +[2073.02 --> 2075.12] where all the variable names have been wiped +[2075.12 --> 2076.30] and all the variables are named +[2076.30 --> 2077.56] A, B and C, et cetera. +[2077.66 --> 2079.60] So you cannot even use the function names +[2079.60 --> 2081.16] or you cannot even use the variable names +[2081.16 --> 2082.98] to try to understand +[2082.98 --> 2084.44] what the program is supposed to do. +[2084.66 --> 2086.18] Imagine that you have no comments +[2086.18 --> 2087.24] inside the code as well. +[2087.68 --> 2089.68] This is basically what reverse engineers have. +[2089.68 --> 2091.18] You don't always have to imagine that. +[2091.46 --> 2092.14] Yeah, of course. +[2092.98 --> 2094.90] Yeah, this might be real life +[2094.90 --> 2096.44] for a lot of people out there. +[2097.00 --> 2097.96] Shout out to them, I guess. +[2098.26 --> 2101.48] But this is exactly what reverse engineering is. +[2101.66 --> 2102.86] Like you receive some source code, +[2103.04 --> 2104.78] whether it is assembly or high-level. +[2105.26 --> 2106.20] A high-level code is, +[2106.60 --> 2107.52] that does a bit of difference +[2107.52 --> 2109.00] because assembly is hard to read. +[2109.06 --> 2109.56] But basically, +[2110.28 --> 2111.38] it's going to be the same thing +[2111.38 --> 2112.40] that you have to go through, right? +[2112.44 --> 2113.30] You receive some code, +[2113.52 --> 2114.88] you have to understand what it does. +[2115.34 --> 2117.66] And the more complex this code is, +[2117.80 --> 2120.00] the more sophisticated its operations are, +[2120.08 --> 2121.72] then the harder time you're going to have +[2121.72 --> 2123.96] to understand exactly what is going on in there. +[2124.48 --> 2126.26] To this, it brings me to the next question +[2126.26 --> 2129.96] that generally Go best practice, let's say, +[2130.16 --> 2132.66] or like the right way to do Go +[2132.66 --> 2135.08] is to write simple, readable code +[2135.08 --> 2136.82] rather than sophisticated +[2136.82 --> 2139.00] and like trinary operations and whatnot +[2139.00 --> 2141.38] and complicated things. +[2141.64 --> 2144.86] Does this in any way help or not? +[2145.22 --> 2146.14] I wish, +[2146.30 --> 2148.98] but it turns out that for the compiler, +[2149.30 --> 2151.30] whether you write the simple way +[2151.30 --> 2152.68] or the trinary operation, +[2152.90 --> 2154.20] if the compiler is smart enough, +[2154.26 --> 2155.06] then at the end of the day, +[2155.06 --> 2155.98] it's going to generate +[2155.98 --> 2157.88] the very same assembly construct. +[2157.88 --> 2160.30] Hopefully, the compiler, +[2160.56 --> 2162.84] if you do this, if, then, else, +[2163.32 --> 2165.30] or if you use the trinary operator, +[2165.98 --> 2167.18] it's going to be able to recognize +[2167.18 --> 2168.06] that it's the same thing +[2168.06 --> 2168.52] and in the end, +[2168.58 --> 2169.66] it's going to generate assembly +[2169.66 --> 2170.94] that does exactly the same. +[2171.08 --> 2172.56] So it's a good thing +[2172.56 --> 2175.18] as far as development practices are concerned. +[2175.58 --> 2178.06] But when you reach the assembly level, +[2178.26 --> 2180.52] then all those helpful things +[2180.52 --> 2181.44] and all those precautions +[2181.44 --> 2182.36] that you have taken +[2182.36 --> 2184.02] to make sure that other people +[2184.02 --> 2185.98] will be able to understand +[2185.98 --> 2186.60] what you are doing, +[2186.60 --> 2188.52] they just get taken out by the compiler +[2188.52 --> 2190.24] because there are things for humans, +[2190.34 --> 2191.42] there are no things for CPUs, +[2191.44 --> 2192.66] and so they have no place +[2192.66 --> 2193.84] in your compiler program. +[2194.34 --> 2195.32] Yeah, that's a great point. +[2195.78 --> 2197.10] Actually, one callback +[2197.10 --> 2198.98] from one question you asked earlier, +[2199.20 --> 2200.86] you asked how many different things +[2200.86 --> 2201.80] can a compiler do +[2201.80 --> 2202.84] when it comes to different languages? +[2203.52 --> 2204.46] An additional example +[2204.46 --> 2205.22] I can give for Go +[2205.22 --> 2207.38] is Go programs, Go functions, +[2207.50 --> 2208.86] they can return any number +[2208.86 --> 2210.78] of return values, right? +[2211.36 --> 2212.34] This is not something +[2212.34 --> 2214.00] that most languages are doing. +[2214.20 --> 2214.98] And so when you look +[2214.98 --> 2215.94] at the assembly language, +[2215.94 --> 2217.32] then at the end of the day, +[2217.42 --> 2219.00] it turns out to be translated +[2219.00 --> 2220.24] in CPU code +[2220.24 --> 2222.32] in very different ways +[2222.32 --> 2223.88] than normal functions +[2223.88 --> 2225.48] are supposed to work, right? +[2225.68 --> 2226.56] When you have a function +[2226.56 --> 2227.70] that can only return +[2227.70 --> 2228.58] single arguments, +[2228.82 --> 2229.74] which is the case +[2229.74 --> 2231.28] for a lot of languages out there, +[2231.36 --> 2232.50] like CC++, etc., +[2232.50 --> 2234.06] then you have a very simple convention. +[2234.20 --> 2235.00] The convention says, +[2235.18 --> 2236.24] oh, the return value +[2236.24 --> 2237.78] will be in register EAX +[2237.78 --> 2238.30] in assembly. +[2238.42 --> 2239.90] This is the rule. +[2240.26 --> 2240.94] It's very simple. +[2241.48 --> 2242.64] When it comes to Go language, +[2242.82 --> 2244.36] then you don't have +[2244.36 --> 2245.10] a single place +[2245.10 --> 2245.88] because you can have +[2245.88 --> 2246.62] several arguments. +[2246.78 --> 2247.64] And so they are returned +[2247.64 --> 2249.48] differently through the stack +[2249.48 --> 2250.30] and you have to go +[2250.30 --> 2251.00] look for it there. +[2251.08 --> 2252.70] It's just much more complex +[2252.70 --> 2253.78] and it's very different +[2253.78 --> 2255.24] from a traditional language. +[2255.80 --> 2256.88] And the difference +[2256.88 --> 2257.64] between the languages +[2257.64 --> 2258.24] are going to be +[2258.24 --> 2259.60] small things like this +[2259.60 --> 2261.06] when it comes to conventions, +[2261.26 --> 2262.38] but actually, +[2262.66 --> 2263.86] some of all of them +[2263.86 --> 2264.84] are going to result +[2264.84 --> 2265.98] in having a source code +[2265.98 --> 2267.08] or an assembly code +[2267.08 --> 2268.10] at the end of the day +[2268.10 --> 2269.26] that is really +[2269.26 --> 2270.56] extremely different +[2270.56 --> 2271.32] from one another. +[2271.32 --> 2272.54] Would you say that +[2272.54 --> 2272.98] the way, +[2273.18 --> 2273.88] from what you see, +[2273.98 --> 2275.50] the way that Go handles +[2275.50 --> 2276.94] on the compiler level, +[2277.04 --> 2278.86] the return of multiple arguments +[2278.86 --> 2280.46] is efficient overall? +[2280.92 --> 2281.16] Yeah, +[2281.26 --> 2282.50] it really feels like it +[2282.50 --> 2283.54] as far as I recall. +[2283.90 --> 2284.56] Like maybe this is +[2284.56 --> 2285.64] me making a mistake, +[2286.04 --> 2287.22] but it feels like +[2287.22 --> 2288.10] the return values +[2288.10 --> 2289.26] from one function call +[2289.26 --> 2290.26] are placed exactly +[2290.26 --> 2290.94] where they should be +[2290.94 --> 2291.42] on the stack +[2291.42 --> 2292.98] so that another function +[2292.98 --> 2293.90] can use them +[2293.90 --> 2294.88] as arguments immediately. +[2295.42 --> 2296.56] And so chain the calls +[2296.56 --> 2297.80] between different functions +[2297.80 --> 2298.28] when you have +[2298.28 --> 2299.88] a function calling. +[2299.88 --> 2301.16] When you call a function +[2301.16 --> 2302.42] by passing an argument +[2302.42 --> 2303.74] which is the result +[2303.74 --> 2304.72] of another function call, +[2304.96 --> 2305.50] it feels like +[2305.50 --> 2306.66] in assembly +[2306.66 --> 2308.18] that these function calls +[2308.18 --> 2308.76] are going to be +[2308.76 --> 2310.24] very close to each other. +[2310.32 --> 2311.08] You won't have to +[2311.08 --> 2312.38] move back stuff +[2312.38 --> 2313.92] from the return values +[2313.92 --> 2314.72] back on the stack, +[2314.84 --> 2315.02] et cetera, +[2315.12 --> 2316.44] just already there. +[2316.98 --> 2317.58] And I think that +[2317.58 --> 2318.44] on that level, +[2318.88 --> 2320.06] they're going to +[2320.06 --> 2321.24] be pretty efficient +[2321.24 --> 2322.40] and to go pretty fast. +[2323.34 --> 2323.56] Nice. +[2324.10 --> 2324.64] Good to know. +[2324.80 --> 2325.68] It's always encouraging. +[2325.68 --> 2329.74] So Go is kind of built +[2329.74 --> 2330.66] in a way that you don't +[2330.66 --> 2332.26] debug this line by line +[2332.26 --> 2333.26] with breakpoints +[2333.26 --> 2333.78] and so on +[2333.78 --> 2334.38] as you do in many +[2334.38 --> 2334.96] other languages, +[2334.96 --> 2336.16] but you do something +[2336.16 --> 2337.24] and you check for errors +[2337.24 --> 2337.90] all the time. +[2338.46 --> 2339.30] I will not ask you +[2339.30 --> 2340.32] whether malware +[2340.32 --> 2341.46] is generally written +[2341.46 --> 2342.42] like this or not +[2342.42 --> 2343.34] or how good are they +[2343.34 --> 2344.74] with their error catching +[2344.74 --> 2345.98] unless you know +[2345.98 --> 2346.96] and then please do share. +[2347.40 --> 2348.40] I do know actually +[2348.40 --> 2349.90] because I end up +[2349.90 --> 2350.54] reading the code, +[2350.74 --> 2350.94] right? +[2350.94 --> 2353.42] So what I learned +[2353.42 --> 2354.56] about Go +[2354.56 --> 2355.76] by reading assembly code +[2355.76 --> 2357.44] and also by trying it myself +[2357.44 --> 2358.00] and to understand +[2358.00 --> 2358.68] what was going on +[2358.68 --> 2359.42] is the fact that +[2359.42 --> 2361.12] the Go language +[2361.12 --> 2362.66] will not allow you +[2362.66 --> 2364.12] to not catch the errors, +[2364.34 --> 2364.58] right? +[2364.94 --> 2365.58] If you have a function +[2365.58 --> 2366.96] that returns to return values +[2366.96 --> 2368.86] and if you do not catch them, +[2368.96 --> 2369.36] then you're going +[2369.36 --> 2369.88] to be in trouble, +[2370.00 --> 2370.14] right? +[2370.18 --> 2371.14] The code is not going +[2371.14 --> 2371.60] to compile. +[2372.32 --> 2374.24] So I think you can probably +[2374.24 --> 2375.92] create this underscore variable +[2375.92 --> 2376.90] that means I don't care. +[2377.02 --> 2377.30] Yes. +[2377.50 --> 2377.68] Right? +[2378.50 --> 2380.00] But as far as I can tell, +[2380.32 --> 2380.92] malware developers +[2380.92 --> 2381.72] at least for the programs +[2381.72 --> 2382.26] that I've seen, +[2382.46 --> 2384.14] they do catch the errors +[2384.14 --> 2385.32] and they check for the errors +[2385.32 --> 2386.72] and they handle them properly, +[2387.28 --> 2388.60] which I think makes sense, +[2388.70 --> 2388.90] right? +[2388.92 --> 2389.96] Because if the language +[2389.96 --> 2391.22] forces you to do it, +[2391.62 --> 2392.58] then you're going to do it. +[2392.64 --> 2393.08] Like, of course, +[2393.10 --> 2394.30] you can circumvent this +[2394.30 --> 2395.24] by, you know, +[2395.28 --> 2396.50] using this special variable +[2396.50 --> 2397.74] and not actually checking. +[2397.96 --> 2400.02] But if the mechanism is there +[2400.02 --> 2400.92] and if the language +[2400.92 --> 2402.96] creates the framework +[2402.96 --> 2403.86] where you kind of +[2403.86 --> 2404.66] have to do it, +[2404.76 --> 2405.94] then it's kind of feel foolish, +[2406.02 --> 2406.38] I think, +[2406.46 --> 2406.96] to not do it, +[2407.00 --> 2407.50] even though you can. +[2407.58 --> 2408.48] Because if you don't want +[2408.48 --> 2409.18] to do those things +[2409.18 --> 2410.34] and just go back to C +[2410.34 --> 2411.64] and play without +[2411.64 --> 2413.10] any safety belts +[2413.10 --> 2413.64] and, you know, +[2413.64 --> 2416.62] just play by your own rules. +[2416.70 --> 2417.12] But if you're going +[2417.12 --> 2418.00] to use the Go language, +[2418.10 --> 2419.18] I think it makes sense +[2419.18 --> 2421.00] to use the language +[2421.00 --> 2421.88] as it was intended. +[2422.06 --> 2422.74] And this is, in fact, +[2422.78 --> 2423.32] what I'm seeing +[2423.32 --> 2424.74] when I look at malware code. +[2425.28 --> 2426.38] That is a little bit +[2426.38 --> 2427.74] sparking joy +[2427.74 --> 2428.56] to know that +[2428.56 --> 2429.52] even hackers use +[2429.52 --> 2430.72] the best practices. +[2431.10 --> 2431.60] But it's true +[2431.60 --> 2432.20] that errors +[2432.20 --> 2434.08] have a lot of information +[2434.08 --> 2434.84] in them. +[2434.84 --> 2435.94] Is that something +[2435.94 --> 2437.76] that, as a developer, +[2438.02 --> 2438.98] you are not seeing? +[2439.24 --> 2440.02] Do you see a lot +[2440.02 --> 2440.96] of co-workers +[2440.96 --> 2442.30] and code you receive +[2442.30 --> 2443.40] where the error checking +[2443.40 --> 2444.28] is just bypassed +[2444.28 --> 2445.30] and not used at all? +[2445.68 --> 2446.86] No, usually those things +[2446.86 --> 2447.84] will not pass +[2447.84 --> 2449.34] peer review, +[2449.50 --> 2450.66] but I just don't know +[2450.66 --> 2451.76] enough whether hackers +[2451.76 --> 2452.60] do peer review, +[2452.72 --> 2452.94] you know, +[2453.04 --> 2454.40] so that is interesting. +[2455.14 --> 2456.00] Well, you know, +[2456.40 --> 2457.34] this is what I've seen. +[2457.44 --> 2458.12] It's going to be +[2458.12 --> 2459.20] anecdotal at best. +[2459.62 --> 2460.08] Of course, +[2460.18 --> 2460.32] you know, +[2460.32 --> 2461.08] there are always going +[2461.08 --> 2462.36] to be hackers out there +[2462.36 --> 2463.18] that do things +[2463.18 --> 2464.22] their own way, +[2464.38 --> 2465.32] their own bad way, +[2465.52 --> 2466.48] just like real-world +[2466.48 --> 2467.72] developers that work +[2467.72 --> 2469.38] on other legitimate projects. +[2470.00 --> 2471.02] And so I can only speak +[2471.02 --> 2472.58] about the few malware +[2472.58 --> 2473.78] programs that I have seen +[2473.78 --> 2474.80] and for which I can say +[2474.80 --> 2475.38] that they looked +[2475.38 --> 2476.38] pretty well developed +[2476.38 --> 2477.60] as far as I was concerned. +[2478.46 --> 2479.64] But, like, +[2479.70 --> 2480.16] there is bound +[2480.16 --> 2480.78] to be somewhere +[2480.78 --> 2481.44] out there +[2481.44 --> 2482.30] that is going to be +[2482.30 --> 2483.62] writing the most +[2483.62 --> 2484.68] despicable Go code +[2484.68 --> 2485.40] you can think of. +[2487.46 --> 2488.38] It does, of course, +[2488.44 --> 2488.88] make sense. +[2488.98 --> 2489.34] But still, +[2489.40 --> 2490.02] I'm happy to hear +[2490.02 --> 2491.48] that generally good practices +[2491.48 --> 2492.06] are followed. +[2492.06 --> 2492.52] everywhere. +[2493.20 --> 2493.76] Another thing +[2493.76 --> 2494.70] that I wanted to ask +[2494.70 --> 2495.40] about that is +[2495.40 --> 2496.48] how does this represent +[2496.48 --> 2497.06] in assembly, +[2497.64 --> 2498.32] given that this is +[2498.32 --> 2499.36] generally not a very +[2499.36 --> 2500.84] common practice? +[2501.10 --> 2501.58] Do you, +[2502.30 --> 2502.60] I guess, +[2502.64 --> 2503.28] because you don't see +[2503.28 --> 2503.84] lots of errors, +[2503.90 --> 2504.86] you basically don't see +[2504.86 --> 2506.22] the representation of this? +[2506.86 --> 2507.00] Well, +[2507.10 --> 2508.10] the way I see it +[2508.10 --> 2508.92] is that +[2508.92 --> 2511.80] when I look +[2511.80 --> 2513.04] at the assembly code, +[2513.62 --> 2514.38] most of the times, +[2514.68 --> 2516.02] since I'm not a Go developer, +[2516.28 --> 2517.16] I have to look up +[2517.16 --> 2517.70] the functions +[2517.70 --> 2518.38] that are being called. +[2518.76 --> 2518.82] Right? +[2519.06 --> 2520.32] Sometimes the names +[2520.32 --> 2521.38] are self-explanatory, +[2521.38 --> 2522.80] but most of the times +[2522.80 --> 2523.78] I have to go to the +[2523.78 --> 2524.62] Go documentation, +[2524.88 --> 2525.52] which, by the way, +[2525.58 --> 2526.80] I think is extremely well done. +[2527.16 --> 2528.18] Every time I look for +[2528.18 --> 2529.22] a function documentation, +[2529.22 --> 2529.96] I find it, +[2530.06 --> 2531.32] which is always a good thing. +[2531.68 --> 2532.30] There are languages +[2532.30 --> 2533.74] where you try to find stuff +[2533.74 --> 2534.98] and you just don't, +[2535.08 --> 2535.20] right? +[2535.26 --> 2536.60] Even those basic functions. +[2536.84 --> 2537.22] But anyway, +[2537.84 --> 2538.72] so I go look +[2538.72 --> 2539.26] at the documentation +[2539.26 --> 2540.76] and then I get information +[2540.76 --> 2542.44] about what arguments +[2542.44 --> 2543.42] this function is supposed +[2543.42 --> 2544.04] to receive +[2544.04 --> 2545.86] and what return values +[2545.86 --> 2546.86] it's supposed to provide. +[2546.86 --> 2548.96] And then when I do +[2548.96 --> 2550.36] this looking up, +[2550.92 --> 2551.94] then I get the information +[2551.94 --> 2553.08] about whether or not +[2553.08 --> 2554.08] the error value +[2554.08 --> 2555.76] is supposed to be returned +[2555.76 --> 2557.74] or whether an error value +[2557.74 --> 2558.44] is supposed to be returned +[2558.44 --> 2558.98] by the function. +[2559.46 --> 2560.56] And when that's the case, +[2560.60 --> 2561.44] then I cannot recall +[2561.44 --> 2562.84] of an instance +[2562.84 --> 2565.16] where I was supposed +[2565.16 --> 2566.58] to see an error value +[2566.58 --> 2567.98] returned from the program +[2567.98 --> 2568.82] and that was not checked. +[2569.26 --> 2571.14] So the way that you would +[2571.14 --> 2571.82] see it in assembly +[2571.82 --> 2572.46] would be like +[2572.46 --> 2573.38] have this function call +[2573.38 --> 2574.34] and then you see +[2574.34 --> 2576.00] some random variable +[2576.00 --> 2577.08] being taken back +[2577.08 --> 2577.70] from the stack +[2577.70 --> 2579.64] and compared with value zero. +[2579.86 --> 2580.24] So basically, +[2580.42 --> 2582.18] if error is equal to nil +[2582.18 --> 2583.64] and then you have a block +[2583.64 --> 2585.16] and whether or not +[2585.16 --> 2586.92] the error is or is not nil, +[2586.98 --> 2587.54] then you can go +[2587.54 --> 2588.10] into that block +[2588.10 --> 2589.40] or go into another one. +[2590.20 --> 2591.38] But that block is here, +[2591.50 --> 2592.10] which means that +[2592.10 --> 2592.62] the attacker +[2592.62 --> 2593.92] or the malware author +[2593.92 --> 2595.26] went through the trouble +[2595.26 --> 2596.48] of actually making sure +[2596.48 --> 2597.22] that the error, +[2597.74 --> 2598.64] like there was no error +[2598.64 --> 2599.36] returned by the function. +[2600.04 --> 2600.82] So this is the way +[2600.82 --> 2602.14] that I observe it. +[2602.14 --> 2603.44] So earlier, +[2603.56 --> 2604.32] you mentioned something +[2604.32 --> 2605.64] that you compared +[2605.64 --> 2606.92] kind of your work +[2606.92 --> 2608.34] to getting code +[2608.34 --> 2609.62] from a colleague, +[2609.84 --> 2610.94] but it's all kind of +[2610.94 --> 2612.62] no parameter names, +[2612.80 --> 2613.64] no documentation, +[2613.86 --> 2614.58] no function names +[2614.58 --> 2615.16] and so on. +[2615.56 --> 2616.46] And that reminded me +[2616.46 --> 2616.98] how sometimes +[2616.98 --> 2617.74] you can use +[2617.74 --> 2619.68] all sorts of AI tools +[2619.68 --> 2620.32] like Codex +[2620.32 --> 2621.82] and Copilot +[2621.82 --> 2622.28] and whatnot +[2622.28 --> 2623.34] to highlight that +[2623.34 --> 2623.70] and says, +[2624.00 --> 2625.02] well, explain what it does. +[2625.68 --> 2626.88] So did you ever have a chance +[2626.88 --> 2627.84] to use one of those? +[2628.64 --> 2629.30] I didn't. +[2629.52 --> 2630.32] I know that GitHub +[2630.32 --> 2631.52] released this project +[2631.52 --> 2632.74] but I personally +[2632.74 --> 2634.68] have a very +[2634.68 --> 2636.04] religious fear +[2636.04 --> 2636.86] of such projects +[2636.86 --> 2637.52] just because, +[2637.58 --> 2637.74] you know, +[2637.80 --> 2638.34] I know that +[2638.34 --> 2639.28] the way it works +[2639.28 --> 2640.56] is that all the source code +[2640.56 --> 2641.56] I write gets uploaded +[2641.56 --> 2642.70] into the cloud +[2642.70 --> 2643.36] and analyzed +[2643.36 --> 2644.18] and gets to feed +[2644.18 --> 2645.66] their machine learning algorithm. +[2646.42 --> 2647.96] And it's kind of stupid +[2647.96 --> 2648.94] because all the code +[2648.94 --> 2649.82] I write ends up +[2649.82 --> 2650.74] being open source anyway, +[2651.10 --> 2651.84] but I don't like that. +[2652.16 --> 2652.84] And if it's on GitHub, +[2653.00 --> 2654.08] it goes to the same place. +[2654.80 --> 2655.20] Exactly. +[2655.62 --> 2656.70] It also ends up there. +[2656.86 --> 2657.50] So overall, +[2657.64 --> 2658.34] like it's not really +[2658.34 --> 2659.26] a good reason for me +[2659.26 --> 2659.96] not to do it, +[2660.28 --> 2662.72] but I didn't try it yet. +[2662.90 --> 2663.50] I've been told +[2663.50 --> 2664.30] by some coworkers, +[2664.56 --> 2665.28] I think they've used it +[2665.28 --> 2665.78] for Python +[2665.78 --> 2666.56] and I've been told +[2666.56 --> 2667.24] that it's amazing +[2667.24 --> 2668.06] and you can pretty much +[2668.06 --> 2669.38] guess what you are thinking, +[2669.62 --> 2670.44] which is kind of scary. +[2670.86 --> 2671.90] But when you write code +[2671.90 --> 2672.74] or when you reverse code? +[2672.74 --> 2673.82] Yeah, exactly. +[2673.98 --> 2674.70] When you write code, +[2674.82 --> 2676.44] I am not aware +[2676.44 --> 2679.56] of a machine learning project +[2679.56 --> 2680.36] that would help you +[2680.36 --> 2681.84] reverse engineer programs, +[2682.28 --> 2684.42] although I am 100% sure +[2684.42 --> 2685.44] that this is possible. +[2685.44 --> 2686.78] Like I've been playing a lot +[2686.78 --> 2688.96] with the image generation AIs, +[2689.06 --> 2689.92] especially mid-journey. +[2690.24 --> 2691.54] I tried the one +[2691.54 --> 2692.22] that generates text, +[2692.34 --> 2693.18] which is called Lex. +[2693.70 --> 2694.94] All those AIs, +[2695.10 --> 2695.96] as far as I'm concerned, +[2696.08 --> 2697.78] produce incredible results. +[2698.26 --> 2699.40] Like if you had told me +[2699.40 --> 2700.42] one year ago +[2700.42 --> 2701.14] that I would be able +[2701.14 --> 2701.88] to type some text +[2701.88 --> 2702.34] and I would get +[2702.34 --> 2703.70] the corresponding image generated +[2703.70 --> 2704.48] and that the image +[2704.48 --> 2705.70] would actually look pretty amazing, +[2706.30 --> 2706.96] I would really not +[2706.96 --> 2708.10] have believed you for real. +[2708.22 --> 2708.78] I would have said +[2708.78 --> 2709.76] that this is science fiction. +[2709.86 --> 2710.58] It's never going to happen +[2710.58 --> 2711.20] in my lifetime +[2711.20 --> 2712.74] or maybe when I'm old +[2712.74 --> 2713.70] and don't understand +[2713.70 --> 2714.42] what's going on anymore. +[2714.42 --> 2717.08] But like we're there. +[2717.20 --> 2717.94] We're there for +[2717.94 --> 2719.20] many complex applications +[2719.20 --> 2720.46] such as understanding +[2720.46 --> 2721.18] human language +[2721.18 --> 2722.26] and generating +[2722.26 --> 2723.48] lots of content. +[2724.02 --> 2725.80] And at this stage, +[2725.90 --> 2727.56] I would be extremely shocked +[2727.56 --> 2728.46] if you told me +[2728.46 --> 2730.02] that the recognizing +[2730.02 --> 2731.46] functions that are +[2731.46 --> 2732.06] actually generated +[2732.06 --> 2733.32] by other computers +[2733.32 --> 2734.52] is not something possible. +[2734.78 --> 2736.36] Like this is 100% +[2736.36 --> 2737.72] going to happen eventually. +[2737.90 --> 2738.64] I don't know +[2738.64 --> 2739.48] who's going to do it. +[2739.82 --> 2740.64] Maybe I should actually. +[2740.64 --> 2741.52] I don't know anything +[2741.52 --> 2742.54] about the pageant, +[2742.66 --> 2743.68] but like this is +[2743.68 --> 2744.84] an extremely worthy project +[2744.84 --> 2745.90] and I think that eventually +[2745.90 --> 2747.12] this is going to help us +[2747.12 --> 2748.84] win so much time +[2748.84 --> 2749.50] when we work +[2749.50 --> 2750.70] on some unknown programs. +[2750.84 --> 2750.94] Yeah. +[2751.08 --> 2752.40] Probably we'll need to have +[2752.40 --> 2753.36] specialized AIs +[2753.36 --> 2754.34] for different languages. +[2754.46 --> 2755.38] We will need one for C, +[2755.46 --> 2756.08] one for C++, +[2756.28 --> 2756.82] one for Go, +[2757.40 --> 2757.84] et cetera. +[2758.18 --> 2760.12] But I cannot imagine +[2760.12 --> 2760.98] that this is not +[2760.98 --> 2761.70] in our future +[2761.70 --> 2762.18] and probably +[2762.18 --> 2763.28] in our near future too. +[2763.98 --> 2764.78] Hopefully they won't +[2764.78 --> 2765.94] sell this too expensive +[2765.94 --> 2767.00] because I want it. +[2767.54 --> 2768.78] So a follow-up question +[2768.78 --> 2769.28] to that. +[2769.28 --> 2771.12] for code generation, +[2771.94 --> 2772.56] some languages +[2772.56 --> 2773.56] are better than others. +[2773.86 --> 2774.32] For example, +[2774.52 --> 2775.84] Go is performing +[2775.84 --> 2776.76] even better +[2776.76 --> 2777.84] than Python and such +[2777.84 --> 2778.50] just because +[2778.50 --> 2779.48] Go has this +[2779.48 --> 2780.24] built-in linter +[2780.24 --> 2781.56] and there's many things +[2781.56 --> 2782.52] that it's not +[2782.52 --> 2783.08] either or, +[2783.22 --> 2783.60] but you know, +[2783.62 --> 2784.68] it's definitely tabs. +[2784.82 --> 2785.52] It's definitely +[2785.52 --> 2786.62] curly brackets +[2786.62 --> 2787.36] and a new line. +[2787.54 --> 2788.78] So the AI +[2788.78 --> 2790.20] has a more +[2790.20 --> 2791.18] consistent data +[2791.18 --> 2792.92] set to be trained on +[2792.92 --> 2793.58] versus Python +[2793.58 --> 2794.04] that you or +[2794.04 --> 2794.82] many other languages +[2794.82 --> 2795.46] that you can write +[2795.46 --> 2796.60] in whatever way. +[2796.60 --> 2797.74] So it just sees +[2797.74 --> 2798.80] lots of different examples +[2798.80 --> 2799.66] and it might +[2799.66 --> 2800.72] in best case +[2800.72 --> 2801.70] generate inconsistent +[2801.70 --> 2803.00] code of, +[2803.58 --> 2803.76] you know, +[2803.82 --> 2804.44] one file is different +[2804.44 --> 2804.92] from another, +[2805.02 --> 2805.78] but sometimes even +[2805.78 --> 2807.68] just sometimes wrong +[2807.68 --> 2809.28] following two different +[2809.28 --> 2810.52] paradigms in one file. +[2811.48 --> 2812.80] Do you think that +[2812.80 --> 2814.22] for the reverse part of it, +[2814.64 --> 2815.50] will this benefit +[2815.50 --> 2816.36] kind of keep rolling +[2816.36 --> 2816.98] or not +[2816.98 --> 2818.48] for the AI perspective +[2818.48 --> 2818.96] of it? +[2819.38 --> 2820.36] Yes and no. +[2820.42 --> 2821.22] So what you are referring +[2821.22 --> 2821.94] to is the fact +[2821.94 --> 2823.20] that the more strict +[2823.20 --> 2824.12] languages are +[2824.12 --> 2824.76] then the easier +[2824.76 --> 2825.38] for an AI +[2825.38 --> 2827.06] it is to understand +[2827.06 --> 2827.82] what is meant +[2827.82 --> 2828.66] and to generate code +[2828.66 --> 2829.56] and I think it makes sense +[2829.56 --> 2831.36] because Python, +[2831.64 --> 2831.98] for instance, +[2832.10 --> 2832.66] as a language +[2832.66 --> 2833.58] is extremely ambiguous, +[2833.86 --> 2834.04] right? +[2834.36 --> 2835.14] Not as ambiguous +[2835.14 --> 2835.84] as JavaScript +[2835.84 --> 2837.02] or maybe as PHP, +[2837.64 --> 2837.78] you know, +[2837.82 --> 2838.16] which are, +[2838.32 --> 2839.14] as far as I'm concerned, +[2839.46 --> 2840.20] they don't qualify +[2840.20 --> 2841.22] as actual languages +[2841.22 --> 2842.02] because they're not +[2842.02 --> 2842.82] rigorous at all. +[2843.34 --> 2844.16] But when you have +[2844.16 --> 2844.90] stuff like +[2844.90 --> 2845.64] duct typing +[2845.64 --> 2846.48] and a variable +[2846.48 --> 2847.14] that you don't have +[2847.14 --> 2848.60] to type at all, +[2849.20 --> 2850.18] well, of course, +[2850.40 --> 2851.02] for an AI +[2851.02 --> 2852.18] to figure out +[2852.18 --> 2852.90] what is going on +[2852.90 --> 2853.66] is going to require +[2853.66 --> 2854.46] much more effort +[2854.46 --> 2855.18] and it's going to +[2855.18 --> 2856.64] probably not be possible +[2856.64 --> 2857.30] all the time +[2857.30 --> 2857.90] because I think +[2857.90 --> 2858.80] that in many cases +[2858.80 --> 2859.54] a lot of stuff +[2859.54 --> 2860.16] gets determined +[2860.16 --> 2860.70] at runtime +[2860.70 --> 2862.12] which is not something +[2862.12 --> 2862.62] that an AI +[2862.62 --> 2863.74] is probably going +[2863.74 --> 2865.32] to be able to do +[2865.32 --> 2866.88] at least not that soon. +[2867.60 --> 2868.88] On the other end, +[2869.12 --> 2870.52] when you have +[2870.52 --> 2871.64] some assembly code, +[2872.00 --> 2872.82] then assembly code +[2872.82 --> 2873.86] is going to be +[2873.86 --> 2875.12] this very strict +[2875.12 --> 2876.08] unique language +[2876.08 --> 2877.32] that probably all AIs +[2877.32 --> 2878.26] will have to work on +[2878.26 --> 2879.74] and I'm not exactly sure +[2879.74 --> 2880.70] how they are going +[2880.70 --> 2881.66] to work their way +[2881.66 --> 2883.20] back up to either +[2883.20 --> 2884.32] recognizing a function +[2884.32 --> 2885.76] or actually generating +[2885.76 --> 2887.52] corresponding high-level code. +[2888.38 --> 2889.20] But the good thing +[2889.20 --> 2890.00] that AIs will have +[2890.00 --> 2890.58] going for them +[2890.58 --> 2891.32] is that assembly +[2891.32 --> 2892.26] is going to be +[2892.26 --> 2893.38] the exact opposite +[2893.38 --> 2894.30] of ambiguous. +[2894.78 --> 2896.20] You have ambiguous +[2896.20 --> 2897.20] and at the exact +[2897.20 --> 2897.90] other end +[2897.90 --> 2899.46] of what possible language +[2899.46 --> 2899.72] you have, +[2899.74 --> 2900.28] you have assembly +[2900.28 --> 2902.32] which is 100% precise. +[2902.56 --> 2903.84] As consistent as it gets. +[2903.94 --> 2904.56] Yeah, consistent +[2904.56 --> 2905.86] and actually dumb +[2905.86 --> 2906.56] in some ways +[2906.56 --> 2907.64] but it's just +[2907.64 --> 2909.30] very simple operations +[2909.30 --> 2910.26] that can only do +[2910.26 --> 2910.92] a single thing +[2910.92 --> 2911.82] in a very defined way. +[2911.96 --> 2913.54] So on that front, +[2913.66 --> 2914.40] I think that this is +[2914.40 --> 2915.02] actually going to be +[2915.02 --> 2916.12] a very, very good thing +[2916.12 --> 2916.80] for the AIs +[2916.80 --> 2917.92] whenever they are ready. +[2918.18 --> 2919.24] So there's already +[2919.24 --> 2920.38] good tools out there +[2920.38 --> 2921.28] that just take binary +[2921.28 --> 2922.32] and translate that +[2922.32 --> 2923.84] into assembly, +[2924.10 --> 2924.88] not 100%, +[2924.88 --> 2926.52] but a very good coverage. +[2927.18 --> 2927.66] And assembly +[2927.66 --> 2928.68] is consistent enough +[2928.68 --> 2929.62] so that means that +[2929.62 --> 2931.64] some IDAPro plugin +[2931.64 --> 2933.26] that uses AI +[2933.26 --> 2934.14] must be developed +[2934.14 --> 2934.80] as we speak +[2934.80 --> 2936.06] to say, +[2936.30 --> 2936.48] like, +[2936.52 --> 2937.56] here's an assembly input, +[2937.56 --> 2938.58] but please translate that +[2938.58 --> 2939.70] to GoCode for me. +[2940.12 --> 2941.18] Well, it's a good question +[2941.18 --> 2943.10] because you would think +[2943.10 --> 2943.56] that someone +[2943.56 --> 2944.64] would be working on this +[2944.64 --> 2945.68] but when you look +[2945.68 --> 2946.54] at the market +[2946.54 --> 2947.88] for reverse engineering tool, +[2947.94 --> 2948.84] it's actually quite small. +[2949.02 --> 2950.50] You have Hexrays, +[2950.70 --> 2951.96] the creators of IDAPro +[2951.96 --> 2953.74] and Hexrays +[2953.74 --> 2954.70] are kind of a +[2954.70 --> 2956.00] old school company. +[2956.20 --> 2956.24] Like, +[2956.26 --> 2958.20] their product is amazing +[2958.20 --> 2959.12] but, +[2959.28 --> 2959.54] like, +[2959.62 --> 2960.06] they've been, +[2960.66 --> 2961.82] they haven't really tried +[2961.82 --> 2963.24] to create +[2963.24 --> 2964.34] any form of disruption +[2964.34 --> 2965.62] in the past 20 years. +[2965.80 --> 2965.94] Now, +[2966.00 --> 2966.54] they have been doing +[2966.54 --> 2967.24] lots of improvements +[2967.24 --> 2967.88] to their products +[2967.88 --> 2969.00] but I think part of it +[2969.00 --> 2969.98] is only because +[2969.98 --> 2970.92] they have been challenged +[2970.92 --> 2971.76] by Ghidra, +[2971.96 --> 2973.12] their open source competitor. +[2973.78 --> 2973.90] Right? +[2974.24 --> 2975.32] It's kind of +[2975.32 --> 2976.82] my opinion that, +[2976.98 --> 2977.12] well, +[2977.14 --> 2977.68] let's make this +[2977.68 --> 2978.64] my unpopular opinion +[2978.64 --> 2979.36] if you want to, +[2979.46 --> 2980.60] but it's kind of my opinion +[2980.60 --> 2981.60] that if Ghidra +[2981.60 --> 2982.48] had never appeared, +[2983.00 --> 2984.08] then IDAPro +[2984.08 --> 2984.72] would basically +[2984.72 --> 2985.24] have stayed +[2985.24 --> 2986.62] kind of the same +[2986.62 --> 2987.64] like for the next decade +[2987.64 --> 2988.10] or so +[2988.10 --> 2989.34] because the developers +[2989.34 --> 2990.50] had kind of no incentive +[2990.50 --> 2991.80] to make it +[2991.80 --> 2992.70] significantly better +[2992.70 --> 2993.38] because just +[2993.38 --> 2994.60] they had no competition there. +[2995.08 --> 2996.40] The way that the compiler +[2996.40 --> 2997.58] is working at the moment +[2997.58 --> 2999.02] as far as I understand +[2999.02 --> 2999.88] is purely +[2999.88 --> 3001.34] through algorithmic means +[3001.34 --> 3002.56] and they do not use +[3002.56 --> 3003.58] any form of +[3003.58 --> 3005.16] machine learning. +[3005.48 --> 3006.14] There is no AI +[3006.14 --> 3007.08] applied to +[3007.08 --> 3007.72] their +[3007.72 --> 3009.42] decompilation process. +[3010.22 --> 3011.18] Maybe they have +[3011.18 --> 3012.30] started working on this +[3012.30 --> 3012.64] but +[3012.64 --> 3014.64] as a company +[3014.64 --> 3015.32] and maybe I'm +[3015.32 --> 3015.90] totally mistaken +[3015.90 --> 3016.32] about this, +[3016.38 --> 3017.06] I don't work there +[3017.06 --> 3018.02] but Hexray +[3018.02 --> 3018.74] doesn't strike me +[3018.74 --> 3019.34] as a company +[3019.34 --> 3020.08] that would be doing +[3020.08 --> 3021.26] groundbreaking R&D +[3021.26 --> 3022.24] that would +[3022.24 --> 3022.86] maybe +[3022.86 --> 3024.30] prepare for the +[3024.30 --> 3025.18] the next generation +[3025.18 --> 3026.08] of decompilers. +[3026.18 --> 3026.66] I think they would +[3026.66 --> 3027.90] rather make incremental +[3027.90 --> 3029.12] changes on their +[3029.12 --> 3029.78] existing product +[3029.78 --> 3030.54] to make it slightly +[3030.54 --> 3031.54] better year after year. +[3032.24 --> 3033.24] I think Ghidra +[3033.24 --> 3034.40] which has a +[3034.40 --> 3035.32] decompiler as well +[3035.32 --> 3036.50] is open source +[3036.50 --> 3037.38] and I think this +[3037.38 --> 3038.24] decompiler is doing +[3038.24 --> 3039.52] pretty well too +[3039.52 --> 3040.84] but I don't think +[3040.84 --> 3041.64] it uses AI +[3041.64 --> 3042.98] in any shape +[3042.98 --> 3043.70] or form either +[3043.70 --> 3045.42] and I'm not aware +[3045.42 --> 3046.52] of any plans +[3046.52 --> 3047.76] to like to start +[3047.76 --> 3048.44] working on this. +[3048.92 --> 3049.48] Probably +[3049.48 --> 3050.98] developing some +[3050.98 --> 3051.84] AI product +[3051.84 --> 3052.96] focused on reverse +[3052.96 --> 3053.34] engineering +[3053.34 --> 3053.96] we require +[3053.96 --> 3054.70] some very +[3054.70 --> 3056.22] specific AI +[3056.22 --> 3056.74] knowledge +[3056.74 --> 3058.90] and my feeling +[3058.90 --> 3059.60] is I don't know +[3059.60 --> 3060.54] everyone in the +[3060.54 --> 3061.08] security field +[3061.08 --> 3062.46] but I tend to +[3062.46 --> 3063.70] like my perception +[3063.70 --> 3064.44] of this is that +[3064.44 --> 3065.14] we have people +[3065.14 --> 3065.68] that are extremely +[3065.68 --> 3066.32] focused and +[3066.32 --> 3067.10] extremely talented +[3067.10 --> 3067.60] and skilled +[3067.60 --> 3069.00] in the specific +[3069.00 --> 3070.02] field of +[3070.02 --> 3070.54] cybersecurity +[3070.54 --> 3071.08] that they are +[3071.08 --> 3071.62] working on +[3071.62 --> 3072.98] but I cannot +[3072.98 --> 3074.30] recall anyone +[3074.30 --> 3074.96] that I've met +[3074.96 --> 3076.22] that was both +[3076.22 --> 3077.26] a great researcher +[3077.26 --> 3077.92] or a great +[3077.92 --> 3078.52] reverse engineer +[3078.52 --> 3078.94] or a great +[3078.94 --> 3079.36] pen tester +[3079.36 --> 3080.64] and a very +[3080.64 --> 3081.50] qualified data +[3081.50 --> 3082.38] scientist as well +[3082.38 --> 3083.34] or someone +[3083.34 --> 3083.78] that would be +[3083.78 --> 3084.48] able to tune +[3084.48 --> 3085.04] a machine +[3085.04 --> 3086.36] learning algorithm +[3086.36 --> 3087.14] that would +[3087.14 --> 3087.82] change our +[3087.82 --> 3088.42] lives forever. +[3089.14 --> 3089.64] So maybe +[3089.64 --> 3090.60] someone is working +[3090.60 --> 3091.06] on this +[3091.06 --> 3091.54] somewhere +[3091.54 --> 3092.06] but if it's +[3092.06 --> 3092.54] the case +[3092.54 --> 3093.16] then I'm +[3093.16 --> 3093.62] not aware +[3093.62 --> 3093.94] of it +[3093.94 --> 3094.90] and like +[3094.90 --> 3095.88] I think +[3095.88 --> 3096.48] if it were +[3096.48 --> 3097.74] public +[3097.74 --> 3098.50] or if it +[3098.50 --> 3099.24] was out there +[3099.24 --> 3099.78] I think I +[3099.78 --> 3100.14] would know +[3100.14 --> 3100.60] about it +[3100.60 --> 3101.30] maybe +[3101.30 --> 3101.80] maybe I +[3101.80 --> 3102.14] don't +[3102.14 --> 3103.10] but overall +[3103.10 --> 3103.52] if there was +[3103.52 --> 3103.88] this big +[3103.88 --> 3104.34] project +[3104.34 --> 3105.06] about to be +[3105.06 --> 3105.40] released +[3105.40 --> 3106.40] I want to +[3106.40 --> 3106.92] believe I would +[3106.92 --> 3107.46] have heard +[3107.46 --> 3107.86] about it +[3107.86 --> 3108.44] but I do +[3108.44 --> 3108.90] still hope +[3108.90 --> 3109.24] that some +[3109.24 --> 3109.58] company +[3109.58 --> 3110.00] somewhere +[3110.00 --> 3110.88] probably in +[3110.88 --> 3111.16] Israel +[3111.16 --> 3112.28] is working +[3112.28 --> 3112.66] on this +[3112.66 --> 3113.12] in a secret +[3113.12 --> 3113.44] lab +[3113.44 --> 3113.78] and that +[3113.78 --> 3114.24] eventually +[3114.24 --> 3114.64] they're going +[3114.64 --> 3114.90] to take +[3114.90 --> 3115.26] the market +[3115.26 --> 3115.78] by storm +[3115.78 --> 3117.04] and make +[3117.04 --> 3117.94] my job +[3117.94 --> 3118.50] a lot +[3118.50 --> 3118.86] easier. +[3119.32 --> 3119.74] Yeah +[3119.74 --> 3120.96] another idea +[3120.96 --> 3121.66] for the +[3121.66 --> 3122.02] followers +[3122.02 --> 3122.78] who are +[3122.78 --> 3123.52] tuning in +[3123.52 --> 3123.84] so in +[3123.84 --> 3124.10] addition +[3124.10 --> 3124.76] to an +[3124.76 --> 3125.06] AI +[3125.06 --> 3125.48] that +[3125.48 --> 3126.10] translates +[3126.10 --> 3126.60] from +[3126.60 --> 3127.98] assembly +[3127.98 --> 3128.54] to +[3128.54 --> 3129.20] go +[3129.20 --> 3130.38] it would +[3130.38 --> 3130.56] be +[3130.56 --> 3131.00] personally +[3131.00 --> 3131.34] interesting +[3131.34 --> 3131.74] for me +[3131.74 --> 3132.16] to have +[3132.16 --> 3132.44] some +[3132.44 --> 3132.80] AI +[3132.80 --> 3133.46] that says +[3133.46 --> 3134.22] this malware +[3134.22 --> 3135.08] is written +[3135.08 --> 3135.40] in the +[3135.40 --> 3135.70] style +[3135.70 --> 3136.04] of +[3136.04 --> 3136.64] and +[3136.64 --> 3137.18] what I +[3137.18 --> 3137.48] mean by +[3137.48 --> 3137.82] that +[3137.82 --> 3138.54] is that +[3138.54 --> 3138.96] already +[3138.96 --> 3139.26] now +[3139.26 --> 3139.62] you can +[3139.62 --> 3139.96] write +[3139.96 --> 3140.24] if you +[3140.24 --> 3140.52] go to +[3140.52 --> 3140.96] codex +[3140.96 --> 3141.50] and other +[3141.50 --> 3141.86] tools +[3141.86 --> 3142.24] like that +[3142.24 --> 3142.52] you can +[3142.52 --> 3142.80] say +[3142.80 --> 3143.64] write a +[3143.64 --> 3144.10] Go program +[3144.10 --> 3144.60] that has +[3144.60 --> 3145.02] this and +[3145.02 --> 3145.36] that and +[3145.36 --> 3145.84] this and +[3145.84 --> 3146.14] do it +[3146.14 --> 3146.40] in the +[3146.40 --> 3146.70] style +[3146.70 --> 3147.04] of +[3147.04 --> 3147.98] and if +[3147.98 --> 3148.62] you mention +[3148.62 --> 3149.14] a GitHub +[3149.14 --> 3149.58] handle +[3149.58 --> 3150.04] of somebody +[3150.04 --> 3150.36] who is +[3150.36 --> 3151.46] a known +[3151.46 --> 3151.88] developer +[3151.88 --> 3152.54] has lots +[3152.54 --> 3153.08] of stars +[3153.08 --> 3153.72] or some +[3153.72 --> 3154.52] other +[3154.52 --> 3155.76] big presence +[3155.76 --> 3157.36] and their +[3157.36 --> 3157.94] type of +[3157.94 --> 3158.48] code has +[3158.48 --> 3158.92] a flavor +[3158.92 --> 3159.28] which is +[3159.28 --> 3159.68] maybe less +[3159.68 --> 3160.14] common in +[3160.14 --> 3160.54] Go because +[3160.54 --> 3160.86] it's so +[3160.86 --> 3161.22] structured +[3161.22 --> 3162.12] you will +[3162.12 --> 3163.10] get their +[3163.10 --> 3163.58] style of +[3163.58 --> 3163.92] code +[3163.92 --> 3164.68] so +[3164.68 --> 3165.56] eventually +[3165.56 --> 3166.26] and next +[3166.26 --> 3166.68] step to +[3166.68 --> 3167.00] this +[3167.00 --> 3167.98] magical +[3167.98 --> 3168.42] plugin +[3168.42 --> 3169.16] would be +[3169.16 --> 3169.84] in whose +[3169.84 --> 3170.42] style is +[3170.42 --> 3170.60] the +[3170.60 --> 3170.98] malware +[3170.98 --> 3171.40] written +[3171.40 --> 3172.04] this will +[3172.04 --> 3172.32] be also +[3172.32 --> 3172.74] interesting +[3172.74 --> 3173.62] and then +[3173.62 --> 3174.12] all those +[3174.12 --> 3175.12] language teachers +[3175.12 --> 3175.56] out there +[3175.56 --> 3176.06] will know +[3176.06 --> 3177.40] oh well +[3177.40 --> 3178.06] I taught +[3178.06 --> 3178.64] that hacker +[3178.64 --> 3179.64] yeah that's +[3179.64 --> 3180.00] a very +[3180.00 --> 3181.92] nice idea +[3181.92 --> 3182.90] actually there +[3182.90 --> 3183.42] have been +[3183.42 --> 3184.52] maybe not +[3184.52 --> 3185.54] rumors but +[3185.54 --> 3186.80] open research +[3186.80 --> 3187.38] projects +[3187.38 --> 3188.46] working exactly +[3188.46 --> 3189.18] on this for +[3189.18 --> 3190.04] probably decades +[3190.04 --> 3190.70] in the +[3190.70 --> 3191.24] intelligence +[3191.24 --> 3191.62] field +[3191.62 --> 3192.56] I think I +[3192.56 --> 3192.96] recall +[3192.96 --> 3193.96] a +[3193.96 --> 3194.96] CCC +[3194.96 --> 3195.54] presentation +[3195.54 --> 3196.68] from maybe +[3196.68 --> 3197.54] 2010 or +[3197.54 --> 3198.04] 2012 +[3198.04 --> 3198.86] one of +[3198.86 --> 3199.26] those years +[3199.26 --> 3200.44] and +[3200.44 --> 3201.28] basically +[3201.28 --> 3202.42] people were +[3202.42 --> 3202.84] already +[3202.84 --> 3203.90] working on +[3203.90 --> 3204.92] obtaining +[3204.92 --> 3205.68] code from +[3205.68 --> 3206.22] open source +[3206.22 --> 3206.70] repositories +[3206.70 --> 3207.38] like basically +[3207.38 --> 3207.88] downloading +[3207.88 --> 3208.34] github +[3208.34 --> 3209.26] and trying +[3209.26 --> 3210.20] to extract +[3210.20 --> 3211.38] maybe some +[3211.38 --> 3212.10] penmanship +[3212.10 --> 3212.78] characteristics +[3212.78 --> 3213.88] from every +[3213.88 --> 3214.60] single developer +[3214.60 --> 3215.60] and they were +[3215.60 --> 3216.26] hoping that +[3216.26 --> 3216.82] they would be +[3216.82 --> 3217.68] able to +[3217.68 --> 3218.54] take any +[3218.54 --> 3219.08] program in +[3219.08 --> 3219.50] the future +[3219.50 --> 3221.10] and be +[3221.10 --> 3221.48] able to +[3221.48 --> 3221.86] tell you +[3221.86 --> 3222.46] that this +[3222.46 --> 3223.12] might probably +[3223.12 --> 3223.58] be the +[3223.58 --> 3223.96] developer +[3223.96 --> 3225.82] that created +[3225.82 --> 3226.18] this +[3226.18 --> 3226.90] executable +[3226.90 --> 3227.68] now +[3227.68 --> 3228.92] it's been +[3228.92 --> 3229.48] 10 years +[3229.48 --> 3230.46] 15 years +[3230.46 --> 3231.44] I'm not +[3231.44 --> 3231.92] sure +[3231.92 --> 3232.76] I haven't +[3232.76 --> 3233.16] heard about +[3233.16 --> 3233.72] this for a +[3233.72 --> 3234.14] while so +[3234.14 --> 3234.48] maybe +[3234.48 --> 3235.74] this either +[3235.74 --> 3236.30] didn't work +[3236.30 --> 3236.84] as well as +[3236.84 --> 3237.38] they expected +[3237.38 --> 3238.36] or this was +[3238.36 --> 3239.14] actually absorbed +[3239.14 --> 3239.84] into some +[3239.84 --> 3240.76] intelligence service +[3240.76 --> 3241.12] somewhere +[3241.12 --> 3242.24] because there are +[3242.24 --> 3243.02] obvious very +[3243.02 --> 3243.86] obvious intelligence +[3243.86 --> 3244.74] applications there +[3244.74 --> 3246.08] that I think +[3246.08 --> 3247.30] would like +[3247.30 --> 3248.24] those types of +[3248.24 --> 3248.62] services they +[3248.62 --> 3249.18] would be they +[3249.18 --> 3249.68] would love to +[3249.68 --> 3250.12] have such a +[3250.12 --> 3250.92] capability because +[3250.92 --> 3251.68] they would be +[3251.68 --> 3252.04] able to +[3252.04 --> 3252.84] identify malware +[3252.84 --> 3253.86] authors they +[3253.86 --> 3254.38] wouldn't need +[3254.38 --> 3255.94] really to have +[3255.94 --> 3256.42] the burden of +[3256.42 --> 3256.94] proof the way +[3256.94 --> 3257.64] that you know +[3257.64 --> 3258.86] police forces +[3258.86 --> 3259.66] have like they +[3259.66 --> 3260.02] would just be +[3260.02 --> 3260.52] able to know +[3260.52 --> 3261.00] what kind of +[3261.00 --> 3262.28] guy is like +[3262.28 --> 3262.78] or what who +[3262.78 --> 3263.38] the guy behind +[3263.38 --> 3263.90] some malware +[3263.90 --> 3264.90] is and just +[3264.90 --> 3265.86] do their usual +[3265.86 --> 3266.76] parallel construction +[3266.76 --> 3268.62] stuff so this +[3268.62 --> 3269.26] is something we +[3269.26 --> 3269.96] know they want +[3269.96 --> 3270.60] and this is +[3270.60 --> 3271.10] something that we +[3271.10 --> 3271.80] also know they +[3271.80 --> 3272.62] spent money on +[3272.62 --> 3273.38] and I remember +[3273.38 --> 3273.80] that some +[3273.80 --> 3274.56] universities were +[3274.56 --> 3275.28] actually working +[3275.28 --> 3275.80] on those types +[3275.80 --> 3277.00] of projects for +[3277.00 --> 3277.82] research not for +[3277.82 --> 3278.46] intelligence but +[3278.46 --> 3279.98] it's kind of those +[3279.98 --> 3280.80] worlds they tend to +[3280.80 --> 3281.36] communicate with +[3281.36 --> 3281.94] each other anyway +[3281.94 --> 3282.60] when there are +[3282.60 --> 3284.82] applications if +[3284.82 --> 3286.22] this came to +[3286.22 --> 3287.60] fruition then +[3287.60 --> 3289.22] this did so in +[3289.22 --> 3290.76] secret the way +[3290.76 --> 3291.42] that they used to +[3291.42 --> 3292.02] be working on +[3292.02 --> 3292.66] this was again +[3292.66 --> 3293.54] algorithmic they +[3293.54 --> 3293.96] were trying to +[3293.96 --> 3294.52] extract their +[3294.52 --> 3295.72] characteristics and +[3295.72 --> 3296.12] they were not +[3296.12 --> 3297.24] using those black +[3297.24 --> 3298.84] box AI capabilities +[3298.84 --> 3299.58] that we have now +[3299.58 --> 3300.94] maybe this is a +[3300.94 --> 3301.64] new avenue of +[3301.64 --> 3302.60] research for those +[3302.60 --> 3304.02] applications maybe +[3304.02 --> 3304.86] we'll know in the +[3304.86 --> 3306.54] future but as far +[3306.54 --> 3307.76] as I know this is +[3307.76 --> 3308.40] like this is an +[3308.40 --> 3309.24] existing problem that +[3309.24 --> 3309.92] people are trying to +[3309.92 --> 3310.22] solve +[3310.22 --> 3312.30] and I haven't +[3312.30 --> 3313.08] seen any signs +[3313.08 --> 3314.18] that they have +[3314.18 --> 3314.94] although it's not +[3314.94 --> 3315.88] exactly sure that I +[3315.88 --> 3316.14] would +[3316.14 --> 3317.38] this episode took +[3317.38 --> 3318.10] very interesting +[3318.10 --> 3319.02] turns like new +[3319.02 --> 3320.14] ideas for tools and +[3320.14 --> 3321.02] projects and whatnot +[3321.02 --> 3322.60] and remember where +[3322.60 --> 3323.46] you heard this first +[3323.46 --> 3324.94] yeah I was not +[3324.94 --> 3325.60] expecting that +[3325.60 --> 3326.76] yeah but that's +[3326.76 --> 3327.58] very cool that's +[3327.58 --> 3328.44] really inspiring to +[3328.44 --> 3328.88] speak with you +[3328.88 --> 3329.68] Ivan thank you +[3329.68 --> 3330.80] for this hour of +[3330.80 --> 3331.44] the conversation +[3331.44 --> 3332.18] of course +[3332.18 --> 3332.84] as the previous +[3332.84 --> 3333.70] episode it also +[3333.70 --> 3334.50] ended with me having +[3334.50 --> 3335.12] lots of open +[3335.12 --> 3335.62] questions +[3335.62 --> 3337.30] please consider +[3337.30 --> 3338.00] joining me again +[3338.00 --> 3338.86] next year for +[3338.86 --> 3339.74] one more episode +[3339.74 --> 3340.44] or 10 more +[3340.44 --> 3344.56] eventually I think +[3344.56 --> 3344.98] it will be pretty +[3344.98 --> 3345.70] soon we're going to +[3345.70 --> 3346.74] expand like the +[3346.74 --> 3347.52] whole knowledge that +[3347.52 --> 3348.12] I have on Go +[3348.12 --> 3348.88] language right +[3348.88 --> 3349.98] if I have to come +[3349.98 --> 3350.92] back then at this +[3350.92 --> 3351.60] point I will really +[3351.60 --> 3352.40] have to look into +[3352.40 --> 3353.82] the language more +[3353.82 --> 3354.86] and maybe try to +[3354.86 --> 3356.32] come with actual +[3356.32 --> 3357.12] research that I can +[3357.12 --> 3357.66] share with you +[3357.66 --> 3358.28] because otherwise +[3358.28 --> 3359.40] I don't want this +[3359.40 --> 3360.14] conversation to be +[3360.14 --> 3360.78] boring but I'll do +[3360.78 --> 3361.20] my best +[3361.20 --> 3362.36] next time we'll +[3362.36 --> 3363.16] talk about generics +[3363.16 --> 3363.44] maybe +[3363.44 --> 3363.84] okay +[3363.84 --> 3364.92] see how that +[3364.92 --> 3366.22] Go proposal goes +[3366.22 --> 3367.00] and I'll work on +[3367.00 --> 3367.20] that +[3367.20 --> 3368.58] yeah and instead +[3368.58 --> 3369.38] of unpopular +[3369.38 --> 3370.16] opinion we have +[3370.16 --> 3370.98] provided two +[3370.98 --> 3372.62] unicorn ideas +[3372.62 --> 3373.56] so you welcome +[3373.56 --> 3374.00] everyone +[3374.00 --> 3376.22] and actually I can +[3376.22 --> 3376.82] pretty much give you +[3376.82 --> 3377.70] my personal guarantee +[3377.70 --> 3378.66] that if anyone in +[3378.66 --> 3379.46] the audience actually +[3379.46 --> 3380.30] implements one of +[3380.30 --> 3380.92] those two things +[3380.92 --> 3381.30] that they're going +[3381.30 --> 3381.86] to be extremely +[3381.86 --> 3383.66] rich so there you +[3383.66 --> 3383.88] go +[3383.88 --> 3385.20] that's almost as +[3385.20 --> 3386.22] good as being on +[3386.22 --> 3386.84] the rank of the +[3386.84 --> 3387.42] most unpopular +[3387.42 --> 3387.82] opinion +[3387.82 --> 3388.36] yeah +[3388.36 --> 3391.08] thanks everyone +[3391.08 --> 3391.60] for joining +[3391.60 --> 3392.16] thank you very +[3392.16 --> 3392.60] much Ivan +[3392.60 --> 3393.42] thank you +[3393.42 --> 3394.22] and was happy +[3394.22 --> 3394.96] being here +[3394.96 --> 3395.54] happy +[3395.54 --> 3396.54] so it was a +[3396.54 --> 3396.98] pleasure speaking +[3396.98 --> 3397.48] to you and +[3397.48 --> 3398.02] see you next +[3398.02 --> 3398.44] time +[3398.44 --> 3404.38] that is our +[3404.38 --> 3404.88] show for this +[3404.88 --> 3405.14] week +[3405.14 --> 3405.80] thanks for +[3405.80 --> 3406.20] listening +[3406.20 --> 3407.52] did you know +[3407.52 --> 3408.70] we've added a +[3408.70 --> 3409.42] super brief +[3409.42 --> 3410.36] Monday news +[3410.36 --> 3411.12] edition to the +[3411.12 --> 3411.94] change log to +[3411.94 --> 3412.78] complement our +[3412.78 --> 3413.74] awesome long +[3413.74 --> 3414.30] form interviews +[3414.30 --> 3415.16] that we publish +[3415.16 --> 3415.62] each Friday +[3415.62 --> 3416.96] give it a +[3416.96 --> 3417.82] listen if you +[3417.82 --> 3418.48] haven't yet +[3418.48 --> 3419.24] the software +[3419.24 --> 3420.26] world moves fast +[3420.26 --> 3421.40] keep up with +[3421.40 --> 3422.16] change log news +[3422.16 --> 3423.40] check it out +[3423.40 --> 3425.14] at changelog.fm +[3425.14 --> 3426.32] thanks once again +[3426.32 --> 3426.96] to our partners +[3426.96 --> 3428.66] fastly and fly.io +[3428.66 --> 3429.48] for helping make +[3429.48 --> 3430.14] go time possible +[3430.14 --> 3431.24] to the mysterious +[3431.24 --> 3432.04] breakmaster cylinder +[3432.04 --> 3432.78] for keeping our +[3432.78 --> 3433.78] beat supply on +[3433.78 --> 3435.14] swole and to you +[3435.14 --> 3435.92] for being part of +[3435.92 --> 3436.28] the go time +[3436.28 --> 3437.46] community connect +[3437.46 --> 3438.22] with us at +[3438.22 --> 3439.00] go time at +[3439.00 --> 3439.92] changelog.social +[3439.92 --> 3440.98] next time on +[3440.98 --> 3442.18] go time Matt +[3442.18 --> 3442.90] gathers an +[3442.90 --> 3443.80] awesome panel to +[3443.80 --> 3444.50] discuss how to +[3444.50 --> 3445.42] make go code more +[3445.42 --> 3446.32] efficient through +[3446.32 --> 3447.44] modern observability +[3447.44 --> 3448.90] practices stay +[3448.90 --> 3449.96] tuned for that we'll +[3449.96 --> 3450.70] have that episode +[3450.70 --> 3451.46] all ready for you +[3451.46 --> 3452.20] next week diff --git a/How Pinterest delivers software at scale_transcript.txt b/How Pinterest delivers software at scale_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..dfe9cb3cf831626493cfac1b650b49154b09ac06 --- /dev/null +++ b/How Pinterest delivers software at scale_transcript.txt @@ -0,0 +1,253 @@ +**Johnny Boursiquot:** Alright, welcome, one and all. Today we have a special guest that is joining us to talk about scale. Scale stuff. But before I introduce him, I want to acknowledge my co-host who decided to join me last-minute, and I welcome him very much... Jon, welcome back. + +**Jon Calhoun:** Thanks for having me, Jonny. + +**Johnny Boursiquot:** So today's guest is Mr. Nishant Roy. He is an engineering manager over at Pinterest, and Pinterest deals with a lot of scale, as you would imagine. There's a lot going on over there, and we figured, hey, why don't we bring Nishant over and talk about some of these things? And obviously, talk about the role Go plays in the mix. But obviously, I have to warn you, this is not going to be an all about Go kind of podcast. Obviously, Go plays a role in the kind of engineering they're doing over there... But obviously, our conversation is going to be a bit broad, as far as things like CI and CD, and pipelines, and \[unintelligible 00:03:28.29\] that kind of scale is concerned, and hopefully that'll be of interest to a lot of you out there who may be in a similar situation. So Nishant, why don't you give us a brief intro, and before we get into it, what brought you to Pinterest, and what are you doing over there? + +**Nishant Roy:** Yeah, thanks so much, Jonny. Nice to meet you, Jon, and Jonny as well. Hi, everyone on the stream. So yeah, I'm Nishant, I've been at Pinterest for just under five years now. I lead the ad serving platform team here. I started out as an intern actually, close to six years ago now, on the same team. The team at that point was just about ten people. My team itself now is 15. So talking about scale, the team has scaled tremendously as well with the product and everything around us. + +What brought me to Pinterest is actually a little different than what I'd say most people's answers are. I met some of the team on campus at Georgia Tech while they were there recruiting, and I really enjoyed talking to both the recruiting team, as well as some of the engineers and product managers who were there. As people, they seemed like people I would love to work with and hang around with, and also, just the problems they were talking about at that point - this is like late 2016 when I met them for my internship... Pinterest was just entering, I think, this hyper-growth phase from both a user perspective and a revenue/monetization perspective... So it just seemed like a really interesting place to go. And coming out as a new grad of college, I didn't really know a whole lot, so I figured it'd be a good place to take what I did know, learn a lot more from those folks, and see what a company going through that hyper growth period does look like. + +**Johnny Boursiquot:** Awesome, awesome. So you've been there for quite a while, so you've had a chance to see the organization, the engineering organization, at least you're part of it, mature and grow as well, yeah? + +**Nishant Roy:** Yeah. And the big change, I think, was obviously going from a pre-IPO company to going public, and seeing how various things change around that, not only from a day-to-day engineering work, but also communications that go out within the company, outside the company... It definitely did really impact how we operate as an engineering org. The biggest thing, obviously, as once you're public, there's a lot more compliance that you have to go through, and being on the ad side of that, we deal with that pretty frequently. + +**Johnny Boursiquot:** Okay. So the life of an engineer day to day at Pinterest, working in your group, looks like what, generally speaking? + +**Nishant Roy:** Yeah, so back when I started, again, the team was much smaller; the overall ads organization was much smaller, so things were moving a lot faster. We had less checks in place for things like -- if we're talking about scale and CI/CD today, we didn't have as robust a system as we do today. People were able to make changes a lot faster... However, that came at the cost of a lack of proper verification in place. So we had a much faster feedback loop in one sense, that you were able to get changes out to production faster... However, the piece of feedback that was missing was, "Is my change actually gonna bring down production, or cause any major issues?" It was not completely missing, but it was definitely less sophisticated than it is today. + +\[06:13\] So at that point, there was less process, essentially. If you had your own idea, go write the code, you essentially just needed one person to sign off on the code review... Less compliance and blocking reviews at that time. Go ahead, put your change out, present your metrics, your experiment, your A/B results essentially to your team leader, your org leader, and as long as you get approval from a couple of people, you were good to go. So that was what a day in the life looked like back then. + +Now there's a lot more process; not necessarily a bad thing. I know process has a pretty negative connotation... But what it means - we have way fewer severe incidents at least, and that means more people at the org have an understanding and a say in what changes are going out and what reason they're going out for. So essentially, like I mentioned earlier, you just wrote up this doc with your results, and got approval from a few folks, and you were good to go. Now, essentially, there's a more robust and involved process to get, where there's a forum that comes together to review the changes, and there's healthy debate around why a certain decision is being made, and what the rationale for that is... And ensuring that everyone from different \[unintelligible 00:07:13.24\] and the product side are on board before any major changes like that go out that may impact pinners, us internally, or our partners or advertisers. + +**Johnny Boursiquot:** Okay. I imagine that some changes require more scrutiny than others, right? + +**Nishant Roy:** Definitely. + +**Johnny Boursiquot:** So if it's a quick bug fix, you probably don't go through that same extensive process every time, right? + +**Nishant Roy:** Yes. So we have a more lightweight process for smaller bug fixes etc. And the criteria for that, essentially, is as long as there's not a significant change to any of the 10 to 15 top-line metrics that the org has decided are vital to monitor, that change requires the old process, essentially; you get approval from one or two folks within your team or your org, and then you're good to go. + +**Johnny Boursiquot:** Okay, so part of being able to deliver changes, big or small, is having some sort of an integration and delivery pipeline, right? So it's the same process, I would assume, whether the change is small or big; you're going through the same -- from a technology standpoint, from an automated process standpoint, it's the same thing. There's no different way of doing things if it's a small change or a different way if it's a big change. + +**Nishant Roy:** That's right. + +**Johnny Boursiquot:** Okay. So with that said, what are the stages of this pipeline? So I'm thinking, generally speaking, most people only really have -- they do their unit tests locally, and there might be some integration tests that happen in the cloud, and maybe there's a staging environment involved... For most people, that's sufficient. Right? So basically, what are the different stages in your pipeline, and why do you have all these different stages? + +**Nishant Roy:** Yeah, and I'm happy to say that this is something that's evolved a lot as we've scaled as well... So just backing up again to what I do - as part of leading the ad serving platform team, our responsibilities are to enable the ads team to continue to grow and deliver and launch more products, new algorithms to improve ad delivery and efficiency on the platform. We're over a $2 billion company now; that all comes from ads, so that's part of what we do. + +The other side of that obviously is, again, going back to scale and compliance, making sure that our systems are scaling at an ideally sublinear cost rate. So for every dollar made, we don't want to be spending an additional dollar; that's one condition. And then also, as the number of engineers grows, and the number of changes grow in our system, we don't want to be having as many or as frequent or as severe outages, at the very least. So given those requirements, having a robust CI/CD system, integration testing, staging environment, etc. incremental rollout became extremely important. + +So what this was like when I joined Pinterest was - I'm proud to say that the ads team was one of the first teams to have a continuous deployment process, so there was a pretty good foundation for me to build on top of. Essentially, when you committed your change, within let's say the next 30 minutes it would hit a canary environment. That canary environment was essentially just, let's say, one host, and the goal of that host was to ensure -- and this is a Go service. So the goal of that host was to just make sure that your service is not panicking and crashing. If it did, it would trigger an alert to an on-caller, who would then go in, look at the cause for the panic, roll back the change, and ensure that it doesn't go out to further stages etc. + +\[10:12\] Once it passed that initial canary test, which is about a 10 to 15-minute test, it would go out to what we call the staging, but let's say like a larger canary essentially, which was at that point let's say about 1% of the cluster, of the entire production cluster. This is all serving production traffic, so now on these 1% of the cluster we can actually monitor more application-level metrics, whether that is number of ads inserted, CPU usage, memory usage, disk usage etc. At that point I want to say we were monitoring probably about 70 to 80 metrics. If any of those metrics showed a significant regression - and we could tune the thresholds for each of those alerts - it would trigger similarly an alert to our team's on-caller, who would have to go in and manually pause deploys, and the whole process of debugging from there on forth. + +What we've evolved a lot since is we no longer require the on caller to go manually pause deploys, or roll back deploys. We have a system in place, it's built on top of Spinnaker, which is an open source platform from Netflix, if I remember correctly... And essentially, your CD pipeline will automatically pause the deploy if any of those metrics show a significant regression. So that was a big win, because one, it reduces stress on our on-callers, two, it actually reduced the number of incidents that could have just been prevented by someone going in and clicking the button at the right time. + +Besides that, what we built in - you mentioned integration testing, and the fact that some people sometimes will run some local testings and push out their changes and wait for feedback in production... We realized that was actually one really big gap, because there was no great way of enforcing that people were actually running that local test... Because that essentially required the code reviewer to follow up and ensure that sufficient testing was performed. And again, like you said, depending on whether it's a small change or a large change, the burden on the reviewer changes based on that as well. + +So we wanted something more uniform, that applies to every single change. Every single code change runs in a reasonable amount of time, doesn't require developers to be sitting there for an hour while their tests run, and then gives us reasonable confidence that if this change goes out, we're not saying that it's definitely not going to cause an outage, but at least we're not going to have a significant outage that's going to bring down the service, bring down the site etc, etc. So one thing we built is what we call a pre-submit test. That was actually one of my first projects when I joined the company. Essentially, every request that comes into Pinterest, we log a sample of that -- or every request that comes into the ad system rather, we log a certain sample of that to a Kafka topic. And then when we want to run this online integration test, essentially every time you put up a new PR, we package your changes, similar to how we would for a production build. Package your changes, create an artifact that this pre-submit test framework then deploys for you to a couple of test hosts. So now that we have that log traffic, we can essentially tail that Kafka log, get some number of requests, send them to these hosts, simulating production traffic without actually affecting any users... That host then emits its own set of metrics, and we can essentially grab those metrics, compare them against a production cluster, or at least a different host that is running the latest version of your main branch... See if there's any significant regression, and obviously, we're losing the thresholds for that, because we're only running this test for anywhere from like 3 to 10 minutes. And if there is a regression, then your commit is essentially blocked from landing until we resolve those discrepancies. + +We wanted to make sure, and I think we successfully achieved this, that this test is able to run in less than 10 to 15 minutes... Because one of the great parts of Go in our relatively smaller codebase at that point was that our entire repository, or entire service at least, was able to build an artifact in about, I want to say 4 to 5 minutes. So if we suddenly introduced this test that ran for 30 minutes, developers would have hated us. Our developer speed just went up almost 10x, so that was one of the things that we wanted to guarantee, is that this test essentially ran in 10 minutes, you did not need to wait for more than 10 minutes for all your Jenkins builds, your unit tests to pass, and for this integration test to pass before you were able to make your change. + +\[13:57\] Those were the two key parts that came out, was this pre-submit integration testing, which allowed us to actually define which metrics need to be monitored, and then secondly was this automated canary analysis essentially, which is, I think, an industry-wide practice at this point... But that really saved us a lot of headache of manual process, and reduced the stress on our on-callers. And on top of that, we've now seen a lot of teams actually adopting these frameworks add in their own metrics. So for instance, I mentioned earlier, when we had this sort of continuous deploy process in the early days, we monitored about 70 to 80 metrics. That number now - I haven't looked at it lately, but it's anywhere between 500 to 700 metrics now, which obviously doesn't come all from the infra team; it comes from a lot of product teams, and machine learning teams being able to onboard their own metrics and have the confidence that these metrics are a) stable enough, b) actually protecting our systems. + +Similarly with the pre-submit test themselves, when we first rolled it out, the infra team just configured 10 metrics, I think, which are mainly system performance-related. Now we're up to about 90 metrics, added by all sorts of product and quality teams to ensure that their particular slice of the ads pie doesn't go down. So for instance, if we're -- I mean, we could even do it to the granularity of checking the number of ads inserted for simulated dark users coming from Canada, or Japan, or something like that. We would obviously have to increase our sample size to get a meaningful number there, but this framework gives you that level of flexibility, which was widely adopted, and has really helped us a lot. + +**Johnny Boursiquot:** So it sounds like you're kind of testing in production, pretty much... + +**Nishant Roy:** Yes and no. So the ACA, the Automatic Canary Analysis is happening in production; our pre-submit testing is not happening in production. It's essentially replicating production traffic and sending it to a couple of dark side of hosts, which is not actually serving any actual user traffic. + +**Johnny Boursiquot:** Okay, so you simulating what would have happened had it been hit by the actual traffic. So I'm assuming that traffic basically was captured from production, actual production environments, and then you just basically replay it against the canary. + +**Nishant Roy:** That's right. We essentially just log 1% of all production traffic to a Kafka topic, keep it around for about two or three days. And that's constantly being refreshed, and we can constantly use that to replay and simulate what would happen, like you said, if that binary was being served to production users. + +**Break:** \[16:11\] + +**Jon Calhoun:** So when people think about testing, I think a lot of the time the first thing that pops in their head is like unit tests, and these smaller things that happen offline... Would you say that introducing all of this stuff has caused developers to focus less on that, because you have these basically production or production-simulated-type tests that are at least a lot more realistic sounding, so would you say that like the unit tests get a little bit less focused then? Or how does that change the dynamic? + +**Nishant Roy:** A little bit, yes. When we first rolled it out, especially in the first couple of years, we did see that. I mean, I personally saw a lot of PRs coming in where folks were just like "Hey, pre-submit tests passed. I'm not gonna write unit tests for this." And to be honest, that was not the worst thing in the world for developer velocity, to some extent. Like, you're guaranteed, or you have a fairly high rate of confidence that your change is safe, so why take that extra time to write unit tests, when you might be doing something, writing a new feature instead? + +We did however -- I'm trying to think of a few examples; I'll come back to you if they come to me... But there were a few instances where things sort of did pass through our pre-submit test framework likely because they weren't impacting something top-line, but something perhaps offline. For instance, if we didn't have proper validation of the data that was being logged for - whether that's offline analysis, or machine learning training jobs, or billing and reporting pipelines, all that stuff, those were things that may not necessarily be monitored by this framework, since at the start, like I said, it was mostly just monitoring ad insertions, and things like that. So we did realize that, while this is great, it's not sufficient, and we do need at least some baseline level of test coverage, or at least local integration testing, to capture those things as well. + +**Jon Calhoun:** So seeing what you're doing now and seeing how it works, at least from my perspective, this sounds like something where at scale, this approach works very well, but if you were just starting up from like a smaller business or something, this is one of those cases where if you tried to mimic what a big company was doing, it wouldn't work at all, because you just don't have the volume or scale... Or if you're only getting 100 web requests a day, you can't trust your production in 1% of your servers to actually give you any real information. So would you say that this is definitely one of those cases where as your company gets bigger, the approaches you can use sort of change and adapt based on your circumstances? + +**Nishant Roy:** \[20:06\] Completely. Like you said, if you don't have a large enough sample size, this sort of testing is not really going to give you meaningful results; it's just going to be a coin flip at that point whether your test passes or fails. So having those more deterministic unit tests that actually test the app behavior on one single request is a lot more important at an early stage. + +**Johnny Boursiquot:** You haven't mentioned what you're using for your observability stack. So a lot of things are being captured, metrics, your team are adding things based on features or products... Like, is this something home-grown, or are you using off the shelf software to provide that observability? + +**Nishant Roy:** Yeah, so for general observability at Pinterest, metrics are essentially stored in an OpenTSDB backend, and we have an internal tool that is called Statsboard I believe there's a blog post about it. If not, I'm sure there will be at some point... It's a super-great team. It's one of my favorite teams to work with. Statsboard essentially very much like Prometheus. I haven't used Prometheus a lot myself, but just a UI to visualize time-series metrics. It also allows you the ability to define alerts based on different thresholds etc. and that's what we're using for observability at a company-wide scale. + +For pre-submit tests specifically we have a slightly more custom solution, because we didn't want to go through the whole hassle of essentially -- since I mentioned that we were doing it is every binary, or every PR, that we package it into a deployable binary for testing would get deployed to a single host, or just one or two hosts. So we needed a better way of essentially isolating metrics for those two hosts, rather than needing to go through this UI, and filtering for those two host names in particular to get the right set of metrics. + +So for those hosts, we used essentially the expvar library from the Go standard library to expose all metrics through an HTTP endpoint, and essentially, we could then scrape that endpoint and get all the metrics that were generated on that host as one large JSON blob, essentially, parse those in our pre-submit test framework and use that for the actual metric regression analysis. + +**Johnny Boursiquot:** Very cool. So how much of your stack is Go-based services? + +**Nishant Roy:** Only the ads team uses Go heavily and for online serving; so the ad delivery and ad logging systems are in Go. Pinterest broadly is more a Python, Java and C++ based infrastructure. So our front, user-facing API is in Python, a lot of our backends are in Java, unless they need to be-- Unless there's like, low latency, high-performance requirements, in which case they're in C++. + +The ads team - I don't know, the historical reason for this; this happened like seven, eight years ago, so like three years before I joined... But my understanding is it was a couple of things. I think we realized that the ads stack needed to be pretty low latency, but also required pretty high developer velocity... So essentially, my understanding is the team was debating between Java versus C++. Java had its latency concerns at that point; garbage collection wasn't as advanced as it is today... C++ had developer velocity concerns at that point, and I think we had a couple folks on the team at that point who felt very strongly about Go. And again, Jon, like you mentioned, things happen differently at an early-stage company versus when you're later... So 2013-2014, Pinterest was probably like four to five years old, I want to say. I don't think we had as refined a process for choosing the frameworks and languages we develop in as we do now. + +So that's kind of how it happened, and we arrived at using Go... And it's been pretty great. I think for the longest time we've hit the happy middleground between developer velocity and performance. And Jonny, when we met at the Baltimore Go meetup, we talked a little bit about the challenges we faced with with Go efficiency, as well as on the garbage collection side... And more recently, I've been hearing through the Go Enterprise Advisory Board that this is becoming more of a known concern for large-scale companies. + +Now, the Go team has put out a few flags recently to allow you to better tune the garbage collector. In Go 19, for instance, I think most recently they added the ability to tune your soft memory limit, as opposed to tuning just the Go GC value, which I think is a great way of allowing developers to essentially better control the impact of garbage collection on their systems. I don't know if you guys have seen this, but think Uber last year or so put out a blog post about how they dynamically tune the Go GC environment variable itself based on system metrics... Which worked really well for them, I think. If I remember correctly, they saved millions of dollars in cloud infra costs... But it's a little hacky. It's not the ideal way that we want to be managing our infra, so I'm glad that the GO team is always listening to those concerns and putting out new features to make it easier. + +**Johnny Boursiquot:** \[24:21\] Right. Adding some official tooling for tweaking the garbage collection process. + +**Nishant Roy:** Exactly. + +**Johnny Boursiquot:** The Java world is sort of notorious for all the flags and optimizations, and all the bells and whistles, all the buttons you have access to for GC tuning... So then hopefully, Go doesn't necessitate that level of customizability. But again, as you say, the fact that the Go team is listening, especially for customers that have -- basically, if you have to tweak GC settings, then you're doing it at a scale that most people simply aren't right, right? + +**Nishant Roy:** Right. + +**Johnny Boursiquot:** It's not often where we need to basically tweak what Go does out of the box. So this may not be a concern for a lot of people, the vast majority of people, but for those that do need it, it's good to have non-hacky ways of going about tweaking those things. So that's pretty cool. + +**Nishant Roy:** Exactly, yup. I spent about six months of my life, I think two years in, strictly just analyzing how GC was impacting our service, and how to improve it. I've written a couple of blog posts about that as well. It was really interesting, it was fascinating, but I know now with this new flag I could have saved at least four months of those six... So here we are. + +**Jon Calhoun:** It's kind of nice that they took their time with it, at least a little bit, because when Go was first released... You have to imagine - at least in my mind - it's young enough that they didn't have enough customers using Go in high enough production environments that they really had enough data to decide what needs to be done, versus what necessarily doesn't... At least that would be my perspective as an outsider. Maybe I'm wrong. + +**Nishant Roy:** No, I agree with that. + +**Jon Calhoun:** So it's nice they've taken their time and they're trying to figure that out as people are using it at scale... But that's something that in my mind would be very hard to have from the get-go, because it's like trying to optimize a page when you have 10 users, for a million users. It's like, I have no idea what my bottlenecks are actually going to be at that point. + +**Nishant Roy:** Yup. And the good news is they did have sufficient tooling for you to understand your system usage, if needed. That's something that came packaged in from -- I guess I don't know how early on, but things like pprof, and memstats, etc. that allow you to actually understand how your system is performing, analyze that data exactly how you need to understand the bottlenecks in your system. + +One thing I'd love to see, if the Go team is listening - if we can have an official guide on how to use these tools a little more; maybe some tutorials for people starting out who have never used flame graphs before, who don't understand how profiling works, or how to best read it. I think that'd be a really great way to further the adoption of these tools, and make it much simpler for everyone to understand how to tune their systems. + +**Johnny Boursiquot:** To that point, there's an interesting dichotomy, I think, within the Go committee right now... Because we've had talks and blog posts and things written on performance tuning and optimization for Go in general, be it in terms of analyzing allocation, memory usage, how many threads you have running at any given time... Like, all that tooling exists. But we also tell developers "Well, don't worry about premature optimization", right? \[laughs\] + +**Nishant Roy:** Right. + +**Johnny Boursiquot:** Which means like "Don't really look under the hood that much, don't run pprof, and don't analyze what your application is doing, because that might be premature optimization", right? "Only go looking for these tools if you suspect you have a problem with performance." But at the same time, a lot of times you end up building things perhaps in a sub-optimal way. And that ends up pretty much receiving more traffic than you anticipated. Or maybe you go through a growth period, kind of like Pinterest did, where once upon a time you were a small organization that was not public, and now you are public, and you're getting a lot more traffic... Those things - you don't go back and rewrite some of those things you were earlier on just because you now have more traffic, right? + +\[27:50\] So you end up with this sort of a legacy built-in performance, suboptimal processes and tooling and services that are doing things that is way more than they were designed to do. Now you kind of have to have an engineering effort to refactor things. + +I'm curious if you've experienced this need to go back and change things and refactor things and make things faster, and obviously, how Go made that easier or harder? + +**Nishant Roy:** Yeah, I think you're right. I mean, it's an impossible decision to make, right? Because on one hand, if you're spending too much time on optimizing early on, you're not going to make it as a startup, especially given this environment we're heading into now... Speed of everything is of most importance. On the other hand, like you said, you end up with these legacy systems that don't then support hypergrowth, or supporting new products. + +I think one big thing that we've seen over the last probably five years to a decade or so is widespread adoption of video. So folks or companies who didn't really optimize early on for, let's say, content delivery for instance, suddenly you're delivering video, which is obviously a lot more expensive, and people are having to reconsider how they built those systems, or in some cases even rebuilding those systems to operate at this new scale that customers now want. + +For us, for my team in particular, the biggest scale challenge we were going through was the team growing. So like I said, my team specifically was ten people, now the entire ads infrastructure org is probably getting close to about 100. The ads team is probably three to four times that number. So the question we were asking ourselves is now that we're going through this phase of growth, building a lot more products for a lot more users, for a lot more partners, how can we continue to keep that same level of velocity that our pinners are used to, our partners are used to, and us as engineers are used to, with this growing org, without causing more incidents or without having people stepping on each other's toes? And I think that is, to some extent, the eternal quest; we're never going to really hit a perfect happy -- I don't know what the right word is, but middleground between all three of those factors. But that is something we're constantly evaluating and tuning for, is how do we build our systems in a way -- either can we build better modularity, can we have more config-driven systems where folks can make changes without needing to understand how the rest of the system works? + +For instance, for rolling out new experimental models, do I need to actually go and read exactly how those models are going to be chosen? Do I need to understand what features have been fed into my model? Do I need to understand how those scores have been used? Or can I just go in and make a simple JSON change or something, and say "For this subset of traffic, which may be country, or surface, or device, for iPhone users coming from Canada on the search page, I want you to use this model, with this percentage of traffic. And I don't care how the rest of the system works." + +So those are the things that have been really on our minds constantly, and something we're looking to continuously improving. In fact, the pre-submit test framework was one way that we did that, is anyone can go and make a config change to add a new metric and add a new slice, without affecting how any of the rest of the system works. + +Our actual ads delivery system isn't fully there yet. We're continuing to improve this for developers, making it easier for them to make changes, a) without bringing down the system, b) without blocking other folks or needing to understand how your change interacts with someone else's. But like I said, it's the eternal quest, as the product offerings from the sales side and the product teams get more and more diverse, we realize that some parts of our system just weren't built to support those sort of products in mind. + +Something as simple as if we need to serve - these are some of the problems that we've solved perhaps, but if we need to serve both video and image ads for the same request, do we have a good way of doing that? Not a particularly hard thing if you're building a system from scratch, obviously, but once you've built it with a certain assumption, just these basic things sort of start to fall apart, and then you need to go and look into "Should I just be redesigning this whole thing, or is there a quick and easy way for me to get this off the ground, and then go in and redo it to unblock the next big thing?" + +\[31:48\] And again, what ends up typically happening - I think this is probably true for most big companies - is you obviously need to get that MVP out, so you do something a little hacky to start out with, and then you retroactively go in and ensure that a) it's not gonna break anything, and b) how do we make this a more pleasant experience for developers and product managers alike? So that's the wheel that's always spinning, and we're trying to stay ahead of it, but we're usually playing catch-up. + +**Jon Calhoun:** Jonny had asked you about existing services where you didn't take all that stuff into consideration... Do you find yourself now when you're building something new that you actually look at things like flame graphs, and garbage collection, and try to optimize those things upfront, now that you know you're at the scale where that stuff actually matters? + +**Nishant Roy:** Yes, and the reason it's become a lot easier is because our amazing infra team Pinterest-wide has actually provided better tooling for us to do those things more automagically out of the box. So rather than us needing to write a manual script to go run a profile on our Go system, we essentially have a central system now that we can essentially enable profiling through. So I think those are the sort of investments, once again, that when the company hits a certain critical mass, and we have a team that we can afford to invest in building tools like that, it becomes a lot easier. But till that point, I'd say no; until we had this company-wide tool, it was still up to us as the ads infra org owning our own services to decide when and where we needed to do this. + +And honestly, a large part of it was driven in a voluntary manner; folks who were interested in thinking about performance, and saving in infra costs, saving in latency, would sort of just on their own time go and run these profiles and identify hotspots. Alternatively, if we hit certain system limitations. So let's say modeling team X wants to launch this new model that is suddenly a lot more expensive... Expensive in terms of either dollars, or milliseconds of system latency that could be affecting the end user experience. It's only at that point that we would then go in and look at, "Okay, you want to add 15 milliseconds of latency to the end-to-end system. We only have 120 total. Can we save this somewhere else, or can we optimize your request in a way that would that brings that number down to minimize our impact?" But this is something that we're now looking forward to be more proactive about. + +**Johnny Boursiquot:** Did you suffer from -- like, during this growth phase, did you suffer from team silos? Basically, a team out of necessity having to create their own tooling, adopt their own ways of doing things, and then now trying to do this company-wide having to disentangle teams from the way they know how to do things, because they built things from scratch, or they bought something and have been using it? How challenging was that to sort of rise up to a common level of tooling for everybody? + +**Nishant Roy:** Yeah, to some extent. So definitely for us being on the ads team, like I said, we were the only major online service in Go at that point, so we were definitely building our own tools, we owned our own deploy systems, our own testing systems etc. There has been a divergence there, in how different teams on ads vs. non-ads do things. + +Within ads itself, I think one thing we saw is that because of a lack of that centralized framework from our side, from the ad-serving infra side, different teams sort of built their own products, even within the ads stack; they wrote their own code. And that caused some divergence in a few different ways. So the code quality is not always at the mark that we would like it to be. There's not exactly the same level of test coverage everywhere, or things might just be done differently. So for something as simple as if we're sending a batched request to a different service, or data back in or model inference backend, that logic might be implemented differently by different teams, which then makes it really hard when those folks leave, when things break, to go in and understand exactly what broke and why. So one thing we're trying to really do as the ads infra team now is standardize those frameworks and tools for at least all the different ads product teams to use. + +So going back to batching as an example, my former manager just essentially took one day and wrote a library that made it really easy to run things in batches in Go. So taking away the requirements, but not the ability... Taking away the requirement for engineers to understand how goroutines work, and do so in a way that is thread-safe. So building that batching library essentially allowed them to use this interface and be guaranteed that there was proper panic recovery, there was essentially an easier way to enable locks, if needed, and you didn't need to write your own concurrency code yourself. Things like that have really helped standardize practices across the different verticals on ads itself, and make it much easier for us to maintain and grow the system going forward. + +**Johnny Boursiquot:** \[36:13\] How do you record and communicate infrastructure decisions within your teams, or even to broader teams, other teams that might learn from them? + +**Nishant Roy:** That's a great question. Overall, Pinterest now has a much better standardized documentation template for these large, new services or new changes that are being rolled out... And all of that information essentially is saved permanently on our internal drives etc. For sort of smaller things... So if my team were to go change how we do a certain thing in the ads stack, I think there's a couple different ways. There's no good central place, I'd say, where you can go and find all the large infra decisions, most likely. What we do is we have a Production Readiness forum, essentially, within ads itself. That involves some senior leadership from the ads team, help from the SRE org, and then representatives from the various infra teams, who will essentially evaluate your change. So you're required to come prepared to this meeting with what you're changing, what the stack of your change is, what are the critical metrics that you're monitoring, what services might you be affecting, what failure scenarios do you foresee, what are the mitigation strategies, etc. And then that one-hour meeting for Production Readiness essentially becomes a forum where everyone can sort of test you, to some extent, and test your plan to see whether your change is -- at least do you have enough of a plan in mind to make sure your change is safe to go out in production. And then that becomes a way to make enough people aware of these large decisions. That's sort of worked well for us so far. + +The gap there is obviously in an environment where all those folks suddenly leave at a similar period of time. Hopefully, that doesn't happen. It would require somebody to go in and read those Production Readiness documents again, to understand the reasoning behind everything. + +For larger product launches, we have an email alias within the company where essentially every time you want to launch an ads experiment that is moving some significant top-line metrics, you need to send out an email to this alias before you get launch approval. And at least that serves as providing everyone in the monetization org some visibility into the changes that were made and why. Those emails are then further linked to very, very detailed 50 to 100-pages-long documents, in some cases, going through all the various steps of analysis that went into it. They're usually about 15-20, but I just saw one that was like 93. It's mostly just graphs and tables, but still. All that knowledge is recorded somewhere. + +I think what's worthwhile is probably this behavior of just publicizing it to everyone once the decision is going out, so that even if the top 10% of the leadership team leaves, there's probably at least enough people who know where to find the rationale behind that decision. So I assume that's what's worked pretty well. + +**Break:** \[38:42\] + +**Johnny Boursiquot:** What does an environment like this look like for a junior developer? If you're looking to make impact within this system, which on its face seems very intimidating if you're not familiar with any such process - so maybe you're out of school, maybe you've only spent a couple of years in industry... And here you are, faced with this process. How do you shepherd a more junior member of the team through this? How do you go about that? + +**Nishant Roy:** Yeah, so having been there myself, one thing that worked really well for me, and I encourage a lot of folks do this within Pinterest and other companies, is the biggest opportunity you have for understanding gaps in the system, whether that's performance-related or stability-related especially, are incidents. So getting involved, when there's an incident -- you don't actually have to do anything. You just see that something's broken, you join the channel, or the Google Meet, or whatever it is where that issue is being discussed... You sit there quietly, listen to what people are saying... They're going to talk about the system. After the incident is done, someone's going to write up a very detailed description of how is the system expected to work, why did it not work that way, and what are we going to do about it. That is one of the most valuable sources of knowledge, because it's usually the system owner or an expert on the system doing that piece... And that shows you what the gaps are. And that I think benefits you in two ways; in terms of impact, and some extent visibility, certain issues reoccur. Even though we have remediation tasks to make sure that issues don't reoccur, some things just happen over and over again. + +I about six months ago saw an incident on the ad side that I remember happening 4,5 years ago, when I had been here only like six or seven months. When I saw some of those metrics, it just sort of jogged my memory, like, "Hey, this happened a long time ago, but it happened before. Let me just go in and see what we did that time, and we can try and redo that and see if it works." + +So I've seen that work for several folks who are new to the company. It's like "Hey, I saw this happen two weeks ago. Here's what Nishant did to fix it that time. Let's try this again." And then suddenly, you're the hero that saved the day. And even if that's not necessarily what happens, as part of the follow-ups, taking on some of those remediation items is often something that is not glorious work necessarily, but can sometimes have really big payoffs. It's like, I made this one change that prevented a malicious user request, or a bad user request, or any corrupt data from bringing the system down for the rest of time, potentially, right? And that teaches you a lot about understanding faults in the system, understanding how we analyze those faults in the system, and how to plug those gaps. + +So I'd say that's probably one of the easiest ways... If you have absolutely no knowledge of a system, just get into these rooms with people who are discussing everything about the system, and hypothesizing about all the reasons why something might break. It's just one of the quickest ways to learn a lot about different parts of the company's tech stack. + +**Johnny Boursiquot:** And if you have a suggestion, big or small, do you usually pair up with somebody more senior, who can help you through the vetting process? + +**Nishant Roy:** Yeah, so for my team, and I think a lot of teams at Pinterest usually, everyone new especially is assigned an onboarding mentor, who's typically the more senior person on the team. So that's usually a good sounding board to start with, like "Hey, I have this idea. What do you think?" + +For my team specifically, we have a bi-weekly team meeting where essentially for 20 minutes every two weeks everyone is encouraged to -- there's other parts to that meeting, it's an hour-long meeting, but in these 20 minutes, it's like open floor, just share anything that you've found interesting, any problem that you faced, or any question that you have. So I saw this thing happen, I have no idea how this works, the way it works. + +One example for me early on was understanding how the RPC library in Go works, for instance. It's a little bit of magic how those functions get registered... So bringing those questions up and giving people the opportunity to get answers to those questions, and just bringing up those suggestions and how they can improve systems, team processes etc. has worked really well for us to essentially get visibility and get answers. + +**Johnny Boursiquot:** Awesome. Jon, have you got anything before we switch it over to unpopular opinions? Because Nishant said he brought the heat, so I'm looking for it... \[laughs\] + +**Jon Calhoun:** Nothing else comes to mind. I'm ready to hear this unpopular opinion. + +**Johnny Boursiquot:** Nice, nice. + +**Jingle:** \[44:08\] + +**Johnny Boursiquot:** Alright, let's hear it, Nishant. + +**Nishant Roy:** This might actually be a really popular opinion for this pod... But my unpopular opinion is that working with non-typed languages, non-compiled languages can be a nightmare, especially as your codebase grows extremely large. I will say, shout-out to the Pinterest team. I think they've done a fantastic job of making our Pythonic API easy to work with, easy to read, great documentation, great testing... However, still, if I just want to go in and read one bit of the code, it's really hard to understand exactly how it works, because I don't know -- for lack of typing, I don't know what's going into this function, unless things are really well named, which at this scale becomes hard to enforce... It becomes a little hard to test; you really need solid unit tests for everything, down to like input parameter validation, which is something you don't need in a typed language. And I think it becomes really easy to like accidentally introduce some bugs. + +I know for a fact that one of my friends in the past who worked at Pinterest introduced one bug where they accidentally reused a variable that was defined elsewhere, overwrote that variable's name, because this file was like 900 lines long, and Python didn't throw up any flags for using this variable, like Go would... And it essentially caused a -- I'm not gonna go into the issues, but it caused an issue with the content users were seeing. And that's something we didn't realize till the change went out and was live for a few days. I mean, it was restricted to internal employees only, but it essentially caused -- it was something that I know that Go would have saved us like three days of wasted time, essentially. So that is my unpopular opinion; industry-wide it may not be an unpopular opinion for folks listening to this pod, or with you two... + +**Johnny Boursiquot:** Good. Jon, what do you think? + +**Jon Calhoun:** I mean, I feel like what he's saying has a lot of truth. What makes those languages great at small scale also makes them challenging at large scale. But I would assume most people getting into those things kind of understand those trade-offs... But I don't know, maybe some people completely disagree. + +**Johnny Boursiquot:** My opinion is that there's always going to be somebody who is of a different opinion... But I think opinions, and perhaps - well, this is not an unpopular opinion, but I think opinions should be allowed to change over time. Right? Once upon a time, I was a die-hard statically-typed language person, then I discovered dynamic typing, I'm like "Oh man, Ruby's the greatest thing since sliced bread. Python is awesome" etc. And then I swinged back around and now I'm a Go die-hard because I know what my types are, and I know what to expect in places. + +So I think it's okay to swing back and forth, but I think the better approach, which I think is harder than most people give it credit, is to have a measured approach to how you make decisions about things, to how you judge things to be nor good or bad. There's no such thing as absolutely good, or absolutely bad. You kind of have to say, "Okay, well, what problem am I solving?" + +So today I might use dynamic typing because I'm able to move a little faster, because maybe I'm building a prototype, maybe I'm building something that doesn't need to be statically-typed for reason X, Y, and Z, and you know what the decision is... Which is why I love decision records, right? Including why you pick a framework, or why you pick a language, or what velocity you think you're going to gain from having made these decisions, right? Because if you look back, three months, six months, six years from now, and look at, "Okay, this is why this decision or these decisions were made. This is why we picked Rust. This is why we picked Lua. This is why -- whatever", and you can explain, you can provide the context within which you were making that decision, the information you had at the time, then I think it's perfectly fine to argue for and make the case for using anything you want to use, right? As long as you have the rationale for it, and your teammates at the time understand what your reasoning is, and they're behind it. + +\[48:06\] The disservice we do ourselves is by basically being fanatics of one thing. And I know sometimes I myself sound like a Go fanatic... Because I love Go. I do it every day. It's my favorite language. Hopefully, I'll be writing Go into my retirement. But I know that Go is not always the right tool for the job, so I have to check myself whenever I see somebody coming in with a Ruby, or Java, or whatever it is, to not say, "Oh, we don't want to use those languages, because they're not Go." That kind of ego I think should be left out of technology decision-making... Because technology is all about trade-offs, and you're gonna experience trade-offs, whether it be a language, or framework, software, you buy off the shelf, or software you build yourself, how you run your software, how you're building infrastructure... Everything is a trade-off. So I think we have to be more gracious with our teammates who may like something different than us, and we have to be level-headed, and understand that there's context for every decision made. + +**Jon Calhoun:** I mean, I can definitely agree with that... There's always trade-offs with all that stuff. It's also interesting -- as Nishant was talking about it, I guess, I was also thinking about the fact that depending on when you join a project, I think it can sometimes skew your opinion of whether something's more complicated. Like, if I'm working on a Ruby project that I've worked on my entire life, like 10 years or something, and I know really well, I don't find it very complicated, even if it's gotten very large... But that's because I know it like the back of my hand at that point. But I could completely imagine cases where somebody new coming in, like "I want to make a change", and then looking at it and being like "I have no idea what's--" Like, it's much trickier. Whereas like in that case, I think sometimes a statically-typed language like Go is a lot easier to jump into with no context, no idea of what everything does, because it's a lot more explicit... But I mean, I completely agree, Jonny, that generally speaking, as long as there's good reasoning behind whatever you choose, there's not "One's better than the other." But I could definitely see the argument at this point being "Okay, we're a large company. At this point, it seems like the statically-typed ones are more stable, they're easier to bring new developers on, so even if we think it's gonna be a little quicker in Python, maybe it makes more sense to use the statically-typed language." + +**Nishant Roy:** Right. And it's so dependent on the stage that you're at... Because tomorrow if I was starting a company, needed to hire five folks, how easy or hard would it be for me to find five great Go developers, versus five great Python developers? It's just night and day. I know I could find five Python developers probably in a week, versus Go - one, it's actually hard to evaluate too, right? Like, if I want to evaluate how well a person's writing Go, it's not as easy as like one quick interview, versus Python's probably a little easier, because I know you can -- even if you're not well-versed in everything, it's going to be easier for you to get started and do 80% of your job, and then ramp up on some of the internals as needed. + +The question then becomes "Is there ever a right stage for a company going from 10 engineers to 100 engineers to 1000 engineers - when and if do you decide to go from Python to Java, to Go, to C++, whatever that is? And is that a good idea? Is that a bad idea?" I don't think there's enough data that exists in the industry to truly answer that question. + +There's a couple of pieces floating around around rewrites essentially not being the best idea in the world always, but I've seen cases where rewrites have actually paid off tremendously. I think Snap rewriting their Android app was one example, where it was a tremendous win for them. So it's very dependent, you're right. And Jonny, that was very well said. I do agree with you. + +**Jon Calhoun:** It kind of reminds me of the discussion around like microservices or anything like that, when people are like "Is it the right time to switch to them?" And you'll see a million stories online about how it was successful, and other stories where it's like "It was the biggest problem we've ever had because we've made that switch." There isn't enough data to really say "This is the way you should do it." + +**Nishant Roy:** Yup. + +**Jon Calhoun:** But if you are starting a new company, I guarantee you can find 10 people who want to learn Go. They might not be experts in it. + +**Nishant Roy:** That's true. + +**Johnny Boursiquot:** That is true. That is true. Yeah, I think that is way more likely than to find actually fully-grown, so to speak, Go developers. That's true. That's true. Jon, did you bring an unpop? + +**Jon Calhoun:** I did not. + +**Johnny Boursiquot:** Fine... + +**Jon Calhoun:** I'm too depressed with the snow outside... \[laughter\] + +**Johnny Boursiquot:** Maybe that is your unpop. There shouldn't be snow at this time of year for you. + +**Jon Calhoun:** There really shouldn't be snow at this time of the year. It's not even Thanksgiving yet. Come on. + +**Nishant Roy:** It's a little early... + +**Jon Calhoun:** My wife now wants to get Christmas decorations out, there's snow outside... It's really hard to say no. + +**Johnny Boursiquot:** Right, right, right... + +**Nishant Roy:** It's basically Christmas once it starts snowing, right? That's how it works? + +**Jon Calhoun:** Sort of... Honestly, the biggest reason why I say no right now is because I'm like "I'm too busy right now." After Thanksgiving it usually dies down a little bit for me, so I'm like "Then I can help with Christmas stuff if you need." I'm not trying to ruin the party. + +**Johnny Boursiquot:** This was a very nice episode. Thank you, Nishant, for coming and talking about the kinds of things you're working on over at Pinterest, and the challenges the team has experienced, and your role in it... I wish you continued success and growth in your endeavor. It was awesome having you. + +And Jon, thank you so much for being my co-host last-minute, always good having you. With that, we will say goodbye to the internet, and listener, we will catch you on the next one. diff --git a/How can we prevent legacy from creeping in_transcript.txt b/How can we prevent legacy from creeping in_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..786400f440eb03efda95107389eacbe79f704f2d --- /dev/null +++ b/How can we prevent legacy from creeping in_transcript.txt @@ -0,0 +1,477 @@ +**Angelica Hill:** Welcome to Go Time! Today we're gonna be talking about how to prevent legacy code from creeping in. I'm joined by a wonderful array of guests, as well as my lovely co-host, Kris. Hello, Kris. + +**Kris Brandow:** Hello! And yes, that is my voice you hear out there, listeners. I am finally back, after four months. It's been quite a while, but I am very excited to be back. + +**Angelica Hill:** It's been far too long. And along for the ride with us, talking about how to work with, avoid issues arising from, as well as just how to deal with legacy code on a day-to-day basis, we have firstly Dominic St-Pierre, who operates a small consulting company in Montreal. He is the maintainer of a FOSS alternative to Firebase called StaticBackend, and is the author of Build SaaS Apps in Go. Welcome, Dominic. Welcome back, I should say. + +**Dominic St-Pierre:** Thank you. + +**Angelica Hill:** Pleasure to have you. + +**Dominic St-Pierre:** Happy to be here. + +**Angelica Hill:** Then we've got Misha, who is a first-time Go Time guest, but also (I'm told) a first-time podcaster in general. Very privileged that you decided to grace us with your first podcast. How are you today, Misha? + +**Misha Avrekh:** I'm good, thank you. + +**Angelica Hill:** \[03:57\] And then we have Jeff, who is an associate backend engineer at the New York Times. I should also say, Misha is also at the New York Times and is a software engineer. Welcome, Jeff. How are you? + +**Jeff Hernandez:** Great. I'm excited to be here. + +**Angelica Hill:** Excited to have you. And then Jon, we've been here before... Jon Sabados, who is also at the New York Times. I promise I'm not biased. They just happen to be people I spend most of my life with... He is staff software engineer. Welcome back. How are you today? + +**Jon Sabados:** Doing well, and good to be back. + +**Angelica Hill:** Awesome! So before we dive into the nitty-gritty, I wanted to ask the group, what is legacy code? How do we even define what we call legacy? Dominic, do you wanna dive in? I know you've been working with something that I'm sure we would all agree is legacy... + +**Dominic St-Pierre:** Yeah, I think in my case it's really a "passion." To me, legacy is when a company starts to have difficulties hiring developers to work on their software, you can think of that as legacy, I think. Also, when there's no unit tests, not really good documentation, it's very old software that nobody other than the original author can maintain anymore... I'm talking about like 20 years old software. So this is one of the things that I do in my consulting. + +So yeah, to me when the tooling starts to not work for you anymore, and things like that. In a nutshell, that's a little bit my definition of what legacy software is. + +**Angelica Hill:** So you said very, very old... So we're on Go Time. Go to me doesn't feel like a very old language. So I wonder, maybe - Jon, I know that you've had a little bit of experience working with this... But at what point did Go applications turn into legacy? + +**Jon Sabados:** Well, that's a good question, because you can actually -- and again, it depends on how you define legacy, but at least in our scenario it's very possible to have fairly recent code that is suddenly legacy. One example of where you might hit that is if you have a system that was implemented in a fashion that really didn't suit the needs, so you're forced into a greenfield rewrite, and supporting two systems. That's an easy definition of legacy code. We've got a new thing that's going to do the stuff, but an old system that we have to support. I might define legacy code as any type of code that your engineers hate to work on. + +**Angelica Hill:** Okay, so any code that I don't wanna work on, I can say "Oh, it's legacy. It was written last week, but it's too old; I don't wanna work on it." \[laughs\] Okay. And then I wonder, Jeff, specifically - I know you've joined the New York Times preferably recently, and I know that you came to a kind of legacy system... Coming in, how do you identify whether something you're working on is legacy or not? Is it that like everyone on the team is saying "Oh, this legacy system..."? Are there indicators when you're joining a new team you're like "Oh, this feels legacy..." + +**Jeff Hernandez:** I feel like part of it is everyone just says that this portion of our services are really difficult to work with, they're kind of considered legacy. But I don't know, I feel like maybe my own definition of what is legacy is more broad than what's been given so far. Legacy is definitely for me something that is already there by the time you join the team, regardless of whether people like to work with it or not. That's kind of like software itself - we're building on top of what has come before us. I guess that would be my own definition of legacy. But I definitely take into account what other people on the team who have been working with this system have to say about it. Our old system is very difficult to work with; it definitely informs my opinion on what I consider legacy. + +**Angelica Hill:** \[07:59\] Yeah, for sure. And then obviously, when we're talking about legacy, I feel like not only are we talking about legacy architecture, legacy infrastructure, but we have legacy dependencies. A lot of our apps have hundreds possibly of external dependencies. At what point and how do we go about defining our dependencies as legacy? I don't know whether Misha, you wanted to speak on how do you think about dependencies and when they become legacy. + +**Misha Avrekh:** Right. Well, for example the project that I work with now, which is written in Go - it's about two years old at this point, so probably not legacy strictly speaking, although there are parts of it that are starting to feel like legacy code... I can talk about that separately later. But yeah, as far as external dependencies that are legacy systems - we have a few of those. The way we've dealt with them is by caching as much as possible. Caching the data that's returned by the external dependencies to reduce the latency, essentially, of dealing with external legacy projects. + +**Angelica Hill:** And is that something that you do for all of your dependencies? Or is that something you kind of have a specific list, or -- like, how do you make that decision? + +**Misha Avrekh:** So on our team we made those decisions specifically based on our latency metrics. It doesn't really matter whether the external dependency was written in Go, or Java, or what have you. If we feel like we need to optimize for latency, we try to cache it. That's been our approach. So yeah, monitoring and testing for latency, in this case. + +**Angelica Hill:** For sure. And then, obviously, when we're talking about legacy, in its many different definitions, because as we can tell from even these initial definitions, there is kind of a fluid way of defining, we can't just identify and then say "Great! We're gonna rebuild, refactor..." We have to maintain it, whether it be for a month, a year, many, many years... So I'd love to hear a little bit about what it is like, both from like a day-to-day, personal, but also from a technical point of view, to maintain legacy code. + +I know Kris, our wonderful co-host, has been kind of spearheading a maintenance series, so I know this really tees into the various different topics you've been exploring... Perhaps for our listeners who aren't fully familiar, I'd love to kind of pass over to you before we dive in, to talk a little bit about how we've been exploring and thinking about maintenance. + +**Kris Brandow:** Yeah. I guess I would start with saying my view of maintenance is definitely not in some ways as maybe negative as everybody else seems to think it is. When I hear the term "legacy", I think there is this feeling of dread that a lot of us have, but the term legacy can actually be a good thing, too. Like, the legacy that you leave from the life that you lived, and things like that... And I take the same view that I take for something like debt. It's like, there's good debt and there's bad debt. There's good legacy, there's bad legacy. And I think that really informs the way that I also see legacy when it comes to code. Anything that's hit production to me is legacy. It's out there, it's something we have to deal with, it's something that we have to maintain, and we have to understand, even if we don't necessarily like that code at the end of the day. + +So I think before it gets to production, when you can still change it and mold it and do what you like with it, I think at that point - yeah, we're still dealing with something that's not legacy, it's not debt yet, it's not something that we have to necessarily deal with in the future... But once it hits production, once it's out there, it becomes very difficult for us to pull back. + +Now, of course, we can get into the semantics of what does it mean when something actually gets to production, but I think that's the way that I at least conceptualize this... Pulling it further back than anybody else has, and really seeing it as all of our code is legacy, and it really comes down to which code do we think is good legacy and which code do we think is bad legacy. + +**Dominic St-Pierre:** \[12:01\] Well, yes, I really like what I'm hearing so far... One thing that jumped to my mind is to me a Go codebase, even though it might not have been designed the proper way or whatnot, it's still easier to refactor, it's still easier to change the status of the codebase... Because there's probably unit tests and whatnot; there's probably lots of tests in a modern codebase. So to me, I understand all of this, but I feel it might not be the same thing here. When you're stuck with something and you have no documentation, no tests to exit yourself out of this situation... Because I really like what you said - everything is legacy; yeah, I understand that. But it seems to be that when the budget is there and when the team is there, you still have options. Whereas I've worked at small companies that just cannot invest into any kind of refactor, or even a rewrite at that. So you still have to maintain the codebase. + +**Kris Brandow:** Yeah, and I think that's definitionally just bad code, at the end of the day, if you can't maintain it in any way... But I'll also say, I've run into lots of circumstances where we have lots of tests and they're not good. They give us a false sense of confidence. Or you have lots of tests and they're written in the wrong way, so whenever you touch something of the codebase, you have to go refactor a few hundred lines of tests. + +So I think even in your example there, it's like - yeah, tests definitely can make a codebase better, and generally they show the mark of a better codebase, but I think there's ways in which even those types of codebases can be a pain to work with. But I would absolutely say, a codebase that has zero tests, no documentation, none of that, is the worst kind of legacy codebases that we can deal with. + +**Dominic St-Pierre:** How do you know then if your test suite is going to be good for the long run? + +**Kris Brandow:** Good design of your software. I think if you've gotta kind of push the boulder up the hill, or really go back to the genesis of the stream... Like, if you're gonna build a dam, it's a lot easier to build a dam across a stream that's two feet wide than it is to build one across one that's 200 or 2,000 feet wide. So yeah, it's like, if we put a lot of the stuff downstream to know if we've done the right thing, we can, through various techniques, understand how good it is. But I think it's much better to try and put in ways of knowing how correct your code is upstream. + +But of course, a dam across two feet of stream is quite useless of a dam... So there's a middle ground here where we need to find the right place to put things. And I've definitely found that right place to be in good design documentation and good design ideology for the software that we build. But I'm sure other people have different opinions on that. + +**Jon Sabados:** And I think I'd add a modern codebase does also not mean tested. There's plenty of fresh code that I see written that engineers oftentimes don't like to write tests, so that is a pervasive problem that seems to plague all codebases, is stuff just isn't tested. + +**Misha Avrekh:** Yeah, one thing I can say we've done to mitigate that problem is to keep an eye on the coverage thresholds and actually building the test coverage thresholds into our testing and deployment pipeline. So if someone adds new code without adding tests, the pipeline actually fails. + +**Jon Sabados:** That is definitely an interesting approach, but I've also seen it backfire horribly in the past. I've seen engineers, and this actually got someone talking to, this is in java land, put an entire class on a single line, because they were just annoyed by the test coverage requirements. So you can kind of get malicious compliance with those code coverage requirements. And also, depending on your language -- Go is not as bad, but in Java land there's all sorts of... Like, you've gotta catch these exceptions that practically are never going to happen, but you might not even be able to write tests to get there. So I think you do have to be a little bit careful with coverage requirements. I think it's a good metric to have and to track, but there's a danger zone when you start failing builds because of it. + +**Angelica Hill:** \[16:10\] I'm interested to hear a little bit more about -- you referenced Java, Go... Are there aspects of Go as a language that make it easier for you to either "avoid" legacy code, or avoid it being a whole huge refactor project? Because in fact, refactoring bit by bit, backwards-compatibility comes to mind... Are there aspects of Go that make it easier for us to avoid legacy creeping in? + +**Jon Sabados:** I think there's some, in that it's a simpler language. Up until recently you didn't have generics, so you're not gonna have people going nuts with generics and creating really complex things, like trying to deduplicate everything through generics... So that helps. But I think the biggest thing that Go has going for it with the lack of legacy and maintainability is that it's a newer language. It hasn't been around for that long, so you don't have projects that have been sitting under that were conceptualized 20 years ago, to carry that forward. But there's a lot of Java projects out there that have been around for a long period of time. So I think there might be a false sense of security saying "Well, Go is going to have less legacy" just because you don't see it as much, but that's also a function of the language being newer and therefore not having the crap that's been around for a long time. + +That's also something we learned every time we write code and every time we develop, so there's a lot of patterns that we might have been doing 20 years ago in Java land, that we've maybe learned better ways than Go. But again, that's not so much a function of the language, just of the newness of it. + +**Jeff Hernandez:** I just wanna add that what Jon was saying kind of made me think about what this topic of legacy will be like for Go in 10-20 years, in terms of when it gets to be old enough to -- the same age as Java is currently, and we have a lot more Go projects around that need maintenance... It's just an interesting thought experiment to be having. + +**Dominic St-Pierre:** Well, I can tell you about stories with .NET, even though -- let's say it started in 2001. So in 2005 there were already projects that you could not even migrate to the next version. So to me, the tooling that the language provides, which is very solid in Go, in my opinion, and the fact that it's backward-compatible - if they can maintain that for a long time, it will diminish the effect of that, I'm certain of that. Because in my case, at the moment, the tooling of Microsoft is my main issue. It's not the code per se; it's not that at all, but the tooling. + +**Jon Sabados:** I guess I'd be interested, because I've never really worked in .NET. I've always been allergic to Windows. But the Java side of things, they strived for a long time to make the JVM backwards-compatible with prior versions of Java, and that didn't necessarily stop the need to have complete rewrites, or like you being stuck on an ancient version of Java because whatever new thing. + +I've actually even seen the new thing with Go, where they're seeing projects that were running on old versions of GAE, that have pinned their version of Go to 1.11, or something, and upgrading those has become a nightmare. So again, it's less likely in Go, and it is for sure easier to generally upgrade versions of Go than my experience has been with Java. But again, in Java land you're also dealing with a ton more dependencies, because there's this massive ecosystem that abounds, where you're not having to do everything. In Go, a lot of that ecosystem doesn't exist, so you're reimplementing a lot of things, and not sucking in as many dependencies. The benefit of that is when it's time to upgrade, you're only worried about your code, not a million different libraries. I'd be curious to see if that easy to upgrade continues as the Go ecosystem grows and as it becomes more complete. + +**Dominic St-Pierre:** \[20:09\] Yeah, but you are talking about the platform. So the Google App Engine is a platform; it's not really Go's fault or responsibility, in a sense. You can still take codebases from the early, early days of Go and compile that with the new compiler. This is something that you just cannot do in the .NET world. So that's mostly what I'm talking about - when the tooling is failing you, it's a different ball game, because you have the code, but the technologies that were suggested to use when a company makes a choice of going with something, and three years later it's not even supported by Microsoft anymore, then it's a different things than that design, I think. + +**Jon Sabados:** Yeah, and I guess there's a different past user story there, and experience, because - again, coming from Java land, everything will compile just fine, but the things that I've encountered, that I've seen, that make upgrading hard isn't so much the compilation, which it sounds like in .NET that might be a thing... But it is the platform that is the hard thing to upgrade, from my past experience. So that's where I think that Go won't be immune to the platform increasing scope and hardness to upgrade with that as it becomes more robust, or there's more things, it's gonna be like GAE and whatnot. So I definitely agree with you - yeah, if your vendor is making stuff that doesn't compile it in three years - on my God, that's terrible. And I'm sorry. + +**Dominic St-Pierre:** Yeah, I can give another example. So you know, TLS 1.1 just vanished recently, but there is still old applications that run under an old version of Windows Server 2008, which is not supported anymore... But because their migration path to the next version of .NET is not an easy one - I'm not saying it's not possible; it's not what I'm saying. But there is lots of investment that a company needs to do to migrate to the next version. So that's another aspect of legacy... When the OS and the bare-bone communications system, the protocol starts to fail you, what do you do? So I had to start writing some very small piece in Go that just do HTTP call, because I needed to use TLS 1.2 or 1.3. So yeah, it's a different ball game, I think. + +**Jon Sabados:** That does bring up is working in legacy with a closed source system is different than working in legacy with open source systems when it comes to those dependencies... Because that is one thing I've always liked about working in all the various different techs I've been in open source. You know if Oracle buys Java, somebody is gonna be like "Well, we can take the JVM and run with it still." OpenJDK... It's been a while since I've been in Java land, so I'm not sure how successful OpenJDK has been, but that is a possibility, right? When you've got closed source, you might be forced into complete rewrites. + +**Angelica Hill:** So do we feel like that's one of the differentiators of Go, in that we have this vibrant, wonderful community where if there's anything that is detrimental, it's causing people to feel that it's legacy in some way in terms of the tools provided? That there's gonna be mass opera in the comminuty saying "Hey, can we fix this, please?" + +**Jon Sabados:** It definitely can help, up until the point that there's enough bifurcation that there's a fork, which hopefully won't happen. It's rare to see languages truly fork like that. So I think largely yes, but that forking is also a potential danger. + +**Kris Brandow:** Languages except for JavaScript... \[laughter\] + +**Jon Sabados:** Oh, God, JavaScript... Yaay! + +**Break:** \[23:54\] + +**Angelica Hill:** So we talked about the pains of maintaining, and the need to - as you've kind of touched upon, Dominic, kinda of hack into these small solutions to make sure your apps continue running... But at what point do you throw your hands up and say "No, we need to completely rewrite this"? Well, one, relying upon having the time and investment to do that, but we'll touch on that later... But if you feel like there is the time and investment to do a rewrite, at what point do you go from maintaining to really advocating that we need to rip this out and plant a completely new tree? + +**Dominic St-Pierre:** One nice thing about legacy software is that at some point they kind of run without any intervention from any developer. So that is something that could help with a complete rewrite. If the company is able to do that and seize the value in doing that, then yes. I think at some point the maintenance -- because after 20 years you are not really maintaining any codebase. You're not changing lots of codebase. It's everything around the software that is failing. So it's not the code. The code is battle-tested. So even though there's no unit test, I trust those old software completely. They have been fixed, there is not really any bugs anymore... They are pretty stable, but it's a huge investment to decide to rewrite the software, especially old applications like that. So yes, I don't really know what to answer to that, but maybe someone else has another point. + +**Jon Sabados:** I don't know that there is a hard and fast rule for when to do a greenfield rewrite other than it should probably be one of your last resorts, because it is an incredibly expensive thing to do. Oftentimes you have to do it bug for bug, depending on your downstream clients, and that can be a nightmare to do. + +And you know, different scenarios can offer different reasons. If you're doing a cloud migration - well, there may very well be a reason to do a greenfield rewrite of fairly recent stuff, in that case, because your underlying platform is changing dramatically. + +**Misha Avrekh:** \[27:41\] Based on the projects I've seen, there definitely needs to be a close alignment between business and product, as it were; business requirements and the tech requirements and what needs to be done on the tech side... Because that's when you start seeing possible solutions, like "Oh, maybe I don't need to rewrite the whole thing. Maybe I need to rewrite just parts of my legacy code, and those will work as standalone pieces, while the legacy thing continues to chug away in the background." And also, of course, the expense part that needs to be understood and underwritten by your business, if that is to happen, if any kind of rewrite is to happen. + +**Jeff Hernandez:** I'm curious as to when it is important for business and product to be aligned, but also taking into account the engineering quality of life. If I'm spending all my time -- like Dominic said, I'm not fixing the code itself, I'm trying to keep the code standing, keep it running, because everything around it is just crumbling... Speaking for myself, I don't feel like I would last very long if that's all I was doing at a company. Like, yes, I'm willing to support legacy systems, but that's not all I wanna be doing. I wanna be creating something new, or at least extending it, or something. I don't wanna just be maintaining forever. I feel like I won't grow as an engineer if that's all I'm allowed to do. + +**Dominic St-Pierre:** Yeah, I can totally relate with what you're saying. \[laughs\] + +**Misha Avrekh:** Yeah, if your engineers -- oh, sorry. + +**Dominic St-Pierre:** I'm alone at the company at the moment, so they are simply not able to hire anyone else. So I'm kind of stuck there. + +**Kris Brandow:** I do think there are some engineers who actually do like just tinkering on old codebases and not adding new features or anything, but just keeping something running. I think there are people that just get an immense amount of joy out of just doing that. But I definitely think trying to put engineers that don't wanna do this sort of thing onto a codebase where that's all of the work is not going to work out very well for anybody involved. + +**Angelica Hill:** And do you feel like that is more in keeping with just like personal preference, or is that perhaps to do with where you are in your progression as a software engineer? I mean, purposefully we have kind of a range of different levels on this call, because I wanted to really get an understanding of what is it like coming it at like Jeff's level, and very early, looking to learn, looking to grow at a very rapid pace, versus someone like Kris, Jon etc. who are at that -- I don't wanna say end phases; you have many more years in you... \[laughter\] But you see what I'm saying. You've worked on enough problems to have learned a massive amount, and perhaps to find a new problem for you to solve is more difficult. Or do you think it's just personality? Go on, Kris. + +**Kris Brandow:** I think it's a different type of engineering, rather than a certain level of it. I think there are people that want to be maintenance engineers in the same way there are people that want to just be pure R&D, pure prototype engineers. They wanna build stuff, they never want any of that to be in production, because they know it's not gonna survive production, and they know that they don't want to have to maintain it. They're like, "I'm just doing research, I'm just proving out an idea. That's what I like doing." And I think there's leveling within those types of engineering, but I don't think they stack on top of each other, or anything." + +I think there's like, okay, maybe you're on one far end of the spectrum, you really like getting one of those old, nasty codebases, and you're like "I'm not gonna add anything new, I'm just gonna fix it up. I'm gonna add tests, I'm gonna document things, and I'm gonna turn this into a really nice codebase to work within." + +At the other end of the spectrum are those prototype people that are like "I just write code that I just clobber together and make something work. But there's no way this will work for the long-term." And I think the majority of engineers fall in the middle. They like doing some new stuff, they like maintaining some old stuff, and that's just kind of it. And I think it's important that we start seeing that as a spectrum, instead of trying to see that as like "Well, who's better than the other?" There's no better, there's just like "We need a bit of all of them." You need to keep your organization balanced. + +I think that's one of the criticisms I have for a lot of tech companies right now, is that they are very much focused on the other end of that spectrum, that is the R&D engineers and the people that want to build newer things, and the results of that is a lot of codebases that don't wind up getting properly maintained because there's no one in the company that wants to do that maintenance. So there's no one in the company that's advocating for it, even though everybody in the company is aware that it needs to happen. + +\[32:19\] I think that happens, a lot of teams sitting there being like "Well, we really want to work on all of this old, legacy stuff, but where is the time to do it? How are we gonna make this happen?" So I think getting some of that balance is good. + +But to answer your original question - yes, I think they are different types of engineering, not different amounts of time that you've been doing software engineering. + +**Angelica Hill:** For sure. And off the back of that, people know that a certain platform or a certain part of that technology, that stack is legacy; it's known about. You're slacking about it, "Oh, this annoying platform. Oh, this annoying..." Like, it's all known. But how do you then advocate for a rewrite, or for the time to maintain? What are, from your experiences, the key things that you have to bring up to get buy-in from product, business etc. to actually give you that time? Because that is, I know, a challenge that many people face. So how do you overcome that? How do you make them care? + +**Jon Sabados:** Yeah, that's a hard problem, because effectively, that maintenance is a cost center, with no -- you don't see the benefit immediately. Generally, my experience has been everyone is short-sighted and more concerned about getting their current feature out than they are about the ability to add another feature six months down the line. And I think the best argument that I've had is working with business and explaining that this tech debt is something that actually has a cost down the line, and will impede future things. + +Most people understand the idea of debt; there's only so much you can have. And explaining it as tech debt and being like "But we will get to the point where all we can do is pay our tech debt, and we won't be able to make any new investments." And oftentimes, that falls on deaf ears, so... + +**Kris Brandow:** I definitely think structuring it as a conversation around debt and the financial aspect of it can help people, but also just like -- I think a lot of it has to do with process and project management, of like "Okay, when we said that we could deliver this feature, we said it would come with this amount of debt, this amount of stuff that we have to do later. Well, here's the later part, and we have to go do that, and you said that we could do it, so you're gonna have to actually stick to what you said you were going to do and give us the time to do this." + +I think it does take a lot of fighting and advocating, but I think part of the problem is that in general, the teams I've been on have been kind of not great at explaining exactly what that technical debt is. It's like, we took on some technical debt, but it's just kind of this mysterious thing. It's like, "Oh, you've gotta go write some tests. You've gotta go do some other stuff." + +So I think it's less that product and business don't want to pay for it, but the cost of it is not known, not quantifiable. It's like, if we talked in debt, but you never found out what the interest rates are. It's like, "Okay, well, you're telling me I have to pay down this debt, but is this debt that's at a 0.5% interest rate, or at a 25% interest rate? Because we're gonna deal with that in very, very different ways. 25% interest rate - we've gotta pay that down immediately; stop everything else, get rid of that. 0.5% interest rate - I'm just gonna let that sit there, because it's not gonna cost me much in the long-run." + +Not being able to talk about things in that kind of quantifiable level I think is what holds back a lot of engineering organizations from being able to pay down that technical debt and handle that legacy. So I think, at least in my experience, the closer that I've gotten to talking in terms that are more concrete, of like "Here's what we need to do. Here's the tactical project. Here's the plan for it. Here's how much time it's gonna take, and here are the benefits that we're gonna get out of it at the other end", I've been very successful in selling that. It's definitely the other side of it when it's more just like -- you can't get a good grasp on it, when it's not quantifiable in terms of the business or in the product team. + +**Jon Sabados:** \[35:58\] Quantifying it can be interesting too, because yeah, there's some things you can be like "Yeah, this is a huge thing that's going to bite us hard." But when you're trying to quantify in exact terms for like "This may or may not be a problem, depending on what we do", which oftentimes maybe the answer is "If it's not a problem, don't worry about it." But I guess quantifying that debt is hard. + +**Misha Avrekh:** Would it be fair to say that that is exactly the different between legacy code and technical debt, where technical debt is something that needs to be repaid and that will potentially grow with time, whereas legacy code, as Dominic said, it might just sit there and work for years until anyone notices? + +**Kris Brandow:** I feel like legacy code is like the technical debt that we haven't paid down for so long that we kind of declare bankruptcy on it. And the way that people think about legacy code - it's that code that's just like... Whenever you make that declaration of a greenfield, you're like, "Okay, we're done with this. And since we haven't planned well for our debt in the first place, we haven't planned well for how we're gonna replace that debt, because we also haven't quantified it for ourselves..." So I think it starts off as "We want to declare bankruptcy, but we don't actually know how to do that", so we have this thing that kind of sits there and lives on continuously... But I'm definitely open to other people's interpretations of that as well. + +**Dominic St-Pierre:** To me it boils down to once the company is near reaching a point where it will affect their bottom line, and when you're not able to find any engineers to work at your company - this is terrible. You've already reached a certain point that you should never have passed, in my opinion. But yeah, before going there... + +The other point from just a minute ago - I think we might see a new kind of engineering type evolving... Because if you think about it, how many software are we creating these days, compared to 25 years, 30 years ago? It's incredible. There will not be enough of us to maintain every single piece of software that's being created at the moment. I think we will see new jobs evolve. And I'm not talking about AI. An AI can help maintain software. I'm looking at you, GitHub Copilot. But yeah, my point is there should probably not be a difficult argument to be said about if a piece of software should be left as legacy like that, because it will impact the bottom line of the company at some point. And that should be the metric that every businesspeople should understand. + +**Kris Brandow:** I think there was also a question about how we might start quantifying what this debt looks like for the business. And I think kind of using their own terms would be helpful here. Businesses tend to have really good -- or at least the ones that survive for a long time, tend to have really good risk management apparatuses, so they have really good ways of talking about things probabilistically, and we as engineers hate talking about things in probabilistic/statistic terms. We're like, "That will absolutely happen" or "That will never happen", and those are the only two options. But I've also found some success in talking about things in a more probabilistic way. This can go as far down as story-pointing, kind of changing it to be like "I have 70% confidence that this will take this amount of time", and just kind of bubbling that all the way up. + +Once we start talking in less of these absolute terms, I think it becomes easier for us to communicate with businesspeople and give them something quantifiable. It's like, "Okay, if we don't do this work, there's an 80% chance that we're gonna run into a problem", versus like "If we don't do this work, there's a 20% chance we're gonna run into this problem." And then we can start doing those types of trade-offs, and then importantly, that also gives us something to track, so we can improve there as well. Because once again, if we're trying to figure out what our interest rates are, or figure out how much things are costing, it doesn't matter if we're just making up terms that don't actually come back to reality; that's going to degrade the trust from product and business. + +\[39:52\] So we have to say "Okay, we have an 80% probability of this thing happening. Okay, did it actually happen, and how many times did it actually wind up happening?" Just going back and tracking all of that and tracing all of that, so then every time the business asks, we can be like "Here's the pages and pages and pages of how we've discussed this and how we came to these numbers, and how we came to this probability, and here's the historical aspect of that, and how much it has occurred historically, so then we're actually able to give you some basis to believe us, and kind of build that trust back up." I think that's how you start making it more quantifiable for people. + +But I think, to your point of that type of engineer that we need to develop, I think one of the types of engineers we do need to develop are people that are kind of focused on going into codebases and doing this type of analysis, and figuring out how to prioritize that debt that we have and come up with debt consolidation and debt paydown plans, and help us actually determine "Hey, we want to declare bankruptcy and build a new thing. This is how we're going to do it so we don't wind up with the old things still being there, haunting us forever." + +**Jon Sabados:** That is definitely an interesting thought, and like you mentioned, that is almost like a new type of engineer, a new type of position, because coming up with all those percentages and whatnot - that's like an FTE almost. It's not something that you can just kind of do on the side, while you're doing your feature work. You need somebody who's focused and dedicated to that. + +**Kris Brandow:** Yeah, absolutely. It's a thing you have to be dedicated to doing, and actually want to push your organization forward with doing in the future. And it's like a whole process and a whole framework that you have to develop as an engineering team. We're lacking this in most of our engineering organizations. I've not been to (I think) any companies that have had this type of robust way of talking about technical debt, or legacy systems, or even just future work. Most of the time it's just like we've got a bunch of story points, we've got a bunch of sprints, and we're gonna take a guess, and we're gonna have retrospectives... But retrospectives are complaining about things that went wrong, not about "Did we actually meet what we said we were going to meet?" So I think there's definitely a lot of area here that we could kind of build up something better for being able to discuss a lot of this stuff. + +**Jon Sabados:** Yeah, and I wonder if that also calls up -- there's oftentimes so much ceremony involved; at least my experience has been a lot of engineers are like "Well, this is kind of killing productive time, because if we spend a bunch of time pointing things, and we do the work based on scrums, but no one's holding us accountable for our sprints, and they actually end up devolving into something like business ends up wanting Kanban anyway, so I'm doing all this ceremony work for nothing... Why do I wanna add more ceremony on to that?" + +**Kris Brandow:** Yeah. I think it's the kind of hole we've dug for ourselves in, building a system where we are supposed to be getting these benefits, but then not actually the back-half of that work. Not actually figuring out, "Is this form of agile we're doing, are sprints working for us? Are storypoints working for us? Are we able to actually get the benefits out of them that we think we're getting? Or is this just kind of --" in reality, yeah, as you're saying, a ceremony. Something we just do because we've done it, not because we're getting any intrinsic value out of it at the end of the day. + +**Angelica Hill:** I feel a whole Go Time episode coming on on this exact topic. So I'm interested to hear a little bit more about -- we've talked about getting to that explosion point where it's falling apart, having bugs every day, making on-call engineers' life hell; that is the point at which you're like "We need to do something." Going back to our original question, like "How do we stop legacy code creeping in?", as we talk about accumulating tech debt etc. what are some ways that we can stop getting to that final point of explosion where everything is falling to pieces? + +**Jon Sabados:** I think one thing if you wanna keep from code becoming legacy, I think that testing is a helpful tool for it. It's by no means a solve it, but if you can define your actual problems that you're trying to solve, and the right tests that exercise those problems, that will also help ensure that you've got discrete chunks doing those things, rather than tangled mess all playing together. + +**Jeff Hernandez:** \[43:58\] It sounds like you're advocating for test-driven development... Is that right? + +**Jon Sabados:** I just might be! + +**Jeff Hernandez:** \[laughs\] + +**Jon Sabados:** It's the cure-all, and there's oftentimes a lot of places where test-driven development doesn't work, especially like lab settings or where you don't know what you're going to do, but it is a tool that helps on more than just like ensuring that your code functions correctly. + +**Misha Avrekh:** In my experience, you have to invest in your infrastructure and your monitoring probably before you invest in actually rewriting your code... In the sense that for example if your deployment pipeline takes hours for your changes to get deployed, you need to rewrite the deployment pipeline first, right? Add tests, add monitoring to make sure that you know what your system is doing, what queries it's executing on the backend, or what external calls it's making, or whatever it is. Probably do that before you get to that explosion point that Angelica just mentioned, and then the explosion point will be a lot less painful. + +**Dominic St-Pierre:** Yeah, I like that. Going small piece by piece, and just being in a better way each week probably than you were last week; it feels like a great way to do that, and also probably less scareful and dangerous, if I can say that... Because when you think about that rewriting as an entire software, it is scary, for sure. + +**Kris Brandow:** I think I would say writing documentation, or just comments, really. I think testing is definitely helpful, but I think if I were to be given a codebase that has really good tests, but no comments and documentation, or a codebase that has documentation and comments, but no tests, I would definitely take the latter... Because good comments and documentation tell me the intent of what this thing is supposed to be doing, which then I can go write the tests for myself. + +But when I only have tests and they don't have good documentation and comments, then it's like, I have to assume that the tests were also written to do the thing that was intended to be done, or kind of try and derive from those tests what's supposed to be handled here. + +But I think just going through and being like, okay, here's a function with this name. Here's a type with this name. Here's a method with this name. Does this do the thing I think it does, and then writing that down and making sure everybody has an understanding of that - a real great way to actually a) start exploring and understanding a legacy codebase, but also helps prevent codebases from becoming legacy. Because then other people can catch the bugs and what you intended, versus what's actually written in the code. + +**Jon Sabados:** Clarity of intent is a good one, because that's also like comments, documentation, and even just the coding style as well can go a long way towards that... Because naming is one of those things that's really hard, and people tend to just eventually give up on, or get frustrated and be like "Ah, whatever." But it is worth that time to take care of your naming, because if you can structure your code in a way where the code declares its intention, that's even better than having comments about saying what it's going to do, what it's intending to do. + +**Dominic St-Pierre:** I think that's another strength of Go, in my opinion, due to the packages and whatnot, compared to other frameworks and languages that I've worked with in the past. So you kind of already have a small sense of what something is going to do in terms of business case and whatnot, because they are probably properly placed into the right packages. + +**Jeff Hernandez:** I'm wondering if people have any examples... Because everything so far sounds like having good engineering standards and protocols for your team and organization is a big help in preventing legacy code, I'm curious if others have examples of maybe not just good documentation or good tests and stuff like that, that could help in preventing legacy code... Because a lot of times that comes from the whole team that has to kind of agree upon a set of documentation, and the coding style... But if there's maybe something that you as an individual can take upon yourself and improve yourself as an engineer... I'm totally not trying to talk about myself. + +**Kris Brandow:** \[47:59\] I've definitely found that, at least for me -- well, the outcome of this is usually writing better documentation, or tests, or whatever... But I think just like taking some extra time to stop and really think about what you're trying to do and what you're trying to implement. It doesn't have to be like a ton of time. I'm not saying like take hours and hours and hours. But just like, sitting there and looking at the thing, and then thinking about the problem you're trying to solve for even just a few minutes longer than you would have before I think can really help you perhaps design something a little better. + +And I think also just many iterations of things. So if you write something and you budget your time and plan on being able to write it 3, 4, 5 times, that gives you the space to like write it once, maybe quick and dirty and look at it and say "But is this the way that it should be written?" and kind of get that feeling of if this is the right thing. + +I think a lot of the times we wind up with legacy code, or the bad legacy code, because we didn't go through enough iterations of thinking about something... And I think it can be very difficult to sit down and do that, because we do have deadlines we're trying to hit, but I think that's one of the things that has definitely made me a better engineer, is like pushing back on myself and my team and saying "I'm going to take the extra few minutes here to think this through a little bit more, because it doesn't feel right", and really just listening to your gut and developing your gut instinct over time, to know "Okay, I've written something. Okay, this is good" versus "I've written something and this is rough." And listening to that feeling inside you that's like "It's rough, but I don't know why", and then following that thread. + +I think a lot of times, once again, we ignore that thread, because we have things to do, we have features to develop, and we have something that works, so why are we gonna follow this thread; but I think really just following that thread at the end of the day has definitely helped me write codebases that are much more resistant to legacy... Even when I'm working with a bunch of people that are just trying to move really fast; at least the parts that I'm working on are a little bit more immune to that bad legacy at the end of the day... Which can help quite a lot, because that kind of thing tends to be contagious and other people tend to pick up on it, because they're working in the code that you wrote, and they're like "This is great. This is amazing. I want the whole codebase to feel like this." It's very infections at the end of the day. + +**Dominic St-Pierre:** Do you think that pair programming could also do this effect? + +**Kris Brandow:** I think it can, if your environment is one that is, I'd say, opportune for pair programming. If you're actually in person, you can actually sit at the same desk, you can actually work through things together at the same time, I think then yeah, pair programming can be very helpful. I think also just group design, group thought processes... Whenever you can get people in the same setting and you have enough psychological safety where people can just ask those questions that they might think are done... Because at the end of the day, that thread you have is very much a "I don't know what the problem is, but this doesn't feel right", so you have to be in an environment where you can express that, and the people you're working with will help you work through that, and have enough trust in you. + +So yes, I definitely think pair programming can be excellent for this. I think whiteboard sessions, if you can be in a space where you can do whiteboarding, is great. But I think also just - maybe not doing the pair programming part of programming, but just hopping on a call and talking through an idea with people, maybe over a Google Doc, that also I think is equivalent and a very helpful way to execute this. + +**Angelica Hill:** So we've talked a little bit about kind of you have this legacy thing, and how do you make it so that you don't get to that final explosion point... But I'm interested to talk a little bit about when/if you get that buy-in to rewrite a legacy codebase, or to change up a system and business, for whatever reason, decide -- they're having a great day, and you're like "You have a year, you have two years, or however long it takes. We need to rebuild this thing and make it better." When you are architecting that new solution, or you're thinking through "How do we build this new thing?", how do you from day one, from the whiteboarding session, build something that is going to avoid being legacy for as long as possible? I.e. is it always having microservices, making sure nothing is deeply coupled? What are some things, some questions to be asked? How do you go about designing a new system, already with legacy in mind? + +**Kris Brandow:** \[52:08\] I would say from the jump understand what went wrong with your old system. I think people run for microservices or for Kafka, for some new, flashy tech, at the end of the day; they're like "Oh, the problem was we weren't using microservices", or something. But it's like, sit down and actually think, why don't you like working on this code? Why have we had to declare bankruptcy on this codebase? + +A thing that I have done, that has hilariously sometimes wound up with me not actually abandoning old codebases when I really wanted to, because the problems I wanted to fix I could just fix in that old codebase, so why make a whole new one, with all new problems... But I think that's definitely where I tend to start, is like "What is it about this codebase I'm working on now that's making it so I want to build a new one?" + +**Angelica Hill:** Jon, what do you think? I know this is something that I'm hoping you're thinking about... + +**Jon Sabados:** I don't know if I've got too much on this one... + +**Angelica Hill:** Also, off the record, for those who didn't see my cheeky smirk, Jon and Jeff in fact are both on my team... So when I make any cheeky comments in their direction, it's slightly more layered than other people who don't have that context may know. And I am their technical product manager, so that's also a context that you should know. I am speaking off the record as a businessperson. + +**Kris Brandow:** Undercover businessperson. + +**Angelica Hill:** I'm undercover. Secret gopher... Well, over gopher, secret product manager. \[laughs\] Jeff, what are some questions that you would hope -- if I was to go to you, Jeff, and be like, "Hey, Jeff, I'm thinking we should completely get rid of our current platform and rewrite it." What would be things that would pop to your head as things you would want to be given time to think about, before we kind of went "Right, this is the new architecture. Let's go. Code." + +**Jeff Hernandez:** Yeah, I'm definitely following what Kris is saying. You definitely have to understand why you're trying to rewrite everything. You don't wanna rewrite just for the sake of "Oh, we have lots of issues with this codebase. Maybe we should just declare bankruptcy and start something new." + +It's definitely something I myself have been interested in. For our team specifically, we're in the middle of rewriting our two platforms, and getting some historical context as to why we went that route. It would be interesting. But I don't wanna trigger Jon's PTSD too much, because it looks like he's going through something when the question came up. \[laughs\] + +**Jon Sabados:** Yeah, and I guess rewrites are tough... I feel like everybody at some point in their career needs to go through the "Let's just do the rewrite because of it, and feel the pain, and watch it fall flat on its face, and waste like six months of our lives..." + +**Kris Brandow:** Six months if you're lucky. + +**Jon Sabados:** Yeah. I've seen entire years go down the drain. So having an understanding of the risk involved is huge. And you know, there are scenarios that do warn it, but they're the exception rather than the rule. + +**Angelica Hill:** So am I right in assuming that if your product partner or the business came to you - I'm speaking to all of you here - and said "Right, Kris, Misha, Dominic, I know you've been working with this 20+ year (in Dominic's case) infrastructure. We're gonna give you five years maximum to rewrite completely." Would there be cases in which you would then, after assessing, be like "No, actually, we don't wanna rewrite." Or would you always, if given the opportunity, want to rewrite? If given the buy-in, the money, the time to do it, would you always jump at that opportunity? + +**Dominic St-Pierre:** I would only rewrite a software that works if the technology is not going to exist sooner. Take an old system like Cobol. In Quebec there's a couple of companies that are still built on top of AS/400 and Cobol and whatnot. I mean, there's no manpower anymore for that, so what do you do? You don't have a choice. But I would be extremely careful to rewrite a software, not built from a long-time software, from scratch. I would still try to see if there is something that can be done, other than a complete rewrite. + +**Kris Brandow:** \[56:22\] I'm in agreement there. I think when it comes to greenfield projects, or when it comes to rewriting legacy projects, just don't. I think that's kind of the hard and fast rule most of the time. Because sure, usually we get to the point we want to rewrite something when there's 20% to 30%, maybe a fifth to a third of the project is just obnoxious to work within. But there's still 70% to 80% of it that is doing things correctly. + +When you do a rewrite, you have to account for all those things that it does correctly, and then replicate them, and then also fix all those bugs. So it's always going to take way, way, way longer to rewrite the thing than it is to fix the problems in the thing that exists. So I think that's the thing that we have to account for. And if you go into it and you do that math and you say "Yup, nope, we want to rewrite because we have X, Y and Z reasons why" - sure, go ahead and go forth. But if it's just like "Here are some individual pain points that I don't like about this system, and it's causing us to slow down some. Let's rewrite the whole thing" - absolutely not. You have to do in-depth research and understanding of the project that you're dealing with before you can rewrite it. You should actually be able to rewrite it before you start rewriting it, which sounds like something that sounds super-obvious when said, but is a thing that I have witnessed over and over again; it's not something that we do as engineers in general. + +**Jon Sabados:** Yeah, there's gotta be some fundamental problem that requires a complete greenfield rewrite. It can't just be bang where everyone is gonna be like -- And also, it shouldn't come from product. It should be tech being like "There's this technical roadblock that means we just cannot do this anymore." Because you know, if you've got this legacy codebase and you just wanna make it better and get there, there's ways that you can boil things out, too. So you can be like, "Well, we're gonna move this small set of functionality to a new thing", and piecemeal things out if you really want an entirely new codebase, versus improving it. But again, that should be, I think, tech-driven, rather than coming from product. + +**Kris Brandow:** If you can't come up with a way of conceivably, actually retiring completely the thing that you're rewriting at the beginning of your rewrite, then you shouldn't do it. Because you're not going to be able to retire it if you don't already have a plan for how to do it. I think that's just like fact at this point. I've never seen anywhere that's been able to successfully retire something... Either it's you've been able to successfully do it because you had a plan, or you were able to successfully do it because you just threw engineering bodies at the problem and basically said "All other work halts until we've finished this movement and this migration", which is incredibly difficult and really requires buy-in all the way up to the CTO level at the end of the day, and probably CEO level, to really say "No, we're actually going to no matter what throw as many bodies as is required to get rid of this thing." Either you have to do that or you have to have a good plan from the beginning of how you're gonna do it. And most of the time we don't have a plan, so we're really saying the latter. And if you're not gonna go and spend that political capital to buy into that, then just don't rewrite the thing, and really sit down and think about what is frustrating you about this codebase. Because I would say for most of the codebases, except for the ones, once again, written in Cobol or in some other really old language that we don't wanna deal with anymore... For the most part, especially if it's a Go codebase and I wanna rewrite this, just stop; just stop and be like "What's bothering me about this?" and then go fix those things. That will take tremendously less time, and you will be happier with the result, because now you will have the codebase you wanted, without the codebase you didn't. You don't have this legacy codebase that someone gets stuck maintaining until your new thing actually does what you promised it would do... Which also will take a lot longer than you think it's going to take. The rewrite always takes longer, and if you can't shut down the old thing until the rewrite is done, well now someone has to keep fixing bugs and keep the old thing operable until the new thing is ready to go in full capacity. So if you really don't wanna work on the old thing anymore, just fix the old thing in place. + +**Dominic St-Pierre:** \[01:00:20.26\] You know how it is, even if everyone is agreeing to not touch the old thing anymore, they will always have a new feature, there will always be bug fixes that you will need to replicate on the new system. It's extremely difficult. + +**Kris Brandow:** Yeah, don't assume that you will not have to add something new to the old system until the old system is completely off. So if your new thing is "I'm gonna go build something new, and it's gonna take me 2,5 years to build it", assume that you have to keep adding stuff to the old thing until that 2,5 year mark hits, and you actually get to turn that old thing off. + +**Angelica Hill:** Moral of the story is greenfield rewrites are not always the right direction to go. You need to sit, take some time, think it through, identify the core issues, and then get really good - nod to your wonderful series, Kris - at maintaining your software and your technology. + +**Jon Sabados:** Yeah. I think I have participated in one successful greenfield rewrite in my 20-year career. + +**Angelica Hill:** And where was that, Jon? + +**Jon Sabados:** That was with us. + +**Angelica Hill:** And who is your product manager? \[laughs\] + +**Jon Sabados:** That is with the Times, yup. + +**Kris Brandow:** I also have a side note about the term "greenfield", which I find hilarious... Because if you want to have an actual greenfield, someone has to maintain that. If you go into nature, you don't find a lot of greenfield -- you find a lot of mud pits and stuff, and all that overgrown grass and stuff. But a real greenfield? You've gotta have good lawn equipment, good agriculture... It takes a lot of work, a lot of maintenance to have a greenfield. And people are just like "We're just gonna start with a greenfield." That's like, you're gonna start with something that's well-maintained and well-manicured. Okay. So it's like, we should really call it mud pit development, and then people will probably -- + +**Angelica Hill:** The grass is always greener on the other side... + +**Kris Brandow:** Because the other person actually maintains their lawn, unlike you... Right? \[laughs\] Their grass is greener because they take care of it. You go over there, it might be green for a little bit, but then your old grass is going to be greener, because that person maintains it now. So it's like, no, just learn how to maintain your grass and then you'll have a really green lawn. And also, don't envy your neighbor. There's lots of lessons in here. I can take this analogy all day long. + +**Angelica Hill:** I feel we're getting into the Canterbury tales here + +**Kris Brandow:** This is what happens when you have two writers on a podcast, Angelica... + +**Angelica Hill:** I know. + +**Kris Brandow:** We just feed off each other. + +**Angelica Hill:** Don't even get me started on Shakespeare and references... We'll be here all night. To rewrite or not to rewrite? That is the question... \[laughs\] Whether it is greener on the other side... Okay, I'm gonna stop now. But I can go on for days. + +Great! Well, regrettably, we have to move on to our next section, because I wanna hear your unpopular opinions. But it's been an absolute pleasure chatting about legacy code, and I'm sure myself and Kris have many an episode as a follow-up we now will like to do. + +**Kris Brandow:** This maintenance mini-series is becoming a series. It's just kind of expanding. + +**Angelica Hill:** Yes. I feel like we should stop calling it mini and just go with "This is a series. Maintenance is important. It's a thing." We should make its own little theme tune; I'm ready, Kris. Let's whip out the keyboard. I'll do the vocals, and I'll rap a little bit; it'll be great. + +**Kris Brandow:** A new jingle. We'll make that a note for Jerod. Hey, Jerod, we need a jingle for the maintenance series. + +**Angelica Hill:** Yeah. And on the topic of jingles... Here's our unpopular opinion one. Smooth transition... + +**Jingle:** \[01:03:34.16\] to \[01:03:50.28\] + +**Angelica Hill:** So... For all you lovelies who have not heard Go Time unpopular opinions before, this is where we ask our lovely guests, and also Kris if we have time, for an unpopular opinion. It does not need to be about Go, it does not need to be about technology. It can be about your aunt's favorite sock collection, it can be about china dolls, it can be about your view on the ethos and life. It can be on anything. + +First I'm gonna turn over to you, Jeff, for an unpopular opinion. Now, the goal of this is to come up with an opinion that you believe will be unpopular, because we're gonna post this opinion, we're gonna tweet about it, have our lovely gopher community vote on it, and then we're gonna tell you what percentage unpopular it was. + +**Kris Brandow:** And if it's popular, you have to come back. + +**Angelica Hill:** Yeah. If it's popular, we're getting you back on Go Time to grill you again, to get a better unpopular opinion. + +**Jeff Hernandez:** What if I just use that as the way to keep coming back on the show? I just keep giving popular opinions, because I like it here. + +**Angelica Hill:** Oh, we'll sniff you out. We're savvy individuals on this show. + +**Jeff Hernandez:** So I've been racking my brain with this ever since I agreed to be on Go Time, so I don't think I have anything Earth-shattering like Jon's previous unpopular opinion... But what I came up with is I do not like any type of yogurt. I feel like that's a very popular snack that people like to eat. I just do not like yogurt at all. I have a story actually about that... + +So I bought Greek yogurt as a substitute for sour cream, because I heard that's like a good, healthy substitute... And it was my first time actually trying it. I opened up the container and it just smelled really funky. I told my college roommate at the time, "Did this go bad?" And he came over and smelled it, and he's like "No, that's just how it smells." I'm like, "Oh... Definitely not for me then." + +**Angelica Hill:** \[laughs\] Wait, are we talking just like plain, or are we talking like mixed in with some granola and fruit? + +**Jeff Hernandez:** Any... + +**Angelica Hill:** You also don't like it? + +**Jeff Hernandez:** I don't like any yogurt. I've never been like a yogurt person. + +**Angelica Hill:** Like, Greek only? Are we talking about Yop, like drinkable yogurt? + +**Jeff Hernandez:** No yogurt. No yogurt. + +**Angelica Hill:** Oh my gosh, this is like an umbrella yogurt ban in the Jeff household. + +**Jeff Hernandez:** Basically, yes. \[laughs\] + +**Kris Brandow:** Fascinatingly, I thought I always hated Greek yogurt, and for some reason I know like Greek yogurt, and I don't know why. So I'm kind of on the opposite spectrum as you. + +**Angelica Hill:** Well, Kris, you know, your taste buds do mature as you age, so... \[laughter\] That's a good unpopular opinion. I love yogurt. I love Greek yogurt. Also, it's very good for your digestion, I'm told, and it has things called active things... \[laughter\] Active -- what are they called? + +**Jeff Hernandez:** Probiotics? + +**Angelica Hill:** Yes. They have active probiotics that are apparently, I've been told, being brainwashed by the media, are very good for my digestion... But I have no scientific facts to back up that view. But I will after this google it and check. Maybe as you get older, like Kris, you'll develop a taste. + +**Jeff Hernandez:** I get all my active probiotics from Yakult, if anyone else has had that. I don't know what I would describe it as, but it's like a small little drink. It's not yogurt, as far as I know... But that's where I receive my active probiotics, since I don't have any yogurt. + +**Angelica Hill:** World's leading probiotic beverage created in Japan, in 1935, and is now sold in over 40 countries. + +**Jeff Hernandez:** And it's not yogurt, so... + +**Angelica Hill:** And it's not yogurt. I actually felt like it was, and I was gonna google it in the hopes that I would catch you out... But it's not. + +**Jeff Hernandez:** I've done my research, so... You won't catch me. + +**Angelica Hill:** Misha! Unpopular opinion. + +**Misha Avrekh:** Let's see... Alright, so my unpopular opinion, that is not about Go, is that CSS is a full-fledged programming language that will someday replace all other programming languages. Now I just need to figure out how to rewrite all of my backend microservices in CSS. + +**Kris Brandow:** \[laughs\] + +**Angelica Hill:** I don't even know how to respond to that one. + +**Jeff Hernandez:** You broke my brain, I think. + +**Misha Avrekh:** That will solve the legacy problem, for sure. + +**Angelica Hill:** Yeah, let's just write everything in CSS... + +**Jon Sabados:** Does it have to be CSS, or can I write it in Sass, and then transpile it to CSS? + +**Misha Avrekh:** \[01:08:20.08\] Sass is fine. + +**Jon Sabados:** Okay. + +**Angelica Hill:** Getting snazzy with it here. + +**Kris Brandow:** Well, you can do math in CSS now, so I can kind of see how you could create a virtual machine out of C but the bytecode is CSS, and then... Oh, God. Now we're gonna wind up with someone writing a C compiler that compiles C into CSS, that can actually -- ah, this is terrible. + +**Angelica Hill:** What have we started...? \[laughs\] I would be surprised if that was not unpopular. That's a very good unpopular opinion. + +**Misha Avrekh:** Thank you. + +**Angelica Hill:** But I do now really want someone to do that. I want to see this compiler. + +**Misha Avrekh:** It's a challenge. + +**Angelica Hill:** Dominic, over to you. In fact, Misha, you can do it, and we can come back on the show and talk about it. + +**Kris Brandow:** You've just gotta write it in Go. + +**Misha Avrekh:** Sounds good. + +**Angelica Hill:** I'm ready. Though I swear that defeats the purpose... If you write it in Go, is it not then fully CSS? + +**Kris Brandow:** You can eventually bootstrap it so it's self-hosting on CSS, but you'd have to write it in some -- + +**Angelica Hill:** Does that count? + +**Kris Brandow:** Yeah. Go is written in Go, but it didn't start off being written in Go. It started off being written in C. + +**Angelica Hill:** Oh, okay... Misha, and the definition for your unpopular opinion that you just came up with - does that count? + +**Misha Avrekh:** Yeah, that would count, yes. + +**Angelica Hill:** Can we do this? Okay. It's official now. \[laughs\] Dominic, hit me. Unpopular opinion. + +**Dominic St-Pierre:** Well, last time I thought it was the highest I could go, so let's see... + +**Angelica Hill:** Clearly not, because you're back. + +**Dominic St-Pierre:** Clearly not, exactly. GitHub Copilot will do more harm than good, I think. + +**Kris Brandow:** Oh, that's not gonna be unpopular. + +**Dominic St-Pierre:** Yeah... + +**Kris Brandow:** \[laughs\] + +**Angelica Hill:** It's more of an opinion that needs to be put out there. + +**Dominic St-Pierre:** Yeah, I know... I was supposed to think about -- maybe I will pass, because like I was saying at the beginning, I caught the Covid yesterday. I was supposed to think about that yesterday, and to be frank, I did not, so... Yeah, let me pass on that. + +**Angelica Hill:** Shout-out to Dominic for doing Go Time with active Covid. + +**Dominic St-Pierre:** Yeah, it was rough. + +**Kris Brandow:** Real trooper. + +**Angelica Hill:** Yeah. Thank you so much. + +**Dominic St-Pierre:** Yeah, that's my unpopular opinion, in fact... Doing things with the Covid. + +**Angelica Hill:** Powering through. Alright, Jon, you are the person who holds one of the most unpopular Go Time opinions ever... Shameless plug to a previous Go Time episode where we go through all of our unpopular opinions of the year. I don't know exactly what number it is, but if you DM me, I'll let you know. It's a great episode... Where Jon said that he hates chocolate. So I'm ready for this next one. It better be a good one. My expectations are sky-high. + +**Jon Sabados:** I'm gonna swing for the fences on this one and say that I think adding generics to Go was a mistake. + +**Kris Brandow:** Ooh... + +**Jon Sabados:** So while there are use cases where generics actually are very handy, I think overall I have seen more harm done with generics than in general... Because I think that it encourages developers to try and deduplicate things that are not necessarily the same. And when you're trying to reuse for the sake of reusability and accomplish different tasks with code, you oftentimes end up with code that is much more of a tangled mess, and harder to maintain, and harder to understand. And I think that generics encourage that. + +**Dominic St-Pierre:** I think like everything in life, balance will return at some point... But in my opinion, it will be good. Just think about all the loops that we are doing, that we will not be forced to do... Like removing an item from a slice, for example, or... I don't know. So I think for the next 6 months it will be very dirty everywhere; everyone will try that, but... I'm not sure. I disagree with that, respectfully. + +**Jon Sabados:** Totally. And I totally get you with those loops and whatnot; there's cases where it totally makes sense. But I think that having those cases is the lesser sin from the abuse of generics that's going to follow. + +**Dominic St-Pierre:** \[01:12:09.21\] Yeah, I see the point. But I hope we will see a balance at some point. + +**Kris Brandow:** It's gonna be channels and goroutines all over again. People are just gonna sprinkle that stuff all over their codebases. It's gonna be "Oh, can I use a channel here? I'll use a channel. Oh, now I need to use seven channels? I'll use all seven of those channels, just because I can. I'll sprinkle goroutines everywhere." It's gonna be that all over again. + +**Angelica Hill:** Need I return to my first app ever, that I think Kris you've seen, where I literally ended up having 317 channels, individually written, partly because I didn't know you could know you could send stuff to one channel, but also because I just had fun. I was like, "I'm gonna create all these--" And then I had pointers to channels... + +**Kris Brandow:** Oh, boy... + +**Angelica Hill:** It was a lot of fun, but clearly not the right approach. + +**Jeff Hernandez:** Too fickle to maintain. + +**Angelica Hill:** Oh yeah, and then I added goroutines onto every single one. That was fun. \[laughs\] Yeah, Kris, I know you loved it when I showed you, like "Look how pretty it is." "Nooo...!" \[laughs\] Kris, do you have an unpopular opinion to finish us up with? + +**Kris Brandow:** I have many an unpopular opinion. I think I hold one of those most coveted unpopular opinions... + +**Angelica Hill:** You absolutely do. + +**Kris Brandow:** Okay, so this one is tech-related, not Go-related... But I think that we should kill off all C-like languages and C-derived languages, and get rid of the van Neumann architecture and move on as an industry. + +**Angelica Hill:** Do you have a bit more as to why? + +**Kris Brandow:** Yes, actually. So when the van Neumann architecture was created, we were still using discrete parts. But even when we made that first transition in the '70s to transistors, we had on the order of thousands of transistors operating at kilohertz. And now, as of a couple weeks ago, you can buy a chip, buy a computer on the market that has over a hundred billion transistors operating at several gigahertz each. So we've had many, many, many orders of magnitude along two different dimensions of improvement in our hardware capabilities, but we still used the same fundamental architecture, that has the same fundamental bottlenecks attached to it... Which ironically is turning our systems, our chips into distributed systems just to keep pace with the improvements that we've made. I think all of this extra work and all of this strife is caused by the fact that we're still using this architecture. And the reason we still use it is because - well, it's the architecture of C, at the end of the day. C very much programs in this mentality that recognizes the van Neumann architecture as being there. It's very difficult to write truly parallel, or good parallel code, or concurrent code in C. + +So jettison the van Neumann architecture, and along with it C and C-like languages... Which unfortunately includes Go as of right now, because while you can write some good concurrent code in Go, it's still pretty difficult. We want languages that look more like Erlang, or at least have the same underpinnings of things like Erlang at the end of the day, to really capture the capabilities of the silicon that we have. Because it is quite incredible; I mean, Apple did just come out with a chip that has 115 billion transistors on it. And it's like -- yeah, we can't really utilize all of those if we're using them in our CPU-based architecture. So more stuff that looks like graphics cards too, because they're very good at capturing this idea of parallelizable and concurrent workloads, but... Yes. No more C, no more von Neumann. Get rid of it, move on as an industry. Let's get better. Let's stop wasting so many CPU cycles waiting for main memory to get back to us. + +**Angelica Hill:** Yeah, that may be an unpopular one... Especially if you add in that very blaze like "Oh, Go is included." + +**Kris Brandow:** I mean, Java, Go Rust, C, C++... All those popular imperative languages, just - goodbye. Time for something new. + +**Dominic St-Pierre:** Let's rewrite Linux in Erlang. It will be fun. + +**Angelica Hill:** So on that fine point, "Let's get rid of Go", regrettably, that is the end of our episode. Thank you all, it was an absolute pleasure having you on, and I hope we'll have you all back very soon, subject to how unpopular your opinions are. diff --git "a/How can we prevent legacy from creeping in\357\274\237_transcript.txt" "b/How can we prevent legacy from creeping in\357\274\237_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..889d70202a4917464f7ce2855d6854be8988254b --- /dev/null +++ "b/How can we prevent legacy from creeping in\357\274\237_transcript.txt" @@ -0,0 +1,1276 @@ +[0.00 --> 4.02] When I hear the term legacy, I think there is this kind of feeling of dread that a lot of us have. +[4.10 --> 6.06] But, you know, the term legacy can actually be a good thing, too. +[6.14 --> 9.26] Like the legacy that you leave from the life that you lived and things like that. +[9.30 --> 13.22] And I kind of take the same view that I take for something like debt. +[13.48 --> 16.50] There's good debt. There's bad debt. There's good legacy. There's bad legacy. +[16.80 --> 21.40] And I think that really informs the way that I also see legacy when it comes to code. +[21.40 --> 24.14] I kind of see it's like anything that's hit production to me is legacy. +[24.52 --> 26.06] It's out there. It's something to deal with. +[26.10 --> 28.58] It's something that we have to maintain and we have to understand. +[28.58 --> 32.20] Even if we don't necessarily like that code at the end of the day. +[35.76 --> 38.40] This episode is brought to you by Sourcegraph. +[38.82 --> 42.94] Sourcegraph is universal code search that lets you move fast, even in big code bases. +[43.56 --> 48.18] Here's CTO and co-founder Byung Liu explaining the problems that Sourcegraph solves for software teams. +[48.18 --> 54.08] Yeah, so at a high level, the problems that Sourcegraph solves, it's this problem of for any given developer, +[54.58 --> 57.22] there's kind of two types of code in the world, roughly speaking. +[57.22 --> 61.66] There's the code that you wrote and understand, like the back of your hand. +[61.78 --> 65.26] And then there's the code that some idiot out there wrote. +[65.48 --> 72.66] Or, you know, alternatively, if you don't like the term idiot, it's the code that some inscrutable genius wrote and that you're trying to understand. +[72.78 --> 76.32] And oftentimes that inscrutable genius is like you from, you know, a year ago. +[76.32 --> 80.72] And you're going back and trying to make heads or tails of what's going on. +[80.88 --> 91.82] And really, Sourcegraph is about making that code that some idiot or inscrutable genius wrote feel more like the code that you wrote and understand kind of intuitively. +[91.82 --> 96.30] It's all about helping you grok all the code that's out there, all the code that's in your organization, +[96.56 --> 102.64] all the code that is relevant to you in open source, all the code that you need to understand in order to do your job, +[102.70 --> 105.86] which is to build the feature, write the new code, fix the bug, etc. +[106.34 --> 106.82] All right. +[106.88 --> 111.46] Learn how Sourcegraph can help your team at info.sourcegraph.com slash changelog. +[111.46 --> 115.32] Again, info.sourcegraph.com slash changelog. +[128.74 --> 129.68] Let's do it. +[130.26 --> 131.32] It's go time. +[131.90 --> 137.04] Welcome to Go Time, your source for diverse discussions from around the Go community. +[137.04 --> 141.58] Subscribe today at go time.fm and follow the show on Twitter. +[141.82 --> 143.38] We are at GoTimeFM. +[143.76 --> 148.78] Special thanks to our partners at Fastly for delivering GoTime super fast all around the world. +[149.10 --> 151.66] Check them out for yourself at fastly.com. +[151.92 --> 153.06] That's all for me. +[153.38 --> 154.10] Here we go. +[163.40 --> 165.36] Welcome to Go Time. +[165.36 --> 171.54] Today, we're going to be talking about how to prevent legacy code from creeping in. +[171.82 --> 178.02] I'm joined by a wonderful array of guests, as well as my lovely co-host, Chris. +[178.34 --> 179.24] Hello, Chris. +[179.52 --> 179.86] Hello. +[180.50 --> 182.74] And yes, that is my voice you hear out there, listeners. +[182.96 --> 185.42] I'm finally back after, what, four months? +[185.96 --> 188.40] It's been quite a while, but I am very excited to be back. +[188.70 --> 190.60] It's been far too long. +[190.60 --> 197.44] And along for the ride with us, talking about how to work with, avoid issues arising from, +[197.90 --> 204.76] as well as just how to deal with legacy code on a day-to-day basis, we have, firstly, Dominic +[204.76 --> 207.90] St-Pierre, who operates a small consulting company in Montreal. +[208.58 --> 214.42] He's the maintainer of an Eskos alternative to Firebase called Static Backend, and is the +[214.42 --> 216.58] author of Build SaaS Apps in Go. +[217.10 --> 218.00] Welcome, Dominic. +[218.08 --> 219.50] Welcome back, I should say. +[219.50 --> 220.22] Thank you. +[220.38 --> 221.34] Pleasure to have you. +[221.46 --> 222.14] Happy to be here. +[222.52 --> 228.74] Then we've got Misha, who is a first-time GoTime guest, but also, I'm told, a first-time +[228.74 --> 229.90] podcaster in general. +[230.08 --> 234.16] So very privileged that you decided to grace us as your first podcast. +[234.56 --> 235.56] How are you today, Misha? +[235.96 --> 236.94] I'm good, thank you. +[237.36 --> 241.48] And then we have Jeff, who is an Associate Backend Engineer at the New York Times. +[241.56 --> 244.94] I should also say Misha is also at the New York Times and is a software engineer. +[245.34 --> 246.38] But welcome, Jeff. +[246.42 --> 246.98] How are you? +[247.54 --> 247.96] Great. +[247.96 --> 249.26] I'm excited to be here. +[249.50 --> 251.04] I'm excited to have you. +[251.46 --> 253.98] And then, John, we've been here before. +[254.48 --> 257.72] John Sabados, who is also at the New York Times. +[258.10 --> 259.40] I promise I'm not biased. +[259.56 --> 262.40] They just happen to be people that I spend most of my life with. +[262.88 --> 264.44] He was a staff software engineer. +[265.02 --> 265.58] Welcome back. +[265.72 --> 266.34] How are you today? +[266.58 --> 267.90] Doing well and good to be back. +[268.42 --> 268.72] Awesome. +[268.72 --> 275.30] So before we kind of dive into the nitty-gritty, I wanted to ask the group, what is legacy code? +[275.54 --> 279.10] How do we even define what we call legacy? +[279.68 --> 282.12] I don't know whether, Dominic, you want to dive in? +[282.18 --> 286.32] I know you've been working with something that I'm sure we would all agree is legacy. +[286.32 --> 287.76] Yeah, yeah. +[287.80 --> 292.42] I think in my case, it's really a passion, quote unquote, because, yeah. +[292.66 --> 301.62] So to me, legacy is when a company starts to have difficulties hiring developers to work on their software. +[301.62 --> 304.90] You can think of that as legacy, I think. +[305.20 --> 309.72] Also, when there's no unit test, not really good documentation. +[310.26 --> 319.20] So very, very old software that nobody other than the original author can maintain anymore. +[319.56 --> 322.68] I'm talking about like 20 years old software. +[322.90 --> 326.40] So this is one of the things that I do in my consulting. +[326.40 --> 327.40] So, yeah. +[327.52 --> 334.98] So to me, when the tooling starts to not work for you anymore and things like that. +[335.06 --> 335.36] So, yeah. +[335.56 --> 339.58] In a nutshell, that's a little bit my definition of what legacy software is. +[339.80 --> 341.46] So you said very, very old. +[341.96 --> 343.26] So we're on Go time. +[343.72 --> 347.14] Go, to me, doesn't feel like a very old language. +[347.82 --> 353.22] So I wonder, maybe, John, I know that you've had a little bit of experience working with this, +[353.22 --> 356.70] but at what point did Go applications turn into legacy? +[357.44 --> 362.92] Well, so that's a good question because you can actually, and again, it depends on how you define legacy. +[363.22 --> 370.06] But it's, at least in our scenario, it's very possible to have fairly recent code that is suddenly legacy. +[370.06 --> 377.98] And like one example of where you might hit that is if you have a system that was implemented in a fashion that really didn't suit the need. +[378.14 --> 382.18] So you're forced into a greenfield rewrite and supporting two systems. +[382.18 --> 384.46] That's an easy definition of legacy code. +[384.58 --> 389.02] We've got a new thing that's going to do the stuff, but an old system that we have to support. +[389.28 --> 395.34] I might define legacy code as any type of code that your engineers hate to work on. +[395.34 --> 399.46] Okay, so any code that I don't want to work on, I can just say, oh, that's legacy. +[399.80 --> 401.78] It was written last week, but it was too old. +[401.82 --> 402.72] I don't want to work on it. +[404.34 --> 404.70] Okay. +[405.36 --> 410.90] And then I wonder, Jeff, specifically, I know you've kind of joined the New York Times preferably recently, +[410.90 --> 414.38] and I know that you came to a kind of legacy system. +[414.90 --> 420.78] Coming in, how do you identify whether something you're working on is legacy or not? +[420.78 --> 426.62] Is it that like everyone in the team is saying, oh, this legacy system, but are there indicators when you're kind of joining a new team, +[426.90 --> 428.88] you're like, oh, this feels legacy? +[429.42 --> 436.22] I feel like part of it is everyone just says that this portion of our services are really difficult to work with. +[436.22 --> 437.62] They're kind of considered legacy. +[438.16 --> 439.42] But I don't know. +[439.46 --> 445.70] I feel like maybe my own definition of what is legacy is kind of more broad than what's been given so far. +[445.70 --> 452.96] I think legacy is definitely something that, for me, is something that is already there by the time you joined the team, right? +[453.34 --> 455.76] Regardless of whether people like to work with it or not. +[456.30 --> 457.88] That's kind of like software itself, right? +[457.88 --> 460.24] We're building on top of what has come before us. +[460.98 --> 464.40] So I guess that would kind of be my own definition of legacy. +[465.20 --> 470.16] But I definitely take into account what other people on the team who have been working with the system have to say about it. +[470.16 --> 474.94] But our old stuff or like our old system is very difficult to work with. +[475.04 --> 478.92] It definitely informs my opinion on what I consider legacy. +[479.64 --> 480.22] Yeah, for sure. +[480.22 --> 490.12] And then obviously when we're talking about legacy, I feel like not only are we talking about legacy architecture, legacy infrastructure, but we have legacy dependencies. +[490.62 --> 496.32] I mean, a lot of our apps have hundreds possibly of external dependencies. +[496.32 --> 502.14] At what point do we and how do we go about defining our dependencies as legacy? +[502.28 --> 509.42] I don't know whether Misha, you wanted to speak a bit on kind of how do you think about dependencies and when they become legacy? +[509.88 --> 510.10] Right. +[510.22 --> 518.52] Well, so, for example, the project that I work with now, which is written in Go, it's about two years old at this point. +[518.52 --> 525.50] So probably not legacy code, strictly speaking, although there are parts of it that are starting to feel like legacy code. +[525.82 --> 528.08] We can talk about that separately later. +[528.26 --> 534.56] But yeah, as far as external dependencies that are legacy systems, we have a few of those. +[534.56 --> 548.94] The way we've dealt with them is by caching as much as possible, caching the data that's returned by the external dependencies to reduce the latency, essentially of dealing with external legacy projects. +[549.54 --> 552.38] And is that something that you do for all of your dependencies? +[553.02 --> 556.18] Or is that something you kind of have a specific list? +[556.52 --> 557.84] Or like, how do you make that decision? +[558.22 --> 563.80] So on our team, we made those decisions specifically based on our latency metrics. +[563.80 --> 571.40] So it doesn't really matter whether the external dependency was written in Go or Java or what have you. +[571.52 --> 576.70] If we feel like we need to optimize for latency, we try to cache it. +[577.10 --> 578.22] That's with our approach. +[578.86 --> 582.58] So yeah, monitoring and testing for latency in this case. +[583.10 --> 583.80] For sure. +[584.24 --> 592.94] And then obviously, when we're talking about legacy in its many different definitions, because, you know, as we kind of can tell from even these initial definitions, +[592.94 --> 601.50] there is kind of a fluid way of defining, we can't just identify and then say, great, we're going to rebuild, refactor. +[602.12 --> 608.14] We have to maintain it, whether it be for a month, a year, many, many years. +[608.14 --> 619.80] So I'd love to hear a little bit about what it is like, both from like a day to day, personal, but also from a technical point of view to maintain legacy code. +[620.20 --> 625.16] And I know kind of Chris, our wonderful co-host has been kind of spearheading a maintenance series. +[625.16 --> 629.80] So I know this really tees into the various different topics you've been exploring. +[630.30 --> 639.02] Perhaps for our listeners who aren't fully familiar, I'd love to kind of pass over to you, Chris, before we dive in to talk a little bit about like how we've been exploring and thinking about maintenance. +[639.84 --> 649.66] Yeah, I guess I would start with saying my view of maintenance is definitely not, I think, in some ways as maybe negative as everybody else seems to think it is. +[649.66 --> 654.38] Like when I hear the term legacy, I think there is this kind of this feeling of dread that a lot of us have. +[654.46 --> 659.96] But, you know, the term legacy can actually be a good thing, too, like the legacy that you leave from the life that you lived and things like that. +[660.00 --> 663.92] And I kind of take the same view that I take for something like debt. +[664.16 --> 667.64] It's like there's good debt, there's bad debt, there's good legacy, there's bad legacy. +[668.14 --> 672.72] And I think that really informs the way that I also see legacy when it comes to code. +[672.78 --> 675.82] I kind of see it's like anything that's hit production to me is legacy, right? +[675.82 --> 679.54] Like it's out there, it's something you have to deal with, it's something that we have to maintain. +[679.66 --> 684.48] And we have to understand, even if we don't necessarily like that code at the end of the day. +[684.64 --> 692.86] So I think like before it gets to production, when you can still change it and mold it and do what you like with it, I think at that point, yeah, we're still dealing with something that's not legacy. +[692.98 --> 694.16] It's not debt yet. +[694.40 --> 697.38] It's not something that like we have to necessarily deal with in the future. +[697.58 --> 701.68] But once it hits production, once it's out there, it becomes very difficult for us to pull back. +[701.84 --> 705.68] Now, of course, we can get into the semantics of what does it mean when something actually gets to production. +[705.96 --> 709.16] But I think that's like the way that I at least conceptualize this. +[709.16 --> 715.18] So I think pulling it further back than anybody else has and really seeing it as like, well, all of our code is legacy. +[715.46 --> 720.66] And it really comes down to like, which code do we think is good legacy and which code do we think is bad legacy? +[721.40 --> 724.94] Well, yes, I really like what I'm hearing so far. +[724.94 --> 734.08] One thing that jumped to my mind is, to me, a Go code base, even though it might not have been designed the proper way or whatnot, it's still easier to refactor. +[734.26 --> 739.10] It's still easier to change the status of the code base. +[739.10 --> 741.98] Because there's probably unit tests and whatnot. +[742.10 --> 746.92] There is probably lots of tests in a modern code base. +[747.48 --> 754.86] So to me, I understand all of this, but I feel it might not be the same thing here. +[754.86 --> 763.72] When you're stuck with something and, yeah, you don't have anything, anything, no documentation, no tests to exit yourself out of this situation. +[764.02 --> 766.58] Because like you said, I really like what you said. +[766.66 --> 767.34] Everything is legacy. +[767.52 --> 768.42] Yeah, I understand that. +[768.42 --> 775.54] But it seems to be that when the budget is there and when the team is there, you still have option. +[775.86 --> 784.66] Whereas I work at small companies that they just cannot invest into any kind of refactor or even a rewrite at that. +[785.14 --> 787.98] So you still have to maintain the code base. +[788.30 --> 794.10] Yeah, and I think that's definitionally just bad code at the end of the day, like if you can't maintain it in any way. +[794.10 --> 799.94] But I would also say I've run into lots of circumstances where we have lots of tests and they're not good. +[800.06 --> 801.48] They give us a false sense of confidence. +[801.70 --> 804.40] Or you have lots of tests and they're written in the wrong way. +[804.46 --> 809.04] So whenever you touch something of the code base, you have to go refactor a few hundred lines of tests. +[809.60 --> 814.68] So I think even in your example there, it's like, yeah, no, tests definitely can make a code base better. +[815.16 --> 817.38] And generally they show the mark of a better code base. +[817.44 --> 821.38] But I think there's ways in which even those types of code bases can be a pain to work with. +[821.38 --> 830.32] But I would absolutely say like a code base that has zero tests, no documentation, none of that is like the worst kind of legacy code bases that we can deal with. +[830.32 --> 835.58] How do you know then if your test suite is going to be good for the long run? +[836.08 --> 838.02] Good design of your software. +[838.72 --> 844.60] Like I think like if you kind of got to push the boulder up the hill or like kind of really go back to the genesis of the stream, right? +[844.60 --> 854.16] It's a lot easier to like if you're going to build a dam, it's a lot easier to build a dam across a stream that's two feet wide than it is to build one across one that's 200 or 2,000 feet wide. +[854.38 --> 862.02] So, yeah, it's like if we put a lot of the stuff downstream to know if we've done the right thing, we can through various techniques understand how good it is. +[862.02 --> 868.08] But I think it's much better to try and put in ways of knowing how correct your code is upstream. +[868.30 --> 872.92] But of course, too, it's like a stream, a dam across two feet of stream is quite useless of a dam. +[873.24 --> 876.50] So there's a middle ground here of where we need to find the right place to put things. +[876.54 --> 883.38] And I've definitely found that right place to be in good design documentation and good design ideology for the software that we build. +[883.74 --> 886.76] But I'm sure other people have different opinions on that. +[886.76 --> 891.00] And I think I'd add modern code base does also does not mean tested. +[891.46 --> 897.66] There's plenty of fresh code that I see written that like engineers oftentimes don't like to write tests. +[897.96 --> 904.88] So that is a pervasive problem that seems to plague all code bases and stuff just isn't tested. +[904.88 --> 916.16] Yeah, one thing I can say we've done to kind of mitigate that problem is to keep an eye on the coverage thresholds and actually building the test coverage thresholds +[916.16 --> 918.68] into our testing and deployment pipeline. +[919.12 --> 924.24] So if someone adds new code without adding tests, the pipeline actually fails. +[925.02 --> 927.30] So that is definitely an interesting approach. +[927.36 --> 930.06] But I've also seen it backfire horribly in the past. +[930.06 --> 941.00] I've seen engineers and this actually got somebody talking to, but like this is in Java land, put an entire class on a single line because they were just annoyed by the test coverage requirements. +[941.00 --> 946.18] And so you can kind of get malicious compliance with those code coverage requirements. +[946.38 --> 949.60] And also depending on your language, like Go, it's not as bad. +[949.70 --> 956.12] But in Java land, there's lots of all sorts of like you've got to catch these exceptions that practically are never going to happen. +[956.12 --> 959.46] But you might not even be able to write tests to get there. +[959.88 --> 963.32] So I think you do have to be a little bit careful with coverage requirements. +[963.54 --> 965.66] I think it's a good metric to have and to track. +[965.98 --> 969.98] But there's a danger zone when you start failing builds because of it. +[969.98 --> 974.54] I mean, I'm interested to hear a little bit more about your reference kind of Java, Go. +[974.54 --> 987.86] Are there aspects of Go as a language that make it easier for you to either avoid quote unquote legacy code or avoid it being a whole huge refactor project? +[988.04 --> 993.46] Because in fact, refactoring, you know, bit by bit, you know, backwards compatibility comes to mind. +[993.64 --> 998.44] Are there aspects of Go that make it easier for us to avoid legacy creeping in? +[998.44 --> 1002.26] So I think there's some in that it's a simpler language. +[1002.38 --> 1004.20] Up until recently, it didn't have generics. +[1004.28 --> 1010.96] So you're not going to have people going nuts with generics and creating really complex things like trying to deduplicate everything through generics. +[1011.20 --> 1012.12] So that helps. +[1012.46 --> 1019.44] But I think the biggest thing that Go has going for it with like the lack of legacy and maintainability is that it's a newer language. +[1019.44 --> 1021.30] It hasn't been around for that long. +[1021.68 --> 1029.86] So you don't have projects that have been sitting under or that, you know, have been that were conceptualized 20 years ago to carry that forward. +[1029.98 --> 1034.66] But there's a lot of Java projects out there that have been around for a long period of time. +[1035.08 --> 1041.62] And so I think there might be a false sense of security saying, well, Go is going to have less legacy just because, you know, you don't see it as much. +[1041.62 --> 1048.62] But that's also a function of the language being newer and therefore not having, you know, just the crap that's been around for a long time. +[1048.98 --> 1052.66] And that's also something like we learn every time we write code and every time we develop. +[1053.02 --> 1060.46] So there's a lot of patterns that, you know, we might have been doing 20 years ago in Java land that, you know, we've maybe learned better ways for that are being let go. +[1060.86 --> 1065.46] But again, that's not so much a function of the language, just of the newness of it. +[1065.46 --> 1082.32] I just wanted to add that what John was saying kind of made me think about what this topic of legacy will be like for Go in like 10 or 20 years in terms of when it gets to be old enough to like the same age as Java is currently. +[1082.68 --> 1087.04] And we have a lot more Go projects around that need maintenance. +[1087.26 --> 1090.90] It's just an interesting thought experiment to be having. +[1090.90 --> 1097.96] Well, I can tell you about stories that, you know, with .NET, even though, let's say it started in 2001. +[1098.16 --> 1105.24] So in 2005, there were already projects that you could not even migrate to the next version. +[1105.62 --> 1113.68] So to me, the tooling that the language provide, which is very solid in Go, in my opinion, and the fact that it's backward compatible, +[1113.68 --> 1119.48] if they can maintain that for a long time, it will diminish the effect of that. +[1119.48 --> 1120.56] Yes, I'm certain of that. +[1120.64 --> 1126.82] Because in my case, at the moment, the tooling of Microsoft is my main issues. +[1127.20 --> 1128.26] It's not the code per se. +[1128.44 --> 1131.02] I mean, it's not that at all, but the tooling. +[1131.64 --> 1136.16] And I guess I'd be interested that because I've never really worked in .NET, always been allergic to Windows. +[1136.16 --> 1146.14] But the Java side of things, you do have, like, you know, but they strive for a long time to make the JVM backwards compatible with prior versions of Java. +[1146.38 --> 1154.66] And that didn't necessarily stop, like, the need to have complete rewrites or, like, you being stuck on an ancient version of Java because, like, whatever new thing. +[1154.66 --> 1165.86] I've actually even seen the new thing with Go where they're seeing projects that were, like, running on old versions of GAE that have pinned their version of Go to, like, 111 or something. +[1166.10 --> 1168.20] And, like, upgrading those has become a nightmare. +[1168.20 --> 1171.76] So, again, it's less likely in Go. +[1171.88 --> 1178.56] And it is, for sure, easier to generally upgrade versions of Go than my experience has been with Java. +[1179.00 --> 1187.58] But, again, in Java land, you're also dealing with a ton more dependencies because there's this massive ecosystem that abounds where you're not having to do everything. +[1187.78 --> 1191.08] Where in Go, a lot of that ecosystem doesn't exist. +[1191.24 --> 1195.04] So you're reimplementing a lot of things and not sucking in as many dependencies. +[1195.04 --> 1201.26] So the benefit of that is when you continue to upgrade, you're only worried about your code, not, like, a million different libraries. +[1201.96 --> 1208.96] I'd be curious to see if that easy to upgrade continues as the Go ecosystem grows and as it becomes more complete. +[1209.28 --> 1211.32] Yeah, but you are talking about the platform. +[1211.48 --> 1213.96] So the Google App Engine is a platform. +[1214.14 --> 1218.90] It's not really Go's fault or responsibility, in a sense. +[1218.90 --> 1225.40] You can still take code bases from, like, early, early days of Go and compile that with the new compiler. +[1225.60 --> 1228.56] This is something that you just cannot do in the .NET world. +[1228.76 --> 1231.16] So that's mostly what I'm talking about. +[1231.30 --> 1242.94] When the tooling is failing you, it's a different ballgame because you have the code, but you have the technologies that were suggested to use. +[1242.94 --> 1255.26] When a company makes a choice of going with something and three years later, it's not even supported by Microsoft anymore, then it's a different thing than bad design, I think. +[1255.60 --> 1263.54] Yeah, and I guess there's a different user or past user story there and experience because, like, again, coming from Java land, like, everything will compile just fine. +[1263.54 --> 1271.68] But the things that I've encountered that I've seen that make, like, upgrading hard isn't so much the compilation, which it sounds like in .NET that might be a thing. +[1272.04 --> 1276.78] But it is the platform that is the hard thing to upgrade from my past experience. +[1277.20 --> 1288.98] And, you know, so that's where I think that Go won't be immune to the platform increasing scope and, like, hardness to upgrade with that as it becomes more robust. +[1289.68 --> 1292.34] Or there's more things going to be, like, GAE and whatnot. +[1292.34 --> 1295.04] So I think I definitely agree with you. +[1295.12 --> 1299.32] Yeah, if your vendor is making stuff that doesn't compile it in three years, oh, my God, that stuff's terrible. +[1299.88 --> 1300.38] I'm sorry. +[1301.04 --> 1302.38] Yeah, I can give another example. +[1302.38 --> 1308.00] So, you know, our TLS 1.1 just vanished recently. +[1308.22 --> 1316.88] But there is still an old application that runs under an old version of Windows Server 2008, which is not supported anymore. +[1316.88 --> 1323.54] But because their migration path to the next version of .NET is not an easy one. +[1323.80 --> 1325.30] I'm not saying it's not possible. +[1325.54 --> 1326.36] It's not what I'm saying. +[1326.46 --> 1331.64] But there is a lot of investment that the company needs to do to migrate to the next version. +[1332.30 --> 1335.30] So that's another aspect of legacy. +[1335.30 --> 1344.30] I mean, when the OS and the bare-bone communication system, the protocol starts to fail you, what do you do? +[1344.38 --> 1355.20] So I had to start writing some very, very small piece in Go that just do HTTP call because I needed to use TLS 1.2 or 1.3. +[1355.20 --> 1358.18] So, yeah, it's a different ballgame, I think. +[1358.64 --> 1366.30] That does bring up, like, is working in legacy with closed-source systems different than working in legacy with open-source systems when it comes to those dependencies? +[1366.72 --> 1368.28] Because that is one thing. +[1368.32 --> 1372.02] That's one thing I've always liked about working in all the various different texts. +[1372.20 --> 1373.54] I've been in its open-source scene. +[1374.90 --> 1376.16] Oracle buys Java. +[1376.84 --> 1380.64] Somebody's going to be like, well, we can take the JVM and run with it still. +[1380.86 --> 1381.54] Open JDK. +[1381.78 --> 1383.26] It's been a while since I've been to Java land. +[1383.26 --> 1387.12] So I'm not sure how successful OpenJDK has been, but that is a possibility, right? +[1387.30 --> 1392.74] When you've got open-source, closed-source, you might be forced into complete rewrites. +[1393.40 --> 1408.14] So do we feel like that's one of the kind of differentiators of Go in that, you know, we have this vibrant, wonderful community where if there's anything that is detrimental is causing people to feel that it's legacy in some way in terms of the tools provided, +[1408.38 --> 1412.98] that there's going to be mass opera in the community be like, hey, can we fix this, please? +[1413.26 --> 1414.72] It definitely can help. +[1415.14 --> 1419.80] Up until the point that there's enough bifurcation that there's a fork, which hopefully won't happen. +[1420.02 --> 1423.62] And it's rare to see languages truly fork like that. +[1423.80 --> 1427.50] So I think largely, yes, but that forking is also a potential danger. +[1427.84 --> 1429.12] Languages except for JavaScript. +[1431.02 --> 1431.92] Oh, God. +[1432.06 --> 1432.40] JavaScript. +[1432.74 --> 1432.96] Yay! +[1432.96 --> 1433.14] Yay! +[1433.14 --> 1433.16] Yay! +[1433.16 --> 1433.26] Yay! +[1433.26 --> 1433.96] Yay! +[1433.96 --> 1444.98] This episode is brought to you by our friends at Square. +[1445.28 --> 1452.00] Millions of Square sellers use the Square app marketplace to discover and install apps they rely on daily to run their businesses. +[1452.00 --> 1456.52] and the way you get your app there is by becoming a Square app partner. +[1457.00 --> 1458.12] Let me tell you how this works. +[1458.56 --> 1463.42] As a Square app partner, you can offer and monetize your apps directly to Square sellers +[1463.42 --> 1466.70] in the app marketplace to millions of sellers. +[1466.70 --> 1470.68] You can leverage the Square platform to build robust e-commerce websites, +[1471.10 --> 1474.96] smart payment integrations, and custom solutions for millions of businesses. +[1475.50 --> 1476.56] And here's the best part. +[1476.88 --> 1480.40] You get to keep 100% of revenue while you grow. +[1480.40 --> 1486.62] Square collects a 0% cut from your sales for the first year or your first 100 Square referred sellers. +[1487.04 --> 1490.70] That way you can focus on building and growing your Square customer base, +[1490.84 --> 1493.02] and you get to set your own pricing models. +[1493.56 --> 1495.40] You also get a ton of support from Square. +[1495.76 --> 1498.06] You get access to Square's technical team using Slack. +[1498.32 --> 1501.92] You get insights into the performance of your app on the app marketplace. +[1502.50 --> 1505.16] And of course, you get direct access to new product launches. +[1505.74 --> 1509.36] And all this begins at changelog.com slash Square. +[1509.36 --> 1512.22] Again, changelog.com slash Square. +[1512.22 --> 1529.34] So you talked about kind of the pains of maintaining and the need to, +[1529.76 --> 1532.94] as you kind of touched upon Dominic, kind of hacking these small solutions +[1532.94 --> 1535.82] to make sure your apps continue running. +[1535.82 --> 1541.40] But at what point do you kind of throw your hands up and say, +[1541.64 --> 1543.44] no, we need to completely rewrite this? +[1543.80 --> 1546.32] Like this needs to have, well, one, +[1546.90 --> 1549.44] reliant upon having the time and investment to do that, +[1549.48 --> 1550.60] but we'll touch on that later. +[1551.20 --> 1555.60] But if you feel like there is the time and investment to do a rewrite, +[1555.74 --> 1560.68] at what point do you go from maintaining to really advocating for, +[1560.84 --> 1563.58] we need to, you know, rip this out and plant a completely new tree? +[1563.58 --> 1568.24] One nice thing about legacy software is that at some point, +[1568.34 --> 1572.82] they kind of run without any intervention from any developers. +[1573.18 --> 1577.28] So that is something that could help with a complete rewrite. +[1577.50 --> 1581.44] If the company is able to do that and sees the value in doing that, +[1581.96 --> 1584.30] then yes, I think at some point the maintenance, +[1585.04 --> 1589.26] because after 20 years, you're not really maintaining any code base. +[1589.40 --> 1591.54] You know, you're not changing lots of code base. +[1591.54 --> 1595.16] It's everything around the software that is failing. +[1595.62 --> 1596.70] So it's not the code. +[1596.76 --> 1597.88] The code is battle tested. +[1598.02 --> 1599.60] So even though there's no unit test, +[1599.92 --> 1603.54] I trust those old software completely. +[1604.10 --> 1605.86] You know, they have been fixed. +[1606.10 --> 1608.70] There is not really any bugs anymore. +[1609.20 --> 1610.20] They are pretty stable, +[1610.20 --> 1614.28] but it's a huge investment to decide to rewrite the software, +[1614.78 --> 1617.44] especially old application like that. +[1617.86 --> 1621.72] So yes, I don't really know how to answer that, +[1621.88 --> 1625.96] but maybe someone else has another point. +[1625.96 --> 1631.36] I don't know that there is like a hard and fast rule for when to do a green or a greenfield rewrite, +[1631.36 --> 1635.24] other than it should probably be like one of your last resorts, +[1635.24 --> 1638.24] because it is an incredibly expensive thing to do. +[1638.76 --> 1640.92] Oftentimes you have to do it bug for bug, +[1641.02 --> 1642.80] depending on your downstream clients. +[1642.80 --> 1645.18] And that can be a nightmare to do. +[1645.18 --> 1649.34] So, and you know, different scenarios can offer different reasons. +[1649.54 --> 1651.50] Like, you know, if you're doing a cloud migration, +[1651.72 --> 1658.12] well, there may very well be a reason to do a greenfield rewrite of fairly recent stuff in that case, +[1658.12 --> 1660.98] because your underlying platform is changing dramatically. +[1661.18 --> 1662.90] Based on the projects I've seen, +[1663.08 --> 1668.34] there definitely needs to be a close alignment between business and product, +[1668.34 --> 1673.32] as it were, business requirements and sort of the tech requirements +[1673.32 --> 1676.00] and what needs to be done on the tech side, +[1676.08 --> 1679.66] because that's when you start seeing possible solutions like, +[1679.80 --> 1682.70] oh, well, maybe I don't need to rewrite the whole thing, right? +[1682.78 --> 1686.70] Maybe I need to rewrite just parts of my legacy code, +[1686.70 --> 1694.26] and those will work as standalone pieces while the legacy thing continues to kind of chug away in the background. +[1695.12 --> 1697.02] And also, of course, the expense part, +[1697.02 --> 1702.50] that needs to be kind of understood and underwritten by your business, +[1702.50 --> 1707.02] if that is to happen, if any kind of rewrite is to happen. +[1707.48 --> 1710.02] I'm kind of curious as to whether, +[1710.30 --> 1714.14] when it is important to be, for business and product to be aligned, +[1714.24 --> 1719.24] but also taking into account kind of the engineering quality of life. +[1719.42 --> 1723.84] Like, if your engineers are feeling like I'm spending all my time, +[1723.84 --> 1727.38] like Dominic said, I'm not like fixing the code itself. +[1727.38 --> 1730.92] I'm trying to keep the code standing and keep it running, +[1731.38 --> 1733.12] because everything around it is just crumbling. +[1733.80 --> 1739.78] Speaking for myself, I don't feel like I would last very long if that's all I was doing at a company. +[1740.32 --> 1741.48] Like, yes, I can. +[1741.48 --> 1746.82] I'm willing to support legacy systems, but I don't want, that's not all I want to be doing. +[1746.96 --> 1749.86] I want to be at least adding, like creating something new, +[1750.22 --> 1751.64] or at least extending it or something. +[1752.16 --> 1755.48] I don't want to just be maintaining forever. +[1755.94 --> 1761.22] And I feel like I won't grow as an engineer if that's all I'm allowed to do. +[1761.22 --> 1765.02] Yeah, I can totally relate with what you're saying. +[1765.54 --> 1767.16] Yeah, if you're an engineer, I suppose, sorry. +[1767.80 --> 1769.82] I'm alone at the company at the moment. +[1769.94 --> 1772.94] So they are simply not able to hire anyone else. +[1773.16 --> 1774.62] So I'm kind of stuck there. +[1774.88 --> 1778.44] I do think there are some engineers who actually do like kind of just +[1778.44 --> 1782.34] tinkering on old code bases and not adding new features or anything, +[1782.44 --> 1784.66] but just like keeping something running. +[1784.78 --> 1789.50] I think there are people that just like get an immense amount of joy out of just doing that. +[1789.50 --> 1792.96] But I definitely think like trying to put engineers that don't want to do that sort of thing +[1792.96 --> 1795.58] onto a code base that is, that's all of the work. +[1795.68 --> 1798.20] It's not going to work out very well for anybody involved. +[1798.60 --> 1804.10] And do you feel like that is more in keeping with just like personal preference? +[1804.10 --> 1808.86] Or is that perhaps to do with like where you are in your progression as a software engineer? +[1809.32 --> 1812.58] I mean, purposefully, we have kind of a range of different levels on this call +[1812.58 --> 1817.68] because I wanted to really get an understanding of what is it like coming in like at Jeff's level +[1817.68 --> 1823.56] and, you know, very early looking to learn, looking to grow at a very rapid pace versus someone +[1823.56 --> 1826.48] like Chris, John, et cetera, who are at that. +[1826.96 --> 1828.06] I don't want to say end phases. +[1828.20 --> 1829.68] You have many more years in you. +[1830.54 --> 1831.62] But you see what I'm saying? +[1831.70 --> 1836.42] Like you've worked on enough problems to have learned a massive amount and perhaps to find +[1836.42 --> 1838.74] a new problem for you to solve is more difficult. +[1839.12 --> 1840.82] Or do you think it's just personality? +[1840.96 --> 1841.50] Go on, Chris. +[1841.50 --> 1845.40] I think it's a different type of engineering rather than a certain level of it, right? +[1845.40 --> 1849.20] I think it's like there are people that want to be maintenance engineers in the same way +[1849.20 --> 1853.94] there are people that want to just be like pure R&D, pure prototype engineers. +[1854.04 --> 1854.94] Like they want to build stuff. +[1855.10 --> 1858.48] They never want any of that to be in production because they know that it's like not going +[1858.48 --> 1859.08] to survive production. +[1859.18 --> 1860.92] And they know that they don't want to have to maintain it. +[1860.96 --> 1862.54] They're like, I'm just doing research. +[1862.62 --> 1864.06] I'm just proving out an idea. +[1864.54 --> 1865.50] That's what I like doing. +[1865.62 --> 1870.40] And I think there's leveling within those types of engineering, but I don't think they kind +[1870.40 --> 1872.16] of like stack on top of each other or anything. +[1872.16 --> 1875.30] I think there's like, okay, maybe you're on one far into the spectrum. +[1875.46 --> 1879.20] You really like getting one of those old, nasty code bases. +[1879.30 --> 1880.86] You're like, I'm just going to, I'm not going to add anything new. +[1880.88 --> 1881.76] I'm just going to like fix it up. +[1881.84 --> 1882.42] I'm going to add tests. +[1882.50 --> 1883.30] I'm going to document things. +[1883.34 --> 1886.46] I'm going to turn this into like a really nice code base to work within. +[1886.72 --> 1889.36] And then there's, you know, the other end of the spectrum is those prototype people +[1889.36 --> 1893.16] that are like, I just write code that I just clopper together and like make something +[1893.16 --> 1893.64] work. +[1893.82 --> 1896.36] But like, there's no way this will work for the long term. +[1896.66 --> 1899.42] And I think, you know, the majority of engineers fall in the middle. +[1899.42 --> 1900.78] Like they like doing some new stuff. +[1900.84 --> 1902.26] They like maintaining some old stuff. +[1902.52 --> 1903.62] And that's just kind of what it is. +[1903.64 --> 1907.22] And I think it's important that we start seeing that as a spectrum instead of trying to see +[1907.22 --> 1909.60] that as like, well, who's better than the other? +[1909.68 --> 1910.72] It's like, there's no better. +[1910.88 --> 1913.06] There's just like, we need a bit of all of them. +[1913.12 --> 1914.82] You need to keep your organization balanced. +[1914.82 --> 1918.32] And I think that's one of the criticisms I have for a lot of tech companies right now is +[1918.32 --> 1923.18] that they are very much focused on that, the other end of that spectrum that is the +[1923.18 --> 1926.00] R&D engineers and the people that want to build newer things. +[1926.16 --> 1930.06] And the results of that is a lot of code bases that don't wind up getting properly maintained +[1930.06 --> 1933.58] because there's no one in the company that wants to do that maintenance. +[1933.78 --> 1936.98] So there's no one in the company that's advocating for it, even though everybody in +[1936.98 --> 1938.62] the company is aware that it needs to happen. +[1938.72 --> 1939.56] And I think that happens. +[1939.62 --> 1943.36] A lot of teams sitting there are being like, well, we really want to work on all of this old +[1943.36 --> 1945.96] legacy stuff, but where's the time to do it? +[1946.04 --> 1947.44] Where and how are we going to make this happen? +[1948.04 --> 1949.48] So I think like getting some of that balance is good. +[1949.54 --> 1952.52] But to answer your original question, yes, I think they are different types of engineering, +[1952.72 --> 1955.92] not different amounts of time that you've been doing software engineering. +[1956.42 --> 1956.92] Yeah, for sure. +[1957.42 --> 1963.70] And kind of off the back of that, like people know that a certain platform, a certain part +[1963.70 --> 1966.08] of their technology, their stack is legacy. +[1966.28 --> 1968.50] And it's all kind of, it's known about like you're slacking about it. +[1968.52 --> 1969.74] Oh, this annoying platform. +[1969.92 --> 1971.86] Oh, this annoying, like it's all known. +[1971.86 --> 1978.40] How do you then advocate for a rewrite or for the time to maintain? +[1978.58 --> 1984.06] Like what are from your experiences, the key things that you have to kind of bring up +[1984.06 --> 1989.16] to get buy in from kind of product business, et cetera, to actually give you that time? +[1989.22 --> 1992.22] Because that is, I know, a challenge that many people face. +[1992.30 --> 1993.74] So how do you, how do you overcome that? +[1993.78 --> 1994.60] How do you make them care? +[1995.08 --> 2000.48] That's a hard problem because it's effectively like that maintenance is a cost center with +[2000.48 --> 2003.58] no, like you don't see the benefits immediately. +[2003.58 --> 2008.82] And generally my experience has been everyone is short-sighted and more concerned about getting +[2008.82 --> 2013.78] their current feature out than they are about the ability to add another feature six months +[2013.78 --> 2014.58] down the line. +[2015.06 --> 2019.80] And I think the best argument that I've had is like working with business and explaining, +[2019.96 --> 2025.16] you know, that this tech debt is something that actually has a cost down the line and will +[2025.16 --> 2027.44] impede future things. +[2027.44 --> 2031.10] And most people understand the idea of debt, you know, there's only so much you can have +[2031.10 --> 2035.70] and explaining it as tech debt and being like, you know, we will get to the point where all +[2035.70 --> 2040.30] we can do is pay our tech debt and we won't be able to make any new investments. +[2040.80 --> 2042.50] And oftentimes that falls on deaf ears. +[2042.76 --> 2048.00] So I definitely think structuring it as a conversation around debt and the financial aspect of it can +[2048.00 --> 2048.42] help people. +[2048.42 --> 2052.90] But also just like, I think a lot of it has to do with process and project management of like, +[2052.90 --> 2057.46] okay, like when we said that we could like, when we said that we could deliver this feature, +[2057.52 --> 2061.30] we said it would come with this amount of debt, this amount of stuff that we have to do later. +[2061.56 --> 2062.88] Well, here's the later part. +[2063.04 --> 2064.48] And we have to go do that. +[2064.50 --> 2065.70] And you said that we could do it. +[2065.78 --> 2070.14] So you're going to have to actually stick to what you said you were going to do and give us +[2070.14 --> 2070.86] the time to do this. +[2070.90 --> 2073.18] I think it does take a lot of fighting and advocating. +[2073.18 --> 2077.86] But I think part of the problem is that in general, the teams I've been on have been +[2077.86 --> 2082.48] kind of not great at explaining exactly what that technical debt is. +[2082.54 --> 2086.04] It's like we took on some technical debt, but it's just kind of this mysterious thing, +[2086.08 --> 2086.18] right? +[2086.38 --> 2087.88] It's like, oh, you got to go write some tests. +[2087.92 --> 2089.36] You got to go do some other stuff. +[2089.80 --> 2093.96] So I think it's less that product and business don't want to pay for it. +[2094.02 --> 2098.44] But the cost of it is kind of like not known, not quantifiable. +[2098.44 --> 2102.52] It's like if we talked in debt, but you never knew or never found out what the interest rates +[2102.52 --> 2102.80] are. +[2102.96 --> 2105.60] And it's like, OK, well, you're telling me you have to pay down this debt. +[2105.72 --> 2110.56] But is this debt that's at a 0.5% interest rate or at a 25% interest rate? +[2110.66 --> 2113.78] Because we're going to deal with that in very, very different ways. +[2113.88 --> 2118.10] Like 25% interest rate, we got to pay that down immediately, stop everything else, get +[2118.10 --> 2119.64] rid of that 0.5% interest rate. +[2119.66 --> 2122.38] I'm just going to let that sit there because it's not going to cost me much in the long +[2122.38 --> 2122.62] run. +[2122.68 --> 2126.94] And not being able to talk about things in that kind of quantifiable level, I think is what +[2126.94 --> 2131.18] holds back a lot of engineering organizations from being able to pay down that technical debt +[2131.18 --> 2132.30] and handle that legacy. +[2132.52 --> 2137.30] So I think, at least in my experience, the closer that I've gotten to talking in terms +[2137.30 --> 2139.70] that are more concrete of like, here's what we need to do. +[2139.80 --> 2140.78] Here's the technical project. +[2140.88 --> 2141.68] Here's the plan for it. +[2141.74 --> 2142.94] Here's how much time it's going to take. +[2143.20 --> 2145.58] And here are the benefits that we're going to get out of it at the other end. +[2145.66 --> 2147.84] I've been very successful in selling that. +[2147.98 --> 2153.28] It's definitely the other side of it when it's more just like you can't get a good grasp +[2153.28 --> 2157.46] on it when it's not quantifiable in the terms of the business or in the product team. +[2157.46 --> 2160.76] And quantifying it can be interesting, too, because there are some things that can be +[2160.76 --> 2163.32] like, yeah, this is a huge thing that's going to bite us hard. +[2163.52 --> 2168.78] But when you're trying to quantify exact terms for like, this may or may not be a problem +[2168.78 --> 2173.80] depending on what we do, which oftentimes maybe the answer in that is, if it's not a +[2173.80 --> 2175.24] problem, don't worry about it. +[2175.46 --> 2177.18] I guess quantifying that debt is hard. +[2177.18 --> 2182.42] Would it be fair to say that that is exactly the difference between legacy code and technical +[2182.42 --> 2187.02] debt, where technical debt is something that needs to be repaid and that will potentially +[2187.02 --> 2193.38] grow with time, whereas legacy code, as Dominic said, it might just sit there and work for +[2193.38 --> 2195.02] years until anyone notices. +[2195.52 --> 2200.22] I feel like legacy code is like the technical debt that we haven't paid down for so long +[2200.22 --> 2202.24] that we kind of declare bankruptcy on it. +[2202.44 --> 2205.72] Like, and the way that people think about legacy code, it's like that code that's just like, +[2205.72 --> 2210.02] whenever you make that declaration of a greenfield, you're like, okay, we're done with this. +[2210.14 --> 2213.22] And since we haven't planned well for our debt in the first place, we haven't planned well +[2213.22 --> 2217.38] for how we're going to replace that debt, because we also haven't quantified it for ourselves. +[2217.90 --> 2224.04] So I think it starts off as we want to declare bankruptcy, but we don't actually know how to +[2224.04 --> 2224.36] do that. +[2224.42 --> 2229.06] And so we have this thing that kind of sits there and lives on continuously, but definitely +[2229.06 --> 2231.34] open to other people's interpretations of that as well. +[2231.34 --> 2239.86] To me, it boils down to once the company is near reaching a point where it will affect +[2239.86 --> 2240.74] their bottom line. +[2241.08 --> 2247.62] And when you're not able to find any engineers to work at your company, this is terrible. +[2247.86 --> 2253.22] I mean, this is, you've already reached a certain point that you should never have passed, +[2253.32 --> 2253.86] in my opinion. +[2253.86 --> 2261.70] But before, yeah, before going there, the other point from just a minute ago, I think +[2261.70 --> 2266.62] we might see a new kind of engineering type evolving. +[2266.62 --> 2272.64] Because if you think about it, how many software are we creating these days compared to 25 years, +[2272.76 --> 2273.52] 30 years ago? +[2273.68 --> 2274.58] It's incredible. +[2274.90 --> 2281.30] There will not be enough of us to maintain every single piece of software that is creating +[2281.30 --> 2281.76] at the moment. +[2281.76 --> 2285.62] I think we will see new kind of job evolved. +[2286.22 --> 2287.92] And I'm not talking about AI here. +[2288.04 --> 2289.96] And AI cannot maintain software. +[2290.38 --> 2292.68] I'm looking at you, GitHub Copilot. +[2292.88 --> 2300.00] But yeah, my point is, I mean, there should probably not be a difficult argument to be said +[2300.00 --> 2305.14] about if a piece of software should be left as legacy like that, because it will impact +[2305.14 --> 2307.20] the bottom line of the company at some point. +[2307.28 --> 2311.14] And that should be the metric that every business people should understand. +[2311.14 --> 2315.38] I think there was also a question about how we might start quantifying what this debt +[2315.38 --> 2316.90] looks like for the business. +[2317.00 --> 2319.28] And I think kind of using their own terms would be helpful here. +[2319.42 --> 2323.46] Like businesses tend to have really good, or at least the ones that survive for a long +[2323.46 --> 2326.48] time, tend to have really good risk management apparatuses. +[2326.48 --> 2329.74] So they have really good ways of talking about things probabilistically. +[2329.74 --> 2335.10] And we as engineers hate talking about things in probabilistic terms or statistic terms, right? +[2335.14 --> 2338.12] We're like, that will absolutely happen or that will never happen. +[2338.18 --> 2339.68] And like, those are the only two options. +[2339.90 --> 2344.20] But I've also found some success in like talking about things in a more probabilistic way. +[2344.34 --> 2349.62] This can go as far as down as like story pointing, kind of changing it to be like, I have 70% +[2349.62 --> 2351.54] confidence that this will take this amount of time. +[2351.54 --> 2353.70] And just kind of bubbling that all the way up. +[2353.80 --> 2358.48] And once we start talking in less of these absolute terms, I think it becomes easier for +[2358.48 --> 2362.34] business people to commit for us to communicate with business people and give them something +[2362.34 --> 2362.98] quantifiable. +[2362.98 --> 2367.98] It's like, okay, well, if we don't do this work, there's an 80% chance we're going to +[2367.98 --> 2371.48] run into a problem versus like, if we don't do this work, there's a 20% chance we're going +[2371.48 --> 2372.32] to run into this problem. +[2372.48 --> 2375.08] And then we can start doing those types of trade-offs. +[2375.12 --> 2380.04] And then importantly, that also gives us something to track so we can improve there as well. +[2380.04 --> 2383.82] Because once again, if we're trying to like figure out what our interest rates are, figure +[2383.82 --> 2387.68] out like how much things are costing, it doesn't matter if we're just making up terms that don't +[2387.68 --> 2391.58] actually come back to reality, that's going to degrade the trust from product and business. +[2391.58 --> 2394.90] So we have to say, okay, we have an 80% probability of this thing happening. +[2395.40 --> 2397.14] Okay, did it actually happen? +[2397.42 --> 2399.72] And how many times did it actually wind up happening? +[2400.08 --> 2403.06] Just going back and like tracking all of that and tracing all of that. +[2403.12 --> 2406.28] So then every time the business asks, we can be like, here's, you know, the pages and +[2406.28 --> 2409.92] pages and pages of like how we've discussed this and how we came to these +[2409.92 --> 2411.66] numbers, how we came to this probability. +[2411.66 --> 2415.98] And here's like the historical aspect of that and how much of the time, how much it has occurred +[2415.98 --> 2420.58] historically so that we're actually able to like give you some basis to believe us and +[2420.58 --> 2422.14] kind of build that trust back up. +[2422.14 --> 2425.22] And I think that's how you start making it more quantifiable for people. +[2425.72 --> 2429.42] But I think to your point of that, like type of engineer that we need to develop, I think +[2429.42 --> 2433.20] one of the types of engineers we do need to develop are people that are kind of focused +[2433.20 --> 2437.60] on like going into code bases and doing this type of analysis and figuring out how to +[2437.60 --> 2441.94] prioritize that debt that we have and come up with like debt consolidation and debt pay +[2441.94 --> 2446.64] down plans and help us actually determine, hey, we want to declare bankruptcy and build +[2446.64 --> 2447.24] a new thing. +[2447.64 --> 2448.94] This is how we're going to do it. +[2449.00 --> 2452.44] So we don't wind up with the old thing still being there haunting us forever. +[2452.94 --> 2455.22] That is definitely an interesting thought. +[2455.34 --> 2459.22] But it is like you mentioned, that is almost like a new type of engineer, a new type of position +[2459.22 --> 2465.58] because coming up with all those percentages and whatnot, that's like an FTE almost. +[2465.76 --> 2470.72] So it's not something that you can just kind of do on the side, you know, while you're doing +[2470.72 --> 2471.52] your feature work. +[2471.76 --> 2474.66] You need somebody who's focused and dedicated to that. +[2475.22 --> 2475.94] Yeah, yeah, absolutely. +[2476.14 --> 2480.90] It's like it's a it's a thing you have to be dedicated to doing and actually want to push +[2480.90 --> 2482.90] your organization forward with doing it in the future. +[2483.26 --> 2486.64] And it's like a whole process and a whole framework you have to develop as an engineering team. +[2486.64 --> 2488.86] Like we're lacking this in most of our engineering organizations. +[2488.94 --> 2494.00] Like I have not been to, I think, any companies that have had this type of robust way of talking +[2494.00 --> 2497.70] about technical debt or legacy systems or even just feature work. +[2497.70 --> 2500.16] Like most of the time, it's just like we got a bunch of story points. +[2500.22 --> 2503.54] We got a bunch of sprints and we're going to take a guess and we're going to have retrospectives. +[2503.54 --> 2507.68] But retrospectives are complaining about things that went wrong, not about did we actually +[2507.68 --> 2509.12] meet what we said we were going to meet? +[2509.56 --> 2513.00] So I think there's definitely a lot of area here that we could kind of build up something +[2513.00 --> 2515.22] better for being able to discuss a lot of this stuff. +[2515.22 --> 2519.52] Well, yeah, no wonder if that also calls out there's oftentimes so much ceremony involved +[2519.52 --> 2523.76] that at least my experience has been a lot of engineers are like, well, this is kind of +[2523.76 --> 2528.50] killing productive time because like, you know, if we spend a bunch of time pointing things +[2528.50 --> 2533.82] when we do the work based on scrums, but no one's holding us accountable for our sprints +[2533.82 --> 2538.06] and they actually end up devolving into something like business ends up wanting con bond anyway. +[2538.30 --> 2540.48] So I'm doing all this ceremony work for nothing. +[2540.96 --> 2543.28] Why do I want to add more ceremony onto that? +[2543.28 --> 2548.44] Yeah, I think that's the kind of hole we've dug for ourselves in building a system where +[2548.44 --> 2552.60] we are supposed to be getting these benefits, but then not actually doing the back half of +[2552.60 --> 2553.12] that work, right? +[2553.24 --> 2557.84] Not actually figuring out like, are these systems like is this form of agile we're doing? +[2557.92 --> 2559.08] Are sprints working for us? +[2559.12 --> 2560.40] Are story points working for us? +[2560.58 --> 2564.76] Are we able to actually get the benefits out of them that we that we think we're getting? +[2564.76 --> 2567.12] Or is this just kind of in reality? +[2567.24 --> 2572.12] Yeah, like you're saying a ceremony, something we just do because we've done it, not because we have +[2572.12 --> 2575.12] we're getting any intrinsic value out of it at the end of the day. +[2575.32 --> 2579.26] I feel a whole go time episode coming on on this exact topic. +[2580.38 --> 2584.36] So I kind of I'm interested to hear a little bit more about we've talked about getting to that +[2584.36 --> 2589.64] explosion point where it's falling apart, you know, having bugs every day, you know, +[2589.64 --> 2594.36] making your on call engineers life hell, that is the point at which you're like, we need to do +[2594.36 --> 2594.78] something. +[2595.12 --> 2600.18] But going back to our original question, like, how do we stop legacy code creeping in as we're +[2600.18 --> 2602.62] talking about accumulating tech debt, etc? +[2603.18 --> 2608.86] What are some ways that we can stop getting to that kind of final point of explosion where +[2608.86 --> 2610.44] everything's falling to pieces? +[2610.96 --> 2616.48] Well, so I think one thing that like, if you want to keep from code becoming legacy, I think +[2616.48 --> 2619.12] that testing is a helpful tool for it. +[2619.12 --> 2621.44] It's by no means a solve it. +[2621.56 --> 2627.28] But you know, if you can define your actual problems that you're trying to solve and write +[2627.28 --> 2633.20] tests that exercise those problems, that kind of it'll also help, you know, ensure that +[2633.20 --> 2638.26] you've got discrete chunks doing those things rather than tangled mess all playing together. +[2638.62 --> 2641.10] Sounds like you're advocating for test driven development. +[2642.30 --> 2644.02] I just might be. +[2645.48 --> 2646.56] It's not the cure all. +[2646.56 --> 2649.88] There's oftentimes a lot of places where test driven development doesn't work, especially +[2649.88 --> 2652.74] like lab settings or where you don't know what you're going to do. +[2653.12 --> 2658.64] But it is a tool that helps in more than just like ensuring that your code functions correctly. +[2658.64 --> 2665.58] In my experience, you have to invest in your infrastructure and your monitoring probably before you invest +[2665.58 --> 2670.70] in your in actually rewriting your code in the sense that, for example, if your deployment +[2670.70 --> 2676.38] pipeline takes hours for your changes to get deployed, you need to rewrite the deployment +[2676.38 --> 2678.02] pipeline first, right? +[2678.44 --> 2683.50] Add tests, add monitoring to make sure that you know what your system is doing, like what +[2683.50 --> 2687.88] queries it's executing on the back end or what external calls it's making or whatever it is. +[2687.88 --> 2693.82] And probably do that before you get to that explosion point that Angelica just mentioned. +[2694.06 --> 2696.94] And then the explosion point will be a lot less painful. +[2697.36 --> 2698.28] Yeah, I like that. +[2698.48 --> 2705.56] I like that going small piece by piece and just being in a better way each week probably +[2705.56 --> 2707.02] than you were last week. +[2707.18 --> 2709.20] It feels like a great, great way to do that. +[2709.20 --> 2714.30] And also probably less careful and dangerous, if I can say that. +[2714.56 --> 2720.12] Because when you think about that, you know, rewriting as an entire software is, it is scary, +[2720.70 --> 2721.14] for sure. +[2721.78 --> 2725.84] I think I would say writing documentation or just comments, really. +[2726.28 --> 2728.10] Like I think testing is definitely helpful. +[2728.10 --> 2734.42] But I think if I were to be given a code base that has really good tests, but no comments +[2734.42 --> 2738.00] and documentation or a code base that has documentation comments, but no tests, I would +[2738.00 --> 2738.86] definitely take the latter. +[2739.48 --> 2744.06] Because good comments and documentation tell me the intent of what this thing is supposed +[2744.06 --> 2747.06] to be doing, which then I can go write the test for myself. +[2747.32 --> 2750.84] But when I only have tests and I don't have good documentation or comments, then it's like +[2750.84 --> 2755.64] I have to assume that the tests were also written to do the thing that was intended to be done +[2755.64 --> 2759.68] or kind of try and make, like derive from those tests what's supposed to be being handled +[2759.68 --> 2760.02] here. +[2760.16 --> 2763.64] But I think just going through and being like, okay, here's a function with this name. +[2763.70 --> 2764.42] Here's a type with this name. +[2764.48 --> 2765.64] Here's a method with this name. +[2765.98 --> 2767.50] Does this do the thing I think it does? +[2767.52 --> 2770.42] And then writing that down and making sure everybody has an understanding of that. +[2770.52 --> 2775.28] Real great way to actually like, A, start exploring and understanding a legacy code base, but also +[2775.28 --> 2778.00] helps prevent code bases from becoming legacy. +[2778.00 --> 2782.48] Because then other people can catch the bugs in what you intended versus what's actually +[2782.48 --> 2783.66] written in the code. +[2783.74 --> 2785.90] So that clarity of intent is a good one. +[2786.14 --> 2791.08] Because that's also like comments, documentation, and even just the coding style as well can go +[2791.08 --> 2792.34] a long way towards that. +[2792.78 --> 2797.64] Because naming is one of those things that's really hard and people tend to just eventually +[2797.64 --> 2800.20] give up on or get frustrated and be like, ah, whatever. +[2800.36 --> 2803.08] But it is worth that time to take care of your naming. +[2803.08 --> 2807.66] Because if you can structure your code in a way where the code declares its intention, +[2808.00 --> 2813.24] that's even better than having to say comments about saying what it's going to do, what it's +[2813.24 --> 2813.96] intending to do. +[2814.42 --> 2818.76] I think that's another strength of Go, in my opinion, due to the packages and whatnot, +[2818.92 --> 2822.78] compared to other frameworks and language that I worked with in the past. +[2822.98 --> 2829.66] So you kind of already have a small sense of what something is going to do in terms of +[2829.66 --> 2831.44] a business case and whatnot. +[2831.44 --> 2836.56] Because they are probably properly placed into the right packages. +[2836.56 --> 2839.22] I'm wondering if people have any examples of... +[2839.22 --> 2843.50] Because everything so far, it sounds like having good engineering standards and protocols +[2843.50 --> 2847.88] for your team and organization is a big help in preventing legacy code. +[2848.02 --> 2856.62] I'm curious if others have examples of maybe not just good documentation or good tests and +[2856.62 --> 2860.66] stuff like that that could help in preventing legacy code. +[2860.66 --> 2866.34] Because a lot of times that comes from the whole team that has to kind of agree upon a set of +[2866.34 --> 2869.52] documentation and a kind of coding style. +[2870.20 --> 2874.46] But if there's maybe something like you as an individual can kind of take upon yourself to +[2874.46 --> 2879.44] kind of improve yourself as an engineer, totally not trying to talk about myself. +[2879.56 --> 2884.58] I've definitely found that, at least for me, what the outcome of this is usually writing better +[2884.58 --> 2886.12] documentation or tests or whatever. +[2886.12 --> 2890.82] But I think just taking some extra time to stop and really think about what you're trying to do +[2890.82 --> 2892.96] and what you're trying to implement doesn't have to be a ton of time. +[2893.04 --> 2894.64] I'm not saying take hours and hours and hours. +[2894.80 --> 2898.62] But just sitting there and looking at the thing and then thinking about the problem you're trying +[2898.62 --> 2902.94] to solve for even just a few minutes longer than you would have before, I think can really +[2902.94 --> 2905.42] help you perhaps design something a little better. +[2905.90 --> 2908.28] And I think also just many iterations of things. +[2908.50 --> 2913.24] So if you write something and you budget your time and plan on being able to write it +[2913.24 --> 2917.88] three, four or five times, that gives you the space to write it once, maybe quick and dirty +[2917.88 --> 2921.52] and look at it and say, but is this the way that it should be written? +[2921.82 --> 2924.58] I kind of get that feeling of if this is the right thing. +[2924.64 --> 2929.90] So I think a lot of the times we wind up with legacy code or kind of the bad legacy code +[2929.90 --> 2934.02] because we didn't go through enough iterations of thinking about something. +[2934.44 --> 2938.94] And I think it can be very difficult to sit down and do that because we do have deadlines +[2938.94 --> 2939.62] we're trying to hit. +[2939.62 --> 2942.88] But I think that's one of the things that has definitely made me a better engineer is +[2942.88 --> 2948.20] pushing back on myself and my team and saying, I'm going to take the extra few minutes here +[2948.20 --> 2951.40] to think this through a little bit more because it doesn't feel right. +[2951.74 --> 2957.04] And really just listening to your gut and developing your gut instinct over time to know, okay, +[2957.08 --> 2957.84] I've written something. +[2957.94 --> 2961.72] Okay, this is good versus I've written something and this is rough. +[2962.16 --> 2966.16] And listening to that feeling inside you that's like, it's rough, but I don't know why. +[2966.28 --> 2968.10] And then following that thread. +[2968.10 --> 2971.38] And so I think a lot of times, once again, we kind of ignore that thread because we have +[2971.38 --> 2971.90] things to do. +[2971.98 --> 2974.60] We have features to develop and we have something that works. +[2974.64 --> 2975.62] So why are we going to follow the thread? +[2975.66 --> 2978.82] But I think really just following that thread at the end of the day has definitely helped +[2978.82 --> 2982.94] me write code bases that are much more resistant to legacy. +[2982.94 --> 2986.44] Even when I'm working with a bunch of people that are just trying to move really fast, +[2986.48 --> 2991.14] at least the parts that I'm working on are a little bit more immune to that bad legacy +[2991.14 --> 2995.34] at the end of the day, which can help quite a lot because that kind of thing tends to be +[2995.34 --> 2999.06] contagious and other people tend to pick up on it because they want, like they're working +[2999.06 --> 3000.70] in the code that you wrote and they're like, this is great. +[3000.76 --> 3001.42] This is amazing. +[3001.78 --> 3002.88] Like, I want my code. +[3002.96 --> 3004.60] I want the whole code base to feel like this, right? +[3004.62 --> 3006.22] It's very infectious at the end of the day. +[3006.72 --> 3009.80] Do you think that pair programming could also do this effect? +[3010.26 --> 3011.02] I think it can. +[3011.02 --> 3015.48] If your environment is one that is, I'd say, opportune for pair programming, right? +[3015.54 --> 3019.06] So like you're actually in person, you can actually sit at the same desk, you can actually +[3019.06 --> 3021.80] work through things together at the same time. +[3021.80 --> 3024.38] I think then, yeah, pair programming can be very helpful. +[3025.00 --> 3028.54] I think also just group design, group thought processes, like whenever you can get like +[3028.54 --> 3032.42] people in the same setting and you have enough psychological safety where people can just +[3032.42 --> 3034.78] like ask those questions that they might think are dumb. +[3035.14 --> 3037.70] Because at the end of the day, you know, that thread you have is like the very much a, +[3038.32 --> 3041.12] I don't know what the problem is, but this doesn't feel right. +[3041.24 --> 3044.74] So you have to be in an environment where you can express that and the people you're working +[3044.74 --> 3047.68] with will help you work through that and have enough trust in you. +[3047.98 --> 3050.88] So yes, I definitely think pair programming can be excellent for this. +[3050.88 --> 3054.30] I think whiteboard sessions, if you can be in a space where you can do whiteboarding +[3054.30 --> 3054.90] is great. +[3055.00 --> 3058.70] But I think also just maybe not doing the pair programming part of programming, but just +[3058.70 --> 3061.56] hopping on a call and talking through an idea with people. +[3061.78 --> 3066.20] Maybe if we're a Google Doc, that also is, I think, equivalent and a very helpful way +[3066.20 --> 3067.46] to kind of execute this. +[3068.16 --> 3073.24] So we've talked a little bit about kind of you have this legacy thing and how do you make +[3073.24 --> 3077.04] it so that you don't get to that final explosion point. +[3077.04 --> 3084.28] But I'm interested to talk a little bit about when you, if you get that buy-in to rewrite +[3084.28 --> 3091.22] a legacy code base or to change up a system and, you know, business for whatever reason +[3091.22 --> 3092.88] decide they're having a great day. +[3093.00 --> 3097.74] They're like, you have a year, you have two years, however long it takes, we need to rebuild +[3097.74 --> 3099.10] this thing and make it better. +[3099.38 --> 3104.34] When you are architecting that new solution, we are thinking through how do we build this +[3104.34 --> 3110.90] new thing, how do you, from day one, from the whiteboarding session, build something that +[3110.90 --> 3115.86] is going to avoid being legacy for as long as possible? +[3116.10 --> 3116.46] I.e. +[3116.46 --> 3120.42] is it always having microservices, making sure nothing is deeply coupled? +[3120.64 --> 3123.32] Like what are some things, some questions to be asked? +[3123.80 --> 3128.26] How do you go about designing a new system already with legacy in mind? +[3128.26 --> 3133.70] I would say from the jump, understand what went wrong with your old system. +[3133.90 --> 3140.10] I think people run for microservices or for Kafka or for some new flashy tech at the end +[3140.10 --> 3140.48] of the day. +[3140.56 --> 3143.46] Like they're like, oh, the problem was we weren't using microservice or something. +[3143.60 --> 3148.22] But it's like, sit down and actually think like, why don't you like working this good? +[3148.30 --> 3151.18] Why, why have we had to declare bankruptcy on this code base? +[3151.50 --> 3156.76] A thing that I have done that has hilariously sometimes wound up with me not actually abandoning +[3156.76 --> 3160.62] old code bases when I really wanted to, because like the problems I wanted to fix, I could +[3160.62 --> 3161.92] just fix in that old code base. +[3161.96 --> 3163.82] So why make a whole new one with all new problems? +[3164.24 --> 3168.50] But I think that's definitely where I tend to start is like, what is it about this code +[3168.50 --> 3172.14] base I'm working on now that's making it so I want to build a new one? +[3172.70 --> 3174.16] John, what do you think? +[3174.32 --> 3177.22] I know this is something that I'm hoping you're thinking about. +[3178.80 --> 3182.32] I don't know if I got too much on this one. +[3182.32 --> 3188.18] But also off the record, for those who didn't see my cheeky smirk, John and Jeff, in fact, +[3188.24 --> 3189.26] are both on my team. +[3189.48 --> 3193.84] So when I make any cheeky comments in their direction, it is slightly more layered than +[3193.84 --> 3196.66] other people who don't have that context may know. +[3197.16 --> 3199.04] And I am their technical product manager. +[3199.52 --> 3201.70] So that's also a context that you should know. +[3201.84 --> 3205.12] I am speaking off the record as a business person. +[3205.56 --> 3206.68] Undercover business person. +[3206.92 --> 3207.62] I'm undercover. +[3207.78 --> 3208.66] Secret gopher. +[3208.66 --> 3210.28] Well, overt gopher. +[3210.44 --> 3211.62] Secret product manager. +[3211.62 --> 3218.72] I don't know, Jeff, like, what are some questions that you would hope if I was to go to you, +[3218.78 --> 3222.92] Jeff, and be like, hey, Jeff, I'm thinking we should completely get rid of our current +[3222.92 --> 3224.28] platform and rewrite it. +[3224.62 --> 3228.22] What would be things that would pop to your head as things you would need to, you would +[3228.22 --> 3233.40] want to be given time to think about before we kind of went, right, this is the new architecture. +[3233.40 --> 3234.74] Let's go code. +[3234.74 --> 3238.78] Yeah, I'm definitely following what Chris is saying. +[3238.78 --> 3242.88] It's definitely you have to understand why you're trying to rewrite everything, right? +[3243.28 --> 3247.54] You don't want to rewrite just for the sake of, oh, we have lots of issues with this codebase. +[3247.70 --> 3251.74] Maybe we should just declare bankruptcy and start something new. +[3252.04 --> 3255.06] It's definitely something I myself have been interested in. +[3255.06 --> 3266.42] And I like for our team specifically, we're in the middle of kind of rewriting our two platforms and getting some historical context as to why we went that route. +[3266.66 --> 3267.72] It would be interesting. +[3267.82 --> 3268.50] But I do not. +[3268.62 --> 3269.10] I don't know. +[3269.12 --> 3273.74] I don't want to trigger John's PTSD too much because it looks like he's going through something. +[3274.00 --> 3274.88] The question came up. +[3274.88 --> 3275.52] Yeah. +[3276.52 --> 3278.70] And I guess rewrites are tough. +[3278.84 --> 3288.32] I feel like everybody at some point in their career needs to go through the let's just do a rewrite because of it and feel the pain and watch it fall flat on its face and waste like six months of their lives. +[3288.72 --> 3289.68] Six months if you're lucky. +[3290.18 --> 3290.50] Yeah. +[3290.70 --> 3294.08] I'm like I've seen like entire years go down the drain. +[3294.08 --> 3299.68] So having an understanding of the risk involved is huge. +[3300.28 --> 3305.82] And you know if there are scenarios that do warrant it but they're the exception rather than the rule. +[3305.82 --> 3322.64] So am I right in assuming that if your product partner or the business came to you, I'm speaking to all of you here, and said, right, Chris, Misha, Dominic, I know you've been working with this 20 plus year in Dominic's case infrastructure. +[3322.64 --> 3327.64] We're going to give you five years maximum to rewrite completely. +[3328.20 --> 3334.16] Would there be cases in which you would then, after assessing, be like, no, actually we don't want to rewrite? +[3334.56 --> 3337.96] Or would you always, if given the opportunity, want to rewrite? +[3338.06 --> 3343.30] If given the buy-in, the money, the time to do it, would you always jump at that opportunity? +[3343.72 --> 3350.04] I would only rewrite a software that works if the technology is not going to exist sooner. +[3350.24 --> 3352.16] Take like old system like COBOL. +[3352.64 --> 3359.46] In Quebec, there is a couple of companies that are still built on top of AS400 and COBOL and whatnot. +[3360.20 --> 3362.94] I mean, there's no manpower anymore for that. +[3363.24 --> 3364.00] So what do you do? +[3364.36 --> 3365.36] You don't have a choice. +[3365.90 --> 3375.78] But I would be extremely careful to rewrite the software not built from a long-time software from scratch. +[3375.78 --> 3382.12] I would still try to see if there is something that can be done other than a complete rewrite. +[3382.36 --> 3383.32] I'm in agreement there. +[3383.54 --> 3390.96] I mean, I think when it comes to Greenfield projects or when it comes to rewriting legacy projects, just don't. +[3390.96 --> 3393.44] Like, I think that's kind of the hard and fast rule most of the time. +[3393.44 --> 3401.02] Because like, sure, there's usually like, usually we get to the point where we wanted to rewrite something when there's like, I would say 20 to 30% of it. +[3401.22 --> 3406.02] Like maybe like a fifth to a third of the project is just obnoxious to work within. +[3406.02 --> 3411.26] But there is still 70 to 80% of it that is doing things correctly. +[3411.48 --> 3419.30] When you do a rewrite, you have to account for all those things that it does correctly and then replicate them and then also fix all those bugs. +[3419.50 --> 3428.34] So it's always going to take way, way, way, way, way longer to rewrite the thing than it is to fix the problems in the thing that exists. +[3428.80 --> 3430.88] So I think that's the thing that we have to account for. +[3430.88 --> 3437.34] And if you go into it and you do that math and you say, yep, nope, we want to rewrite because we have X, Y, and Z reasons why. +[3437.74 --> 3438.76] Sure, go ahead and go forth. +[3438.90 --> 3445.74] But if it's just like, here are some individual pain points that I don't like about this system and it's causing us to slow down some, let's rewrite the whole thing. +[3445.98 --> 3446.70] Like, absolutely not. +[3446.78 --> 3452.18] Like, you have to do like in-depth research and understanding of the project that you're dealing with before you can rewrite it. +[3452.22 --> 3456.30] Like, you should actually be able to rewrite it before you start rewriting it. +[3456.30 --> 3462.60] Which sounds like something that sounds super obvious when said, but is a thing that I have witnessed over and over and over again. +[3462.72 --> 3464.88] It's not something that we do as engineers in general. +[3465.62 --> 3470.54] Yeah, there's got to be some fundamental problem that requires like a complete greenfield rewrite. +[3470.74 --> 3473.48] It can't just be pain, which I think everyone's getting at. +[3473.74 --> 3475.08] But also, it shouldn't come from product. +[3475.20 --> 3481.24] It should be tech being like there's this technical roadblock that means, you know, we just cannot do this anymore. +[3481.24 --> 3490.66] Because, you know, if you want to, if you've got like this legacy code base and you just want to make it better and get there, there's ways that you can like boil things out too. +[3490.82 --> 3499.96] So, you know, if you can be like, well, we're going to move this small set of functionality to a new thing and piecemeal things out if you really want an entirely new code base versus improving it. +[3500.22 --> 3504.38] But again, that should be, I think, tech driven rather than coming from product. +[3504.38 --> 3514.32] If you can't come up with a way of conceivably actually retiring completely the thing that you're rewriting at the beginning of your rewrite, then you shouldn't do it. +[3514.46 --> 3519.26] Because you're not going to be able to retire it if you don't already have a plan for how to do it. +[3519.70 --> 3521.58] I think that's just like fact at this point. +[3521.66 --> 3524.82] Like I've never seen anywhere that's been able to successfully retire something. +[3524.82 --> 3540.76] Either it's like you've been able to successfully do it because you had a plan or you were able to successfully do it because you just threw engineering bodies at the problem and basically said all other work halts until we finished this movement and this migration, which is incredibly difficult. +[3540.76 --> 3554.92] And really requires buy-in of, you know, all the way up to the CTO level at the end of the day and probably the CEO level to really say, no, we're actually going to, no matter what, throw as many bodies as is required to get rid of this thing. +[3555.24 --> 3558.12] Either you have to do that or you have to have a good plan from the beginning of how you're going to do it. +[3558.38 --> 3559.48] And most of the time we don't have plans. +[3559.54 --> 3560.90] So we're really saying the latter. +[3561.08 --> 3569.06] And if you're not going to go spend that political capital to buy into that, then just don't rewrite the thing and really sit down and think about like what is frustrating you about this code base. +[3569.06 --> 3577.34] Because I would say for most of the code bases, except for the ones, you know, once again, written maybe in COBOL or in some other really old language that we don't want to deal with anymore. +[3577.50 --> 3581.54] For the most part, especially if you're like, if it's a Go code base and you're like, I want to rewrite this. +[3581.64 --> 3582.52] Just stop. +[3582.78 --> 3584.84] Just stop and be like, what's bothering me about this? +[3584.88 --> 3586.02] And then go fix those things. +[3586.48 --> 3594.92] That will take tremendously less time and you will be happier with the results because now you will have the code base you wanted without the code base you didn't. +[3595.02 --> 3599.00] You don't have this legacy code base that someone gets stuck maintaining until your new thing. +[3599.06 --> 3604.20] It actually does what you promised it would do, which also will take a lot longer than you think it's going to take. +[3604.64 --> 3605.86] The rewrite always takes longer. +[3605.94 --> 3616.06] And if you can't shut down the new thing, the old thing until the rewrite is done, well, now someone has to keep fixing bugs and keep that old thing operable until the new thing is ready to go in full capacity. +[3616.56 --> 3620.68] So if you really don't want to work on the old thing anymore, just fix the old thing in place. +[3620.68 --> 3621.74] You know how it is. +[3621.78 --> 3627.70] Even if everyone is agreeing to not touch the old thing anymore, there will always have new feature. +[3627.80 --> 3631.74] There will always have bug fixes that you will need to replicate on the new system. +[3631.84 --> 3632.94] It's extremely difficult. +[3633.38 --> 3639.68] Don't assume that you will not have to add something new to the old system until the old system is completely off. +[3639.68 --> 3651.48] So if your new thing is I'm going to go build something new and it's going to take me two and a half years to build it, assume that you have to keep adding stuff to the old thing until that two and a half year mark hits and you actually get to turn that old thing off. +[3651.86 --> 3656.98] Moral of the story is Greenfield rewrites are not always the right direction to go. +[3656.98 --> 3663.64] You need to sit, take some time, think it through, identify the core issues, and then get really good. +[3663.90 --> 3668.92] Nod to your wonderful series, Chris, at maintaining your software and your technology. +[3669.58 --> 3677.10] Yeah, I think I have participated in one successful Greenfield rewrite in my 20-year career. +[3677.40 --> 3678.32] And where was that, John? +[3678.52 --> 3679.46] That is with us. +[3680.00 --> 3681.40] And who was your product manager? +[3681.50 --> 3682.46] That is with the Times. +[3682.46 --> 3683.32] Yep, that is. +[3683.32 --> 3692.24] I also have a side note about the term Greenfield, which I find hilarious because if you want to have an actual Greenfield, someone has to maintain that. +[3692.52 --> 3694.54] Like if you go into nature, you don't find a lot of Greenfield. +[3694.58 --> 3698.22] You find a lot of mud pits and stuff and a lot of overgrown grass and stuff. +[3698.28 --> 3699.38] But a real Greenfield? +[3699.50 --> 3701.90] Now you got to have like good lawn equipment, good agriculture. +[3702.34 --> 3705.72] It takes a lot of work, a lot of maintenance to have a Greenfield. +[3705.88 --> 3708.72] And people are just like, we're just going to start with a Greenfield. +[3708.72 --> 3711.70] It's like you're going to start with something that's well-maintained and well-manucured. +[3711.70 --> 3716.60] Okay, so it's like we should really call it like mud pit like development. +[3716.92 --> 3718.32] And then people are probably... +[3718.32 --> 3721.28] The grass is always greener on the other side. +[3721.40 --> 3725.76] Because the other person actually maintains their lawn unlike you, right? +[3725.90 --> 3728.62] Their grass is cleaner because they take care of it. +[3728.80 --> 3735.34] You go over there, it might be green for a little bit, but then your old grass is going to be greener because that person maintains it now. +[3735.72 --> 3736.82] So you're always going to want to... +[3736.82 --> 3740.08] It's like, no, no, just learn how to maintain your grass and then you'll have a really green lawn. +[3740.44 --> 3741.54] And also don't envy your neighbor. +[3741.54 --> 3742.58] There's lots of lessons in here. +[3742.60 --> 3744.28] I can take this analogy all day long. +[3744.44 --> 3744.88] You know, this is... +[3744.88 --> 3747.28] I feel like we're getting into the Canterbury Tales here. +[3747.84 --> 3750.58] This is what happens when you have two writers on a podcast, Angelica. +[3750.78 --> 3751.18] I know. +[3751.24 --> 3752.24] We just feed off each other. +[3752.76 --> 3755.42] Don't even get me started on Shakespeare in reference. +[3756.06 --> 3757.38] We'll be here all night. +[3758.14 --> 3760.54] To rewrite or not to rewrite, that is the question. +[3762.56 --> 3764.28] Whether it is greener or the other side. +[3764.34 --> 3765.18] Okay, I'm going to stop now. +[3765.70 --> 3766.76] I could go on for days. +[3766.76 --> 3767.24] Great. +[3767.24 --> 3767.48] Great. +[3767.80 --> 3769.26] Well, I... +[3769.26 --> 3775.16] Regrettably, we have to move on to our next section because I want to hear your unpopular +[3775.16 --> 3775.80] opinions. +[3775.80 --> 3778.70] But it's been an absolute pleasure chatting about Legacy Code. +[3778.88 --> 3781.42] And I'm sure myself and Chris have many an episode. +[3781.80 --> 3783.96] As a follow-up, we now will like to do. +[3784.50 --> 3787.24] This maintenance miniseries is becoming a maintenance series. +[3787.68 --> 3788.68] It's just kind of expanding. +[3788.92 --> 3789.20] Yes. +[3789.32 --> 3793.30] I feel like we should stop calling it mini and just go with, this is a series. +[3793.86 --> 3794.82] Maintenance is important. +[3795.44 --> 3795.98] It's a thing. +[3796.30 --> 3798.16] We should make its own little theme tune. +[3798.26 --> 3799.08] I'm ready, Chris. +[3799.12 --> 3800.28] Let's whip out the keyboard. +[3800.46 --> 3802.90] I'll do the vocals and I'll like rap a little bit. +[3802.94 --> 3803.42] It'll be great. +[3803.54 --> 3804.08] A new jingle. +[3804.20 --> 3805.24] We'll make that a note for Jared. +[3805.46 --> 3807.36] Hey, Jared, we need a jingle for the maintenance series. +[3808.16 --> 3808.40] Yeah. +[3808.84 --> 3812.62] And on the topic of jingles, here's our unpopular opinion one. +[3812.94 --> 3813.88] Smooth transition. +[3813.88 --> 3819.08] Unpopular opinion. +[3819.34 --> 3820.26] You want to. +[3820.30 --> 3822.04] I actually think she'd probably leave. +[3825.16 --> 3826.92] Unpopular opinion. +[3830.20 --> 3837.62] So for all you lovelies who have not heard Go Time Unpopular Opinions before, this is +[3837.62 --> 3843.34] where we ask our lovely guests and also Chris, if we have time, for an unpopular opinion. +[3843.34 --> 3845.16] It does not need to be about Go. +[3845.28 --> 3846.72] It does not need to be about technology. +[3846.90 --> 3849.78] It can be about your aunt's favorite sock collection. +[3850.04 --> 3851.40] It can be about China dolls. +[3851.68 --> 3854.84] It can be about your view on the ethos and life. +[3855.38 --> 3856.26] It can be on anything. +[3856.78 --> 3862.28] First, I'm going to turn over to you, Jeff, for an unpopular opinion. +[3862.58 --> 3867.96] Now, the goal of this is to come up with an opinion that you believe will be unpopular +[3867.96 --> 3869.92] because we're going to post this opinion. +[3869.92 --> 3874.22] I'm going to tweet about it, have our lovely Gopher community vote on it, and then we're +[3874.22 --> 3876.78] going to tell you what percentage unpopular it was. +[3877.36 --> 3879.00] And if it's popular, you have to come back. +[3879.26 --> 3879.40] Yeah. +[3879.42 --> 3883.74] And if it's popular, we're getting you back on Go Time to grill you again to get a better +[3883.74 --> 3884.82] unpopular opinion. +[3885.26 --> 3889.24] What if I just use that as a way to keep coming back on the show and just keep keeping popular +[3889.24 --> 3890.64] opinions because I like it here? +[3891.50 --> 3892.84] We'll sniff you out. +[3893.16 --> 3895.30] We're savvy individuals on this show. +[3895.30 --> 3899.60] So I've been racking my brain with this ever since I agreed to be on Go Time. +[3899.86 --> 3906.14] So I don't think I have anything earth-shattering like John's previous unpopular opinion. +[3906.32 --> 3913.18] But what I came up with is I do not like any type of yogurt. +[3913.80 --> 3917.48] I feel like that's a very popular snack that people like to eat. +[3917.80 --> 3919.72] I just do not like yogurt at all. +[3919.90 --> 3921.54] I have a story actually about that. +[3921.54 --> 3927.52] So I bought Greek yogurt as a substitute for sour cream because I heard that's a good, +[3927.66 --> 3929.42] healthy substitute. +[3929.96 --> 3932.16] And it was my first time actually trying it. +[3932.42 --> 3936.18] I opened up the container and it just smelled really funky. +[3936.84 --> 3941.50] And I told my college roommate at the time, did this go bad? +[3941.78 --> 3943.18] And he came over and smelled it. +[3943.18 --> 3944.58] He's like, no, that's just how it smells. +[3944.70 --> 3946.70] I'm like, oh, definitely not for me. +[3946.70 --> 3953.38] Wait, are we talking just like plain or are we talking like mix in with some granola and +[3953.38 --> 3953.68] fruit? +[3953.78 --> 3955.12] You also don't like it? +[3955.20 --> 3955.64] Are we talking? +[3955.96 --> 3956.94] I don't like any yogurt. +[3956.98 --> 3958.58] I've never been like a yogurt person. +[3958.70 --> 3959.70] Like Greek only? +[3960.06 --> 3961.14] Are we talking about yope? +[3961.32 --> 3962.66] Like drinkable yogurt? +[3963.06 --> 3963.80] No yogurt. +[3963.98 --> 3964.44] Oh my gosh. +[3964.44 --> 3968.34] This is like an umbrella yogurt band in the Jeff Household. +[3968.80 --> 3969.58] Basically, yes. +[3970.16 --> 3972.62] Fascinatingly, I thought I always hated Greek yogurt. +[3972.84 --> 3974.60] And for some reason, I now like Greek yogurt. +[3974.70 --> 3975.48] I don't know why. +[3975.48 --> 3978.50] So I'm kind of on the opposite end of the spectrum as you, dude. +[3979.04 --> 3983.12] Well, Chris, you know, your taste buds do mature as you age. +[3987.46 --> 3989.00] That's a good unpopular opinion. +[3989.46 --> 3990.22] I love yogurt. +[3990.48 --> 3991.30] I love Greek yogurt. +[3991.42 --> 3993.70] Also, it's very good for your digestion, I'm told. +[3994.20 --> 3997.62] And it has things called active things. +[3998.48 --> 3999.48] But I also... +[3999.48 --> 3999.50] Active. +[4000.14 --> 4000.62] Active. +[4000.88 --> 4001.40] What are they called? +[4001.66 --> 4002.06] Probiotics. +[4002.28 --> 4002.64] Probiotics? +[4003.06 --> 4003.46] Yes. +[4003.46 --> 4008.98] They have active probiotics that are apparently, I've been told by being brainwashed by the media, +[4009.12 --> 4010.36] are very good for my digestion. +[4010.36 --> 4013.94] But I have no scientific facts to back up that view. +[4014.36 --> 4015.10] Purely from like... +[4015.88 --> 4018.82] But I will, after this, Google it and check. +[4019.08 --> 4024.06] But maybe as you get older, like Chris, you'll develop a taste. +[4024.06 --> 4028.16] I get all my active probiotics from Yipolt, if anyone else has had that. +[4028.56 --> 4030.50] I don't know what I would describe it as. +[4030.56 --> 4032.46] It's like a small little drink. +[4033.02 --> 4034.70] It's not yogurt, as far as I know. +[4035.22 --> 4039.64] But that's where I receive my active probiotics, since I don't have any yogurt. +[4039.64 --> 4048.14] World's leading probiotic beverage created in Japan in 1935 and is now sold in over 40 countries. +[4048.56 --> 4050.24] And it's not yogurt, so... +[4050.24 --> 4051.06] It's not yogurt. +[4051.22 --> 4055.12] I actually felt like it was, and I was going to Google it in the hopes that I would catch you out. +[4055.52 --> 4056.50] But it's not. +[4056.66 --> 4058.02] I've done my research, so... +[4058.02 --> 4058.22] Oh. +[4059.22 --> 4060.06] It won't catch me. +[4061.20 --> 4063.02] Misha, unpopular opinion. +[4063.02 --> 4064.22] Let's see. +[4064.44 --> 4064.80] All right. +[4064.96 --> 4074.76] So, my unpopular opinion that is not about Go is that CSS is a full-fledged programming language +[4074.76 --> 4078.42] that will someday replace all other programming languages. +[4078.96 --> 4085.16] Now I just need to figure out how to rewrite all of my backend microservices in CSS. +[4086.10 --> 4088.10] I don't even know how to respond to that one. +[4088.62 --> 4089.72] You broke my brain, I think. +[4089.72 --> 4092.08] That will solve the legacy problem, for sure. +[4092.08 --> 4094.24] Yeah, let's just write everything in CSS. +[4095.08 --> 4099.88] Does it have to be CSS, or can I write it in Sass and then transpile it to CSS? +[4100.32 --> 4101.04] Sass is fine. +[4101.62 --> 4103.30] I'm getting snazzy with it here. +[4103.54 --> 4109.62] Well, you can't do math in CSS now, so I can kind of see how you could create a virtual machine +[4109.62 --> 4113.28] out of CSS, where, like, the bytecode is CSS, and then... +[4113.28 --> 4113.70] Oh, God. +[4113.86 --> 4117.88] Now we're going to wind up with someone writing a C compiler that compiles C into CSS that can +[4117.88 --> 4118.38] actually... +[4118.38 --> 4119.22] Nah, this is terrible. +[4119.66 --> 4120.08] This is... +[4120.08 --> 4120.50] Oh, no. +[4120.56 --> 4121.60] What have we started? +[4122.08 --> 4126.18] I would be surprised if that was not unpopular. +[4126.82 --> 4128.44] That's a very good unpopular opinion. +[4129.00 --> 4129.36] Thank you. +[4130.04 --> 4132.18] But I do now really want someone to do that. +[4132.40 --> 4133.52] I want to see this compiler. +[4134.02 --> 4134.90] It's a challenge. +[4135.86 --> 4136.90] Dominic, over to you. +[4137.38 --> 4141.70] In fact, Misha, you can do it, and then come back on the show, and we can talk about it. +[4141.78 --> 4142.60] You just got to write it in Go. +[4142.66 --> 4143.14] Sounds good. +[4143.42 --> 4143.86] I'm ready. +[4144.46 --> 4146.12] Wait, I swear that defeats the purpose. +[4146.12 --> 4150.70] I swear, if you write it in Go, is it not then fully CSS? +[4151.04 --> 4155.62] You can eventually bootstrap it so it's self-hosting on CSS, but you have to write it in some... +[4155.62 --> 4156.20] Does that count? +[4156.34 --> 4156.58] Yeah. +[4156.98 --> 4159.06] Go was written in Go, but it didn't start off being written in Go. +[4159.14 --> 4160.16] It started off being written in C. +[4160.58 --> 4161.08] Oh, okay. +[4161.70 --> 4165.86] Misha, in the definition for your unpopular opinion that you just came up with, does that +[4165.86 --> 4166.16] count? +[4166.42 --> 4166.74] Yes. +[4166.92 --> 4167.62] Yeah, that would count. +[4167.84 --> 4168.60] Can we do this? +[4168.94 --> 4170.68] Okay, it's official now. +[4172.86 --> 4174.00] Dominic, hit me. +[4174.18 --> 4174.88] Unpopular opinion. +[4175.40 --> 4179.58] Well, last time I thought it was the highest I could go, so let's see. +[4180.24 --> 4181.88] Clearly not, because you're back. +[4182.02 --> 4182.84] No, clearly not. +[4182.90 --> 4183.16] Exactly. +[4183.78 --> 4187.30] GitHub co-pilot will do more harm than good, I think. +[4187.38 --> 4188.58] Oh, that's not going to be unpopular. +[4189.08 --> 4189.30] No? +[4189.48 --> 4189.74] Yeah. +[4190.66 --> 4193.30] It's more of an opinion that needs to be put out there. +[4193.38 --> 4194.00] Yeah, I know. +[4194.00 --> 4196.20] I was supposed to think about it. +[4196.20 --> 4200.62] Maybe I will pass, because like I was saying at the beginning, I caught the COVID yesterday, +[4200.86 --> 4206.14] so I was supposed to think about that yesterday, and to be frank, I did not. +[4206.34 --> 4208.60] So yeah, let me pass on that. +[4208.84 --> 4212.34] Shout out to Dominic for doing GoTime with active COVID. +[4212.60 --> 4213.50] Yeah, it was rough. +[4214.18 --> 4214.80] Real trooper. +[4215.12 --> 4216.36] Yeah, thank you so much. +[4216.60 --> 4218.32] Yeah, that's my popular opinion, in fact. +[4219.40 --> 4220.90] Doing things with the COVID. +[4221.44 --> 4222.12] Powering through. +[4222.12 --> 4230.24] All right, John, you are the person who holds one of the most unpopular GoTime opinions ever. +[4231.10 --> 4236.32] Shameless plug to a previous GoTime episode where we go through all of our unpopular opinions of the year. +[4236.60 --> 4239.84] I don't know exactly what number it is, but if you DM me, I'll let you know. +[4239.88 --> 4240.68] It's a great episode. +[4240.68 --> 4244.64] Where John said that he hates chocolate. +[4245.16 --> 4247.30] So I'm ready for this next one. +[4247.36 --> 4248.26] It better be a good one. +[4248.34 --> 4250.30] My expectations are sky high. +[4250.56 --> 4255.38] I'm going to swing for the fences on this one and say that I think adding generics to Go was a mistake. +[4255.88 --> 4256.36] Ooh. +[4256.36 --> 4266.16] So while there are use cases where generics actually are very handy, I think overall I have seen more harm done with generics than in general. +[4266.16 --> 4273.48] Because I think that it encourages developers to try and deduplicate things that are not necessarily the same. +[4274.14 --> 4285.54] And when you're trying to reuse for the sake of reusability and accomplish different tasks with code, you oftentimes end up with code that is much more of a tangled mess and harder to maintain and harder to understand. +[4285.54 --> 4289.20] And I think the generics encourage that. +[4289.68 --> 4293.32] I think like everything in life, balance will return at some point. +[4293.72 --> 4296.12] But in my opinion, it will be good. +[4296.22 --> 4301.40] Just think about all the loops that we are doing that we will not be forced to do. +[4301.56 --> 4305.96] Like, you know, removing an item from a slice, for example, or I don't know. +[4305.96 --> 4312.86] So I think for the next six months, it will be like very dirty everywhere. +[4313.16 --> 4314.24] Everyone will try that. +[4314.36 --> 4316.06] But I'm not sure. +[4316.32 --> 4317.76] I disagree with that, respectfully. +[4318.70 --> 4319.18] Totally. +[4319.30 --> 4323.78] And I totally get you where like those loops and whatnot, where there's cases where it totally makes sense. +[4324.08 --> 4329.30] But I think that having those cases is a lesser sin from the abuse of generics that's going to follow. +[4329.84 --> 4330.10] Yeah. +[4330.24 --> 4331.02] Yeah, I see the point. +[4331.44 --> 4334.20] I hope we will see a balance at some point because... +[4334.20 --> 4336.28] It's going to be channels and good routines all over again. +[4336.40 --> 4338.82] People are just going to sprinkle that stuff all over their code bases. +[4339.68 --> 4340.00] Yep. +[4340.28 --> 4341.78] Oh, can I use a channel here? +[4341.84 --> 4343.12] I'll use a channel because I can't. +[4343.14 --> 4344.40] Oh, now I need to use seven channels? +[4344.50 --> 4346.38] I'll use all seven of those channels just because I can. +[4346.42 --> 4347.60] I'll sprinkle good routines everywhere. +[4347.78 --> 4349.50] It's just going to be that all over again. +[4349.64 --> 4349.82] Yep. +[4350.14 --> 4353.90] Need I return to my first app ever that I think, Chris, you've seen, +[4353.90 --> 4360.28] where I literally ended up having like 317 channels individually written, +[4360.88 --> 4363.34] partly because I didn't know you could send stuff to one channel, +[4363.34 --> 4364.92] but also because I just had fun. +[4365.08 --> 4366.40] I was like, oh my God, I could create all this... +[4366.40 --> 4368.42] And then I had like pointers to channels. +[4368.80 --> 4369.46] Oh boy. +[4369.66 --> 4372.16] It was all fun, but clearly not the right approach. +[4372.36 --> 4373.12] Difficult to maintain. +[4373.30 --> 4373.54] Oh yeah. +[4373.60 --> 4376.26] And then I added like go routines onto every single one. +[4376.90 --> 4378.54] So that was fun. +[4381.10 --> 4383.32] Yeah, Chris, I know you loved it when I showed you. +[4383.52 --> 4385.10] Like, look how pretty it is. +[4385.30 --> 4385.60] No. +[4388.22 --> 4391.60] Chris, do you have an unpopular opinion to finish us up with? +[4391.60 --> 4393.66] I have many an unpopular opinion. +[4393.66 --> 4396.92] I think I hold one of those most like coveted unpopular opinions. +[4396.92 --> 4397.88] You absolutely do. +[4397.88 --> 4398.58] Unpopular opinions. +[4399.58 --> 4400.04] Okay. +[4400.16 --> 4402.04] So this one is tech related, not Go related. +[4402.56 --> 4410.28] But I think that we should kill off all C-like languages and C-derived languages +[4410.28 --> 4414.28] and get rid of the von Neumann architecture and move on as an industry. +[4414.88 --> 4416.70] Do you have a beat more as to why? +[4417.24 --> 4417.94] Yes, actually. +[4417.94 --> 4421.98] So when the von Neumann architecture was created, like we were still using discrete parts. +[4422.08 --> 4426.32] But even when we made the first transition in the 70s to transistors, we had on the order +[4426.32 --> 4429.48] of thousands of transistors operating at kilohertz. +[4429.74 --> 4434.54] And now, as of a couple of weeks ago, you can buy a chip, buy a computer on the market that +[4434.54 --> 4439.98] has over 100 billion transistors operating at several gigahertz each. +[4439.98 --> 4445.44] So we've had many, many, many, many, many orders of magnitude along two different dimensions +[4445.44 --> 4447.74] of improvement in our hardware capabilities. +[4447.74 --> 4453.28] But we still use the same fundamental architecture that is the same fundamental bottlenecks attached +[4453.28 --> 4459.92] to it, which ironically is turning our systems, our chips into distributed systems just to keep +[4459.92 --> 4461.82] pace with the improvements that we've made. +[4462.12 --> 4465.96] And I think all of this extra work and all of this strife is caused by the fact that we're +[4465.96 --> 4470.00] still using this architecture, and the reason we still use it is because, well, it's the +[4470.00 --> 4471.32] architecture of C at the end of the day. +[4471.40 --> 4476.86] C very much programs in this mentality that recognizes the von Neumann architecture as being +[4476.86 --> 4477.06] there. +[4477.14 --> 4482.24] It's very difficult to write truly parallel or good parallel code or concurrent code in +[4482.24 --> 4482.48] C. +[4483.04 --> 4489.28] So jettison the von Neumann architecture and along with it, C and C-like languages, which +[4489.28 --> 4491.22] unfortunately includes Go as of right now. +[4491.22 --> 4495.78] Because while you can write some good concurrent code in Go, it's still pretty difficult. +[4496.20 --> 4500.48] You know, we want languages that look more like Erlang, or at least have the same underpinnings +[4500.48 --> 4505.16] of things like Erlang at the end of the day to really capture the capabilities of the silicon +[4505.16 --> 4505.74] that we have. +[4505.96 --> 4507.22] Because it is, you know, quite incredible. +[4507.34 --> 4511.76] I mean, Apple did just come out with a chip that has 115 billion transistors on it. +[4512.08 --> 4516.80] And it's like, yeah, we can't really utilize all of those if we're using them in our CPU +[4516.80 --> 4517.56] based architecture. +[4517.56 --> 4521.64] So more stuff that looks like graphics cards, too, because they're very good at capturing +[4521.64 --> 4524.66] this idea of parallelizable and concurrent workloads. +[4524.92 --> 4527.76] But yes, no more C, no more von Neumann. +[4528.16 --> 4528.98] Get rid of it. +[4529.12 --> 4530.10] Move on as an industry. +[4530.38 --> 4530.88] Let's get better. +[4530.98 --> 4534.74] Let's stop wasting so many CPU cycles waiting for our main memory to get back to us. +[4535.00 --> 4539.10] Yeah, that may be an unpopular one, especially if you add in that like very blase, like, +[4539.14 --> 4540.50] oh, you know, Go is included. +[4541.02 --> 4544.76] I mean, it's like, it's all C, like, you know, Java, Go, Rust, C, C++. +[4544.76 --> 4548.68] You know, all those popular imperative languages, just goodbye. +[4549.16 --> 4549.92] Time for something new. +[4550.16 --> 4551.98] Let's rewrite Linux and Erlang. +[4552.30 --> 4552.84] It will be fun. +[4553.28 --> 4556.14] So on that fine point, let's get rid of Go. +[4556.64 --> 4558.82] Regrettably, that is the end of our episode. +[4559.38 --> 4560.12] Thank you all. +[4560.18 --> 4562.22] It was an absolute pleasure having you on. +[4562.28 --> 4565.24] And I hope we'll have you all back very soon. +[4565.36 --> 4567.46] Subject, how unpopular your opinions are. +[4567.46 --> 4573.80] That is Go Time for this week. +[4574.12 --> 4575.30] Thanks for hanging with us. +[4575.74 --> 4581.34] Don't forget to follow GoTimeFM on Twitter so you can get in on our unpopular opinion polls. +[4581.78 --> 4585.94] And if this is your first time listening, subscribe now at GoTime.fm. +[4586.32 --> 4592.10] There you'll also find lists of recommended episodes, listener favorites, and a request form +[4592.10 --> 4595.06] so you can let us know what you want to hear about on the pod. +[4595.06 --> 4598.34] Special thanks to Fastly for being our CDN partner. +[4598.54 --> 4601.00] Our episodes reach you fast because of Fastly. +[4601.28 --> 4603.26] And to Breakmaster Cylinder for the fresh beats. +[4603.64 --> 4605.16] Finally, thank you for listening. +[4605.36 --> 4606.06] We appreciate you. +[4606.54 --> 4609.50] Next week, yours truly is guest hosting. +[4609.82 --> 4615.36] I've got Ian Lopshire and Chris Brando answering my newbie Go Curious questions. +[4615.86 --> 4616.64] Stay tuned. +[4616.78 --> 4619.80] We'll have that one ready for you next time on Go Time. +[4619.80 --> 4620.10] Go Time. +[4620.10 --> 4620.14] Go Time. +[4620.14 --> 4622.14] Go Time. +[4622.14 --> 4622.58] Go Time. +[4622.58 --> 4623.14] Go Time. +[4623.14 --> 4623.58] Go Time. +[4625.06 --> 4626.12] Go Time. +[4631.40 --> 4635.96] Go Time. +[4638.10 --> 4638.68] Go Time. +[4638.74 --> 4650.66] Go Time. +[4650.66 --> 4650.74] Go Time. +[4650.74 --> 4650.84] Go Time. +[4651.38 --> 4653.40] Go Time. diff --git a/Inside GopherCon_transcript.txt b/Inside GopherCon_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..1bdbb0442db023cecee58a12b1d1cde7281214dd --- /dev/null +++ b/Inside GopherCon_transcript.txt @@ -0,0 +1,283 @@ +**Angelica Hill:** Hello, and welcome to Go Time! Today we are going to be talking about GopherCon. We're going to talk a little bit about kind of where did GopherCon come from, how is it put together every year... We'll be hearing about some of the logistics, the difficulties, the challenges, but also some of the fun stuff about putting a conference of this scale together. We're very, very lucky and excited to have our wonderful guest, Erik St. Martin, who is a GopherCon organizer, as well as being co-author of Go In Action. Welcome, Erik. Lovely to have you. + +**Erik St. Martin:** Thanks. I'm glad to be here. + +**Angelica Hill:** How are you doing? + +**Erik St. Martin:** I'm doing pretty good. Yeah, I'm actually happy to be on. I miss hanging out with all of you on air. + +**Angelica Hill:** I'm so excited. This is I think the first time I've hung out with you on air, so I'm excited. Oh, and then Johnny... He's my wonderful co-host... He's "Oh, fine. I'll leave you two to it. I don't want to interrupt this beautiful moment." \[laughs\] And he's back! Awesome. Great. Welcome back, Johnny. I was sort of upset, I was like, "Oh, I'm sorry. I'm sharing this beautiful moment with you as well, Johnny." + +**Erik St. Martin:** Johnny's "Oh, it's Erik. I'm outta here..." \[laughter\] + +**Johnny Boursiquot:** I'm outta here... + +**Angelica Hill:** Well, welcome. I'm excited. I haven't co-hosted a show with you in a while. How are you doing? + +**Johnny Boursiquot:** It's been a while. I'm doing great, and I'm actually looking forward to this episode as well, because I'm one of the GopherCon OGs, as you'd say. I was at the very first one, and I think I've been in most of them, maybe for one, but we'll pretend that's not true. + +**Angelica Hill:** Well, lovely to have you both as a co-host on the show, but also as a co-chair of this conference, as well as being on the Organizing Committee. So we're going to start right at the very basics for all our lovely newbie gophers... What is GopherCon, and how did it first become a thing? + +**Erik St. Martin:** Yeah, so it's interesting... We've talked about this before on (I think) Go Time and Changelog as well... It came around as a really interesting story. At the time, Brian Ketelsen and I worked together at a company; this is 2011 or so... And we were both using Go, and we were attending lots of conferences... And so early 2013, we kept talking about the fact that like "I really wish there was a Go conference. I really, really want to go to a conference and meet other people who develop Go." We both live in Tampa, really big in the tech industry for healthcare, which is not, a bleeding edge technology adopter... So it was basically just Brian and I covering all of Tampa at that time with Go. It was still very, very early. And so mid-2013 rolls around, like April-May or so, and some folks in Denver, from the Denver Go meetup group were talking about "I really wish there was a Go conference." And Brian and I were "I know, this is what we've been saying." And so it was basically "Why don't you guys do it?" And we're "Well, how hard could it be? We should. We should totally do it." It turns out after the fact - really hard. It's actually a lot that goes into it. We knew it'd be a ton of work, we totally underestimated just how much work it would take. But yeah, so that's sort of how things came about. + +It was really interesting, because we were hoping to just get, 100-200 people together. At the time, we didn't know how many gophers there were; the language was still super-new... And so we planned to hopefully get like two or three hundred people together, and then it just sort of exploded from there. We had to keep calling the venue back to get more room. We maxed out at 750 people the first year... So that was really crazy for us, because - I mean, anybody in the conference circuit kind of knows, like... I mean, most conferences take years to even hit those numbers. So we were just really, really in shock. + +\[08:15\] It was really a humbling experience though too, like the number of sponsors who came on... And same thing, they just wanted a conference for this technology too, and so many of them weren't even worried about ROI at the time. They're just "The ROI is a Go conference starting, right? We need to hire people, we need these events to exist." And so that was really awesome. + +Then like all of these people that we really admired in the Go community were there helping us stuff bags, swag bags... It turns out stuffing 750 swag bags takes a long time. Who knew...? So yeah, it was just really interesting to have -- I'm trying to even remember... Steve Francia, Bill Kennedy, Damian Gryski, Dave Cheney... There were so many. I'm gonna leave people out here, which I always hate doing those things... But trying to remember all of the people who were like back there - we'd sit here forever. But it was just kind of crazy, all of these people who were like "No, no, you should be giving talks. Why are you helping the stuff bags?" But yeah, I think everybody was just really excited to just see this thing that happen, and anything they could do to make it an event that would be remembered, they were willing to do. + +And the same is true about MC-ing, right? At the time, Brian and I were running around just putting out fires, because there's so much stuff that happens behind the scenes that people don't even know, and like things that you need to manage while you're on site... And so we'd be like putting out a fire and then running up on stage to introduce the next speaker. And Kelsey Hightower came over and was like "It looks like you guys are kind of a little busy here... Do you want me to take over introducing the speakers for you, so that you guys can handle everything?" And so he did, and that's just like another example of people in the community just trying to pitch in to make it something. + +I think that all of that probably sparked the atmosphere that exists there today, where we care more about the community than anything else, and everybody just sort of contributing to it being a great time. + +**Angelica Hill:** So how has it evolved? The image I'm getting in my brain of you kind of scrapping this together, wonderful gophers helping in any way they can, to like the very put-together, huge conference that it is today... I would love to hear what are the big evolutions, what are the big changes that you've seen? + +**Johnny Boursiquot:** So Erik, before you answer that, as an attendee to the first one, I can tell you what my experience was... So I'm sitting probably like in the first or second row, and basically, I was there with my friend, Mark Bates, who's also a prominent member of the community... And we're sitting in like the first or second row, basically in this ballroom filled with 750 people... This was a single-track conference at the time; a massive number of people... Like, can you imagine as a speaker, right? For a brand -- effectively a new technology, right? Go had been around for a while, but it hadn't really hit its stride. But you're about to give a talk to a roomful of 750 people. So I can only imagine what the first speaker felt... But as an audience member, the whole thing was so smooth, with Erik running around, and Brian running around... I didn't know there was so many fires behind the scenes being put out. For me, it was just basically things -- the schedule went according to plan, the speakers were all on point, all the talks were great... To me, this was just a well-organized conference. Now, how do you beat that? Erik, how do you beat that? + +**Erik St. Martin:** \[12:00\] Yeah, so it's kind of funny... There's always like the magic that happens behind the scenes. You're right, it was a very, very early technology then. April 2014 is when the very first GopherCon was held, and so we're talking pre-Docker, pre-Kubernetes, like all of these big technologies... Most of the HashiCorp products hadn't been released yet... Speaking of which - actually, Mitchell Hashimoto bought the first two tickets to the very first GopherCon. + +**Johnny Boursiquot:** Nice... + +**Erik St. Martin:** That's kind of a funny story there. But yeah, so all of these big technologies that we look at that are written in Go, most of these hadn't really taken off yet, or hadn't even been released. So it was still very, very early. But as an example of things that happen behind the scenes that people are unaware of... So we talked about the Kelsey thing... So I had this idea the first year - it was like, okay, well, most people will stay for the talks, and many people will leave the next day. Flights all over the world will be taking off at various times... Let's book the space for the next day too, and people can just hang out and hack on stuff, right? This is the thing we all know now as community day, right? That day, it was just called Hack Day. And the whole idea was just hang out, hack on stuff together, before leaving for your flight the next day, right? + +And so we get down there in the morning, and the escalator just keeps coming. More people, more people. And there's like one of those "Oh, crap" moments. So many people had chosen to stay for the next day, just as part of the conference... And you look at each other and you're like "We're going to have to feed these people." And so we're just running back with the catering crew, like "What do we have that's for like events, in the next couple of days, that you can like reorder, that we can--" Yeah... And so there's things like that, that happen, and some of them are amusing after the fact, some of them are stressful. + +But yeah, I think that with the first year there was a lot that wasn't thought about as far as the viability of it as, for lack of a better term, business. Because though it's not a for-profit company or whatever, bills have to be paid, right? And so you can't just sort of go in and be like "Oh, yeah..." And one of the big things we did was we dramatically underestimated just how expensive running events are. And the best way I can compare this for people is anybody who's ever gotten married, and you see -- you're like "How much per person for a plate of food...?" It's the same thing for events, and sometimes worse, where Wi-Fi for three days is $40,000 or something, right? The price tags on these things are just astronomical. + +And so there was a lot of evolution that had to go in there, with budgets, and how are these things going to be paid, and how do we account for the prices of these things going up over time, how do we account for things like attrition... Most people don't realize this, but when you book these hotels to account for everybody staying there, you're committing to a number of hotel rooms being sold. And if people don't book those hotel rooms, you get hit with a giant bill. And most people don't realize that. And sometimes you're going to be wrong, and you will get hit with a bill, and so you have to kind of account for having budgets for these things. There's a lot that goes into just trying to figure out how many rooms are you going to need which nights of the week... Because if you guess too high, you get hit with an attrition bill. If you get too low, your attendees don't have a place to stay. And so there's like a lot of this sort of complexities there. + +\[15:56\] So there's that side of things, thinking about those... But then the two other ways I'd say that things kind of evolve was one, the vision of like a high production value, right? I always love the idea of people feeling like they've made it. People who speak to thousands of people, you want them to kind of feel really excited about stepping up on stage, like "I got here!" And so the idea of that was really fascinating for me; I really wanted to make sure that that happened, and just a means of the production quality for videos coming out for people to have that type of stuff... + +The other was I took off a month of work before that event, for all of the stuff that had to just be done... And there was so much work that was being done at night after work, and things like that. And so realizing that long-term, there was no way that we could manage that; not while having day jobs. And so this is where many of you who have been in the GopherCon circle, this is where Heather Sullivan and Convention Designs comes in. + +The first year we had worked with them as a local contact; they were referred to us by the hotel, our planner at the hotel, for the ancillary events. So we wanted to do kind of like an offset party, and like a dinner for the speakers, and things like that... So that's how we first came in contact with their company. + +And going into 2015 - that's where that relationship started to build even more, and they came in and help with a lot more of the logistics side of things, negotiating hotel contracts, and things like that. Managing travel, and working with the convention center, and all of that stuff, figuring out what we needed for insurance, and paramedics, and all of these things you don't really think about. Since then, every year, that relationship has built more and more, where they've taken on more and more of some of the backend logistics side of things so that we're able to focus on more of the content and the bigger picture... And there's still a lot that I do behind the scenes too, but all of that is -- like, going through event banquet orders and hotel contracts with a fine tooth comb... Not exactly exciting for me. So I'm really, really glad to have that relationship where Heather and I are very yin and yang when it comes to a lot of these things, like complement each other... So that was probably a big one, too. + +So this really -- an event, once you get to this size, is a year-round effort, and I have a full-time job. There is no way that I'd be able to work all of this stuff full-time. I'm really grateful for that relationship, and most of you have met her over the years, and her amazing team that helps us with so many aspects of the conference. + +**Break:** \[19:08\] + +**Angelica Hill:** So we've heard a little bit about kind of the logistics, putting together an event... But of course, GopherCon would be nothing without its participants and its amazing speakers... So we'd love to hear a little bit about how do we get these amazing, wonderful people on stage? I know, Johnny, you've been a speaker, you've been part of the talk review committee, and now as a chair... I'd love to hear a little bit about what is that process? What is that process for the speaker, as well as kind of the committee, and getting to that final lineup? + +**Johnny Boursiquot:** So there's two sides to this... And having been a speaker, and now being part of basically the people who get to decide who the speakers are going to be, these are very different things... So I'll tell you from a speaker's standpoint - even submitting a talk right is a terrifying process. \[laughs\] Despite all the advice that you read as a potential speaker, and you watch YouTube videos on how to go about it, even the GopherCon Committee puts out blog posts on "This is what we're looking for" and things like that, it's still a terrifying process. And this is something we definitely have to acknowledge. You always have basically this self-doubt that sort of travels with you, all the way from the time you hit submit for that proposal, all the way to the point where you deliver the talk and you're walking off stage and you're still thinking "Did I do a good job?" + +\[22:05\] So it's a very scary proposition for a lot of people, and basically, this is me giving a shout-out to you - if you've ever submitted a proposal to GopherCon, even if you didn't get accepted, kudos to you, because we know it can be a very stressful process. And for those who have been accepted and have given talks over the years, big shout-out to you as well, because we know the preparation process, rehearsing over and over again, and making sure you crossed your T's, you dotted your I's, you're going on stage, whether you're doing a big sort of general audience talk, which is very scary, or you're doing a tutorial in a smaller room - we know it's very challenging... Unless you do -- even when you've done it a few times, it's still scary, it's still sort of inducing of sort of that fight or flight sort of emotion... + +**Erik St. Martin:** And let's be clear here, too - the tutorial sessions and stuff like that are more about being able to go more in depth, having more time; it's much more specific, so you can drill in. But those rooms aren't small either, right? Pre-pandemic, you're talking 450-500 people. That's not a small talk... + +**Johnny Boursiquot:** That's the size of another conference. \[laughs\] Exactly. But the thing is, we have this process... So moving on now to the sort of selection process. It's always a combination of what is something that -- basically, the value threshold was always high for us. It's "Is somebody going to do this talk just--" Sometimes we had proposals for talks that the topic itself is an interesting one, but there might not be enough there, enough meat on the bone, so to speak... It could be based on the proposer's comfort level with the talk, which sometimes comes through in the description and how they propose things... It could be that the material is interesting, but there's not enough interest for it... There's a lot of things that goes into the selection of the talks and the topics that we focus on. + +For example, we've heard so many talks on gRPC, that this year in our selection process we're like "Ahh..." Yeah, we have a couple more gRPC submissions, but there's nothing in there that's sort of really new; there's not going to be new ground covered, so we might sort of basically deprioritize those in order to give other people a chance to introduce something new, right? + +It's always about what is the value that we want the community to derive out of this conference when they walk away from it, and we want that to be as high as possible. All the machinations that goes on behind the scenes, between the coordination, just getting all of us in the same room at the same time, scheduling-wise, just to get the chairs together to pick the darn talks is challenging... + +**Erik St. Martin:** Yeah, there was five, six meetings... This was after the shortlist was created... + +**Johnny Boursiquot:** After the shortlist, yeah. + +**Erik St. Martin:** To Johnny's point, too - to everybody who's ever submitted a talk... Like, we know emotionally there's a hit there, when your talk doesn't get accepted. But in a normal year - we're going to talk pre pandemic, because everything after is a mess... But pre-pandemic, we would get 300+ submissions, for 25 slots. Those are hard numbers. A lot of great talks aren't going to be accepted... And additionally, the process is intentionally set up so that none of us can favor certain talk types, and stuff like that. So there's a whole committee of people who review these things blindly, so we don't know who you are, or your expertise, and things like that, when they're reviewed. So two things come of that - so we rotate out the committee, and so that fresh perspectives, where it's not like a bunch of us who have been here ten years being like "Ah, do we really want an Intro to Go talk?" That wouldn't be fair. And so as we're swapping that out, that means the perspectives kind of change, and we get a consensus across the group, what topics are interesting or not. + +\[26:03\] But I think the key thing here that Johnny pointed out is that because of the way this blind process works, we don't know anything about you. And so if you're super-vague about the talk that you're giving, why are you the person to give that perspective, right? And so if we can't look at your outline and abstract and kind of be like "Oh, this person really knows what they're talking about, they know what the takeaways should be..." We don't know if you're a good public speaker, we don't know how well the topic... If you're like "I'm going to talk about the compiler." What specifically about the compiler? That's complex. How are you going to fit that into a 25-minute talk? So you have to give enough detail that somebody who doesn't know who you are is able to look at that proposal and be like "This person's clearly thought about how long it's going to take to deliver this information, that there's a clear path to delivering it, they know what their takeaways are going to be", things like that. + +And topics that have been covered in the past aren't necessarily bad if you also are talking about like how you're building off of the talks that have been given in the past, or how your perspectives on that technology have changed over time, and things like that, where there's some new information, and it's not effectively the same talk being given by a new person every year. + +And sometimes there is room for some of those topics that are just going to be regular topics that every new generation of gophers coming to the conference is going to need to see or hear. Then there's some of the technologies that are -- I don't want to call them fads... They're just - they're trends, right? Like, this year, WebAssembly is just really hot, and there's lots of proposals about WebAssembly, or something, right? I don't want to call it a fad, because a fad makes it sound like it dies off... But people get really excited about one technology in this like one or two-year period, and so there's usually like an influx of talks about that topic. + +**Johnny Boursiquot:** This year is GopherCon is all blockchain talks, folks. I'm kidding. I'm so kidding. \[laughs\] + +**Erik St. Martin:** Web 3. + +**Johnny Boursiquot:** It's Web 3. \[laughs\] No, no, no... We won't do that to you. + +**Angelica Hill:** So just so there's absolute clarity, the process is that we have this group of enthusiastic gophers, ideally at different levels, different experiences, that come together to form a kind of paper review committee. Is it that the chairs reach out to these people? If there are wonderful Go Time listeners who are like "Oh, I'd love to volunteer. I'd love to be part of this process", how would they go about doing that? + +**Erik St. Martin:** Yes, there's obviously the GopherCon channel that's inside Slack. Myself, Heather, we all hang out there... So I don't always get to read all of the scrollback, but if you ping myself or Heather Sullivan who's in there, or Johnny, or Angelica, or anybody else who works with us in the CFP committee and things like that, we can definitely do that. + +Obviously, there's a limited number of slots there... Often we'll reach out to people, especially like newer speakers and things like that; we'll do that. Because ideally, you want people who kind of -- there's two needs there, right? So one is fresh perspectives on what topics are important and exciting to the community, and you want to make sure you kind of have your finger on the overall pulse of that. It's easy for us to kind of get in our own little world and be excited about things like Web 3, and NF Ts... \[laughter\] But the other is that you also want people who understand what is a good talk, right? Sometimes with these -- if there's enough information there, we will reach back out to people who have submitted talks, too. We're like "Okay, I think I have the gist of what they mean, and this could be a really engaging talk. Let's follow up with them and get them to give us just a little more detail, or to clarify some things." + +\[30:02\] It's really hard to do at scale, giving everybody feedback... But if you're like on the 10-yard line, and we're like "I think this is a talk that could be super--" Especially if you've got a topic that's not done very often, that's really needed. I'll pick on a topic area that I'm super-passionate about, which is like security, right? There's not a lot of talks about security, but definitely needed. So if you submitted a talk in that category, that was so close, we might actually reach out and be like "Yeah, could you clarify a couple of these things?" and stuff like that. I think I kind of went off the beaten path here from the original topic we were talking about... + +**Angelica Hill:** The beauty of Go Time. + +**Johnny Boursiquot:** Well, I can add some flavor to that, too... So one, reaching out to us - if you're interested in sort of helping out at any level here... So you don't have to be a speaker at the conference to actually help out the conference, right? So you can be a part of the review committee... And again, we're always looking for a fresh perspective to help us sort of get our shortlist, get the talks that we should be paying close attention to. So you can always reach out to us and say, "Hey, I want to take part in the next time you do this", and we'll definitely reach out. + +And sometimes we reach out to people, because we notice you. One of the things that people out there may not realize is that as active members of the community, we ourselves are always keeping an eye out for new blood, for fresh blood. So at some point, Erik's gonna retire from this; at some point, I will retire from this. And so will Brian, and so will all the people that you are familiar with, or you might have heard, and you see names popping up over and over again - these people will go away at some point, and we need to sort of be grooming, so to speak, the next generation. Who's going to take over GopherCon after Erik is done? Who's gonna be passionate enough to be able to step in his shoes and carry this forward, right? Because I'd hate to see GopherCon go away just because Erik decides to take a break. Right? + +**Erik St. Martin:** What are you trying to say here, Johnny? Are you kicking me out? + +**Johnny Boursiquot:** \[laughs\] + +**Angelica Hill:** These are the five reasons that I feel that I would qualify... \[laughter\] + +**Erik St. Martin:** It's a mutiny... I knew this was a setup! \[laughter\] + +**Angelica Hill:** This is actually your exit interview... + +**Johnny Boursiquot:** This is your exit interview. \[laughs\] We keep an eye out for people that are active members of the community, and we notice when somebody is, for lack of a better term, a nice member, a nice person to interact with, and sort of talk to... + +**Angelica Hill:** I think that Johnny cut out... + +**Erik St. Martin:** Yeah. + +**Angelica Hill:** I think the moral of the story is that we've got our eyes out. If you're doing good work, if you're interested, if you're active, that doesn't go unnoticed. A big part of the work of the chairs and the work of those who are organizing GopherCon is to be in the Slack channels, to be going to the smaller meetups, to be keeping an eye out for interesting topics, people, and then engaging them. And I'll take advantage of the fact that Johnny is frozen and therefore can't hear this... Also, these are people who you can reach out to if you're even dabbling at the idea of submitting to GopherCon. + +Personal story - when I was going into the GO community and I was like "Oh my gosh, GopherCon - this is so big. This is so cool. I aspire one day to speak at this amazing conference" and was talking to Johnny, and he was like "Angelica, just do it. Just put a proposal together and submit. If you don't get accepted, try again." And these people are so giving with their time, are so willing to help... + +There is also a more formalized program around mentorship, and helping coach the speakers to be ready for stage... So even if you may have apprehension around like "Am I going to be okay at speaking? Am I going to be alright?" there are programs around helping you get on stage and be set up to succeed. + +Johnny was the reason why I ended up speaking at GopherCon, and it's a huge part of that kind of mentorship programming... And serendipitously, Johnny seems to now be back. So I'm gonna pick on you, Johnny, to talk a little bit about -- you missed my whole spiel about how you were just a really bad mentor to me, and you never encouraged me to do anything... + +**Johnny Boursiquot:** \[34:12\] Oh, wow... \[laughs\] + +**Angelica Hill:** So this is your chance to defend your reputation. + +**Johnny Boursiquot:** Redeem myself. + +**Angelica Hill:** Yeah. + +**Erik St. Martin:** The beatings will commence until morale improves. But I think also -- like, we talked about how we got here, and it's interesting... When this conference was started, we always believed that people who knew how to run tech events would come build the big Go conference. But I think that whole caring about people and the community-first is the huge thing that probably made GopherCon what it is... Because from conference one, there were many of us, from organizing crew, to other speakers and stuff like that, that were in one of the hotel suites until all hours of the night, while some of the speakers did dry runs of their talks, and things like that. And over the years, it's become a much more formal thing, but that's always been prevalent. + +And I know we're talking a lot about how do I get involved on the content side of things... So volunteering - if you've got a lot of public speaking experience and you want to mentor other people who are new to public speaking, that's a great way. But also, you can carve your own path. There's so much at the conference that have come from other people coming to us. Andy Walker came to us one year with the idea of like "How do we teach people how to have a good experience here, how to make the best out of a conference?" and things like that. That's when the whole Gopher Guides program was started. And he's ran with that, and it's been amazing, especially considering so many attendees are first-time attendees... Some of them have never been to another conference before. And so all it really takes is just being passionate about the community and bettering the experience for other people. And if you've got ideas, please come to us, because that's what matters more than anything to us. + +**Break:** \[36:17\] + +**Angelica Hill:** We've talked a little bit about kind of how to get involved, the speakers, moral of the story being if you have even an inkling of a thought of a dream that you would like to be involved in or talk at GopherCon, just do it. But I want to hear a little bit about this year, because obviously, we're coming off the back of the pandemic, a period of virtual hybrid conferences... This is going to be kind of the first in-person GopherCon in a hot second. I myself am extremely, extremely excited, given that I did my virtual keynote, next year was like reviewing papers, now I'm co-chairing... And this is going to be my first in-person GopherCon, ever. + +**Johnny Boursiquot:** Wow...! + +**Angelica Hill:** So I am extremely excited. I will be a first-time in-person GopherCon attendee. + +**Johnny Boursiquot:** That is amazing. I hadn't realized that. + +**Erik St. Martin:** Yeah, me neither. + +**Angelica Hill:** Get ready, I'm gonna whip out the gopher onesie if I can find one on Amazon... \[laughter\] I'm ready. But tell us. Tell us a little bit about GopherCon. I know we had a bit of change; I want to hear about like Chicago. I'm certainly very excited about Chicago. Never been. The only thing I've been told is try the Deep Dish pizza... But Erik, can you -- + +**Erik St. Martin:** That's a lot of bread though... + +**Angelica Hill:** I love bread. \[laughs\] + +**Erik St. Martin:** This isn't my unpopular opinion, but really? Why don't we need that much bread? + +**Johnny Boursiquot:** That's a lot of carbs. \[laughs\] + +**Erik St. Martin:** They're gonna like put me back on the plane when I hit Chicago. "What did you say about our pizza...?" \[laughter\] + +**Angelica Hill:** But tell us, this year, what are you excited about? What should we be looking forward to? + +**Erik St. Martin:** I'm just excited to get people back together. And I know there's a lot to kind of work around and navigate right now, because the pandemic is far from over... So I know there's a lot to navigate there, but two years of sort of not being connected - it almost feels the way it did back in 2013, when this started; you just so desperately want to get together in the same room with some other gophers. + +So yeah, that's what I'm most excited about, is just to have people there, even though it's not going to be the same scale as it was pre-pandemic... But there's also some opportunity to reminisce a little bit too, right? Like, that tight-knit group that came out of the first year, and everything, forging new relationships... So yeah, I'm pretty excited. I'm excited about some of the topics for talks, and things like that, too. I probably hit on a few of those, but then I'm like playing favorites here, which is like "How much should I do that?" \[laughs\] + +**Johnny Boursiquot:** \[42:19\] I'm looking forward to seeing people again, face-to-face, honestly. For me, going to a conference, and perhaps there's other people out there who share the sentiment - going to the conference, I honestly go less for the talks, and more for the interaction with people. I get to see people face-to-face that perhaps I only know in avatar form on Twitter, or something like that, right? I get to actually have conversations with people, go out with people, knock a few beers back with people, and just really get to know the people that I interact with, that are part of this community, that I enjoy being a part of so much. + +And one of the things that I think really I don't want us to take for granted is that -- basically, once I started sort of getting more involved in the community, and this basically started taking hold after the first one for me, because the community was so small at the time, I immediately saw the opportunity, based on the people that I had interacted with at the conference and beyond. I immediately saw an opportunity in the form of basically... Having been part of other communities before Go, I saw of some ways that some things kind of got derailed, and I saw some ways that interactions would be sort of sour, especially for beginners. There's that elite mentality that "Oh, RTFM" or "Don't ask, quote-unquote, stupid questions." Really, something I'm hoping technologies as a whole we're getting out of, although there's still quite a bit of that out there, but that whole sort of commonality around between other communities that I saw, I didn't see or sense any of that. Even though I was new to the community and new to the language, I didn't see or sense any of that negativity early on. So I'm like "Okay, maybe this is an opportunity to keep it that way, or to make it better, and kind of keep it moving in that direction." + +And I think the people that are really involved, the people you know now, that have been longtime gophers - I think there was this implicit sort of decision on the part of all of us that says, "Hey, you know what - let us all collectively make this community what we'd like it to actually be in the future. Let's try to keep it a community that is welcoming, and nice to beginners, friendly to beginners, people that are still learning." Because whether we like it or not, there are more new Go developers currently than there are old hats like myself, or Erik, or folks on the committee. There's way more new people. That has been the case for the last, I'd say five plus years, right? Every year, more and more new people come into the community. And we want that first interaction to be a positive one. + +We want you to come for the technology, but stay for the people. Because the people are really what's going to make the whole experience much better for you, whether it be a colleague, a co-worker or open source contributor that you happen to be working on something with; whatever that interaction is, we want it to be positive. And the Code of Conduct stuff that we have for the conference itself, the Go community Code of Conduct stuff in general, the interaction, the rules that sometimes not everybody likes for the Slack channel, Gopher Slack - we know that these things can sometimes be annoying in certain contexts. But overall, they are positive instruments for making sure that both the conference spaces that we're in, whether it be GopherCon in the US, or GopherCon Brazil, or GopherCon UK, or any other Go conference that GopherCon itself has inspired - we want all those spaces to be safe spaces for everybody, welcoming spaces for everybody, great learning environments for those who are brand new to the language and brand new to the community, and old hats as well. + +I think this is something that's very important to us as community leaders, and I think that's something that, you know, whether Erik wanted it to be or not, that is what we've become, and I hope you do like where we're going, Erik, as far as conferences go. + +**Erik St. Martin:** \[46:23\] Yeah. I mean, I like the community aspect... Prior to the pandemic, I had dreams about another conference I wanted to start too, for similar reasons. Conferences - you know, they're very community-focused, versus ones that very much are sales machines... And they are what they are, and that's not like a hit on them. I just... I'm passionate about being in the communities where it's just that - a bunch of people that are just trying to support each other and help lift each other up. + +So I like where we're going, but it's interesting - you said about the number of new people... I mean, all of the OG gophers, so to speak, if most of them walked into a new Go conference, the only people who are probably going to recognize them are other OG gophers, right? + +**Johnny Boursiquot:** Right. Exactly. \[laughs\] + +**Erik St. Martin:** We've hit that point of sort of like critical mass. We all looked up to and admired, and have since gone on to do other things. People wouldn't even recognize their name, right? It's not good or bad. It's just the evolution of a technology once you hit a point of like critical mass, where it just explodes. Most people don't care about those details anymore. Right? Like, the number of people - as sad as it sounds - who cares about who the original people who wrote the Go programming language is, it's mostly OG people. Most of the people don't care, because they don't need to know that to get stuff done. They just want a language that works, and does what they need it to do, and it does that, in spades. So yeah... + +**Angelica Hill:** So I want to kind of press on the community aspect a little bit harder; not to be a skeptic, but there are many communities that say they're very inclusive, they say they're very diverse, they say they're very welcome to all, but in practice they turn out not to be. So I'd love a little bit to hear about the actual, in-practice, how do you do that? How do you foster an inclusive, welcoming community? What are almost some like tips and things that you can do, some things that you are cognizant about? I would also like to press that one step further, and directly ask you, Erik - I know that the move to Chicago was rooted in serving our community, the Go community, so I'd love to hear a little bit about that, because I think that is an example of where these concepts of like we care about our community actually go into practice. + +**Erik St. Martin:** Yeah, so that's a hard one, right? Because I don't claim to be an expert here. And we're all doing -- I can't say all, but most of us are doing our best to do what's right for other people, and to advocate for them, and things like that. And we're gonna screw up, right? Like, it happens. We're human, and we're going to make mistakes. I think the big thing there is to be accountable, and be able to admit fault and figure out those things. So I think a lot of it comes through example, if your people or other people look up to and they want to mimic your behavior. And if you're a jerk, then they're a jerk, too. And these things are important. + +I think the other part is about -- we like to talk about being an ally, but I think for most people, it's very fair weather, right? Like, it's easy to be an ally when it requires words, right? It's much harder to do those things when there's real sacrifice needed, or you're exposing yourself to a ton of risk. And so the move to Chicago - that's one of those scenarios. I don't know what will come of that. There's a lot of complexities in all of that, a lot of which I can't get into currently... But those things come secondary to doing what's right. + +\[50:22\] I think that when we look at the tech community and things like that, and even just people who mean a lot to me, that are very close to me, that are impacted by all of these changes - looking at Florida, it just doesn't align with our values, right? Like, we put up a code of conduct, and we talk about how we all need to be welcoming and accepting, and all of these things, and then it's like "Well, let's hold our event in this state that just keeps passing--" And I should clarify, the state I live in. Like, I have to live here. Moving out of state - speaking of hard things to do, right? A lot of complexities there. But, yeah, I mean, passing these laws that are very, very harmful to people in our community. + +Some people may argue that, "What does that really do?" Right? Ron DeSantis isn't going to be like "Oh, GopherCon's leaving. We should reevaluate these laws." But it may not matter to them, but it matters to us, all of us who are part of the community. And it's a slightly different use of the phrase, but the law of two feet, right? Like, we're speaking with our feet. We remove ourselves from the situations that we don't want to be in, and move towards places. And so I think Chicago aligns a lot more with our values. + +**Johnny Boursiquot:** Do you think we'll swing back over to San Diego at some point? That was my favorite spot, by the way... When we had it there. + +**Erik St. Martin:** Yeah, so one of the things -- like, when we talk about the complexities and logistics and stuff, what a lot of people don't realize is the number of years in advance. I think we're contracted up to like 2025, like, going into the pandemic... And so there's a lot of complications in all of that. So yeah, San Diego is definitely on the list. We love San Diego, too. + +And I should also clarify too, when we talk about sort of like acceptance and things like that too, and people doing their best, is there's a lot - and we don't have a lot of time here to go into the risks associated with running conferences and things like that... It's important to remember that while we made this decision, people's situations are completely different, and we have to understand there's a lot that goes into these things, in the contracts, and how long people are contracted out, and things like that. It's not an easy decision to make... And with any regard to conferences and these types of things. + +So we have to have a critical eye on speaking with our feet, and not attending conferences that are just putting people with sexual harassment issues and things like that on stage, and stuff like that. We have to be very careful about -- we don't want to support conferences like that. But on the flip side, we have to understand that some of these things come with real, real consequences. You're signed out several years later, you've got minimum commitments and things like that, and great, you decide to go somewhere else. And then you get sued for everything, right? + +And so we have to kind of give people a little space and support too when they don't do these things, and try to understand that - are these the type of people who would do something like... I don't know. Like I said, these things are complicated, and I don't claim to be an expert. We made the decision because like this conference was founded on the idea of the community being the most important thing, and so whatever comes from all of this, it's worth it to me to do the right thing. But also, I recognize that that's very, very hard. + +\[54:11\] And these things take a long time, right? So even if somebody wants to do something, you don't just like make this decision on a whim. You have people reviewing contracts, and lawyers, and all of this stuff. They're not quick decisions. Even if you were already doing it, you can't make an announcement until you've found a new home. So there's a lot of these things we're sometimes giving people the time to sort this stuff out. + +Don't get me wrong, publicly asking people like "Hey, what are you going to do about this? We don't agree with this scenario." Like, absolutely. Call people on that stuff. Let them know that that stuff's important to you. But also, you've gotta understand that these things -- if we have to sign a contract five years out, how long do you think it takes us to actually -- changing the wheels on a moving bus is a difficult thing. + +**Angelica Hill:** Well, regrettably, we are coming to near the end of our time. I feel like there are so many more questions I want answers... But I have one more question to ask, which is - for anyone who's thinking about like "Should I, shouldn't I go to GopherCon this year?", why should they go? + +**Erik St. Martin:** Because it's awesome. + +**Angelica Hill:** Okay. \[laughs\] + +**Johnny Boursiquot:** Clearly. + +**Angelica Hill:** D'oh... + +**Erik St. Martin:** I think, to Johnny's point, the networking aspect of conferences is huge. And I know we have digital means of communication, but it's not quite the same. And the interesting thing that I always find about attending conferences - and GopherCon is a conference that does this for me, but lots of conferences do as well - I always come out super-invigorated, and just excited about something that I want to just dive into... And having those conversations... + +So I think, absolutely, if you can, please attend. If you can't, that's awesome, too; we totally understand, we're in the middle of some of the most challenging times for people, and everybody's situations are different... We recognize there's still risk in the pandemic, and we do our best to mitigate that stuff as best as possible... Also kind of recognize that budgets for companies to attend conferences and stuff like that, either through travel restrictions or just budget slashing and preparation for recessions and things like that aren't there... So obviously, if you're not in a position, we hope to see you in a future year... But we would love to get a bunch of people together this year. I know I'm super-excited about just getting together in-person with other gophers. + +**Johnny Boursiquot:** I just can't wait to hang out with my friends again, and then make new ones. I'm tired of the endless Zooms, and the Hangouts, and the virtual thing... And I will acknowledge that for some, the virtual aspect of a conference is actually a good thing. It's allowed people to basically virtually attend these events, who normally wouldn't be able to travel, because it's not sponsored by their employer, or because it's just a little too much, especially if you're coming from abroad, and things like that... So we do see the value of having the virtual aspect/component for that, and hopefully, we can keep doing the hybrid thing, where we stream the talks as they happen. Obviously, the recordings are always available, usually within a matter of 2, 3, 4 weeks after the conference itself. + +Ideally, we want to be able to support kind of both things for those who are able to make it on-site, and sort of get that full, immersive experience of actually physically hanging out with people. Hopefully, we can support both. I think that's sort of the best of both worlds kind of thing. But for me, really, it's the getting to hang out with people again... And I don't know, I enjoy the physical human contact. That's huge for me, and it's definitely what I'm looking forward to the most. + +**Angelica Hill:** \[57:56\] Awesome. Well, I'm very excited. I would very much encourage anyone, I would mimic what you both said - if you can't go this year, please look to GopherCon in the future as a conference that would welcome you, that is wonderful to attend. As I said, I'm very excited to see where my involvement in GopherCon goes, because I feel like my first keynote was like "Anyone can be a gopher." My next one will be "How GopherCon changed my life and taught me I was secretly technical." + +**Johnny Boursiquot:** Nice. + +**Angelica Hill:** So I would say it can truly change your life. And I mean that from the bottom of my heart, from anyone who remotely knows me. Awesome. So now we're on to the -- not the best part of this discussion, because I think the whole chat was great; but the part that I'm intrigued to hear, which is unpopular opinions. + +**Jingle:** \[58:45\] to \[59:03\] + +**Angelica Hill:** Erik, what is your unpopular opinion? + +**Erik St. Martin:** Is it the Brooklyn-style pizza is way better than Chicago-style? And also pineapple on pizza is good. Fight me. \[laughter\] + +**Angelica Hill:** Catch me outside. + +**Erik St. Martin:** So let's use one for the industry... I actually -- you know, as a manager and things like that, and doing interviews, and I've had discussions about people who've been interviewing for new jobs, and there's always that fun one, like "Why should we hire you?" sort of ordeal. Or "Why do you want to work here?" So I'm gonna go with my unpopular opinion is like I dislike those types of questions. It's kind of narcissistic, in a manner... To be succinct in it, "To make good money", or "To get a paycheck" is a perfectly acceptable answer to "Why do you want to work here", right? You don't need to have passion, and all... Those things are great, right? Like, for some of us who are invigorated by tech, and we want most of our life to be that, that's awesome. But it's perfectly okay if you're just somebody who just wants a good-paying job, so that you can enjoy the rest of your life outside of work. And I think we need to do away with all of the crazy, like "You need to write an essay about why our company is the only company you've ever wanted to work for in your entire life, and stroke our ego more." It's just -- yeah, it's gross. + +**Angelica Hill:** Do you have a response to that, Johnny? I see you nodding your head. + +**Johnny Boursiquot:** I understand the spirit in which the question is asked... But like Erik says, I don't think I like the sort of connotation of it. If somebody is asking that, they're really -- why don't they just ask sort of the underlying thing, which is "Why do you want to work here? What value is there going to be basically for us mutually to benefit from? What can we do for you, what can you do for us?" kind of thing. Because this is all it is at the end of the day, right? It's an exchange of my time for your dollars? Let's not pretend it's more than what it actually is. + +\[01:01:02.07\] If I work somewhere and I happen to like the company - great, that's even better. But to tell you the truth, perhaps that is my unpopular opinion - I don't have a problem working for a company that is disliked. If the money is THAT good, right? I don't know. Because at the end of the day, you may have a social justice agenda, or you may have some sort of agenda, and that's fine. It's good to have beliefs, and have values, and believe in things... But I have one too, which trumps yours. In my case feeding my family trumps all of your values, I don't care what they are. For me, that is number one. And if that means working for a company that perhaps you happen to dislike, I don't have a problem with that. You do, but I don't have a problem with that. + +And again, I know that's gonna be unpopular with some folks, but you kind of have to give people space to make the decisions that are right for them and their family; not everything I do or say you'll agree with, and not every decision I make you'll agree with, and that's okay, too. Give each other some space and some grace. + +**Erik St. Martin:** The ability to walk away from your job because you disagree with something that company does is a position of extreme privilege, that not everybody has. And more power to the people who are in a position to do so, and are willing to take that risk. And it's the same thing with like the moving of a conference, right? Awesome for people who can do that, but it doesn't necessarily make somebody evil because they're not in a position to do so, or are uncomfortable with the risks, right? Like, some of us are much more risk-averse than others, and stuff like that. And there's nothing inherently wrong with making decisions that are good for your family. + +It's the same thing when people quit a job, and they feel really terrible, like they owe the business their life. It's like, you should never feel guilty for doing what's best for you and your family, especially if it doesn't harm other people. Like, why even judge somebody for that? If it harms nobody, and it enriches that person's life, then it was a good choice. + +**Angelica Hill:** Moral of the story - your choices are yours. Other people's choices are their own. + +**Johnny Boursiquot:** Yes... + +**Angelica Hill:** Let us all be accountable to our own choices, because we don't know what's going on with other people. And I would hope just listening to this conversation gives you an insight into the wonderful human beings that are behind GopherCon, and how considerate and thoughtful they are about the world. It has been an absolute joy to talk to you both. I'm hoping that I get you both on with me very, very soon again. And thank you for jumping on on a Friday... + +**Erik St. Martin:** I'm always happy to come back. + +**Angelica Hill:** I will hold you accountable to that, Erik. I know you're very, very busy, and now you've said it. You're coming back from GopherCon very soon... + +**Erik St. Martin:** As long as nobody's like "Keep that man off the air", I'm happy to come back and discuss other things. Even if it's not GopherCon-related. + +**Angelica Hill:** Well, as long as Go Time lets me come, I'll get you on, too. And if not, we can just start another podcast. Don't tell Jerod. \[laughter\] + +Jerod Santo: I heard that... + +**Angelica Hill:** Thank you both. I could babble to you all day, and it gives me such happiness just to see your faces, and I cannot wait to see them at GopherCon in real life! + +**Erik St. Martin:** Me too. + +**Johnny Boursiquot:** Yay! + +**Angelica Hill:** Have a great rest of your Fridays, and a great weekend. diff --git a/Instrumentation for gophers_transcript.txt b/Instrumentation for gophers_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..6763c3d17ccdbcf96c1b859248ae6067d55de20d --- /dev/null +++ b/Instrumentation for gophers_transcript.txt @@ -0,0 +1,351 @@ +**Mat Ryer:** Hello, and welcome to Go Time. I'm Mat Ryer, and today we're talking about instrumentation, and instrumenting your Go code, specifically. I'm joined today by Johnny Boursiquot. Hello, Johnny. + +**Johnny Boursiquot:** Hello there. + +**Mat Ryer:** How's it going? It's been a while. + +**Johnny Boursiquot:** Yeah, it's been a while. Yeah, we haven't been in a show together for a minute. + +**Mat Ryer:** Yeah. + +**Johnny Boursiquot:** How are you, mate? + +**Mat Ryer:** Yeah, not for a while. Well, you don't just have to tolerate me on your own, Johnny. Don't worry. We have some guests today to dig into this subject. We've got Björn Rabenstein. He's here, he's an engineer at Grafana Labs and long-time Prometheus contributor. Hello, Björn. Welcome to Go Time. + +**Björn Rabenstein:** Hello. I'm glad that I'm finally on this podcast, of which I'm a great fan. + +**Mat Ryer:** Oh, well, we're great fans of yours as well. And we're also great fans of Bartek Płotka, who's also joining us. Hello, Bartek. + +**Bartlomiej Płotka:** Hello. Nice to be here. + +**Mat Ryer:** Bartek is a engineer at Red Hat, a maintainer of many open source projects in Go. You've probably seen his name around GitHub here and there. Yeah, maintainer of Prometheus and Thanos, and you're authoring Efficient Go. Is that right, Bartek? + +**Bartlomiej Płotka:** That's correct. Hopefully, published this year. + +**Mat Ryer:** Yeah, that's exciting. How is it writing a book? Is it what you thought it was going to be? + +**Bartlomiej Płotka:** \[04:06\] Definitely not. Definitely not. Lots of pros and cons. Lots of learnings. I think it's worth it, but maybe just once a lifetime. \[laughs\] + +**Johnny Boursiquot:** One book will do. Yeah. + +**Mat Ryer:** That's how you feel now. Maybe you'll get the buzz. + +**Bartlomiej Płotka:** Exactly. + +**Mat Ryer:** Yeah, you'll get a bug for it. Well, we'll see. Okay. So maybe we could just start at the beginning. For people not familiar, what is instrumentation? What's it useful for? Bartek, maybe you could kick us off. + +**Bartlomiej Płotka:** Right. So how I see the instrumentation as, essentially, it's about generating the signals that will tell us later on how your application is behaving when you don't have visible access to how it's running. So maybe you are putting that in production, in your cloud, maybe on your friend's machine, just spinnig up the process out of your Golang code that you created, and you have to have some kind of - well, those magic words, "monitoring" or "observability", so the ability to really derive the state of this application only from its observability or monitoring signals. So in order to generate those signals, we need instrumentation. So essentially, we need instruments that tells us what is happening remotely. This is as simple as that. + +Now, of course, this can go even more complex, because you can have either manual instrumentation, so you are directly adding a code statement to your Go application, to add a log line, to add a metric, to create a span, tracing span, or maybe there is a Golang run time logic that creates profiling. And probably we can discuss this in this episode. But there are also automatic instrumentation mechanisms, where you can deduce the application state essentially from, let's say, a closed box pattern, where you can essentially ask operating system, so essentially, a kernel. This is what is very popular nowadays, with this eBPF solutions that allows you to really understand what process is doing without any manual statement. So essentially, this is the categorization - manual instrumentation, automatic, all to derive some signals. + +**Mat Ryer:** Great. And so what's this useful for then? So we find out that we can find out things going on inside-- or what's like a specific example of something that you might want to instrument and then report on? + +**Bartlomiej Płotka:** So feel free, Björn, to chime in, but there are plenty of things. And I think first and foremost is something that we can read in a popular SRE book, right? So essentially, the Google team, who created this SRE movement, site reliability engineering - we can read that monitoring is like a key component in building any application in the cloud. And the reason for it is that you don't know if it's running if you don't observe some signals. So health monitoring is kind of the first thing you do. You want to ensure that it's actually doing any work if you are not looking on it, if you are not actively poking using its -- I don't know, HTTP requests, or looking into its work items. So health is the first thing you want to check out, probably. + +**Mat Ryer:** And so that's like a real endpoint running on the server that just returns some "okay." And so if you can reach that, then you know the server is at least up. + +**Bartlomiej Płotka:** Yeah, that's the basic solution. That is not the best, because if you're not looking, then it might be down in those periods when you are not poking it for 200 \[unintelligible 00:07:41.18\] + +**Mat Ryer:** Yeah. + +**Bartlomiej Płotka:** So there are many, many different methods of how you can process this information about healthiness. + +**Björn Rabenstein:** \[07:51\] I mean, what we got with the age of cloud or the age of distributed systems was also a notion of that just being up or down - it's not that easy anymore, right? Your server consists of many microservices, and every microservice has many instances... Some of them are always down, because once you have enough of them, they will never all be up. So you get into this whole area of -- I mean, we have to tolerate a bit of downness, and then you start to think that just probing for up or down for this binary result isn't enough anymore. This is where your start to want perhaps some metrics about your running tasks. And that's where the instrumentation also, I think, gets into the game, right? If you just run a probe to see if something is up, you arguably don't really need instrumentation; you just check out if your endpoint is up, right? + +I mean, this is what I call the founding myth of Prometheus. I have to talk about Prometheus, of course... I don't know if it's literally true. I was joining the Prometheus team very early in its history, but not from the beginning, right? So I also got this just from stories. So this is why it's all myth in the distant past... But, I mean, sometimes you get nice stories from that, right? And the mythological version is that the first idea for Prometheus was actually, "We need to instrument our code for metrics", and then the initial founder started to create an instrumentation library. And I like to believe it was the Go library they created first. It might have been the Java library, but let's assume it was the Go library, because it's a Go podcast, and then that might even be true. + +**Mat Ryer:** Would you assume that was the Java one, on Java Time? + +**Björn Rabenstein:** I don't know. I think I will never be on Java Time if there's anything like that. \[laughter\] We'll see. So this was the initial spark of inspiration. There was an instrumentation library where you could instrument for metrics and it would expose in a separate endpoint - I think Bartek mentioned that shortly - that you have a separate endpoint where you not just collect "Is it up?", but you can collect the data from the inside of your running binary. And then the next question was, "Okay, how do we scrape this data? And where do we collect it? How do we store it? How do we read it? How do we curate it?" And this is where then the Prometheus server got invented, essentially. But it all started with instrumentation... And maybe with the Go library, which still exists, and Bartek is the current maintainer of it, and I'm the past maintainer of it. + +**Mat Ryer:** Wow. + +**Björn Rabenstein:** It's one of the most used Go packages in the universe, I think. + +**Mat Ryer:** Wow, in the universe. Wow. I mean, that's assuming Go only exists on this planet. We can't say for sure that-- + +**Johnny Boursiquot:** Or just in this universe. Because there is a multiverse, from what the movies tell me. + +**Mat Ryer:** Yeah. Still not bad though if you're the best Go package in this particular universe, even if-- + +**Johnny Boursiquot:** In this particular-- yeah. + +**Mat Ryer:** Yeah, even if there's more stars in other universes. + +**Björn Rabenstein:** Is there Go on Mars? There must be Go on Mars, right? + +**Mat Ryer:** Yeah, there must be by now. Yeah. That's great. + +**Björn Rabenstein:** And then there might be a Client Golang -- a Prometheus Golang might be on Mars as well. Who knows? + +**Mat Ryer:** Probably. + +**Björn Rabenstein:** It's running in all the German -- like, on the rail stations, they use it to monitor those. + +**Mat Ryer:** Do they? + +**Björn Rabenstein:** So it's running everywhere around me. It's pretty cool. + +**Mat Ryer:** Well, if only you'd instrumented it and it was reporting that somewhere, so you could see all the places it was running, that'd be awesome. Actually, I would like that as a feature, please, because for -- yeah, all packages, we ought to have that, so we can see how it's been used. + +**Björn Rabenstein:** The popularity of this library is actually -- for me, it was a huge surprise. And that's also, I think, an important topic about instrumentation. Back then - we're talking about the year 2012 - it was very uncommon for normal developers to even think about instrumentation. I mean, if they really thought about it, they would realize even putting a printf statement for debugging into their code, or emitting a logline - that's already instrumentation. Instrumenting code for profiling is instrumentation. Luckily, we get this for free in Go. + +\[11:48\] So they kind of did instrumentation, but they would never think about instrumentation for monitoring. This was completely an ops concern, and developers would never think about ops concerns. And teaching them that they have to instrument their code for things like monitoring was a big deal, right? I didn't expect that would get traction so quickly. But of course, the DevOps movement, blah, blah, blah, all those things. So that might have helped. + +Also, developers might have pretty quickly realized that if you instrument your code for even more things, like metrics or tracing or profiles, that it even helps during development. That's also an important thing, that monitoring per se is everywhere in the stack, right? It's not just in the end when you add it as an afterthought, and you need it to run your system. + +**Johnny Boursiquot:** Different people have different needs, right? So when we talk about the operation -- sorry, if your organization is large enough to have a dedicated ops team, or the people who care about the VMs, and basically making sure that the CPU is not spiking, or there's enough memory, or whatever it is; managing the lower level set of concerns, so at the infrastructure operational level... And then as you go up the stack, if you will, then you have a different set of concerns, right? You have the developers that are instrumenting maybe to find out, "Okay, how--" Let's say you wanted to sort of send traces out, and you want to know, "Okay, in this long-running process, I'm going to have a sub-span, to tell me how long this particular part of the whole thing is taking", and things like that. So that gives you that sort of that debuggability, the understanding of what your system is doing on any given request kind of thing. + +But then I think there's an even higher level to that, where you instrument things and you can tell whoever asks that, "Okay, the service is up." I can hit the health endpoint or the readiness endpoint... Hopefully, it's not just a ping to get a TCP response, but maybe it's something you do and something meaningful that says, "Okay, the system is operational, it's actually ready to process requests successfully", kind of thing. + +But beyond that, you're going to have the - can you answer "Is this thing doing what the business wants out of it?" So you're not just saying, "Is this thing up?" If your thing is up, but maybe you're failing process, 1 out of your 100 credit card transactions, and you can't explain why, at scale, that's millions of dollars potentially, right? So there are some things that the business is going to care about, that instrumentation can help you answer. But you have all these different layers of observability. Different parties are going to be interested in different things, depending on who you ask. + +I saw people who use Nagios as a tool for instrumentation. Well, I'm not going to throw my developers into a Nagios dashboard and say, "Hey, go answer me some questions about business KPIs or something", right? But yeah, I'm going to use different tools. But I think the observability, really - I think it's an all-encompassing term now, which I think has gotten sort of diluted a little bit, ever since we transitioned from calling things monitoring to now using the more trendy term 'observability', and I think that's going to mean different things with different people. But instrumentation remains, I think, at the end of the day, whether it's automatic or manual - for the developer specifically, that remains the ultimate source of truth, beyond logging. + +**Bartlomiej Płotka:** Yeah. And I would like to kind of add something to your points, both Björn and Johnny, about instrumentation. I think it's very underestimated how much work it takes to really build a solid instrumentation library... Because the amount of work you put, Björn, in this Client Golang that you were surprised is popular - well, it's because it's so hard to produce one. And especially when I joined maintenanceship of this package or module, really, there's so much work in terms of making sure the code is efficient, because suddenly so many applications are just importing this package. Modules... So suddenly, the amount of dependencies really matters. The efficiency of the code matters. The API scope scrape of this really, really matters. And I was just today trying to -- well, what we did, let's be honest... We made a little bit of a mistake on a new \[unintelligible 00:16:03.06\] Golang. So a Go team member really helped us to essentially move to different runtime metrics for the Golang processor. + +\[16:15\] So in Client Golang you are able to essentially expose really interesting information about your Golang process, and around garbage collection routine, and memory, and essentially heap allocations, like how many tiny objects and big objects you are kind of allocating; very low-level information. So we kind of consumed that, we merged this contribution with that, clearly looking on maybe how many metrics you are exposing, and we suddenly expose maybe twice more metrics, and suddenly we are impacting many Golang services, because suddenly they are scraping twice more metrics. + +So the amount of thoughts that had to go through those instrumentation libraries is enormous. It's not only metrics, and it's even worse when you are talking about logging instrumentation and tracing instrumentation, because it's just even more data that you are passing through. You want to pay for this application compute power for doing normal work and not monitoring, right? And sometimes, if you'd implement it wrongly, if you use those libraries wrongly, if you make maybe wrong decisions, you can pay much more for your observability than for your real application. That's a risk here, right? + +**Mat Ryer:** Hm. So it really matters, doesn't it? It's amazing, actually, to think of -- I genuinely probably naively thought that that library was quite simple, that it would just have had some helpers that would collect buffering maybe, and then sending off batches, or making them available or something. But yeah, now when you mention -- like, often I will tell people, "Don't worry too much about optimizing yet. Get something built that works first, and you can optimize it later when it becomes a problem, if it ever does." This is one of those times where it really matters that this isn't an expensive thing, because - yeah, people are calling these methods all the time, aren't they? They use that client everywhere. It's almost in every code path. + +**Björn Rabenstein:** You definitely don't want to think, you don't want to consider if you use that or not; you just want to use it and it doesn't have a significant penalty. Similar to when I realize in a Go binary profiling is kind of always on, and you just have this endpoint if you just anonymously import this package, and stuff like that. It was weird, "Isn't that expensive? Should I switch that off in my real production binary?" But no, it's all designed to be, I don't know, a fraction of a percent of your total resources. So you just leave it on, you don't ask questions. And when you actually get in trouble, it's there, and it doesn't matter if it's developer trouble or ops trouble; it's just there, and that's so cool and so good to have. + +**Bartlomiej Płotka:** Right. But there are configuration variables for those Golang profiling that increases the frequency of-- increasing the sampling or frequency of CPU profiling that can trash your application completely; or you can actually reduce it. So there's lots of things you can control here. + +**Björn Rabenstein:** And in what you just described, Bartek, with those new Go runtime metrics, now we also need to consider if we make this user configurable, because it's so many metrics that if you don't need them, it might be too expensive. So, of course, there are traders offs, but the ideal state in some utopia is you just have everything on and it costs you nothing. And we try to get as close as possible to that. + +**Break**: \[19:38\] + +**Mat Ryer:** Bartek, you mentioned earlier that you made a mistake in the client. What was that specifically? + +**Bartlomiej Płotka:** Right. Let's talk about mistakes, yes. + +**Mat Ryer:** If you don't mind... + +**Bartlomiej Płotka:** Of course. Of course. + +**Mat Ryer:** That makes me feel better. + +**Bartlomiej Płotka:** Yeah, so I think that the mistake is really around -- we didn't have time to review exactly properly everything, and that change was pretty big. And as maintainers of Client library, I think we just get that from Bjorn's hands, so we are kind of pretty new. And essentially, usually, it's really about memory runtime metrics that Golang allows you to achieve. So usually what you could do - you could create this runtime.memstats structure, and you can just program runtime.readmemstats and provide the structure, and that will fill you the structure with statistics. And this is how we used to kind of get those metrics, and give you kind of a very easy one-line statement that you provide in the beginning of your application, that continuously exposes information, up to date, so Prometheus can scrape it and have up-to-date heap information; how many allocations you did on the heap, what's your GC latency, and so on. + +Now, what happened in Go 1.17 (I think) was that runtime/metrics package was introduced, that kind of gave even more information about internal Go runtime, and we wanted to switch, essentially, to this new format... So totally go away from this memstats logic. + +However, what's happening with especially metrics is that we all care about cardinality. We make sure the metric is stable, because this is what allows us to aggregate over time, and also over different series, and kind of compare to each other. This is the key concept behind metrics. And also, the cost of the metrics system really depends on the uniqueness of those series. So yeah, we just introduce a new unique series, so new names... And this is kind of what we could avoid, what we are trying to fix right now, is to add essentially some optionality, some configuration pieces. Especially because this library is so used, so popular, we need to make sure it by default doesn't increase your metric count twice. + +So this is generally kind of what we try to achieve. So we have a community who wants to have this new runtime, amazing, granular metrics... So we kind of approve it and merge it, and then suddenly, the rest of the world was like, "Oh, I need to pay twice more. I don't understand really." So this is really the trade-offs we have to make for such a popular module. It's much more stressful and much more work than if you could just break the API on every release. + +**Mat Ryer:** \[24:20\] Yeah. That is a big responsibility. Not only because it took off and became popular, but also the nature of it, the fact that it's found its way everywhere. I mean, it really is probably everywhere now. So it really does matter the decisions you take. How do you balance that? Is it a tough balance? Is there a disagreement about what can get in? Are there some people that are really paying attention to that cost? + +**Bartlomiej Płotka:** Definitely, yeah. After merging this feature, we got like 12, let's say, issues, and very friendly ones. So we were very, very -- I mean, it was amazing community contributions to just report some problems. So I think that's the pros and cons of such a popular module, is that you're going to get early feedback, but also you have to make sure you're not making mistakes a lot. + +But generally speaking - yeah, I think there are basic rules when you are writing any Golang code, especially that is consumed by others, like library modules and packages that saves your life. For example, any new feature you should add under a new API, under new methods. You don't change the existing code. Whenever we want to change something, we are having this in comments // depricated kind of special word that is actually handled by the IDEs, like Golang and others, that automatically points you that this code is -- well, it's for compatibility reasons there, but there is a better solution somewhere else. So there are certain rules that you can go with. But of course, at some point, you need to consider V2, or V3, or V100. That's a big chunk of work unfortunately, nowadays, with Go. + +**Mat Ryer:** Yeah, yeah. I like that you've got nice issues though, at least; like, someone's like, "Thank you so much. I love the recent PR, but--" + +**Bartlomiej Płotka:** "But I need to pay millions more", yeah. + +**Mat Ryer:** "My kid didn't need to go to college anyway. He's probably smart enough, didn't need it." But that's nice that they're nice. Be nice on GitHub, everybody. That's the lesson there. + +**Bartlomiej Płotka:** Yeah. And this is why I solved those issues, because you were nice to me. Yeah, exactly. + +**Mat Ryer:** There you go. Yeah. + +**Björn Rabenstein:** Even for features that are really non-controversial, if a package is used so frequently, you are really, really, really concerned about not breaking anyone, and that sometimes is weird, right? Also, when we added exemplars... Exemplars is also a huge thing, right? Perhaps we can talk about that later, but for now, it's just a new feature that we didn't think about before. Now we had to add new methods to an interface, and stuff like that, but we didn't want to break anyone... So now you need interface upgrades, like little Go tricks; they look weird if you just look at them in isolation. But if you realize, "Okay, I couldn't break the users of the old interface", then it perhaps makes sense. But you get this kind of cruft in the library. And I think we have been talking about a legendary V2 for many years. It has never happened... But also, like, you conserve weird coding patterns, back then in 2012, 2013... + +**Mat Ryer:** Yeah. + +**Björn Rabenstein:** Guess what we all thought? + +**Mat Ryer:** For consistency. + +**Björn Rabenstein:** ...channels are the coolest thing, right? So we put channels in the function signatures, essentially used them as concurrency-safe iterators, which, you know, that's not what they're supposed to be used for. But they're still there, because if you now just change it \[unintelligible 00:27:42.22\] is this much cleaner signature for the function. Now, please, all the code in the world, please change to that new function. That's pretty hard... So we still have those channels there. And if you think that's weird - yes, I agree. \[laughs\] + +**Mat Ryer:** \[27:58\] Yeah. The cruft you mentioned is interesting, because in Go especially we pay a lot of attention to writing very easy-to-read, maintainable code, and sometimes we'll sacrifice performance, in the right place, for readability. And then whenever you have to then optimize, of course, you're doing different things, right? You're doing more complicated things, or you're finding little ways to save memory, or avoid allocations, those kinds of tricks. And then you end with sort of mess, and unusual bits, and things you wouldn't be very proud of maybe. But really, that's just the reality, kind of how it evolves. And I think the stability in the API is worth it, and I'm sure the community thanks you for the attention paid at that level... Because yeah, if this was a package that was breaking all the time, and you never knew what -- you'd end up with all kinds of horrible things when like a dependency used a different version, and things like this. So I think yeah, I can speak for everybody when I say thanks for that. + +**Bartlomiej Płotka:** Yeah, lots of trade-offs that are there. And I think what's also cool about this is it's just one module. These days it's so popular to create instrumentation libraries with like 10 modules, because you want to be, I don't know, generic, or have different versioning across. It's really hard to consume. But I have actually a question, Björn, to you. So you kind of started this library, I presume... Have you been designing the APIs with optimization in mind and you did benchmarking, or you rather created the API so it functionally works, and only then maybe after a couple of years you were maybe optimizing implementation? + +**Björn Rabenstein:** I didn't start it, right? I took over from the founders of the Prometheus project, Matt and Julius. They started this, and when I joined the team in 2013, I think, then one of the tasks I took over was this instrumentation library. And even back then, it started essentially with a whole rewrite, but I tried to keep the spirit intact. + +Yeah, we thought a lot about optimizing, because we knew this is in every path, if you instrument your code; but often, we were wrong. And micro-benchmarks is a cool tool. Actually, my first GopherCon talk - that was at the second GopherCon ever, or third, I don't know. Back then, it was really, really exciting. It went quite well. This was essentially about how to implement a float in a concurrency-safe way, so that it doesn't kill your performance. It's a little lecture. I mean, this is perhaps why this talk went well, because back then we were all learning, and it was a little lecture on how micro-benchmarks can help you, and also covered those things. But yeah, it's of course all the stories about premature optimizations, and everything. + +It's interesting how also later you realize, when your programs get faster and they want to increment this counter really often, how that is problematic. Then you have histograms where you want to increment pockets, and then you realize they have to be incremented all together, because otherwise you can get inconsistent results, if you collect your data at the wrong time. And thus, this created a whole talk for GopherCon UK, about all this concurrent programming, how you can do a lock-free atomic increment of all those pockets in the histogram. And yeah, it's kind of, even from an academic point of view, it's interesting, but it all comes with this warning, "Please do this for fun, but not in your production code." But if you do it in your production code, it should be really well justified; is this really in a path where all this complexity is worth it, and all the risk you take by doing this in the weird way? If you like, you can just go into the histogram code in Prometheus Client Golang. It's really weird, but it's also fun. + +**Johnny Boursiquot:** So speaking for -- I like to sort of put myself in the shoes of a user of these instrumentation libraries, not necessarily a maintainer of them; that's sort of the lens we've been using for a little bit... So if I wanted to instrument my code -- so we've already teased out what instrumentation means to different tiers, with all the people that are concerned about observability, right? So if I'm new to observability, and I'm looking to figure out, "Okay, what does it even mean to instrument my code? What is it that I should be looking at? How do I figure out whether I should be measuring requests per second versus latency? How do I even approach this world?" And obviously, there are different libraries and different things; there's Prometheus, I have to figure out where that fits in my stack... There's OpenTelementry, and there's metrics, and there's this, and there's that, there's traces... How do I -- as a developer, when do I use what tool, and for what? + +**Björn Rabenstein:** \[32:48\] I think there is a problem that - who knows if you use Prometheus for metrics collection? Perhaps you use something else, and then you want to instrument not just for metrics, but for all the other things. And of course, you have a lot of choices to make. And then there are a lot of efforts to unify this. OpenTelementry is definitely a huge effort of tying up all the loose ends, and it's really hard, right? + +So the first question in a practical context is you should look at what your organization you're working in is doing. Often, they use some framework, right? They use -- whatever. Let's start with a web router. You have your favorite Go web router, or whatever. And then from there on, you can kind of inform that decision. At SoundCloud, we had -- most microservices at SoundCloud were running on the JVM, and they had... This is also like somewhere in the public, people talk about it, they advocate for that, right? I think it was called JVMKit, where they had all that framework, how to write a microservice at SoundCloud. And then they put instrumentation in there. + +So just merely by using this framework, which you would use anyway, you were instrumented for traces, for metrics... You've got some logging, some standardized logging... It's all in there. We phrased this coin -- yeah, no, we coined the phrase that at SoundCloud it was easier to get monitoring than to not get monitoring. You couldn't avoid it, right? And that's -- for many developers, you're already in some framework. It may be a little software framework, or the framework of your organization. And there might be something -- or if there's nothing, you might just put the pieces in there and you could look for the right thing. + +I mean, rarely you have like a greenfield approach. If you do -- I mean, that's also fun, of course. You start your first Go program, and then you could just link in the Prometheus Client Golang... It's like, they're very minimalist programs in the documentation where you can just link it in, expose an HTTP endpoint, and do nothing else and you already get metrics for all the Go runtime stuff, for some process metrics, that's more like an OS-level stuff... So you don't have to do anything to already get a ton of useful metrics out of your program. And then you can go on, you can do really low-level metrics instrumentation. I also have a talk about this somewhere in my portfolio. We might link this. But there's also a set of middlewares. I don't know, perhaps - Bartek, you can talk about the middlewares we have, because that's also very nice how Go does HTTP things, it plays very well with that. + +**Bartlomiej Płotka:** Yeah. It's hard to mention, because you don't need to reimplement every possible metric instrumentation. So for example, if your application is doing an HTTP request, which probably 90% of applications do, you don't need to create a special separate HTTP request metric, or actually logline and trace as well. There are libraries that abstract this for you. Say you are using a standard HTTP library to create a server - you can just put a middleware, which is essentially a wrapper over your HTTP handlers, that will instrument automatically with very consistent metrics that are already -- someone thought through those to make sense, to make something reasonable. + +\[36:08\] And actually, the other plus of it is that you are getting lots of observability for free. So someone probably already built some Grafana dashboards, and maybe alerts, and maybe recording rules for this information, right? So we think you already created from HTTP package in Prometheus Client Golang, which already does that... And I'm maintaining a gRPC library, because we are using gRPC a lot. It's a very popular protocol, and we have literally the same middleware. They are called interceptors, but essentially they are like wrappers over gRPC things that add metrics. But also, we have logging and tracing in this module we could link. + +Break: \[36:49\] + +**Mat Ryer:** Well, that brings us to a new segment that I've just literally made up, but the editors do a great job when I do this... It's time for "Explain it quickly". So I'm going to challenge you, Bartek, to explain exemplars as quick as you can. Exemplars. We mentioned them earlier. What are they? + +**Bartlomiej Płotka:** Okay, quickest. + +**Mat Ryer:** Go! + +**Bartlomiej Płotka:** I think the easiest way to mention this - it is essentially an information that allows you to show an example situation that triggered some metric increment or metric latency observation, size observation, whatever you are measuring. And you can essentially -- usually, it's just a string, some characters and some timestamp, and actually an exact value of this example situation. Usually, we put there a trace ID to correlate with other signals like tracing, but we can put request ID, for example, to correlate with logging. You can put anything. But it's pretty useful for recognizing an example situation that is represented by this metric increase/decrease observation. + +**Mat Ryer:** \[40:32\] Cool. Congratulations. That was very well explained, and pretty quick as well, which really gets into the spirit of things, so thank you for that. Okay. But you don't include an example in every case, right? Do you do this kind of randomly, or how do you decide when to emit an exemplar? + +**Bartlomiej Płotka:** Björn, go for it. I can see you want to explain that. + +**Björn Rabenstein:** Do I? \[laughs\] I'm not even sure if I'm the most qualified person for that, but anyway... if you use metrics to measure something that is essentially just a whole bunch of events, like your HTTP request, then metrics means you count them all together, you kind of aggregate along certain dimensions, and then you have this gigantic number of requests, but you can't save them all. So this is where an exemplar is coming from; it's just an example. So you pick one, and then of course, which one do you pick? Do you pick a one that represents a normal request, or do you pick one that's like an exceptional one? Probably you should do one of each, or something, but not too many of the regular ones, because they are boring. You just want them for reference. + +That's actually a super-interesting question. We have a similar thing when you do like logging or tracing, where you do try to collect every single event, and you realize you can't really do this. So the naive thing is I just sample every thousandth - complicated word... Then you realize, "No, that's just getting me the regular ones", but I want those rare ones that have like a long latency, or they fail, for some reason. So it's a super-interesting decision, and I think this is a hot topic of research right now, which exemplars do I actually want to have. + +The current -- like, if you have a normal Prometheus histogram and put exemplars there, they just put the last observation that fell into each pocket, which I think is a pretty good heuristic, because you get exactly one exemplar for every latency band you're interested in... Which is a good start, I think, but I think this can be improved upon, and it will be interesting what the future brings here. + +**Bartlomiej Płotka:** And I think it's important to mention exactly what you did say about sampling... We have to make sure the example you put there, for example, trace ID - it is the one that will be sampled and provided in your tracing backend, right? So those systems have to be, at some point, connected. And I made some demo. It is very possible with Client Golang to just talk to your tracing library to tell what trace ID you can put, as long as it's sampled. But of course, there is a problem if you have tail sampling and other complex scenarios. So it's definitely not a solved problem. + +**Björn Rabenstein:** Yeah, we had those discussions a lot like you should... The tracing system should just sample those that Prometheus has as examplers. That's the one direction. Or the other is you somehow want to tell your binary to pick the exemplars according to what your tracing system has sampled. So it's a bit of - yeah, hen and egg problem. + +**Bartlomiej Płotka:** Anyway, if you are using Client Golang, and you are using tracing, and you are not using exemplars, you are missing out. It's a very underestimated feature. Please use it. + +**Mat Ryer:** There you go. + +**Johnny Boursiquot:** When should you adjust your sampling? In the beginning are you just basically saying, "Okay, I don't want any sampling. I want everything, because I don't understand the system yet. I want to see -- I want to detect patterns", or something like that, and then maybe before production you pick something that makes sense, given amount of volume and cost, because the more data you collect, the more expensive it gets. So what decisions are you making about sampling, and at what point in time? + +**Björn Rabenstein:** Bartek, you should be in a better position to answer. + +**Bartlomiej Płotka:** \[44:07\] Sure. So it's really tough situation, because it's really around tracing instrumentation and libraries. Usually, sampling is not application-driven. It's really communicated or done on a collector level, or agents. And so there are different phases you can do that. So it has to be collaboration between every of those signals, which we -- what works for us is that literally we use that on our production system, is you have whatever sampling suits you, as long as it's not tail sampling. So then it's very easy to essentially provide a proper exemplar to the proper metric. As long as it's tail - yeah, this is kind of impossible. + +**Johnny Boursiquot:** Can you define tail sampling? + +**Bartlomiej Płotka:** Yeah, of course. So tail sampling is when you decide that you save-- + +**Mat Ryer:** You can do this slowly, Bartek. This isn't part of the segment "Do it fast." You can just take your time on this one. + +**Bartlomiej Płotka:** So data sampling is when you decide if you are throwing a trace span, or you are saving it in your backend system, after even the request happened, because as we know, traces are really bound to some request. And it's very useful, because you can also then have a heuristic that will decide if you sample or not based on the result of the request. If it was slow, or if it's a failure, then probably it's interesting to sample that. If it's just a success, it's boring. Who wants to check successes, right? So \[unintelligible 00:45:31.12\] along those lines are tail sampling. + +**Mat Ryer:** Yeah. Actually, the more I learn and hear about this, the more complicated it seems. Should everyone pay attention to this now? What point in a career should a young engineer really start paying attention to this? Because like you said, Björn, we used to -- we didn't really do it at all. It was kind of handled usually by somebody else in some other way. And Johnny, you mentioned that you use monitoring to find out if the system is doing what the business needs it to do. So it's pretty important, isn't it? Should this be a fundamental piece that everyone gets to grips with in some small way, at least? + +**Björn Rabenstein:** I wrote this little piece in this 97 Things Every SRE Should Know About The Third Age of SRE, and it kind of ends with that sentence where we are really in the third age of SRE, if SRE is taught at universities. Just as if you're a computer scientist student and even before you get your bachelor you should have a few courses about, let's call it SRE... I mean, I don't know. Words are hard, right? And those things would definitely include that you need to instrument everything, and it's just essential. + +If you are looking for your first job, you might even check out if that company you're working for is already doing this, because if you're experienced with that already, you could even evangelize this in an organization, and it's great fun, because it's so satisfying to see all the progress you made by just introducing this, right? I mean, that was fun back at SoundCloud, when nobody knew if the site is even up, and then even the simplest things you do, they are so much relief. So that can also be quite good. But if you want to learn something in your first job and you want a role model organization to work for, that might be a good topic to look at if they are doing proper monitoring. + +**Mat Ryer:** Oh, good question for the interview there maybe. Johnny, aren't you an SRE? Aren't you literally an SRE right now? + +**Johnny Boursiquot:** Yeah. And I just realize where I've come across Björn before. We were both co-authors in the same book, 97 Things Every SRE Should Know. \[laughter\] That's how we met. + +**Mat Ryer:** Okay. + +**Johnny Boursiquot:** Yeah. + +**Mat Ryer:** That's a nice way to meet. + +**Johnny Boursiquot:** I know, right? + +**Mat Ryer:** The other 95 should -- you should all get together with the other 95 and have a big party. + +**Björn Rabenstein:** But the weird thing is there are actually 98 things in the book, and mine is the 98th. So perhaps I don't really belong to it. + +**Mat Ryer:** Is one of them a section on off-by-one errors? + +**Björn Rabenstein:** Yeah. + +**Johnny Boursiquot:** And the funny thing is, it's not one per author. There's fewer than 97. + +**Mat Ryer:** Oh, that'll save you effort. + +**Johnny Boursiquot:** \[48:10\] There's 98 in this case. Yeah. And then some of us wrote like two or three. \[laughs\] + +**Mat Ryer:** I'll save the flies + +**Johnny Boursiquot:** Oh, yeah, yeah. That's funny. + +**Mat Ryer:** Oh, it sounds good. Well, I recommend people get that book. By the way, I've read it, and yeah, I genuinely do recommend it. Guess what? It's that time... It's time for Unpopular Opinions. + +**Jingle:** \[48:28\] to \[48:43\] + +**Mat Ryer:** Who has an unpopular opinion for us? + +**Björn Rabenstein:** I have a very pressing unpopular opinion. It's because we already use that word that I don't like, and I hope I didn't use it, that's the O word, which is the word observability, right? I mean, my unpopular opinion is that I think this word, if it ever had a reason to be used in our profession here, it has lost all meaning by now. + +I can very well understand why people came up with this word, because -- I mean, I already talked about this paradigm shift that people sense when this word developers thought monitoring was just an ops concern, and ops people were monitoring by just probing the HTTP ports or something, or pinging the server, or something like that, and staring at dashboards. And when people realized this is not enough, and then you started to realize there's so much more... You need it for other things, like just keeping your site up... We didn't even talk about things like capacity planning. And we talked about business a bit... And then all those different signals you have. + +So I totally understand that some people felt this urge, "We need a new word for that", right? But then -- I don't know... Observability, arguably, is coming from control theory, and it's -- I mean, I think control theory is super-relevant for monitoring, but it's a rather specific theory, and just taking a single technical term from a theory and telling everyone, "This is now the new word..." + +I'm more a fan of, instead of overgeneralizing a very specific term, just take the existing terms and appreciate and acknowledge that we have all very different ideas of what that term actually means, and try to find like a wider, more common understanding of that term, right? Now, of course, the problem was also that then people used this word observability to essentially say, "So you are just doing monitoring, but I'm doing observability of my product. My software project is an observability project, and yours is just a monitoring project." And then it became a bit of like a word to fight over. What that means is then the marketing people will pick it up, right? And then someone came up with the legendary three pillars of observability, which sounds so great. And, I mean, it evolves as a concept. It really conveyed some insight and some widening of ideas. But of course, everyone picked it up, the marketing people picked it up, and then it was used so much that it became like -- then people had to discuss why three pillars of observability isn't even enough. So you run in circles. You have the same thing, and -- I don't know. Of course, I was brought up at Google, essentially, where they just called it monitoring, and they had tracing before people knew what it was, and all those things... And we just called it monitoring and it just worked fine, right? + +So for me, observability is like, if you want to use the term, it's something like a property of your system you want to have, but for me, it's a subset of monitoring, while most people think monitoring is something that might be part of observability, or not even. I don't know. What's his name? I always blank out on names. + +**Bartlomiej Płotka:** \[51:59\] No worries. But I just realized, Bjorn, that I was always saying that monitoring is a subset of observability, and now we are saying it's opposite. I guess I was close. + +**Björn Rabenstein:** Yeah... Cory Quinn. Cory Quinn said-- he always says on his podcast, "Observability is the hipster word for monitoring." And I kind of like that, but I would say observability is the marketing word for monitoring, right? And if engineers call something a marketing word, it's probably doesn't mean anything good in a way, right? So I think if I use the word, it might mean exactly what I want, or it might mean nothing... And yeah, I just stopped using the word. I just use monitoring. And if you ever hear me use the word monitoring, I use it in the widest sense. It's not just Nagios probes, and it's not just metrics. It's everything. + +**Johnny Boursiquot:** Do you use the words serverless? + +**Björn Rabenstein:** I do use the word serverless, but it-- + +**Johnny Boursiquot:** So, yeah... \[laughs\] + +**Björn Rabenstein:** Bartek and I just discussed yesterday -- or was it today? Yeah. No, it was today. It's a long day. We discussed how there are serverless applications that are not really serverless, and then there are the true serverless applications that nobody really caters for, because it's so rare that you have this pure use case. But yeah, I mean, words are fluid, right? We have to embrace that. But it doesn't become better if you add more words to the mix, and then very aggressively push them. + +**Mat Ryer:** Well, we'll be tweeting out that opinion to find out, once and for all, if that is indeed unpopular or not. Bartek, have you got an unporpular opinion? + +**Bartlomiej Płotka:** Oh, yeah. Do we have a time for my unporpular opinions? + +**Mat Ryer:** We do. It depends. How long is it? + +**Bartlomiej Płotka:** It depends, yeah. Let's see. Alright, so, quickly... + +**Mat Ryer:** Let's measure it. + +**Bartlomiej Płotka:** So I was thinking a lot -- I have many am power opinions, but some of it that I would truly want to, I don't know, inspire others is that... Stay with me on this. It might sound childish and unprofessional for the start, but... I think the Go language is the best foundation and platform to program every software in the world. And I really mean it, from embedded systems, robots, browsers, mobile devices, machine learning... Configuration even, like infrastructure as a code... And I mean it, because in my experience, I was working with many programming languages - Pascal PHP, C++, Python... I mean, yeah, many, many. Even niche ones, like \[unintelligible 00:54:17.11\] And the reason why I stayed and I have been working with Go maybe seven, eight years, and I still love it is -- I mean, there are many reasons already stated in \[unintelligible 00:54:28.11\] so let's not repeat that, but the point I want to make is that it's very powerful to keep this Go language as a foundation to create more specialized tooling, right? And you can have a lot of benefits by just keeping this foundation to inherit simplicity, reliability patterns, tooling, like for example IDE integrations, documentation building, auto-completion... + +Some of it is that, for example, like - yeah, JSON as a language for configuration. Maybe it's good, but I don't have hints when I kind of point my fields of this JSON on JSON nothing kind of hints me what the structure actually implements. And Golang already had that, right? So why not reusing this and building on the shoulders of giants? And I really mean that. + +We used configuration in my past job to build kind of configuration for infrastructure in Go, and it wasn't perfect, because there are some features of the language that maybe are too much. You have for loops, you have error checking, and you have a little bit of boilerplate, but if you refine the language, maybe specialize it a little bit, then it might just fit the use case just very well. Why we are creating totally new languages from scratch, like CUE or Jsonnet if Go is almost perfect? A little bit is needed to do to perfection. + +And maybe the last example, embedded systems I mentioned. Of course, garbage collection is not very efficient here, but there are already implementations of manual allocations. I think Dgraph created a special allocator. And there's even Vlang, which also kind of removed GC, but kept some of the features that Golang did. + +\[56:14\] So we might have maybe a version of Go that has Rust-like memory ownership, and that'll be amazing, because we keep other features instead of kind of living with totally different decisions that the Rust community made for other stuff, that maybe I'm happy with. So that's my unporpular opinion here. + +**Mat Ryer:** Well, I'll genuinely be interested to see if that is unpopular with our audience, because maybe a lot of people would agree with you. I don't know. + +**Björn Rabenstein:** The Go fans should all agree, right? + +**Mat Ryer:** Go fans. + +**Björn Rabenstein:** But I disagree, for one. I think it's a truly unpopular opinion, because I disagree with it. \[laughs\] + +**Bartlomiej Płotka:** Great. + +**Mat Ryer:** Fair enough. Okay. I'm afraid that's all the time we have for today, but this has really flown, and that's how you know we've had a good conversation... So thank you so much for that. We heard all about monitoring there, and observerbility... The myth of Prometheus - very interesting, the origins there of that story, and as well as exploring the exemplars and the good kind of practices and things. So I think it was honestly genuinely very helpful. Thank you so much to Bartek Plotka. Bartek, will you come back sometime on Go Time? + +**Bartlomiej Płotka:** I would love to. Yeah. + +**Mat Ryer:** We'd love to have you. Björn Rabenstein, also, will you come back? + +**Björn Rabenstein:** Yes, sure. It's one of my favorite podcasts. And if I'm in there myself, even better. + +**Mat Ryer:** Is it? + +**Björn Rabenstein:** It's boring for me then, but it's a great honor... + +**Mat Ryer:** Yeah. If you remember everything you've said, that's boring. So if you can try and forget it... Or just listen to it in a couple of years. But yeah, if you just said that you'll be back, that'd be great. + +**Björn Rabenstein:** Yeah, I'll be back. + +**Mat Ryer:** Thank you. + +**Björn Rabenstein:** I should say it with the Terminator voice. Yeah, "I'll be back." + +'**Mat Ryer:** Yeah, that's just your normal voice... Isn't it, Björn? + +**Björn Rabenstein:** Sometimes, yeah. + +**Mat Ryer:** And of course, Johnny Boursiquot. Thanks, Johnny. It's always good hanging out. Björn, good luck with finding John Connor. We'll see you next time on Go Time. + +Outro: \[58:05\] to \[59:55\] diff --git a/MLOps in Go_transcript.txt b/MLOps in Go_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..ef4ae13979223fabef412e103e49354c6521dd4d --- /dev/null +++ b/MLOps in Go_transcript.txt @@ -0,0 +1,333 @@ +**Natalie Pistunovich:** Hi-hi! I'm joined today by my co-host, Johnny. Hi, Johnny. How are you doing? + +**Johnny Boursiquot:** Hello, happy new year! This is like my first episode since the year started. It's the first episode of 2022. + +**Natalie Pistunovich:** Wait, is this really just the first episode of this podcast for the year? + +**Johnny Boursiquot:** No, I'm sure we released some already. + +**Natalie Pistunovich:** But this is the first one recorded in the new year. + +**Johnny Boursiquot:** For me, at least. + +**Natalie Pistunovich:** I'm not sure there were other episodes, but we will find out. Mystery. And we are joined today by Mike Eastham. Hi, Mike! It's great to have you. + +**Mike Eastham:** Yeah, it's great to be here. And you are joining us from the overseas side, or from Johnny's side of the overseas; so you two are based in the U.S, but you are based in Brooklyn... + +**Johnny Boursiquot:** Maryland. Not too far, actually. Both East-Coasters, yeah. + +**Natalie Pistunovich:** Is there a time difference between you two? + +**Johnny Boursiquot:** Nope. + +**Natalie Pistunovich:** Okay. Well, we have six hours apart to here, to Berlin, so that's 9 PM here. Fun way to start the year for me. \[laughter\] So Mike, you are a tech lead and a software engineer at Tecton. + +**Mike Eastham:** Yeah. + +**Natalie Pistunovich:** Or TectonAI, I should say, because the name of the product, as you told me, is Tecton. And Tecton is building systems to operate and manage feature pipelines and datasets for production machine learning applications for all sorts of customers that we all know... And you used to be a googler. At Google you were working on indexing and serving infrastructure for web search. + +**Mike Eastham:** \[04:08\] Yup. + +**Natalie Pistunovich:** Fun. So can you tell us a little bit on where you work and what you do there? + +**Mike Eastham:** Yeah, so Tecton is a company that was founded in 2019; the co-founders were all at Uber prior to starting the company, where they worked on a feature store product (or internal product) called Michelangelo. I think we'll get into it a little bit later in the podcast what it means to have a feature store product, but that's kind of where they met each other and got to know each other. + +**Natalie Pistunovich:** And I will pause you for one second to say that all the different names that you're mentioning, for example Project Michelangelo, we will include in the show notes. So if you wanna read more about this one product, you can find it in the show notes. + +**Mike Eastham:** Yeah, definitely. So yeah, the company was started around 2019; I joined in March of 2019, I was like the first engineer to join outside of the founders, so I've kind of worked on pretty much every part of the product since pretty near the beginning, when it was just very early prototypes, to where we are now. + +So the company is like 55 people now, around there. We've got a couple different sub-teams, and the engineering teams. So I guess we're kind of a mid-sized company now. Our main product is still this proprietary feature store that we've been working on since the beginning, but we've also started contributing to another open source feature store product called Feast. Kind of the lead maintainer and starter of the project works at Tecton now, so we're kind of working on both of those. + +**Natalie Pistunovich:** So both the products, the enterprise and the open source one, are feature stores. + +**Mike Eastham:** That's right. + +**Natalie Pistunovich:** What's a feature store? + +**Mike Eastham:** I think it's a very buzzwordy term right now, so I think you'll probably see a lot of people trying to explain it in different ways, or having a different idea about what it is... + +**Johnny Boursiquot:** We just went through the whole serverless thing, so we're accustomed to -- + +**Natalie Pistunovich:** New year, new term. + +**Mike Eastham:** Yeah. I'll do my best to give my take on what it is. Feature stores - they're part of this broader concept called MLOps, which you mentioned before. It's all about kind of applying a devops style mindset to getting machine learning applications into production. Feature stores aim to solve one piece of that pipeline, and that's managing feature definitions and feature data. + +So you could kind of break down into a couple different components what a feature store is. I'd say the primary thing is that it acts as a source of truth for your feature definitions within your organization or your team, whoever is working on your model. Basically, the idea is that you just have a single place where you define what a feature is, instead of having multiple different copies potentially of the same feature definition, or just having a bunch of different systems for all of your features if they're all spread all over the place. + +In our case, we have data scientists and data engineers define those features generally as a SQL query, which is kind of a tool that those types of roles are typically familiar with. The idea is that they can do it in one place and not have to prototype it using SQL, and then later hand it over to another team that builds the production version of that feature using Java or Go or something like that, and then having those things be out of sync. The first aspect is just having a single place where the feature is defined. + +**Natalie Pistunovich:** I will pause you just here to go one question deeper, and ask, for those who really focus on backend web servers, and that little world - what is a feature? + +**Mike Eastham:** Yeah, that's an important question. You can probably have a very mathematical definition of this, but generally, the features are sort of the engineered inputs that go into a machine learning model. So when you're in the process of making a machine learning model, you sort of have an algorithm that takes in some values and it uses those to predict an outcome. And generally, when you're building these systems, you kind of engineer those values to encode some sort of domain knowledge about the system you're trying to model... + +\[08:16\] So if you have a bunch of, let's say raw data about users in a system, and maybe you're trying to predict a product they might be interested in buying, as someone who kind of knows a bit about or predicts what people might wanna buy, you might say, "Okay, I think their age might be an important determinant of what things they might be interested in." So then you can build a feature that based on your raw data, let's say like a database of all of your user information, just extracts the age as a number. So you're kind of like distilling down things that you think might be important to the model training system. And that process is super-important for building models that actually perform well... Because the model training algorithms are now actually quite sophisticated, and they can produce really good results, but you're always gonna get better outcomes if you have better data going into the training. + +**Johnny Boursiquot:** Is the objective of a feature to make it a reusable part of the workflow for building these models? Or what is the objective of having a feature? + +**Mike Eastham:** Yeah, so I think definitely in an MLOps context you might say that you want it to be reusable... You can certainly build machine learning systems where all of the work is just done in a silo, and you don't reuse it across different models ever, but you're probably gonna have better outcomes for your organization -- if you ever have more than one model, you're probably gonna have better outcomes if you can share that work across different models. So that's definitely something that people try to use feature stores to achieve. + +**Johnny Boursiquot:** Now, how granular -- you gave an example about having a deterministic way of getting somebody's age, giving your data model, your schema, your data store, wherever you're storing your user information. Is the point to have these things very granular, to ease their reuse? Because they are very small in scope; they do one thing and one thing well kind of thing, and then you can just chain them as part of your model? + +**Mike Eastham:** Yeah, I think some of that depends a bit on what types of techniques you're using to build the model. So from the perspective of someone who's building a feature store, we kind of just wanna make sure that we support whatever types of features that people wanna encode in the system... So you can have really generic things like age, but you can also get even super-specific features that are something that's specific to a certain product and user combination. So I think there's like a variety of answers to that. + +**Johnny Boursiquot:** So a feature then is part of the set of inputs that you're feeding into your model? + +**Mike Eastham:** Yeah, so you use it both for training, so you'll provide basically a bunch of previous examples of feature values, as well as the outcome. An example would be I might have a bunch of features about a prospective shopper, like let's say their age, maybe the last ten things they had searched for would be in there, encoded somehow, things of that nature, and then I would have an outcome which would be like "Did they buy a product, or what products did they buy?" + +And then during model training, you provide all these examples, and then a model is the result of that. And then later on, you can basically use that model to predict the outcome from just the inputs. + +An example I gave before - if you're trying to predict while they're shopping what it is that they might be interested in buying, you have just those feature values as inputs; you feed that into your model serving system, and then you get out a predicted outcome. + +**Johnny Boursiquot:** \[11:50\] So I'm trying to figure out where features fit in the greater set of tools that you would use; if we're talking personalization here, for example, I do know in having used some commercial personalization products that there's different algorithms that are used and different models that are used -- or rather different approaches to training and being able to produce recommendations... Like the example you gave - if I've liked a few things, searched for a few things, then "This should be the next set of things that you might like." Kind of like going on Amazon and you search for one thing and you see other recommendations and you're like, "Oh, I'm seeing things I didn't know I wanted." Or you go to Netflix and you watch something and all of a sudden it's just recommending other things. + +So these kinds of recommender systems - it sounds like these are a kind of problem that the features that you're creating would feed into sort of helping these kinds of problems. So I'm trying to imagine in my head where the layer of things needed to have a fully-operational recommendation engine, what unit -- where's the feature at? Is it one of the first things that you do as somebody who's building a model, or once you train a model? What's the workflow? Where does that enter the picture? + +**Mike Eastham:** Yeah, so I think it is one of the first things you would start thinking about, but I think another kind of important aspect of MLOps in comparison to DevOps is that you have to go through a lot of cycles. So you might start thinking about features at the beginning, but to really effectively build ML systems, you wanna be able to quickly go through a lot of cycles where you build the features, you build a model, you see how it performs, and then that can give you more information about "Maybe I need to tweak my feature definitions" or possibly "I just need to get more data", or something of that nature. + +So I think it is one of the first things you would think about, but you also kind of have to be continuously monitoring what the values of your features are looking like, if you need to change the definitions, if you need that nuance throughout the lifetime of the model. + +**Johnny Boursiquot:** So what makes features sophisticated enough or complex enough to require their own data store? + +**Mike Eastham:** So there's a couple of interesting requirements from a systems engineering perspective when it comes to feature data. One of the biggest ones is that I mentioned that you kind of end up using feature values in two main places; you use them once when you're building the model or training the model, as part of your process of deploying a machine learning system, but you also need to use it in production once you have the model deployed, in order to actually make the predictions. So it's the same sort of logic that defines the feature values, but you need to apply it in two different places. And you have different performance requirements for those two different use cases. + +Generally, when you're making a prediction, in the context of most recommender systems, for instance - you have a user generally that's kind of sitting there at the other end of the connection, waiting for search results to be ranked, or something like that. You obviously want that to happen quickly; so in that kind of online context, when you're making a prediction, the latency is very important. Lots of times people have requirements of something like ten milliseconds in order to retrieve feature values, because that's only one piece of this whole process that has to happen to make a new prediction. + +Whereas when you're training the model, the latency is not super-important. I mean, you obviously don't want people sitting around for hours, waiting for training to happen, but the throughput is more important in that case. Generally, the more examples you can put into training, the better results you're gonna get. So that's a different set of priorities for those two different contexts that you need the feature values. + +In addition to the performance requirements being different, in the online case you're generally looking for the freshest data, so you want the most recent version of the feature value. That doesn't make sense with the age example I've been using, but if you're talking about a feature that's like the last product that someone bought, you want that to be updated very quickly after someone makes a purchase. + +\[16:07\] So you want those to be fresh, and have the most recent copy, whereas in the training or offline contexts, you want to get the correct value for like a historical point in time. So you wanna be retrieving the feature value as of the example that you're computing features for. So the interface looks a little bit different in addition to the requirements for performance. + +**Natalie Pistunovich:** Would you say that the reason that feature stores became a thing rather recently is that we expect faster everything? Like you said, fast search results, and everything? For example, feature stores were not really around five years ago. + +**Mike Eastham:** Yeah, I think that's a big part of it, the performance requirements. I also just think that people are trying to deploy these machine learning systems in a much larger variety of businesses and products than they were maybe 5-10 years ago. I think really large tech companies have had some form of what I'm talking about for quite a long time - Google, Facebook - but now there's really a lot of smaller organizations that are trying to build these types of things, and running into these problems that they can't put as many engineers on as a Google or a Facebook. So there's a lot more interest in trying to find these off-the-shelf systems that help out. + +**Johnny Boursiquot:** So given that this falls in the category of MLOps, as you've been referring to it, is this meant for the people that operate these systems? I'm trying not to conflate the people actually building and training those models with the people actually running those things. So there's a seam there, but I'm not quite sure on which side of the fence basically the feature store users and operators fall. + +**Mike Eastham:** Yeah, that's a good question. The answer is that both of those categories are users of feature store systems typically, and really, one of the most important things that a feature store can do is provide a clean seam between those two different roles. + +I briefly mentioned before, but before people were kind of integrating this feature store concept into their workflows, oftentimes the way a feature would make it into production would be a data scientist would kind of do the exploration in the data, come upon insights and figure out what feature values they wanted, and they would generally test those out by doing something like writing SQL, or maybe using ad-hoc Python scripts, and cutting data up with Pandas, or something along those lines, that would produce data that was good enough to do the training, but maybe wouldn't actually meet the performance requirements of doing the actual online survey piece. And then often what would happen is that the SQL queries or the Python scripts or whatever they are would kind of just get handed to a data engineering team, they would have to decipher whatever it was that was going on and translate that into a system that could compute the features in real-time. + +There oftentimes were mistakes made in that process, people didn't quite understand what was going on in one version or the other... And if you have discrepancies between the training system and the online system, it really hurts the performance of your model, typically. + +So the whole idea with the feature store is that we can just have the data scientists define their features once in a way that they understand... So we support SQL, we also support defining them in Python, if people prefer that... But once they've done that, it already is meeting the performance requirements that are kind of needed on the data engineering side. These behind the scenes do a lot of pre-materializing feature values, and things like that, to make sure that they could be served promptly when they're needed by the model serving system. + +So we have one part of the interface that's focused towards the data scientists that are doing exploration, defining new features, and then another piece that is focused towards people that are operating the serving systems and integrating them into CI/CD pipelines, and things like that. + +**Natalie Pistunovich:** \[20:11\] Yeah. A company I used to work for in the past, they had exactly what you described. There was the data science/engineering team, who were building models and taking data and crunching and coming up with new ideas and new rules, kind of... You know, "if this, then perform this action." Then they would just hand over those Python chunks of code to the Go team, which is the backend team, and then "Now translate this into the system. That's exactly what it's solving." So it's been a couple of years, but yeah, that's very much in my personal world of context at least. + +**Johnny Boursiquot:** That's not error-prone at all... \[laughs\] + +**Natalie Pistunovich:** Yeah, even just translating from Python to Go, you would lose some of the precision that they found, or something interesting... + +**Mike Eastham:** Yeah... I mean, sometimes model training systems can even latch onto little quirks of behavior you wouldn't really think about when you're doing this translation... But even if there's some edge cases where certain numerical libraries they're using, or something behaved differently, it can really cause a lot of problems that are quite difficult to find. + +**Natalie Pistunovich:** Yeah. + +**Break:** \[21:16\] + +**Natalie Pistunovich:** So we asked you to define some things, Mike, and I guess we've covered a little bit what was your definition of MLOps, but maybe you can say that for us, how do you see this, and how do you see this different from the devops and the infrastructure as a thing. You mentioned, for example, how many lifecycles... + +**Mike Eastham:** Yeah. From my perspective, the most significant difference is that with devops generally you're talking about deploying servers, or some other programs into production... You've kind of got this pipeline that's going from some code in your Git repository or wherever it is, and then you're compiling it, so these artifacts that end up in production. And of course, there's some steps there to make sure maybe you're doing canary deployments to make sure you're not breaking things, and you're testing, and making sure everything works reliably... But the actual transformation of the code to the server is actually pretty straightforward and fast. + +\[24:04\] With MLOps, it's really kind of an extension of the same thing, it's just that that process of transforming the raw data into the final product, which is like a model in this case, is really quite a lot more complex. So it takes longer, either because there's manual steps that are involved, or because some of these training systems actually take a significant amount of computational power and time to produce the results. + +You've got different kind of disciplines involved, with different stages of that pipeline... There's data scientists, data engineers, so multiple different stakeholders and people that are concerned with that pipeline. So really, there's a lot more complex pieces involved, is the main difference that I see. + +In addition to taking longer, it's also -- with a traditional continuous deployment pipeline it pretty much only goes one way. So you have the code, you produce the server, you deploy it, and then that's kind of it. But with MLOps, you always continuously have this loop back, because you're producing more data from the system that has the model deployed, which kind of gets fed back into the beginning of this process, to become more training data for the model, which we'll then deploy later. So you kind of have this circular nature that makes it a little bit interesting as well. + +So yeah, I think that's really the biggest difference... I think a couple other things are -- I mentioned this a little bit already, but you really have a lot of different kind of skills and roles involved with most ML systems than you do with a typical software system... So you have people that are just straight up software engineers, they're data engineers, and then there are people who are data scientists, who are technical, but they understand things more on like statistics, and math, and they have certain software tools that they wanna use, like Pandas, SQL, but don't necessarily understand the whole -- I think a lot of data scientists wouldn't be super-happy if you asked them to set up microservices by themselves, and things like that. + +So I think you kind of have to spend more time thinking about the different roles that are involved in making this system operate well, and make sure that you can kind of accommodate what they wanna do at different stages in the pipeline. + +**Natalie Pistunovich:** There's two names for this, right? There's MLOps and there's AIOps. + +**Johnny Boursiquot:** Ooh... + +**Natalie Pistunovich:** \[laughs\] + +**Johnny Boursiquot:** I wanna say, it sounds like we're splitting hairs, but I don't know enough to know for sure... \[laughs\] + +**Natalie Pistunovich:** So I did a tiny bit of research on Google trends, and things like this... AIOps is more common in the Americas; MLOps is more common in Europe and Asia. + +**Johnny Boursiquot:** But they really mean the same thing? + +**Natalie Pistunovich:** It kind of means the same thing. I mean, of course, you can go into splitting hairs, but this is kind of practically interchangeable. + +**Johnny Boursiquot:** Right. + +**Natalie Pistunovich:** I was also curious if you write that as one word, or two words. + +**Mike Eastham:** Well, I typically use MLOps, actually... + +**Natalie Pistunovich:** As one words, and you're based in the U.S. + +**Mike Eastham:** Yeah. + +**Natalie Pistunovich:** You're breaking the statistics. + +**Mike Eastham:** But I usually do like a capital M, capital L, capital O, and then the rest lower-case. I don't know what the specifics are on that, but I think some people do it all-caps, and there's a bit of a disagreement there. + +**Johnny Boursiquot:** Any other way, then it's just Emlops... \[laughter\] Or mlops. + +**Mike Eastham:** Yeah, it could be confusing. + +**Johnny Boursiquot:** "What is this mlops thing I keep seeing everywhere?" \[laughs\] "Is that a new product?" + +**Natalie Pistunovich:** I'm happy to say that very few people actually write with a space, but I guess that kind of makes sense, because you already write devops as one word, so it kind of continues that one. But if anybody was curious about that, here's all the information you need, and the terminology. + +**Johnny Boursiquot:** I will say, I am curious where Go fits into this wonderful world of mlops. \[laughs\] + +**Mike Eastham:** \[27:58\] Yeah, great question. So I can kind of give you some perspective on how we're using it at my company, and then maybe a little bit on how I think it's going in the outside ecosystem... Although I'm obviously more familiar with what we're doing. + +So I've mentioned before we have this online surveying interface for the system where low-latency is one of the big requirements... And so that's primarily where we're applying Go at Tecton. + +So we have basically a server that we deploy for every customer, that is taking in online requests for feature values, and we basically have a bit of a hybrid between pre-materializing the data and then -- so we have some of it prematerialized, but we do some kind of final aggregation steps at serving time, and estimate incremental updates easier when people need frequently-updated values. + +So we basically have a bunch of prematerialized data if we get this request coming in, and we do final aggregations, and then return that result to the user. So it's a pretty straightforward thing it's doing, but latency is really important, both keeping it low and then also keeping it consistent, because typically, people are concerned with how their tail latency is looking... So you wanna make sure that you control that and keep everything pretty consistent. So that's where we've deployed Go. + +The reason we chose Go for this use case - for me, the biggest thing came down to the tools that Go gives you, kind of during performance tuning, and being able to dig into performance problems; things like the built-in tracing and profiling we've used really extensively, in order to meet our performance targets for this component. I think it's worked out really great for us. + +**Natalie Pistunovich:** Was Go the first choice for this, or did you start with something and switched? + +**Mike Eastham:** Go was the first thing we used for this component, yeah. I'm trying to remember, it's been a while now... I don't think we super-seriously considered other options. We kind of by necessity have a large part of the system written in Python, because our SDK that we give to data scientists - they're kind of expecting Python there. I mean, we probably briefly considered it, but I don't think it would have worked with the performance requirements we have. + +Interestingly, we actually do have a Python interpreter embedded in this Go server. We allow people to do -- we call them on-demand transformations. They're able to put in transformations that run basically during serving time, and you might wanna use that functionality, for instance, to compute a feature which is like a combination of something in the user's history, but also a query that they're making, that you're actively making a prediction for. + +So you might wanna have a feature which is like -- a silly example would be what the difference in the length of the query and the last query was, or something like that. And you would do that inside the Python interpreter on the feature server. So it is a little bit of a hybrid, but 95% of it is all written in Go. + +**Johnny Boursiquot:** Now you have me curious about this thing you've embedded inside of the -- so this is not like you're getting a raw Python string literal and then you're shelling out to some Python runtime, and then getting results back. This is like you're actually interpreting the Python code and translating that into what you need to then do. + +**Mike Eastham:** So we are running a full Python interpreter; we're not translating the Python into some -- + +**Johnny Boursiquot:** Okay. + +**Mike Eastham:** Yeah, that would be pretty cool, but we're not doing that. + +**Johnny Boursiquot:** I'm like, "You need to open source that thing." \[laughs\] + +**Mike Eastham:** I have actually kind of looked into this. There are projects floating around that will let you basically compile Python down to LVM bytecode or something, and then you could potentially link that into a Go binary. We haven't really seriously investigated that. + +Most of these transformations are doing quite simple things, they're not super performance-intensive, so we haven't really seen a need... But it's something we could look at doing, potentially. + +**Johnny Boursiquot:** \[32:10\] So this was done really to solve not really a technical problem, but more of a meeting your internal users, where they are, kind of thing... Because if Python is what they know and that's their tool, that's the thing that they know best, supporting that to some degree was sort of the goal; not really because it couldn't have been done in Go, right? + +**Mike Eastham:** Yeah, that was definitely the main thing, that our customers writing these are primarily familiar with Python. There's also a bit of a nice thing technically with Python, in that you can just take the function definition as a string and just send it to the server, so you don't have to worry about linking it into the binary and restarting the server every time there's a new configuration, basically. That's something I'm sure we could have figured out with Go, but we just haven't really had a strong need to do it. + +I know there are other systems, like -- I may be misspeaking here, but I believe Airbnb has published, they have an internal feature store product where I think they... I think they have an equivalent concept where they have people write the transformations in Rust maybe, and they can make this whole thing super-fast... But I think that's kind of like an outlier case. Most of our users, when they're writing these things, are doing very simple additions or very small transformations that really only take a few hundred microseconds, generally so, we haven't seen it yet. + +**Natalie Pistunovich:** Would you say that Go is generally a good choice for machine learning ops? MLOps? Mlops? + +**Johnny Boursiquot:** Mlops... \[laughs\] + +**Mike Eastham:** Yeah, I think so... I think you'll find at the various layers of the MLops stack there are always gonna be these components that are pretty latency-sensitive. Across most of the customers I've talked to, generally, pretty low latency for the end-to-end system is a pretty important requirement. And I really like Go for that purpose. I mean, it's not necessarily that the language itself is somehow super-fast compared to other languages out there. In fact, we've had issues with the garbage collector that we might not have had in other languages... But really for me the tools that it gives you to identify these problems and deal with them make it a really good choice. So even if there are issues that come up, I am confident that we can solve them quickly, without spending a ton of time trying to figure out what's going on. + +**Johnny Boursiquot:** And that is worth the price of admission, my friend. + +**Mike Eastham:** Yeah, definitely. I would say the other thing too is that there's -- I mean, in the broader MLOps ecosystem I think a lot of people are using Kubernetes as part of the solutions that they're deploying, and I've found that just like being familiar with Go and the ecosystem is really helpful in that space, because it's pretty easy to just -- if you're confused at why something is behaving in a certain way, you can just go look at the code and figure it out. So I think that's pretty helpful too, being familiar with the ecosystem. + +**Natalie Pistunovich:** You mentioned the tooling is a thing that is super-useful for you in your world as a developer, in the world of mlops. Interesting. I would really expect that you would also mention something like the speed of the language, because it is really fast. + +**Mike Eastham:** It is certainly fast compared to something like Python... In my previous life as a software engineer at Google we were mostly using C++, so compared to that, it's not intrinsically faster. For me, I guess, given my background, I just look at the tooling and the ease of making everything work as the main advantage. But I guess if you're coming the other way, from having written a lot of Python, it would probably be a different story. + +**Johnny Boursiquot:** It's all relative. + +**Mike Eastham:** Yeah. + +**Break:** \[36:01\] + +**Natalie Pistunovich:** When did you start writing Go? What got you into this? + +**Mike Eastham:** Yeah, so it was at Google... I was trying to remember before I came on here... I think it was probably around 2015 was the first time... And it was not the primary thing I did there; like I said, it was mostly C++, but I worked a lot on the basically experiments system for web search at Google, and a big piece of that was that all of the different frontends that were serving a Google search would be basically creating kind of structured application logs of everybody's interactions constantly... And they would all get crunched down into a bunch of business metrics every day. And those metrics were what drove all the experiments that were going on at Google, basically. There'd be like thousands of these experiments running at a time. And the whole pipeline for crunching all of those logs down into metrics was at the time I was involved with it written in Go. And it still is. Before that, it had been all written in this domain-specific language called Sawzall, which was actually created by Rob Pike. So there's a bit of a connection to Go there as well. + +But yeah, at the time I was working on it everything was in Go. That system had -- there were a bunch of different teams at Google that would contribute their own metric to it. So the image search team would have metrics having to do with images, there would be some that were common across all the different search properties, but basically there were a ton of people contributing these metrics. There was a library for creating them, and I spent some time working on that library, because we were changing how things were logged in search frontends, and so we had to make sure that we weren't breaking people's metrics when we were doing that. So that's how I got involved writing Go, basically... It was trying to not break people that were downstream of the changes I was making. + +**Natalie Pistunovich:** And that's also, would you say, the main language for you now, right? Some Python, but mostly this. + +**Mike Eastham:** Yeah, we've got definitely a mix. All of our online serving stuff is written in Go; there's the Python for the customer-facing stuff... And then just for variety we have Kotlin that's involved in some of the more management and orchestration systems that are kind of off to the side. + +**Natalie Pistunovich:** And feast is the open source project, right? + +**Mike Eastham:** Yes. + +**Natalie Pistunovich:** And it has some Go. + +**Mike Eastham:** It has some Go. And frankly, I'm less familiar with the exact composition of everything on the Feast side... They're a lot more Python-focused, because they basically run a lot more of the feature store system inside the client, whereas Tecton's architected with more things living on the backend... But they do have some Go mixed in there as well. + +**Natalie Pistunovich:** \[40:12\] Okay. So anybody who's listening and wants to dive into this a little bit by contributing code, can do this in the repo of Feast. + +**Mike Eastham:** Yeah. Feast is a good place to start. + +**Natalie Pistunovich:** I love the way he alalyzed the documentation, I have to say, of Feast. It's a great job at that. + +**Mike Eastham:** Yeah, they've done a really good job with all that stuff. + +**Natalie Pistunovich:** Cool. Okay, and Johnny, as we laughed earlier, quickly skimming through everything, and now he's becoming the expert... \[laughs\] + +**Johnny Boursiquot:** Ingesting data, yeah. Next week interview me on mlops. I'll have all the answers. \[laughter\] + +**Natalie Pistunovich:** But then I will include the question on the difference from mlops to aiops. \[laughter\] + +**Johnny Boursiquot:** Right... + +**Natalie Pistunovich:** Then I guess comes the fun part... Unless, Johnny, you want to ask anything before we start talking about the unpopular parts. + +**Johnny Boursiquot:** Let's get to the unpopular stuff. + +**Natalie Pistunovich:** Alright. + +**Jingle:** \[41:07\] to \[41:22\] + +**Natalie Pistunovich:** I don't know if I'm supposed to say this, and I'm sorry if not, but the performer of this wonderful song has a birthday today. Finally turning 18. So if you see Mat on Twitter... \[laughs\] Wish him a happy birthday. But only if you're listening to the live recording. Actually, you know what - even if you're listening afterwards, just say hi to Mat. You don't need to say happy birthday. Just say "Hey, Mat." + +**Johnny Boursiquot:** He's finally an adult. + +**Natalie Pistunovich:** \[laughs\] Maybe he'll grow up. + +**Johnny Boursiquot:** Maybe he'll grow up. \[laughs\] That's an unpopular opinion, I guess. Mat, we love you very much. It's all in good vibes. Mike, do you have an unpopular opinion for us? + +**Mike Eastham:** Yes, I have an unpopular opinion. It's not tech-related, which I hope is okay, but... + +**Johnny Boursiquot:** It's all good. + +**Natalie Pistunovich:** That's not popular. + +**Mike Eastham:** I really just don't like maple syrup, is my unpopular opinion. Even on pancakes, on waffles - I just don't do it. That's my opinion. + +**Johnny Boursiquot:** Okay... + +**Natalie Pistunovich:** Johnny, I saw a tweet from you from the weekend that you were making pancakes. + +**Johnny Boursiquot:** Yeah, yeah. You know, I have young kids, and they love themselves a pancake, especially the syrup, so... \[laughs\] + +**Mike Eastham:** It seems almost universal... Just not for me. + +**Natalie Pistunovich:** What don't you like about it, Mike? + +**Mike Eastham:** How sticky it is is a big part of it. + +**Natalie Pistunovich:** Is it more or less sticky than honey? + +**Mike Eastham:** Somehow it feels more to me... I don't know if that's chemically accurate, but that's my perception anyway. And we've got a lot of Canadians working at Tecton actually; in fact, the CEO is Canadian, so... I hope if they hear this I don't get ostracized. + +**Johnny Boursiquot:** \[laughs\] They'll be like, "Oh, you don't like syrup, huh?" Well, there you have it. Unpopular opinion. + +**Natalie Pistunovich:** Yeah. I think the most funny opinions we have is in the context of food, so... Great choice for an unpopular opinion. There will be a Twitter survey; we will see how unpopular is your unpopular opinion. You might get into the top five, or low five + +**Mike Eastham:** Are you supposed to have the most people disagree with you? + +**Johnny Boursiquot:** I mean, it's a goal... We all have life goals, so why not this one? + +**Mike Eastham:** Okay. Hopefully I do well then. + +**Johnny Boursiquot:** Mm-hm. + +**Natalie Pistunovich:** Cool. Alright, well that was a very interesting conversation, and I think it's a rather new field, at least in our context of gophers, this mlops in general, but also feature stores in specific... And we will be providing links for you to learn if you wanna dive into that, or if you wanna contribute code to Feast. And we also want to say thank you to Mike for joining us and for teaching us so many interesting things. + +**Johnny Boursiquot:** Indeed. + +**Mike Eastham:** Thank you for having me. + +**Natalie Pistunovich:** Thanks, everyone. + +**Johnny Boursiquot:** Alright, y'all. diff --git a/Making Go more efficient_transcript.txt b/Making Go more efficient_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..d899bc551ac47ca3d49d635f7e7a12b01f1a9e52 --- /dev/null +++ b/Making Go more efficient_transcript.txt @@ -0,0 +1,325 @@ +**Mat Ryer:** Hello, and welcome to Go Time. I'm Mat Ryer, and today we're talking about making Go more efficient, and how best do we do that. I'm joined by Bartek Płotka, author of Efficient Go, which was released last month, principal engineer at Red Hat, and a Prometheus and Thanos maintainer. Hey, Bartek. Welcome back! + +**Bartlomiej Płotka:** Hello. Hello, everyone. How are you? + +**Mat Ryer:** Very good. Thank you very much. And you're calling in from London, right? + +**Bartlomiej Płotka:** That's correct. Yeah. + +**Mat Ryer:** Yeah. Cool. I'm also joined by Kemal Akkoyun. + +**Kemal Akkoyun:** Hello, Matt. + +**Mat Ryer:** Hello. You're a software engineer at Polar Signals, right? And also a Prometheus and Thanos maintainer, and you're currently building an open source eBPF-based continuous profiling solution called Parka. Welcome. + +**Kemal Akkoyun:** Hello. + +**Mat Ryer:** Great to see you. This is your first time on GitHub, I think, isn't it? + +**Kemal Akkoyun:** Yeah, it is. + +**Mat Ryer:** Yeah. Good. + +**Kemal Akkoyun:** A bit nervous. + +**Mat Ryer:** Yeah. Oh, don't be. I am a bit as well, just because of the sheer brainpower on this episode... I'm out of my depth. I'm also joined by Christian Simon. Christian is a software engineer at Grafana Labs, working on observability databases like Loki, Mimir, and the new one, Flare. Hello, Christian. + +**Christian Simon:** Yeah. Hi there, Mat. How are you? + +**Mat Ryer:** I'm good, thanks. Welcome to Go Time. It's nice to have you here. Let's get to know our guests a little bit, shall we? Kemal , what's your favorite drink? + +**Kemal Akkoyun:** I guess I have two... One is single malt scotch, and the other one is coffee. + +**Mat Ryer:** Oh, nice. Yeah. Do you have them together? there's the Irish coffee that we call it, if we mix them together... + +**Kemal Akkoyun:** No. I have them separately, one in the morning and one in the afternoon. + +**Mat Ryer:** Okay, we won't ask which is which... What about you, Bartek? That sounds pretty good, doesn't it? + +**Bartlomiej Płotka:** Yeah, definitely. That's a +1 from my side as well, coffee and whiskey... And I got approval from my personal trainer as well. He said "Whiskey... Is it clean at least?" and I was like "Yep, clean. Just ice" and she was like "Okay, that's fine." So I think it's good. + +**Mat Ryer:** Oh, wow. Yeah, that sounds like a good personal trainer... Christian, do you have a favorite tipple? + +**Christian Simon:** Yeah. I think when I'm at home, I'm definitely more on the juice surely, \[unintelligible 00:05:10.08\] basically, as I'm originally from Germany, so that's the drink there... + +**Mat Ryer:** I see. + +**Christian Simon:** But yeah, I think given the feedback from the personal trainer from Bartek, I think I might need to change that and revisit. + +**Mat Ryer:** Yeah, no wonder you're not ripped... You're not drinking enough whiskey. Oh, you're from Germany, so -- okay, that's why we're doing an episode on making Go more efficient. Makes sense. That's the classic stereotype, everybody, which I don't recommend... + +**Christian Simon:** Yes. People haven't lived in Germany. + +**Mat Ryer:** \[laughs\] I don't know... Well, you have, haven't you, Kemal? Because you live there now. Do you like it? + +**Kemal Akkoyun:** I really like it. I like Berlin. It's been five years... Learning to be a German. + +**Mat Ryer:** Ja... + +**Kemal Akkoyun:** Slowly. + +**Mat Ryer:** \[05:58\] Nice. I think Berlin is a great city as well. GopherCon EU was there this year, so I got to go and hang out in Berlin, and we had a great time. That's the small talk over; the small talk section is now finished, okay? No more small talk now. Let's get down to business. I want to know just something initially... making code obviously more efficient - I think it's kind of the case or it's obvious -- You don't use as many resources... It's actually sometimes like the greener thing to do as well, because you're using literally less energy, especially when you get at scale. But for sort of like people that are new to writing Go or they're just getting started, do they have to worry about the efficiency stuff from the very beginning? Is this something that you should just get it working first, and then you can make it more efficient later? + +**Bartlomiej Płotka:** So there are two schools actually, right? One is saying that premature optimization is the root of all evil; it's like a very old saying that got very, very, very popular, especially in the '80s and '90s... And that was kind of claiming that engineers are playing too much. They're having too much fun of making those optimizations, and efficiency improvements... That at the end it didn't matter, and it actually make things worse, because the code was unreadable, unmaintainable, and changing it to be faster for different circumstances was harder, so at the end it was less efficient. So that's one school... And I think there are good truths to be learned from that. But nowadays, I'm kind of -- you know, maybe discussing in my book as well, but also in my talks, that sometimes we are too afraid of touching the efficiency topic. So we are always kind of putting that as like last step in your development practices last priority and those are excuses to not do this maybe a little bit extra additional work to make things faster and more cheaper to run. + +So we can go deeper, but generally, yes and no. Generally, especially if you're learning, focus on readability, maintainability, and those kinds of first things that -- you have to learn how to make your code functionally running, so kind of like fulfilling your functional requirements, and then you can optimize it later. However, with time, and if your team is more kind of productive, and can learn more, and can use tools that we'll hopefully talk about, around how to make things more efficient, you can kind of like move this step sooner in your development process. However, you have to know what you are doing, so if you're new to Go, definitely don't even aim to make like code more efficient until you know you have to do it, and then you can learn how to do it. So that's the balance \[unintelligible 00:08:40.07\] + +**Christian Simon:** I would definitely agree. I guess, as with many other things, I think a balance of the two is probably good, and kind of the better you understand your ability, your team's abilities, but also maybe the problem space... Like, let's say you attempt to rewrite something you think you understand quite good - I guess then maybe efficiency is some of the reason of that rewrite... While maybe if you're trying something completely new, you might just have the wrong assumptions of what you're optimizing against... So there's no one size-fits-all, I guess. + +**Kemal Akkoyun:** Yeah, I totally agree for both of the statements... But I guess as Bartek put that, I belong to the first school. The original quote is from Donald Knuth, and it says 97% of the time. So I think that 3% sometimes makes the difference. And you can always determine that 3% by measuring things. In Go we have lots of tools around that, that you can actually measure and benchmark and decide maybe even from the beginning what to optimize. + +**Mat Ryer:** \[09:45\] Yeah, so this is interesting then, and I kind of like what we're going to get into... Like, this idea that we don't really know what we can do to make it more efficient necessarily. And sometimes, even experienced engineers will have a really good idea, it'll be really obvious, but it turns out that it doesn't necessarily make an impact when he makes certain changes. So how do we know when we've got the opportunity, when we've found that 3% to make something more efficient? How do we know? + +**Bartlomiej Płotka:** Well, we need to make some measurements. We need to observe our system under load, under kind of requirements we set it to be running with, and essentially observe what is causing the problem... And then generally, at what level of efficiency we are in. So those two things that generally I in my book am kind of mentioning - you need to assess your efficiency, you need to be able to assess your efficiency... Because perhaps you are actually fast enough; it's actually some other component that is causing the problem. Or something might be outside of requirements; maybe a request is going to your HTTP server written in Go, and it's actually a malformed request, and that's why things are slow, so maybe some safeguard check is enough. + +So this assessment is one thing, and you have to kind of measure that. And that generally, in kind of observability space, we are going into. You have to measure it while running; either running on production, or locally, under your benchmark, we call it. So test, and experiment... + +So assessment is one point. And the second point is bottleneck analysis. So once you know that things are too slow, so you generally know you are, for example, one second behind your expectations of the HTTP request, for example, you can go and find some tooling around observability that will tell you the bottleneck. So what part of my code is actually a main contributor for the problem? Maybe it's too slow. Maybe it's using too much memory. And that's a bottleneck, right? And there can be multiple contributors, right? Of course, it's usually multiple. It's not like only one of your programs using all the resources. But generally, there is this rule called Pareto, which is saying that it's usually 80% of resources used by 20% of your code, things like that. So generally, there is some bottleneck that this thing that you should focus first, to have the most impact. And bottleneck analysis is, again, something you can get from observability. We could mention it later on. + +**Mat Ryer:** Does anyone disagree with that? + +**Kemal Akkoyun:** Everyone agrees. Everyone works on observability and monitoring, so who can disagree...? + +**Mat Ryer:** Yeah, exactly. Good point. Okay, so observability then... So this is literally like monitoring resource usage, or the time it takes to do things... Is there a limit, really? How do we know what to measure? + +**Kemal Akkoyun:** I guess it depends on what kind of program or service that you're writing, right? Let's say you have an online system, and it's a request serving system, like HTTP or gRPC. Maybe you are putting this program as a service, and then we can talk about maybe service-level objectives. And what actually affects customers, like, what are your error rates? What are your latencies? Then from that, we can come up with certain metrics and start measuring or instrumenting our codebase. + +This is one example, but we also have different types of systems... Maybe they're not online request-based systems, but they're batch processors, or offline systems, or some event processing system - then you need to consider other stuff. So everything starts with knowing your basic purpose of your system. + +**Bartlomiej Płotka:** Maybe I will play like a devil's advocate here, Kemal... And I agree with you, but I think that the general misconception these days is that as a developer, you are giving me some solutions for SRE, for DevOps, for operator, that show me some kind of metrics of SLOs, and service availability... But I am developer; I don't care if the service is up or not. I care if my code is faster, because I'm paid for making the software better. Right? So how those tools kind of relate to me? + +**Kemal Akkoyun:** By tools you mean like observability \[unintelligible 00:13:58.11\] + +**Bartlomiej Płotka:** Like metrics of HTTP requests on my SLO dashboard - how can I use it for efficiency improvements? + +**Kemal Akkoyun:** \[14:11\] This is your North Star metric, right? The most important one. Because think about - you are having a website, and you are actually serving a website, and the most important thing for your user is how fast it actually renders. Because psychology shows us there are some attention spans, and they are getting lesser and lesser every day... And so you need to be fast, right? + +Or even before that, if you're serving that request with a success, you need to know that. That's why I always advise to start with these type of high-level metrics. But then - let's take latency, right? HTTP request latency; you're serving this website, but then you can drill down, like what does this mean for the system? What contributes to this latency? If you think of a tracing observability signal, it's a span, right? In that span, you have different spans, like different methods or different components in your system that takes time. And now you need to get in there and understand why it is taking that much time. Maybe you are doing a calculation that it's not that efficient in your hot path. Or maybe you're allocating too much memory, and you need to know about it. But it always starts with your high-level goal, and drilling down using different observability signals, and maybe instrumentation. + +**Christian Simon:** Yeah, great. Exactly right. I was kind of challenging you... But generally - yeah, I agree; you can use the same metric, kind of pipeline. So the same client instrumentation in Golang, where you are either importing some library with this instrumentation, or you have instrumentation built-in, maybe Prometheus, maybe something else, maybe OpenTelemetry, and you have that built in... And the same thing that your SREs are using, your operators, or people who are running your software; maybe you are running it, and that's ideal. But sometimes you're not. And then as a Go developer, you could use exactly the same metric, for example HTTP latency, as your guidance, as your assessment value, if you are improving your code or not. So that's kind of like an important part of observability/monitoring here, that kind of can be composed into something useful for efficiency. + +**Mat Ryer:** So say that you've got this setup, then, and you've got a good idea of some kind of metric on your code. And hopefully, what it's doing in production; because that's the other thing - like, it might work great on your machine, but once it's running at scale, for whatever reason, just the way it interacts in ways you just wouldn't expect maybe... So the environment's very similar. I think it's probably good advice. So so you've got that - do you like check that every time you do a PR? Is that something that you keep an eye on things to see if you've made a difference? Or do you just do it kind of periodically, like once a quarter you'll just have like "Let's do a performance check"? How do you approach it? + +**Kemal Akkoyun:** We haven't jumped into that, but right now, I'm working on a continuous profiling tool. We haven't even basically defined the profiling, but let's say we define - profiling is kind of an observability signal. It's an action that you collect the resource usages of your processes that's running on your machines, right? And what makes it fun is this continuous profiling part. So you are profiling all these processes, and then you release a software; maybe you scale up, you scale down, but you always have these snapshots of memory usage or CPU usage in a timespan, right? And then you can actually check between versions if you actually changed something. So when you have a continuous signal like that, you can actually optimize or do something about any of these actions, right? + +\[17:56\] Same goes for metrics. You have a latency metric, and you can define some SLO-based alert on that. And that alerting part actually makes the difference, right? So all of a sudden, you release a version, and your latency has spiked, and it actually triggered an alert. Now that you know -- since you put that alert in there, you know that you are actually exceeding your allowed boundaries, your error budgets maybe. And then you can actually get in there and try to optimize, try to do something about that. But it's the fun part of that, right? We mentioned the profiling side of that, and we mentioned the metrics side of that, and we mentioned alerts, right? They come all together, actually. You get the alert, you check what's going on with the process, down to the method and function, and then you decide to do something about it. + +**Bartlomiej Płotka:** I think it's great. I think you're kind of explaining the production assessment level, right? So when you're actually building your PR, merging, everything deploys, you release it to different environments, and finally it gets to production. And that's your assessment. That's probably that ideal assessment in terms of kind of how real reliable it is, because that's actually production; that's what matters. Things are not very easy if you maybe distribute the software to different productions, or maybe to different customers. Probably it varies. That's why I think it's very useful to think about this problem the same way as we think about testing, about functional testing. + +We also have unit tests that always are running on PR. We have integration tests that sometimes are running on PR. It depends how much you want to pay. And there are production tests, or load tests maybe, and end-to-end tests. There are different kinds of levels, and no one is saying you should always only implement unit tests and nothing else; or only implement end-to-end, and nothing else. There is a graduation process where things closer to the code cost less, but it's not always reliable, and then things farther away, towards production, is more reliable, but it's kind of sometimes too late, because the feedback loop is very, very large. This is exactly the same for assessments. You have a micro benchmark, which is kind of a unit test, then you have macro benchmarks, which is kind of integration end-to-end, and then we have production assessment. And the problem is these days ideally we run on every PR micro benchmarks. But we don't, because it's still very expensive, it's still unreliable... So we hope maybe hardware will get cheaper... I mean, there are some kinds of solutions that do that, sometimes we do that, but generally, it's not achievable yet. + +What we have in Prometheus, for example - we have macro benchmarks. So we spin up a test cluster with Prometheus server, which is kind of like written in Golang, some time-series for metrics. And then we can kind of spin up this setup on the PR. So we say, "Hey, prompt test kind of start" and then it starts, it gives you dashboards, it gives you results, and then you can shut it down, because it costs money for the CNCF organizations, so they can't run after hours But generally, ideally, the space is better, and we run everything on the PR, but that's not the case always, right? + +**Christian Simon:** I guess for us, we also have that kind of \[unintelligible 00:21:10.06\] tools before we cut a release. At Grafana Labs we usually cut a weekly release, and then go through different stages, like more of the internal tools, and then into production. And yeah, we would really only alert on SLO. So if we think it's worth, we would compare our benchmarking in the PR, but there's not an automated way, and there's not a good way of kind of raising a question mark about a PR once things would go out of hand. I was wondering how that works with the \[unintelligible 00:21:42.28\] Do you get that kind of percentage assessment and then the reviewer needs to make a call? Or is it getting kind of hard-rejected in the PR pipeline? + +**Bartlomiej Płotka:** \[21:55\] Yeah, it's very manual right now, because again, it's kind of a community project, so it's not like tied to any production, one production kind of environment. So it will be hard, but it's kind of manual these days. But it's good enough, right? However, I think that if you look on the Vitess project, I think it's kind of in Go as well, and they have an amazing page around benchmarks. So they do nightly benchmarks, like all micro benchmarks in a codebase. In Go, benchmarks are kind of built in, the micro benchmarks. So you use go test to run your benchmarks, and they are very nice and easy to write. And if you have them in the codebase, you run a specific go test command, it will just run everything. And they kind of try to have hardware, so kind of virtual machines, which are stable, but also... By the way, running virtual machine is also sometimes unreliable, because \[unintelligible 00:22:46.12\] the characteristic is different. There are so many complexities. But sometimes, if you run it long enough, it can give some data. So on their website, they are having the results for every nightly builds, and it's pretty sweet. + +**Break:** \[23:01\] + +**Kemal Akkoyun:** I think I'm going to disagree with Bartek, at least partially, about testing this stuff... So here comes my first unpopular opinion. So even though testing and doing end-to-end tests or synthetic benchmarks on your development machine or on your CI, your development environment is not the same with your production environment. I am a big fan of collecting these metrics in production, or just profiling the actual workload in the production. Because the CPU, the machines, the network stack that actually this software runs on is totally different than what you have in the CI or in your local machine. So if you can do this on a production environment, that data is also the actually holy grail, right? + +On top of that, one other thing is with the prom bench, or with the Prometheus project, we actually know what is the hot path, so that's why we have those benchmarking tools. We know what to optimize, and we know beforehand what shouldn't go out of range, so that we have that in CI. So it's really helpful. + +But most of the other projects, you don't have this information, and you actually don't know what to optimize. So there's a lot of unknown unknowns. For those types of things, actually doing this on production, with certain metrics, or like certain facilities to collect this data is more valuable to me. + +**Mat Ryer:** Yeah. Could you have like a budget, so that you say, "We're going to add features, we're going to add extra work, so we know that's going to slow it down, but we want to set a budget for how much we could do that?", something like that, those kinds of techniques. Does that work? + +**Christian Simon:** I think I'm always a fan of having maybe two feature sprints, and then one tech debt clean-up to maybe bring those new inefficiencies or newly identified - yeah, hot paths, maybe fix them, make them more efficient... I think that seems to work quite well, I guess. And we also notice in the project if we run too many of the feature sprints, then I guess we're gonna pay in terms of getting a bit more red graph when we look at our SLO budgets. + +**Mat Ryer:** Yeah. It's quite nice, because you can kind of free yourself and focus on the features, and then you know that at some point you'll then go back and have that dedicated time to pay off the tech debt, or just see what the impact is. + +**Christian Simon:** And you don't know how a particular feature is used... So I'm working on continuous profiling as well on Flare, but before, on Loki, we implemented deletions, and we had an idea of what kind of requests to expect, but I guess within the first month we had very interesting requests, in terms of their size, and just the number of them, that we kind of didn't kind of optimize for, and that then required another sprint of making them actually scale to that level. + +**Mat Ryer:** Was it just that people wanted to use a new feature, so they're like "Right, what can we delete? Let's do it." Because I'm like that when I get a new phone; I'm like "Yeah, I'm using all these features, because they're new." You know, it's an expensive phone, so... + +**Christian Simon:** Yeah, let's delete every second line of your logs. + +**Mat Ryer:** \[27:53\] Yeah, why not? You halve your bill... + +**Bartlomiej Płotka:** Yeah, I think it always kind of depends on the case... But a wonder - let's say you have those weeks for the feature development, and then you deploy those features, and the SLO is... Or like some kind of like this measurement assessment is threat. What do you do - wait until the tech debt, or you revert the feature and work on optimizations? This is where this thing is kind of tricky, right? + +**Mat Ryer:** What do you do though, Christian? + +**Christian Simon:** Yeah, I guess we look at error budgets, and I guess as long as we are within the budget, I guess we can kind of push it down the line, I would say... But yeah, I think if we kind of exceeded our error budget, then this definitely becomes not \[unintelligible 00:28:31.03\] but definitely will kind of drain out all the feature work. + +**Bartlomiej Płotka:** Yeah, makes sense. It makes sense, yeah. Definitely. And it's kind of like a data-driven approach, where you kind of express everything in number... Like, "Hey, how much can I sacrifice?" However, the problem is usually people don't set the goals, don't set the performance requirements, or efficiency requirements. And for maybe a service like yours, you have some latency, but what about the memory used, or other costs, like CPUs? It's so hard sometimes, and we don't always do it, even at Red Hat; it's like, you've got the feature, but what's your requirement? What's your goal? Are you happy with spending thousands, or it should be like $2,000 per month, or it should be $10,000? Ideally, you have that goal, because you can calculate those budgets then, both on latency and the resources, which is cost, right? + +I think that's the trouble that many engineers have; it's very hard to set the goals, because it sets you up for failure, and this kind of like psychology behind that. I think we should do it more. I think we should set more goals, to know our budgets, essentially. + +**Christian Simon:** Yeah. I think it doesn't help that the billing cycle, for example from cloud providers, is not really matching up maybe with your kind of release cycles; they're often delayed, or kind of over longer periods... So we definitely kind of got that with a bit of delay... And then it was even more important, because I guess it aggregated over that timeframe. + +**Bartlomiej Płotka:** Yeah, that's a cool problem to keep in mind. I guess there are tools like OpenCost, or even CNCF tools that kind of calculates, gathers some observability about your projects, and even Go applications, and then they kind of try to calculate the costs. So yeah, maybe that will help, but... Yeah, generally a hard problem to solve. + +**Mat Ryer:** Yeah. So that's really interesting... Kemal, you mentioned earlier really the best way to get this data is to do it in production. Is there a cost to that? Because you're doing extra work to get that observability data out. Does that itself, not slow things down? + +**Kemal Akkoyun:** Of course, there's a cost attached to that. There is no free lunch, right? So that's why when it comes to observability, you actually -- or at least I tend to start with metrics, right? That's the low-cost observability signal that you can collect from your systems, because it's just an aggregation. And if you are actually monitoring your system using Prometheus, it's just in certain scraping intervals, you just like get your aggregated metrics and just store them, and you build your alerts or dashboards on top of that. + +This, as far as I've seen, in most of the cases, our metrics are super-useful, and so it does the job. But in most of the cases, there is this in-depth situation; maybe you need more information about your system, maybe that's where the logs comes in, or any eventing system that you collect from your system comes in, and you check what's going on with the events. + +If you have tracing data - this is really nice, because then right now, you can actually propagate your context between other services or other context boundaries, and you can have a broader picture of the system. This is another observability signal. But again, there's another cost attached to that. First of all, it's really expensive to store the tracing data, and the other cost is you actually need to instrument your code. And tracing instrumentation tends to be a little harder when you compare it to the metrics, because you need to go in there and you actually semantically determine which function boundary that you're crossing, and you need to just propagate that context, or you need to do this to all the systems that you have in your architecture, right? But metrics, in that sense, it's easier. And we know logging - everyone likes to omit the logs, even though I find them super-noisy, and it's really hard to take actions against them when you have hundreds of instances of the same process, right? + +**Mat Ryer:** \[32:33\] We are the one deleting them then, all that time... + +**Kemal Akkoyun:** I tend to just not store them... Even though I really like writing log lines, because while developing it's a single process, I know what's going on, and it helps me a lot. And then on top of that, you have profiling, and it's really easy to profile your applications using Go. But profiling is specific to resource usages, right? To actually debug and introspect and understand what's going on, maybe troubleshoot a problem, you actually need to combine all of them, because every signal has its own use case. + +**Mat Ryer:** Yeah, something that you cover in your book, Bartek... And by the way, congrats on the book. + +**Bartlomiej Płotka:** Thanks. + +**Mat Ryer:** I don't get a kickback from it, so I can just say, it is available now... So if you want to learn about writing efficient Go, get it. In your book, I know that you mention this trade-off between features and performance. And we've kind of talked about this when you sort of build up tech debt as you go along adding features. Does this argument happen a lot, where you're making the case to not have a feature because of the cost it's going to have? And then what about like product people, or sales people, or whatever else they're working at, obviously incentivized really for the opposite thing maybe? How does it go down? + +**Bartlomiej Płotka:** Yeah, totally. It's kind of an everyday question almost, if we should invest in some amazing feature... And especially deletion is an interesting one, because we usually - at least in our ecosystem of like Prometheus, and the Thanos project, as well as Mimir, and Cortex, and Flare even... You know, all those kind of like distributed systems - or not distributed; like time-series database, or sort of kind of like storages for different observability data, we kind of store it in immutable storage. So once you kind of like try to delete something, it can cause problems, and it can cause costs. So that's kind of like the question - when we should add, how much of the flexibility in deletion we should add, and sort of those stories. It really, really depends, and what really helps is to have this budget of how much cost you want to kind of add to the system... Even like developing a feature is a cost; spending engineering time, and testing, and maintaining it. + +I really kind of try to recommend into using the same pattern for discussing how much you want to invest in efficiency of something, if we should make it faster or not, or we should maybe disable some feature because it was kind of slowing the system... This is exactly the same discussion as you would have should you add some feature, because again, it's just worth thinking about efficiency. Again it's still kind of some cost; it will be some cost of like maintaining, and so on. So everything, ideally - ideally, in an ideal world, you know the cost of it, so you know if deletion is... How much deletion requests will cost, kind of roughly CPU time, and if you have the capacity in your cluster. So this should be kind of like the core discussion to have, and if users are willing to pay for that, sure. But if, for example, you want to kind of minimize the cost, maybe you should disable this feature. It really depends on what user you are in, and what project this is in... But generally - yeah, speaking about data, numbers is kind of helpful. And treating everything as the whole package. Not just features as readability, and maybe complexity of the code, or API, or configuration, but also add efficiency on top, like roughly the cost, essentially - it just helps to make the decision. + +**Christian Simon:** \[36:16\] Yeah. And I guess it's what the business leaders in the end care about; I guess they want to kind of maintain that margin, and the closer to those numbers you can get from your system, from your production system, the easier you will be able to flag issues. If, for example, that deletion feature is for free, and they actually ramp up a lot of costs, I guess that that will kind of target your whole margin as a product, and I guess you want to know about that before it really turns it down into a negative. + +**Mat Ryer:** Yeah, the best performance improvement I remember I ever did, in my whole career - I made something about a thousand times faster by removing a time.sleep that I'd left in there by mistake... \[laughter\] They were thrilled. They were like "This is really fast now." I looked like a hero. Instead, I was of course an idiot at the beginning. So do you recommend that, put some time sleeps in early, and then you can take them out over time? + +**Bartlomiej Płotka:** Maybe it's a genius move, right? You just put them in, and then remove... Ta-daa! It's an improvement. + +**Mat Ryer:** Yeah. + +**Bartlomiej Płotka:** I mean, this is not a joke. Literally, I was hearing some stories from AWS, or some other service, and... I would need to dig it up, the source of it... But generally, they were having a service, and they added sleep to the service, to not give too fast expectations for people. Because they knew they want to implement additional complexity, additional features. Imagine they were doing, for example, like this database, and they want to add deletion later on, but they didn't yet, so that's why it's so fast. But they know where they will add it, so they added sleep. So once they added this deletion or some other feature, it gets so slow they have to remove the sleep, and now everyone's happy. + +So literally, those things are recommended, honestly, because if you are kind of like -- again, this is again this budget and goal thing; if you are agreeing with your customer that your service is responding under one second, having it respond in 100 milliseconds, if they pay for one second, sometimes it's not smart, because people kind of depend on it forever, even though you kind of agreed on one second. So those slowdowns on purpose - they kind of make sense sometimes, right? + +**Mat Ryer:** Yeah, but isn't that the kind of thing that gets leaked at some point, and then there's a big outrage? I don't know if this story is true, but I heard a story of like Uber checking your battery on your phone, and charging you more if your battery was really low. I don't know if that's true or not... But it's a scandal. Imagine if they said, "Yeah, this service is deliberately slow." + +**Bartlomiej Płotka:** I mean, yeah, let's not go there... But the AWS a story - I think it's fair; like, you paid for one second, and that's fair. They want to make this project sustainable. But what you mentioned is kind of like maybe less legal or less clear. But there are things where... How's it called...? Like when companies want you to buy a new phone, and they will put time.sleep into your keyboard, so it's kind of slow on purpose - things will happen like that. So yeah, that's unfortunate... No, we are to improve efficiency in this podcast, not add problems of efficiency. + +**Kemal Akkoyun:** Yeah. But I guess it's legit if it doesn't exceed your error budget, or if it's in your service level agreements, so you can like read the battery levels, or put the sleep in there and slow things down. + +**Bartlomiej Płotka:** Yeah, the problem is now you are buying a phone and you have a list of efficiency goals... Like, your keyboard will be always going from bottom in half a second. It's the same with SLOs, like how much granular those kinds of numbers can be. It won't be granular to your local library method in Golang, right? It will be kind of like very high-level. So that's a problem of the requirements really, and that's why every team should have some rough numbers, because you are developing your library, and the SLO is impacted. Is it your fault? Is it somebody else's fault? Like, how fast I should be? If one second is the goal, then your program is doing many things. My library on my team is developing like "Should it be -- how much of this second?" So those are not easy questions, but again, with things like metrics, with things like profiling, you could kind of find this number easier. + +**Mat Ryer:** Yeah, another valid use case for sleeping in code I've seen is literally to add a random sleep to sort of jitter requests around as well. And I think that kind of makes sense, because then it's about you're really helping... And as long as, like you say, it's within the budget, maybe it's okay. + +**Break:** \[40:49\] + +**Mat Ryer:** So how does Go compare to other languages and other technologies when it comes to observability profiling, and being able to actually introspect things that's going on inside the programs? Is it good? Are there good things and bad things? Anything you wish you'd change? Something you love? Christian, what do you reckon? + +**Christian Simon:** So I generally would say it comes with batteries included. So I think compared to other languages, when you install the Go version, that is the most recent one - like, everything is already on your machine. You have basically the runtime, it takes care, for example, of the profiling, and you basically just need to tell it to do it for your particular, either mini benchmark... + +I think in all languages you often rely on some external tooling, or kind of intercepting somewhere... And yeah, I guess that makes it kind of available for any program built on top of Go, or built with Go. It just sets a certain standard, I would say. + +**Kemal Akkoyun:** I guess I can jump in here... We mentioned the signals, right? For example, one of those signals are metrics, and Bartek and I, we maintain Prometheus client Golang library, which is one of the first Prometheus libraries, right? Surprise, because Prometheus is written in Golang... + +**Mat Ryer:** Good choice... + +**Kemal Akkoyun:** So from the beginning, the support for metrics for Go was a first-class citizen. On top of that, the Go runtime itself also provides facilities to actually make your programs more introspectable, or observable, right? One of them is profilers. And Go, from the beginning, actually had pprof, which is a project from Google itself, and it's embedded in the runtime itself, and you can actually expose this profiling data pretty easily. You can actually write, and then you run your tests with just additional certain flags. You can actually dump CPU profiles, memory profiles, or you can just like add a package and add an endpoint, and then you can expose a profiling pprof endpoint, and then pprof can scrape and collect profiling for that. This was there from the beginning, so profiling -- Go is kind of built with profiling in mind. + +We also mentioned metrics, and one other thing that we can mention - there is an ongoing PR and a discussion around structure logging, for example. Go has a standard logging library, but now they're trying to improve that, and there are certain controversial opinions around that... But still, I think as a community, Go always thinks about observability and how to make the programs more observable, more introspectable. + +**Mat Ryer:** So sounds pretty good then... Are there any downsides? Anything you wish it did differently? + +**Kemal Akkoyun:** \[43:49\] Kind of... I will jump again; I'm super-passionate about like profiling, right? Go itself has a profiler built in, and you can actually expose certain types of profiles out of the box, right? And they're super-useful. But the approach that you have with Go, for example - it's an interrupt-based profiler, right? And you are using certain operating system facilities, and I'm talking specifically about Linux in here. You need to set \[unintelligible 00:44:19.07\] and then you need register a call, and then the operating system itself is sending you some interrupts. And in those interrupts you actually dump the stack of your process itself. + +It's actually -- you're changing things a little bit when you want to observe the system itself. So I think this is a downside. And right now, with the improvement of our technology in different parts of the stack, you can actually collect these profiles without paying this cost attached to this. + +**Bartlomiej Płotka:** But I guess you could turn off those profiles. There's native profiles if you are a user of eBPF, or some other solution that gives you those information from other sources... I guess -- are you able to turn it off? Isn't that a solution? + +**Kemal Akkoyun:** You can definitely turn it off. But I already told you that I'm a big fan of doing things in production, right? I want to profile my applications on production, so I'm always thinking about in that case, right? How can I actually profile this process in production, without paying any cost? + +**Bartlomiej Płotka:** Yeah. I guess kind of like your -- you are profiling on production; you need profiling for production use. So that's your production. So that's the cost you're paying. And then you kind of account for that. But yeah, I know what you mean - the more it costs, your observability stack can sometimes cost more than your compute power on a normal application. So this is like very -- you need to be very careful what signals you are deriving, observability signals you are kind of getting off your application, because you can get crazy here. + +So yeah, things like metrics and profiles are my favorite ones for like efficiency. You can make them kind of cheap enough for it to be profitable still, but I know your point. And all those kinds of things that are happening in the profilers, either moving into kernel, or kind of like in a Go runtime, making them kind of like cheaper to run, and more accurate in the same time, are amazing things. And you know, in every Go release we have some improvements of that, so it's pretty sweet. + +**Mat Ryer:** Yeah, maybe we should chat a little bit more about eBPF. This is something that I see a lot, and I see people getting quite excited about it. Liz Rice, I think, has an eBPF book, and it's called... It's "Runtime security", I think, isn't it? + +**Bartlomiej Płotka:** Something like that, yeah. And you know, it works on security, and that was the main maybe hype for eEBF... But definitely, it's used much more. + +**Kemal Akkoyun:** There is another book, I think, "What is the eBPF", if I'm not mistaken? + +**Christian Simon:** Yeah, so that's also the first Google result I've found. + +**Mat Ryer:** Well, there's definitely a talk... We'll find out and put this in the show notes. But what is eBPF? What actually is it? + +**Kemal Akkoyun:** So I guess I kind of alluded that this is the new solution in the different part of the stack, right? eBPF is actually - yeah, Bartek is actually showing us the book... So you will add the link, I guess... + +**Bartlomiej Płotka:** Yeah. + +**Kemal Akkoyun:** So eBPF is a new technology in the kernel space. It was actually invented as filtering certain packages, but the acronym became a technology itself. Basically, it's a virtual machine and a verifier that attach to your kernel, and you write your C programs, or you can also write the programs in Rust, and you can target this virtual machine, and then the kernel can load these programs in your runtime when the program is run. So what it helps - it actually really opens up a lot of possibilities for observability tooling, security tooling, or network tooling. Since our concern is observability tooling, with eBPF - there are subsystems in Linux, right? One of them is perf, which is for collecting certain hardware measurements from your hardware, from your host machine, right? + +\[48:17\] Using perf subsystem you can subscribe to the perf events, and using those perf events you can actually capture the stack traces of your running processes. Then you don't need to actually register something; you just need to subscribe to this perf event, and you can aggregate, and this happens outside the process, right? This happens in the kernel space. So it doesn't actually change what's going on in your program itself. We cannot say that there is no cost. Of course, there is a cost. Of course, it's changing stuff, because you are running some program in your CPU; but compared to what you get from interrupt-based profilers, this is so minimal, even negligible. + +**Mat Ryer:** Are you excited about eBPF? + +**Kemal Akkoyun:** I am totally excited. I'm going crazy about it. I'm also working as a Prometheus maintainer, and mostly working for client Golang; it's an instrumentation library. And recently, I gave a talk in Prometheus Days, and I was focusing on how we can do stuff using eBPF without instrumenting the code, and mixing and matching those things. And on top of that, I'm helping to build an open source continuous profiling project, especially for Go. It's called Parca, and we are using eBPF to actually profile the processes. So I'm all sold out, I don't know. + +**Mat Ryer:** You're all-in. + +**Kemal Akkoyun:** I'm all-in + +**Mat Ryer:** I think it's going to be a popular opinion, but fear not... It wouldn't be Go Time if we didn't explore your unpopular opinions... + +**Jingle:** \[49:51\] + +**Mat Ryer:** I promised last time that I would kick us off, because I forgot mine... And I've brought two. I've brought a serious one, and a very serious one. And my very serious one is, I think from now on the past tense of screenshot should be screenshat. So we should say like "Oh, did you send that thing to Bobby Booksmith?" "Yeah, I screenshat it to him." That kind of thing. If you've got no objections, we can probably sign that into law now... + +And my other one was the optional second return arguments that you sometimes have in Go, like when you're checking a map, or doing type positions - I don't think they should be optional. I think you should either have to explicitly use an underscore to say "I'm ignoring this", or you should use it... Because -- I just don't think you should ever not use it, basically. I don't know if they'll be popular or unpopular... But we will find out. Don't forget, we test these on Twitter @GoTimeFM. So you can vote on a poll and decide whether this is indeed popular or not. Anybody else got popular and unpopular opinions? + +**Bartlomiej Płotka:** Yeah, I can do that. So my unpopular opinion is that if you see in some software -- if I see in some software three extremely generic abstractions, and they usually are together, things like input, processor and output, source, function, exporters, collector, processors, publishers, in most cases those mean troubles. That's my opinion. If I see this, I'm running in a different direction, or try to find a replacement software. + +I might be kind of exaggerating, but one kind of example is like generic data pipeline, or pub/sub systems that support every data format, and maximum flexibility, and you can configure thousands of intermediate steps as the processor, or between input and output... Or maybe, you know, sometimes I saw this Go interface called processor, with one method - do. And bytes argument. And those things are crazy, but I saw them... + +\[52:17\] And the reason why it's bad is because the best abstraction is generally opinionated and small, while the complexity of implementation can be higher, because it's hidden under this abstraction, right? And there is a saying even - keep interfaces shallow, implementation deep. And those things with over-generic abstraction happens, but they kind of cause so many problems. It end up your program in YAML at the end, because there are so many kinds of configuration for this inputs, and processors, and outputs; it poses limitations to implementations. It's simply inefficient. + +At the same time, the standard library kind of functions - that's kind of another popular opinion, I guess, but standard libraries sometimes are inefficient, because they're very generic. It has to check different edge cases that you might not have. For example, parsing integer from string, if you don't have negative numbers, your function can be much faster, if it's, of course, in the critical path. So things like that... + +Anyway, I was just seeing so many failures, so many projects abandoned even, because they were following these very over-generic abstractions... And I want you to be cautious here. That's my opinion. + +**Mat Ryer:** Very nice. By the way, some of my best friends are YAML programmers, so let's not... You know... + +**Bartlomiej Płotka:** I mean, if you like it sure. + +**Mat Ryer:** Is it not just a failing of the naming? It's just that we struggle to name it? Or do you mean about the input and output types as well, like you want those to be very specific, ideally? + +**Bartlomiej Płotka:** Yeah, it depends on the case, but generally, you want to have naming, but also underlying kind of configuration options to be opinionated, to be limited; there has to be kind of like defined abstraction. Otherwise, it's over-open, so it leads to abuse of the system as well. So if you want to have any arbitrary input that works with any processor, you end up with like reading PDF, filtering using, I don't know, a processor made for raw bytes, and then outputting as a Helm chart. I don't know, like something crazy like that. And people will do it, and then claim success... It's like a waterfall of different problems if you are naming it kind of over-generic, but also having the interface very generic. + +**Mat Ryer:** Yeah, it makes sense. I understand how it happens, because you want to be so flexible, and really smart, and you want to take on the pain, so that it's easy for people to use... But I kind of agree. Like, if you can be more specific, then do. That's a pretty good one. We'll test it and see, Bartek. And if it's not unpopular, you are contractually obliged to return to Go Time and do another episode. I don't make the rules, but... + +**Bartlomiej Płotka:** Sure. + +**Mat Ryer:** Okay. Do we have any other unpopular opinions? + +**Christian Simon:** Yeah, so I have an unpopular one on the Go execution tracer. So while it is very colorful, and I think you can learn a lot how maybe your program is running, I think I've never found the problem, and I always started tracing when I was just desperate. And so my unpopular opinion is instead of starting it, go outside and do a walk in the park, and then maybe you see the missing error check, or whatever it is. + +**Mat Ryer:** Hey, that's interesting... Is that unpopular, Kemal? + +**Bartlomiej Płotka:** Not in my book, I guess... + +**Mat Ryer:** Okay. Well, we'll test it, Christian, and we'll see. Kemal, did you bring an unpopular opinion today? + +**Kemal Akkoyun:** I brought a bunch of them. I don't know where to start, but I guess maybe I can just keep it simple and focus on the things that we talked today... + +**Mat Ryer:** It'll keep Bartek happy... + +**Kemal Akkoyun:** \[55:50\] Yeah. This one for Bartek. So I think Go should give certain APIs to control memory for certain use cases, right? Go runtime does a lot of assumptions for us for the memory safety, and it gives us guarantees, and it's really nice with the garbage collection... But sometimes, you really want to reduce the allocations. There are certain ways to do that, but it would be really nice to have some granular control on certain memory parts, so that you can optimize your program for your hot path. Because as far as we see with the Go programs, it's mostly about allocations and the GC pressure. + +I know there are some improvements recently. Actually, there are some knobs that you can turn and adjust your garbage collection. On top of that, there is this PR proposal about the Arena support, but still, it would be nice to have more control. I think Bartek is gonna like this, because I know for a fact that he made a couple of packages that are on this, to actually make this happen. + +**Bartlomiej Płotka:** I feel like our opinions were not super-unpopular, because I think I agreed with all of it... And I think there should be some legal agreement we are assigning with you, Mat, that those words cannot be used against us, or something like that. + +**Mat Ryer:** Right... \[laughs\] + +**Bartlomiej Płotka:** But yeah, I would agree with Kemal, Arena is especially good, but I would even like bring a Rust ownership model to like a portion of your Go code. So you could strictly control their scope of living kind of lifecycle of your variables, of your data... And maybe only for 5% of your code, because this is the critical path that matters. And then use garbage collection for something else. I kept looking -- I mean, people look as crazy if I say that, but I think yeah, I would even go further as that. + +**Mat Ryer:** Hm... Yeah, so it kind of make sense then... Like, most people won't have to know about it or even think about it. But when you're doing efficiency work, when you really do want to control it and optimize it, that sort of makes sense. A bit like how you can have Assembly in your Go code, I guess... Because that's a bit unusual, to have assembly code in there; people do it sometimes for very good reasons, but it's a trade-off between - this is now very difficult to maintain, especially for Go programmers, who typically don't write much Assembly... And same, then - yeah, for those that need it, they'd be able to use it. + +**Bartlomiej Płotka:** Yeah. I think we're going in this direction though, so it's good. Arena already is claiming some kind of part of your memory, where it's outside of garbage collection. So it's a precedence. Once that is done, we can go further, deeper into bad things, but also nice for efficiency sometimes, yeah. + +**Mat Ryer:** Well, these all do sound quite popular. Does anyone want to have a go at a properly unpopular one? + +**Bartlomiej Płotka:** No. Before the legal contract, no. + +**Mat Ryer:** \[laughs\] + +**Kemal Akkoyun:** I can go... + +**Mat Ryer:** Kemal, come on. Let's dig deep. Find an unpopular one. + +**Kemal Akkoyun:** Okay. I think the Go team should be an independent third party, instead of something that only Google people actually control. There are a lot of other kinds of stakeholders actually depending on Go. For example, one of the big ones is the CNCF, the Cloud Native Foundation. We work on the open source project that actually donated to that foundation, so... It would be nice to have some representation in the Go team to adjust things, right? + +**Mat Ryer:** Hm... Interesting. I know that that has been a subject before, and I think it is quite a controversial one... But let's have that chat on Twitter. We'll certainly put that clip of you, and then vote, and then that's it; that's really how we do our observability data collecting... + +Excellent. Well, thank you so much. This was great. And we should do this more. You know, we touched on profiling. We really didn't go deep on profiling, and I think that could be also great for a future episode. But thank you so much, Bartek, Christian, Kemal. Absolute pleasure. And thanks for listening. Thanks to our live listeners. Remember, this goes out live, and you can see us on YouTube... And also, this is edited nicely as well for the podcast. So if you're listening to it on a podcast thing, good for you; but there's some gold that was chopped from it... I suspect one of my unpopular opinions will have been cut out... So if I said "I've got two unpopular opinions" and then I only said one, you'll know that one was cut out. But we'll see... Thanks very much. See you next time on Go Time! diff --git a/Making the command line glamorous_transcript.txt b/Making the command line glamorous_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..7401f8fa7fbdddc730f6075a53d124bb73425e81 --- /dev/null +++ b/Making the command line glamorous_transcript.txt @@ -0,0 +1,673 @@ +**Jerod Santo:** So we have Toby Padilla here from Charm. Toby, thanks so much for joining us on the show. + +**Toby Padilla:** Thank you for having me. + +**Jerod Santo:** Charm - y'all build tools to make the command line glamorous. I landed on your website and I thought "I am the target audience for these folks." Really cool stuff. + +**Toby Padilla:** Thanks. Yeah, we like to be glamorous. + +**Jerod Santo:** How many people are in this Charm group? Because I just reached out to all y'all and said "Hey, come on the show", and I think there was a handful of people that responded. There was some talk on your side, "Who should actually be the representative?" and it ended up being yourself. But who else is involved? + +**Toby Padilla:** There's six of us. There's myself, there's the other co-founder, Christian Rocha. There's Muesli, who's been with us since almost the very beginning. There's Carlos Becker, who's the GoReleaser guy. There's Ayman Bagabas. He does a lot of the Soft Serve development and a lot of our DevOps stuff... And then we've just hired Julie Zhang, who worked with us at BetaWorks. She was a VC over at BetaWorks, and we brought her onto the business side two weeks ago. We also have Pengu. Gotta shout out to Pengu; he's our intern. Amazing open source guy, friend of Muesli. + +**Jerod Santo:** Shout-out to Pengu. So I know Carlos Becker... He was on Go Time last year. + +**Toby Padilla:** Yup. + +**Jerod Santo:** And ever since then, him and I have been working together to get him a Go Time T-shirt shipped to his location, because -- + +**Toby Padilla:** He mentioned today that he hasn't got his T-shirt yet, so yeah... \[laughs\] + +**Jerod Santo:** \[04:11\] No... But we've tried over and over again. I think he's in Brazil, or somewhere... + +**Toby Padilla:** He is. + +**Jerod Santo:** ...and you know, shipping things is hard these days. We've just actually decided to start from scratch, hopefully try a second time... It just got stuck somewhere in customers... Who knows. Who knows what happened. But yeah, Carlos and I have been talking for a while, ever since his Go Time appearance. + +**Toby Padilla:** Carlos is a good guy, we're excited to have him on. He actually started as a contributor, and a person who is just kind of around the Charm project, and then we brought him on as a contractor. Around four months ago I believe he came on full-time. + +**Jerod Santo:** So y'all have a bunch of tools... It's kind of hard to decide where to start and which one to go into... And that was one of the things - when I first landed on your site, I thought "Oh, here's an easy Changelog News link-up. And then I started scrolling and I'm like "Holy cow, there's a whole bunch of different projects here. Which one do I link to?" And then I was just like -- I'm into command line stuff, so I was like, "Well, we just need to bring him on the show and talk in more depth." + +But one thing I've already keyed in on... I'm sure Adam's keyed on it, because he's all about the Founders Talk... You said "my co-founder", and I'm looking at a whole bunch of open source tools, and I'm just very curious, how do you guys approach Charm? You said there is an incorporation. So tell us about, I guess, what is this entity. + +**Toby Padilla:** So we are actually a seed-funded startup, and we started because we wanted to make the command line more fun; we wanted to make it glamorous. And that sounds insane, but it's something that we wanted to do. But it actually means a few things to us. Christian and I - we're at BetaWorks together, and like I said, Julie was there too, and we were doing various projects for them... We have a long back-story with them; we've done multiple startups, I was VP at TweetDeck, which was a BetaWorks-founded company... So we've been in the BetaWorks space for a while, but we wanted to do something together, because the entire time we've known each other, even when we were doing other startups, we were exchanging Vimrc tips, Bash tips, all of these things... We were like "You know, we're really passionate about the command line. Why don't we see if we can turn it into a company? It needs a rethinking, it needs a rebranding, and we wanna just work on it. We want to do this. It's a fun thing to do." + +So what we wanted to do was a) make it more fun to work on the command line, b) bring modern product thinking to the command line. If you look at the history of the command line, it was very much rooted in 1970's and '80s mainframe technology. It's designed around multiple users per machine; it's not really set up for the internet... It doesn't have a ton of built-in encryption; it does, sort of, now maybe at the disk level or something like that... But we wanted to modernize it. We wanted to make it look nice, because it doesn't always look nice, right? Sometimes it looks really raw; and looks can be important, so we wanted to -- Christian is a designer by trade, and he's a really great designer actually, so we wanted to sort of apply that to the command line. And then we wanted to make it easy to take all of these concepts and for us to build tools with it, but also to have other developers build tools with it. + +So we built a bunch of stuff. You were referencing -- there's a lot of projects, right? And it sometimes can be hard to talk about everything, but we kind of have two separate areas that we focus on. One is apps, and one is libraries. And the apps are focused on solving things that we see are big problems on the command line; use cases that we think are important. For example, Glow is meant to solve reading documentation on the command line, or discovering documentation on the command line. So that's a first-party app that we've built. + +An example of a library is Bubble Tea. We built Bubble Tea to build Glow, and now it powers all of our apps, but it also powers hundreds of apps that people in the open source community have built. And Bubble Tea is a TUI framework based on the Elm architecture. Christian was a big Haskell guy, loves functional programming; we all kind of came into Go, we'd been doing it a bit before we started Charm... We were like, "Let's bring other ideas into this space." So one of the very first things he built was this Elm-based TUI framework, so we could build some really cool, exciting command line stuff. + +\[08:12\] We've got a bunch of other projects as well. The libraries are focused on really the front-end, so they're sort of -- Lip Gloss, Bubble Tea, Bubbles, Harmonica, are all frontend apps. They have interesting names, I know. But then we also have backend apps. So Charm is a bunch of things; it's not just the company; it's also a tool, it's a set of frameworks, and it's the server; we make everything self-hosted. So Charm libraries are based on storage, data storage, and encryption, and identity, things of that nature. And then on top of that, recently we've been really focused on SSH and pushing SSH forward... Because we think that's one of the most exciting, interesting, powerful technologies on the command line is SSH. So we're doing a lot around SSH identity, SSH access, remote access via SSH... So we've built this thing called Wish, which is a framework for building SSH tools; you can wire Bubble Tea apps up into Wish, and so you can SSH into something and get a Bubble Tea interface and have it do whatever. + +We also have Git middleware inside of Wish. One of the things that inspired a lot of what we're doing is the Git SSH interface; kind of how GitHub does this. So you push, and they know who you are based on your public key; you add your public key to their server. So we built Soft Serve, which is a TUI-based self-serve Git host. And that, again, is using Wish and Bubble Tea for the interface. + +**Adam Stacoviak:** Wow. Blown my mind here; it's just so many toolings... I love the direction. It's crazy just to see somebody come out swinging like this on the command line. You almost feel like it's boring, in a way, from the outsider's perspective, those who think like "Oh, the command line is just kind of a drag..." But you know, this is really cool tooling. I'm super-impressed. + +**Toby Padilla:** Thank you. That is sort of a perception that's there, and it's partially true and it's partially not true; it's difficult... One of the reasons we've built this is because Christian and I have been on the command line each for like 20 years apiece, and neither of us considered ourselves experts... Because the learning curve for this stuff is just a sheer cliff wall. You spend your life learning about that... And that shows you the power, but it also scares a lot of people off. So it's like, "Command line? No. I don't want anything to do with it." And then a lot of the tools are really dated. Some of the dated tools are excellent; I'm a huge Vim guy. Not the most newest editor out there... But still, I use it every day. + +There's kind of -- on the command line you get these things that look like lifer technologies, where it's like you invest your life in learning this thing, and then it'll pay off in the long run. I spent a lot of time moving to Vim from TextMate, or something like that. And it was difficult at first, but I knew "Suffer through this and it'll pay off 10 years from now, 20 years from now, when you're SSH-ing into a server and you're just like kicking ass, editing text", right? So that's one. Or piping commands together; learning basic, Unixy workflow stuff there. All of that stuff - it's a big investment, but it pays off in the long-run... Versus learning a proprietary stack, or even a proprietary bit of technology. I'll use a web example for this; and this is by no means to bash the project, but D3 - I always look at D3 and it's like "Hey, you can do awesome stuff with D3 in terms of visualizations." And then I start learning it and I'm like "This is so much to learn." It's so much to learn, just this like very specific framework, in this very specific spot. It's not gonna be knowledge I can apply across the computing world. So when we think about command line stuff, doubling down into your knowledge there should be applicable across a whole range of tools. + +**Jerod Santo:** So I tend to agree with you, but I wonder how big the audience is for tooling like this. I think of the command line as a subset of technical people, or maybe developers... Let's just call them tech-oriented people; of those, there is a subset who prefers the command line. And of those, I am one, and many of our listeners are; that's why they listen to the Changelog and they hang out on Hacker News and Changelog News, and GitHub, and these things... But when you're building -- did you say you're VC-backed? Is that what you've said? + +**Toby Padilla:** We are, yes. + +**Jerod Santo:** \[12:16\] Okay. So when you're talking to venture capitalists and try to make a business case for this type of a company, don't you butt up against that feeling of it being old and very niche, and saying "Well, we're gonna bring it to the future." But is the future interested in the command line? I personally am, so I'm easy to sell on the vision, but I'm wondering if other people are harder to sell on the vision. + +**Toby Padilla:** It can be like that... There are some good examples of doing what we're talking about doing. So GitHub is a great example of modernizing the command line. They took Git, which is an extremely powerful, and let's say edgy tool; it requires a lot of knowledge to do it right... And they brought a bunch of what we call modern product thinking to this. They brought social interaction, they brought a nice web interface; they invented some sort of terminology in use cases, like PRs, pull requests. That's a GitHub thing, that's not a Git thing. + +So Git was designed for the Linux kernel and mailing and patches and all these things; really small audience to do that kind of stuff. But GitHub mainstreamed this, and almost every developer - not every developer, but lots and lots of developers are on GitHub because they kind of did what we wanna do. But then they kind of stopped. They did a lot of innovation, they've been acquired, they're part of Microsoft, they're really focused on GitHub Actions and stuff, which is really neat... But it's an example of building a huge, multi-billion-dollar business off the back of a command line tool. So we think that we can do this. + +If you look at DevOps and things of this nature, the command line is really important there. SSH keys are really important there. The workflows for deployment are often in these developers tools that live, in a sense, on the command line. So there's maybe not a trillion people, but especially myself, coming from the consumer internet side, music, TweetDeck, things like this, where it's like "Oh, you've gotta get a billion users", or whatever... We don't necessarily need a billion users. We're happy with 30 million hardcore developers, and ultimately, our business model is to sell into enterprises anyway, and solve enterprise use cases and developer workflows inside of companies. So we think that getting that developer mindshare will help us in that business sense. + +**Jerod Santo:** Well, shout-out to your designer, because when you land on Charm.sh, you accomplished what many people fail at, which is to immediately have a recognizable and distinguished and interesting look and feel. You mentioned the somewhat -- you didn't call them silly, did you? You said you have lots of different names... I can't remember the adjective you used; I don't wanna -- + +**Toby Padilla:** We like to be unique with our naming, yeah. + +**Jerod Santo:** Okay, so the unique naming convention... Tell us what inspired this. You've got Bubble Tea, Lip Gloss, you mentioned Harmonica, Soft Serve, there's like an ice-cream theme going on... Kind of like a desserty theme going on... Talk about that for a minute. + +**Toby Padilla:** So Christian is the co-founder and designer, so he's doing all of this design. 3D backgrounds... So you'll see a lot of 3D models there, which people don't necessarily put in the command line design sensibilities, right? From the get-go, we wanted our brand to be "My Little Pony meets Fist of the North Star." And I think that we're trying to accomplish that. We want it to be less intimidating than the command line has traditionally been, but even more powerful. You can do just this amazing, crazy stuff with this. But by making it friendly, it's kind of changing the culture around the command line. We don't wanna be 1970s MIT. We don't want to be like Gravy or Unix-y those guys did a lot of amazing stuff, but it's 2020 now. So updating the language that we speak to people with is a big part of the project. And we actually think about culture a lot, both in terms of our product, our branding, but also the team, and how we run it. We like to do things a little differently internally as well. + +**Adam Stacoviak:** \[15:58\] Do you see, I guess, when you get to a certain point, when you SSH let's say into your production server, do you envision where this absent tooling that lives -- I'm not sure how you divide them, because I'm still learning the product graph you have... But do you envision -- because I'm looking at Wish, and "Make SSH apps", and you've got the parrot doing the dance, like you can SSH into a certain server... Do you imagine that someday, somebody would replace or make their production server easier to navigate, and maybe even Linux tooling that you have to constantly google, or whatever, kind of be front and center? Is that what you hope for at some point, where you enable a future generation to build on top of the tooling you currently have, basically? + +**Toby Padilla:** Exactly. That's what we're doing. I would say the best example of a Wish app is Soft Serve. So if you SSH to git.charm.sh, you will go into our Git repos. So this is a Soft Serve instance, anyone can access it if you have a public key. If you don't have a public key, it'll ask you for your password, and I think you can just enter it and it'll let you in. So custom SSH, right? We're letting everybody in; come on in. We're seeing you for the first time? Come on in. This was something we could do with Wish, because it's custom SSH servers sort of divorcing the user accounts from the machine in the application. + +SSH is really powerful, because you know who somebody is. Everything is, obviously, encrypted, so that's nice... If you think about setting up a server today on the web, you get HTTPS, and you have to manage those certificates. Maybe CertBot manages it for you, but you still have to set that up; you have to do a DNS entry, or whatever. You have to think about HTTPS and your domain name. With SSH you don't do that. + +One of the things Wish will do is actually make the server keys for you, just when you start it up for the first time, so you don't have to think about that. You should store them someplace secure, but it will generate them for you. But you just fire it up and you have encryption. And you also have identity, in a sense. Because people, when they come in to SSH, are presenting their public key to you. So you can say "Okay, this public key is attached to this identity." And we've got really far down the rabbit hole with that with Charm, where we allow you to link multiple public keys to one identity, and then we actually do a bunch of encryption stuff where we make some metric encryption keys, encrypt them for all of your public SSH keys, and then store those encrypted on the server. We don't store private keys at all; so your SSH private key stays on your machine. + +**Jerod Santo:** Sure. + +**Toby Padilla:** You pull down all these symmetric keys, decrypt them with your private key, and then you can decrypt all of this encrypted data that's been pushed up to the Charm cloud. + +**Adam Stacoviak:** So end-to-end then. You're end-to-end. + +**Toby Padilla:** Yup, very end-to-end. We don't want your data. Data is a liability. It's not our business model; we don't want it, we don't wanna know what you're doing, we don't wanna see what you're doing. We are very big on privacy. For instance, we don't put analytics into our apps. You privacy is more important than our business metrics, period. + +There was a year period we didn't know how many people were using Glow, because we didn't have any Charm cloud functionality. It was just a markdown reader that was not networked. So before we invented the Charm cloud and all this encryption stuff, Glow was simply a Markdown reader; you'd pass it a Markdown file and it would style it. It would style it with a style sheet. So that's what Glamour is. It's a library that styles up Markdown according to an ansi style sheet. And it will also detect your background; so if you have a dark background or a light background, it'll give you the right colors there. + +GitHub, by the way, has used Glamour in their official command line client to render Markdown. So if you use cli/cli on GitHub, all of the Markdown rendering is happening with Glamour. But we didn't have metrics, we didn't have analytics, and that was fine, because we wanted people to have privacy. We also don't think that's even a good way to develop product. You should develop products as the product owner or developer because you have a vision for it, you have a need you're trying to solve, and you can just bank that out. There was no analytics or A/B testing that went into Git, or Vim. This was -- somebody had a vision and they made this incredible tool. And even going to the consumer side, all of my favorite products have been created by usually one person. + +\[20:07\] I went to work at TweetDeck because I loved TweetDeck. And Ian, the TweetDeck founder, just made this thing. He was just like "I wanna look at tweets in columns. Make an app to do that." LastFM - RJ was a college student, and he invented scrobbling as his university project. He was like "I wanna track all my playcounts. That blew my mind. I was a huge music fan, and I was like "I've never seen anything this cool in my life." Delicious. Joshua is like, "Hey, tags. Freakin' cool. Let's just build this thing." + +So you see these amazing products come out, not because of analytics, not because of studying user behavior, but because "I have a vision that I need to make this thing", and that to us tends to be the most exciting stuff to build. + +**Adam Stacoviak:** We'll obviously go deep into all the different product there is, but going back to Jerod's question with BetaWorks... Obviously, BetaWorks has a track record, so I think there's potentially some inherent de-risking of maybe putting you in a team together to do stuff like this... But what is your trajectory then? How do you passionately go into this world - very well done, by the way, but how do you passionately go into this world thinking out the other end something commercial will come to play? Something valuable will come to bear of this. You mentioned Delicious - those are great examples of just scratching an itch, essentially. How will this itch turn into something commercial? + +**Toby Padilla:** Well, we were lucky in that we were at BetaWorks, because they're all about just making cool stuff, and seeing what happens. They never go into something saying "Let's turn this into a business at first." Having success is important, but it's always like "Let's build something interesting and see where it goes." Technology is great about that. You can invent stuff that doesn't have a clear path to monetization at first, and then it turns into that. Again, Git wasn't clearly monetizable, and GitHub didn't monetize it for a long time. But eventually, it turned into a real business that was worth billions of dollars. So we came from this VC firm that was very freedom-loving, and very creative, and very art-loving... So we're like "We wanna build stuff on the command line." And John Borthwick who's the CEO of BetaWorks, he's like "Toby, Christian - you guys have done a lot of stuff for us. We believe in you. Let's do some pre-seed and get started." We also had the support of Alex Chung, who's the Giphy founder. Giphy is another BetaWorks investment/incubation project. + +So Alex is another one of our angels... So just having this crew of people who had worked together before, building experimental things... Giphy was an experiment. It was a weekend experiment that Alex built. He's like GIFs are cool; let's make a GIF search engine. And then he figured out how to monetize that. But that was sort of the history and the DNA of the early team and investors. + +**Adam Stacoviak:** It's good to get that trust right away though; to have that "Hey, you've done some good work here. Let's do that." So you said you're a pre-seed... Is that right? + +**Toby Padilla:** We're seed now. So we've raised a few rounds of pre-seed. We closed our seed round in December of 2020. So our seed round was led by Cavalry; they're a German-based seed fund. They're awesome. We just clicked with them immediately. I have a fairly extensive investor network in Europe, because LastFM and TweetDeck are both European startups, so we kind of knew them through that network. BetaWorks, one of our lead investors... Also, Fuel Capital. They're a Bay Area-based investment firm; they're really cool. Chris Howard over there has been supportive of us from essentially day one. So he was in the pre-seed and seed round as well. Kevin Carter - he's in all of our rounds as well; he's a really great guy. And Tokyo Black, the Looker founders, and Ben Porterfield, they've been investors. We have a bunch of other angels as well. + +**Adam Stacoviak:** Gotcha. What's the total funding so far? + +**Toby Padilla:** Total, 3.6 million. + +**Adam Stacoviak:** Okay. You've got a burn rate, you've got a /open page on Charm.sh? I love that, by the way, when people have /open and it's like "Here's all the things we're doing", because they're so just transparent about what they're doing. + +**Toby Padilla:** We are transparent. We're not uber-transparent. We like surprises. It's like, "Hey, look at this cool thing that we've built. You didn't know about it. Here's something new." Our burn rate is low. We have six people. + +**Adam Stacoviak:** \[24:07\] Gotcha. + +**Toby Padilla:** The team is a lot like code - the less code you write, the better. The fewer people you hire, the better. Don't bloat your team. Don't increase your burn. Hire the right people. The way we hire is we don't do technical interviews. We find people we want to work for us, and then we try to convince them to join us. And the way we find them is they're often just contributing to us. And it's like, "Okay, here's some of our top contributors. Here's somebody who built something really cool with our technology. Can we get them to join us?" Or it's somebody else that we know through our network. + +So we're very proactive in our hiring. So when you find these people who have really demonstrated the ability to make something from zero to whatever, and ship it and support it and make it successful, you bring them on. One Carlos or one Muesli is worth 50 non-Carloses or Mueslis. GoReleaser is an amazing technology. We were quite excited when Carlos agreed to join us, or even talk to us. We were like "We're fan boys. We love it. We use it all the time." So it's really cool to just proactively seek out the right team members. It means you can keep the team a lot smaller like that, too. + +**Adam Stacoviak:** What that tells me though is like -- when we have this listening audience paying attention to these show we do, and we ask a question like that, we're not just probing, "Hey, what's your actual burn rate, Toby?" I don't really care, necessarily. I kind of wanna know, but what I wanna hear is the wisdom behind it. Because it's one thing to be able to produce this kind of tooling and have this kind of inertia and this innovation, but it's another thing to commend it with wisdom. + +We can almost respect and appreciate what you're doing way more because you have this innovation side of you, but you're also doing it with discipline. You're not just like, "Let's just throw money at it, hope it works. Let's just hire everybody who's ever creative", like some companies might do. Or you see some companies just acquire a bunch of talent, and it's like "Wow, they're just vacuuming up everybody." Whereas here, you're doing it with a bit more discipline. I think that's an admirable trait, honestly. + +**Toby Padilla:** Thank you. And actually, even speaking to the burn rate - we need to keep it low, because we like time. Time is innovation. We need time to innovate, to throw things against the wall to see what sticks, to iterate on ideas, to try new things. So think very much like an R&D project, where you can't have these really short-term time horizons where you need to raise the next round next week, or whatever, because you're not going to do the right things; you're not going to make the most innovative thing that you could possibly make. Sometimes that takes a little while. You don't ever wanna go heads down for two years, and go on these secret missions. Everyone has a tendency to do that, myself included. It's like, "Oh man, I'm gonna make my Magnum Opus. See you in two years, everybody." That's always a temptation. But you mix short-term thinking in that and you try to come out. + +So we did a long development cycle on this. We did a recent release with the self-hosted Charm Cloud, Wish, Soft Serve... A bunch of stuff. That was like 11 months in the work. And that was quite frankly as far as we should ever go without having launched anything. And it was starting to get to the point where it's like, "Okay, we're just iterating, and iterating, and iterating, and we need to get this out there into the real world." So you want time to try exciting, fun, new things, but you don't wanna bury yourself under an ocean and try to swim up, because that's just not gonna be good. And it doesn't make good product that way either; you need to get it out there in the real world. So finding that balance requires -- it's helpful to have a low burn rate, so you have some freedom. + +**Break:** \[27:45\] + +**Jerod Santo:** So let's talk SSH. You brought up Wish earlier, we talked about how you can SSH right into git.charm.sh. for what is kind of a tabbed interface, with a menu, of your different projects. I'm assuming this was all built custom by you guys to provide this interface. Before we get into how to build these things, are you imagining a world where a bunch of these SSH apps exist as an alternative -- almost like an alternative view from your website, sort of? I mean, it's its own thing, but it's a content viewer, which a web browser also is... I'm thinking about Changelog.com and how we have news and podcasts and search and things on our website, but we could have a cool SSH interface where you're doing similar things... And I'm just curious - are you imagining a world where a bunch of SSH apps exist, that are kind of like links customized in terminal browsers? Or obviously, GitHub used theirs for you to do Git interactions, to use the Git service, so different kinds of services... I guess just start by saying what do you see people building with SSH apps? + +**Toby Padilla:** So SSH can do a lot of stuff, and what you're looking at there is a TUI. So it is a menu-ing system, it's to browse through. One of the nice things is that it keeps you on the command line. So you're gonna git clone something from the command line. So going back and forth between a web browser and the command line. + +One of the innovative things that GitHub did was this Flash-based clipboard of the SSH clone URL, remember? And I think maybe that's natively supported in the browser, but it wasn't at first. + +**Adam Stacoviak:** I remember that, yeah. + +**Toby Padilla:** So that was just a way to get something from a web browser into the command line. So there's some value to being on the command line. But I think there's a few things that can happen with SSH. One, like you alluded to, it's a Git API. You can start to do really interesting things by layering -- you add the Git middleware to Wish, and we give you all the hooks for when projects are pushed, and when pulls are happening, and how authentication can happen... You can start to build applications around that. + +For instance, Soft Serve is configurable by Git. So when you first run Soft Serve, the very first thing you do as the host is clone the config repo. So you clone the config repo, and it has a Markdown file, a readme file, which is what you see when you first get into git.charm.sh. And it has a yaml that lets you configure access, customize that menu, add collaborators, add their public keys, all of these things. So all of that happens via Git. So you can imagine other configurations happening like this. + +\[32:12\] We're also about to release a version of Soft Serve; we haven't released it, but probably in the next couple of days... That adds non-interactive SSH functionality. So TUIs are great; they're great for discoverability, they're great for having some kind of UI on the command line for remote controlling various things... But on the command line, the real spirit of it is composing things; so piping the output of something into something else. And it turns out if you run SSH with a command, by default, it does not issue you a TTY, which means that you do not get any interactivity to it. You have to put -t to get in. So if you ssh -t git.charm.sh soft-serve, you'll go straight into Soft Serve. If you don't add the t, right now it doesn't do a whole lot. The new update that we're doing, you can say ssh git.charm.sh soft-serve/main.go, and it will pipe out that Go source code directly to your command line. You can add the -s flag (I believe), and it will style it up. It'll do syntax highlighting on it. + +So at that point, you can start baking these things into your command line workflows. You can start exploring grepping things, manipulating the output of these files directly from a Git repo into other command line tools and pipelines, which we think is a really interesting idea... Because there's not a great way to do that now. Yes, you can curl something if GitHub has the web interface for it, but this is very discoverable, and it's very native to the way that the command line works and the way Git works. + +**Jerod Santo:** So the other thing that's interesting about this particular TUI - which, just for the listeners' sake, that's TUI, terminal user interface... + +**Toby Padilla:** It's actually text. + +**Jerod Santo:** Oh, text. + +**Toby Padilla:** Text user interface, yeah. + +**Jerod Santo:** Well, thank you for correcting me. I don't even know what it is; I just assumed it was terminal. + +**Adam Stacoviak:** Good guess... + +**Jerod Santo:** Because where else are you gonna look at your text? + +**Toby Padilla:** Why wouldn't it be terminal user interface, yeah. Absolutely. + +**Jerod Santo:** I'm ready to rebrand that sucker... No. So the text user interface is -- at the command line there are certain things that you come to think about, like pagination, or piping things into Less or something, and there's standardized key combinations that usually work... Of course, usually depending on which variant of Unix you're on, or whatever. But here you're kind of reinventing a few things. You actually have tab-based selection, which navigates between like a menu item and a content item, and I'm assuming that stuff was invented by you all inside of the build text-based UIs tool, which - I can't remember which one that one's called... + +**Toby Padilla:** That's Bubble Tea. + +**Jerod Santo:** That's Bubble Tea, thank you. There's a lot of tools here. + +**Toby Padilla:** It's Bubble Tea, and Bubbles is a selection of components built for Bubble Tea. So some of that stuff is a bubble. It's a bubble that anyone can use. Some of it is custom components that we've built for this application. + +**Jerod Santo:** Okay. So I guess what I'm driving at is if people are all adopting Bubble Tea at least in order to build some things, we can expect certain common paradigms inside of a Bubble Tea-based TUI, right? + +**Toby Padilla:** Yeah. + +**Jerod Santo:** Okay. Because I would be concerned that everything would be ad-hoc, and... + +**Toby Padilla:** And super Wild West style? + +**Jerod Santo:** Yeah. I like predictabilities of the command line. It seems like this might lose it if people are all built on different ways. + +**Toby Padilla:** So we're seeing a lot of particular components being used quite a bit in other people's apps. So the viewport used all over the place. You'll see we have a little help menu down at the bottom - that's part of a component. That's something everybody seems to like, because they're baking it in... There's challenges. We're inventing a lot of this stuff; it's kind of like being in the 1980s and inventing a windowing system. So we have to take the best practices that we've learned elsewhere, try to come up with good ideas, iterate and experiment. There are downsides to it that we haven't completely solved for yet. Accessibility is one of those. How do you get a screen reader to read a TUI? Whenever we launch anything, there's always a comment saying "This doesn't work for screen readers." And that's true. It doesn't work great with screen readers. So how do we improve that in the future? We don't have an answer now, but it's on our radar. That's something important to do. + +\[36:20\] Easy things, we've kind of fixed. And it's not even easy, but like I've mentioned before, do you have a light background or a dark background? When possible - and this isn't possible with Wish, because you can't do this over SSH... But on a TUI that you're running native to your machine, we detect your background color. Like, "Okay, you're on a light color. We'll give you things that work with this." Since you can't do that over SSH, we picked a color theme that works on white and dark. It looks slightly better on dark, because we're all on dark, but it still looks quite nice on a light background. + +So there's challenges that windowing systems and the web have come to solve over the decades that they've been in existence, that we want to solve, at least to some degree, but aren't all solved yet, to be completely honest. There's some things that we still need to work on. But because it's so early, that's also exciting. It feels neat; you can come up with a new idea that works for the terminal and makes sense for the terminal. + +Some of these terminals are pretty cool. Kitty is the terminal that we all like to use; a lot of us like to use it. And Kitty is a really great terminal. The person who made it has really pushed forward what he wants to be standards in terms of OSC codes; I believe that's what they are. So ansi escape codes that do things. So he has native notifications that will pop up, like a windowing notification, and he can trigger that from an ansi code. He's got windowing stuff, he's got sounds, he's got -- Copy to Clipboard is another one of these things that he's added... And so when you start looking at that, I think there can be standards that happen as these applications become more rich, that we can all embrace, and solve for some of these problems, especially as a new generation of people come into these tools, with different expectations. Because I think in the past people were like, "Whatever, I'm a gray beard Unix developer. It works for me, so that's fine." + +**Jerod Santo:** One thing I'm seeing, maybe a bug report, or -- I'm just curious; this is the kind of stuff that falls out, is with this two-paned UI that you have here. You have the menu on the left, and then you have the content on the right, which effectively looks like you're rendering Markdown readmes or something that exists elsewhere... + +**Toby Padilla:** It's a Markdown readme, yeah. + +**Jerod Santo:** Yeah. You're using your Lip Gloss...? + +**Toby Padilla:** Glamour. + +**Jerod Santo:** Glamour. Dang it! + +**Toby Padilla:** It's Bubble Tea, Glamour, Lip Gloss... It's got everything; it's got all of our stuff in it. + +**Jerod Santo:** It's bubbles all the way down. + +**Adam Stacoviak:** The works... + +**Jerod Santo:** Yeah. So the git clone command in my terminal - I'm just using macOS Terminal by the app. I can highlight git clone, copy that, it works. But when I try to highlight, for instance, the description for "Bubble Tea, the fun, functional, stateful way to build terminal apps", and I go multi-line, it doesn't highlight the text, it highlights the text plus things in the menu on the left... So this is a circumstance where the new UI hits the real world of a terminal that's controlled by Apple Inc... + +**Toby Padilla:** Yup. + +**Jerod Santo:** ...and these are things that have to be figured out, and ironed out over time. + +**Toby Padilla:** Exactly. And maybe the answer to that is we change the layout for something, right? Because there might not be a good answer to that that exists today. But that's exactly the kind of stuff you're gonna hit when you start doing this. + +I think that terminal in particular is often challenging to develop for. You'll see, the lines have spaces between them probably when you're looking at this... That is a bug in OSX Terminal or macOS Terminal. + +**Jerod Santo:** \[laughs\] + +**Toby Padilla:** Its performance isn't great... But it's the terminal of choice for everybody, right? It's the default terminal for millions and millions of people, so we have to -- so we're always like "Go download Kitty! Go get it now, because everything works super-awesome in it." But we can't rely on that. + +So hopefully, we push this world forward enough that even Apple pays attention. And we've heard that there's a bunch of people at Apple using our stuff, so it's a good first step... + +**Jerod Santo:** That's cool. + +**Toby Padilla:** ...but ideally, the terminal world starts to evolve as well. + +**Jerod Santo:** It's just an interesting world, because in the web you've got browser vendors, right? And then really, where you are -- and the "Download Kitty" was like the old "Go get Firefox" back in the day, right? + +**Toby Padilla:** Yeah. Exactly, yeah. + +**Jerod Santo:** \[40:20\] And it's like, "Well, you can't expect everybody to do that." Maybe we'll get to 20% market share, but still, the other 80% are running IE9, back in the day. + +**Toby Padilla:** That'd be lovely, but it's probably 0.1%, unfortunately. Yeah, exactly. + +**Jerod Santo:** So it's the same exact problem, but a different space. Adam, you were gonna say something... + +**Adam Stacoviak:** I was just gonna say, I can see the coming terminal wars, man... At some point maybe there's the terminal to rule them all. If apps become more prevalent, if this innovation keeps pushing forward, and we hackers find more and more ways to use this and it becomes mainstream... + +**Jerod Santo:** Or at least developer mainstream... + +**Adam Stacoviak:** Yeah. Like I was saying to Toby the first time, if in the future this evolves to the point where when we SSH into let's say our VPS, for some reason... Like, if I wanna do that, I've got my identity attached to it, it's got a particular UI. Maybe there's the common everyday command line, but what if we know our stack and so there's something with the Kubernetes painted glass; we wanna see what's going on ad nauseam across the deck. Maybe there's something like that where it's aware. I don't know, maybe that's a thing. + +**Toby Padilla:** We've been thinking about possibly non-text-based UIs. You can use SSH as a protocol for identity and data transfer above and beyond Terminal. And in fact, X Windows does this; there's an X Forwarding command on SSH that allows you to forward your X Windows over SSH. So that's an old, ancient idea people probably aren't even aware of; that maybe is a new life ahead of that. You can have a UI, having the SSH key as your identity; SSH is a great protocol for encryption, but it doesn't necessarily mean that you have to look at everything inside of a terminal. So we would love to explore this. Even thinking about a mobile app... Maybe you generate some keys on an iOS app, and it identifies with an SSH-based API to figure out who you are. + +And talking quickly about Charm and our server, this is how this works - you use SSH to get a jwt back, and then we use that jwt on a bunch of HTTPS APIs. And we do that for sort of performance reasons, and so other people can build APIs using technology that they're familiar with. But we've been doing a lot of work, and we're going to do a big release around this soon with jwt's. + +So this is moving, "Hey, here's who I am. Here's my public key SSH." Okay, we know who you are; that server who knows who you are issues a signed credential saying "We assert that this person is who they claim they are. Now, web server, whatever, you can rest assured that this person is who they claim that they are." You can imagine this in iOS apps, you can imagine it in other non-text-based apps. + +Again, going back to the Unix philosophy of doing one thing and one thing well, and composable apps... By doing SSH well, it opens up the door to lots of possibilities. + +**Adam Stacoviak:** Are you familiar with the Raspberry Pi, by any change? + +**Toby Padilla:** Yeah, we love Raspberry Pi. In fact, one of our -- not all of our stuff works on it, but it's sort of like a target platform for us... Because one of the thigns we pitched BetaWorks when we started is "Hey, there's a bunch of really (what we call) low-power machines out there that the command line is a really great pairing with." Whether that's a Raspberry Pi, or a machine that's like 15 years old. There's a lot of hardware out there that if you try to run the latest version of macOS on it or Windows, it's just gonna slow to a crawl. But the command line interface is gonna be rocket fast on this thing, so let's target these lower-powered platforms and make cool stuff for them, so you can do cool things. + +I have like five Raspberry Pi's, and I have them all hooked up to my TVs, and they're more or less just Kodi servers. But I would love to actually use it as like a computer in a way that's more meaningful. So if you start making these rich command line apps, then these low-powered Raspberry Pi-esque machines become a viable computing option, and you can do neat things on them. So that's been sort of a Raspberry Pi thought that we had from the get-go. + +**Adam Stacoviak:** \[44:24\] And I was even thinking like that, because there's a lot of people who have multiple Raspberry Pi's in their home lab; there's maybe one for their surveillance, or their automation, things like that... In my case I have a Pi-hole and that's pretty much it. I've got a Plex on a whole different one. It was on Raspberry Pi, but now it's on a beefy Linux server, so I cn't really call that a Raspberry Pi... But it's still command line; when I SSH into it, I have Plex running via Docker, it's running on ZFS, all that good stuff. But I was just thinking, when I psuedo raspi, rsp pi, like raspberry pi configuration, when I go into that, that's some sort of text application. If that world gets easier to build those kinds of things, that kind of config, is that also a future where you can see this is solving for -- that config gets a lot easier for that developer or that team to maintain and build, because this becomes the React for SSH apps, or command line apps. + +**Toby Padilla:** That's exactly what we're thinking. It's such a good answer to remote-controlling something over the command line... Because historically, you're editing a lot of files. You're doing into Etsy and editing like a million flat files to get the configuration you want, because you don't have X Windows (or whatever) necessarily all the time. So you have to do that. So offering a graphical interface for that is good. And you've seen stuff like that. There's things that have existed, but they look like DOS. Writing the code for that is like you're doing C++ or something like that. + +You mentioned React, but that's similar to Elm, in that it's sort of like a reactive, functional design pattern. But that's what Bubble Tea is. If you wanna develop these TUIs in a hyper-modern fashion, that's what Bubble Tea allows you to do - use functional programming to build these. + +What people have learned over the past ten years in terms of building Web UIs, taking those practices and being able to build a TUI with it - we think it's cool. And that use case of having a bunch of Raspberry Pi's and you're controlling them through SSH is exactly the kind of thing that we wanna solve. + +For instance, we launched another project, and it was the first thing Carlos built. So when Carlos started, a week later he shipped his first project, and it was called Wishlist. He loves Wish, and so he's really excited about Wish... So Wishlist started as a proxy to other SSH server and other Wish apps. So he built the party parrot thing, and all of this stuff over Wish. So you would SSH into Wishlist and you would see a list of all of your other Wish apps, or any SSH server, and you could go into it. + +Last week we released local mode. So now you can run Wishlist on a local machine, and it looks at your SSH configuration file, finds all of your hosts, and gives you a graphical interface for jumping into all of these hosts. + +**Adam Stacoviak:** Hm. That'd be super-cool. + +**Toby Padilla:** So discovery, topography discovery of your network. You've got all your Raspberry Pi's in there and you don't have to -- yes, you can have complete it now but having a TUI jump off and discover all of your hosts is a good thing. It makes it easier to do this. + +**Adam Stacoviak:** You're getting me excited, because my cat config lab when I do my SSH config, it's organized; I have all my machines in there, I've got local DNS resolving for machines because I run Raspberry Pi, the Pi-hole, and so because I have that, I've got local DNS essentially for all my machines. So I don't' actually SSH to an IP address anymore; I've now resolved those to an actual domain name that makes sense, that's only internal. For example, my Raspberry Pi, or one of them, is PiMax.home.lan. When I wanna go there, I ssh.pimax.home.lan, and I'm in. But I have all of my hosts files on all my machines the same way... And I kind of wish there was a way -- build this tool next; make it where I can command all my host files from one machine... + +**Jerod Santo:** "Build this one next..." \[laughs\] + +**Adam Stacoviak:** ...because I essentially go to each one of them and manually update them. If I can manage all my hostfiles, all my config files that manage those hosts... + +**Jerod Santo:** You should set up a local DNS server and just point all their DNS to your DNS server. Them it can resolve those, and then you manage it at the DNS server. + +**Adam Stacoviak:** \[48:30\] I guess you could, but... + +**Toby Padilla:** You can actually do that with Pi-hole. + +**Adam Stacoviak:** You can? + +**Toby Padilla:** Yeah, yeah. Pi-hole has an option, so it's running its own local DNS server... And you can add your hosts manually to Pi-hole. + +**Adam Stacoviak:** I've done that already, but what I mean by that is whenever I go to my ssh config file, each host has to be in there for each machine, so that it knows what my hostname is, what my username is for that machine, that kind of thing. + +**Toby Padilla:** Right. + +**Adam Stacoviak:** That thing is tedious. + +**Jerod Santo:** No, if you SSH to a domain -- like, when you SSH to git.charm.sh, there's a DNS lookup that happens. So those machines would just -- + +**Toby Padilla:** But I think he means editing the SSH config. You've gotta shove all of those. You make a new host - great. You've added its Pi-hole, you've added to your DNS; you can get to it. But you have to know its name. + +**Adam Stacoviak:** Right. + +**Toby Padilla:** If you can put it in the SSH config, that starts looking nice. + +**Jerod Santo:** He's writing it into his configs; so he knows the names. Okay, I see what you're saying, but it's -- + +**Adam Stacoviak:** Right. So each computer has that SSH hosts file, that config file, so each computer needs to know, so I can SSH and autocomplete to iMac Pro, or 2 Mac Pro or to whatever the different machines are out there. Each machine -- the DNS is done by Pi-hole, but what isn't done is each machine knowing all the different hosts it has available. Like he's saying, he can -- + +**Toby Padilla:** I will say this is an area that we're exploring heavily. So check out Wishlist for now; that'll get you a nice TUI of browsing through your SSH config. But you can imagine that evolving for server discovery. + +**Adam Stacoviak:** I love it, because it's gonna look at that file, see all the hosts I have defined in there, and give me a list of the things it can navigate as a result. + +**Toby Padilla:** That is exactly what it does. And you can also run it as a server. So it could be a bit of a bastion host that you can configure, and have it be a jumping off point to a bunch of hosts on that subnet, or whatever. + +**Adam Stacoviak:** I'm listening, Toby, I'm listening. Keep going. + +**Jerod Santo:** \[laughs\] + +**Adam Stacoviak:** Tell me more. + +**Jerod Santo:** Well, let's drill down into Bubbles and Bubble Tea, because you mentioned this is like your React's Elm architecture, declarative, easier, a modern way of building these things... And Bubbles are the components, and Bubble Tea is the framework. Now, this is a Go thing. So does this lock you into Go, is my guess...? + +**Toby Padilla:** Yes. And all of our frameworks, all of our libraries and all of our apps are Go. And we actually had a lot of questions; I just got a question about it yesterday, "Do you have a Python SDK?" And the answer is no, we don't. I mean, we're a small team, so we have limited resources... But we think Go is a good language to build command line tools in, for a number of reasons. One is you get a compilable binary that you can ship, a single file, and you don't have to force all of your users to install all of the dependencies that you've used to develop, a la Node.js or Python or something like that. So it's sort of a natural fit for this type of tooling. It's just a nice language; it's got a really good standard library... So a lot of stuff that you might want to do to make a really cool command line app is just straight up available in the Go standard library. It's got an HTTP server, it's got an HTTP client. It's got a lot of stuff inside of it. So that was why we focused on it. + +We like lots of languages, and we come from backgrounds where we -- I was a Clojure developer for six years. I was doing Rust actually for two years before I moved to Go... And Muesli - I follow him on GitHub, and I got jealous of all the cool libraries that he was starting all the time. I'm like, "I want that." Because he's a Golang developer, and he's like "Super-cool crypto thing", or whatever. And I'm like, "Hey, that looks neat. I'm in Rust, writing my own bencoding library, because such a thing doesn't exist." At some point I just thought "You know, I love Rust, it's really cool, it's a super-exciting language", but I jumped ship. I'm like, I wanna be more productive, I wanna use all the fun toys, and so I started doing Go. + +\[52:19\] And so some of us enthusiastically, some of us begrudgingly, have moved to Go over the years, just because it is kind of the answer to building these type of tools. So when people ask us, "Hey, can I do a Node.js version of your stuff?", our answer is "No. Use Go, but come into our Slack and we'll help you." So we're very enthusiastic about onboarding people into Go, teaching people about Go. We'll answer any question, no matter how basic, about Go, in our Slack. So Charm.sh/slack. Come on in and ask us about Go, because we love to answer it, and we'd love to see more people adopting it, especially for command line stuff. + +**Adam Stacoviak:** Are y'all in the main Gopher Slack, by any chance? + +**Toby Padilla:** I'm not a huge chatty person, so I'm not in a lot of Slacks, but I'm sure that Muesli and Carlos are. + +**Adam Stacoviak:** I was gonna say, it might make sense to have a channel just for Charm, that could be a welcome -- sort of a tie even, where maybe that channel could be shared between your Slack and that Slack... That way you can absorb some of the help at large, in the main Gopher Slack. + +**Toby Padilla:** We'll take some of the workload; we'll happily take that, yeah. + +**Adam Stacoviak:** That'd be a good onboarding practice or a good growth strategy. + +**Toby Padilla:** I like that. That's really good. + +**Jerod Santo:** So one of the things I've been musing on is some sort of a Changelog command line tool... And I actually was considering Go as the language of choice for that, because I really like the way GitHub developed their command line tool, which originally was written in Ruby and they rewrote in Go... And I thought, "If I'm gonna just cargo-cult a bunch of stuff off of somebody and not have to do the entire job, I would probably just look at their source code and go from there." And then I've found Bubble Tea and Bubbles, and it seems like these tools are pretty much custom-built for doing exactly what they've already done. So how would you go about writing a command line app with Bubble tea? + +**Toby Padilla:** So the readme is a good place to start, and it kind of walks you through a sort of semi-tutorial of how to do things. It's highly based, like we've said, on Elm. There's some concepts that will help you, like immutability, and you can hack this in Go. You can either do it the Go way, or you can do it the Elm way. So we let you do either things; some of us do it certain ways, and some of us do others. But essentially, there's very few things you have to implement. You have to implement an update method, which is sort of like changing the state of things, and you implement a view method, which returns a string. Period. And the string will look at the state that was updated in the update method, and return the correct string. So you've logged in, so username is now set to Toby. When the string renders, it renders out my name, or whatever. + +And then there's a concept of commands. And commands are -- if you're doing something that's blocking, or a long-running process, you do a command, and when it returns, it calls update with the result, the model that's the result. And that's it. That's all Bubble Tea is, is those three concepts. + +So you learn about the view, you learn about update, and you learn about commands, and then you can start building Bubble Tea apps. And then you can learn about the Bubbles. So if you go into the Bubbles repo, you'll see spinners, progress bars, input forms, viewports, pagers, that kind of thing. So you can just start stringing that stuff together. It takes a while to master this, and because it's so new, we're kind of inventing design patterns as we go. What should be composed? What should be inside of something else? How do you bubble up state? ...this type of stuff. But as we've built bigger and bigger applications, we've learned better design patterns and we've tried to bake that into the documentation and bake that into some of the examples. + +**Jerod Santo:** So what's missing from the Bubbles, if we're getting into the components? Or is there anything glaring, where you're gonna end up writing your own components because there's big gaps in the library? + +**Toby Padilla:** \[56:12\] I think you'll be pretty good with just a basic TUI. I tend to like to invent crazy things, so when I'm making Bubble Tea apps, I rarely start with a Bubble. I just start banging away and making my own stuff, and then often it'll get replaced by a Bubble when somebody else takes over the project, somebody more sophisticated at doing this than I am... But you can do it either way. You can start composing these things that already exist. You can get a long way just trying things out on your own as well. + +**Jerod Santo:** So the string that returns - is this including all of the layout information, or is this string like a whole block of ASCII that comes back, or what is it usually comprised of that creates the UI? + +**Toby Padilla:** That's a great question, and that's actually what Lip Gloss does. Lip Gloss is our styling and layout library. Lip Gloss. + +**Jerod Santo:** So many libraries... + +**Toby Padilla:** I know, we've got a million of them. So Lip Gloss will do the coloring, and it'll do the layout of all of your stuff. So when you do your view, you'll use a bunch of Lip Gloss inside of the view... And it returns a string. So Lip Gloss always spits out a string. So as it renders, it has essentially what looks like CSS style sheets, and then you style up your stuff with that, and return it in your Bubble Tea view. + +**Jerod Santo:** Okay. So I'm looking at Lip Gloss code right now... It very much almost looks like -- is this Go? + +**Toby Padilla:** It is Go. It's almost like a DSL that was written in Go. + +**Jerod Santo:** Yeah, because it doesn't feel like Go code that I would normally read, except for the fmt.println. + +**Adam Stacoviak:** Fumt. + +**Jerod Santo:** Fumt, yes, I know... But I always say "format", I don't know why. Okay, so you've got Lip Gloss for the way it looks and feels, and you've got Bubbles for pre-built stuff, and then Bubble Tea is like what everything plugs into. + +**Toby Padilla:** Then toss it behind Wish, so that it's remotely accessible with SSH... Yup. + +**Jerod Santo:** What about local storage? I mean, these are things that you would just do maybe BoltDB, or something; you'd just use whatever is available in Go, if you're not trying to do -- + +**Toby Padilla:** Well, for persistence we offer Charm. So Charm is a lot of things... It was already a lot to take in. Charm itself is a whole ecosystem. It's charmbracelet/charm on GitHub, and you'll see it's a bunch of things. It's a tool that you use, it's a client, it's a server, and it's a set of libraries. So Charm KV is a Charm-managed BadgerDB that stores your data encrypted in the Charm Cloud and synchronizes it across every machine that you've linked your account to. Charm FS is a fully encrypted Golang fs.fs implementation with writeability, that stores all of your data in the Charm Cloud, and makes it accessible to any machine that you've linked your Charm account to. + +So if you need a key-value store that's accessible across multiple machines, use Charm KV. If you need to access files or store files in the cloud, encrypted, use Charm FS. And CharmCrypt is there if you want just some straight up encryption library, using those intermediary symmetric keys I talked about earlier. And if you run Charm server, or Charm Serve, that's your own Charm Cloud. And you can "charm host = whatever host you're running that on", and instead of using our default servers, all of our tools and any tool built with our libraries will then start storing the data on your own personal server. + +**Break:** \[59:58\] + +**Adam Stacoviak:** So the one thing when you ssh.git.charm.sh, when you git in there - for me at least in Terminal, and maybe this has been talked about and I just missed it, but I don't think so, it's not interactive. And I would imagine the -- I hope at least, Jerod, the plan for this future SSH app for Changelog would be interactive. We have obviously news links... + +**Toby Padilla:** It should be interactive. So if it's not, that's definitely a bug. + +**Adam Stacoviak:** Okay. So it's not interactive -- + +**Jerod Santo:** What do you mean by interactive? + +**Toby Padilla:** Meaning that I can't click the links, for example. There is on selecting the git clone... + +**Jerod Santo:** Oh, mouse support. + +**Adam Stacoviak:** Yeah. There's mouse support to navigate... + +**Toby Padilla:** Oh, you have to use the keyboard to navigate. + +**Adam Stacoviak:** Well, I've got that. So I'm navigating it, but it's not like I can click the links. I have to -- for one, you can't play an mp3 in there. Like, you don't have any, so you're not playing them. I would imagine we have mp3's to play... + +**Jerod Santo:** Well, I noticed also an image on one of them. I think Bubbles has an image -- or no, it's Wishlist. Wishlist has an image in the Markdown that says image:screencast arrow and it points out to a gif that we can't look at. + +**Adam Stacoviak:** Right. + +**Toby Padilla:** So that goes back to the earlier conversation about terminal capabilities... + +**Jerod Santo:** If you had Kitty, you'd be loving this right now. \[laughter\] + +**Toby Padilla:** If you're on Kitty -- we don't actually default to it. We don't display images, because most people aren't gonna be able to see it. + +**Adam Stacoviak:** Gotcha. + +**Toby Padilla:** Some terminals will show you an image. Most won't. Or most meaning macOS Terminal will not show you an image. So we don't try to render that image, because most people aren't going to support it. + +**Adam Stacoviak:** Gotcha. + +**Toby Padilla:** \[01:04:08.16\] So yeah, it's not a rich experience. It's not a multimedia, rich experience. I would love to see that. I like to watch a lot of videos, and I would love to be able to do something with video, and build some sort of video app. That's not today, but we can all try to get there as command line aficionados over the next 5-10 years... Because I think it's ideal. + +One of the reasons we're doing this is because we have super web fatigue. Lots of things about the web suck. I'm sick of trackers, ads... I don't like the advertising business model. I don't like all this JavaScript that's running on my machine, and these bloated web pages, and everything's bloated, everything's slow... There's ads autoplaying, and stuff... So stripping all that away, getting to the point where you're not tracked everywhere you go is one of the reasons we moved to the command line. But we lost a lot of stuff along the way, right? We lost YouTube. Bye-bye, YouTube. Like, that's important; we need that. So it's really early days in terms of that. + +There are solutions to this that are more or less interesting. So you could run a server that serves a custom web page that displays video, and then managing that server is certainly doable on the command line. But there's not a great way to make rich user interfaces. Some people would like that. I don't know if you guys are familiar with the Gemini project, but their entire philosophy is -- they're basically like Gopher 2.0, text-based web replacements, HTTP replacement, and they also have web fatigue, probably more than we do... So they've gone into this world where everything's supposed to be stripped down; it's just text only, right? If it's not text-only, they don't even want a part of it. So there's cultures out there that embrace just that... But I think it's really important to look forward on how we can make this even richer. + +**Adam Stacoviak:** Maybe I'll reframe my question there, or at least the statement I was trying to get to, which is - you know, I can't click the Git links, or the different links in the readmes, so that's like a stopper for me... In terms of what we're trying to build, or aspire to build, is then if we want to have an SSH app that is built with Bubble Tea, styled with Glamour, all that good stuff, behind Wish, and we're doing all this fun stuff, and we want to allow our listeners, our audience, essentially, to see our podcasts, see the index of all of our shows, but then also be able to play them - is that a possibility? And if it's not, what's the hurdles to get to that point from the interface standpoint, with going this route? + +**Toby Padilla:** It is not if you go through Wish. So SSH itself is limiting some of this stuff; there are solutions -- if you guys build a Changelog command line app, and you have people download it, you could have it open stuff; you could have it open something on the users' machines. You can have it open a web page, you can have it do lots of things. + +Over Wish, that's running on the server, so you're not gonna be able to say "Hey, open this file", or whatever. You can do that though, as long as you build a tool that they will run on their own machine. You can do that numerous ways. Sometimes the terminal itself - and I don't know if the OSX terminal does this, where it'll just take a hyperlink and make it clickable. A lot of terminals do that, so you can just open it in the browser. But you could also just execute in the background the open command, or the xdg-open and it will open up the appropriate application for that mind type, so you can open up an mp4, or whatever. It would probably open it up into Apple Music, or something like that. + +So there's definitely ways to handle it if you build a local tool. If you're trying to make it remotely accessible, that's limited. And that's actually a challenge we've had. Because those git clone URLs - you can't execute that. Your back almost in the GitHub world, to some degree, where you're hopping and pasting it into another window. Now, you're still on the terminal, so that's nice, but we're looking at options there. + +\[01:05:07.20\] SSH, to some degree, as the viewer, is limiting there, in that it doesn't support linking. Hypertext linking - that's not a thing on SSH. So do we write our own SSH client replacement that offers a bunch of cool, rich options? I don't know, maybe we do. + +**Adam Stacoviak:** Maybe you should just write your own terminal, the one that rules them all... And the reason why I kind of get there, why I wanted to sort of option click, is because that's the experience you get in VS Code, for example. Traditionally, even in an editor, you're not necessarily interacting with the links in it. So maybe pushing the boundaries of, say, playing a video, or seeing an image, or playing an mp3 - maybe that's pushing the boundaries of the protocol. But from an interaction standpoint, I would desire or expect to be able to option click a link, because that's what you could do in other editors. + +So if this emulates, at least visually, even though it is terminal, and you're innovating that part of it, my expectation as a user is like "Can I option click these links and do different things?" But the answer is no right now. So it's more like understanding what the limits are, and how do you -- if you're trying to innovate the command line, how far down to the metal do you have to go? Do you have to kitty it? Do you have to make your own kitty? You know what I mean? Or can you just adopt -- + +**Toby Padilla:** \[laughs\] Who doesn't want to...? + +**Adam Stacoviak:** Right. But you know, is that how far you have to go? + +**Toby Padilla:** It's a great question. We actually asked ourselves yet initially, like "Do we want to make a terminal?" And we explicitly said "Not yet. Not right now." Because there's a lot of other stuff that has to get built at the application level. We're getting to the point where that's sort of the limiting factor. + +**Adam Stacoviak:** It seems like the apps are limited though by the current standards, and if you could evolve those current standards, maybe the process of building these apps, and the libraries and the different things that you're building on top of it might -- maybe I'm wrong, because I'm not doing the work, but maybe it might be easier to just take a stab at it for a month, do a month project and see what it would take to do that... Or maybe longer. + +**Toby Padilla:** Building a terminal is a big task because it's got like 20 years of backwards-compatibility that if you don't support, people would be raging on you right away... Like, "Oh, it breaks --" Because there's so many command line apps. Like "Oh, Vim doesn't work through Tmux for me anymore", or whatever. There's a lot of stuff that people are doing out there that you need to support, and it's actually a pretty big job to do it. We are -- and this is talking about a sort of secret project that we were looking at at one point, that we decided not to go down... We're like "Hey, VR is kind of cool. Maybe we should do a VR terminal." That's kind of an interesting concept, right? It's a new platform... How do you bring the command line to this new platform? How do you make a really good mobile command line? How do you make a really good mobile terminal? + +So I think there's desktop terminals, I think there's other platforms that could have really cool terminals... I think maybe we even explore some of that stuff at some point in time. But as of right now, we've been really focused on -- I mean, even though Bubble Tea and Bubbles are some of our biggest projects... Bubble Tea is actually our biggest project at this point. It's just passed Glow in terms of stars last week. A lot of our effort and engineering research has gone into the SSH side of things; so the identity side, the encryption side, how you handle all of the Git protocols over SSH... + +**Adam Stacoviak:** Fundamentals. + +**Toby Padilla:** Yeah, because it's a little more separated from the UI aspect. The UI is great, and it's what gets people in the door, it gets people coming to us, and it's really nice to show off, and it's super-fun building this stuff. And we need to keep pushing that forward. But a lot of the value of what we're creating I think is at the identity level, and really pushing SSH forward. So that's another area for us. There's deep rabbit holes everywhere, and you can start going down them... And as we get bigger, we'll probably simultaneously dive into multiple rabbit holes. But for now, we can kind of like dip into a couple of them. + +**Jerod Santo:** \[01:09:12.27\] So if we redirect back to that side of things, the SSH side - you have Soft Serve, which is a Wish-based Git server, correct? + +**Toby Padilla:** Yup. + +**Jerod Santo:** You have Wishlist, which is a Wish-based SSH directory. + +**Toby Padilla:** Yup. + +**Jerod Santo:** Are there other Wish-based apps out there, or even an idea for apps that people could wrap their heads around? + +**Toby Padilla:** Yes, so Charm is also Wish-based. + +**Jerod Santo:** Okay, Charm. Yup. + +**Toby Padilla:** And there's a version of Glow that's in beta now, that we'll soon be launching, that's based on the new Charm. So Glow sits on top of Charm as well, and the use case in Glow -- + +**Jerod Santo:** Glow is the Markdown rendering thing. + +**Toby Padilla:** Glow is the Markdown rendering thing. + +**Jerod Santo:** And Charm is the -- + +**Toby Padilla:** Charm is the encryption and identity and cloud storage. + +**Jerod Santo:** Okay. + +**Toby Padilla:** So in Glow, if you open up a Markdown in Glow, you can hit s. And s is stash. Say you just downloaded Glow for the first time, you ran it in a project directory, they'll discover all the Markdown in that directory... And then you're like, "Hey, this is interesting. I wanna save it for later." S. Stash. "I don't know what that is. It seems cool." Stash it. "Okay, great. It's stashed." What we did behind the scenes there is generated you a new SSH key pair, generated you encryption keys, encrypted all of that Markdowns file, hit the Charm server with your key, created a new account for you on the Charm server, and then pushed up that encrypted Markdown for your account onto our server, or a self-hosted server, if you want that. + +**Jerod Santo:** Okay. + +**Toby Padilla:** So all of that SSH mechanic behind the scenes is happening through Wish. And so Charm is a Wish app and Wish library that applications like Glow can use to satisfy any kind of storage or encryption or persistence or identity use cases. There are third-party Wish apps that are starting to come out. We saw two Wordle clones launch in the last three weeks over SSH. Play Wordle over SSH. And they're really great. One of them, Clidle, was on the top of our Golang on Reddit all last week. I don't have its URL off the top of my head, but it's a great implementation of Wordle over SSH. And people loved it. Something about this resonates with people; it's kind of like the BBS days, or something... Just the fact that you can be on the command line, you can use SSH, and all of a sudden you're playing Wordle, is kind of a neat idea. And you can even score stuff, because it keeps your public key, or whatever... I know that they're working on a lot of this stuff, but it's like "Hey, identity, plus remote access..." I didn't install anything; I didn't install Wordle. I just had SSH lying around and I was able to use this thing. + +So we saw two of those launch... And then some kind of like toy kind of stuff... But it's pretty new. Wish is like -- I think a month-and-a-half it's been out in the wild, or something like that. So it's all relatively new. But it's exciting to start seeing people not just build stuff with it, but gain some success. So we love it when people build with our tools, and then they go to the top of Hacker News, or they're trending on GitHub, or something. They're able to build something that resonates with the community, with the tools we're giving them; that's pretty neat. + +There's a lot of popular Bubble Tea apps. Bubble Tea has almost 500 applications that have been built with Bubble Tea. It's been out there a little longer. There's a thing called Slides this guy Maas created, and it's the second most popular Bubble Tea app besides Glow, in terms of being just like a pure Bubble Tea app... And it's like PowerPoint for the terminal. Like I mentioned, the official GitHub client is using Glamour, it's not using Bubble Tea, but that's like a very big distribution of one of our libraries. Min.io and their official command line tool is using Bubble Tea. That's a new thing. Supabase - I know you had the founder on recently; they in their official command line client are using Bubble Tea. + +\[01:12:55.15\] So we're starting to see good Bubble Tea distribution and some pretty large, pretty hot projects... And a bunch of fun stuff, too. There's just fun projects that people make with Bubble Tea. Mergestat is a great one; this is "Treat your Git repo like a SQL database, and run SQL against it." He had a functioning tool, and then he kind of discovered our stuff and started layering more and more. He tweeted that, he's like "I just added a bunch of Bubble Tea stuff to mergestat. I'm gonna add all the Charm stuff to it." Like, yes, please do. That's awesome. So we're seeing people get excited about it, build it, and then achieve some success with it, which is neat. + +**Jerod Santo:** I'm over here, googling all these names that you are saying... Actually, mergestat I think is by the AskGit guy, who we've talked about in the past. Small world... QueryGit like it's SQL - pretty cool stuff. Definitely hook us up with a list for the show notes. + +**Toby Padilla:** I will. + +**Jerod Santo:** I did find Clidle, but you'll probably even think of more after you hang up with us, and say "Oh, here's another cool one." So we'll have like a nice list of things using either Wish, Charm, and/or Bubble Tea out there in the wild. + +**Toby Padilla:** Cool. + +**Jerod Santo:** Because you know, sometimes just seeing those things are the inspiration you need for the idea, and actually make it click. I think Bubble Tea right now clicks for me in a way which I'm not sure Wish does, because I've seen the examples you're giving and I'm like "These are cool." But then I can't go anywhere from there beyond without more interactive things than is currently available with SSH. That being said, I'm just a guy with very small ambitions and can't think of cool things. I see them and I'm like "Oh, now it all makes sense!" So the more examples... Like you said, it's only been out there for 30 days, so more time, more people will start to play with this and come up with cool ideas. If you're listening and you have a cool idea for something like an SSH-based app, definitely holler at us, let us know about that. Let us know if you would like to try it with Wish, or with Charm, or with Bubble and Bubble Tea... All these things. + +**Adam Stacoviak:** So many names... + +**Toby Padilla:** Lip Gloss, Harmonica... + +**Jerod Santo:** Lip Gloss, maybe some Glamour in there... I'm not sure if Glamour does any more -- oh yeah, style sheets-based Markdown rendering for your CLI apps. Yes, Glamour... + +**Adam Stacoviak:** As you were naming those examples, I was seeing how our desire for what we wanna do is certainly the edge case, right? + +**Jerod Santo:** I think we could use Bubble Tea for sure, and build a really cool, interactive Changelog command line tool, where you could play mp3s. As long as we can get that stuff done in Go, right? Like, to play the audio, which probably has some cross-platform libraries that you'll have to use to interact with the audio libraries of various operating systems. Bubble Tea is just for the interface, right? You can build full-fledged Go applications and use Bubble Tea for your interface, right? + +**Toby Padilla:** Yes, you can. One thing I'm most excited about not having seen a lot of people build with it yet, but it's kind of futuristic and weird, is the Git stuff. Building Git-based apps. So what can you do with Git? Could you clone a repo, and you guys are pushing new episodes into it all the time, and you're just fetching the new versions? Or is there a collaborative option there? For instance, this episode was maybe a repo, and I could push up these links to it, and you guys could add stuff to it, too... Or something like that. + +So I think Git-based interaction - it's super-unexplored. And to me, that's the most exciting part of Wish, is like "Oh, you can start building these Git-based apps", and using Git as a way to control things, and as a way to collaborate. That seems interesting. + +**Adam Stacoviak:** And if you can probe it like a database, which it basically is, then maybe you can use it sort of like a pseudo-database almost. We have a web app... Maybe Postgres can pull data from that Git repository via WebHooks with GitHub, or some other way, to essentially be -- + +**Jerod Santo:** We already do that for our show notes. Our show notes are synced with a GitHub repo. + +**Adam Stacoviak:** \[01:16:53.20\] Right. It's very similar to that, yeah. + +**Toby Padilla:** You can make a Wish app that does that. So you would get, in the go code, someone's pushing something, here's what they're pushing, and then in the Go code you could use your Postgres driver to just populate the database with the file, or whatever got pushed into Git. So from a CMS perspective it's kind of interesting, and that's how we're doing the CMS bit of Soft Serve, is with all the Git pushing. + +**Jerod Santo:** Soft Serve being the self-hostable Git sever for the command line. Each time we say a noun, I have to describe which one it is... Because y'all have so many projects, man. It's crazy. + +**Toby Padilla:** \[laughs\] + +**Adam Stacoviak:** It is challenging to keep track of what -- + +**Toby Padilla:** Imagine having to make 3D models for all of these things... \[laughter\] + +**Jerod Santo:** Well, you're doing that on Hard mode. You took that upon yourselves. + +**Toby Padilla:** We did. + +**Jerod Santo:** But that's the fun part, right? + +**Adam Stacoviak:** Let me ask you one question here... You named a few examples out there. When you were in the final stages - and these were Wish apps, right? That you've just described. + +**Toby Padilla:** Mm-hm. + +**Adam Stacoviak:** I'm trying to keep track, sorry about that... When you were in the final days, I suppose, of pushing and the team was dreaming, "Oh my gosh, when we get this out there, people will build this." What were those dreams? You described some things people are using. GitHub using Glamour, or Lip Gloss, one of those two, and I think you mentioned Supabase, what they used... What were some of the apps you imagined would be built with Wish? + +**Toby Padilla:** We actually built Wish and Soft Serve because we had a very specific need for it. So we had built Charm, the new version of Charm, the self-hostable version of Charm, all that Charm KV stuff - all of that, and we needed to launch it, and we wanted to move Glow over to this backend. And the way doing all of this works in Golang is difficult... Because if you look at Golang code, when you import a library, you put a Git URL; it's literally like github.com/charmbracelet/charm/kv, whatever. Whatever library you bring in, it's literally a Git URL. So it was not ready for primetime, but we wanted to get it in front of people, and so we needed another place to put this, other than our public GitHub profile. And we weren't ready to overwrite some of these apps with the new versions, and we wanted to bring them in... This is kind of complicated inside of Go to make it happen, so we were like, "Okay, we need a new Git place where we can pull this stuff." So we started building our own Git server, and like "Let's make it on the command line. Let's make it over SSH. Let's do all this." + +So we created Wish -- essentially, we back-ported it into Charm. But we created Wish so we can make Soft Serve, so we could launch Charm. We wanted an alternative Git space that wasn't GitHub. Because there's times you need a Git repo -- Git is very much about promotion; you put it out there, everyone's gonna see it, but you sometimes wanna put something out where people can use it, but you don't necessarily wanna broadcast it out to everyone. So that's why we made Soft Serve... It's like your second Git host. So that was why we made Wish. But after we made it, we were like, "Wow, it's kind of crazy. We just wired Bubble Tea up into SSH", and we're wiring identity up, and we can do all this really cool stuff with Git. + +And then our wheels started spinning. We were like, "Okay, it turns out Git is awesome for CMS." Some of these things I've just told you guys about, like the podcasting, and the notes, and editing - that's my mind in Wish land. I'm like, "Now, these things are possible." So it's time to start thinking about how you can leverage Git as an API. How do you use Git as the way you manage an application, or what can you do from a collaboration standpoint with Git, that's not just writing code. So I think it's really early days there... But it seems incredibly exciting; it seems interesting, because it's not how you would typically do something. Being able to configure a yaml, check out the config repo, edit the yaml, and push it back to Soft Serve, and then Soft Serve behaves differently is kind of an interesting, new concept. To use Git as a way of configuration seems pretty powerful, and I think we'll see more and more of that. + +**Adam Stacoviak:** \[01:21:04.20\] I'll throw this one out there, too... I don't know if this is a one-to-one, because I almost would imagine just managing my own Bash profile, or the ZSH profile, and I was thinking of like Oh My Zsh. This is something that a lot of people install, it has config essentially, where you can choose your theme, and a bunch of stuff... And I wonder, would Oh My Zsh be a good candidate for some sort of app like this, where rather than editing your actual profile, the ZSH profile file itself, would it make more sense to put that in the config? I don't know. Is that where you see that too, where some tooling or dev tools have config locally, and it makes more sense, rather than opening up the file? I don't know. What do you think? + +**Toby Padilla:** We're looking at dotfiles as well. We've got some stuff in the works there, because it's a natural fit for what we're doing... And it's a lot of what you say - imagine you've got a Windows machine... More realistically - you have a MacBook and you've got a Linux-based VPS running on DigitalOcean, or something... And you've got Vim on both of them, or you've got some configuration that's slightly different. Tmux works different on Mac than it does on Linux, or something like this. So you have different versions of the same dotfile. How do you manage that? Well, immediately you think Git is useful there, right? Because the way I'm managing my dotfiles now is in branches on a Git repo, and it's quite hellacious, but it's a problem we're looking to solve... And we've actually got some prototype stuff running; I can't talk too much about it yet, because it hasn't been launched, but we're close to launching some stuff that plays in this area. + +**Adam Stacoviak:** Gotcha. So it helps in the dotfiles area. Maybe even an actual dotfiles manager. + +**Toby Padilla:** Potentially. + +**Adam Stacoviak:** And branches on a Git repository, which would be so cool. Each branch is a machine, for example. This machine is that branch, and... + +**Toby Padilla:** Yeah, and that's what I do now, and managing that manually is not fun. + +**Adam Stacoviak:** Yeah, I bet. + +**Toby Padilla:** So tooling around that is definitely desirable. + +**Jerod Santo:** Well, let me suggest a name for this SkunkWorks project... I would call it Cupcake. + +**Toby Padilla:** You know what - I will tell you what its name is... And its name is Donut. + +**Jerod Santo:** Ohh...! + +**Adam Stacoviak:** So close! + +**Jerod Santo:** Pretty close. + +**Adam Stacoviak:** Donut's cool. + +**Toby Padilla:** So you guys are getting the scoop on that; we haven't announced it yet. + +**Adam Stacoviak:** Alright, so charm.sh/ -- I'm just kidding. Don't insert name here, but... That's where it's gonna live, right? Of course... + +**Toby Padilla:** Probably, yes. + +**Jerod Santo:** Donut sounds like dotfile, so... + +**Toby Padilla:** Doesn't it? It looks like a dot, too. + +**Jerod Santo:** That's true. + +**Adam Stacoviak:** That is so cool. Okay, okay... I'm digging it. + +**Jerod Santo:** I'm interested, yes. + +**Adam Stacoviak:** I think that between Donut and what you're doing, I can definitely see why you are so passionate about innovating in this way. + +**Jerod Santo:** Yeah. + +**Adam Stacoviak:** My hope is that commercial opportunity is in the future, just because sustain... But not so much that -- I just hope that you have enough runway, and can sustain enough runway to keep innovating, because these are definitely cool things. I'm just not sure - except from the Charm Cloud aspect; that's where I can see a commercial offering for you when you get more and more people to adopt SSH-based apps... And why run your own when you can run it there? So give me managed; I can run my own, I can run my VPS, or do it on a Raspberry Pi, because I'm sure you've thought about that... + +**Jerod Santo:** Right. + +**Adam Stacoviak:** But do I have to? Maybe I can pay $10/month, or enterprise pay $50 or $100, or whatever the number is. That's where I can see something, at least in the near future. But you know, I'm loving the direction, for sure... And the naming, and the fun behind it - that's what really makes it fun. It makes your job, I'm sure, more fun, it makes the team have a lot more fun with meetings... Calling a new project Donut is a lot more cool than something that's lame, basically... You know? + +**Jerod Santo:** Something that's lame... \[laughs\] Something like Cupcake. + +**Toby Padilla:** Always better than something that's lame. + +**Adam Stacoviak:** Right, right, right. Anything else in closing? What have we not asked you, that you wanna share before we close out? + +**Toby Padilla:** I think we covered a lot here... Thank you for having me on; this was a lot of fun. + +**Adam Stacoviak:** It was a lot of fun, yeah. + +**Toby Padilla:** \[01:24:59.07\] I would say if people wanna follow us, come to @charmcli on Twitter, and @charmbracelet on GitHub. That's the best way to stay up to date. Or like I said, join our Slack, charm.sh/slack. But other than that, this was great. + +**Jerod Santo:** One last thing that I will say before we close out... I've just thought of this, because I'm looking up Charm's license... I would just say that since Charm is long-term probably the enterprise commercial play, and it can be self-hosted, and yet maybe you want them to host it for you - I just want to say, you've got street cred because it's out there, it's open source, it's MIT-licensed... So you really are putting your money where your mouth is in that regard; even with that piece that is probably going to become the commercially viable part of all this stuff, which is just free and open; really cool stuff. So props to you on that aspect as well. + +**Toby Padilla:** Thank you. + +**Adam Stacoviak:** I would say even more props too on all the other tooling being able to wire up too a self-hosted Charm. + +**Jerod Santo:** Yeah. + +**Adam Stacoviak:** Like, have it as a cloud that's managed is cool, but then a single config, it sounds like -- I was thinking like Ruby days... And I've even forgot all the terminology. I haven't done it in so long. But wherever you point Ruby Gems to... What's that file, Jerod? + +**Jerod Santo:** RubyGems.org. + +**Adam Stacoviak:** Yeah, exactly. In the top of the file you can just point it wherever you want to get your gems. It seems like a similar change in terms of a config. Just point to wherever. + +**Toby Padilla:** It is. It's just an environment variable, Charm host. You set your Charm host environment variable... And in fact, once Glow is launched, the new one, that points to Charm Cloud - even Glow, when you stash, it'll start stashing it to your Charm Cloud. Anything that's built with our stuff will just magically work on your self-host. That's important to us. + +And from a commercial standpoint, it's not necessarily a bad idea, because what you need as a self-hosted person is quite different than what an enterprise of 10,000 people need, right? And Go actually makes this easy, because we design all of our stuff with interfaces, implement those interfaces on SQLite on the self-hosted side, and then for instance, on our industrial strength Charm Cloud we've got Postgres backing it, and we just implement all the same interface methods, in a different, more sort of industrial-grade backend. + +So we think through some of this stuff -- or stats for monitoring and for all of this stuff... We don't even implement that on the self-hosted side; on our servers, that interface has an implementation that points to Grafana. + +**Adam Stacoviak:** Pretty cool. + +**Jerod Santo:** So before we let you go, we'd be remiss not to mention the footer of the website, where you take haters... + +**Adam Stacoviak:** Oh, yes. + +**Jerod Santo:** ...and you redirect them to /dev/null. Tell us the story here. You're sending the haters to /dev/null? + +**Toby Padilla:** Where else would they go? + +**Jerod Santo:** \[laughs\] + +**Toby Padilla:** Where else would we want them? + +**Jerod Santo:** Good question. \[laughs\] Well said, well said. + +**Adam Stacoviak:** And you've trademarked it. + +**Jerod Santo:** That's right. + +**Adam Stacoviak:** Well, I want the T-shirt... Put that on a T-shirt, so I can wear it. + +**Toby Padilla:** We are working on swag. There is stuff coming soon. + +**Jerod Santo:** Cool. + +**Adam Stacoviak:** Give me a sticker, give me a T-shirt, give me a pen, some sort of glossy pen... I'd be so excited about that. + +**Toby Padilla:** We will. + +**Jerod Santo:** Give me a cupcake. + +**Adam Stacoviak:** Yeah, give me a cupcake. + +**Toby Padilla:** We were thinking about Keycaps. We all have mechanical keyboards, so something like that... + +**Jerod Santo:** Oh, that would be cool. + +**Adam Stacoviak:** Yeah, for sure. + +**Jerod Santo:** Well, good luck to you guys, and keep it up. It's really cool stuff. I'm definitely gonna check out Bubble Tea for this SkunkWorks Changelog command line thing that we've been playing around with... So we'll give it a shot. + +**Toby Padilla:** Awesome. And just let us know if you have any questions. We're happy to help you. + +**Adam Stacoviak:** Alright, Toby, thanks for your time. + +**Toby Padilla:** Alright. Thank you very much. diff --git a/Mastering Go_transcript.txt b/Mastering Go_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..7c363b662d1411f84e54377e047477d9ec9c7d38 --- /dev/null +++ b/Mastering Go_transcript.txt @@ -0,0 +1,269 @@ +**Johnny Boursiquot:** Yes, yes, it is indeed Go Time. I will be your host today. I'm Johnny Boursiquot, and joining me is the author of the book "Mastering Go", Mr. Mihalis Tsoukalos. Hopefully, I've done your name justice... + +**Mihalis Tsoukalos:** Yes. + +**Johnny Boursiquot:** Awesome, awesome. So I became aware of your book - or rather the first edition of your book, maybe a year or two ago. And then it seemed like one of those very thick reference manuals. I was like, "Okay, this is one of those things you've gotta get on your shelves", like old-school, those giant tomes that you'd put on your shelves and you'd use them as reference every once in a while... Because nobody reads a book front to back these days. So I'm thinking, "Okay, maybe I'll put this book on my shelf." + +Lo and behold, I never happened to do that, and then the publisher reaches out to me and says "Hey, would you like a review copy of this book? And I'm like "Sure, yeah. It's not like I don't have a dozen other things to do", but hey, I'm a sucker for reviewing Go books, and I was looking forward to this one as well... And I got the book in the mail, started going through it, and I was basically nodding in few places, I'm like, "Yup, yup, that's how I would explain that." In other places I'm like, "Okay, yeah, that's a different way of explaining it." But the thing is, as I was flipping through the book - and I certainly looked at different areas of the book to see how you explain things, and see "Are there some tricks and tips and things that I can steal from you to explain certain things as well in teaching people Go?" And when the opportunity came up for -- I think I was talking about it on Twitter or something, and you chimed in... I'm like, "Okay, so he seems friendly enough. Why don't I just have him on the show, so we can actually talk about the book?" But not just about the book, but really -- the title, basically, I find intriguing, and I wanted to talk about what it is like to master Go. We've gotta go meta with this, right? + +\[04:09\] But yeah, for those who don't know who you are, who you are on Twitter, the Twitter handle is @mactsouk, and you're a systems engineer and technical writer. So obviously, the writing is important. You're writing books, and everything. This is the third edition of your book. I wanted to start out with basically -- well, I'll give you a chance to sort of add any flavor you want to your intro. Like, what else have you been up to? + +**Mihalis Tsoukalos:** Okay, so hello everyone, thanks for listening to us, and Johnny, thanks for the opportunity to come to your podcast. I'm a Unix passionate, I like Unix. I don't know why, but I like Unix, so you know... And systems stuff. I used to be an Oracle DBA, and I used to work with Cisco IOS... You know, the operating system of Cisco, not the Apple version of iOS. And Unix, and setting up SendMail DNS, with BIND, all these things that some people find miserable, but I really like them. + +**Johnny Boursiquot:** Yeah, it used to be a rite of passage, you know? Set up your own mail server, and everything, your own DNS... + +**Mihalis Tsoukalos:** Yeah, SendMail with m4 configuration files, and those things. And working with versions of Unix other than Linux, like the IBM version AIX, or the Hewlett Packard version of Unix, HP-UX, and Solaris, my favorite one... So I'm a systems person, I like Unix, and I like writing software... So every time I see a systems programming language, I like to try it. I like C very much, but C is not perfect. I don't like C++ though... It's too complex for me. + +So the idea was to find something that I can write systems software, but without having to worry about C stuff. That's how I started getting interested in Go. And when I learned that Rob Pike was involved - that was a big plus. + +So I started working more with Go, and learning Go more, and after some time I just wanted to write a book that I would like to read when I was learning Go. But the thing is that first I wrote a series of articles for Computer Magazine in the U.K. Ten, or maybe twelve, I don't remember... A series of articles that at the end became something like the table of contents of the book. Because you know, you have to prepare. + +And that's how I started getting involved with Go. You know, you can't trust a programming language too early, but you have to work with it, see how it goes, and see if you want to keep working with it. + +**Johnny Boursiquot:** Yeah, I definitely agree with that sentiment. The fact of the matter is when you say you want to write a book that you would have liked to learn Go with, the approach you took with the book - I definitely didn't think this was a beginner book. I think that's something that usually when I give references to people for learning material, I usually ask them "What level are you at? Are you an existing programmer who is learning to pick up a new language, or are you brand new to programming? Or are you basically a die-hard functional programming developer?" We analyze your preferences. And when I picked up this book, the first place I went - in the intro it basically says who this book is for. And right then and there, it says "Hey, this is an intermediate book." + +So then I was like, okay, great. That means when I start going through that book, I'm not going to see things like "This is what a conditional is." It'll just say "This is how Go does conditionals." Or things of that nature. + +\[07:56\] So when I went through the book, once I understood sort of the frame of reference that you were using to actually write the book, then the gripes I might have had with it, I started basically saying "Well, okay, so if Mihalis is making these assumptions about the reader, then this is the lens with which I have to look at this book and ingest some of this book." So I think definitely something I wanna put out there - a lot of times folks just pick up a book, and they'll give a review, and not really understanding basically who the book was targeted for... If you're a beginner and you've picked up this book - yeah, you'll still learn a lot. But there are some things that you're not gonna go deep into. There's some language features that you already ran to without having to explain the background on them, because you assume that the reader is coming in with sort of a basic understanding of at least what Go is, and what Go does. + +So that was the approach I took with the book... And for those who are looking to pick it up as well, that's what they should expect. That said, I kind of wanted to give you a chance to narrate a little bit on -- yes, this is a book you would have liked to have had, but you've been working with Go for a while; obviously, you're gonna have a biased answer for this, but do you believe a book like this - maybe not necessarily your book, but a book like this - is the right entry point for somebody who's maybe coming from another language and looking to get into Go? Is this a book for just learning Go, period, or are you targeting one level deeper? ...not just for the person who's starting to learn to pick up Go, but as you mentioned, you like systems programming, right? Will a systems programmer find this Go book in particular a lot more their speed, or should they look at something else more targeted at systems programming? + +**Mihalis Tsoukalos:** It depends. Generally speaking, if you have some experience with programming, you should pick up Mastering Go very quickly, because you have to have some experience. If you have the experience, then you can work with it very easily... Because programmers know that the key point to learning something is practice. Some things we find them difficult, and we have to look again, and other people find other things difficult, and they have to look at them multiple times to understand them. + +With Go -- I mean, one thing that surprised me with Go is that you didn't have the tilde, and fork, and things that you do in C. It took me a while to understand it, and understand that I have to work with goroutines instead. You have to practice to understand that, but in order to look for a fork, you have to know about fork. So if you know fork, then you can understand what is going on there. + +Again, I think that the key point is practice. If you want to practice, then you will be okay. And the cool thing with the book is that even I use it as a reference sometimes, many times. Because you can't remember everything. + +**Johnny Boursiquot:** God must give you some special kind of pride to be using your own book as a reference. + +**Mihalis Tsoukalos:** Yes, yes, yes. This is a good point, not as a pride, but as a thing that you can -- you know, it is useful. It's not pride, because I have read many books, and I have used them as a reference as well. Many people have done that. But it's good to find it useful, because you cannot remember everything. You can't have everything in your mind. + +**Johnny Boursiquot:** Right, right. Yeah, Google-driven development is a thing... I partake in Google-driven development myself, so no shame in my game. Nobody can remember everything, and the nuance of everything. Every once in a while I'll still go to -- you know, my favorite resource in the Go community is gobyexample.com, for example... For example, for example. Gobyexample.com is a great resource that I regularly reference, even to this day, after having been doing Go for like five years, or something. I'll be like, "Okay, how do I initialize this again? What's the technique for doing this again?" + +\[12:00\] One of the things I constantly refer to is "How do I properly seed for randomization?" I always have to go look up "Okay, what's the package duran something?" These things, even though you use them frequently -- like, sometimes it's "Okay, what's the syntax? What is this type that I use all the time, that I need to learn a little bit more about? What is this package in the standard library that I have used a dozen times and I've only ever tried one or two functions from this package? What else does it offer?" + +So this is sort of a continuous journey of mastering Go, which is something I definitely wanna get into and talk about in a bit... Sort of at the philosophical level, what does it mean to master Go. But this is a 750-page book, so it's a chunky book. Lots of topics covered. What do you think was the most interesting parts for you to research and write about? + +**Mihalis Tsoukalos:** The Go garbage collector. + +**Johnny Boursiquot:** Ooh... + +**Mihalis Tsoukalos:** I really enjoyed reading about that and learning about it, and then writing about that. I really enjoyed it. And I also enjoyed lots of external libraries like Viper, and Cobra, and \[unintelligible 00:13:06.02\] and Gorilla. Because they extend the language in a good way. Especially Cobra and Viper are really helpful. As I'm working with some REST APIs, I use gorilla/mux, and I'm trying also to learn \[unintelligible 00:13:24.24\] to understand what their differences are, and which one is better for what job. + +I mean, that's the experience part. It's one thing to learn about something, but it's another thing to be able to choose based on facts. That's the difficult part. + +**Break:** \[13:43\] + +**Johnny Boursiquot:** Speaking of the research and writing process, how long did the -- at least the first edition; I'm assuming with each edition you're not rewriting everything, you're just updating a few parts, some things that have changed, maybe some idioms have changed and whatnot... Roughly how long did it take you to write and publish the first version? + +**Mihalis Tsoukalos:** Given the experience that I had with the articles, and another book that I have written for Packt is the Go Systems Programming, which helped a lot; it's one edition, it's the first book... It took me - and a lot of preparation though - six to eight months. But with a lot of preparation. You know, you cannot start from zero and expect to have the book ready in six months. But a lot of preparation, and a lot of code. You have to have some code ready. + +**Johnny Boursiquot:** \[16:07\] So did you enjoy -- like, in my mind, having collaborated on book projects and things, in my mind there's the thing that I'm writing about, and then there's the writing about the thing that I'm writing about. Those are very different disciplines, very different skillsets. Writing about something you know and doing the thing you know are very different things. I can go all day and feel like I know what I'm doing, but the process of writing, even if I've written before, writing a book and working with an editor, working with people that are gonna give you feedback on the writing process, the tools and resources that you use as a writer - these things are not your day-to-day typical programmer tool. There's always sort of a learning process that's also taking place with regards to the act of writing a book, or the act of creating a training, or a course, or the act of creating something for relaying and conveying information to another human being. That in itself is a separate skillset that you have to develop. I wonder how you feel about that skillset, having gone through it a few times. How do you feel about that? + +**Mihalis Tsoukalos:** For this skillset, the articles that I'm writing for Computer Magazine have helped me a lot, because for the article you have to present something, and you have a limited space. So you have to have a beginning, a middle and an end, and that helps a lot. So I'm trying to follow the same approach. As I've told you before, you cannot remember everything. If I read something afterwards, I try to be sure that it makes sense, at least to me, and then to the readers. Because you know what they say, "Write for your friends, but edit for your enemies." + +**Johnny Boursiquot:** \[laughs\] I'd never heard of that one before. I'm gonna use that. + +**Mihalis Tsoukalos:** So when I'm editing, I'm trying to edit and be a hard judge for my own writing. And I think that helped me a lot with magazines, but also for the books. But have in mind that I have written some articles for magazine that I'm not very proud of. \[laughs\] It's not perfect, it's a process. + +**Johnny Boursiquot:** Right, right, right. + +**Mihalis Tsoukalos:** You don't begin perfect, it's a process. You learn, and you make mistakes, and then you make more mistakes, you correct something, and you try. We're not perfect, we're just -- you know, work in progress. + +**Johnny Boursiquot:** Don't expect to have something good, ready to ship the first time around, the first go-around, right? + +**Mihalis Tsoukalos:** Yeah. + +**Johnny Boursiquot:** ...which sounds very much like the engineering process as well. You never quite get something right the first time. That's why we have even idioms around the act of writing software. You write a test, you get a failure, we make the test pass, and rinse and repeat. We often have mnemonics and things around these things, so I think we shouldn't expect any different from a different kind of creative process. It's just ones that you're not writing code. Although, I must say, I'm envious a bit, because -- I'm sort of picking your brain in this whole thing, because I too wanna write books on Go and whatnot, and I'm picking your brain and saying "Okay, what should I run away from?" \[laughs\] If I were to take on such a project, what should I run away from, or watch out for? What are the pitfalls to taking on a writing project? ...not necessarily exclusively about Go, but really about any task of this nature. Basically, you're teaching people about how to use a tool, this tool being a programming language. What do you think the biggest takeaways for you were about writing this book and creating this project? + +**Mihalis Tsoukalos:** The key to writing the book is to enjoy the process first. Because if you don't enjoy the process, it's difficult. The second is that you have to think of the reader as a friend; not as money, but as something that you're trying to help... You know, just help. Try to explain and don't assume that something that is easy for you should be easy for the other people, and vice-versa. Sometimes you have to explain something twice maybe... It's a process. You have to enjoy the process. Once you enjoy the process, then everything becomes easier. + +\[20:22\] The other thing that I told you before - write for your friends, edit for your enemies. You should be edit hard, for your own good. That's the way to do -- you have the skeleton, you have the text, but you have to edit hard. I'm still learning that. I really enjoy books like The C Programming Language, or The Unix Programming Environment - they have the necessary thing only. + +**Johnny Boursiquot:** Right, right. Everything you need, nothing you don't. + +**Mihalis Tsoukalos:** Yeah. And they have written these books with troff and Unix tools, and that's really amazing. \[laughter\] + +**Johnny Boursiquot:** Yeah... + +**Mihalis Tsoukalos:** Not even latex + +**Johnny Boursiquot:** Yeah, not even modern tools that we have. That is amazing. + +**Mihalis Tsoukalos:** And they have created classic books, just for the love of it. + +**Johnny Boursiquot:** Indeed, indeed. Yeah, I always tell folks that based on all the publishers that I talk to, and having been part of book projects and everything else, I'm like, "Okay, people who do this - it's not done for money." This is definitely a labor of love, the advance that you get really is not -- if you were to map that to your hourly wage, whatever you're getting out with your employer or consulting, whatever it is, the amount of time you put in, it wouldn't even come close. When you do this because it's a labor of love, it's not something that -- unless you have a book that's immensely popular, whatever it is, and you make bank, and that's fine... But the vast majority of people writing technical books like this - you're really just sharing something with the community, for the love of it. I can definitely appreciate that. + +So again, the book is called Mastering Go, and I wanted to talk about what we believe mastering Go today looks like. If somebody started out saying, "Hey, you know what - I'm gonna pick up this language, and I wanna master it." What does that journey look like for somebody who's basically starting today and saying "I wanna master this language." What are the things they need to look out for? What are the approaches that they should have? What should they expect? + +**Mihalis Tsoukalos:** Look, my book is here to help people - not master Go, but like Go. If you like Go and you use it, you're going to master it. The thing that mastered everything is the marketing stuff mainly. You cannot master something, but... + +**Johnny Boursiquot:** Right, right, right. + +**Mihalis Tsoukalos:** No book will help you master anything. + +**Johnny Boursiquot:** Right. \[laughs\] + +**Mihalis Tsoukalos:** It's like swimming. You cannot master swimming by reading about swimming. You have to swim. So I just try to make people love Go, understand what they can do and what they can't do with Go, and where Go is -- what are the advantages and disadvantages of Go, and then Go out and write software. That's the key. Without practice, you won't do anything. I have been inspired by The C Programming Language book, I have been inspired by the book The Unix Programming Environment, by the Stevens books, the Advanced Programming in the Unix Environment series of books, or the TCP/IP, network programming books by Stevens... These books were my inspiration. They didn't teach me everything, but they inspired me to write things. I have written an FTP client in C a long time ago, which was -- it was pretty challenging, but rewarding. I mean, that's the idea. + +**Johnny Boursiquot:** Yeah. You're gonna learn by bumping your head against things, and researching, and figuring out how it's done... Yeah. That's really key to mastering anything, right? You have to keep doing it. + +**Mihalis Tsoukalos:** Yes. + +**Johnny Boursiquot:** \[23:59\] For me, whenever I hear the term "master this, master that", I'm like -- yeah, like you said, this is marketing. I get it. But to me, at some point I believe you can reach what we can consider the realm of mastery. When somebody can present a problem and say "Hey, how would you approach this in Go?" and immediately ideas start popping in your head, and you're like "Oh yeah, I could use a buffer for this, I could use some goroutines for that, I'm gonna communicate over these channels..." So immediately in your head you start thinking "How can I approach this?" So I think sort of getting to these ideas flowing into your mind at a rapid pace - to me, I think that is what shows a sign of mastery. + +**Mihalis Tsoukalos:** Exactly. + +**Johnny Boursiquot:** As you become so comfortable with the language that it's no longer a matter of "Okay, how do I piece together this syntax?" but I have a problem that I wanna solve. I wanna go from problem statement to possible solution and something that I can test, and not having to worry about the actual writing of the syntax. You can't type fast enough, because the language at this point becomes your pen. You're writing it down. So you have the idea in your mind and you're basically just trying to get your physical hand to keep up with your mind. + +**Mihalis Tsoukalos:** Yes, exactly. + +**Johnny Boursiquot:** So to me, when you've mastered Go or when you're reaching a level of mastery, I think that's when you know; when the language sort of fades away, it gets out of your way, and it's just about now communicating your design or your idea, you're communicating that an bringing something to life. That's where I think -- if somebody's looking for that signal that you can recognize in your own self, "How good am I with this thing?", I think you have to ask yourself "How often do I have to reference basic concepts? How often do I have to look up some documentation on something I use often enough?" So that's how I know I still have a lot to learn about Go, because I'm still looking up some basic stuff sometimes. + +**Mihalis Tsoukalos:** Me too, yes. + +**Johnny Boursiquot:** I'm like, "Okay, yeah, I've been doing it for a while, but I expect - hopefully, if I'm lucky, I'll still be doing it for years and years to come, so I expect that ease of just using the language and it kind of fading away as a tool, and not really thinking about it too much... I expect that to come in another 5, 6, 7 years." So I expect it to be a long journey. + +Personally, that's why I love Go, because - can I stick with this language for a decade? Can I use this language and be happy using it for a decade or more? To me, that's gonna give me enough time to get to the mastery level, and that's the journey that I'm on... Like, "Can I spend a decade on this language?" + +**Mihalis Tsoukalos:** And the most important point is that a master knows when not to use it. + +**Johnny Boursiquot:** Mm-hm, mm-hm... + +**Mihalis Tsoukalos:** You cannot solve everything with a tool. You have to know when not to use it. It's equally important. + +**Johnny Boursiquot:** Yeah, agreed. So obviously, you've been a part of the Go community for a while... Personally, I'm big on the Go community overall; I think technology is technology, but the people is what makes it fun. I'm interested in knowing - do you have any stories about the community, or people, some of your favorite gophers, or anything like that? Do you have anybody in the community you wanna give a shout-out to? ...however you wanna do that, but I just wanted to ask, at the very least. + +**Mihalis Tsoukalos:** \[laughs\] Lots of people's writings have helped me a lot, and especially the official Go Blog is really helpful. I have read many things from Mat, who's not here with us today, and Bill Kennedy has helped me a lot... His writings are very good, and he's always ready to help and willing to help people, and the community is really -- I mean, I don't know if other programming languages have such an active and helpful community. + +**Johnny Boursiquot:** \[27:52\] Yeah, I totally agree. These are definitely some of my favorite people as well. I happened to also have the distinct pleasure of co-hosting this podcast with Mat every once in a while, along with my fantastic other co-hosts that grace the mic here every once in a while... And again, for me that's part of the joy; I get to hang out with people that share this love, and basically we want the community to be a welcoming one, a friendly one, a helpful one... So we try to live that, and hopefully as people new to the community come in and they feel that love, they feel that welcome, they stick around. + +I usually say -- pretty much every year we have a GopherCon, which is the big, official, I'd say the biggest Go conference we have year-round. And over the years we've gotten a lot more conferences spread throughout the world as well, but GopherCon is sort of the big one, and the one that usually gets a lot of people showing up on-site, and everything. During the pandemic we had to do it online the last couple of years, but it's gonna be in-person again this year... + +**Mihalis Tsoukalos:** Oh yes, that's great. + +**Johnny Boursiquot:** So I am looking forward to that. I can't tell you how much I'm looking forward to that, to see people I haven't seen in quite some time, or only online here and there... But yeah, at GopherCon I usually see the makeup of the audience; it's usually, consistently, there are more newbie Go developers, brand new folks in the Go community than there are veterans. This means that there are more people coming into the community than there are people that have been doing it for a while. Every year, consistently. + +So that means Go still has a lot of time, a lot of growth in front of it. We're just gonna continue to attract a ton of new people into the community, and to me, it's important that with every layer of growth, right? Every ring in the tree if you will, people that come in and that we add to the community, that we continue to carry on that kindness, and be nice to people... + +**Mihalis Tsoukalos:** Exactly, exactly. + +**Johnny Boursiquot:** ...be helpful... All of these things that basically this community has as core values. So we definitely want that to keep growing, along with the number of people that we have coming into the community. So to me, that's huge. Obviously, you play your own role in that, in creating learning materials and interacting with people online, so definitely you're part of that... So on behalf of the community, I say thank you. + +**Mihalis Tsoukalos:** Thank you. \[laughs\] + +**Johnny Boursiquot:** I did wanna ask a couple more questions, and then we're gonna switch it over to unpopular opinions. I hope you brought an unpopular opinion... + +**Mihalis Tsoukalos:** Yes. + +**Johnny Boursiquot:** Nice, nice. Obviously. Go is not the only language you've used; obviously, you've done some C, and whatnot; you even built an FTP client in C. Who knows, you should do that in Go and compare notes, and see "Okay, what was that experience like?" But anything else that you're actively using and learning about today beyond just Go and old-school C? + +**Mihalis Tsoukalos:** I'm using Python a lot nowadays, because I'm doing some excess things with time series and stuff, and Python is really good at that. Big data, and things. + +**Johnny Boursiquot:** Okay. + +**Mihalis Tsoukalos:** For the first time I wrote Java last year. + +**Johnny Boursiquot:** Oh. \[laughs\] + +**Mihalis Tsoukalos:** I don't like Java, but I had to... + +**Johnny Boursiquot:** Well, I'm sorry... \[laughter\] + +**Mihalis Tsoukalos:** It's not a bad programming language, but it's not my favorite. + +**Johnny Boursiquot:** I'm kidding, I'm kidding. Yeah, no shade to the Java developers out there. We love you, too. + +**Mihalis Tsoukalos:** I don't like the syntax. It's not a bad programming language, but I don't like the syntax. I never liked the syntax. But as we said before, it's a common secret that a good programmer can write in any programming language. You can pick up a programming language really easily, especially when you don't have to deal with graphics, and that stuff; if it's command line utilities... + +**Johnny Boursiquot:** Except Haskell. Man, every time I pick up Haskell, I'm like "I don't know what the heck is going on..." \[laughter\] I digress... + +**Mihalis Tsoukalos:** Yes. I have tried to learn Elixir and Erlang, but... You know, they require a different kind of thinking. They are good, especially Erlang; Erlang was used for telephone initially, but it's high availability, and all these things; it's very powerful. But you know, you have to like it. Or you have to read it again, and like it in the process. + +\[32:14\] So I mainly write Python and Go, and I'm learning other languages. I'm going to write a tutorial about the V programming language. I haven't heard of it before, but... You know, it's a good way of learning it. + +**Johnny Boursiquot:** Yeah, yeah. + +**Mihalis Tsoukalos:** And I like Rust, but the syntax of Rust is really bizarre, I don't know... Have you tried Rust? + +**Johnny Boursiquot:** You know, I've played around in some playgrounds, and actually I've read in the early days, when Rust was in pre-1.0, I tried to get into it... And since then there's been some improvements in terms of some things that were kind of weird in the beginning; they've refined those things. So I need to give it another go, I have to give it another try. A lot of things have been sort of smoothed out... + +But if I add one more thing on my plate, I'll probably just fall apart. There's so many balls in here right now; that's my excuse right now, and it is a legitimate one. I have so many projects going on, I'm like "Okay, before I can add one more thing to my plate, I have to take one off." I have to be done with one project before I take on another one. So that's how I'm keeping myself honest. But yeah, definitely, Rust is on my to-learn list. Basically, I try to pick up a new language every 2-3 years. The last one I picked was Go, and I've stuck with that for the last 4-5 years or so... And now it's time to pick up something new. Definitely not leaving behind Go, for sure; not by a long shot. Go is still my daily bread, it's still what I use for work every day, it's still what I use for my side-projects... It's my thing. But yeah, it's always good to learn something new; it brings in a different lens with which to look at problems. So definitely, Rust is on my list. + +Before diving into Go, I had actually started picking up Elixir as well, only because some of the same few folks that were in the Ruby community basically were either working on Elixir, or somehow were involved in that community as well... I did a ton of Ruby back then, so naturally, I heard about Elixir and everything else, and started to learn about Erlang, and the BEAM virtual machine, and all that stuff. I learned a good deal about that stuff. + +And then I met Go, and pretty much all others were cast aside once I learned about Go... But definitely, I think I'm at a point where I'm adding something new to the mix. Always be learning. That's how you grow. Always be learning. + +**Mihalis Tsoukalos:** Yes, yes. Usually, when I want to learn a programming language, I start writing small Unix command line utilities first, to see how it goes. I usually do that. + +**Johnny Boursiquot:** Right. + +**Mihalis Tsoukalos:** It's my way of learning. And then I might go to a REST API server, or client, and see how it goes... Because that's the kind of things that I usually work with, so that's making sense, to write something totally different. + +**Johnny Boursiquot:** Yeah, indeed. Alright, so we're gonna switch it over to Unpopular Opinion... And I'm told I need to play a tune, so here we go. I'm gonna start playing my tune. + +**Mihalis Tsoukalos:** Okay. + +**Jingle:** \[35:27\] to \[35:44\] + +**Johnny Boursiquot:** Alright... So you brought some heat? Let's hear it. + +**Mihalis Tsoukalos:** Okay, I have two smaller ones. The first one is that most meetings can be replaced by emails. + +**Johnny Boursiquot:** Right. You just send an email. Yeah, I don't think that's gonna be unpopular at all. \[laughter\] A lot of meetings could have been an email. + +**Mihalis Tsoukalos:** And the second one is that C is the best programming language ever. + +**Johnny Boursiquot:** Ooh...! + +**Mihalis Tsoukalos:** \[36:08\] \[laughs\] + +**Johnny Boursiquot:** Okay, okay... Alright, alright... Let me mull on that for a second. + +**Mihalis Tsoukalos:** If I have to choose a single programming language to live with, that would definitely be C. + +**Johnny Boursiquot:** That would be C, huh? Okay, okay... + +**Mihalis Tsoukalos:** Yes. Not C++, just C. + +**Johnny Boursiquot:** Just C. You like managing your own memory, and all that stuff, huh...? + +**Mihalis Tsoukalos:** No, I don't anymore... \[laughter\] I don't anymore. But you know, back then it was not easy to have garbage collection or automatic memory management. Otherwise we wouldn't have Unix. That would be a shame. + +**Johnny Boursiquot:** Okay, okay... I mean, yeah, if you didn't have a choice... You know, if you're gonna write C, that's as close to the metal (as they say) as you're gonna get, unless you wanna start writing assembly, or something... + +**Mihalis Tsoukalos:** Yes, exactly. \[laughs\] Exactly. + +**Johnny Boursiquot:** I can see that... Okay, okay... Well, gosh. + +**Mihalis Tsoukalos:** What's your favorite programming language full-time? I mean, if you have to pick one. + +**Johnny Boursiquot:** I mean, clearly, it's gonna be Go. That's the language I use. You make me repeat myself. I use it for my day job, I use it in my side projects, I use it for my side-side projects... Go is my go-to. That's the first thing I think of, even for things that I used to write sort of a quick Bash or a shell script for, or whatever... Or I used to rely on Ruby for a lot of my little scripts, knick-knacks here and there... Because the compiler in Go is so fast, I get that feedback immediately, and I get type safety, and all that stuff... Even the random little scripts that I used to write - I've replaced all that stuff with just Go. + +So it's quite possible that I'm now using Go as a hammer to hit every nail, but it makes it so easy for me to do that. But obviously, again, right tool for the job, so depending on what project I'm working on, if there's already a language that's being used that's not Go, I adapt to that. So in my repertoire, I have at least 6-7 different languages, so every once in a while I'll come across something that "Okay... Man, I haven't written Lua in ages, but this thing is written in Lua, and now I've gotta brush off my Lua." Or be it Java, or reading Python, whatever it is. + +I think the key takeaway here is sometimes you don't get to pick the best tool for the job... And perhaps that is my unpopular opinion - you don't always get to pick the best tool for the job. Sometimes the tool that was picked is maybe even the worst possible tool that could have been picked for the job... But that was the tool that was used, so now you have to deal with it. So the unpopular opinion is "Suck it up, buttercup! It is what it is. Sometimes you've just gotta do what you've gotta do." + +**Mihalis Tsoukalos:** Sometimes people choose the tools for you, and you have to use them. + +**Johnny Boursiquot:** Yeah, exactly. + +**Mihalis Tsoukalos:** That's true, that's true. + +**Johnny Boursiquot:** That is exactly it. Mihalis, thank you so much for being on the show. I'm definitely enjoying the book. + +**Mihalis Tsoukalos:** Thank you. + +**Johnny Boursiquot:** The first thing I did when I got the book, I went straight to the generics stuff and see how did you explain generics, and you did a pretty decent job of that. So yeah, for those looking to buy the book, it's available on all the major publisher sites, and Amazon, and all that stuff. + +**Mihalis Tsoukalos:** Yes. + +**Johnny Boursiquot:** I think it's already out, and ready to be shipped, or something. This is the third edition of the book, I've been enjoying it, so thank you for your contribution to the Go community. Looking forward to more from you, and looking forward to your active participation in the community as well. + +**Mihalis Tsoukalos:** Thank you very much, and thank you very much for your kind words. I really enjoyed being here, and maybe we can talk again in the future. + +**Johnny Boursiquot:** Yes, indeed. Well, if you keep writing these books, I'm gonna keep having you on, to come and talk about them. \[laughter\] + +**Mihalis Tsoukalos:** Thank you, thank you very much. + +**Johnny Boursiquot:** Alright, y'all, it's been awesome, so let's put on some outro... diff --git a/Mat's GopherCon EU diary_transcript.txt b/Mat's GopherCon EU diary_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..b32eb3b0b97a9656d4c6862bc77036e182d28d40 --- /dev/null +++ b/Mat's GopherCon EU diary_transcript.txt @@ -0,0 +1,739 @@ +**Mat Ryer:** Hello, and welcome to my Go Time diary. I'm on my way to GopherCon EU. I'll be co-hosting there, with a group of great people. I'm gonna be looking forward to the talks, and the hacking days, there's conference contributor chats that are gonna be, we're gonna play Gophers Say... We're going to do a panel with the Go team... So it's a packed few days. I'm very excited. It's gonna be hard work, but I thought what I'd do is record my thoughts along the way, and sort of bring you with me on the journey and give you a little peek behind the scenes. So let's go! + +Okay, I've just arrived at the airport... Yeah, the drive was okay. I wouldn't say the driver had good breath... But he more than made up for it with his erratic driving. And he did suggest to -- instead of bringing me to the actual airport, just dropping me off at a nearby roundabout... But we both agreed in the end that that was absolutely insane. He wanted to avoid the charges, but... I decided to cover them for him. So here we go, I'm gonna now head into the airport. I'm on my way. + +Okay, through security... Didn't even get checked, didn't even pat me down, or anything, so going to the gym was a waste of time again... And now I'm gonna find my gate. Ah, I think I'm on time. I might just make this. + +\[06:06\] Somebody was just a little bit lax on the old escalator etiquette there, leaving her bag way behind her as she stalled at the top... And it was inevitable I was gonna run into it, because you know, the stairs are moving... So I did fell a little bit. Being British, I apologized, even though we both know it's 100% her fault... + +All boarded now, ready for take-off. + +\[flight attendant making an announcement\] \[07:03\] + +A bit bumpy landing, but I wasn't even scared. My tummy went funny a bit on the way down, but I didn't even cry. I definitely didn't cry. + +I just love the vibes here in Berlin. It's very cool; there's graffiti everywhere. Everyone's got this sort of punk kind of -- almost like a dystopian vibe, which I just think is great. Someone told me the motto of Berlin is "Poor, but sexy", and that resonates with me. I grew up poor, and I was a sexy kid... No, don't say that last bit. But I don't know, everyone just seems very nice and welcoming. Someone just gave me free olives. Do you know what I mean? Just as an example... Free olives. I can't complain -- oh... Yeah? Oh, okay. Oh no, I do -- I've gotta pay for the olives. But still... + +An old lady just dropped her shopping bag, and I helped her pick it up, and she just went "Wanker!" Which is absolutely -- I can't believe it. + +Here I am at the Go Contributors Summit, and there's a lot of gophers assembling now, as we speak. I'm gonna talk to a few of them. I'm here with Egon. Hello, Egon. + +**Egon Elbre:** Hello, Mat. + +**Mat Ryer:** How are you doing? + +**Egon Elbre:** I'm doing well. + +**Mat Ryer:** Is this the first time in Berlin + +**Egon Elbre:** Yeah, it's the first time. + +**Mat Ryer:** And how do you find it so far? + +**Egon Elbre:** It's nice. The street art is amazing. + +**Mat Ryer:** Yeah, isn't it? It's like the whole city is tattooed. + +**Egon Elbre:** Yeah, exactly. + +**Mat Ryer:** And what are you gonna be speaking about? + +**Egon Elbre:** I'm not sure which table I'm going to attend. The static analysis one seems enticing, and yeah, the Factory Automation I'm like, wondering what it's about. + +**Mat Ryer:** Yeah, I don't think it's got anything to do with Java factories. I don't think you need to worry about that. + +**Egon Elbre:** I really hope so... + +**Mat Ryer:** \[laughs\] Okay. We'll catch up with you later then. Thank you. + +**Egon Elbre:** Yeah, okay. + +**Jingle:** \[09:56\] + +**Mat Ryer:** Alejandra is here... Aren't you? + +**Ale Kennedy:** Of course I am! + +**Mat Ryer:** \[laughs\] How was the tour? You went on a walking tour of Berlin this morning, right? + +**Ale Kennedy:** Yes, it was a lengthy tour... But I'm gonna tell you, being from Miami, the weather was perfect for a walk. + +**Mat Ryer:** Oh, really? + +**Ale Kennedy:** I wasn't hot at all, I didn't get tired... It was wonderful. We saw so many places... Victoria was the one leading the tour; it was amazing. She knew so much about it. It was really like the insiders' tour, or something like that. + +**Mat Ryer:** Oh, nice. + +**Ale Kennedy:** I didn't feel like a tourist. I felt like I was walking with a friend, and she just was showing me around. It was very nice. + +**Mat Ryer:** That sounds great, yeah. So what was your favorite place you've visited? + +**Ale Kennedy:** Um, I would have to say that -- I don't know if it's my favorite, but the most impressive one for me was the Holocaust Memorial that was created for Peter Eisenman. It's very impressive, it's overwhelming, beautiful... So many feelings at the same time that one place can bring. + +**Mat Ryer:** Yeah... + +**Ale Kennedy:** There's nothing I can tell you that would reflect the same thing, because you would be standing in the same place and you would tell me a whole different story. I think that's the beauty of that place. + +**Mat Ryer:** Wow. And the city is so beautiful as well, isn't it? + +**Ale Kennedy:** The city is really beautiful. Today I was talking to my daughter, she's somewhere in Cambridge and whatnot, and I was telling her, if Berlin was a person, it would be a cool, old, tattooed dude. That would be Berlin, if he was a man. + +**Mat Ryer:** It's like tattooed, isn't it? The whole city. + +**Ale Kennedy:** Yeah, it's like the whole city is wearing whole sleeves of tattoos, and it gives it character. I guess the mom in me gets annoyed, because I see beautiful buildings with writings, and then I have to get over that, and then see it for what it really is. There's a lot of people voicing things, and that's amazing. + +**Mat Ryer:** One of the nice things about GopherCon EU is that it moves around as well. So we had one in Iceland, we had one in Tenerife, and now Berlin... So it's great, you get to visit these places that you would otherwise maybe not go to. + +**Ale Kennedy:** That is amazing. You have a different take on it, because usually the organizers give you a little bit of an insight of wherever it is that you are. They kind of like curate the experience for anybody that goes. + +**Mat Ryer:** Yeah. Great. Well, I hope you enjoy the rest of the time, and I look forward to hosting with you. + +**Ale Kennedy:** Will do, thank you so much. + +**Mat Ryer:** It's gonna be fun. + +**Ale Kennedy:** Alright! I'll see you on Saturday. Thank you. + +**Mat Ryer:** Yeah, see you then. Thank you. + +**Ale Kennedy:** Bye! + +**Jingle:** \[12:45\] + +**Mat Ryer:** I've just tracked down Ole Bulbuk. Ole, hello. + +**Ole Bulbuk:** Hi, Mat. Nice to meet you. + +**Mat Ryer:** Nice to see you again. You're doing a table at the Contributors Summit on long-term maintenance. What is that? I only do quick projects and then leave it for someone else to worry about... What is long-term maintenance? + +**Ole Bulbuk:** Yeah... \[laughs\] That's when you come like three years after a few guys have been working on it, and then find a total mess, and think "Oh dang, now I have to take care of it..." + +**Mat Ryer:** \[laughs\] Yeah. Do you think Go has a kind of particular advantage when it comes to long-term maintenance and legacy code + +**Ole Bulbuk:** I think it has quite some advantages. There is often only one way to do something in Go, so there's only one way to mess it up usually, too. That's great. And it has been done with scalability in mind; that's absolutely fantastic. + +**Mat Ryer:** Yeah. + +**Ole Bulbuk:** For example, no circular dependencies between packages, which in the long run is really great, even if it hurts sometimes in the short run... And yeah, I think it's a rather pragmatic and simplistic way to work with code, that you say "Simple is better", and so on. This is really great and helpful. + +**Mat Ryer:** \[14:15\] Yeah. It pays dividends over time especially, doesn't it? I know what you mean. And what about the backwards-compatibility promise? That must be a big part of this, too. + +**Ole Bulbuk:** Yeah, of course. If you have a language that evolves really quickly, every year an incompatible version - well, then you either stick with a very old version, that has security problems after a while, or you are constantly evolving the whole thing, and always -- or you are still doing it like two weeks before, and you should really catch up now, and so on. + +**Mat Ryer:** Yeah... We're not naming any names. Swift \*cough-cough\* + +**Ole Bulbuk:** Yeah... Or JavaScript, or so... But yeah. And this is really nice, of course, for Go, that you can count on it and build something solid, and maintain it without a whole team working on a rather little codebase all the time... And you can do it just as a side project when there is not much functionality to be added, but you can really maintain it easily. + +**Mat Ryer:** Yeah, that's a really good point. Well, enjoy the conference, and I'll catch up with you later. + +**Ole Bulbuk:** Thank you! + +**Jingle:** \[15:31\] + +**Break:** \[15:46\] + +**Mat Ryer:** I'm here with Christen -- Christian... Oh, okay, let me say that again... + +**Christian Haas:** Can I say who I am? + +**Mat Ryer:** Yeah, yeah... Can you pronounce it okay? + +**Christian Haas:** Yes, I hope so... + +**Mat Ryer:** Okay. + +**Christian Haas:** So Mat is here with Christian Haas... + +**Mat Ryer:** Thank you. \[laughter\] Christian, you're doing a table discussion about -- it's called "How to assert", right? + +**Christian Haas:** Yeah. It's called "How to assert", because I was given to have it a name with only three words. And I also realized it should have been better called "How to assert in tests." + +**Mat Ryer:** Ah, right. + +**Christian Haas:** So \[unintelligible 00:17:12.07\] It came out of a Twitter discussion. + +**Mat Ryer:** Oh, yeah. You could have also had "Assert yourself, baby." + +**Christian Haas:** It could have been, yes, and a totally different discussion would have come out of it. + +**Mat Ryer:** Yeah. That would be a very different Terminator. I don't think Terminator was written using test-driven development... Otherwise it may be a little more out there + +**Christian Haas:** Well, he shoots first, and I know he asked first "Have you seen John Conner" and then shoots + +**Mat Ryer:** That's true, yeah. + +**Christian Haas:** So there's still a question before it. Yet it's way beyond what we are discussing here on our table here... + +**Mat Ryer:** You're talking about then testing in Go, and specifically about the assertions; are you talking about the code structure, and things like that? + +**Christian Haas:** Essentially, the code structure, almost... The question was "Do we manually assert?" Like, in an if my got values equals the want value compared to do it right in assert library, whatever is A B and or Testify, or any other of the assertion libraries that are out there. + +**Mat Ryer:** \[18:09\] Yeah, that's interesting. I mean, at one time Testify as a cert package was the most imported package in Go... And when we made it, we kind of were used to writing tests like that, and we just found the three lines to be verbose and repetitive, because you still wanna print out what the values are, and things. But it has, of course, the benefit of -- you know, it's very readable, and you don't have to learn anything new. So that's kind of the trade-off, isn't it? + +**Christian Haas:** That's the trade-off. And the question, or at least the curiosity that came now to this question was "What are the reasons that people would still choose such an assertion library?" Because it can go back and forth. Yes, you have just laid out the reasons - you want to have it readable, some say they don't want to learn a new library... Yet my question is "Okay, what kind of reasons do the people have that do use an assertion library?" And I would also like to bring in the four rules of simple design - are we breaking them, or are we actually adhering them if we use either or the other? + +**Mat Ryer:** Yeah. And Testify has, of course, the Require sister package, which is where it aborts early... + +**Christian Haas:** That's right. + +**Mat Ryer:** That's also a choice that people make. Some people prefer assert where you can just keep going and make lots of other assertions... + +**Christian Haas:** Correct. + +**Mat Ryer:** ...and other people like to just have one failing thing at a time. + +**Christian Haas:** Right. So this is then the question "Do you want a checker on properties as a whole, as a group we have the property that I'm going to test consists of several smaller things that I would not expect, or do we want to require this particular thing not to be nil because if I would continue then it would be lost anyway + +**Mat Ryer:** Yeah, right. Exactly. And when we designed is, it actually returns booleans, so that you can write them inside if statements. You can say like "If this is not nil, basically, then..." And then that will fail if it is, but otherwise it returns true and you carry on into that block, and stuff. But yeah, I quite like the abort early style myself, because I love the fact that you get like a to-do list from your tests, and you just get the next thing that you have to fix. + +**Christian Haas:** Right, right. This helps in test-driven development as it is. Then again, the question is "Do we want to immediately break up because you want to stop?" And I've heard people saying "Well, if this one fails already, we will be pretty sure anyway what the problem is", and then you can skip the next requires or the next expectations, because if the first one is already broken, you have to look at it anyway. + +**Mat Ryer:** Yeah. Very cool. Well, I'll come back and chat here after you've had the conversation and get your sense of what people think. I'd be interested. + +**Christian Haas:** Yeah, more and more people are coming in for this - not roundtable, but rather rectangular table discussion... But yeah, all good here. + +**Mat Ryer:** Yeah. Good. I like the pedantry. + +**Christian Haas:** Thank you. Thank you, Mat. + +**Mat Ryer:** Thank you! + +**Christian Haas:** This was Mat Ryer, with Christian... \[laughter\] See you around. + +**Mat Ryer:** Christian Haas, thank you. + +**Christian Haas:** Alright, thank you. + +**Jingle:** \[21:11\] + +**Mat Ryer:** Oh no, I can't believe it... Bill Kennedy is here. Hi, Bill... + +**Bill Kennedy:** What?! I can't believe it. Mat Ryer is here. + +**Mat Ryer:** \[laughs\] So you're doing a roundtable on educating gophers... And you've been doing that for a long time now, haven't you? + +**Bill Kennedy:** On the Go side, since 2014... Yeah... + +**Mat Ryer:** Yeah, wow. That's the first GopherCon. + +**Bill Kennedy:** Yes. I got to even speak at the first GopherCon, which was pretty cool. + +**Mat Ryer:** Yeah, I remember it. + +**Bill Kennedy:** That's because nobody knew me yet, and they needed speakers, and I was dumb enough to say "Sure..." Which was actually the first conference I ever talked at. + +**Mat Ryer:** Was it? + +**Bill Kennedy:** Yeah. + +**Mat Ryer:** Oh, wow. + +**Bill Kennedy:** 800 people. It was with the Go team, if you remember, all sitting there. + +**Mat Ryer:** Yeah, yeah. + +**Bill Kennedy:** Yeah, that was a little nerve-wracking. + +**Mat Ryer:** Is it on YouTube still? + +**Bill Kennedy:** I imagine Brian and Erik still have that website... That's interesting, because I've never looked at the -- I did a cool little video with the Gopher in Walmar-- shopping stuff, and I've played that first, and everybody laughed... + +**Mat Ryer:** \[22:12\] Yes, I remember it. + +**Bill Kennedy:** I've gotta find that video. I don't think I have the video itself anymore, but it will be part of the talk somewhere. + +**Mat Ryer:** I must have that somewhere. Yeah, because I just randomly lived in Denver, and was getting into Go, and then the conference was just down the street from where I lived, just completely randomly. So I was at the first GopherCon in 2014. + +**Bill Kennedy:** You were living in Denver? + +**Mat Ryer:** Yeah, yeah, I lived in Denver. + +**Bill Kennedy:** I didn't even know that. + +**Mat Ryer:** Yeah, yeah. + +**Bill Kennedy:** But you were coding in Go before you walked in there, or you were just like -- + +**Mat Ryer:** No, we'd built Testify before that conference. + +**Bill Kennedy:** Oh... + +**Mat Ryer:** Yeah, because I started doing it when it was 0.56; before it even got to 1.0. I remember oserror type, and stuff like that. + +**Bill Kennedy:** Yeah. I started in 1.2. 1.2 just got (I think) released, and that's when I started. See, what's interesting is you started before 1.0 release, right? + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** Which had to be -- what made you do that? Because that's a big risk, right? Like, I just heard Google announced another potential programming language to replace C++, and now my brain can't remember what the name of this thing is. It's a weird name... They're announcing it. And my brain's going "I don't wanna go anywhere near that right now..." So you're like that person that saw an announcement... + +**Mat Ryer:** Yeah, well, I was building something for App Engine, and it just supported Java, Python, or this new Go thing with an experimental badge. And that was before it was proper -- I think it was before 1.0 that it supported that. And I just wanted to do some really simple database stuff, and I've found it very easy to do in Go, because it was very clear -- I don't know why; it was just easy to pick up. But do you think teaching Go is easier than it would be teaching other languages? + +**Bill Kennedy:** I've taught C++ before. I taught at a vocational level. So when I was maybe in my mid-20's and my wife at the time go pregnant, she couldn't work, I had to pick up a second job. So I picked up a job teaching vocational at night. I was able to make an extra thousand dollars a month teaching these classes, three hours Monday through Thursday. + +So I was teaching C++ at the time, because that's what I was coding in. And when I look back on that, I don't think teaching Go was necessarily any easier than C++. I think it's about understanding what your student needs in order to be productive. If you're doing training, I believe it needs to be practical. They need to be able to walk out of that room maybe at the end of the week and go back to work and feel more empowered... + +**Mat Ryer:** Right... + +**Bill Kennedy:** And sometimes I have to tell people "Don't go crazy refactoring right now. Just moving forward, and if you have this opportunity..." But I think any training, I don't care what it is that you're teaching, if the student walks out of that and doesn't have the ability to use it -- and maybe not on Monday, but... Then what are you doing in there, right? + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** And I think the same thing -- I say this all the time for the stage talks. If you're giving a talk about tech X, and I'm using tech X, and I don't walk away from that talk knowing something that I can practically use... People like academic talks. + +**Mat Ryer:** \[25:57\] Yeah, I know, they do. I know, they do. And there's value in it. I quite like the academic ones, but there's nothing like that being able to actually do something with what you've learned. For me, learning in the context of some real problem that you have is the way I learn... And I did my book -- the Blueprints book was about building real little projects. And some people complained "Oh, it didn't go into all the depth of all the concepts and stuff first", and it's like, "Yeah, I didn't." I deliberately didn't. + +**Bill Kennedy:** No, but that would be a 10-volume book too, which people don't understand. You have maybe 20 pages, 25 pages to explain something in a book, and you have to really decide what level... Or you go off to tangents. But how cool is it in a conference -- I've been in really good talks, when the person in front of me opens up their laptop, brings up the code and they start... Like, for me -- + +**Mat Ryer:** That's a win. + +**Bill Kennedy:** That's the win, right? Or somebody coming up and asking you questions about their job. "Hey, I'm doing..." To me, that's what I wanna see, at least for me, in trainings at talks; that's what I wanna see. + +**Mat Ryer:** No, that makes a lot of sense. So do you have to figure out the level of your audience before you're teaching? It's difficult at a conference because you're gonna have people at all levels. How do you deal with that? + +**Bill Kennedy:** If I'm doing the 20 to 30-minute talk, then no, because at that point you've already described hopefully what it is you're gonna explain and show, and if it's one of these multi-track, you've labeled "This is for intermediate. This is what I'm gonna do. Either it's for you or not", whatever. + +If you're in a workshop, then I think it's really important to try to gauge who your audience is. Now, if you're super-lucky, 80% are gonna be at the same level. Regardless of what it is, they'll be at the same level. If you're super-unlucky, there's an even distribution of beginners, intermediate, and advanced, which is what's gonna happen to me tomorrow... + +**Mat Ryer:** Yeah... \[laughs\] + +**Bill Kennedy:** Right? Because there's so many new -- everybody I'm meeting here, it's their first conference, they're mostly new to Go... So you're gonna have the mix. That's the challenge. But I always say this to everybody - it is better to overwhelm someone, than underwhelm. Always. + +So I'm always gonna try to -- again, if it's blended, I'm gonna target the intermediate; I'm gonna throw inside advanced stuff. I might just say, "Okay, now for all of you that are -- listen to this. It's not gonna make any sense to you, but..." And I pull them back in. + +And then the other thing which is critical, if you're in a room -- it's a long day conference and you have this mix. What I'll tell the more advanced students is "There's material here for you, but there are gonna be moments where you're gonna say "I know this." So what you need to do is understand that you're gonna be a teacher at work. At some point you're wearing two hats: you're a student, and you're a teacher. So what I want you to do on those moments where you know this, I want you to focus on how I'm teaching it." + +**Mat Ryer:** Right... + +**Bill Kennedy:** And whether you like the way I'm doing it, then use it. If not, then how would you do it different? + +**Mat Ryer:** Hm... Either way, there's something there, yeah. + +**Bill Kennedy:** But focus your brain on what I'm saying and why I'm saying it; not for you to learn, but to learn how to teach. + +**Mat Ryer:** Yeah. That's a really good point. Are there enough people teaching Go? Do you think we need more? + +**Bill Kennedy:** There are a ton of new books coming out from new people... And I only know that because -- and you may get the emails, too... "Will you review, review, review...?" Right? + +**Mat Ryer:** Yeah... + +**Bill Kennedy:** \[29:48\] And I think that's good. The interesting thing though is Go has not really changed in ten years. Okay, we added generics, and there's been mass improvements in some internal runtimes. But when people say "Bill, your ultimate Go is on video. It's like two years old, it's gotta be outdated", I laugh. It's not. The semantics haven't changed, we've got backwards compatibility, promises... And I threw generics in the book, and I'm not gonna teach it, because 1) I don't have time, and 2) I don't want you to really use it, unless... Right? + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** So we've got a new book coming out. I don't know -- I try to push people away from the general books. I go "We already have them." We're lacking - and I say this all the time, and nobody wants to... We're lacking standard library package books. If somebody's listening to this and they're thinking about a book, do not write another Go in Action, and Go PL. It's been done, we don't need another one of those. You're -- I don't wanna say wasting time, but you're wasting time. What we need is a catalog of standard library books. How do you use a strings package, bytes package, time package? That would be -- and these are nice books; these are 40, 50-page book, and you can sell them individually, and you could have this whole collection. + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** And every day, I would pick up one of them - because I have to google time, and bytes, all the time, because I don't retain that stuff... Right? I just need to look at something. + +**Mat Ryer:** Yeah. And it's still protected by the backwards-compatibility promise, so you've got longevity in the material as well + +**Bill Kennedy:** That's right. You've got it. So on the new book, what I did is I self-published, so I can update it when it's absolutely necessary. Because once you go to print, you're kind of locked in. + +**Mat Ryer:** What do you do - go around to people's houses and give them new pages to sell a tape in + +**Bill Kennedy:** Well, they can always download the PDF new, which is good... + +**Mat Ryer:** Oh, so they get that for the life of it. + +**Bill Kennedy:** For the life of it. You just grab all the new PDF. If you wanna get another print - that I can't do anything about, because that's Amazon cost... So we have that. We haven't had conferences, really... I mean, we've had conferences, but for some reason me personally - I don't watch conference talks. Maybe I've watched a handful of them over the years, but I watch it for entertainment. I rarely watch it because I'm... I don't know why. That's interesting. + +**Mat Ryer:** Yeah. You don't do that to learn, you mean... + +**Bill Kennedy:** No. Somebody I know is giving this talk, and I wanted to see it... Or maybe a handful of times... But I don't have patience for video. I love reading, because I can read very, very fast... The video, I have to put on 2x, and try to parse it. + +So video for conference talks for me are not where I go. And I've been asking people where they've been getting, and mostly it's been books, and then searching online. So you've got the books, I think they're valuable, I'd love to see a whole other set. Video - I do video, because-- And I think that stuff is in. But I've never looked at Udemy, I've never looked at Pluralsight, I've never seen anything negative about them, so then it ends up becoming a cost, I think. + +My big problem with video is I'm gonna be more expensive than you going to Udemy. Even though I give away a ton of video for free on scholarships... People don't really know that they can just ask me, and I will give them discounts. I do it all day. + +**Mat Ryer:** I hear stories all the time though, that they saved -- "Oh yeah, Bill just gave me this thing, because I'm a student, I couldn't afford it", and they got in touch, and... + +**Bill Kennedy:** I didn't do this for you not to see it, or read it. Right? But at the same time, I'd like to believe that the quality of material I'm providing is going to be better than anything else out there. + +**Mat Ryer:** That's why you do it, right? + +**Bill Kennedy:** At least I have that belief system. So I don't mind the sales team trying to sell three or four sets of videos for -- I don't even know what they sell it for. Maybe $800. But I don't honestly know how many times we actually sell that free \[laughter\] I'm constantly being told that "Bill, this isn't a charity. This is a business." + +**Mat Ryer:** \[34:14\] \[laughs\] Yeah. That's why you need those businesspeople, isn't it? + +**Bill Kennedy:** Yeah, that's why I need Ed, and Miguel, and all those guys. But I also didn't do it for it sit on a shelf... So if you want material that I have, all you ever have to do is reach out and we'll figure it out. + +**Mat Ryer:** Yeah, that's great. + +**Bill Kennedy:** So I think that stuff is really good. And then what else is there? The books, video... And then I guess googling everything. + +**Mat Ryer:** Yeah. What's the big challenge educating? What's the thing that people struggle with the most, in your experience, when they're learning Go? + +**Bill Kennedy:** I break up my training in two parts. I call it a micro-level understanding and a macro-level understanding. So the micro-level understanding is when we get down to lines of code. That's the readability, the simplicity, the syntax, the idioms. And you need that. If you really wanna engineer -- there's programming and there's engineering. And I tell people that are working with me, "Do some programming right now. Just give me code that works. I couldn't care less if I didn't see a single error, if. I'm not looking for that right now. Once we've got code that works, now we'll talk about engineering. We'll engineer." + +So at a micro-level, it's critically important to have those idioms and understand what it means for something to be readable. And then learn how to refactor for simplicity. But after a week of Ultimate Go, I don't think you have the ability to go back and be as productive as you could. It's not necessarily a practical -- I almost call it an academic-level class... But you need it, because it sets the stage for the next Ultimate service which I'm teaching here, which is your macro-level. + +**Mat Ryer:** Right. + +**Bill Kennedy:** Now we're gonna talk about architecture, now we're gonna talk about project structure. Now we're gonna talk about macro-level idioms, and practical things like building and logging and configuration. And I don't wanna have to have the micro-level conversation when I'm doing that. + +**Mat Ryer:** Yeah, yeah. + +**Bill Kennedy:** But I think you need both. So the hardest thing is if I get somebody in my macro-level class, that hasn't taken my micro-level class, I have to fight to not take tangents, and just say "You know what - I'll get you the video. Let's do that." And on the micro-level class, I have to remind myself that - I know that there are gonna be people here feeling like this was a great class, but it wasn't a lot that I could do, other than maybe change the way I wrote some functions here and there. + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** I don't think people understand how much they get out of the micro-level class until their product's in production, and now they start hitting the tooling a little bit, and now all the little things I said... Because I've gotten that fever. You know, data class and then we started having problems in production, and like your voice just started flying all over my head. + +So I think as a teacher there, it's -- I don't know, I'm always fighting to make sure I'm giving the person what they need... And the struggles are just different. Everybody's struggling with different... + +**Mat Ryer:** Yeah. But can people learn -- I feel like you'd be able to learn the macro stuff. You have to sort of be okay with not having the complete picture down to that detail, but you can still learn concepts, and retrofit that. Or do you find that you do need that foundation to build on? Is it more that way around? + +**Bill Kennedy:** \[37:56\] I think with Go we've already learned that within three weeks you can build just about anything you want, and it will run, and it will run pretty dang good. And if you wanna leave it like that, you can. We've learned that already. So what comes down to you is what level engineer do you wanna be? You just wanna be somebody who's programming, and does the bare minimum? And that's fair... + +**Mat Ryer:** Yeah, sometimes that's appropriate. Like, if you just have to solve a problem that just you have right now, you wanna process these files and change them, or... And that's great. + +**Bill Kennedy:** That's it. And it's fair. And you're the only one who's gonna worry about this... It's totally fair. We got it to -- remember back in the day when the community got crazy, like if you weren't writing idiomatic Go, you were evil? + +**Mat Ryer:** \[laughs\] + +**Bill Kennedy:** Like, what the hell, guys? So I'm glad I don't see that as much as we were in the beginning, where maybe there were more academics than there were people just saying "Tell me what to do." + +**Mat Ryer:** Yeah... + +**Bill Kennedy:** So I think that's fair. But if you wanna be a better engineer, and really think about readability, and maintainability, then -- + +**Mat Ryer:** Yeah. You've got a big team, or multiple teams working on a codebase that's gonna have a long life, then it's a different game. + +**Bill Kennedy:** Yeah. And you don't wanna know... On all of my teams, when I start a new project, I tell everybody "I don't wanna know who wrote any line of code when I'm looking at it. If I know who wrote this, we've got a problem." + +**Mat Ryer:** That's interesting, yeah. That's funny, because with `go fmt` and the fact that there isn't lots of different ways to do things in Go - there tends not to be - you do get sometimes this effect where you look at some code and it feels like you've written it. And that's such a -- + +**Bill Kennedy:** It is. But let's talk about where that can go wrong. Variable declarations. + +**Mat Ryer:** Okay, yeah. + +**Bill Kennedy:** Right? So I'm teaching all the time, var for zero value construction. I don't wanna see short variable declaration operators, I don't wanna see empty literals unless we're on returns, right? So there I can pick out that "This is it." + +**Mat Ryer:** Yeah. Because there's multiple ways to do it, that's why... Isn't it? + +**Bill Kennedy:** To me, it was a necessary war to keep the compiler simple, but there's a war there. It's one of the first things I teach in the micro - I don't care what you do here, but have a plan. + +The other thing is I hate the else clause. I hate it. Use a switch. Use the naked switch. It's much more readable. So the moment I see an else clause in a piece of code, I freeze, actually. It's horrible. I get stressed out... Like, "This has gotta go." + +**Mat Ryer:** Have you ever considered making like a Bill linter, where you put your opinions into this tooling? + +**Bill Kennedy:** No, because the static check linter is really good, and everybody should be supporting Dominic on that project. We do at Ardan with the GitHub contributions. + +**Mat Ryer:** Ah yeah, sponsors. + +**Bill Kennedy:** Yeah. Everybody who's using that tool should be giving them something every month, because what he's doing is amazing. Now, I did start writing what we were calling the Ardan Playbook... + +**Mat Ryer:** Okay, yeah... + +**Bill Kennedy:** ...which was all of this -- and I got about ten pages in, and it was so tedious to write... + +**Mat Ryer:** Why? + +**Bill Kennedy:** Because this wasn't kind of conversational things that I did. This was literally-- + +**Mat Ryer:** Complaining. \[laughs\] + +**Bill Kennedy:** Not even complaining. It was just very dry. It was "Do this, and not this. Then this, and not this", and after about like a day, my brain just started to shut down, and I couldn't -- and I have a little bit of it out there, but... I did put a lot of it in the Ultimate Go Notebook, but it's not like formally listed, like I was trying to do with the Playbook... I don't know. + +**Mat Ryer:** Yeah, I did a talk at Gotham Go in New York City called "Things in Go I don't use", or "Things in Go I never use", or something. Else is one of them, and I make that same point, of like -- you know, you sort of hide the happy path in this indentation, and things like this... + +\[42:02\] So yeah, I agree. Anything we can do... It's really thinking about the usability, the user experience of your code. Because that's why you get away with a lot when you're the only one that's going to be working on it. As soon as you've got more people that are gonna be maintaining this code, then you've gotta think about them. + +**Bill Kennedy:** But it's bigger than that... Because look, here's the reality - I've been in this industry for 30-something years. To date, that I know of, I've got three projects still in production. One that's over 20 years old, one that's over 10 years old, and one that's a couple of years old. So what I don't want for people is when they finally finish and they retire -- their legacy is that nothing they've done over the last 30 years is still adding value on the planet. This is not what you want. At least that's not what I want. + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** So if you're not writing code for the next person... If the code you're writing, you're not thinking about the next person, two things happen when you leave - either the next person says "This is shit, and we're throwing it away and rewriting it", which could also now mean it ain't in Go anymore... Right? + +**Mat Ryer:** True. + +**Bill Kennedy:** Or the other thing that's gonna happen is it just gets completely abandoned... And now it's almost like it never existed. So what keeps you up at night - finishing in this industry and not having anything to show for it. This isn't like we're building a house that's gonna be here a thousand years from now... + +**Mat Ryer:** That's true, that's true. It can be so transient, can't it? + +**Bill Kennedy:** Yeah. So for me, that's like the nightmare situation. How many late nights did you have at the office at a previous job because you had to finish this code? And guess what \[unintelligible 00:43:54.02\] today? You realize how insignificant that was, because it's not even running anymore, and you ruined your day... + +**Mat Ryer:** Yeah, absolutely. I've experienced that. And I have to remember it when I'm now working late to do this. It is worth remembering, in a few years this won't be important. I do have to keep -- I completely agree. But with education people though, a legacy - that's a multiplier of what contribution you're making. And I don't know that you get visibility into it. You probably have no idea the impact you've had by all the people you've taught... But that's gonna be big, if you think about it... + +**Bill Kennedy:** There are times when somebody comes up to me and they say -- so here's the most recent... Just to have this conversation. And I know everybody hates blockchain, but Ethereum is moving now to their new proof of stake blockchain, right? They're very close to putting it in production. I met one of the developers on that team, and he was like "Bill, we've all watched your videos when we were beginning to build all this code..." And my brain went "Whoa..." That's kind of surreal, right? Because I'm not directly coding, let's say Ethereum blockchain proof of stake, but there is some aspect of what -- + +**Mat Ryer:** Yeah, your contribution. That's the thing, there's gonna be loads more that you'll never even hear about. There's gonna be loads of that, and you won't hear about most of it. So you should be happy, I think, that that's the difference you're making when you educate gophers. + +**Bill Kennedy:** And that's for anybody, because again, everybody's an educator at the end of the day, whether you're doing in front of people or you're just doing it at work. You're constantly educating. + +\[45:46\] I think one of the big jobs that are missing in the industry, and we can't have it because there's no way to measure it, is that every company should have the tech lead who's a floater. This is what I want; if I ever sold Ardan, I wanna work at different companies for like a month to three months. I have no responsibility for product, but I would -- let's say I went to Grafana, I sat down with you and your team, and every day I pair-programmed with somebody different, for a couple of days, to help them learn how to refactor, look at the code, think about it, and make everybody a better - at least based on my design philosophies and guidelines... But try to make everybody better on that team, and get everybody rowing the boat in the same direction. And then move on to the next team, without having any responsibility for it. But the problem with that role is after you do that for a year, you don't have a single commit, right? + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** You have no record of your existence. + +**Mat Ryer:** That's right. But maybe that problem needs to be solved then. Maybe there ought to be a credit system or some way of recording that this was influenced... I mean, Mark Bates does it in his projects. He has a shoulders.txt, and he's basically listing the people that otherwise you would know, that have helped to make -- + +**Bill Kennedy:** I didn't know that. I love the idea of the shoulders.txt. + +**Mat Ryer:** Yeah, standing on the shoulders of these giants. + +**Bill Kennedy:** Yeah, because then at least at review time you could see your name across -- even if you put the number of hours that you provided, that would be... It's a role that's missing. I think you could get your teams up to speed within a couple of months, while being productive. You're not losing a week to training. + +**Mat Ryer:** Yeah, and it's relevant. They're learning in the context of the problems that they're really solving, and that to me is always the most valuable of learning anyway. + +**Bill Kennedy:** On the projects that I'm in at Ardan my mornings are usually a couple hours of pair programming with whoever I'm working with, and then in the afternoon I set them off to go back and do usually more programming than engineering, and then we do the engineering in the morning. + +**Mat Ryer:** Right. I see. + +**Bill Kennedy:** So "Go code this. Just make this work." + +**Mat Ryer:** That's nice, because it gives people permission to just crack on and do something, which is what often -- we need that permission to do anything, especially if you're learning it and you feel like you don't know how to structure this, or whatever. Like, stick everything in Main, get going; get something working, and then we'll iterate on it. I really like that approach. + +**Bill Kennedy:** But here's the cool part -- and I might take questions during the day on little things, but here's the cool part. The cool part is after usually about four weeks, they're engineering while they're doing the pro-- + +**Mat Ryer:** Yes, that's right. + +**Bill Kennedy:** So there's less and less refactoring. That's what I'm looking for. If I'm four weeks in and I'm still doing the same level of refactoring I was, this ain't working out. And usually, by week eight it's almost like my life is really good now. "This is good, this is good, this is good. Go." And that's what I think this role should be. It's an educational role in the company. + +And then -- okay, we didn't document the handbook, but I've taught the handbook to everybody over a couple months. So the next person that comes on, now you do the same thing and your code is consistent. + +**Mat Ryer:** It's like an up-leveler. You're being an up-leveler as a role. + +**Bill Kennedy:** Yeah. + +**Mat Ryer:** But it is something that's important and really valuable. But you're right, if you can't really measure the impact of it, what's the incentive for people to be doing that? Maybe it's not there. + +**Bill Kennedy:** Well, the bean counters who don't see it, just look on spreadsheets and go "We're paying this person X amount a year, and..." + +**Mat Ryer:** Where's the code...? + +**Bill Kennedy:** Yeah, they didn't add anything to the bottom line. Let's get rid of them. + +**Mat Ryer:** Right. Yeah, well, that's a mistake... + +**Bill Kennedy:** Yeah... So to me, that would be the role I would want after I'm done with Ardan. I think that would be so much fun. + +**Mat Ryer:** Yeah, I think that would be fun as well. It sounds great. + +**Bill Kennedy:** And you have no responsibilities, so you go home and your phone's never gonna ring. + +**Mat Ryer:** Yeah. \[laughs\] + +**Bill Kennedy:** That's the key. + +**Mat Ryer:** Living the dream. + +**Bill Kennedy:** You're living the dream, yeah. + +**Mat Ryer:** Bill, thank you so much. It was so great to talk to you, and I'm excited for your talk. You're actually gonna be speaking now, unexpectedly, at this conference, right? + +**Bill Kennedy:** \[50:11\] Yeah, it tends to happen most of the time when a speaker can't make it to the conference. + +**Mat Ryer:** Yeah. \[unintelligible 00:50:15.14\] unfortunately is not able to make it here... But you've stepped in. What are you gonna be talking about? + +**Bill Kennedy:** So I only have half an hour, which if you know me... + +**Mat Ryer:** Not enough. + +**Bill Kennedy:** ...I just get going in about 30 minutes. + +**Mat Ryer:** Yeah. Just start 30 minutes before backstage. + +**Bill Kennedy:** Yeah, it's gonna be bad... And I feel like when you're on-stage, which - I really don't like talking on-stage. + +**Mat Ryer:** Yeah, that surprises me. + +**Bill Kennedy:** I think it's a different experience. It's more stressful... I'm not a good speaker on-stage; I'm good in the workshop, because I can be really fluid and flexible. + +**Mat Ryer:** Do you rely on that interaction more then with people? Because often you don't get so much of that when you're delivering to an audience. + +**Bill Kennedy:** I think a good speaker talk is always really well prepared. The slide deck - it's well prepared; you know your message, you practice it a hundred times. + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** That's not me. I'm more like "What's going on in the room? Let's feed the room." + +**Mat Ryer:** Right. + +**Bill Kennedy:** And I only have 30 minutes here. So what I decided is to take a small bit of Ultimate Go that I've done and just show in a practical way how you can use benchmarking for profiling, in some compiler switches, to find non-productive allocations, and fix them, and do all live coding, and... + +**Mat Ryer:** Ah, again very practical stuff that people will be able to use immediately. + +**Bill Kennedy:** Practical stuff. And because everybody here is so new - because I've been sheltered for the last 2,5 years - it's probably gonna be a lot of new material. It's new to them. + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** I mean, I've been doing this particular exercise for half a decade... So I always have to worry, like, how many people have seen this? And my guess is it's gonna be very, very little. So at that point, it should be a good talk. And you can see my style; I can't do slide decks. + +**Mat Ryer:** Well, I can't wait to see it. Thanks so much for talking to me today. + +**Bill Kennedy:** Alright... Thanks, Mat. + +**Mat Ryer:** Thanks, Bill. + +**Jingle:** \[52:18\] + +**Break:** \[52:38\] + +**Mat Ryer:** Well, that beautiful sound there was just created by Egon. How have you done that? + +**Egon Elbre:** Well, I came here and did a TinyGo workshop, and I took some buttons and mapped them to a trumpet \[unintelligible 00:53:44.26\] then control it with Ableton + +**Mat Ryer:** Yeah. So you had to assemble this yourself on the breadboard. + +**Egon Elbre:** Yeah, exactly. + +**Mat Ryer:** It's got three switches. Can you explain what these components are here? + +**Egon Elbre:** \[54:00\] Yeah, it's an Arduino Nano RP-something... And then it's got three buttons connected to it. And yeah, then there's some magic to map the three buttons to some midi note. + +**Mat Ryer:** Ah, so you've got Go code that's running on there... + +**Egon Elbre:** Yeah, exactly. + +**Mat Ryer:** ...and that is translating into the midi notes, which is then played out through Ableton + +**Egon Elbre:** Yes, exactly. + +**Mat Ryer:** So it's not too many pieces you have to put together to make that work. + +**Egon Elbre:** No, not really. It's 90 lines of code, and I have different scales and stuff here already... + +**Mat Ryer:** That is so cool! And how was your -- this roundtable discussion yesterday; remind us what you talked about. + +**Egon Elbre:** Oh, it was excellent. There was a static analysis discussion. We went through what linters everybody is using, and how do you build them, and what could be made easier to make linking easier to write. And of course, what's the other thing to do? A debugger. Pun not intended... But why are debuggers so powerful, but nobody is using them properly? + +**Mat Ryer:** Yeah. + +**Egon Elbre:** So... Lots of interesting topics. + +**Mat Ryer:** Well, yeah, the interesting thing -- because I've written some linter tools myself, and it's not easy, is it? + +**Egon Elbre:** Yeah. But there are packages, like tools analysis, and the rule card, rules that do help you make it easier. But there's probably -- we could do something where we do like a structural pattern detection on bad examples and good examples, and then figure out how to convert it into a linter. + +**Mat Ryer:** Oh, that's interesting then. So almost like using good code as training data for... + +**Egon Elbre:** Exactly, yeah. And if it's wrong, then you just add an example to one or the other, and it will get better. + +**Mat Ryer:** Oh, that's kind of like GitHub Copilot, but without the machine learning and the SaaS. + +**Egon Elbre:** I guess so, yeah. \[laughter\] + +**Mat Ryer:** And we mean SaaS in both ways. GitHub Copilot is quite sassy sometimes... I asked GitHub Copilot, just in the comments, "Are you alive?" and it said "Yes." + +**Egon Elbre:** So... What else do you need? + +**Mat Ryer:** Exactly. I'm sold... But thanks, Egon... Enjoy. Maybe you can play a bit more for us... + +**Egon Elbre:** Sure. \[56:32\] Okay? + +**Mat Ryer:** Nice. + +**Mat Ryer:** Oh, I can't believe it. @deadprogram, Ron Evans, in real life, and in this timeline. + +**Ron Evans:** Here in Berlin, of course. I wouldn't miss GopherCon EU. I mean, I don't get out much... + +**Mat Ryer:** Yeah. And what are you doing here? + +**Ron Evans:** Well, we have our TinyGo Hardware Hack Session today... + +**Mat Ryer:** Amazing. + +**Ron Evans:** We brought all sorts of hardware from our secret lair, for people to check out lending library style And then using that, they can hack on hardware, fly drones, program gopherbots, program IoT controllers, and even our TinyGo music jam. + +**Mat Ryer:** Oh, that I can't wait for. I was talking to BuBu earlier He was making a drone fly around, before someone nearly squashed it. + +**Ron Evans:** Well, he's actually been consuming the batteries for the drones at a very rapid pace, and... He's definitely getting his drone pilot license after this, I'm telling you. + +**Mat Ryer:** Yeah. \[laughs\] So let's have a quick look at some of the stuff you've got here then. I mean, obviously, it's a podcast, so people can't see it. We're gonna have to describe it... But there's these cool pins, and stuff. + +**Ron Evans:** Oh, these are really awesome TinyGo pins made by ConejoNinja my awesome collaborator and colleague. He makes wonderful toys of all different kinds... So he 3D-printed these cool, little, tiny gophers, which - you've gotta take one, in your appropriate color... + +**Mat Ryer:** Adorable... Yes... + +**Ron Evans:** We have these awesome pins, TinyGo stickers... But the thing he made that is the most amazing is these life-size gopherbot helmets. + +**Mat Ryer:** I know... I can't believe this. + +**Ron Evans:** \[58:05\] We've got those photos that we took a few minutes ago, that you'll have to share... Because honestly, we look better in those helmets than we've ever looked. + +**Mat Ryer:** I know... To be fair - yeah, I feel like it's sort of like Daft Punk, but imagine Daft Punk was struck by lightning. + +**Ron Evans:** Daft Gopher. + +**Mat Ryer:** \[laughs\] + +**Ron Evans:** GopherPunk. + +**Mat Ryer:** Yeah. Yeah, they're just so great. And then there's lots of components and things for people to build... + +**Ron Evans:** Yes, we have fully assembled gopherbots for the people who want to explore the software side of hardware... And then we have all these IoT sensor kits, with individual sensors for people who actually want to touch metal. They want to plug in cables that control individual LEDs. We've got that for you as well. + +**Mat Ryer:** Yeah. As I'm talking, the robot's ears are flapping. + +**Ron Evans:** Yes, the ears move, they're on servos... That was one of the things people asked about the original tiny gopherbots, is "Do they move?" I'm like, "No, they don't move. It's an emotional support robot." You hold it close to you and you tell it what you need to tell it, and it just listens. It doesn't respond. It's just a perfect listener. + +**Mat Ryer:** \[laughs\] Yeah, nice. + +**Ron Evans:** But you know, for conejoninja, he really had to go to the next level... He had to go larger, first of all; full life-size. If you want to become one of the gopherbots, then... You know, it's got the cyberman style... + +**Mat Ryer:** You could get away with wearing them around Berlin, I think. + +**Ron Evans:** I think -- I saw someone wearing something very similar last night... + +**Mat Ryer:** Yeah... + +**Ron Evans:** But he's got these awesome TinyGo-controlled ears that have small servos... So eventually, we'll hook up a brain-computer interface and then they'll move based on your mood. + +**Mat Ryer:** Amazing. So how does it work in code then? Is there an API for these particular things? Do you send messages? What's the actual interface like for the programmer? + +**Ron Evans:** Well, it's all just Go code. And we do have APIs... We have a low-level API called the Machine Package, which in TinyGo lets you actually control the different individual LEDs at the very low level... + +**Mat Ryer:** Right. + +**Ron Evans:** And then we also have a drivers package, which has got somewhat higher-level, for things like displays, and sensors... If you wanna read the temperature of a particular known sensor, you don't wanna have to figure out the low-level protocol details. You just wanna read the temperature and then do something with that information. + +**Mat Ryer:** I see. But that's cool that you still have the option of that really low-level stuff. That is very appealing, especially if you spend a lot of time in higher-level software. The fact that this also reaches out into the real world I think is really -- it gets really exciting. + +**Ron Evans:** And we have another layer on top of that, with some packages... There's a couple of packages specifically for displays, the TinyDraw and TinyFont packages. This is actually treating these LEDs on the helmet as if it was a kind of small display... + +**Mat Ryer:** Yeah, it's scrolling text, right? + +**Ron Evans:** ...and it's compatible with the same exact APIs as those little, tiny... These very, very small displays that are about the size of a postage stamp, and then you have these large displays and you could use the same types of software, and the same drivers, and the same TinyDraw and TinyFont package, regardless of the size of the display. + +**Mat Ryer:** So how is the text scrolling then? Is that part of the driver tech, or is someone actually redrawing each time, and then coming up with the logic to toggle the LEDs individually, and stuff? + +**Ron Evans:** I don't know. But the programmer himself, conejoninja is right next to me and knows all the answers to these wonderful things. + +**Mat Ryer:** Well, that sounds like a great segue. Hey, how's it going? + +**conejoninja**: Fine, thank you. + +**Mat Ryer:** Yeah, so how does the text scroll on this display? Do you have to calculate what it's gonna be and then update things individually? + +**conejoninja**: No, we just have these fantastic libraries, TinyFont and TinyDraw. With TinyFont, you just say "I want to print this text, in this position." So we just move the X position... + +**Mat Ryer:** Ah, I see. Okay, so you have some kind of loop that you manage then, with a sleep in there, I assume, for the delay. So you're drawing it and offsetting it. That is very cool, because it looks very familiar. It just looks like a -- we're used to seeing this kind of scrolling text display, but you're controlling that yourself, which is really cool. And how long did it take to build these helmets? + +**conejoninja**: A couple of months... But I've been busy. + +**Mat Ryer:** Yeah. Well, I can see you've been busy. Are you worried about them coming to life? Have you thought about that? + +**conejoninja**: Yes. + +**Mat Ryer:** Yeah, a little bit... + +**conejoninja**: \[unintelligible 01:02:44.10\] + +**Mat Ryer:**Yeah, that's the plan, isn't it? \[laughs\] Well, this is it. I look forward to these. Because this is gonna be like The Terminator, but they're adorable... Yeah, so... In a way, be nice. + +**Ron Evans:** We're all gopherbots now. + +**Mat Ryer:** \[laughs\] Thank you very much. Great stuff. And if anyone's interested, check this out online, because the TinyGo project is really great. + +**Ron Evans:** TinyGo.org. + +**Mat Ryer:** Okay, there you go. + +**Jingle:** \[01:03:16.26\] + +**Mat Ryer:** Well, that's a wrap from GopherCon EU, and... I'm a little bit exhausted, but it was such a good -- I've met some amazing people, had some great conversations... I hosted with Ale, Natalie, Jessica and Vee. Co-hosting there was great fun. The Gophers Say podcast that we recorded live was great fun, and it went better than I had hoped... And yeah, the talks were so good, the conference, all the staff... Natalie and Donna and the team that put this together... I don't know how they do it; it's just so much work, and they're just so much fun. So if you get the opportunity in the future to go to GopherCon EU, I really recommend it. And they do a diversity sponsorship/scholarship thing, too. So if you feel like there's no chance of you getting there, reach out on Twitter to them, because they make it available to everybody. + +Great. Okay. Well, that's it from me, from Berlin, and from GopherCon EU. I had a great time. I'm off back to London now... \[wanker!\] You're welcome! diff --git a/Might Go actually be OOP_transcript.txt b/Might Go actually be OOP_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..fcc07cecd90a96ce4121fad080d57edc17134e40 --- /dev/null +++ b/Might Go actually be OOP_transcript.txt @@ -0,0 +1,305 @@ +**Natalie Pistunovich:** Good evening, afternoon, morning or night, everyone who's joining us live, or listening to this later. This episode Ian and me are being joined by Ronna to talk about object-oriented programming. And we'll start by introducing Ronna, I guess. Ronna, you are an engineering manager at Delivery Hero, a Google Developer expert for Go, a Women Who Go organizer in Berlin, Go Time's unpopular opinion Hall of Famer, and after 20 years in tech, you know that the sum of the opportunities that were given to you - this is why you're basically giving opportunity to others. This is an interesting thing about your bio. We'll definitely ask you more about this. + +Now we're here to talk about these workshops that you've been crafting since 2017, that you've been giving to meetup Women Who Go Berlin. The most recent one that you'll happen to give at GopherCon Europe later this month, on the topic of actually object-oriented design in Go. + +So as a preparation for this episode, we were figuring out how each of us pronounces the short term for object-oriented programming. Ronna, how do you say that? The three-letter acronym. + +**Ronna Steinberg:** You mean OOP? + +**Natalie Pistunovich:** OOP? Well, I guess it's two versus one, Ian. You pronounce this the same. + +**Ian Lopshire:** \[04:13\] Yeah, OOP. + +**Ronna Steinberg:** I have to say, for the past month or so I've been writing OO a lot, and it looks like "Uh-oh..." the emoji. \[laughs\] And it feels very symbolic to the state of object-oriented. + +**Natalie Pistunovich:** I somehow always thought it's Oop. + +**Ronna Steinberg:** You thought it was Oop? + +**Natalie Pistunovich:** I guess it's because last time I used this term was in university, and it's been a while since. + +**Ronna Steinberg:** And you used to say Oop? + +**Natalie Pistunovich:** Yeah, we used to say Oop. + +**Ronna Steinberg:** Ian, what do you say? + +**Ian Lopshire:** I definitely say OOP. I don't think I've ever heard someone call it Oop before, until ten minutes ago. + +**Ronna Steinberg:** This is definitely a debate worth having. + +**Natalie Pistunovich:** Yeah, that can be a poll. Maybe this can even be my unpopular opinion. There. I thought I had none. So we can say is sort of one big unpopular opinion, coming with the claim, Ronna, that Go can be an object-oriented programming language. + +**Ronna Steinberg:** Well, I mean, I definitely don't think it's not... \[laughs\] So what happened was that I wanted to share with my team how to do object-oriented programming with Go, and then I kind of realized by looking at actual interview questions/answers/feedback how things are happening in my organization, what people think Go is and what people think Go isn't. They definitely think that Go is not an object-oriented programming language... So I decided that "Okay, I'm going to give a workshop about this." Try to start like if you want to do it, how you should do it. And this has always been how I do things. When I teach Go, I try to teach people -- if they want to do something, I try to teach them how to do it, not whether they are right or wrong to think or do things in a certain way. + +So it's more about how to use the tools that you have with the language properly, and especially now with generics going into the language, solving a last kind of edge case that we didn't have... So yeah, I posted that I was going to do that internally in the organization, and immediately, a lot of people that I had never met reached out to give me honest feedback about the workshop that I have never given yet. \[laughs\] And then I kind of knew that "Okay, I have a topic that is definitely worth exploring." So yeah, I guess it is my unpopular opinion. + +**Natalie Pistunovich:** So tell us about some of that feedback or opinions that you got. + +**Ronna Steinberg:** \[laughs\] It's quite amazing... So everybody has an opinion. Some people were with me and wanted to share that they agree that Go can be object-oriented, which was really nice. What was really nice about it was that everybody can find a resource to support their claim, which I found amazing. It's brilliant. I can't even begin. Now, I'm not an academic, obviously, and what I do, I do by trial and error. But I did read a lot about it for the past few months since we kind of decided that I was going to do this, and... The answer is I don't know, but I am leaning towards yes, it is an object-oriented programming language. And it's because of multiple features that it has. + +To explain why I think it's an object-oriented language, I should probably explain why I think it's definitely more object-oriented, for instance, than other languages that are considered object-oriented... And I'm looking at you, Java. And here's my case. Listen up. \[laughs\] Java - yes, sure, it has inheritance, but you can only inherit once. So that means that if you want to express that idea that class A or an object of type A is also an object of type B, or a class B, you can only do that once. And if you have a case C, that A is also C, you cannot express that in Java. + +\[08:19\] Now, I came from C++, so I kind of had this multiple inheritance. I used to be very expressive with inheritance, and when I look at Java, when I look at Ruby - Pythonistas, I'm sorry, I'm not familiar enough with your language to make a case... But when I look at Java and when I look at Ruby I see that you can't really express A is C if you already expressed that A is B, unless you use interfaces or generics. + +Now, interfaces are -- Go has this dramatic effect that interfaces are implicit in Java, and a class has to be aware, to be able to instantiate A as type interface I, A has to be aware of I at the time of writing. So that means that to plug A as I, you have to do something. It could be generics, it could be some sort of a proxy class that we'll implement and derive from... So there are some ways. + +In Ruby we also see how modules are kind of replacing generics, and what people used to do in other languages to sort of be able to very expressive that something is also something else. So the answer is that I think that Go is object-oriented. I'm leaning towards yes. I do see why people don't think so, but yeah, that's where I am as a person who works with Go. + +**Ian Lopshire:** Can we go through some of the reasons people said it wasn't? Like, because it lacks inheritance... Like, what features are missing that people are like "It's not object-oriented"? + +**Ronna Steinberg:** I think it's about the styling. So yes, most people talk about the lack of hierarchy, the composition over -- what's the word I'm looking for...? + +**Ian Lopshire:** Inheritance? + +**Ronna Steinberg:** I guess... I mean, I think there is a general term for inheritance, but yeah. + +**Ian Lopshire:** Okay. + +**Ronna Steinberg:** I think people do a lot of procedural coding with Go. You don't have constructors, for instance. And why do you not have constructors? Because anything can be a type. An integer can be a new type... You can define a new type in pretty much any way that you want. So you don't really have constructors, because you don't have classes, and anything can be a type. And then everything can also have methods. But the truth is that without constructors, without a formal way of working with objects, people get lost... And then they look down at people who come in from other languages, who need these things. + +It is very difficult to define a type and allow anybody to make any changes, so you cannot really tell -- how do you tell if it's corrupt or not? How do you write any kind of defensive code in that situation? + +We also have these best practices -- I mean, I don't like the word "best practices", because I think that I usually use the word "common practice" to explain... Because it's not always best. There's a case for pretty much anything. But the common practices are "Everything is public, anyone can do anything with \[unintelligible 00:11:24.15\] So it is a bit tricky for people coming into the language to know exactly what they're supposed to do and how. And in that sense, we're not making their lives easier. We're just making it harder, instead of doing this gatekeeping. "Oh, no, you're coming from Java. You probably don't know how to do Go. Never mind... Forget about it." This will take some time... And it doesn't matter; they could have 20 years of experience, but you'll still look at them like "But do you know Go? Do you?" + +So that's where I am... I am curious though if we're getting any remarks from our listeners. Maybe they have some opinions... + +**Natalie Pistunovich:** Not just yet, but maybe any minute now. + +**Ronna Steinberg:** Hopefully. One can only hope. + +**Natalie Pistunovich:** \[12:11\] I'd be curious to hear what are some of the feedback that you've got over email, like what are the pros -- the people who agreed or disagreed. You said that they were all backing their claims, which is wonderful; it means you have generally a great discussion... + +**Ronna Steinberg:** Well, \[unintelligible 00:12:25.00\] + +**Natalie Pistunovich:** "Expected practices", or what was the term you used? Common practices? That's great, so tell us about them. + +**Ronna Steinberg:** Well, the first comment that I got - that was really funny - was "Aren't generics enough OOP now? Really?" Something like that. \[laughs\] I thought that was hilarious. So people do refer me a lot to the Go FAQs that specifically say that the answer to whether Go is object-oriented or not is "Yes and no." \[laughter\] No, actually, is it "Yes and no"? I don't remember. + +**Natalie Pistunovich:** I actually prepared it for you. Let me read this out. + +**Ronna Steinberg:** Oh, amazing. + +**Natalie Pistunovich:** "Yes and no. Although Go has types and methods and it allows an object-oriented style of programming, there is no type hierarchy. The concept of interface in Go provides a different approach that we believe is easy to use, and in some ways more general. There are also ways to embed types in other types to provide something analogous, but not identical to subclassing. Moreover, methods in Go are more general than in C++ or Java: they can be defined for any sort of data, even built-in types such as plain, “unboxed” integers. They are not restricted to structs (classes). Also, the lack of a type hierarchy makes “objects” in Go feel much more lightweight than in languages such as C++ or Java." + +**Ronna Steinberg:** Well, I kind of agree... \[laughs\] But then I went into the rabbit hole, because what I remembered from my university was my professor, Jeff Rosenschein, explaining that object-oriented is about being able to send messages to objects. And I checked the origins of the idea and it appears to be the case that that is what they were going for. It wasn't so much about hierarchy. Hierarchy came in later. So if we go back to basics, maybe Go is fully object-oriented. And I think to answer that, you probably need to know more than I do about language design. But again, from practice, I'm leaning towards yes. + +In the workshop we do things very Java-like. Exactly the kinds of things that you're not supposed to do... Or it's not really that you're not supposed to do, but we really try to express the same types of ideas, but keeping loyal to, I guess, the Go way of doing things still. So yeah, at the very least it's gonna be a lot of fun, I think. I try to at least add some value in that regard, like give people a good taste of how they can do things and how they can get creative. + +I also think that there is a lot of room for people to decide within their teams their own best practices, their own practices, how they want to do things. I do think there should be a discussion, and I also think that it should be a little bit more lively than it is now, where I think the veterans sort of dictate how people are going to use the language. The users of the language that are coming in are the future; they are going to also decide for themselves how they want to do things, how they want to program. They should have the space to do that. And if they're coming from other languages and they work in a similar way, that's fine. + +You asked me about quotes that I got... Somebody quoted Rob Pike to me. I tried, but I couldn't find the quote... But somebody quoted Rob Pike to me, so apparently Rob Pike did say something about object-oriented \[unintelligible 00:16:06.22\] or something of this sort... But yeah, I didn't get to the source. + +\[16:17\] What I do see is a lot of people -- this is where I think this is coming from... How people write APIs, how people write their handlers - it's all pretty much functional programming. And the way that this is designed, most people don't actually use an interface in these situations, and even if they do, they kind of don't use object, they use the functions as an interface... Which you can do. You can infer. So they're not used to actually working with objects. Or not objects. Let's say structs. Let's just use the word "structs", with fields and actual -- the classic way of doing things. + +I think that's where it's coming from, because it became a language that is so strong in that space. And in that space you really don't need a lot of running objects. Whatever objects you have are probably going to die soon anyways. And if we think also, those APIs - it sounds funny, but objects don't live long in APIs, unless they live the entire lifetime of the application... Like the server, for instance - that could be an object. + +Their structs are limited to fields, like configurations, and they don't have many methods. But then if you look for instance at how they set up their repositories, or how they set up their models, or how they set up these things, because we didn't have generics up until very recently, they don't really use objects there. Or they generate a lot of code, but they don't actually know what's in it, or care so much about it, just to save store, or whatever. But with generics, you can actually store any kind of model into any kind of database or repository. So we are that much stronger in that sense. Maybe now that people are actually going to write code for themselves, they're actually going to spend more time thinking about what they want to express, and how... Hopefully. + +**Break:** \[18:21\] to \[20:13\] + +**Ian Lopshire:** So you going through that, what you've just said, made me think maybe my idea of what OOP is is pretty heavily influenced and biased by writing Go. I haven't written anything but Go in a long time. Probably two or three years at this point. + +So going through that - you're right, I think a lot of the OOP stuff that was taught in school I don't use anymore, but I still think of what I'm doing as object-oriented. And maybe I'm wrong. In my mind, object-oriented is just encapsulating a set of data and having methods on it that -- it's like self-contained, right? And maybe that's not what it actually is. Maybe I'm totally misunderstanding the concept. + +**Ronna Steinberg:** That's interesting... You said it. You actually said it, "encapsulation". So it's about encapsulation, abstraction, and later came in generalization. So to encapsulate, you don't need hierarchy. Or you can have a composition, essentially, and that would work pretty well too, to encapsulate. But essentially, you don't even need to have any kind of composition, I think, to be able to encapsulate information. It is useful, but the combination of information can be across a lot of things... We know this because we have distributed systems. We live in a world where everything is distributed; all of our data is everywhere, and we kind of need to aggregate it... So even a composition is nice to have, something that we are very used to and we do, but we don't actually need to have it. + +If we have encapsulation, which I think we do in Go, and we are able to define abstractions, which we have interfaces for... And by the way, there is a case - it's kind of interesting, I've found this code snippet that was exactly what I needed for this workshop at the time, of C++ code. I really loved it, because it starts with a license that "Do whatever you want with this code, as long as you don't blame me." \[laughs\] And then this code - it's an event that you can register listeners to, regardless of their type. Essentially, it works with templates, generics... So that code uses generics for a case that we don't need. When we write this code in C++, we have to use this kind of code with templates, to plug in something into something else to even be able to do that. We're just plugging some functionality into something else that it's not aware of, again, because interfaces -- well, C++ doesn't have interfaces, it has classes, but you can define functions as pure ritual so it doesn't have to have all the implementation. But whatever object you have, you cannot pass in as another unless it implements, it extends, or is essentially a class. But in this case, you define your own event handler. Obviously, nobody knows -- you can't really pass in or use in this code any class that you didn't write yourself. So you get stuck, and this is why you use generics. In many, many languages it's for this case, and that was why for many years people would say "Why do you need generics when you have interfaces?" But the truth is that you do need generics for the cases where the behavior of the class is derived by the type. For instance, next node in a linked list... The next is going to return the same type of node, and if your node is holding an integer, or a float, or whatever type of value, it's going to be different. A repository that stores any kind of model, a map... + +By the way, Go always had generics -- actually, that would be my unpopular opinion. Go always had generics, because map is a generic type. It maps from a key that is a generic type to a value that is a generic type... I see you're sipping, you're drinking, so I know that you disagree... + +**Ian Lopshire:** No, I a hundred percent agree. + +**Ronna Steinberg:** Oh, you do. Okay. \[laughs\] The slide says that we're always generic... We just couldn't define our own, so it was like good for the 80% of the use cases, but then we had our own use cases that we needed it, and we just didn't have anything. So you will talk now, and I will sip a drink. + +**Ian Lopshire:** \[laughs\] No, I definitely agree with you that the maps and the slices are generic, literally since day one. And like you said, it is that 80% use case, right? I don't know how to tie this back to OOP stuff now... + +**Ronna Steinberg:** Why I think it plays a role in this is because what people did in those cases is they would write a lot of functional code, procedural code etc. and did not use necessarily generic types. I think we've had a lot of workarounds. I think now we can eliminate them. I think the language is very mature to be object-oriented. + +**Ian Lopshire:** That makes sense to me. + +**Ronna Steinberg:** Natalie, \[unintelligible 00:25:21.26\] + +**Natalie Pistunovich:** No... One thing is on a Zoom call, another thing is on a podcast. It's a whole new level. I've found the quote from Rob Pike about Go being yes or no object-oriented. So it's a bit of a thread... So it starts with "Whenever someone from Java, or C++, or C\# (I'm looking at you, Ronna) comes to Go, they look for class, finds track and stops looking. But this misses two fundamental differences between Go and traditional object-oriented languages. The first one is that it's not only structs. Any concrete type can have methods, integers, booleans, slices, even funcs, but the more important idea is the separation of concepts. Data and behavior are two distinct concepts of Go, not conflated into a single notion of class. + +This is the insight (which goes all the way back to SmallTalk) on which the object-oriented type system, including the interface model, is built. Stopping at "struct == class" misses much of what makes Go work." So what are your thoughts? + +**Ronna Steinberg:** I agree... It sort of works very well with everything that I learned recently about this, honestly. It feels like Go is more naturally object-oriented than languages that I worked with before. I had this discussion with my boyfriend, because of the workshop and the amount of resistance that I was getting. + +**Natalie Pistunovich:** Does he think Go is object-oriented? + +**Ronna Steinberg:** Yes. So his answer was "Well, of course it is." It wasn't even a question in this mind. And then he told me "Well, you know how the JavaScript people say that a prototype is the best way to define object-oriented." And my take-away from this is that every night I go to sleep next to a man who quotes the JavaScript people... \[laughter\] That's it. I thought it was a pretty cool observation, to be honest... But still, concerning. I'll keep you posted. + +**Ian Lopshire:** The quote there, where it says it separates the idea of data and behavior - I think that's the important part of that area... But I'm having a hard time articulating exactly why in my head. + +**Ronna Steinberg:** Do you think that that is a good way to express object-oriented, abstracting away the data from the functionality? + +**Ian Lopshire:** Yeah, I think going to what you were saying, where Go is almost a more pure version of object-orientation, I think that's what leads to it; the idea that -- I don't know, I'd have to really sit down and think about it, but that's where my mind is going. + +**Ronna Steinberg:** \[27:57\] I think that I've found something that a person that wrote me about Go not being an object-oriented language might be related to... So I searched for Rob Pike and object-oriented and I've found Wikipedia. On Wikipedia he is criticizing object-oriented. So it's possible that that was what that quote was about. + +**Natalie Pistunovich:** Do you want to read this out and share this later in our show notes, so everybody can go back to this, too? + +**Ronna Steinberg:** Yeah. I think it's funny... \[laughs\] I mean, clearly... Yes, I will send it to you, and we can include that. So he criticized object-oriented for being incredibly heavy, I guess... And that sort of works pretty well with those quotes about objects being lightweight in Go. I think I agree. I think it feels very, very natural. + +By the way, something that I really love is we don't have enumerators, but we do have -- like, it's very easy for me when I look at pseudo-code with those red/black trees, and coloring notes in a graph, and stuff like that - this is the kind of stuff that I teach to people who don't have (let's say) the classical background in university. I teach that stuff to people because they need to pass interviews, unfortunately... \[laughs\] So this is where this is coming from. But what I'd really like to do - because I take pseudo-code with white, black, grey, whatever, red, colors. I don't define a new type by alias integers to do that... And I get very, very expressive. I basically write pseudocode in Go. And I don't think that you could do that with a non object-oriented language. I can pretty much express any pseudocode that I see, without actually giving much thought to what is written there. I just essentially copy and paste it and make the code work. I just make the types work, and the functions work, and if the function happens to be a method, that's what it's going to be. And if the function happens to be a procedural function, that's what it's going to be, and I just make it work, whatever it's necessary. I don't think I would have been able to do that if Go was not object-oriented. Also, you should not be testing people on red/black trees. \[laughter\] + +**Ian Lopshire:** I definitely could not code up a red/black tree without googling something right now. + +**Ronna Steinberg:** Yeah, obviously. No, I still do these exercises with people so they get used to the idea. And then if they can do this, they can do anything... Stuff like that, essentially. But yeah, there are way too many edge cases in a red/black tree. It's not okay. \[laughs\] \[unintelligible 00:30:56.00\] + +**Ian Lopshire:** Someone should write a generic library for it once, and then never touch it again. \[laughter\] + +**Ronna Steinberg:** That sounds very good. So that's a rant about people who give impossible tests to complete beginners in coding. + +**Natalie Pistunovich:** Or tell us how you do this right, when you're hiring for your team. + +**Ronna Steinberg:** So I don't know that I do this right, let's start with that... But I try to decide what I want, and then I don't test people on stuff that I don't want. What I will do - for instance, if I want to know how thorough somebody is going to be, then I will dive with them on something that they do know, and check how familiar they are with all the details. + +Let's start with the very basics - if somebody tells me that they have worked with MySQL, then I will test them heavily on MySQL until I reach something that either they don't know, or... To get a sense of how far into the rabbit hole they went. Or any other topic. This one is kind of normal. + +\[32:05\] But it all depends on their profile. It really all depends on their profile. We said earlier that I am the sum of the opportunities that were given to me... And I don't think I'm a bad developer. We all want to hire somebody who cannot do what we can't do. \[laughs\] Essentially, I feel like we are raising the bar constantly to a level that people cannot actually meet. And not because they are not good enough. + +For instance, if I can't do something and I am going to search for somebody who's going to be able to do it, that's a little bit unfair. And I feel like we are doing this all the time; we're sort of like raising the bar, raising the bar, raising the bar, to essentially maybe make up for our own disadvantages. + +So I try to not be unreasonable when I interview somebody. I do try to see if they can learn, there in the interviewing process, for instance. So I will present them with something that they will have to think about, internalize, and then spit out some information that could be wrong, could be right... It could be very small things that tell me if they get it. + +For instance, I work with billing people, and I'm not necessarily familiar with those concepts, so I will introduce something that has something to do with what we do in the domain... And it's fine that they don't know this. And then I will ask them "How would you do it, given those constraints? What would you do?" So I taught them something, and now I can see how they actually learned and what they actually do with this information. And it's not really just problem-solving, it's more of try not to introduce them to problems that they already know, essentially. It could be very, very small things. + +So yeah, I am the sum of the opportunities that were given to me, and I am a huge believer that if you give people an opportunity, they're not gonna let you down, especially if they've been begging for work since forever... And I do think that you can bank on people's loyalty in those cases. I think for a lot of people just having their foot in the door is such a big deal for them that they will \[unintelligible 00:34:23.10\] So I do see a lot of value in these people. + +Even now, if I go for a job interview, I am never going to apply for something that I already know. I want to grow just as much as everyone else, so I'm always going to go and try to convince somebody that they need to invest in my potential. + +I am very, very similar to a complete beginner, because we all have to go around and essentially convince people to give us a chance to prove that we can do something that we have not shown before that we can do. It's so obvious to people that "Okay, Ronna after 20 years is going to be able to do this." But it shouldn't be. I mean, at the end of the day, it really shouldn't be. I fail just as much as everyone else. So that's what I go with. It's a very long answer to your question, Natalie. + +**Break:** \[35:23\] to \[38:26\] + +**Natalie Pistunovich:** You mentioned briefly about your background, starting with C++, and also a little bit on how you've come to the idea of doing this workshop. Do you want to elaborate a little bit on giving the context of what brought you on the shorter-term and on the longer-term to this idea? + +**Ronna Steinberg:** I don't know, there was this jobs fair in April where I sat down with the organizer of GopherCon Europe, and I asked, point-blank, "Why have you never asked me to do a workshop? You know I do this..." \[laughs\] And then she told me "Well, it's a lot of work..." and I said "Well, yeah, but this is something that I do..." \[laughs\] But the topic itself is really something that I did want to teach my team. My team writes mostly procedural code with Go... + +**Natalie Pistunovich:** Why did you want to do it? How did you even come to think of it? + +**Ronna Steinberg:** Because that's what they do, and I actually don't think that those patterns that they are using serve my team very well. Actually, it's not really my team, it's my former team, because since about a month ago we reorganized... We used \[unintelligible 00:39:26.05\] in a way that was not really working well for us. I mean, I'm not saying that there is no way to make it work, but the way that we did this was not really working well for us... And I was trying to show different approaches to maybe redesign a portion of the code that I felt would just be easier to understand with classic object-oriented... Because that is how we used to see the world. + +\[40:01\] We understand delegation. For instance, if I tell you, Natalie, to breathe in, you don't have to think about operating your lungs. You just breathe in. And if I tell you to breathe deeply, then you will breathe deeply, and if I tell you to stop breathing, you will until you panic. \[laughter\] + +**Natalie Pistunovich:** Defer. Breathe out. + +**Ronna Steinberg:** \[laughs\] Yeah. You'll defer some breath, hopefully. Unless something is broken. \[laughter\] I do feel that when the code gets too complex and you don't understand it anymore, then you really need to start thinking in terms of objects. It makes things so much easier. People do understand those concepts very well, because they mimic the way that we think, the way that we work, the way the world works. + +So that's where it came from. It came from a real problem in our codebase. And then I kind of realized that if people don't agree that it's an object-oriented language, then I want them to see why I think it is, and at least give them the option. + +Also, I feel like we failed them in some way. If people work with Go and they don't see the benefits at all, and they don't think it even exists... It's a massive part of the language that they're not utilizing. + +Go is very simple, in the sense that -- for instance, if we compare it to my first language... It's not really my first language. Let's say my somewhat first language. My first language is Pascal, but if you point a gun to my head, I won't be able to -- + +**Natalie Pistunovich:** Same here! + +**Ronna Steinberg:** Yeah! We're the products of the same education system. \[laughs\] + +**Natalie Pistunovich:** Some people had -- was it Java...? + +**Ronna Steinberg:** \[unintelligible 00:41:48.11\] + +**Natalie Pistunovich:** Which is not really the uncle of Go, unlike Pascal, which is... + +**Ronna Steinberg:** Yeah, that's true. That's actually true. + +**Natalie Pistunovich:** Although I did start with Dr Scheme, to be very, very accurate. + +**Ronna Steinberg:** Well, the only reason that I know that is because of Carmen Andoh, who went around telling everybody that Go is like a child of the branches of -- + +**Natalie Pistunovich:** Pascal... + +**Ronna Steinberg:** Yeah... And then I learned also that Ruby is also somewhat a child of Pascal, which explains why I like Ruby... Although I don't remember Pascal at all, and I can't really tell if there are similarities or not. I have no idea. + +**Natalie Pistunovich:** Try to take it with a piece of paper. Maybe it will refresh your memory more than with a screen. That's how it was for me at least, writing Pascal code. Not sure I ever actually executed any. + +**Ian Lopshire:** Did you get to write it? + +**Ronna Steinberg:** I don't remember... + +**Natalie Pistunovich:** For better and worse... The education system... \[laughter\] It's the whiteboard of the early times. Anyway, yes, it's the uncle of Ruby. I did not know that. + +**Ronna Steinberg:** Yeah, things that you learn when you go to conferences. Random bits of information, of trivia that could be useful in the future or not. + +So back to C++, it has a million features that you're never gonna use. JavaScript as well. Nobody uses -- let's say what it is; it's not a common practice, in that case it's a best practice not to utilize all the features in those languages, because that is not maintainable code. But Go has been written in a way that should allow us to utilize all the features. So the idea that people don't do it is... Well, it's just sad. I don't know what to say else about it. + +**Natalie Pistunovich:** What are some good or bad use cases for using that? You mentioned billing is your close-to-home example... Do you have some other use cases you would say definitely use it, or definitely don't? + +**Ronna Steinberg:** So the quote that Rob Pike, that whomever sent me the message was basing it on - he said that (I don't know) a professor used multiple classes to perform something that was a simple look-up. And I think this is it. And I also understood why it happened... But again, it's not Go, actually. + +\[44:09\] So again, when we go back to Java - you are not able to say that A is B, unless A is aware of B, and knows that it implements B. So you can't really say that A is B. And then in those languages you really have to work extra-hard to express the idea "A is B." And that could create those intermediate layers between code. It's just a proxy; it just invokes more code, it invokes more code, it invokes more code. And then at the end of the day, if you want to perform a simple look-up, it can look like something that will create this very chaotic codebase, for something very simple. But then in Go -- I always say that everything is explicit in Go, except for the things that aren't... And what I mean by the things that aren't, are like stringers, for instance, where we invoke some functionality by performing some type assertion that nobody is aware of somewhere. + +But yeah, generally speaking, Go is explicit. If you have a package, and it's well-written, then it should meet the open/close principles, and then you should be able to wrap this type or extend the functionality of whatever it is that you want. I understand why people struggle with that. I think the biggest issue is how we don't write packages very well. + +By the way, for the workshop I had to revise my code multiple times, because I realized my design choices were sub-par, let's say. + +**Natalie Pistunovich:** Less common practices? \[laughter\] + +**Ronna Steinberg:** For instance, it's really funny... So I actually wanted to show that the code was extendible by actually creating an extra package that will use that package... And then I realized that \[unintelligible 00:46:00.14\] I mean, it should extend, but through an interface, not direct. And as I was doing that, the reason that I realized that was because I had a third package that actually did need that interface... And it needed that interface to be in the in-between layer. And then I started thinking "So how do I make -- like, if I wanted to teach somebody that, how would I actually do it?" And I'm still struggling with figuring out what is the exact problem with that code that I could tell somebody like "If you see this, then that is your problem, and that's what you need to change." But it's really funny, because I was actually writing something to show that it's going to be extendable, and then it wasn't. And then at some point I hit an end. + +So obviously, I fixed it, but generally, it's a different type of language. A package doesn't have to expose an interface. If you write code in Java, if you write code in C++, you don't want to express that A is B. But at the time that you write the package, or you write the class, or you write anything, you have to know how the user is going to use it. You have to know. And then you have to extend or implement or do something so that a user who is using your package is going to be able to plug in your A as B, your A as I, your A as whatever. So you have to think about how people are going to use your class. + +And then in Go you don't have to do it, but then you can write very easily an unusable package, or an unextendable package. So how to write it well, by the way, is something that I still don't know that we are very good at teaching. I'm trying to sort of figure that out as well, as I go. And how do I define that a package is good? + +\[48:00\] I love that when people explain the open-close, they always talk about curl command. Nobody is going to rewrite curl. It doesn't require redesigning. Well, why? Why is it so good? What makes it so good? And what makes a good package? It's a very, very difficult question. Regardless, by the way, if you have object-oriented or not; even if it's just a bunch of functions, it is very difficult to define when you're done, to know exactly when you're done. + +**Natalie Pistunovich:** One last question before we switch to the fun unpopular opinion, which we were missing so much throughout this entire episode... You've been kind of giving us sprinkles of information on the workshop. So other than this is being on object-oriented programming in Go, what else can you tell us about it? + +**Ronna Steinberg:** Well, I can tell you that I will take the learners through a maze of object-oriented... \[laughs\] And when I say a maze, I mean quite literally. We are going to navigate through a maze. I mentioned Jeff Rosenschein earlier. He was my professor to Intro to CS, and our first exercise way back when, in 2003 - that's what I'm taking people through. It's heavily inspired by his work, so credit to him. + +I thought it was a very good way of exploring object-oriented. And I like to take people through journeys that I've found very good for myself. If I had an a-ha moment, I try to share it with the other people as well. And Natalie, you know this -- Ian, I can actually ask you, what is your a-ha moment in your career? Situations when you said "Wow, now I get it. Now I know." + +**Ian Lopshire:** That's a real on-the-spot question. + +**Ronna Steinberg:** I know, it's a difficult one. When did you realize that you can do this, you can code, you know what you do, you've got this? + +**Ian Lopshire:** I mean, that moment was -- so I was in school, and I went to interview for an internship... And doing that interview. They did the whole whiteboard thing. It was eight hours of interviews. I believe that was like an a-ha moment. I feel good at the end of that, and that made me feel like "Oh yeah, I can do this. I can do this." But before that, I had no confidence. + +**Ronna Steinberg:** That's the thing, so for you it was trial by fire. + +**Ian Lopshire:** Exactly, yeah. + +**Ronna Steinberg:** That's very interesting. I think for a lot of people it is just like succeeding in something that they didn't think they could before. + +**Ian Lopshire:** I didn't think I was ready for that kind of set of interviews... But that's a reason why I tell people to just -- you see a lot of those posts where it's like "Oh, you know, I'm trying to do these things before I apply for jobs and interview." And my advice is always just like "Just go interview." If you do poorly, you've learned something, and you can go do the next one a little bit better. Don't wait, just go. + +**Ronna Steinberg:** I think Natalie the next time we need to bring somebody who doesn't agree with us... \[laughter\] Because we are all so on the same page, it's a problem. + +**Ian Lopshire:** A hundred percent. + +**Natalie Pistunovich:** Yeah, we'll pull some from your emails. A thought that came to mind throughout this episode... Let's see if I can make this into a sort of an unpopular opinion... Probably not, so I'm not gonna make the tune just yet. But thinking of AI-generated tools, they get inspired from existing code, right? And we can have this conversation if there's features of OOP in Go or not, but if there's not a lot of examples out there, the different models that generate code will not be creating this a lot. So assuming that the trend of code is not just a human writing it, but more like a human guiding it, it means that it will - at least the way I see it, it means that a lot of the code or the language will kind of fall deeper into their template, or into their niche, or little box, rather than spreading out of it, like you did. So this can be a fun thing to think about as a person who's researching AI and code. + +**Ronna Steinberg:** \[52:14\] Wow, I love it. You know, a few days ago I had that thought, AI is not going to write code OOP maybe, because they don't really understand the world like we do, and they don't have that restriction. A bot doesn't tell a person to breathe, you know what I mean? They don't need to. They can go as deeply into the mechanics of how to breathe, so their understanding is going to be very different, of those models. + +**Natalie Pistunovich:** Yeah. But maybe there'll be some way of doing this guided, and then we're saved from those boxes. + +**Ronna Steinberg:** Exactly. So that is very interesting what you have there, because you know, something that I said earlier about how do you know that a package is complete - maybe a bot can do that for you. + +**Natalie Pistunovich:** Static analysis. Or dynamic. + +**Ronna Steinberg:** Maybe a bot can analyze if something is open-closed. + +**Natalie Pistunovich:** Yeah, that definitely is a fun conversation to have... And if anybody wants to chat about this, we are on the Go Time Slack channel, in the Gophers Slack, so reach out and maybe this will be some future episode. But until then, let's do the tune for the Unpopular Opinion. + +**Jingle:** \[53:25\] to \[53:42\] + +**Natalie Pistunovich:** And now it's time for the unpopular opinion. Ronna, what do you have for us? + +**Ronna Steinberg:** I mean, I felt like the entire show was about -- I have an arsenal of things that I need to convince people about. So I feel that we're going to be much stronger if we collected opinions about Go from people who are not professional Gophers. And instead of teaching them, learn from them a little bit. I do see other languages evolve in many different directions. I think people understand today how to work with languages in a very different way than what we used to do. The evolution of best practices, all of those things - it's tremendous. + +Also, I think that Go added a lot of value to other languages just by existing. Just because Go introduced these features that we discussed. I think we influenced the industry, but I think that we should be open also to be influenced back. + +**Natalie Pistunovich:** So your unpopular opinion is that Go should be open to be influenced by non-gophers. + +**Ronna Steinberg:** Yes. + +**Natalie Pistunovich:** Alright. Let's see how that poll works. This will be a poll, and then let's see if that brings you further into the Hall of Fame of the unpopular opinions. + +**Ronna Steinberg:** Well, Mat said after he put me in the Hall of Fame - he then said that it doesn't exist everywhere. I think that the other people with unpopular opinions are not jealous but as far as I'm concerned, I want that title fair and square, so... \[laughs\] But hey, let's see. It might be as unorthodox as the old one. + +**Natalie Pistunovich:** Cool. Alright, well, that was fun and interesting, and I hope this will bring to the workshop that will generate enough code that it will train the AI to do some OOPy Go. Thanks Ronna, thanks Ian, thanks everyone who joined. + +**Ronna Steinberg:** Thank you...! diff --git "a/Might Go actually be OOP\357\274\237_transcript.txt" "b/Might Go actually be OOP\357\274\237_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..fb8a9495c54a2497edf9359a4d352695a7bc7ca2 --- /dev/null +++ "b/Might Go actually be OOP\357\274\237_transcript.txt" @@ -0,0 +1,3567 @@ +[0.00 --> 4.06] The answer is I don't know, but I am leaning towards yes. +[4.26 --> 6.66] It is an object-oriented programming language, +[6.98 --> 10.58] and it's because of multiple features that it has. +[11.00 --> 13.46] And to explain why I think it's an object-oriented language, +[13.58 --> 18.02] I should probably explain why I think it's definitely more object-oriented, +[18.40 --> 22.98] for instance, than other languages that are considered object-oriented. +[22.98 --> 24.98] And I'm looking at you, Java. +[30.00 --> 35.66] This episode is brought to you by our friends at Square. +[35.94 --> 37.98] Square is the platform that sellers trust. +[38.46 --> 42.72] There is a massive opportunity for developers to support Square sellers +[42.72 --> 45.30] by building apps for today's business needs. +[45.66 --> 48.48] And I'm here with Shannon Skipper, Head of Developer Relations at Square. +[48.88 --> 51.54] Shannon, can you share some details about the opportunity for developers +[51.54 --> 52.88] on the Square platform? +[52.88 --> 53.62] Yeah, absolutely. +[53.86 --> 56.52] So we have millions of sellers who have unique needs, +[56.84 --> 59.92] and Square has apps like our point-of-sale app, like our restaurants app, +[60.00 --> 63.72] but there are so many different sellers, tuxedo shops, florists, +[63.86 --> 66.28] who need specific solutions for their domain. +[66.54 --> 69.96] And so we have a Node SDK written in TypeScript +[69.96 --> 73.72] that allows you to access all of the backend APIs and SDKs +[73.72 --> 77.62] that we use to power the billions of transactions that we do annually. +[77.84 --> 82.20] And so there's this massive market of sellers who need help from developers. +[82.28 --> 86.94] They either need a bespoke solution built for themselves on their own Node stack, +[86.94 --> 90.80] where they are working with Square dashboard, working with Square hardware, +[90.80 --> 93.62] or with the e-com, you know, what you see is what you get builder. +[93.86 --> 94.90] And they need one more thing. +[94.96 --> 96.28] They need an additional build. +[96.54 --> 99.06] And then finally, we have the app marketplace where you can make a Node app +[99.06 --> 102.68] and then distribute it so it can get in front of millions of sellers +[102.68 --> 104.58] and be an option for them to adopt. +[104.58 --> 105.40] Very cool. +[105.48 --> 105.74] All right. +[105.76 --> 108.46] If you want to learn more, head to developer.squareup.com +[108.46 --> 113.52] to dive into the docs, APIs, SDKs, and to create your Square Developer account. +[113.82 --> 115.62] Start developing on the platform seller's trust. +[116.06 --> 118.32] Again, that's developer.squareup.com. +[118.32 --> 133.54] Let's do it. +[134.12 --> 135.16] It's go time. +[135.82 --> 137.14] Welcome to go time. +[137.60 --> 141.02] Your source for diverse discussions from all around the go community. +[141.34 --> 143.70] GopherCon Europe is right around the corner, +[143.90 --> 146.06] and you know we'll be there doing that go time thing. +[146.06 --> 149.36] It's just two weeks away, and tickets are still on sale. +[149.52 --> 152.34] Get yours now, and we'll see you in person or online. +[152.60 --> 154.84] Special thanks to our partners at Fastly. +[155.04 --> 156.94] Everything we ship here at Changelog is fast +[156.94 --> 160.92] because Fastly serves it up super fast everywhere on Earth. +[161.26 --> 163.10] Check them out at Fastly.com. +[163.38 --> 165.04] Okay, here we go. +[167.16 --> 170.00] Good evening, afternoon, morning, or night, +[170.30 --> 173.74] everyone who's joining us live or listening to this later. +[173.74 --> 178.52] This episode, Ian and me are being joined by Rona +[178.52 --> 183.08] to talk about object-oriented programming. +[184.04 --> 187.28] And we'll start by introducing Rona, I guess. +[187.80 --> 190.24] Rona, you are an engineering manager at Delivery Hero, +[190.94 --> 192.40] a Google developer expert for Go, +[192.88 --> 195.04] a woman who Go organizer in Berlin, +[195.86 --> 198.74] Go Times Unpopular Opinion Hall of Fame-er, +[198.74 --> 200.94] and after 20 years in tech, +[201.32 --> 204.04] you know that the sum of the opportunities +[204.04 --> 206.04] that were given to you, +[206.54 --> 209.18] this is why you're basically giving opportunity to others. +[209.54 --> 212.14] So this is an interesting thing about your bio. +[212.26 --> 213.86] We'll definitely ask you more about this. +[214.34 --> 216.20] Now we're here to talk about this workshop +[216.20 --> 220.38] that you've been crafting since 2017 +[220.38 --> 221.78] that you've been giving to your meetup +[221.78 --> 223.66] at Women Who Go Berlin, +[223.66 --> 226.02] the most recent one that you happened to give +[226.02 --> 228.94] at GopherCon Europe later this month +[228.94 --> 232.44] on the topic of actually object-oriented design in Go. +[232.82 --> 235.32] So as a preparation for this episode, +[235.56 --> 240.14] we were figuring out how each of us pronounces +[240.14 --> 243.94] the short term for object-oriented programming. +[244.66 --> 245.50] Rona, how do you say that? +[245.78 --> 247.08] The three letters acronym. +[247.56 --> 248.56] You mean OOP? +[248.90 --> 249.34] OOP? +[250.04 --> 252.26] Well, I guess it's two versus one, Ian. +[252.26 --> 253.40] You pronounce this the same. +[253.68 --> 254.62] Yeah, OOP, yeah. +[254.90 --> 255.86] I have to say, +[256.20 --> 258.36] for the past like a month or so, +[258.42 --> 259.60] I've been writing OO, +[260.16 --> 260.96] like a lot, +[261.64 --> 263.30] and it looks like OO, +[263.52 --> 265.14] like the emoji, +[265.60 --> 267.82] and it feels very symbolic +[267.82 --> 270.84] to the state of object-oriented. +[271.16 --> 273.08] I somehow always thought it's OOP. +[273.30 --> 274.02] You thought it was OOP? +[274.26 --> 276.96] I guess it's because last time I used this term +[276.96 --> 277.82] was in university, +[278.12 --> 279.24] and it's been a while since. +[279.38 --> 280.44] And you used to say OOP? +[280.92 --> 281.96] Yeah, we used to say OOP. +[281.96 --> 283.10] Ian, what do you say? +[283.50 --> 284.56] I'd definitely say OOP. +[284.78 --> 287.42] I don't think I've ever heard someone call it OOP before, +[287.56 --> 287.86] actually, +[287.94 --> 289.68] until 10 minutes ago. +[290.00 --> 292.90] This is definitely a debate worth telling. +[293.34 --> 294.96] Yeah, that can be a poll. +[295.04 --> 298.00] Maybe this can even be my unpopular opinion in there. +[298.16 --> 298.92] I thought I have none. +[299.40 --> 301.36] So we can say this episode is sort of +[301.36 --> 303.12] one big unpopular opinion +[303.12 --> 305.08] coming with a claim, Rona, +[305.34 --> 309.16] that Go can be an object-oriented programming language. +[309.16 --> 310.38] Well, I mean, +[310.62 --> 312.38] I definitely don't think it's not. +[313.38 --> 315.62] So what happened was that +[315.62 --> 317.30] I wanted to share with my team +[317.30 --> 319.76] how to do object-oriented programming with Go. +[320.20 --> 322.38] And then I kind of realized +[322.38 --> 325.44] by looking at actual interview questions, +[325.68 --> 326.06] answers, +[326.36 --> 327.54] how people like feedback, +[328.20 --> 330.50] how things are happening in my organization, +[330.50 --> 332.32] what people think Go is +[332.32 --> 333.54] and what people think Go isn't, +[333.64 --> 335.32] they definitely think that Go is not +[335.32 --> 338.20] an object-oriented programming language. +[338.34 --> 338.96] So I decided, +[339.14 --> 339.50] okay, +[339.70 --> 341.68] I'm going to give a workshop about this, +[341.76 --> 342.14] sort of like, +[342.40 --> 343.94] try to sort the mess. +[344.02 --> 344.12] Like, +[344.16 --> 345.16] if you want to do it, +[345.20 --> 345.34] like, +[345.40 --> 346.24] how you should do it. +[346.62 --> 348.38] And this has always been, +[348.42 --> 348.60] like, +[348.62 --> 349.52] how I do things. +[349.72 --> 350.42] I try, +[350.52 --> 351.22] when I teach Go, +[351.28 --> 352.58] I try to teach people, +[352.88 --> 353.04] like, +[353.06 --> 354.10] if they want to do something, +[354.16 --> 355.46] I try to teach them how to do it, +[355.50 --> 357.24] not whether they're right or wrong +[357.24 --> 360.40] to think or do things in a certain way. +[360.40 --> 363.90] So it's more about how to use the tools +[363.90 --> 366.28] that you have with the language properly. +[366.44 --> 367.86] And especially now with generics +[367.86 --> 368.98] going into the language, +[369.50 --> 373.10] solving a kind of like a last kind of edge case +[373.10 --> 374.42] that we didn't have. +[375.14 --> 375.62] So, +[375.96 --> 376.26] yeah, +[376.60 --> 377.24] I posted, +[377.42 --> 379.42] I was going to do that internally +[379.42 --> 380.60] in the organization. +[380.78 --> 381.22] And immediately, +[381.60 --> 383.26] a lot of people that I've never met +[383.26 --> 385.10] reached out to, +[385.16 --> 387.00] to give me honest feedback +[387.00 --> 388.50] about the workshop +[388.50 --> 390.38] that I have never given yet. +[390.40 --> 391.20] And then, +[391.36 --> 392.84] then I kind of knew that, +[392.96 --> 393.12] okay, +[393.16 --> 393.32] like, +[393.38 --> 394.18] I have a topic +[394.18 --> 396.22] that is definitely worth exploring. +[397.06 --> 397.32] So, +[397.84 --> 398.08] yeah, +[398.38 --> 400.52] I guess it is my unpopular opinion. +[401.18 --> 401.28] So, +[401.60 --> 403.26] tell us about some of that feedback +[403.26 --> 404.36] or opinion. +[405.70 --> 406.18] So, +[406.68 --> 407.78] it's quite amazing. +[408.08 --> 408.22] So, +[408.28 --> 409.62] everybody has an opinion. +[410.38 --> 411.76] Some people were with me +[411.76 --> 412.60] and wanted to share +[412.60 --> 413.14] that they are, +[413.34 --> 413.52] you know, +[413.56 --> 414.26] that they agree +[414.26 --> 416.30] that Go can be object-oriented, +[416.62 --> 417.68] which was really nice. +[417.82 --> 419.04] What was really nice about it +[419.04 --> 420.32] was that everybody can find +[420.32 --> 421.82] a resource to support their claim, +[422.32 --> 424.54] which I found amazing. +[425.58 --> 426.70] It's brilliant. +[426.94 --> 427.10] Like, +[427.40 --> 428.52] I can't even begin. +[428.80 --> 429.00] Now, +[429.06 --> 429.86] I'm not an academic, +[430.08 --> 430.40] obviously. +[430.80 --> 431.16] And, +[431.28 --> 432.22] what I do, +[432.22 --> 432.66] I do, +[432.74 --> 434.16] I do by trial and error. +[434.62 --> 434.82] But, +[434.92 --> 436.06] I did read a lot about it +[436.06 --> 437.26] for the past few months +[437.26 --> 439.36] since we kind of decided +[439.36 --> 440.62] that I was going to do this. +[440.62 --> 441.12] and, +[441.12 --> 442.26] the answer is, +[442.30 --> 442.96] I don't know. +[443.66 --> 444.02] But, +[444.44 --> 446.12] I am leaning towards yes. +[446.70 --> 448.00] It is an object-oriented +[448.00 --> 449.36] programming language. +[449.68 --> 450.22] And, +[450.22 --> 451.44] it's because of +[451.44 --> 452.44] multiple features +[452.44 --> 453.28] that it has. +[454.18 --> 454.28] And, +[454.38 --> 455.12] to explain why +[455.12 --> 456.20] I think it's object-oriented +[456.20 --> 456.70] languages, +[457.00 --> 458.22] object-oriented language, +[458.34 --> 459.80] I should probably explain +[459.80 --> 461.24] why I think it's +[461.24 --> 463.20] definitely more object-oriented, +[463.58 --> 464.18] for instance, +[464.32 --> 465.84] than other languages +[465.84 --> 466.70] that are considered +[466.70 --> 468.14] object-oriented. +[468.14 --> 468.86] And, +[468.98 --> 469.74] I'm looking at you, +[469.86 --> 470.16] Java. +[470.88 --> 471.12] And, +[471.18 --> 471.96] here is my case. +[472.02 --> 472.50] Listen up. +[473.42 --> 473.90] Java, +[474.48 --> 474.72] yes, +[474.82 --> 475.00] sure, +[475.06 --> 475.86] it has inheritance. +[476.52 --> 476.90] But, +[477.34 --> 479.00] you can only inherit once. +[479.48 --> 479.68] So, +[479.76 --> 480.52] that means that +[480.52 --> 481.96] if class A, +[482.08 --> 483.08] you want to express +[483.08 --> 483.98] the idea that +[483.98 --> 484.70] class A, +[484.86 --> 486.18] or an object of type A, +[486.26 --> 488.42] is also an object of type B, +[488.54 --> 489.38] or class B, +[489.98 --> 491.34] you can only do that once. +[491.40 --> 491.48] And, +[491.56 --> 493.06] if you have a case C, +[493.82 --> 495.32] that A is also C, +[495.78 --> 497.48] you cannot express that in Java. +[497.48 --> 498.16] Now, +[498.24 --> 499.44] I came from C++, +[499.76 --> 500.54] so I kind of, +[500.58 --> 500.90] you know, +[500.98 --> 502.30] had this multiple inheritance. +[502.74 --> 504.22] I used to be very expressive +[504.22 --> 505.10] with inheritance. +[505.64 --> 505.72] And, +[505.82 --> 506.60] when I look at Java, +[506.78 --> 507.62] when I look at Ruby, +[508.28 --> 508.74] Pythonistas, +[508.92 --> 509.28] I'm sorry, +[509.42 --> 510.96] I'm not familiar enough +[510.96 --> 511.52] with your language +[511.52 --> 512.00] to know, +[512.26 --> 513.00] to make a case. +[513.60 --> 513.82] But, +[513.92 --> 514.70] when I look at Java, +[514.80 --> 515.58] when I look at Ruby, +[516.14 --> 517.50] I see that +[517.50 --> 519.06] you can't really express +[519.06 --> 520.52] A is C, +[520.62 --> 521.66] if you already express +[521.66 --> 522.62] that A is B, +[523.12 --> 524.60] unless you use interfaces +[524.60 --> 525.54] or generics. +[525.54 --> 526.58] Now, +[526.88 --> 529.02] interfaces are, +[529.46 --> 531.58] the Go has this dramatic effect +[531.58 --> 532.96] that interfaces are implicit. +[533.74 --> 534.78] In Java, +[535.52 --> 537.26] a class has to be aware, +[537.54 --> 538.50] to be able to instantiate +[539.18 --> 541.52] A as type interface I, +[542.36 --> 543.92] A has to be aware of I +[543.92 --> 545.02] at the time of writing. +[545.58 --> 546.06] So, +[546.16 --> 548.46] that means that to plug A as I, +[548.68 --> 549.96] you have to do something. +[549.96 --> 551.86] It could be generics, +[552.24 --> 553.26] it could be +[553.26 --> 555.70] some sort of a proxy class +[555.70 --> 556.96] that we'll implement +[556.96 --> 558.30] and derive from. +[559.08 --> 559.34] So, +[559.40 --> 560.20] there are some ways. +[560.76 --> 560.88] And, +[560.96 --> 561.32] in Ruby, +[561.42 --> 562.44] we also see +[562.44 --> 563.64] how modules +[563.64 --> 565.04] are kind of replacing +[565.04 --> 565.94] generics +[565.94 --> 567.50] and what people used to do +[567.50 --> 568.72] in other languages +[568.72 --> 570.18] to sort of be able +[570.18 --> 571.40] to be very expressive +[571.40 --> 572.24] that something +[572.24 --> 573.56] is also something else. +[574.38 --> 574.70] So, +[575.14 --> 576.32] the answer is that +[576.32 --> 577.16] I think that Go +[577.16 --> 579.74] is object-oriented. +[580.46 --> 581.96] I'm leaning towards yes. +[582.30 --> 583.38] I do see why people +[583.38 --> 584.28] don't think so. +[585.18 --> 585.66] But, +[585.74 --> 585.88] yeah, +[585.96 --> 586.86] that's where I am +[586.86 --> 588.08] as a person +[588.08 --> 590.28] who works with Go. +[590.54 --> 591.06] Can we go through +[591.06 --> 591.68] some of the reasons +[591.68 --> 592.80] people said it wasn't? +[592.90 --> 593.04] Like, +[593.42 --> 594.48] because it lacks inheritance? +[594.68 --> 594.80] Like, +[594.84 --> 595.86] what features are missing +[595.86 --> 596.72] that people are like, +[596.76 --> 597.88] it's not object-oriented? +[598.20 --> 599.44] I think it's about +[599.44 --> 600.12] the styling. +[600.12 --> 600.52] So, +[600.62 --> 600.88] yes, +[601.02 --> 602.50] most people talk about +[602.50 --> 603.48] the lack of hierarchy, +[604.24 --> 606.46] the composition over, +[607.08 --> 607.42] what's the word +[607.42 --> 607.98] I'm looking for? +[608.54 --> 608.98] Inheritance? +[609.52 --> 610.10] I guess. +[610.36 --> 610.70] I mean, +[610.86 --> 611.82] I think there is a, +[611.84 --> 612.20] like, +[612.28 --> 613.50] general term for inheritance, +[613.64 --> 613.92] but yeah. +[614.40 --> 614.66] Okay. +[614.82 --> 616.10] I think it's about, +[617.06 --> 618.12] I think people do a lot +[618.12 --> 619.60] of procedural coding with Go. +[619.90 --> 621.06] You don't have constructors, +[621.16 --> 621.64] for instance, +[621.72 --> 621.96] right? +[622.44 --> 624.12] And why do you not have constructors? +[624.16 --> 625.54] Because anything can be a type. +[625.88 --> 627.32] An integer can be a new type. +[627.40 --> 628.64] You can define a new type +[628.64 --> 629.34] with pretty much +[629.34 --> 630.74] in any way that you want. +[631.30 --> 631.50] So, +[631.56 --> 632.64] you don't really have constructors +[632.64 --> 633.66] because you don't have classes +[633.66 --> 635.22] and anything can be a type. +[635.58 --> 636.96] And then everything can have +[636.96 --> 637.78] also methods. +[638.54 --> 638.98] But, +[639.06 --> 640.10] the truth is that +[640.10 --> 641.62] without constructors, +[642.00 --> 642.64] without a sort of +[642.64 --> 643.70] a formal way +[643.70 --> 645.08] of working, +[645.20 --> 645.44] I think, +[645.50 --> 646.24] with objects, +[646.94 --> 647.82] people get lost. +[648.52 --> 649.02] And then, +[649.10 --> 649.42] you know, +[649.56 --> 650.62] they look down at people +[650.62 --> 652.02] who come in from other languages +[652.02 --> 654.98] who need these things. +[655.58 --> 656.54] It is very difficult +[656.54 --> 658.22] to define a type +[658.22 --> 659.62] and allow anybody +[659.62 --> 660.60] to make any changes. +[660.60 --> 660.80] So, +[660.88 --> 662.02] you cannot really tell. +[662.50 --> 663.22] How do you tell +[663.22 --> 664.44] if it's corrupt or not? +[664.72 --> 665.28] How do you write +[665.28 --> 666.80] any kind of defensive code +[666.80 --> 667.70] in that situation? +[668.34 --> 669.54] We also have these, +[669.58 --> 669.78] you know, +[669.88 --> 670.38] best practices. +[670.50 --> 670.66] I mean, +[670.98 --> 671.92] I don't like the word +[671.92 --> 672.66] best practices +[672.66 --> 674.18] because I think that +[674.18 --> 675.74] I usually use the word +[675.74 --> 676.62] common practice +[676.62 --> 677.46] to explain +[677.46 --> 679.38] because it's not always best, +[679.46 --> 679.62] right? +[679.66 --> 680.20] There's a case +[680.20 --> 681.24] for pretty much anything. +[681.74 --> 683.00] But the common practices +[683.00 --> 684.40] are everything is public. +[684.64 --> 685.74] Anyone can do anything +[685.74 --> 686.38] with the guy +[686.38 --> 687.16] and if they don't read +[687.16 --> 687.88] the documentation, +[688.06 --> 688.36] tough. +[688.96 --> 689.40] So, +[690.20 --> 691.32] it is a bit tricky +[691.32 --> 692.24] for people coming +[692.24 --> 692.96] into the language +[692.96 --> 693.96] to know exactly +[693.96 --> 694.66] what they're supposed +[694.66 --> 695.46] to do and how. +[696.18 --> 696.84] And in that sense, +[696.88 --> 697.54] we're not making +[697.54 --> 698.34] their lives easier. +[698.76 --> 699.78] We're just making it harder +[699.78 --> 700.54] instead of doing +[700.54 --> 701.20] this gatekeeping. +[701.30 --> 701.66] Oh, no, +[701.70 --> 702.52] you come from Java. +[702.66 --> 703.46] You probably don't know +[703.46 --> 703.92] how to go. +[704.66 --> 705.38] Never mind. +[705.66 --> 706.64] Forget about it. +[707.08 --> 708.52] This will take some time +[708.52 --> 709.76] and it doesn't matter. +[709.92 --> 710.96] They could have 20 years +[710.96 --> 711.76] of experience +[711.76 --> 713.92] and still look at them +[713.92 --> 714.14] like, +[714.54 --> 715.36] but do you know Go? +[715.90 --> 716.32] Do you? +[717.00 --> 717.44] So, +[717.60 --> 718.48] that's where I am. +[718.62 --> 719.68] I am curious though +[719.68 --> 720.64] if we're getting any +[720.64 --> 721.84] remarks from our listeners. +[722.32 --> 722.54] So, +[722.88 --> 723.58] maybe they have +[723.58 --> 724.40] some opinions. +[724.68 --> 725.58] Not just yet. +[725.68 --> 726.32] Not just yet. +[726.36 --> 727.70] But maybe any minute now. +[727.90 --> 728.28] Hopefully. +[728.74 --> 729.74] One can only hope. +[730.78 --> 731.64] I'll be curious to hear +[731.64 --> 732.94] what are some of the +[732.94 --> 734.60] feedbacks that you got +[734.60 --> 735.58] over the email. +[735.76 --> 735.80] Like, +[735.84 --> 736.70] what are the +[736.70 --> 738.38] pros +[738.38 --> 739.20] or the people +[739.20 --> 740.26] who agreed or disagreed? +[740.36 --> 741.18] You said that they were +[741.18 --> 742.60] all backing their claims +[742.60 --> 743.34] which is wonderful. +[743.52 --> 744.42] It means you have +[744.42 --> 745.34] generally a great +[745.34 --> 746.44] discussion. +[746.70 --> 746.98] Well, +[747.10 --> 747.42] I mean, +[747.60 --> 748.46] so there is +[748.46 --> 749.50] expected practices. +[749.64 --> 750.42] What was the term you used? +[750.58 --> 751.16] Common practices. +[751.86 --> 752.80] And that's great. +[752.92 --> 753.02] So, +[753.26 --> 753.78] tell us about them. +[754.08 --> 754.22] Well, +[754.26 --> 755.04] the first comment +[755.04 --> 755.96] that I got was, +[756.42 --> 757.50] that was really funny, +[758.00 --> 758.22] was, +[758.38 --> 759.28] aren't generics enough? +[759.54 --> 760.34] OOP now? +[760.52 --> 760.70] Really? +[761.36 --> 762.28] Something like that. +[763.32 --> 764.64] I thought that was hilarious. +[765.40 --> 765.68] So, +[765.90 --> 767.70] people do refer me a lot. +[768.38 --> 770.04] to the Go FAQs +[770.04 --> 771.38] that specifically say +[771.38 --> 772.12] that the answer +[772.12 --> 773.12] to whether Go +[773.12 --> 774.30] is object-oriented +[774.30 --> 774.90] or not +[774.90 --> 776.66] is yes and no. +[778.02 --> 778.92] And then, +[779.28 --> 779.62] no, +[779.70 --> 779.98] actually, +[780.14 --> 781.28] is it a yes and no? +[781.52 --> 781.76] Or, +[781.92 --> 782.48] I don't remember. +[782.82 --> 784.08] I actually prepared it for you. +[784.16 --> 785.04] Let me read this out. +[785.18 --> 785.72] Oh, amazing. +[786.28 --> 786.94] Yes and no. +[787.24 --> 788.36] Although Go has +[788.36 --> 789.24] types and methods +[789.24 --> 790.82] and allows an object-oriented +[790.82 --> 791.78] style of programming, +[792.06 --> 793.34] there is no type hierarchy. +[793.92 --> 795.72] The concept of interface +[795.72 --> 796.58] in Go provides +[796.58 --> 797.62] a different approach +[797.62 --> 798.40] that we believe +[798.40 --> 799.42] is easy to use +[799.42 --> 800.12] and in some ways +[800.12 --> 800.72] more general. +[801.48 --> 802.68] There are also ways +[802.68 --> 803.68] to embed types +[803.68 --> 804.48] in other types +[804.48 --> 805.38] to provide something +[805.38 --> 806.14] analogous +[806.14 --> 807.08] but not identical +[807.08 --> 808.56] to subclassic. +[808.98 --> 809.54] Moreover, +[810.06 --> 810.76] methods in Go +[810.76 --> 812.04] are more general +[812.04 --> 812.72] than C++ +[812.72 --> 813.50] or in Java. +[814.16 --> 815.14] They can be defined +[815.14 --> 816.94] for any sort of data, +[817.18 --> 818.04] even built-in types +[818.04 --> 818.70] such as plain +[818.70 --> 819.62] and boxed integers, +[819.62 --> 820.58] and they're not +[820.58 --> 821.26] restricted to +[821.26 --> 822.40] classes +[822.40 --> 823.52] slash structs. +[824.36 --> 824.72] Also, +[825.10 --> 826.50] the lack of type hierarchy +[826.50 --> 827.40] makes objects +[827.40 --> 828.08] in Go feel +[828.08 --> 828.90] much more lightweight +[828.90 --> 830.12] than in languages +[830.12 --> 831.14] like C++ +[831.14 --> 831.78] or Java. +[832.16 --> 832.40] Well, +[832.44 --> 833.30] I kind of agree. +[834.44 --> 835.18] I mean, +[835.64 --> 836.18] but then, +[836.26 --> 836.48] you know, +[836.56 --> 837.96] I went into the rabbit hole +[837.96 --> 839.02] because what I remembered, +[839.28 --> 840.12] what I remembered, +[840.28 --> 840.54] you know, +[840.60 --> 842.12] from my university +[842.12 --> 843.86] was my professor, +[844.12 --> 845.12] Jeff Rosenstein, +[845.70 --> 846.36] explaining +[846.36 --> 848.16] that object-oriented +[848.16 --> 849.18] is about being able +[849.18 --> 849.94] to send messages +[849.94 --> 850.76] to objects. +[851.28 --> 852.40] And I checked, +[852.58 --> 853.26] played the origin +[853.26 --> 854.14] of the idea +[854.14 --> 855.50] and it appears +[855.50 --> 856.56] to be the case +[856.56 --> 858.18] that that is what +[858.18 --> 861.24] they were going for. +[861.58 --> 862.42] It wasn't so much +[862.42 --> 863.18] about hierarchy. +[863.42 --> 864.60] Hierarchy came in later. +[865.46 --> 865.62] So, +[866.34 --> 867.14] if we go back +[867.14 --> 867.74] to basics, +[867.94 --> 868.40] maybe Go +[868.40 --> 870.20] is fully object-oriented. +[871.08 --> 871.54] And I think +[871.54 --> 872.34] to answer that, +[872.86 --> 873.78] you probably need +[873.78 --> 874.28] to know more +[874.28 --> 874.96] than I do +[874.96 --> 876.80] about language design. +[877.40 --> 878.14] But again, +[878.14 --> 879.46] from practice, +[879.72 --> 881.02] I'm leaning towards yes. +[881.72 --> 882.64] And in the workshop, +[882.96 --> 884.54] we do things +[884.54 --> 885.62] very Java-like. +[886.42 --> 887.42] Exactly the kinds +[887.42 --> 887.82] of things +[887.82 --> 888.38] that you're not +[888.38 --> 889.18] supposed to do +[889.18 --> 889.98] or it's not really +[889.98 --> 890.44] that you're not +[890.44 --> 891.04] supposed to do. +[891.16 --> 892.20] But we really +[892.20 --> 893.18] try to express +[893.18 --> 894.50] the same types +[894.50 --> 895.16] of ideas +[895.16 --> 896.54] that keeping +[896.54 --> 898.82] loyal to, +[898.96 --> 899.32] I guess, +[899.38 --> 900.02] the go-way +[900.02 --> 901.02] of doing things still. +[901.94 --> 902.14] So, +[902.78 --> 902.98] yeah, +[903.10 --> 903.48] I mean, +[903.56 --> 904.18] at the very least, +[904.24 --> 904.64] it's going to be +[904.64 --> 905.16] a lot of fun. +[905.16 --> 907.30] I try to at least +[907.30 --> 908.24] add some value +[908.24 --> 909.42] in that regard. +[909.64 --> 910.42] I give people +[910.42 --> 911.22] a good taste +[911.22 --> 913.18] of how they can +[913.18 --> 913.72] do things +[913.72 --> 914.50] and how they can +[914.50 --> 915.26] get creative. +[915.78 --> 916.58] But I also think +[916.58 --> 917.48] that there is +[917.48 --> 918.40] a lot of room +[918.40 --> 919.00] for people +[919.00 --> 920.20] to decide +[920.20 --> 921.38] within their teams +[921.38 --> 922.84] their own +[922.84 --> 923.82] best practices, +[924.00 --> 924.80] their own practices, +[925.06 --> 926.60] how they want +[926.60 --> 927.32] to do things. +[927.84 --> 928.68] I do think +[928.68 --> 929.34] there should be +[929.34 --> 930.78] a discussion. +[930.98 --> 931.68] I also think +[931.68 --> 932.30] that it should be +[932.30 --> 932.88] a little bit +[932.88 --> 934.42] more lively +[934.42 --> 935.66] than it is now +[935.66 --> 936.64] where I think +[936.64 --> 937.20] the veterans +[937.20 --> 938.22] sort of dictate +[938.22 --> 939.02] how people +[939.02 --> 939.46] are going to +[939.46 --> 940.00] use the language. +[940.42 --> 940.92] The users +[940.92 --> 941.72] of the language +[941.72 --> 942.44] that are coming +[942.44 --> 943.78] in are in the future. +[943.90 --> 944.34] They are going +[944.34 --> 945.60] to also decide +[945.60 --> 946.68] for themselves +[946.68 --> 947.40] how they want +[947.40 --> 947.94] to do things, +[948.00 --> 948.40] how they want +[948.40 --> 948.92] to program. +[949.08 --> 949.52] They should have +[949.52 --> 950.56] a space to do that. +[951.32 --> 951.74] And if they're +[951.74 --> 952.22] coming from +[952.22 --> 952.74] other languages +[952.74 --> 953.72] and they work +[953.72 --> 954.82] in a similar way, +[954.88 --> 955.34] that's fine. +[955.74 --> 956.52] You asked me +[956.52 --> 957.22] about quotes +[957.22 --> 957.84] that I got. +[958.14 --> 959.12] Somebody quoted +[959.12 --> 960.10] Rob Pike to me. +[960.10 --> 961.44] I tried. +[961.56 --> 962.10] I couldn't find +[962.10 --> 962.48] a quote. +[962.86 --> 963.28] But somebody +[963.28 --> 964.14] quoted Rob Pike +[964.14 --> 964.46] to me. +[964.54 --> 965.02] So apparently +[965.02 --> 965.74] Rob Pike +[965.74 --> 966.86] did say something +[966.86 --> 968.16] about object-oriented +[968.16 --> 970.32] look like this, +[970.68 --> 971.40] I guess, +[971.78 --> 972.40] object-oriented +[972.40 --> 973.60] or something +[973.60 --> 974.28] of the sort. +[974.80 --> 975.22] But yeah, +[975.28 --> 976.00] I didn't get +[976.00 --> 976.72] the source. +[977.44 --> 978.36] What I do see +[978.36 --> 979.42] is a lot of people, +[979.62 --> 979.78] you know, +[979.82 --> 980.24] sort of, +[980.78 --> 981.26] this is where +[981.26 --> 981.86] I think this is +[981.86 --> 982.36] coming from. +[982.96 --> 983.76] How people write +[983.76 --> 984.26] APIs, +[984.60 --> 985.50] how people write +[985.50 --> 986.52] their handlers, +[986.88 --> 987.86] it's all pretty much +[987.86 --> 988.80] functional programming. +[988.80 --> 989.72] The way that this +[989.72 --> 990.26] is designed, +[990.78 --> 991.36] most people +[991.36 --> 992.22] don't actually +[992.22 --> 993.26] use an interface +[993.26 --> 994.48] in these situations +[994.48 --> 995.62] and even if they do, +[996.16 --> 996.70] they kind of +[996.70 --> 997.88] don't use object, +[998.00 --> 998.82] they use the functions +[998.82 --> 999.90] as an interface, +[1000.36 --> 1001.46] which you can do, +[1001.62 --> 1002.32] you can infer, +[1002.88 --> 1004.02] so they're not used +[1004.02 --> 1006.16] to actually working +[1006.16 --> 1007.64] with objects +[1007.64 --> 1009.46] or not objects, +[1009.62 --> 1010.18] objects is not, +[1010.36 --> 1010.58] like, +[1010.66 --> 1011.52] let's say struts. +[1011.68 --> 1012.76] Let's just use +[1012.76 --> 1013.52] the word struts +[1013.52 --> 1014.22] with fields +[1014.22 --> 1014.90] and actual, +[1015.12 --> 1015.40] you know, +[1015.70 --> 1016.90] the classic sort of way +[1016.90 --> 1017.80] of doing things. +[1017.80 --> 1019.38] And I think +[1019.38 --> 1019.76] this is, +[1019.90 --> 1020.50] I think that's +[1020.50 --> 1021.30] where it's coming from +[1021.30 --> 1022.00] because we came +[1022.00 --> 1022.70] with a language +[1022.70 --> 1023.70] that was so strong +[1023.70 --> 1024.58] in that space +[1024.58 --> 1025.72] and in that space +[1025.72 --> 1026.84] you really don't need, +[1026.98 --> 1027.44] you know, +[1027.48 --> 1028.60] a lot of running objects +[1028.60 --> 1029.40] and whatever objects +[1029.40 --> 1029.92] you have +[1029.92 --> 1031.16] are probably going +[1031.16 --> 1032.46] to die soon anyways. +[1033.04 --> 1033.64] And if we think +[1033.64 --> 1035.02] also like those APIs, +[1035.74 --> 1035.98] you know, +[1036.02 --> 1036.68] it sounds funny, +[1036.78 --> 1037.22] but I mean, +[1037.26 --> 1038.40] objects don't live long +[1038.40 --> 1038.96] in APIs +[1038.96 --> 1040.78] unless they live +[1040.78 --> 1041.96] the entire lifetime +[1041.96 --> 1042.92] of the application. +[1042.92 --> 1045.06] and like the server, +[1045.24 --> 1045.62] for instance, +[1045.72 --> 1046.48] that could be +[1046.48 --> 1048.16] an object. +[1048.56 --> 1049.22] Their struts +[1049.22 --> 1050.32] are limited to fields +[1050.32 --> 1051.38] like configurations +[1051.38 --> 1052.76] and they don't have +[1052.76 --> 1053.44] many methods, +[1053.52 --> 1053.72] right? +[1054.22 --> 1055.12] But then if you look, +[1055.24 --> 1055.54] for instance, +[1055.66 --> 1056.68] at how they set up +[1056.68 --> 1057.66] their repositories +[1057.66 --> 1059.10] or how they set up +[1059.10 --> 1059.76] their models, +[1059.92 --> 1060.46] how they set up +[1060.46 --> 1061.04] these things, +[1061.44 --> 1062.80] because we didn't +[1062.80 --> 1063.60] have generics +[1063.60 --> 1065.08] up until very recently, +[1065.74 --> 1066.34] they don't really +[1066.34 --> 1067.44] use objects for that +[1067.44 --> 1068.16] or they generate +[1068.16 --> 1069.12] a lot of code, +[1069.12 --> 1069.74] but they don't +[1069.74 --> 1070.84] actually know +[1070.84 --> 1071.56] what's in it +[1071.56 --> 1072.70] or care so much +[1072.70 --> 1073.30] about it +[1073.30 --> 1075.04] just to save, +[1075.22 --> 1075.40] store, +[1075.52 --> 1075.80] whatever. +[1075.96 --> 1076.66] But with generics, +[1076.70 --> 1077.18] you can actually +[1077.18 --> 1078.64] store any kind +[1078.64 --> 1079.26] of model +[1079.26 --> 1080.62] into any kind +[1080.62 --> 1081.90] of database +[1081.90 --> 1083.44] or repository. +[1084.18 --> 1085.26] So we are +[1085.26 --> 1086.18] that much stronger +[1086.18 --> 1088.02] in that sense. +[1088.20 --> 1088.40] Now, +[1088.72 --> 1090.42] maybe that's going +[1090.42 --> 1090.58] to, +[1090.78 --> 1091.92] maybe now the people +[1091.92 --> 1092.64] are actually going +[1092.64 --> 1093.42] to write code +[1093.42 --> 1094.10] for themselves. +[1094.10 --> 1094.64] They're actually +[1094.64 --> 1096.52] going to spend +[1096.52 --> 1097.26] more time thinking +[1097.26 --> 1098.10] about what they +[1098.10 --> 1099.10] want to express. +[1099.12 --> 1099.68] and how, +[1100.18 --> 1100.56] hopefully. +[1113.16 --> 1114.30] This episode +[1114.30 --> 1115.38] is brought to you +[1115.38 --> 1116.06] by Honeycomb. +[1116.20 --> 1116.96] Find your most +[1116.96 --> 1117.80] perplexing application +[1117.80 --> 1118.56] issues. +[1118.86 --> 1119.70] Honeycomb is +[1119.70 --> 1121.24] a fast analysis +[1121.24 --> 1122.22] tool that reveals +[1122.22 --> 1122.80] the truth about +[1122.80 --> 1123.90] every aspect +[1123.90 --> 1125.02] of your application +[1125.02 --> 1125.78] in production. +[1126.24 --> 1126.88] Find out how users +[1126.88 --> 1127.66] experience your code +[1127.66 --> 1128.36] in complex +[1128.36 --> 1129.48] and unpredictable +[1129.48 --> 1130.24] environments. +[1130.58 --> 1131.26] Find patterns +[1131.26 --> 1132.28] and outliers +[1132.28 --> 1133.12] across billions +[1133.12 --> 1133.88] of rows of data +[1133.88 --> 1134.64] and definitively +[1134.64 --> 1135.44] solve your problems. +[1135.88 --> 1136.62] And we use Honeycomb +[1136.62 --> 1137.36] here at Change. +[1137.40 --> 1138.02] That's why we welcome +[1138.02 --> 1138.44] the opportunity +[1138.44 --> 1139.40] to add them +[1139.40 --> 1140.18] as one of our +[1140.18 --> 1141.20] infrastructure partners. +[1141.70 --> 1142.22] In particular, +[1142.22 --> 1143.12] we use Honeycomb +[1143.12 --> 1143.88] to track down +[1143.88 --> 1144.86] CDN issues recently, +[1145.18 --> 1145.68] which we talked +[1145.68 --> 1146.60] about at length +[1146.60 --> 1147.84] on the Kaizen edition +[1147.84 --> 1149.06] of the Ship It podcast. +[1149.30 --> 1149.98] So check that out. +[1150.20 --> 1150.70] Here's the thing. +[1150.92 --> 1151.78] Teams who don't use +[1151.78 --> 1152.18] Honeycomb +[1152.18 --> 1153.32] are forced to find +[1153.32 --> 1154.20] the needle in the haystack. +[1154.20 --> 1155.34] They scroll through +[1155.34 --> 1156.44] endless dashboards +[1156.44 --> 1157.48] playing whack-a-mole. +[1157.72 --> 1158.88] They deal with alert floods, +[1159.08 --> 1159.70] trying to guess +[1159.70 --> 1160.74] which one matters, +[1161.08 --> 1162.06] and they go from tool +[1162.06 --> 1162.92] to tool to tool +[1162.92 --> 1163.76] playing sleuth, +[1163.98 --> 1164.60] trying to figure out +[1164.60 --> 1165.72] how all the puzzle pieces +[1165.72 --> 1166.34] fit together. +[1166.68 --> 1167.94] It's this context switching +[1167.94 --> 1169.04] and tool sprawl +[1169.04 --> 1169.96] that are slowly killing +[1169.96 --> 1171.00] teams' effectiveness +[1171.00 --> 1172.36] and ultimately hindering +[1172.36 --> 1173.00] their business. +[1173.40 --> 1173.92] With Honeycomb, +[1174.00 --> 1174.98] you get a fast, +[1175.28 --> 1175.84] unified, +[1176.12 --> 1177.62] and clear understanding +[1177.62 --> 1179.02] of the one thing +[1179.02 --> 1180.16] driving your business. +[1180.42 --> 1180.84] Production. +[1181.36 --> 1181.94] With Honeycomb, +[1181.94 --> 1182.88] you guess less +[1182.88 --> 1183.84] and you know more. +[1184.30 --> 1184.90] Join the swarm +[1184.90 --> 1186.70] and try Honeycomb free today +[1186.70 --> 1188.30] at honeycomb.io +[1188.30 --> 1189.42] slash changelog. +[1189.58 --> 1190.14] Again, +[1190.28 --> 1191.40] honeycomb.io +[1191.40 --> 1193.06] slash changelog. +[1210.02 --> 1210.54] So, +[1210.74 --> 1211.62] you're going through that, +[1211.62 --> 1212.76] like what you just said, +[1212.82 --> 1213.28] made me think +[1213.28 --> 1214.30] maybe my idea +[1214.30 --> 1216.42] of what OOP is +[1216.42 --> 1218.02] is pretty heavily influenced +[1218.02 --> 1218.56] and biased +[1218.56 --> 1219.54] by writing Go. +[1219.76 --> 1219.98] Like, +[1220.02 --> 1220.88] I haven't written anything +[1220.88 --> 1223.08] but Go in a long time. +[1223.44 --> 1223.62] Right? +[1223.76 --> 1224.02] So, +[1224.50 --> 1224.70] like, +[1224.86 --> 1225.74] probably two or three years +[1225.74 --> 1226.28] at this point. +[1226.80 --> 1227.02] So, +[1227.10 --> 1227.70] going through that, +[1227.84 --> 1228.20] you're right. +[1228.30 --> 1228.46] Like, +[1228.86 --> 1230.00] I think a lot of the, +[1230.00 --> 1230.22] like, +[1230.32 --> 1231.42] OOP stuff that was taught +[1231.42 --> 1231.88] in school, +[1232.46 --> 1233.40] I don't use anymore, +[1233.40 --> 1234.36] but I still think of what +[1234.36 --> 1235.76] I'm doing as object-oriented. +[1236.06 --> 1236.94] And maybe I'm wrong. +[1237.14 --> 1237.30] Like, +[1237.98 --> 1238.18] like, +[1238.24 --> 1238.80] in my mind, +[1239.20 --> 1239.90] object-oriented +[1239.90 --> 1241.28] is just like encapsulating +[1241.28 --> 1242.58] a set of data +[1242.58 --> 1243.96] and having methods on it +[1243.96 --> 1245.68] that is like self-contained, +[1245.74 --> 1245.92] right? +[1245.92 --> 1246.84] And maybe that's not +[1246.84 --> 1247.48] what it actually is. +[1247.54 --> 1248.04] Maybe I'm totally +[1248.04 --> 1249.58] misunderstanding the concept. +[1250.10 --> 1250.26] So, +[1250.54 --> 1251.22] that's interesting. +[1251.44 --> 1252.00] You said it. +[1252.06 --> 1252.92] You actually said it. +[1253.58 --> 1254.86] Said encapsulation. +[1254.86 --> 1255.66] So, +[1255.70 --> 1256.80] it's about encapsulation, +[1256.92 --> 1257.40] abstraction, +[1258.00 --> 1259.48] and later came in +[1259.48 --> 1260.24] generalization. +[1261.12 --> 1261.52] So, +[1261.88 --> 1263.00] to encapsulate, +[1263.10 --> 1263.92] you don't need hierarchy. +[1265.00 --> 1265.38] Or, +[1265.74 --> 1266.16] I mean, +[1266.36 --> 1267.64] you can have a composition, +[1267.96 --> 1268.36] essentially, +[1268.64 --> 1269.76] and that would work +[1269.76 --> 1270.66] pretty well, +[1270.76 --> 1270.96] too, +[1271.20 --> 1272.00] to encapsulate. +[1272.70 --> 1272.94] But, +[1273.10 --> 1273.60] essentially, +[1273.72 --> 1273.86] like, +[1273.90 --> 1274.96] you don't even need to have +[1274.96 --> 1275.94] any kind of composition, +[1276.22 --> 1276.72] I think, +[1277.14 --> 1278.36] to be able to encapsulate +[1278.36 --> 1278.98] information. +[1279.10 --> 1280.06] It is useful, +[1280.30 --> 1281.34] but the combination +[1281.34 --> 1282.58] of information +[1282.58 --> 1283.20] can be, +[1283.26 --> 1283.56] you know, +[1283.62 --> 1284.64] across a lot of things. +[1284.96 --> 1285.88] We know this because +[1285.88 --> 1287.22] we have distributed systems. +[1287.84 --> 1288.62] I feel like +[1288.62 --> 1289.66] we live in a world +[1289.66 --> 1290.82] where everything is distributed, +[1290.96 --> 1291.06] like, +[1291.14 --> 1291.80] all of our data +[1291.80 --> 1291.98] is, +[1292.06 --> 1292.14] like, +[1292.16 --> 1292.52] everywhere. +[1293.02 --> 1294.02] And we kind of need +[1294.02 --> 1294.90] to aggregate it. +[1295.16 --> 1295.50] So, +[1295.70 --> 1296.46] even a composition +[1296.46 --> 1297.78] is sort of nice to have, +[1297.88 --> 1298.54] something that we're +[1298.54 --> 1299.58] very used to when we do, +[1300.10 --> 1301.42] but we don't actually +[1301.42 --> 1302.56] need to have it. +[1302.88 --> 1303.12] So, +[1303.28 --> 1304.60] if we have encapsulation, +[1304.92 --> 1306.54] which I think we do in Go, +[1307.12 --> 1308.58] and we are able to +[1308.58 --> 1309.82] define abstractions, +[1310.00 --> 1311.04] which we have +[1311.04 --> 1311.94] interfaces for. +[1312.48 --> 1312.66] And, +[1312.66 --> 1313.26] by the way, +[1313.34 --> 1314.32] there is a case, +[1314.32 --> 1314.92] so, +[1315.58 --> 1316.34] it's kind of interesting +[1316.34 --> 1317.84] as I found a code snippet +[1317.84 --> 1318.84] that was exactly +[1318.84 --> 1319.46] what I needed +[1319.46 --> 1321.12] for this workshop +[1321.12 --> 1322.38] at the time, +[1322.54 --> 1323.86] it was the C++ code. +[1324.00 --> 1324.64] I really loved +[1324.64 --> 1326.34] because it starts +[1326.34 --> 1327.22] with a license +[1327.22 --> 1328.12] that do whatever +[1328.12 --> 1328.92] the hell you want +[1328.92 --> 1329.56] with this code +[1329.56 --> 1331.04] as long as you don't blame me. +[1331.04 --> 1333.62] And then this code, +[1333.62 --> 1335.26] that it's an event +[1335.26 --> 1336.34] that you can register +[1336.34 --> 1337.22] listeners to, +[1337.70 --> 1339.18] regardless of the type. +[1339.40 --> 1339.96] So, essentially, +[1340.22 --> 1341.62] it works with templates +[1341.62 --> 1342.58] or generics. +[1342.90 --> 1343.38] So, +[1343.46 --> 1344.24] that code +[1344.24 --> 1345.64] uses generics +[1345.64 --> 1346.84] for a case +[1346.84 --> 1347.80] that we don't need. +[1348.36 --> 1349.40] When we write +[1349.40 --> 1350.00] this code +[1350.00 --> 1351.56] in C++, +[1351.82 --> 1352.60] we have to have, +[1352.70 --> 1353.52] we have to use +[1353.52 --> 1354.68] this kind of code +[1354.68 --> 1355.38] with templates +[1355.38 --> 1356.82] to plug in something +[1356.82 --> 1357.78] into something else, +[1357.78 --> 1359.08] to even be able +[1359.08 --> 1359.76] to do that, +[1359.82 --> 1360.64] to just plug in +[1360.64 --> 1361.38] some functionality +[1361.38 --> 1362.32] into something else +[1362.32 --> 1363.34] that it's not aware of. +[1363.44 --> 1363.62] Again, +[1363.72 --> 1364.64] because interfaces, +[1365.24 --> 1365.38] well, +[1365.50 --> 1367.00] C++ doesn't have interfaces. +[1367.22 --> 1368.06] It has classes +[1368.06 --> 1369.52] where you can define +[1369.52 --> 1370.84] functions as pure virtual, +[1371.44 --> 1372.20] so it doesn't have +[1372.20 --> 1373.48] to have all the implementation. +[1374.36 --> 1374.84] But, +[1375.08 --> 1376.80] whatever object you have, +[1376.84 --> 1377.76] you cannot pass in +[1377.76 --> 1378.30] as another +[1378.30 --> 1379.62] unless it implements, +[1380.08 --> 1381.40] it extends +[1381.40 --> 1382.72] or is essentially +[1382.72 --> 1384.04] a class. +[1384.62 --> 1385.62] But in this case, +[1385.62 --> 1387.18] you define your own +[1387.18 --> 1388.84] event handler, +[1389.08 --> 1389.52] obviously, +[1389.66 --> 1390.54] nobody knows, +[1390.98 --> 1391.54] you can't really +[1391.54 --> 1392.42] pass in +[1392.42 --> 1393.66] or use in this code +[1393.66 --> 1394.40] any class +[1394.40 --> 1394.74] that you didn't +[1394.74 --> 1395.44] write yourself. +[1395.96 --> 1396.38] So, +[1396.46 --> 1397.16] you get stuck +[1397.16 --> 1398.18] and this is why +[1398.18 --> 1399.34] you use generics +[1399.34 --> 1400.42] in many, +[1400.52 --> 1401.08] many languages. +[1401.42 --> 1402.24] It's for this case +[1402.24 --> 1403.46] and that was why +[1403.46 --> 1404.36] for many, +[1404.50 --> 1405.04] many years +[1405.04 --> 1405.96] people would say, +[1406.08 --> 1406.12] well, +[1406.16 --> 1406.66] why do you need +[1406.66 --> 1407.26] interfaces? +[1407.74 --> 1408.30] Why do you need +[1408.30 --> 1409.06] generics when you +[1409.06 --> 1409.78] have interfaces? +[1410.26 --> 1411.14] But the truth is +[1411.14 --> 1411.84] that you do need +[1411.84 --> 1413.08] generics for the cases +[1413.08 --> 1414.84] where the behavior +[1414.84 --> 1416.26] of the class +[1416.26 --> 1417.00] is derived +[1417.00 --> 1419.04] by the type. +[1419.70 --> 1419.88] So, +[1420.10 --> 1420.70] for instance, +[1421.20 --> 1422.68] next node +[1422.68 --> 1424.12] in a linked list. +[1424.76 --> 1424.92] So, +[1425.42 --> 1426.22] the next is going +[1426.22 --> 1426.76] to return +[1426.76 --> 1427.58] the same type +[1427.58 --> 1428.00] of node +[1428.00 --> 1428.74] and if your node +[1428.74 --> 1429.20] is holding +[1429.20 --> 1429.72] an integer +[1429.72 --> 1430.46] or a float +[1430.46 --> 1431.04] or whatever +[1431.04 --> 1432.08] type of value, +[1432.32 --> 1432.92] it's going +[1432.92 --> 1433.90] to be different. +[1434.42 --> 1435.20] A repository +[1435.20 --> 1436.00] that stores +[1436.00 --> 1437.16] any kind of model. +[1437.70 --> 1438.22] A map +[1438.22 --> 1439.18] with a key +[1439.18 --> 1440.26] value. +[1440.38 --> 1440.98] By the way, +[1441.48 --> 1442.42] go always +[1442.42 --> 1443.28] hot generics. +[1443.28 --> 1443.88] Actually, +[1443.98 --> 1444.56] that would be +[1444.56 --> 1445.90] my unpopular +[1445.90 --> 1446.34] opinion. +[1446.80 --> 1447.24] Go always +[1447.24 --> 1447.94] hot generics +[1447.94 --> 1448.94] because map +[1448.94 --> 1449.82] is a generic +[1449.82 --> 1450.20] type. +[1450.86 --> 1451.26] It maps +[1451.26 --> 1452.20] from a key +[1452.20 --> 1453.56] that is a generic +[1453.56 --> 1455.30] type to a value +[1455.30 --> 1456.06] that is a generic +[1456.06 --> 1456.46] type. +[1456.88 --> 1457.40] I see you're +[1457.40 --> 1458.38] sipping your drink +[1458.38 --> 1458.62] again, +[1458.70 --> 1459.06] so I know +[1459.06 --> 1459.66] that you disagree. +[1460.70 --> 1460.96] No, +[1461.08 --> 1462.18] I 100% agree. +[1462.68 --> 1462.84] Oh, +[1462.86 --> 1463.18] you do? +[1463.36 --> 1463.52] Okay. +[1464.12 --> 1464.82] The slices +[1464.82 --> 1465.92] were always +[1465.92 --> 1466.36] generic. +[1466.72 --> 1467.44] We just couldn't +[1467.44 --> 1468.40] define our own, +[1468.54 --> 1469.28] so it was like +[1469.28 --> 1470.52] good for the 80% +[1470.52 --> 1471.22] of the use +[1471.22 --> 1471.58] cases, +[1471.68 --> 1472.22] but then we +[1472.22 --> 1472.94] had our own +[1472.94 --> 1473.70] cases that we +[1473.70 --> 1474.18] needed it, +[1474.50 --> 1474.92] and we just +[1474.92 --> 1475.32] didn't have +[1475.32 --> 1475.68] anything. +[1476.14 --> 1476.86] So you will +[1476.86 --> 1477.40] talk now, +[1477.46 --> 1477.86] and I will +[1477.86 --> 1478.46] sip a drink. +[1480.10 --> 1480.50] No, +[1480.58 --> 1481.12] I definitely +[1481.12 --> 1481.58] agree with you +[1481.58 --> 1482.54] that the maps +[1482.54 --> 1483.10] and the slices +[1483.10 --> 1484.62] are generic. +[1484.74 --> 1485.42] They've literally +[1485.42 --> 1486.14] since day one, +[1486.86 --> 1487.26] and like you +[1487.26 --> 1487.38] said, +[1487.40 --> 1487.94] it is like that +[1487.94 --> 1488.90] 80% use case, +[1488.96 --> 1489.12] right? +[1489.50 --> 1489.80] I don't know +[1489.80 --> 1490.28] how to tie this +[1490.28 --> 1491.70] back to OOP +[1491.70 --> 1492.38] stuff now. +[1492.62 --> 1492.74] No, +[1492.80 --> 1494.24] because why I +[1494.24 --> 1495.02] think it plays +[1495.02 --> 1496.00] a role in this +[1496.00 --> 1497.30] is because what +[1497.30 --> 1498.90] people did in +[1498.90 --> 1499.66] those cases +[1499.66 --> 1500.52] is they would +[1500.52 --> 1501.88] write a lot +[1501.88 --> 1502.36] of functional +[1502.36 --> 1503.06] code, +[1503.76 --> 1505.02] procedural code, +[1505.34 --> 1505.74] et cetera, +[1506.22 --> 1507.64] and did not +[1507.64 --> 1508.92] use necessarily +[1508.92 --> 1510.92] generic types. +[1511.54 --> 1512.56] I think we've +[1512.56 --> 1513.26] had a lot of +[1513.26 --> 1514.02] workarounds. +[1515.08 --> 1515.66] I think now +[1515.66 --> 1516.88] we can eliminate +[1516.88 --> 1517.14] them. +[1517.24 --> 1517.70] I think the +[1517.70 --> 1518.76] language is very +[1518.76 --> 1519.84] mature to be +[1519.84 --> 1520.54] object-oriented. +[1521.00 --> 1521.20] No, +[1521.26 --> 1521.70] that makes sense +[1521.70 --> 1522.00] to me. +[1522.26 --> 1522.74] Natalie, +[1523.22 --> 1524.00] can you hear me? +[1524.14 --> 1524.46] No. +[1524.86 --> 1525.58] One thing is on +[1525.58 --> 1525.94] Zoom call, +[1526.04 --> 1526.56] another thing is +[1526.56 --> 1527.26] on a podcast. +[1528.00 --> 1528.34] It's a whole +[1528.34 --> 1528.72] new level. +[1528.90 --> 1530.10] I found a +[1530.10 --> 1530.66] quote from +[1530.66 --> 1531.20] Rob Pike +[1531.20 --> 1532.50] about Go +[1532.50 --> 1534.24] being yes +[1534.24 --> 1534.76] or no +[1534.76 --> 1536.22] object-oriented. +[1536.70 --> 1537.60] So it's a +[1537.60 --> 1538.22] bit of a +[1538.22 --> 1538.88] thread. +[1539.36 --> 1539.94] So it starts +[1539.94 --> 1540.46] with whenever +[1540.46 --> 1541.08] someone from +[1541.08 --> 1541.52] Java or +[1541.52 --> 1542.20] C++ or +[1542.20 --> 1542.68] C Sharp +[1542.68 --> 1544.82] comes to +[1544.82 --> 1545.02] Go, +[1545.16 --> 1545.58] they look for +[1545.58 --> 1545.94] class, +[1546.06 --> 1546.54] find struct +[1546.54 --> 1547.00] and stop +[1547.00 --> 1547.44] looking. +[1547.84 --> 1548.26] But this +[1548.26 --> 1548.78] misses two +[1548.78 --> 1549.38] fundamental +[1549.38 --> 1549.92] differences +[1549.92 --> 1550.54] between Go +[1550.54 --> 1551.54] and traditional +[1551.54 --> 1552.50] object-oriented +[1552.50 --> 1552.98] languages. +[1553.58 --> 1554.18] The first +[1554.18 --> 1554.70] one is that +[1554.70 --> 1555.04] it's not +[1555.04 --> 1555.84] only structs. +[1556.06 --> 1556.74] Any concrete +[1556.74 --> 1557.66] type can have +[1557.66 --> 1558.24] methods. +[1558.24 --> 1558.88] integers, +[1559.18 --> 1559.48] balls, +[1559.74 --> 1560.06] slices, +[1560.28 --> 1560.88] even funks. +[1561.32 --> 1561.94] But the +[1561.94 --> 1562.64] more important +[1562.64 --> 1563.86] idea is the +[1563.86 --> 1564.42] separation of +[1564.42 --> 1564.84] concepts. +[1564.98 --> 1565.40] Data and +[1565.40 --> 1566.00] behavior are +[1566.00 --> 1566.84] two distinct +[1566.84 --> 1567.42] concepts of +[1567.42 --> 1567.64] Go, +[1568.14 --> 1568.84] not conflated +[1568.84 --> 1569.44] into a single +[1569.44 --> 1569.98] notion of +[1569.98 --> 1570.42] class. +[1570.82 --> 1571.88] This is the +[1571.88 --> 1572.58] insight which +[1572.58 --> 1573.18] goes all the +[1573.18 --> 1573.70] way back to +[1573.70 --> 1574.70] small talk on +[1574.70 --> 1575.50] which the +[1575.50 --> 1576.32] object-oriented +[1576.32 --> 1577.10] type system, +[1577.24 --> 1577.76] including the +[1577.76 --> 1578.46] interface model, +[1578.58 --> 1578.98] is built. +[1579.40 --> 1579.70] So stopping +[1579.70 --> 1580.30] at struct +[1580.30 --> 1581.52] equals equals +[1581.52 --> 1582.58] class misses +[1582.58 --> 1583.86] much of what +[1583.86 --> 1584.50] makes Go +[1584.50 --> 1584.94] work. +[1584.94 --> 1585.66] So what +[1585.66 --> 1585.88] are your +[1585.88 --> 1586.20] thoughts? +[1586.60 --> 1587.18] I agree. +[1587.78 --> 1588.38] It sort of +[1588.38 --> 1589.00] works very +[1589.00 --> 1589.46] well with +[1589.46 --> 1590.10] everything that +[1590.10 --> 1591.02] I learned +[1591.02 --> 1592.14] recently about +[1592.14 --> 1592.52] this. +[1592.74 --> 1593.12] Honestly, +[1593.58 --> 1594.80] it feels like +[1594.80 --> 1595.72] Go is more +[1595.72 --> 1597.24] naturally object-oriented +[1597.24 --> 1599.16] than languages +[1599.16 --> 1600.00] that I worked +[1600.00 --> 1601.38] with before. +[1601.96 --> 1602.42] I had this +[1602.42 --> 1603.06] discussion with +[1603.06 --> 1603.66] my boyfriend +[1603.66 --> 1604.52] because of the +[1604.52 --> 1605.42] workshop and the +[1605.42 --> 1605.76] amount of +[1605.76 --> 1606.64] resistance that I +[1606.64 --> 1607.12] was getting. +[1607.50 --> 1608.14] Does he think +[1608.14 --> 1610.00] Go is object-oriented? +[1610.36 --> 1610.66] Yes. +[1611.04 --> 1611.86] So his answer +[1611.86 --> 1612.64] was, of course +[1612.64 --> 1613.06] it is. +[1613.62 --> 1614.88] It wasn't even a +[1614.88 --> 1615.58] question in his +[1615.58 --> 1615.78] mind. +[1616.24 --> 1616.76] And then he +[1616.76 --> 1617.42] told me, well, +[1617.48 --> 1618.18] you know how the +[1618.18 --> 1618.94] JavaScript people +[1618.94 --> 1620.00] say that a +[1620.00 --> 1621.30] prototype is the +[1621.30 --> 1622.04] best way to +[1622.04 --> 1623.58] define object-oriented. +[1623.94 --> 1624.88] And my takeaway +[1624.88 --> 1625.98] from this is that +[1625.98 --> 1627.12] every night I go +[1627.12 --> 1628.08] to sleep next to +[1628.08 --> 1628.84] a man who +[1628.84 --> 1629.48] quotes the +[1629.48 --> 1630.32] JavaScript people. +[1632.32 --> 1633.04] That's it. +[1633.84 --> 1634.84] I thought it was +[1634.84 --> 1635.58] a pretty cool +[1635.58 --> 1636.48] observation, to be +[1636.48 --> 1637.18] honest, but +[1637.18 --> 1638.76] still concerning. +[1638.76 --> 1640.38] I'll keep you +[1640.38 --> 1640.78] posted. +[1642.48 --> 1643.70] The quote +[1643.70 --> 1644.30] there where it +[1644.30 --> 1645.48] says it +[1645.48 --> 1646.66] separates the +[1646.66 --> 1647.58] idea of data +[1647.58 --> 1648.36] and behavior, I +[1648.36 --> 1648.68] think that's +[1648.68 --> 1649.34] the important +[1649.34 --> 1649.92] part of that +[1649.92 --> 1650.30] area. +[1650.74 --> 1651.24] But I'm having +[1651.24 --> 1652.08] a hard time +[1652.08 --> 1652.72] articulating +[1652.72 --> 1654.74] exactly why in +[1654.74 --> 1655.16] my head. +[1655.50 --> 1656.06] You think that +[1656.06 --> 1656.72] that is a good +[1656.72 --> 1657.84] way to express +[1657.84 --> 1658.90] object-oriented? +[1659.42 --> 1660.18] Yeah, I do. +[1660.36 --> 1661.12] Extracting away +[1661.12 --> 1662.36] the data from +[1662.36 --> 1663.38] the functionality? +[1663.80 --> 1664.46] Yeah, I think +[1664.46 --> 1665.42] going to what +[1665.42 --> 1665.84] you're saying +[1665.84 --> 1666.66] where Go is +[1666.66 --> 1667.50] almost a more +[1667.50 --> 1668.70] pure version of +[1668.70 --> 1669.56] object-orientation. +[1669.64 --> 1670.22] I think that's +[1670.22 --> 1671.24] what leads to +[1671.24 --> 1671.56] it, right? +[1671.62 --> 1672.42] The idea that, +[1672.92 --> 1673.26] I don't know, +[1673.32 --> 1673.70] I'd have to +[1673.70 --> 1674.16] really sit down +[1674.16 --> 1674.64] and think about +[1674.64 --> 1675.68] it, but that's +[1675.68 --> 1676.36] where my mind +[1676.36 --> 1676.80] is going. +[1677.08 --> 1678.12] I think that I +[1678.12 --> 1680.10] found something +[1680.10 --> 1681.80] that the person +[1681.80 --> 1683.46] that wrote me +[1683.46 --> 1684.68] about Go not +[1684.68 --> 1685.62] being object-oriented +[1685.62 --> 1687.42] language might be +[1687.42 --> 1688.48] related to. +[1689.04 --> 1689.98] So I searched +[1689.98 --> 1690.72] for Ron Pike +[1690.72 --> 1691.42] in object-oriented +[1691.42 --> 1692.22] and I found +[1692.22 --> 1692.74] Wikipedia. +[1693.96 --> 1695.02] And Wikipedia, +[1695.52 --> 1696.50] he is +[1696.50 --> 1697.26] criticizing +[1697.26 --> 1698.56] object-oriented. +[1699.12 --> 1700.18] So it's +[1700.18 --> 1700.86] possible that +[1700.86 --> 1701.76] that was what +[1701.76 --> 1702.78] that quote +[1702.78 --> 1703.50] was about. +[1703.84 --> 1704.30] If you want to +[1704.30 --> 1704.86] read this out +[1704.86 --> 1705.96] and share this +[1705.96 --> 1706.64] later in our +[1706.64 --> 1707.32] show notes, +[1707.40 --> 1708.18] everybody can go +[1708.18 --> 1708.86] back to this +[1708.86 --> 1709.06] too. +[1709.48 --> 1710.34] Yeah, I think +[1710.34 --> 1710.84] it's funny. +[1711.98 --> 1712.80] I mean, +[1712.94 --> 1713.46] clearly, +[1714.06 --> 1714.90] yes, I will +[1714.90 --> 1715.70] send it to you +[1715.70 --> 1716.94] and we can +[1716.94 --> 1717.58] include that. +[1718.06 --> 1719.56] So he criticized +[1719.56 --> 1720.54] object-oriented +[1720.54 --> 1721.24] for being +[1721.24 --> 1722.30] incredibly heavy, +[1722.92 --> 1723.38] I guess. +[1723.38 --> 1724.08] and that +[1724.08 --> 1724.80] sort of +[1724.80 --> 1725.60] works pretty +[1725.60 --> 1726.08] well with +[1726.08 --> 1726.78] those quotes +[1726.78 --> 1727.94] about objects +[1727.94 --> 1728.84] being lightweight +[1728.84 --> 1729.58] and go, +[1730.02 --> 1730.96] I think I agree. +[1731.46 --> 1731.98] I think it +[1731.98 --> 1732.60] feels very, +[1732.70 --> 1733.30] very natural. +[1733.72 --> 1734.14] By the way, +[1734.20 --> 1734.76] something that +[1734.76 --> 1735.72] I really love +[1735.72 --> 1737.42] is we don't +[1737.42 --> 1738.92] have enumerators, +[1739.12 --> 1739.62] but we do +[1739.62 --> 1740.12] have, +[1740.46 --> 1740.72] like, +[1741.18 --> 1742.36] it's very +[1742.36 --> 1743.40] easy for me +[1743.40 --> 1743.90] to, +[1744.36 --> 1744.82] when I look +[1744.82 --> 1745.70] at pseudocode +[1745.70 --> 1746.16] with those +[1746.16 --> 1746.84] red-black +[1746.84 --> 1748.72] trees and, +[1748.78 --> 1749.24] you know, +[1749.38 --> 1751.48] coloring nodes +[1751.48 --> 1752.06] in a graph +[1752.06 --> 1752.46] and stuff +[1752.46 --> 1752.94] like that. +[1753.28 --> 1753.76] This is +[1753.76 --> 1754.44] kind of stuff +[1754.44 --> 1755.12] that I teach +[1755.12 --> 1755.62] to people +[1755.62 --> 1756.24] who don't +[1756.24 --> 1756.90] have the, +[1757.80 --> 1758.18] let's say, +[1758.22 --> 1758.70] the classical +[1758.70 --> 1759.24] background +[1759.24 --> 1760.14] in the university +[1760.14 --> 1761.90] and I teach +[1761.90 --> 1762.52] that stuff +[1762.52 --> 1763.84] to people +[1763.84 --> 1764.46] because +[1764.46 --> 1766.14] they need +[1766.14 --> 1767.36] to pass +[1767.36 --> 1768.58] interviews. +[1769.30 --> 1769.48] So, +[1769.76 --> 1770.84] unfortunately, +[1771.46 --> 1773.04] this is where +[1773.04 --> 1773.60] this is coming +[1773.60 --> 1773.92] from. +[1774.28 --> 1774.76] But what I +[1774.76 --> 1775.58] really like +[1775.58 --> 1776.02] to do, +[1776.12 --> 1776.56] because I +[1776.56 --> 1777.30] don't, +[1777.74 --> 1778.30] I can't +[1778.30 --> 1778.88] code those, +[1779.20 --> 1779.86] so I take +[1779.86 --> 1780.56] pseudocode +[1780.56 --> 1781.54] with white, +[1781.76 --> 1782.20] black, +[1782.44 --> 1782.76] gray, +[1782.92 --> 1783.30] whatever, +[1783.58 --> 1784.40] red colors. +[1784.94 --> 1785.76] I define a +[1785.76 --> 1786.78] color that is, +[1787.24 --> 1787.48] that, +[1787.68 --> 1788.36] I don't define +[1788.36 --> 1788.90] a new type +[1788.90 --> 1789.46] of alias +[1789.46 --> 1789.98] integers +[1789.98 --> 1791.40] to do that +[1791.40 --> 1791.84] and I get +[1791.84 --> 1792.16] very, +[1792.30 --> 1793.06] very expressive +[1793.06 --> 1794.52] and basically +[1794.52 --> 1795.62] write pseudocode +[1795.62 --> 1795.98] and go. +[1796.76 --> 1797.76] And I don't +[1797.76 --> 1798.28] think that you +[1798.28 --> 1798.96] could do that +[1798.96 --> 1801.04] with a non-object +[1801.04 --> 1801.80] oriented language. +[1801.96 --> 1802.84] I can pretty +[1802.84 --> 1803.52] much express +[1803.52 --> 1804.44] any pseudocode +[1804.44 --> 1805.16] that I see +[1805.16 --> 1806.24] without actually +[1806.24 --> 1806.84] giving much +[1806.84 --> 1807.56] thought to +[1807.56 --> 1808.16] what is written +[1808.16 --> 1808.44] there. +[1808.58 --> 1808.86] I just +[1808.86 --> 1810.14] essentially copy +[1810.14 --> 1810.70] and paste it +[1810.70 --> 1811.00] and make +[1811.00 --> 1811.36] the code +[1811.36 --> 1811.72] work. +[1812.48 --> 1812.72] Like, +[1812.80 --> 1813.14] I just +[1813.14 --> 1813.54] make the +[1813.54 --> 1814.32] types work +[1814.32 --> 1814.82] and the +[1814.82 --> 1815.20] functions +[1815.20 --> 1815.96] work and +[1815.96 --> 1816.42] if the +[1816.42 --> 1816.74] function +[1816.74 --> 1817.32] happens to +[1817.32 --> 1817.84] be a method, +[1818.02 --> 1818.50] that's what's +[1818.50 --> 1819.10] going to be. +[1819.68 --> 1820.48] And if the +[1820.48 --> 1821.28] function happens +[1821.28 --> 1821.78] to be a +[1821.78 --> 1822.08] procedural +[1822.08 --> 1822.62] function, +[1822.80 --> 1823.18] that's what +[1823.18 --> 1823.48] it's going +[1823.48 --> 1824.02] to be and +[1824.02 --> 1824.34] just, +[1824.48 --> 1824.64] you know, +[1824.68 --> 1825.08] make it +[1825.08 --> 1825.40] work, +[1825.88 --> 1827.00] whatever is +[1827.00 --> 1827.82] necessary. +[1827.96 --> 1828.22] I don't +[1828.22 --> 1828.54] think I +[1828.54 --> 1828.80] would have +[1828.80 --> 1829.26] been able +[1829.26 --> 1829.68] to do +[1829.68 --> 1830.44] that if +[1830.44 --> 1830.86] Go was +[1830.86 --> 1831.54] not object +[1831.54 --> 1831.94] oriented. +[1832.46 --> 1832.86] Also, +[1833.14 --> 1833.58] you should +[1833.58 --> 1834.06] not be +[1834.06 --> 1834.88] testing people +[1834.88 --> 1835.36] in red, +[1835.44 --> 1836.22] black trees. +[1836.22 --> 1839.14] I definitely +[1839.14 --> 1839.76] could not +[1839.76 --> 1840.34] code up a +[1840.34 --> 1840.46] red, +[1840.52 --> 1840.88] black tree +[1840.88 --> 1841.68] without +[1841.68 --> 1842.56] Googling +[1842.56 --> 1842.84] something +[1842.84 --> 1843.24] right now. +[1843.36 --> 1843.54] Yeah, +[1843.58 --> 1844.06] obviously. +[1844.60 --> 1844.74] No, +[1844.84 --> 1845.54] I still do +[1845.54 --> 1846.36] these exercises +[1846.36 --> 1846.92] with people +[1846.92 --> 1847.44] so they get +[1847.44 --> 1847.94] used to the +[1847.94 --> 1848.52] idea and +[1848.52 --> 1848.80] then if +[1848.80 --> 1849.08] they can +[1849.08 --> 1849.54] do this, +[1849.60 --> 1849.82] they can +[1849.82 --> 1850.44] do anything, +[1850.62 --> 1851.00] stuff like +[1851.00 --> 1851.30] that, +[1851.42 --> 1851.80] essentially. +[1852.46 --> 1852.76] But yeah, +[1852.80 --> 1853.24] there are way +[1853.24 --> 1853.68] too many +[1853.68 --> 1854.22] edge cases +[1854.22 --> 1854.56] in the +[1854.56 --> 1854.70] red, +[1854.78 --> 1855.20] black tree. +[1855.96 --> 1856.40] Okay. +[1857.98 --> 1858.70] Somebody should +[1858.70 --> 1859.72] simplify this. +[1859.92 --> 1860.42] Someone should +[1860.42 --> 1860.72] write a +[1860.72 --> 1861.18] library, +[1861.54 --> 1862.06] generic library +[1862.06 --> 1862.62] for it once +[1862.62 --> 1863.14] and then never +[1863.14 --> 1863.72] touch it again. +[1863.72 --> 1865.46] That sounds +[1865.46 --> 1865.90] very good. +[1866.00 --> 1866.72] So let's rant +[1866.72 --> 1867.42] about people +[1867.42 --> 1868.44] who give +[1868.44 --> 1869.62] impossible tests +[1869.62 --> 1871.52] to complete +[1871.52 --> 1872.14] beginners in +[1872.14 --> 1872.42] coding. +[1872.94 --> 1873.78] Or tell us +[1873.78 --> 1874.24] how you do +[1874.24 --> 1874.70] this right +[1874.70 --> 1875.30] when you're +[1875.30 --> 1875.78] hiring for +[1875.78 --> 1876.16] your team. +[1876.66 --> 1877.68] So I don't +[1877.68 --> 1878.16] know that I +[1878.16 --> 1878.70] do this right. +[1878.92 --> 1879.38] Let's start +[1879.38 --> 1879.76] with that. +[1880.04 --> 1880.74] But I try +[1880.74 --> 1881.32] to decide +[1881.32 --> 1882.06] what I want +[1882.06 --> 1883.28] and then I +[1883.28 --> 1883.82] don't test +[1883.82 --> 1884.38] people and +[1884.38 --> 1884.90] stuff that I +[1884.90 --> 1885.28] don't want. +[1885.40 --> 1885.96] So what I +[1885.96 --> 1886.62] will do, +[1886.78 --> 1887.24] for instance, +[1887.34 --> 1887.88] like if I +[1887.88 --> 1888.34] care, +[1889.02 --> 1889.46] if I want +[1889.46 --> 1890.06] to know how +[1890.06 --> 1891.00] thorough somebody +[1891.00 --> 1891.54] is going to +[1891.54 --> 1891.82] be, +[1891.98 --> 1892.80] that I +[1892.80 --> 1893.70] will dive +[1893.70 --> 1894.58] with them +[1894.58 --> 1895.02] on something +[1895.02 --> 1895.34] that they +[1895.34 --> 1895.82] do know +[1895.82 --> 1896.34] and check +[1896.34 --> 1896.96] how familiar +[1896.96 --> 1897.46] they are +[1897.46 --> 1897.82] with all +[1897.82 --> 1898.30] the details. +[1899.38 --> 1900.56] So if +[1900.56 --> 1901.46] somebody tells +[1901.46 --> 1901.88] me, I +[1901.88 --> 1902.12] don't know, +[1902.14 --> 1902.50] let's start +[1902.50 --> 1902.78] with the +[1902.78 --> 1903.44] very basics. +[1903.64 --> 1904.04] If somebody +[1904.04 --> 1904.76] tells me +[1904.76 --> 1905.92] that they +[1905.92 --> 1906.82] have worked +[1906.82 --> 1907.38] with my +[1907.38 --> 1907.90] SQL, +[1908.22 --> 1908.96] then I +[1908.96 --> 1910.14] will test +[1910.14 --> 1911.12] them heavily +[1911.12 --> 1911.86] on my +[1911.86 --> 1912.56] SQL until +[1912.56 --> 1912.86] I reach +[1912.86 --> 1913.54] something that +[1913.54 --> 1914.42] either they +[1914.42 --> 1914.94] don't know +[1914.94 --> 1916.24] or to get +[1916.24 --> 1917.04] a sense of +[1917.04 --> 1917.76] how far +[1917.76 --> 1918.46] into the +[1918.46 --> 1918.96] rabbit hole +[1918.96 --> 1919.52] they went. +[1920.28 --> 1921.26] Or any +[1921.26 --> 1921.66] topic, +[1921.76 --> 1922.10] any other +[1922.10 --> 1922.46] topic. +[1922.46 --> 1923.22] this one +[1923.22 --> 1923.72] is kind +[1923.72 --> 1924.78] of normal. +[1925.18 --> 1925.52] But it +[1925.52 --> 1926.00] all depends +[1926.00 --> 1926.28] on the +[1926.28 --> 1926.78] profile. +[1927.12 --> 1927.38] It really +[1927.38 --> 1927.84] all depends +[1927.84 --> 1928.10] on the +[1928.10 --> 1928.38] profile. +[1929.22 --> 1930.06] So we +[1930.06 --> 1930.70] said earlier +[1930.70 --> 1931.18] that I +[1931.18 --> 1931.54] am the +[1931.54 --> 1931.86] sum of +[1931.86 --> 1931.96] the +[1931.96 --> 1932.48] opportunities +[1932.48 --> 1932.86] that were +[1932.86 --> 1933.32] given to +[1933.32 --> 1933.60] me. +[1934.34 --> 1935.04] And I +[1935.04 --> 1935.40] don't think +[1935.40 --> 1935.80] I'm a +[1935.80 --> 1936.50] bad developer. +[1937.08 --> 1937.54] We all +[1937.54 --> 1938.12] want to +[1938.12 --> 1939.02] hire somebody +[1939.02 --> 1939.56] who cannot +[1939.56 --> 1940.08] do what +[1940.08 --> 1940.72] we can't +[1940.72 --> 1940.90] do. +[1941.46 --> 1941.90] Essentially, +[1942.24 --> 1942.72] like I feel +[1942.72 --> 1943.08] like we are +[1943.08 --> 1943.62] raising the +[1943.62 --> 1944.50] bar constantly +[1944.50 --> 1945.20] to a level +[1945.20 --> 1945.58] that people +[1945.58 --> 1946.36] cannot actually +[1946.36 --> 1946.82] meet. +[1947.28 --> 1948.04] Not because +[1948.04 --> 1948.84] they're not +[1948.84 --> 1949.44] good enough. +[1949.44 --> 1950.58] So for +[1950.58 --> 1950.86] instance, +[1950.86 --> 1951.32] if I +[1951.32 --> 1951.80] can't do +[1951.80 --> 1952.50] something and +[1952.50 --> 1952.92] I am going +[1952.92 --> 1953.50] to search for +[1953.50 --> 1954.08] somebody who +[1954.08 --> 1954.46] is going +[1954.46 --> 1956.24] to be able +[1956.24 --> 1956.86] to do it, +[1956.92 --> 1957.76] that's a little +[1957.76 --> 1958.34] bit unfair. +[1959.06 --> 1960.00] And I feel +[1960.00 --> 1960.48] like we are +[1960.48 --> 1961.00] doing this +[1961.00 --> 1961.54] all the time. +[1961.62 --> 1961.92] We're sort +[1961.92 --> 1962.50] of like raising +[1962.50 --> 1962.90] the bar, +[1963.00 --> 1963.30] raising the +[1963.30 --> 1963.46] bar, +[1963.52 --> 1963.84] raising the +[1963.84 --> 1964.16] bar to +[1964.16 --> 1964.70] essentially +[1964.70 --> 1966.22] think maybe +[1966.22 --> 1967.12] make up for +[1967.12 --> 1967.96] our own +[1967.96 --> 1969.04] disadvantages. +[1969.92 --> 1970.86] So I +[1970.86 --> 1971.76] try to +[1971.76 --> 1972.30] not be +[1972.30 --> 1972.78] unreasonable. +[1973.50 --> 1973.92] When I +[1973.92 --> 1974.22] interview +[1974.22 --> 1974.60] somebody, +[1974.70 --> 1975.24] I do think +[1975.24 --> 1975.64] about, +[1976.28 --> 1977.02] I do try +[1977.02 --> 1977.92] to see if +[1977.92 --> 1978.32] they can +[1978.32 --> 1979.80] learn during +[1979.80 --> 1980.86] the interview +[1980.86 --> 1981.60] process, +[1981.70 --> 1982.04] for instance. +[1982.22 --> 1983.16] So I +[1983.16 --> 1983.82] will present +[1983.82 --> 1984.22] them with +[1984.22 --> 1985.66] something that +[1985.66 --> 1985.96] they will +[1985.96 --> 1986.60] have to +[1986.60 --> 1987.60] think about +[1987.60 --> 1988.74] and sort +[1988.74 --> 1989.92] of internalize +[1989.92 --> 1990.92] and then +[1990.92 --> 1991.68] spit out some +[1991.68 --> 1992.98] information that +[1992.98 --> 1993.78] could be wrong, +[1993.84 --> 1994.40] could be right, +[1994.48 --> 1994.92] but you know, +[1995.38 --> 1995.98] it could be +[1995.98 --> 1996.70] very small +[1996.70 --> 1997.34] things that +[1997.34 --> 1998.22] tell me if +[1998.22 --> 1998.86] they get it. +[1999.50 --> 2000.20] For instance, +[2000.74 --> 2001.46] I work with +[2001.46 --> 2001.98] billing. +[2002.42 --> 2003.02] People are not +[2003.02 --> 2004.04] necessarily familiar +[2004.04 --> 2004.80] with those +[2004.80 --> 2005.36] concepts, +[2005.36 --> 2006.60] so I will +[2006.60 --> 2007.06] introduce +[2007.06 --> 2008.02] something that +[2008.02 --> 2008.80] has something +[2008.80 --> 2009.36] to do with +[2009.36 --> 2010.00] what we do +[2010.00 --> 2010.54] in the domain +[2010.54 --> 2011.80] and it's +[2011.80 --> 2012.80] fine that +[2012.80 --> 2013.16] they don't +[2013.16 --> 2013.62] know this +[2013.62 --> 2014.06] and then I +[2014.06 --> 2014.68] won't ask +[2014.68 --> 2014.86] them, +[2014.98 --> 2015.38] so how +[2015.38 --> 2016.12] would you +[2016.12 --> 2016.58] do it? +[2016.68 --> 2016.86] Like, +[2016.94 --> 2017.50] given those +[2017.50 --> 2018.18] constraints, +[2018.82 --> 2019.20] what would +[2019.20 --> 2019.56] you do? +[2019.92 --> 2020.60] So I taught +[2020.60 --> 2021.30] them something +[2021.30 --> 2022.62] and now I +[2022.62 --> 2023.34] can see +[2023.34 --> 2024.20] how they +[2024.20 --> 2024.94] actually learn +[2024.94 --> 2025.92] and what +[2025.92 --> 2026.42] they actually +[2026.42 --> 2026.86] do with +[2026.86 --> 2027.58] this information +[2027.58 --> 2028.76] and it's +[2028.76 --> 2029.22] not really +[2029.22 --> 2030.04] just problem +[2030.04 --> 2030.44] solving, +[2030.76 --> 2031.20] it's more +[2031.20 --> 2031.78] of trying +[2031.78 --> 2032.28] not to +[2032.28 --> 2032.64] introduce +[2032.64 --> 2032.98] them to +[2032.98 --> 2033.32] problems +[2033.32 --> 2033.56] that they +[2033.56 --> 2034.10] already know, +[2034.54 --> 2034.92] essentially. +[2034.92 --> 2036.74] So it +[2036.74 --> 2037.00] could be +[2037.00 --> 2037.28] very, +[2037.38 --> 2037.90] very small +[2037.90 --> 2038.34] things. +[2039.34 --> 2039.48] Yeah, +[2039.60 --> 2040.12] so I am +[2040.12 --> 2040.56] the sum of +[2040.56 --> 2041.08] the opportunities +[2041.08 --> 2041.38] that were +[2041.38 --> 2041.80] given to +[2041.80 --> 2042.12] me and +[2042.12 --> 2043.16] I'm a +[2043.16 --> 2044.10] huge believer +[2044.10 --> 2044.78] that if you +[2044.78 --> 2045.20] give people +[2045.20 --> 2045.80] an opportunity +[2045.80 --> 2046.32] they're not +[2046.32 --> 2046.72] going to let +[2046.72 --> 2047.14] you down, +[2047.54 --> 2048.36] especially if +[2048.36 --> 2048.98] they've been +[2048.98 --> 2049.90] begging for +[2049.90 --> 2051.16] work since +[2051.16 --> 2051.60] forever. +[2052.34 --> 2053.12] And I do +[2053.12 --> 2054.12] think that you +[2054.12 --> 2054.80] can bank on +[2054.80 --> 2055.56] people's loyalty +[2055.56 --> 2056.86] in those cases. +[2056.86 --> 2057.98] I think for a lot +[2057.98 --> 2058.38] of people, +[2059.18 --> 2060.00] just having +[2060.00 --> 2061.24] their foot in +[2061.24 --> 2062.50] the door is +[2062.50 --> 2063.02] such a big +[2063.02 --> 2063.54] deal for +[2063.54 --> 2063.86] them, +[2064.22 --> 2064.56] that they +[2064.56 --> 2065.94] will pay +[2065.94 --> 2066.64] space. +[2067.22 --> 2068.38] So I do +[2068.38 --> 2069.22] see a lot +[2069.22 --> 2069.80] of value +[2069.80 --> 2070.92] in these +[2070.92 --> 2071.20] people. +[2071.70 --> 2072.76] And even +[2072.76 --> 2073.50] now, if I +[2073.50 --> 2074.12] go for a +[2074.12 --> 2074.76] job interview, +[2075.12 --> 2076.56] I am never +[2076.56 --> 2077.32] going to apply +[2077.32 --> 2077.80] for something +[2077.80 --> 2078.34] that I already +[2078.34 --> 2078.62] know. +[2078.72 --> 2079.24] I want to +[2079.24 --> 2079.96] grow just as +[2079.96 --> 2080.28] much as +[2080.28 --> 2080.90] everyone else. +[2080.98 --> 2081.30] So I'm +[2081.30 --> 2081.86] always going +[2081.86 --> 2083.54] to go and +[2083.54 --> 2084.38] try to +[2084.38 --> 2085.00] convince somebody +[2085.00 --> 2085.58] that they +[2085.58 --> 2086.00] need to +[2086.00 --> 2086.48] invest in +[2086.48 --> 2087.46] my potential. +[2087.46 --> 2088.82] I am +[2088.82 --> 2090.80] very, very +[2090.80 --> 2093.50] similar to +[2093.50 --> 2093.84] a complete +[2093.84 --> 2095.24] beginner because +[2095.24 --> 2096.06] we all have +[2096.06 --> 2097.34] to go around +[2097.34 --> 2098.32] and essentially +[2098.32 --> 2099.12] convince people +[2099.12 --> 2099.86] to give us +[2099.86 --> 2101.34] a chance to +[2101.34 --> 2102.00] prove that we +[2102.00 --> 2102.36] can do +[2102.36 --> 2103.16] something that +[2103.16 --> 2104.40] we have not +[2104.40 --> 2105.16] shown before +[2105.16 --> 2105.72] that we can +[2105.72 --> 2105.92] do. +[2106.68 --> 2107.14] It's so +[2107.14 --> 2107.68] obvious to +[2107.68 --> 2108.16] people that +[2108.16 --> 2108.48] okay, +[2108.76 --> 2109.16] Rona after +[2109.16 --> 2109.88] 20 years is +[2109.88 --> 2110.32] going to be +[2110.32 --> 2110.80] able to do +[2110.80 --> 2111.14] this. +[2111.68 --> 2112.10] But it +[2112.10 --> 2112.62] shouldn't be, +[2112.92 --> 2113.18] right? +[2113.44 --> 2113.84] I mean, at +[2113.84 --> 2114.20] the end of +[2114.20 --> 2114.68] the day, +[2115.28 --> 2115.98] it really +[2115.98 --> 2116.60] shouldn't be. +[2116.60 --> 2117.60] I fail just +[2117.60 --> 2118.06] as much as +[2118.06 --> 2118.72] everyone else. +[2119.10 --> 2119.52] So that's +[2119.52 --> 2119.90] why I go +[2119.90 --> 2120.34] with it. +[2120.34 --> 2120.70] It's a very +[2120.70 --> 2121.40] long answer +[2121.40 --> 2121.70] to your +[2121.70 --> 2121.98] question, +[2122.10 --> 2122.30] Loli. +[2133.56 --> 2134.54] This episode +[2134.54 --> 2135.38] is brought +[2135.38 --> 2135.74] to you by +[2135.74 --> 2136.34] our friends +[2136.34 --> 2136.98] at Acuity, +[2137.26 --> 2137.76] a new +[2137.76 --> 2138.38] platform that +[2138.38 --> 2138.90] brings fully +[2138.90 --> 2139.70] managed Argo +[2139.70 --> 2140.60] CD and +[2140.60 --> 2141.00] enterprise +[2141.00 --> 2141.86] services to +[2141.86 --> 2142.22] the cloud +[2142.22 --> 2142.64] or on +[2142.64 --> 2143.02] premise. +[2143.58 --> 2143.80] And I'm +[2143.80 --> 2144.02] here with +[2144.02 --> 2144.34] two of the +[2144.34 --> 2144.88] co-founders +[2144.88 --> 2145.38] from Acuity, +[2145.68 --> 2146.32] Jesse Suin, +[2146.32 --> 2146.64] and +[2146.64 --> 2147.24] Alexander +[2147.24 --> 2148.24] Matusenchev. +[2148.50 --> 2148.74] So the +[2148.74 --> 2149.32] Acuity platform +[2149.32 --> 2150.24] is in +[2150.24 --> 2150.72] beta right +[2150.72 --> 2151.06] now. +[2151.32 --> 2151.58] You guys +[2151.58 --> 2151.94] have some +[2151.94 --> 2152.60] big ideas +[2152.60 --> 2153.00] you're executing +[2153.00 --> 2153.60] on around +[2153.60 --> 2154.20] Argo CD, +[2154.58 --> 2155.20] managed Argo +[2155.20 --> 2155.56] CD, +[2155.84 --> 2156.02] Kubernetes +[2156.02 --> 2156.78] native application +[2156.78 --> 2157.20] delivery, +[2157.54 --> 2157.74] and the +[2157.74 --> 2158.30] power of +[2158.30 --> 2158.68] GitOps. +[2158.74 --> 2158.96] Help me +[2158.96 --> 2159.54] understand the +[2159.54 --> 2159.96] what and +[2159.96 --> 2160.52] the why +[2160.52 --> 2161.00] of what +[2161.00 --> 2161.20] you're doing +[2161.20 --> 2161.62] right now. +[2161.88 --> 2162.44] So we +[2162.44 --> 2163.20] started Acuity +[2163.20 --> 2164.28] because we +[2164.28 --> 2165.10] saw what was +[2165.10 --> 2165.88] happening in +[2165.88 --> 2166.50] the Kubernetes +[2166.50 --> 2167.02] community, +[2167.26 --> 2167.84] the challenges +[2167.84 --> 2168.32] that people +[2168.32 --> 2168.90] were facing +[2168.90 --> 2170.22] about developer +[2170.22 --> 2170.96] experience. +[2171.36 --> 2171.82] And having +[2171.82 --> 2172.72] run Argo +[2172.72 --> 2173.48] CD for +[2173.48 --> 2174.18] Intuit for +[2174.18 --> 2174.46] a couple +[2174.46 --> 2174.78] of years, +[2174.78 --> 2175.36] we knew it +[2175.36 --> 2175.68] took a +[2175.68 --> 2176.44] small team +[2176.44 --> 2177.00] to build +[2177.00 --> 2177.46] this and +[2177.46 --> 2177.88] scale it +[2177.88 --> 2178.36] and provide +[2178.36 --> 2179.12] a performance +[2179.12 --> 2180.16] solution for +[2180.16 --> 2181.06] the developers. +[2181.58 --> 2181.90] And so +[2181.90 --> 2182.80] at Acuity, +[2182.94 --> 2183.14] in the +[2183.14 --> 2183.74] Acuity platform, +[2183.92 --> 2184.14] what we're +[2184.14 --> 2184.84] trying to do +[2184.84 --> 2185.22] is, +[2185.34 --> 2185.80] the first thing +[2185.80 --> 2186.08] we're trying +[2186.08 --> 2186.44] to do is +[2186.44 --> 2187.48] actually provide +[2187.48 --> 2188.22] Argo CD +[2188.22 --> 2189.38] as a fully +[2189.38 --> 2190.36] managed solution +[2190.36 --> 2191.30] to our users. +[2191.30 --> 2191.36] developers. +[2191.66 --> 2192.36] But that is +[2192.36 --> 2192.90] just actually +[2192.90 --> 2193.42] the start +[2193.42 --> 2193.82] of things. +[2193.94 --> 2194.12] And we +[2194.12 --> 2195.06] actually want +[2195.06 --> 2195.88] to take +[2195.88 --> 2196.34] the next +[2196.34 --> 2197.36] steps on +[2197.36 --> 2198.12] improving +[2198.12 --> 2198.82] the whole +[2198.82 --> 2199.42] GitOps and +[2199.42 --> 2199.74] developer +[2199.74 --> 2200.56] experience and +[2200.56 --> 2201.48] providing new +[2201.48 --> 2201.96] tools and +[2201.96 --> 2203.32] ecosystems around +[2203.32 --> 2204.10] Argo and +[2204.10 --> 2204.34] the Argo +[2204.34 --> 2204.76] project. +[2205.10 --> 2205.28] Yeah, +[2205.38 --> 2205.70] that's right, +[2205.76 --> 2205.90] JC. +[2206.08 --> 2206.80] So Argo CD +[2206.80 --> 2207.34] is just +[2207.34 --> 2207.74] the beginning, +[2208.12 --> 2208.86] but every +[2208.86 --> 2210.06] company eventually +[2210.06 --> 2211.20] needs way more +[2211.20 --> 2212.16] tools integrated +[2212.16 --> 2212.78] into the +[2212.78 --> 2213.50] DevOps platform. +[2213.96 --> 2214.42] And that's +[2214.42 --> 2214.78] what we're +[2214.78 --> 2215.18] hoping to +[2215.18 --> 2215.86] deliver with +[2215.86 --> 2216.26] Acuity +[2216.26 --> 2216.64] platform. +[2217.14 --> 2217.50] So we're +[2217.50 --> 2217.92] hoping to +[2217.92 --> 2218.70] provide a +[2218.70 --> 2219.14] great user +[2219.14 --> 2219.96] interface that +[2219.96 --> 2220.82] enable developers +[2220.82 --> 2221.96] to achieve +[2221.96 --> 2222.38] what they +[2222.38 --> 2223.12] need in a +[2223.12 --> 2223.46] matter of +[2223.46 --> 2223.82] just a few +[2223.82 --> 2224.12] clicks. +[2224.58 --> 2224.94] But we +[2224.94 --> 2225.42] also want +[2225.42 --> 2225.88] to make +[2225.88 --> 2226.42] Argo CD +[2226.42 --> 2227.18] enterprise +[2227.18 --> 2227.44] ready. +[2227.92 --> 2228.24] What that +[2228.24 --> 2228.98] means is +[2228.98 --> 2229.94] our customers +[2229.94 --> 2230.54] will get +[2230.54 --> 2232.18] audits and +[2232.18 --> 2232.90] insightful +[2232.90 --> 2234.04] analytics out +[2234.04 --> 2234.54] of the box +[2234.54 --> 2235.14] without +[2235.14 --> 2236.04] configuring +[2236.04 --> 2236.46] anything. +[2236.94 --> 2237.36] That's what +[2237.36 --> 2238.26] we did at +[2238.26 --> 2238.80] Intuit and +[2238.80 --> 2239.16] we learned +[2239.16 --> 2239.64] that it was +[2239.64 --> 2240.16] not so easy +[2240.16 --> 2240.52] to do. +[2240.88 --> 2241.20] And that's +[2241.20 --> 2241.50] what we're +[2241.50 --> 2241.96] hoping to +[2241.96 --> 2242.46] solve for +[2242.46 --> 2242.86] multiple +[2242.86 --> 2243.70] organizations. +[2244.22 --> 2244.48] Very cool. +[2244.56 --> 2244.84] Thank you, +[2244.88 --> 2245.08] Jesse. +[2245.24 --> 2245.60] Thank you, +[2245.66 --> 2245.96] Alex. +[2246.06 --> 2246.46] Again, +[2246.54 --> 2246.84] listeners, +[2247.10 --> 2248.34] this is a +[2248.34 --> 2249.08] closed beta. +[2249.42 --> 2249.72] Check it +[2249.72 --> 2250.04] out. +[2250.20 --> 2250.62] Acuity +[2250.62 --> 2251.62] dot IO +[2251.62 --> 2252.10] slash +[2252.10 --> 2252.86] changelog. +[2252.92 --> 2253.44] Head there +[2253.44 --> 2253.88] and see +[2253.88 --> 2254.14] what this +[2254.14 --> 2255.08] platform is +[2255.08 --> 2255.80] all about. +[2256.14 --> 2256.42] Again, +[2256.58 --> 2257.52] acuity dot IO +[2257.52 --> 2257.90] slash +[2257.90 --> 2258.48] changelog. +[2258.54 --> 2259.30] Links are +[2259.30 --> 2259.72] in the +[2259.72 --> 2260.20] show notes. +[2260.76 --> 2260.96] And by +[2260.96 --> 2261.52] our friends +[2261.52 --> 2262.06] at Launch +[2262.06 --> 2262.48] Darkly, +[2262.76 --> 2263.18] fundamentally +[2263.18 --> 2263.92] change how +[2263.92 --> 2264.26] you deliver +[2264.26 --> 2264.86] software, +[2265.32 --> 2265.52] innovate +[2265.52 --> 2266.20] faster, +[2266.54 --> 2266.76] deploy +[2266.76 --> 2267.16] fearlessly, +[2267.66 --> 2267.92] and take +[2267.92 --> 2268.30] control of +[2268.30 --> 2268.78] your software +[2268.78 --> 2269.12] so you can +[2269.12 --> 2269.56] ship value +[2269.56 --> 2269.98] to customers +[2269.98 --> 2270.90] faster and +[2270.90 --> 2271.56] get feedback +[2271.56 --> 2272.00] sooner. +[2272.46 --> 2272.60] Launch +[2272.60 --> 2273.06] Darkly is +[2273.06 --> 2273.38] built for +[2273.38 --> 2274.04] developers but +[2274.04 --> 2274.50] empowers the +[2274.50 --> 2274.90] entire +[2274.90 --> 2275.50] organization. +[2276.00 --> 2276.28] Get started +[2276.28 --> 2276.80] for free and +[2276.80 --> 2277.36] get a demo +[2277.36 --> 2277.78] at +[2277.78 --> 2278.78] launchdarkly.com. +[2278.78 --> 2279.62] Again, +[2279.96 --> 2280.94] launchdarkly.com. +[2280.94 --> 2301.52] You mentioned +[2301.52 --> 2302.42] briefly about +[2302.42 --> 2303.42] your background, +[2303.60 --> 2304.10] starting with +[2304.10 --> 2305.28] C++ and also +[2305.28 --> 2306.06] on a little bit +[2306.06 --> 2306.96] on how you +[2306.96 --> 2307.64] come to the idea +[2307.64 --> 2308.16] of doing this +[2308.16 --> 2308.56] workshop. +[2308.56 --> 2308.98] Do you want +[2308.98 --> 2310.10] to elaborate +[2310.10 --> 2310.68] a little bit +[2310.68 --> 2311.66] on kind of +[2311.66 --> 2312.36] giving the +[2312.36 --> 2312.98] context of +[2312.98 --> 2314.06] what brought +[2314.06 --> 2315.62] you on the +[2315.62 --> 2316.26] shorter time +[2316.26 --> 2316.70] and on the +[2316.70 --> 2317.24] longer term +[2317.24 --> 2317.72] to this +[2317.72 --> 2318.18] idea? +[2318.56 --> 2319.22] I don't +[2319.22 --> 2319.48] know. +[2319.90 --> 2320.26] There was +[2320.26 --> 2321.42] this jobs +[2321.42 --> 2322.78] fair in +[2322.78 --> 2323.74] April where +[2323.74 --> 2324.48] I sat down +[2324.48 --> 2324.88] with the +[2324.88 --> 2325.92] organizer of +[2325.92 --> 2326.34] Graphicon +[2326.34 --> 2327.34] Europe and +[2327.34 --> 2328.02] I asked +[2328.02 --> 2329.30] her point +[2329.30 --> 2329.66] blank, +[2330.04 --> 2331.04] so why +[2331.04 --> 2331.34] have you +[2331.34 --> 2331.98] never asked +[2331.98 --> 2333.02] me to +[2333.02 --> 2333.72] do a +[2333.72 --> 2334.16] workshop? +[2334.92 --> 2335.58] You know, +[2335.66 --> 2336.34] I do this. +[2336.34 --> 2337.78] And then +[2337.78 --> 2338.30] she told +[2338.30 --> 2338.54] me, +[2338.62 --> 2338.76] well, +[2338.78 --> 2339.20] it's a lot +[2339.20 --> 2339.62] of work +[2339.62 --> 2339.96] and I +[2339.96 --> 2340.18] said, +[2340.28 --> 2340.56] yeah, +[2340.74 --> 2341.22] I mean, +[2341.26 --> 2341.64] this is +[2341.64 --> 2342.26] something that +[2342.26 --> 2342.78] I do. +[2343.80 --> 2344.12] But the +[2344.12 --> 2344.74] topic itself +[2344.74 --> 2345.26] is really +[2345.26 --> 2345.84] something that +[2345.84 --> 2346.18] I didn't +[2346.18 --> 2346.62] want to +[2346.62 --> 2347.10] teach my +[2347.10 --> 2347.32] team. +[2347.40 --> 2347.74] My team +[2347.74 --> 2348.90] works mostly +[2348.90 --> 2349.74] procedural code, +[2349.80 --> 2349.98] right? +[2350.08 --> 2350.34] Mostly +[2350.34 --> 2351.34] procedural code +[2351.34 --> 2352.06] with Go. +[2352.52 --> 2352.74] Why? +[2352.82 --> 2353.22] Why did you +[2353.22 --> 2353.62] want to do +[2353.62 --> 2353.76] it? +[2353.80 --> 2354.08] How did +[2354.08 --> 2354.38] you even +[2354.38 --> 2354.72] come to +[2354.72 --> 2355.02] think of +[2355.02 --> 2355.18] it? +[2355.28 --> 2355.58] Because +[2355.58 --> 2355.90] that's +[2355.90 --> 2356.20] what they +[2356.20 --> 2356.84] do and +[2356.84 --> 2357.54] I actually +[2357.54 --> 2358.12] don't think +[2358.12 --> 2358.52] that those +[2358.52 --> 2359.14] patterns that +[2359.14 --> 2359.66] they're using +[2359.66 --> 2360.62] serve my +[2360.62 --> 2361.08] team very +[2361.08 --> 2361.38] well. +[2361.56 --> 2361.90] Actually, +[2362.06 --> 2362.46] it's not +[2362.46 --> 2362.86] really my +[2362.86 --> 2363.06] team, +[2366.34 --> 2366.82] months ago +[2366.82 --> 2367.50] we +[2367.50 --> 2368.20] reorganized. +[2368.56 --> 2369.12] We used +[2369.12 --> 2369.62] middleware +[2369.62 --> 2370.98] patterns in +[2370.98 --> 2371.76] a way that +[2371.76 --> 2372.56] was not +[2372.56 --> 2373.22] really working +[2373.22 --> 2373.70] well for +[2373.70 --> 2373.96] us. +[2374.62 --> 2374.98] I mean, +[2375.02 --> 2375.30] I'm not +[2375.30 --> 2375.64] saying that +[2375.64 --> 2375.96] there is +[2375.96 --> 2376.58] no way to +[2376.58 --> 2377.10] make it +[2377.10 --> 2377.82] not work. +[2378.20 --> 2378.72] There is +[2378.72 --> 2379.28] no way to +[2379.28 --> 2379.60] make it +[2379.60 --> 2379.86] work, +[2380.24 --> 2381.04] but the +[2381.04 --> 2381.40] way that +[2381.40 --> 2381.76] we did +[2381.76 --> 2382.22] this was +[2382.22 --> 2382.70] not really +[2382.70 --> 2383.28] working well +[2383.28 --> 2383.72] for us. +[2384.40 --> 2385.34] And I +[2385.34 --> 2385.86] was trying +[2385.86 --> 2386.32] to show +[2386.32 --> 2386.66] the team +[2386.66 --> 2387.14] different +[2387.14 --> 2387.68] approaches +[2387.68 --> 2388.64] to maybe +[2388.64 --> 2389.74] redesign +[2389.74 --> 2391.32] really a +[2391.32 --> 2391.98] portion of +[2391.98 --> 2392.48] the code +[2392.48 --> 2393.44] that I +[2393.44 --> 2394.08] felt would +[2394.08 --> 2394.88] just be +[2394.88 --> 2395.84] easier to +[2395.84 --> 2396.60] understand +[2396.60 --> 2397.46] with classic +[2397.46 --> 2398.16] object oriented +[2398.16 --> 2399.30] because that +[2399.30 --> 2400.06] is how we +[2400.06 --> 2400.62] used to see +[2400.62 --> 2401.10] the world. +[2401.56 --> 2402.18] We understand +[2402.18 --> 2402.78] the legation. +[2402.92 --> 2403.30] For instance, +[2403.38 --> 2404.12] if I tell +[2404.12 --> 2404.94] you not only +[2404.94 --> 2405.58] to breathe +[2405.58 --> 2405.86] in, +[2405.96 --> 2406.30] you don't +[2406.30 --> 2406.68] have to +[2406.68 --> 2408.42] think about +[2408.42 --> 2409.62] operating your +[2409.62 --> 2409.88] lungs. +[2409.98 --> 2410.24] You just +[2410.24 --> 2410.74] breathe in. +[2411.66 --> 2412.58] And if I +[2412.58 --> 2413.06] tell you to +[2413.06 --> 2413.74] breathe deeply, +[2413.96 --> 2414.32] then you +[2414.32 --> 2414.92] will breathe +[2414.92 --> 2415.30] deeply. +[2415.40 --> 2415.72] And if I +[2415.72 --> 2416.16] tell you to +[2416.16 --> 2416.88] stop breathing, +[2417.18 --> 2417.60] you will +[2417.60 --> 2418.02] until you +[2418.02 --> 2418.40] panic. +[2420.58 --> 2420.98] Defer. +[2421.32 --> 2421.82] Breathe out. +[2424.74 --> 2425.34] You will +[2425.34 --> 2425.78] the first +[2425.78 --> 2426.20] sun breath, +[2426.38 --> 2426.76] hopefully. +[2428.40 --> 2428.92] Unless +[2428.92 --> 2429.52] something is +[2429.52 --> 2429.94] broken. +[2432.22 --> 2432.98] I do +[2432.98 --> 2433.70] feel that +[2433.70 --> 2434.44] when the +[2434.44 --> 2435.04] code gets +[2435.04 --> 2436.24] too complex +[2436.24 --> 2436.72] and you +[2436.72 --> 2437.38] don't understand +[2437.38 --> 2438.04] it anymore, +[2438.26 --> 2438.92] you really, +[2439.18 --> 2439.66] really need to +[2439.66 --> 2440.44] start thinking +[2440.44 --> 2442.16] in terms +[2442.16 --> 2443.66] of objects. +[2443.92 --> 2444.72] It makes +[2444.72 --> 2445.68] things so +[2445.68 --> 2446.36] much easier. +[2446.56 --> 2447.10] People do +[2447.10 --> 2447.80] understand +[2447.80 --> 2448.98] those concepts +[2448.98 --> 2449.66] very well +[2449.66 --> 2450.46] because they +[2450.46 --> 2451.72] mimic the +[2451.72 --> 2452.20] way that we +[2452.20 --> 2452.62] think, the +[2452.62 --> 2453.06] way that we +[2453.06 --> 2453.54] work, the +[2453.54 --> 2453.88] way the +[2453.88 --> 2454.66] world works. +[2455.62 --> 2456.00] So that's +[2456.00 --> 2456.28] where it +[2456.28 --> 2456.62] came from. +[2456.66 --> 2456.90] It came +[2456.90 --> 2457.22] from a +[2457.22 --> 2457.82] real problem +[2457.82 --> 2458.36] in our +[2458.36 --> 2458.96] code base. +[2459.32 --> 2459.78] And then +[2459.78 --> 2460.70] kind of +[2460.70 --> 2461.62] realized that +[2461.62 --> 2462.30] if people +[2462.30 --> 2462.94] don't agree +[2462.94 --> 2463.36] that it's +[2463.36 --> 2463.74] an object +[2463.74 --> 2464.08] oriented +[2464.08 --> 2464.66] language, +[2464.78 --> 2465.64] then one, +[2466.02 --> 2466.48] I want +[2466.48 --> 2466.84] them to +[2466.84 --> 2467.38] see why +[2467.38 --> 2467.70] I think +[2467.70 --> 2468.16] it is +[2468.16 --> 2469.32] and at +[2469.32 --> 2469.64] least give +[2469.64 --> 2469.94] them the +[2469.94 --> 2470.28] option. +[2470.92 --> 2471.40] And also, +[2471.54 --> 2472.02] I feel like +[2472.02 --> 2472.42] we failed +[2472.42 --> 2472.76] them in +[2472.76 --> 2473.12] some way. +[2473.24 --> 2473.82] If people +[2473.82 --> 2474.46] work with +[2474.46 --> 2474.90] Go and +[2474.90 --> 2475.46] they don't +[2475.46 --> 2476.26] see it +[2476.26 --> 2476.96] and they +[2476.96 --> 2477.50] don't see +[2477.50 --> 2478.14] the benefits +[2478.14 --> 2478.70] at all +[2478.70 --> 2479.80] and they +[2479.80 --> 2480.30] don't think +[2480.30 --> 2480.70] it even +[2480.70 --> 2481.26] exists, +[2481.26 --> 2482.52] it's a +[2482.52 --> 2483.26] massive part +[2483.26 --> 2483.54] of the +[2483.54 --> 2484.56] language that +[2484.56 --> 2485.00] they're not +[2485.00 --> 2485.54] utilizing. +[2486.54 --> 2487.20] Go is very +[2487.20 --> 2488.16] simple, like +[2488.16 --> 2488.78] in the sense +[2488.78 --> 2489.20] that, for +[2489.20 --> 2490.18] instance, if +[2490.18 --> 2490.96] we compare it +[2490.96 --> 2491.62] to my first +[2491.62 --> 2492.22] language, it's +[2492.22 --> 2492.76] not really my +[2492.76 --> 2493.30] first language, +[2493.40 --> 2494.18] let's say my +[2494.18 --> 2495.36] somewhat first +[2495.36 --> 2495.80] language. +[2496.12 --> 2496.58] My first +[2496.58 --> 2497.10] language is +[2497.10 --> 2497.98] Pascal, but +[2497.98 --> 2499.00] if you point a +[2499.00 --> 2499.38] gun to my +[2499.38 --> 2499.78] head, I won't +[2499.78 --> 2500.36] be able to. +[2500.52 --> 2501.08] Same here. +[2501.54 --> 2501.80] Yeah. +[2502.46 --> 2502.86] They're the +[2502.86 --> 2503.62] products of the +[2503.62 --> 2505.04] same education +[2505.04 --> 2505.88] system. +[2506.28 --> 2506.90] Some people +[2506.90 --> 2508.48] had Java, +[2508.48 --> 2510.26] which is +[2510.26 --> 2510.76] not really +[2510.76 --> 2511.90] the uncle +[2511.90 --> 2513.88] of Go, +[2514.08 --> 2514.82] unlike Pascal, +[2514.94 --> 2515.36] which is. +[2515.64 --> 2516.20] Yeah, that's +[2516.20 --> 2516.46] true. +[2516.74 --> 2517.22] It's actually +[2517.22 --> 2517.60] true. +[2517.90 --> 2518.28] Although I +[2518.28 --> 2518.68] did start +[2518.68 --> 2519.16] with doctor +[2519.16 --> 2519.76] scheme to +[2519.76 --> 2520.22] be very, +[2520.36 --> 2521.02] very accurate. +[2521.58 --> 2522.14] Well, I +[2522.14 --> 2523.10] mean, the +[2523.10 --> 2523.64] only reason +[2523.64 --> 2524.10] that I know +[2524.10 --> 2524.38] that is +[2524.38 --> 2524.86] because of +[2524.86 --> 2525.48] Carmenando, +[2525.72 --> 2526.36] who went +[2526.36 --> 2526.94] around telling +[2526.94 --> 2527.90] everybody that +[2527.90 --> 2529.34] Go is like +[2529.34 --> 2530.46] a child of +[2530.46 --> 2531.16] the branches +[2531.16 --> 2531.90] of Pascal. +[2532.18 --> 2532.34] Yeah. +[2532.98 --> 2533.58] And then I +[2533.58 --> 2534.12] learned also +[2534.12 --> 2534.90] that Ruby is +[2534.90 --> 2535.96] also somewhat +[2535.96 --> 2536.60] a child of +[2536.60 --> 2537.10] Pascal, which +[2537.10 --> 2537.70] explains why I +[2537.70 --> 2538.14] like Ruby. +[2538.14 --> 2538.76] Although I +[2538.76 --> 2539.28] don't remember +[2539.28 --> 2540.12] Pascal at all +[2540.12 --> 2540.80] and I can't +[2540.80 --> 2541.50] really feel like +[2541.50 --> 2541.88] there are +[2541.88 --> 2542.66] similarities or +[2542.66 --> 2542.94] not. +[2543.06 --> 2543.62] I have no +[2543.62 --> 2543.90] idea. +[2544.50 --> 2545.30] Try to take +[2545.30 --> 2545.66] it with a +[2545.66 --> 2546.00] piece of +[2546.00 --> 2546.28] paper. +[2546.46 --> 2546.76] Maybe it +[2546.76 --> 2547.20] will refresh +[2547.20 --> 2547.58] your memory +[2547.58 --> 2548.02] more than +[2548.02 --> 2548.28] with a +[2548.28 --> 2548.56] screen. +[2550.04 --> 2550.54] That's how +[2550.54 --> 2551.02] it was for +[2551.02 --> 2551.34] me, at +[2551.34 --> 2551.72] least, writing +[2551.72 --> 2552.42] Pascal code. +[2552.76 --> 2553.04] I thought you +[2553.04 --> 2553.58] ever actually +[2553.58 --> 2554.48] executed any. +[2554.82 --> 2555.06] Did you +[2555.06 --> 2555.26] get to +[2555.26 --> 2555.62] write it? +[2556.08 --> 2556.46] I don't +[2556.46 --> 2556.86] remember. +[2557.34 --> 2557.96] For better +[2557.96 --> 2558.36] and worse, +[2558.58 --> 2559.24] education +[2559.24 --> 2559.70] system. +[2560.56 --> 2561.10] It's the +[2561.10 --> 2561.76] whiteboard of +[2561.76 --> 2562.52] the early +[2562.52 --> 2562.92] times. +[2563.54 --> 2564.32] Anyway, yes, +[2564.46 --> 2564.76] it's like +[2564.76 --> 2565.30] uncle of +[2565.30 --> 2565.54] Ruby. +[2565.62 --> 2565.94] I did not +[2565.94 --> 2566.40] know that. +[2566.62 --> 2567.18] Yeah, things +[2567.18 --> 2567.66] that you +[2567.66 --> 2568.90] learn when +[2568.90 --> 2569.24] you go to +[2569.24 --> 2569.68] conferences. +[2571.12 --> 2571.56] Random +[2571.56 --> 2572.74] bits of +[2572.74 --> 2573.48] information, +[2573.90 --> 2574.64] of trivia +[2574.64 --> 2575.22] that could +[2575.22 --> 2575.66] be useful +[2575.66 --> 2575.92] in the +[2575.92 --> 2576.36] future or +[2576.36 --> 2576.54] not. +[2577.26 --> 2578.10] So, back +[2578.10 --> 2578.58] to C++, +[2578.74 --> 2579.26] it has a +[2579.26 --> 2579.98] million features +[2579.98 --> 2580.36] that you're +[2580.36 --> 2580.70] never going +[2580.70 --> 2581.06] to use. +[2581.56 --> 2582.24] JavaScript as +[2582.24 --> 2582.78] well, nobody +[2582.78 --> 2584.20] uses, let's +[2584.20 --> 2584.88] say, what it +[2584.88 --> 2585.08] is. +[2585.12 --> 2585.58] It's not a +[2585.58 --> 2586.28] common practice. +[2586.36 --> 2586.88] In that case, +[2586.90 --> 2587.46] it's the best +[2587.46 --> 2588.38] practice not to +[2588.38 --> 2589.84] utilize all the +[2589.84 --> 2590.90] features in +[2590.90 --> 2591.58] those languages +[2591.58 --> 2592.54] because that +[2592.54 --> 2594.52] is not +[2594.52 --> 2595.50] maintainable code. +[2596.18 --> 2596.56] So, +[2596.78 --> 2597.32] letting go, +[2597.32 --> 2598.08] Go has +[2598.08 --> 2598.52] been written +[2598.52 --> 2599.08] in a way +[2599.08 --> 2599.60] that should +[2599.60 --> 2600.18] allow us +[2600.18 --> 2601.00] to utilize +[2601.00 --> 2601.80] all the +[2601.80 --> 2602.20] features. +[2603.14 --> 2603.76] So, the +[2603.76 --> 2604.32] idea that +[2604.32 --> 2604.96] people don't +[2604.96 --> 2606.38] do it is, +[2606.94 --> 2607.26] well, she +[2607.26 --> 2608.06] said, I +[2608.06 --> 2608.42] don't know +[2608.42 --> 2608.72] what to +[2608.72 --> 2610.30] say else +[2610.30 --> 2610.82] about it. +[2610.82 --> 2611.42] What are +[2611.42 --> 2611.88] some good +[2611.88 --> 2612.94] or bad +[2612.94 --> 2613.80] use cases +[2613.80 --> 2614.64] for using +[2614.64 --> 2615.00] that? +[2615.34 --> 2615.86] Like, you +[2615.86 --> 2616.26] mentioned, +[2616.52 --> 2617.02] billing is +[2617.02 --> 2617.80] your close +[2617.80 --> 2618.10] to home +[2618.10 --> 2618.60] example. +[2619.06 --> 2619.24] Do you +[2619.24 --> 2619.66] have some +[2619.66 --> 2621.50] other use +[2621.50 --> 2621.94] cases you +[2621.94 --> 2622.18] would say +[2622.18 --> 2622.90] definitely use +[2622.90 --> 2623.12] it or +[2623.12 --> 2623.48] definitely +[2623.48 --> 2623.94] don't? +[2623.94 --> 2625.24] So, the +[2625.24 --> 2625.84] quote that +[2625.84 --> 2627.18] Rob Pike, +[2627.50 --> 2627.80] that I +[2627.80 --> 2628.08] think +[2628.08 --> 2628.70] whatever, +[2629.00 --> 2629.34] whomever +[2629.34 --> 2630.10] sent me +[2630.10 --> 2631.08] the message +[2631.08 --> 2631.80] was basing +[2631.80 --> 2632.34] it on, +[2632.58 --> 2633.08] he said +[2633.08 --> 2633.60] that, I +[2633.60 --> 2633.88] don't know, +[2633.96 --> 2634.40] there were +[2634.40 --> 2635.26] multiple, +[2635.76 --> 2636.56] a professor +[2636.56 --> 2637.30] used multiple +[2637.30 --> 2638.18] classes to +[2638.18 --> 2638.82] perform something +[2638.82 --> 2639.24] that was a +[2639.24 --> 2639.96] simple lookup. +[2640.50 --> 2641.26] And I think +[2641.26 --> 2641.90] this is it. +[2641.98 --> 2642.50] And I also +[2642.50 --> 2643.36] understand why +[2643.36 --> 2643.96] it happened. +[2644.84 --> 2645.30] But again, +[2645.38 --> 2645.92] it's not +[2645.92 --> 2646.24] Go, +[2646.40 --> 2646.86] actually. +[2647.68 --> 2648.70] So, again, +[2648.76 --> 2649.22] when we go +[2649.22 --> 2649.98] back to +[2649.98 --> 2651.52] Java, +[2652.34 --> 2652.82] you are not +[2652.82 --> 2653.56] able to +[2653.56 --> 2654.30] say that +[2654.30 --> 2655.14] A is B +[2655.14 --> 2655.88] unless A is +[2655.88 --> 2656.54] aware of B +[2656.54 --> 2657.42] and knows +[2657.42 --> 2657.66] that it +[2657.66 --> 2658.32] implements B. +[2658.82 --> 2659.22] So, you +[2659.22 --> 2659.78] can't really +[2659.78 --> 2660.36] say that A +[2660.36 --> 2660.76] is B. +[2661.36 --> 2661.82] And then in +[2661.82 --> 2662.32] those languages, +[2662.32 --> 2662.74] you really +[2662.74 --> 2663.54] have to work +[2663.54 --> 2664.98] extra hard +[2664.98 --> 2666.70] to express +[2666.70 --> 2667.24] the idea +[2667.24 --> 2667.66] that A +[2667.66 --> 2668.00] is B. +[2668.66 --> 2669.08] And that +[2669.08 --> 2669.70] could create +[2669.70 --> 2670.18] those, you +[2670.18 --> 2670.88] know, those +[2670.88 --> 2672.02] intermediate layers +[2672.02 --> 2673.08] between code +[2673.08 --> 2673.64] that is just +[2673.64 --> 2674.56] a proxy, +[2674.80 --> 2675.16] which just +[2675.16 --> 2675.78] invokes more +[2675.78 --> 2676.14] code and +[2676.14 --> 2676.66] invokes more +[2676.66 --> 2676.98] code. +[2677.40 --> 2678.02] And then at +[2678.02 --> 2678.36] the end of +[2678.36 --> 2678.66] the day, +[2678.74 --> 2679.08] if you want +[2679.08 --> 2679.70] to perform +[2679.70 --> 2680.52] a simple +[2680.52 --> 2680.92] lookup, +[2681.26 --> 2682.00] it can look +[2682.00 --> 2683.02] like something +[2683.02 --> 2683.46] that will +[2683.46 --> 2684.12] create this +[2684.12 --> 2684.98] very chaotic +[2684.98 --> 2686.02] code base +[2686.02 --> 2687.44] for something +[2687.44 --> 2687.96] very, very +[2687.96 --> 2688.36] simple. +[2689.18 --> 2689.66] But then +[2689.66 --> 2690.78] in Go, +[2691.06 --> 2691.72] I always say +[2691.72 --> 2692.12] that everything +[2692.12 --> 2692.80] is explicit +[2692.80 --> 2693.50] in Go +[2693.50 --> 2693.96] except for +[2693.96 --> 2694.32] the things +[2694.32 --> 2694.78] that aren't. +[2695.46 --> 2695.90] And what I +[2695.90 --> 2696.40] mean by the +[2696.40 --> 2696.82] things that +[2696.82 --> 2697.30] aren't are +[2697.30 --> 2698.04] like stringers, +[2698.16 --> 2698.60] for instance, +[2698.74 --> 2699.86] where we +[2699.86 --> 2700.88] invoke some +[2700.88 --> 2702.24] functionality by +[2702.24 --> 2703.26] performing some +[2703.26 --> 2704.42] type assertion +[2704.42 --> 2706.22] that nobody +[2706.22 --> 2706.90] is aware of +[2706.90 --> 2707.18] somewhere. +[2708.12 --> 2708.34] But yeah, +[2708.42 --> 2708.84] but generally +[2708.84 --> 2709.34] speaking, +[2709.44 --> 2709.80] Go is +[2709.80 --> 2710.24] explicit. +[2710.90 --> 2711.62] If you have +[2711.62 --> 2712.46] a package and +[2712.46 --> 2712.88] it's well +[2712.88 --> 2713.28] written, +[2714.00 --> 2715.16] then it +[2715.16 --> 2716.18] should meet +[2716.18 --> 2716.90] the open +[2716.90 --> 2718.12] close principles +[2718.12 --> 2719.98] and then you +[2719.98 --> 2720.98] should be able +[2720.98 --> 2722.34] to wrap this +[2722.34 --> 2722.78] type with +[2722.78 --> 2723.16] whatever, +[2723.68 --> 2723.98] you know, +[2724.02 --> 2724.94] or extend +[2724.94 --> 2725.76] the functionality +[2725.76 --> 2726.68] of whatever it +[2726.68 --> 2727.10] is that you +[2727.10 --> 2727.30] want. +[2727.86 --> 2728.46] I understand +[2728.46 --> 2728.94] why people +[2728.94 --> 2729.56] struggle with +[2729.56 --> 2729.76] that. +[2730.08 --> 2730.62] I think that +[2730.62 --> 2731.26] is the biggest +[2731.26 --> 2732.54] issue is +[2732.54 --> 2734.08] how we +[2734.08 --> 2734.54] don't write +[2734.54 --> 2735.30] packages very +[2735.30 --> 2735.58] well. +[2735.64 --> 2735.98] By the way, +[2736.06 --> 2736.38] for the +[2736.38 --> 2736.74] workshop, +[2737.04 --> 2738.08] I had to +[2738.08 --> 2739.44] revise my +[2739.44 --> 2740.10] code multiple +[2740.10 --> 2741.06] times because +[2741.06 --> 2741.76] I realized that +[2741.76 --> 2742.40] my design +[2742.40 --> 2743.02] choices were +[2743.02 --> 2744.38] so far, +[2744.84 --> 2745.30] let's say. +[2745.52 --> 2746.08] Less common +[2746.08 --> 2746.56] practices. +[2748.14 --> 2749.14] For instance, +[2749.30 --> 2749.90] it's really +[2749.90 --> 2750.32] funny. +[2751.30 --> 2752.10] So I +[2752.10 --> 2752.78] actually wanted +[2752.78 --> 2753.38] to show +[2753.38 --> 2754.28] that the +[2754.28 --> 2754.72] code was +[2754.72 --> 2755.44] extendable. +[2755.78 --> 2756.68] So by +[2756.68 --> 2757.66] actually creating +[2757.66 --> 2758.08] an extra +[2758.08 --> 2759.02] package that +[2759.02 --> 2759.66] will use that +[2759.66 --> 2760.08] package. +[2760.08 --> 2760.94] and then I +[2760.94 --> 2761.58] realized that +[2761.58 --> 2761.74] no, +[2761.82 --> 2762.26] that package +[2762.26 --> 2762.62] should take +[2762.62 --> 2762.84] in an +[2762.84 --> 2763.26] interface, +[2763.38 --> 2763.64] for instance. +[2763.80 --> 2763.86] Like, +[2763.90 --> 2764.12] I mean, +[2764.50 --> 2764.84] you should +[2764.84 --> 2765.56] extend that +[2765.56 --> 2765.94] through an +[2765.94 --> 2766.40] interface, +[2766.52 --> 2767.12] not direct. +[2767.82 --> 2768.62] And as I +[2768.62 --> 2769.16] was doing +[2769.16 --> 2769.60] that, +[2769.94 --> 2770.38] the reason +[2770.38 --> 2770.94] that I +[2770.94 --> 2771.84] realized that +[2771.84 --> 2772.44] was because +[2772.44 --> 2773.00] I had a +[2773.00 --> 2773.84] third package +[2773.84 --> 2774.42] that actually +[2774.42 --> 2775.54] did need +[2775.54 --> 2776.58] that interface +[2776.58 --> 2777.54] and it needed +[2777.54 --> 2778.24] that interface +[2778.24 --> 2778.70] to be in +[2778.70 --> 2779.34] the in-between +[2779.34 --> 2779.78] layer. +[2780.34 --> 2780.80] And then I +[2780.80 --> 2781.46] started thinking, +[2781.64 --> 2782.12] so how do +[2782.12 --> 2782.58] I make, +[2782.72 --> 2782.84] like, +[2782.88 --> 2783.42] if I wanted +[2783.42 --> 2784.02] to teach +[2784.02 --> 2784.84] somebody that, +[2785.02 --> 2785.64] how would I +[2785.64 --> 2786.70] actually do it? +[2786.96 --> 2787.22] And I'm +[2787.22 --> 2787.94] still struggling +[2787.94 --> 2788.26] with, +[2788.38 --> 2788.46] like, +[2788.52 --> 2789.08] figuring out, +[2789.16 --> 2789.26] like, +[2789.28 --> 2789.84] what is the +[2789.84 --> 2791.60] exact problem +[2791.60 --> 2792.00] with that +[2792.00 --> 2792.84] code that +[2792.84 --> 2793.36] I could tell +[2793.36 --> 2793.66] somebody, +[2793.76 --> 2793.86] like, +[2793.88 --> 2794.36] if you see +[2794.36 --> 2794.70] this, +[2794.76 --> 2795.28] then that is +[2795.28 --> 2795.78] your problem +[2795.78 --> 2796.14] and that's +[2796.14 --> 2796.36] what you +[2796.36 --> 2796.66] need to +[2796.66 --> 2797.20] change. +[2797.92 --> 2798.54] But it's +[2798.54 --> 2798.98] really funny +[2798.98 --> 2799.34] because I +[2799.34 --> 2799.84] was actually +[2799.84 --> 2800.66] writing something +[2800.66 --> 2801.66] to show +[2801.66 --> 2802.58] that it's +[2802.58 --> 2803.10] going to be +[2803.10 --> 2803.64] extendable +[2803.64 --> 2804.10] and then it +[2804.10 --> 2804.70] wasn't. +[2804.78 --> 2805.16] And then at +[2805.16 --> 2805.58] some point, +[2805.64 --> 2805.76] like, +[2805.82 --> 2807.08] I hit an end. +[2807.40 --> 2807.96] So obviously, +[2808.08 --> 2808.26] you know, +[2808.32 --> 2809.10] I fixed it. +[2809.66 --> 2810.96] But generally, +[2811.46 --> 2812.70] it's a different +[2812.70 --> 2813.60] type of language. +[2813.60 --> 2814.70] So a package +[2814.70 --> 2815.50] doesn't have +[2815.50 --> 2816.44] to expose +[2816.44 --> 2817.26] the interfaces +[2817.26 --> 2817.88] that, +[2818.16 --> 2818.34] like, +[2818.38 --> 2818.94] an interface. +[2819.10 --> 2819.24] Like, +[2819.30 --> 2819.90] if you write +[2819.90 --> 2821.22] code in Java, +[2821.32 --> 2821.62] if you write +[2821.62 --> 2822.44] code in C++, +[2822.98 --> 2823.80] you don't want +[2823.80 --> 2824.48] to express it +[2824.48 --> 2824.80] easy. +[2825.46 --> 2826.26] But at the time +[2826.26 --> 2826.86] that you write +[2826.86 --> 2827.36] the package +[2827.36 --> 2827.92] or you write +[2827.92 --> 2828.40] the class +[2828.40 --> 2828.86] or you write +[2828.86 --> 2829.24] anything, +[2829.24 --> 2830.26] you have to +[2830.26 --> 2831.58] know how +[2831.58 --> 2832.28] the user is +[2832.28 --> 2832.90] going to use +[2832.90 --> 2833.08] it. +[2833.52 --> 2834.28] You have to +[2834.28 --> 2834.56] know. +[2834.88 --> 2835.46] And then you +[2835.46 --> 2836.26] have to +[2836.26 --> 2836.92] extend +[2836.92 --> 2838.40] or implement +[2838.40 --> 2839.10] or, +[2839.18 --> 2839.48] you know, +[2840.00 --> 2841.42] do something +[2841.42 --> 2842.92] so that +[2842.92 --> 2843.54] a user +[2843.54 --> 2844.18] who is +[2844.18 --> 2844.74] using your +[2844.74 --> 2845.10] package +[2845.10 --> 2845.48] is going +[2845.48 --> 2846.10] to be able +[2846.10 --> 2846.72] to plug +[2846.72 --> 2847.56] in your +[2847.56 --> 2848.50] A as B, +[2848.78 --> 2850.04] your A as I, +[2850.16 --> 2850.68] your A as +[2850.68 --> 2850.96] whatever. +[2851.86 --> 2852.46] So you have +[2852.46 --> 2853.36] to think +[2853.36 --> 2854.04] about how +[2854.04 --> 2854.58] people are +[2854.58 --> 2855.34] going to +[2855.34 --> 2856.70] use your +[2856.70 --> 2857.16] class. +[2857.62 --> 2858.26] And then in +[2858.26 --> 2858.42] Go, +[2858.48 --> 2858.78] you don't +[2858.78 --> 2859.36] have to do +[2859.36 --> 2859.54] it, +[2859.66 --> 2859.98] but then +[2859.98 --> 2860.30] you can +[2860.30 --> 2860.90] write very +[2860.90 --> 2861.34] easily +[2861.34 --> 2862.08] a usable +[2862.08 --> 2862.62] package +[2862.62 --> 2863.14] or +[2863.14 --> 2863.38] an +[2863.38 --> 2863.88] extendable +[2863.88 --> 2864.46] package. +[2865.28 --> 2865.60] So how +[2865.60 --> 2866.00] to write +[2866.00 --> 2866.46] it well, +[2866.56 --> 2866.94] by the way, +[2867.02 --> 2867.28] is something +[2867.28 --> 2867.62] that I'm +[2867.62 --> 2867.98] still, +[2868.12 --> 2868.56] I still +[2868.56 --> 2869.20] don't know +[2869.20 --> 2869.82] that we +[2869.82 --> 2870.24] are very +[2870.24 --> 2870.62] good at +[2870.62 --> 2871.18] teaching. +[2871.42 --> 2871.94] I'm +[2871.94 --> 2872.44] trying to +[2872.44 --> 2872.82] sort of +[2872.82 --> 2873.34] figure that +[2873.34 --> 2873.78] out as +[2873.78 --> 2874.38] well as +[2874.38 --> 2874.76] I go. +[2875.22 --> 2875.68] How do +[2875.68 --> 2876.40] I define +[2876.40 --> 2877.42] that a +[2877.42 --> 2878.46] package is +[2878.46 --> 2878.90] good? +[2879.28 --> 2879.66] And I +[2879.66 --> 2880.54] love when +[2880.54 --> 2881.32] people explain +[2881.32 --> 2881.72] the open +[2881.72 --> 2882.12] clause, +[2882.22 --> 2882.70] they always +[2882.70 --> 2883.82] talk about +[2883.82 --> 2885.08] CRL or +[2885.08 --> 2885.56] curl or +[2885.56 --> 2886.08] whatever the +[2886.08 --> 2886.50] command. +[2886.84 --> 2887.24] Nobody's +[2887.24 --> 2887.60] going to +[2887.60 --> 2888.66] rewrite curl. +[2889.22 --> 2889.62] It doesn't +[2889.62 --> 2890.92] require redesigning. +[2891.12 --> 2891.56] Why? +[2891.72 --> 2892.24] Why is it +[2892.24 --> 2892.82] so good? +[2892.90 --> 2893.44] What makes +[2893.44 --> 2894.52] it so good? +[2894.90 --> 2895.54] And what +[2895.54 --> 2895.94] makes a +[2895.94 --> 2896.76] good package? +[2897.14 --> 2897.64] It's a +[2897.64 --> 2897.90] very, +[2898.06 --> 2898.64] very difficult +[2898.64 --> 2899.34] question. +[2900.12 --> 2900.62] Regardless, +[2900.84 --> 2901.24] by the way, +[2901.24 --> 2902.50] if you have +[2902.50 --> 2903.14] object-oriented +[2903.14 --> 2903.54] or not, +[2903.62 --> 2904.22] even if it's +[2904.22 --> 2905.08] just a bunch +[2905.08 --> 2905.70] of functions, +[2906.22 --> 2906.74] it is very +[2906.74 --> 2907.36] difficult to +[2907.36 --> 2908.66] define when +[2908.66 --> 2909.22] you're done, +[2909.36 --> 2910.16] to know exactly +[2910.16 --> 2910.52] when you're +[2910.52 --> 2910.76] done. +[2911.14 --> 2912.36] One last +[2912.36 --> 2913.26] question before +[2913.26 --> 2914.22] we switch to +[2914.22 --> 2914.98] the fun and +[2914.98 --> 2915.78] popular opinion, +[2916.08 --> 2917.24] which we were +[2917.24 --> 2918.06] missing so much +[2918.06 --> 2918.54] throughout this +[2918.54 --> 2919.22] entire episode. +[2919.66 --> 2920.34] You've been +[2920.34 --> 2920.98] kind of giving +[2920.98 --> 2921.88] sprinkles of +[2921.88 --> 2922.72] information on +[2922.72 --> 2923.60] the workshop. +[2924.04 --> 2924.94] So other than +[2924.94 --> 2925.66] this as being +[2925.66 --> 2926.56] an object-oriented +[2926.56 --> 2927.60] programming in Go, +[2927.94 --> 2928.68] what else can you +[2928.68 --> 2929.22] tell us about +[2929.22 --> 2929.38] it? +[2930.12 --> 2930.56] Well, +[2930.68 --> 2931.34] I can tell +[2931.34 --> 2931.82] you that I +[2931.82 --> 2932.46] will take the +[2932.46 --> 2933.38] learners through +[2933.38 --> 2935.28] a maze of +[2935.28 --> 2937.26] object-oriented. +[2937.94 --> 2938.66] And when I +[2938.66 --> 2939.66] say a maze, +[2939.80 --> 2940.80] I mean quite +[2940.80 --> 2941.24] literally. +[2942.12 --> 2943.12] We are going +[2943.12 --> 2943.66] to navigate +[2943.66 --> 2944.68] through a maze. +[2945.18 --> 2945.88] I mentioned +[2945.88 --> 2946.72] Jeff Rosenstein +[2946.72 --> 2947.70] earlier, +[2948.00 --> 2948.98] who is my +[2948.98 --> 2950.02] professor to +[2950.02 --> 2951.00] intro to CS. +[2951.00 --> 2953.10] and our +[2953.10 --> 2954.66] first exercise +[2954.66 --> 2955.96] way back +[2955.96 --> 2957.34] when in +[2957.34 --> 2959.00] 2003, +[2959.54 --> 2960.40] that's what +[2960.40 --> 2960.94] I'm taking +[2960.94 --> 2961.74] people through. +[2962.08 --> 2962.52] So it's +[2962.52 --> 2963.86] heavily inspired +[2963.86 --> 2964.94] by his work. +[2965.38 --> 2965.96] So credit +[2965.96 --> 2966.44] to him. +[2967.30 --> 2968.26] I thought it +[2968.26 --> 2968.94] was a very +[2968.94 --> 2970.56] good way of +[2970.56 --> 2972.30] exploring object-oriented. +[2972.70 --> 2973.60] And I like to +[2973.60 --> 2974.38] take people through +[2974.38 --> 2975.42] journeys that I +[2975.42 --> 2976.28] found very good +[2976.28 --> 2976.94] for myself. +[2977.08 --> 2977.72] Like if I had +[2977.72 --> 2978.54] a moment, +[2978.80 --> 2979.46] I try to share +[2979.46 --> 2979.92] it with other +[2979.92 --> 2980.70] people as well. +[2981.00 --> 2982.00] And Natalie, +[2982.18 --> 2982.66] you know this. +[2982.80 --> 2983.36] Ian, actually, +[2983.42 --> 2984.14] I can ask you, +[2984.48 --> 2985.14] what is your +[2985.14 --> 2986.26] aha moment in +[2986.26 --> 2986.84] your career? +[2987.08 --> 2988.30] Like situations +[2988.30 --> 2989.00] when you said, +[2989.12 --> 2990.20] wow, and now +[2990.20 --> 2990.70] I get it. +[2990.80 --> 2991.42] Now I know. +[2991.90 --> 2992.44] That's a real +[2992.44 --> 2993.22] on-the-spot +[2993.22 --> 2993.62] question. +[2993.68 --> 2994.08] I know. +[2994.18 --> 2994.74] It's a difficult +[2994.74 --> 2995.14] one. +[2996.24 --> 2996.94] When did you +[2996.94 --> 2997.66] realize that you +[2997.66 --> 2998.38] can do this? +[2998.50 --> 2999.22] You can code. +[2999.46 --> 3000.10] You know what +[3000.10 --> 3000.48] to do. +[3000.72 --> 3001.62] You've got this. +[3001.98 --> 3002.82] I mean, that +[3002.82 --> 3004.42] moment was, you +[3004.42 --> 3004.84] know, so I was +[3004.84 --> 3005.62] in school and I +[3005.62 --> 3006.22] went to interview +[3006.22 --> 3006.98] for an internship, +[3007.54 --> 3007.94] you know, and +[3007.94 --> 3009.28] going, doing +[3009.28 --> 3009.80] that interview. +[3009.94 --> 3010.48] And they did +[3010.48 --> 3010.90] like the whole +[3010.90 --> 3011.62] whiteboard thing, +[3011.66 --> 3012.08] you know, it was +[3012.08 --> 3013.00] eight hours of +[3013.00 --> 3013.46] interviews. +[3013.88 --> 3014.62] Leaving that was +[3014.62 --> 3015.24] that like aha +[3015.24 --> 3015.96] moment was like, +[3016.04 --> 3016.70] I felt good at +[3016.70 --> 3017.12] the end of that +[3017.12 --> 3017.56] and that made +[3017.56 --> 3018.86] me feel like I +[3018.86 --> 3019.26] was like, oh +[3019.26 --> 3019.78] yeah, I can do +[3019.78 --> 3019.94] this. +[3020.00 --> 3020.52] I can do this, +[3020.58 --> 3020.78] you know. +[3021.16 --> 3021.92] So for you. +[3022.18 --> 3023.04] But before that, +[3023.08 --> 3023.72] I had no +[3023.72 --> 3024.26] confidence. +[3024.94 --> 3025.52] That's the thing. +[3025.52 --> 3026.38] So for you, it was +[3026.38 --> 3027.24] trial by fire. +[3027.62 --> 3027.94] Exactly. +[3028.12 --> 3028.28] Yeah. +[3028.46 --> 3028.90] That's very +[3028.90 --> 3029.28] interesting. +[3029.44 --> 3030.14] I think for a lot +[3030.14 --> 3031.10] of people, it is +[3031.10 --> 3032.58] just like succeeding +[3032.58 --> 3033.50] in something that I +[3033.50 --> 3034.14] didn't think they +[3034.14 --> 3034.74] could before. +[3035.20 --> 3035.90] I didn't think I +[3035.90 --> 3036.66] was ready for like +[3036.66 --> 3037.78] that kind of set of +[3037.78 --> 3038.52] interviews, you know, +[3038.54 --> 3040.80] but that's a reason +[3040.80 --> 3041.44] why I tell people +[3041.44 --> 3042.58] just you see a lot +[3042.58 --> 3043.22] of those posts where +[3043.22 --> 3044.40] it's like, oh, you +[3044.40 --> 3044.98] know, I'm trying to +[3044.98 --> 3045.94] get do these things +[3045.94 --> 3046.98] before I apply for +[3046.98 --> 3047.78] jobs in an interview. +[3047.78 --> 3048.60] And my advice is +[3048.60 --> 3049.30] always just like, just +[3049.30 --> 3049.72] go interview. +[3049.86 --> 3050.82] Like if you do +[3050.82 --> 3052.20] poorly, like you've +[3052.20 --> 3052.98] learned something and +[3052.98 --> 3054.00] you can go do the +[3054.00 --> 3054.46] next one a little +[3054.46 --> 3054.82] bit better. +[3054.98 --> 3056.12] Like, don't wait, +[3056.24 --> 3056.66] just go. +[3056.94 --> 3057.92] I think Natalie, the +[3057.92 --> 3058.86] next time we need to +[3058.86 --> 3059.60] bring somebody who +[3059.60 --> 3060.30] doesn't agree. +[3062.92 --> 3063.86] Because we are all +[3063.86 --> 3064.90] so on the same page. +[3064.96 --> 3065.54] It's a problem. +[3066.20 --> 3066.56] 100%. +[3066.56 --> 3067.50] Yeah. +[3067.94 --> 3068.64] Yeah, we'll pull +[3068.64 --> 3069.30] some from your +[3069.30 --> 3069.70] emails. +[3070.36 --> 3071.60] A thought that came +[3071.60 --> 3073.08] to mind throughout +[3073.08 --> 3073.82] this episode. +[3074.68 --> 3075.64] See if I can make +[3075.64 --> 3076.98] this into sort of an +[3076.98 --> 3077.96] unpopular opinion. +[3078.64 --> 3079.24] Probably not. +[3079.32 --> 3079.84] So I'm not going to +[3079.84 --> 3080.56] make the tune just +[3080.56 --> 3080.84] yet. +[3081.40 --> 3082.66] But thinking of an +[3082.66 --> 3083.98] AI-generated tool, +[3084.00 --> 3085.82] they get inspired +[3085.82 --> 3086.82] from existing code, +[3086.88 --> 3087.08] right? +[3087.48 --> 3089.78] And we can have +[3089.78 --> 3091.04] this conversation if +[3091.04 --> 3091.94] there's features of +[3091.94 --> 3094.72] Go, of OOP in +[3094.72 --> 3095.62] Go or not. +[3095.86 --> 3096.80] But if there's not a +[3096.80 --> 3097.82] lot of examples out +[3097.82 --> 3099.96] there, the different +[3099.96 --> 3101.98] models that generate +[3101.98 --> 3103.78] code will not be +[3103.78 --> 3105.26] creating this a lot. +[3105.56 --> 3107.34] So assuming that the +[3107.34 --> 3108.42] trend of code is not +[3108.42 --> 3109.48] just human writing it, +[3109.52 --> 3110.44] but more like human +[3110.44 --> 3112.64] guiding it, it means +[3112.64 --> 3113.44] means that it will, +[3113.90 --> 3114.46] at least the way I +[3114.46 --> 3115.22] see it, it means +[3115.22 --> 3116.40] that a lot of the +[3116.40 --> 3117.70] code or the +[3117.70 --> 3118.60] languages will kind +[3118.60 --> 3119.68] of fall deeper +[3119.68 --> 3121.74] into their template +[3121.74 --> 3122.96] or into their +[3122.96 --> 3124.86] niche or little box +[3124.86 --> 3126.14] rather than spreading +[3126.14 --> 3127.42] out of it like you +[3127.42 --> 3127.64] did. +[3127.88 --> 3129.24] So this can be a fun +[3129.24 --> 3130.46] thing to think +[3131.28 --> 3132.78] about as a person +[3132.78 --> 3133.84] who's researching AI +[3133.84 --> 3134.30] and code. +[3134.80 --> 3135.12] Wow. +[3135.12 --> 3136.00] I love it. +[3136.12 --> 3136.98] So, you know, a +[3136.98 --> 3139.24] few days ago I had +[3139.24 --> 3140.46] that thought, AI is +[3140.46 --> 3141.32] not going to write +[3141.32 --> 3142.90] code or OOP maybe +[3142.90 --> 3143.92] because they don't +[3143.92 --> 3144.74] really understand the +[3144.74 --> 3145.44] world like we do. +[3145.52 --> 3146.16] They don't have that +[3146.16 --> 3146.64] restriction. +[3147.36 --> 3148.10] You know, a bot +[3148.10 --> 3149.00] doesn't tell a person +[3149.00 --> 3149.60] to breathe. +[3149.82 --> 3150.32] You know what I mean? +[3150.98 --> 3151.88] They don't need to. +[3152.02 --> 3153.38] They can go as +[3153.38 --> 3154.96] deeply into the +[3154.96 --> 3156.30] mechanics of how to +[3156.30 --> 3157.68] breathe so that +[3157.68 --> 3158.88] their understanding is +[3158.88 --> 3160.04] going to be very +[3160.04 --> 3161.44] different of those +[3161.44 --> 3161.86] models. +[3161.86 --> 3162.74] Yeah. +[3163.12 --> 3164.22] Maybe there will be +[3164.22 --> 3164.92] some way of doing +[3164.92 --> 3166.60] this guided and then +[3166.60 --> 3167.78] we're saved from +[3167.78 --> 3168.42] those boxes. +[3168.62 --> 3169.04] Exactly. +[3169.32 --> 3170.24] So that is very +[3170.24 --> 3172.32] interesting is what +[3172.32 --> 3173.88] what you have there +[3173.88 --> 3175.36] because, you know, +[3175.42 --> 3176.36] like something that I +[3176.36 --> 3177.58] said earlier about how +[3177.58 --> 3178.66] how do you know that +[3178.66 --> 3179.94] a package is complete? +[3180.08 --> 3181.18] Maybe a bot can do +[3181.18 --> 3181.90] that for you. +[3182.16 --> 3183.60] Static analysis or +[3183.60 --> 3184.04] dynamic. +[3184.32 --> 3185.58] Maybe a bot can +[3185.58 --> 3187.36] analyze if something +[3187.36 --> 3188.16] is open closed. +[3188.70 --> 3188.88] Yeah. +[3189.00 --> 3189.18] Yeah. +[3189.30 --> 3190.50] That definitely is a +[3190.50 --> 3191.64] fun conversation to +[3191.64 --> 3191.96] have. +[3192.60 --> 3193.64] And if anybody wants +[3193.64 --> 3194.54] to chat about this, +[3194.64 --> 3195.46] we are on the go +[3195.46 --> 3197.14] time Slack channel +[3197.14 --> 3197.70] and the go for +[3197.70 --> 3198.02] Slack. +[3198.26 --> 3198.96] So reach out and +[3198.96 --> 3199.86] maybe this will be +[3199.86 --> 3201.18] our some future +[3201.18 --> 3201.70] episode. +[3202.06 --> 3202.70] But until then, +[3202.82 --> 3203.60] let's do the tune +[3203.60 --> 3204.54] for the unpopular +[3204.54 --> 3204.94] opinion. +[3210.94 --> 3212.24] I actually think +[3212.24 --> 3213.18] she'd probably leave. +[3216.32 --> 3217.00] Unpopular +[3217.00 --> 3218.10] opinion. +[3221.64 --> 3223.66] And now it's time +[3223.66 --> 3224.32] for the unpopular +[3224.32 --> 3224.80] opinion. +[3224.98 --> 3225.60] Rona, what do you +[3225.60 --> 3226.14] have for us? +[3226.58 --> 3227.58] I mean, I felt like +[3227.58 --> 3229.08] the entire show +[3229.08 --> 3230.50] was about, I have +[3230.50 --> 3232.06] an arsenal of, you +[3232.06 --> 3234.44] know, of things that +[3234.44 --> 3235.34] I need to convince +[3235.34 --> 3236.84] people about. +[3237.64 --> 3240.64] So I feel that we're +[3240.64 --> 3242.00] going to be much +[3242.00 --> 3243.76] stronger if we +[3243.76 --> 3245.26] collected opinions +[3245.26 --> 3246.78] about go from people +[3246.78 --> 3247.30] who are not +[3247.30 --> 3248.44] professional gophers. +[3248.44 --> 3250.38] And instead of +[3250.38 --> 3251.34] teaching them, +[3251.92 --> 3252.90] learn from them a +[3252.90 --> 3254.08] little bit, I do +[3254.08 --> 3255.36] see other languages +[3255.36 --> 3256.68] evolve, you know, +[3256.70 --> 3257.40] in many different +[3257.40 --> 3258.02] directions. +[3258.26 --> 3260.32] I think people +[3260.32 --> 3261.98] understand today +[3261.98 --> 3263.44] how to work with +[3263.44 --> 3265.36] languages in a very +[3265.36 --> 3266.50] different way than +[3266.50 --> 3268.02] what we used to do. +[3268.56 --> 3269.40] The evolution of +[3269.40 --> 3270.58] best practices, all +[3270.58 --> 3271.38] of those things, +[3271.80 --> 3272.54] it's just, it's +[3272.54 --> 3272.86] tremendous. +[3273.06 --> 3273.82] And also I think +[3273.82 --> 3275.52] that go added a lot +[3275.52 --> 3276.46] of value to other +[3276.46 --> 3277.68] languages just by +[3277.68 --> 3278.62] existing, just +[3278.62 --> 3279.56] because, you know, +[3279.84 --> 3281.42] go introduced these +[3281.42 --> 3282.14] features that we +[3282.14 --> 3282.70] discussed. +[3283.12 --> 3283.76] I think we +[3283.76 --> 3285.08] influenced the +[3285.08 --> 3286.62] industry, but I +[3286.62 --> 3287.10] think that we +[3287.10 --> 3287.78] should be open +[3287.78 --> 3288.48] also to be +[3288.48 --> 3289.36] influenced back. +[3289.48 --> 3290.34] So your unpopular +[3290.34 --> 3291.36] opinion is that we +[3291.36 --> 3292.30] go should be open +[3292.30 --> 3293.18] to be influenced +[3293.18 --> 3294.76] by non-gophers. +[3295.14 --> 3295.56] Yes. +[3296.02 --> 3296.48] All right. +[3296.60 --> 3297.28] Let's see how that +[3297.28 --> 3297.84] poll works. +[3298.22 --> 3299.06] This will be a poll +[3299.06 --> 3299.86] and then let's see +[3299.86 --> 3301.08] if that brings you +[3301.08 --> 3302.68] further into the +[3302.68 --> 3304.02] hall of fame of +[3304.02 --> 3304.46] the unpopular +[3304.46 --> 3304.96] opinions. +[3305.26 --> 3305.92] Well, I mean, +[3306.12 --> 3307.28] Matt said after he +[3307.28 --> 3307.96] put me in the +[3307.96 --> 3309.04] hall of fame, he +[3309.04 --> 3309.88] then said that it +[3309.88 --> 3310.70] doesn't exist +[3310.70 --> 3311.24] everywhere. +[3311.72 --> 3312.58] I think so that +[3312.58 --> 3314.84] the other people +[3314.84 --> 3315.62] with unpopular +[3315.62 --> 3316.72] opinions are not +[3316.72 --> 3316.94] jealous. +[3317.60 --> 3318.66] But I mean, as +[3318.66 --> 3319.22] far as I'm +[3319.22 --> 3321.22] concerned, it's +[3321.22 --> 3324.80] I won that title +[3324.80 --> 3325.56] fair and square. +[3328.06 --> 3329.10] But let's see. +[3329.18 --> 3329.56] Let's see. +[3329.66 --> 3330.60] I mean, it might +[3330.60 --> 3331.82] not be as +[3331.82 --> 3333.22] unorthodox as the +[3333.22 --> 3334.22] old one. +[3334.22 --> 3336.48] Oh, all right. +[3336.62 --> 3337.80] Well, that was +[3337.80 --> 3339.08] fun and interesting +[3339.08 --> 3340.78] and I hope this +[3340.78 --> 3341.48] will bring to the +[3341.48 --> 3342.16] workshop that will +[3342.16 --> 3342.92] generate enough +[3342.92 --> 3343.50] code that will +[3343.50 --> 3344.38] train the AI to +[3344.38 --> 3344.90] do some +[3344.90 --> 3346.28] OOPI Go. +[3347.10 --> 3347.78] Thanks, Rona. +[3347.98 --> 3348.42] Thanks, Ian. +[3348.52 --> 3348.96] Thanks, everyone +[3348.96 --> 3349.38] who joined. +[3349.70 --> 3350.26] Thank you. +[3354.70 --> 3355.60] What do you think +[3355.60 --> 3357.22] about Go and OOP? +[3357.58 --> 3358.84] Let us know in the +[3358.84 --> 3359.36] comments. +[3359.72 --> 3360.48] Links to the +[3360.48 --> 3361.48] discussion in the +[3361.48 --> 3361.94] show notes. +[3361.94 --> 3363.40] I also want to +[3363.40 --> 3364.14] share with you an +[3364.14 --> 3364.92] awesome conversation +[3364.92 --> 3365.64] we had on the +[3365.64 --> 3366.44] changelog recently. +[3366.72 --> 3367.58] Ken Konser wrote +[3367.58 --> 3369.60] up 16 lessons he +[3369.60 --> 3370.44] learned doing +[3370.44 --> 3371.46] security audits for +[3371.46 --> 3372.54] tech startups and +[3372.54 --> 3373.28] we sat down with +[3373.28 --> 3374.06] him for a thorough +[3374.06 --> 3374.88] discussion of his +[3374.88 --> 3375.26] findings. +[3375.78 --> 3376.78] Here's one moment +[3376.78 --> 3377.76] from that episode +[3377.76 --> 3379.10] where Ken shares +[3379.10 --> 3380.72] some pro tips for +[3380.72 --> 3381.72] pen testing teams. +[3382.34 --> 3383.84] And honestly, we +[3383.84 --> 3384.78] would also ask the +[3384.78 --> 3385.92] devs, we would +[3385.92 --> 3386.44] say like what +[3386.44 --> 3387.24] keeps you up, like +[3387.24 --> 3388.02] where in the code +[3388.02 --> 3388.66] keeps you up at +[3388.66 --> 3388.86] night? +[3389.12 --> 3389.68] We wouldn't treat +[3389.68 --> 3390.40] that as God's +[3390.40 --> 3391.72] truth, but you +[3391.72 --> 3392.56] know, developers +[3392.56 --> 3393.48] have a surprisingly +[3393.48 --> 3395.52] good sense, even +[3395.52 --> 3396.28] without security +[3396.28 --> 3398.24] knowledge, of what +[3398.24 --> 3399.14] parts of the code +[3399.14 --> 3400.82] are scary and +[3400.82 --> 3401.40] they're kind of +[3401.40 --> 3401.94] worried about. +[3402.32 --> 3402.76] They definitely +[3402.76 --> 3403.74] have blind spots. +[3404.34 --> 3405.02] That's definitely +[3405.02 --> 3405.76] true, but in +[3405.76 --> 3407.08] terms of like, we +[3407.08 --> 3407.42] were talking about +[3407.42 --> 3408.30] business logic, a +[3408.30 --> 3408.70] lot of times +[3408.70 --> 3409.04] they'll be like, +[3409.08 --> 3409.62] yeah, this part +[3409.62 --> 3410.54] is super gnarly, +[3410.68 --> 3411.94] like there's a ton +[3411.94 --> 3413.06] of logic here and +[3413.06 --> 3414.26] it kind of works, +[3414.38 --> 3414.98] but like it also +[3414.98 --> 3415.64] breaks a decent +[3415.64 --> 3416.90] amount and it's +[3416.90 --> 3417.30] an important +[3417.30 --> 3418.16] functionality for the +[3418.16 --> 3418.70] app, so please +[3418.70 --> 3419.22] check that out. +[3419.60 --> 3420.06] So those two +[3420.06 --> 3420.82] things really helped +[3420.82 --> 3421.26] prioritize. +[3421.72 --> 3422.56] That scary +[3422.56 --> 3423.44] intuition reminds +[3423.44 --> 3424.14] me of severance, +[3424.28 --> 3424.62] honestly. +[3424.84 --> 3425.82] It's like, well, I +[3425.82 --> 3426.70] can easily spot the +[3426.70 --> 3427.62] scary numbers here. +[3429.56 --> 3430.54] Continue listening +[3430.54 --> 3432.32] at changelog.fm +[3432.32 --> 3433.64] slash 494. +[3433.80 --> 3434.62] That's episode +[3434.62 --> 3436.26] number 494. +[3436.60 --> 3437.26] Thanks again to +[3437.26 --> 3437.82] our partners at +[3437.82 --> 3438.62] Fastly for having +[3438.62 --> 3439.68] our CDN covered, +[3439.96 --> 3440.82] to the mysterious +[3440.82 --> 3442.00] Breakmaster Cylinder for +[3442.00 --> 3442.76] keeping our beat +[3442.76 --> 3444.36] supply on swole, and +[3444.36 --> 3445.12] to you for listening. +[3445.32 --> 3446.12] We appreciate you. +[3446.34 --> 3447.50] That is all for now. +[3447.70 --> 3448.34] We'll talk to you +[3448.34 --> 3449.80] again next time on +[3449.80 --> 3450.34] GoTime. +[3450.34 --> 3454.26] GoTime. +[3454.72 --> 3455.18] Goigi. +[3455.18 --> 3456.18] GoTime. +[3461.32 --> 3462.72] GoTime. +[3462.72 --> 3465.10] Game on! diff --git a/Migrations without migraines_transcript.txt b/Migrations without migraines_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..eba5c1c938389665fd2cf80a06a6d92e6e28a4ce --- /dev/null +++ b/Migrations without migraines_transcript.txt @@ -0,0 +1,255 @@ +**Jon Calhoun:** Hello everyone, and welcome to Go Time. Today I am joined by Mike Fridman. Hey, Mike. How are you? + +**Mike Fridman:** Hello! + +**Jon Calhoun:** I am also joined by Vojtech Vitek... + +**Vojtech Vitek:** Hello there! + +**Jon Calhoun:** Vojtech, how are you? + +**Vojtech Vitek:** I'm good. + +**Jon Calhoun:** And today we're gonna be talking about schema migrations. So we're gonna be talking about tools that you can use to manage these, different processes around them, some mistakes you can make, and maybe some tips on how to avoid them hopefully... So we're gonna be diving into all of that. + +Let's just start off by talking about what are migrations. Do any of you wanna take that one? + +**Mike Fridman:** Yeah, sure. I can take a crack at it. So at a very high level, migrations - or the type of migrations that we're gonna talk about - are schema migrations. So that's taking some SQL, and then evolving your database over time. + +**Jon Calhoun:** So for anybody who started -- I guess it depends on when you started. But for somebody who started in the day where you would just pop into PHP MyAdmin and just type in some SQL and change your database - why do we need something to track migrations? Why is that an important thing? Vojtech, do you wanna take that one? + +**Vojtech Vitek:** Sure. I think about database migration as a Git for your SQL schema. You need to make sure that the schema is correct and valid in all environments, including the local host, for all the developers' machines, and then for development staging, production, whatever environment you have. You wanna make sure that your database schema is always in the same state. + +**Jon Calhoun:** That makes sense. I'm sure we've all ran into some of those issues where something seems to be working locally and it's not somewhere else, and database migrations can definitely be one of those cases where if you've done something locally and forgotten you've done it, it can have a big impact on whether it's gonna work somewhere else. + +\[04:11\] So to start here, I wanna talk a little bit about the tooling around it, because one of the questions we get asked a lot is "How should people be running these migrations? What sort of tooling should they be using? How do these tools even work?" And I think a lot of times this stems from the fact that people come from a framework like Ruby on Rails, where all of that is kind of baked into it, and they don't even think about it... But then in Go, that's not really the case, because we're not using a big framework. So what are some of the tools out there? And I guess, if you can elaborate a little bit on what they're actually doing behind the scenes. + +**Mike Fridman:** I guess maybe we should - maybe, as a suggestion - take it one step back and understand why we're doing this in the first place, before we get to a tool... So before you even jump into a tool, as a developer, you're usually developing some backend application, and you wanna make incremental changes to your database, because when your requirements come up, or you understand users better... And if you're a single developer on a single system, you can probably get away by applying migrations directly, doing an ad-hoc deployment, and then just calling it a day. But as soon as you get into a team or an organization and you have multiple developers working in parallel, there needs to be some sort of a process to coordinate how you incrementally make changes to your database over time. Then you wanna sort of decouple that from your application, and then be able to eventually roll that out. + +So some of the tools that you would use would be Golang Migrate - I think that's the most popular one, probably because of SEO... I think there's other ones such as Dbmate... Me and Vojtech maintain one called Pressly Goose. But at a very high level, the way these things work is you usually have a migrations directory, and that directory is going to contain SQL files, or Go files, depending if the tool supports it... And those files are usually gonna be incrementing integers, so like 1, 2, 3, or they might be timestamps, like year, month, date, hour, minute, seconds... So they're large numbers, but they're still incrementing over time. + +And then these tools - their only responsibility is to read those files and determine whether those migrations have been applied or not, and then keep track of that with usually a metadata or a schema history table in a database. + +**Jon Calhoun:** So when you talk about like a history table in a database, this means -- and I think a lot of people don't quite realize this... But this would mean that actually your tool is gonna create a table in their database, that keeps track of these things. So whenever it's actually running the migrations, it's gonna say "Has migration 01 run?" And if so, it's gonna skip it, otherwise it's gonna run it. Is that correct? + +**Mike Fridman:** Yeah, that's pretty much what it does... Depending on the tool. Different tools work slightly differently, but conceptually, that's what happens - you read your files, you read where you are in terms of diversion that you're currently at, and then any new migrations should be applied. + +**Jon Calhoun:** So presumably, this could lead to a case where -- I guess one way to look at it is I think I've seen people change migrations that have already run, and that can be confusing to them... And I think that's in part because they don't quite understand how they work... And I suppose this could also lead to cases where you need to undo migrations. So what would that process look like? + +**Mike Fridman:** Yeah, so usually when you write these SQL files, you'll usually create an up migration. So that'll be something like "Create table, alter table, add an index", maybe you even drop an index; that could be your up migration. And then you provide the corresponding way to roll that change back. So it'll be something like drop table, drop a column, drop an index, whatever that may be. + +So depending on the tool you use - for example, Golang Migrate, you have two separate files. So you'll have something like "01 create table up", and then you'll have "01 create table down". And when you do your up migrations, that creates the table. When you do your down migration, that drops the table. So you kind of have a way to go forwards and backwards in your migrations. + +**Jon Calhoun:** \[08:09\] So when we're looking at these migrations, another question that I've seen pop up is "Is it possible for them to run not directly in the order that's listed?" If you're sorting (I guess) alphabetically... I'm not sure what -- basically, they're sorted by the number, like the 01, 02, 03... Is it possible for these to not run in the order that they're listed in the directory? + +**Mike Fridman:** It depends on the tool. This is actually an interesting one, because the number one requested feature in Goose was being able to apply out of order migrations. Because from what I understand - and it's been like eight years since I've done any Rails work... In Rails you have time-based migrations, and you kind of just apply the migration and call it a day. You don't think about it too much. + +So depending on the tool, if you're tracking every single migration that you have historically applied, you can start to determine whether you have out-of-order migrations. And the way you get ot out-of-order migrations can be, for example, two developers checking out the repo on the same day, creating let's say a sequential number or a time-based number, and then trying to apply those. + +Depending on the tool, you'll get conflicts. If you're trying to apply migration 26 when another developer already applied 26, the tool will tell you to fail. But the tool can also detect that you have out of order migrations, and allow you to apply those migrations if you want to. Or if you don't supply the correct flag, it'll just fail and say "Hey, this is out of order. You need to either fix it or provide the flag." Golang Migrate does not do this, because it only keeps track of the latest migration that was applied. So you literally have your metadata table, you'll have one entry in there, and it'll say "Migration 14 has been applied", and that's it. But you don't know historically when the other migrations have been applied. + +**Jon Calhoun:** Okay. + +**Vojtech Vitek:** Personally, I think it's a better idea to use the sequential versions. I think there are pros and cons to both approaches. If you have the timestamp migrations, I think it's a little bit easier for the developers to work with it, because you generate the file just once, the timestamp will be generated at the time where you execute the command to create the migration file, and then you call it a day. + +But then if you have multiple people working on different feature branches at the same time, some of the feature branches might take much longer to get merged into the main branch, and eventually you will end up in a state where perhaps you push something into development, or staging environment, sooner than it would get applied to production. And this may happen with the timestamp migrations. However, I think it's just a very edge-casey thing that's gonna happen if you run into issues with this approach. + +I prefer the sequential increments for the versioning, just because I like to think about this history for the schema linearly. So every single change has to go in order, and this is ensured by the sequential numbers. + +**Jon Calhoun:** I think it's one of those things that -- like, out of order sort of made sense with Rails, where people expect things to kind of magically work. They don't wanna have to go in and merge some conflict where two people both generated Migration 24, or something... Because when you're using migrations that are expected to go in a sequential order, I suppose the issue there is that if you and I check out a branch at the same time, Vojtech, and then you check yours in sooner, when I go to check mine in, all of a sudden my migration number is gonna be wrong, and I'm gonna have to fix something at that point to actually get it to work. And I think one of the reasons people like the out of order is that they don't really have to worry about it... If they're time-based at least, you can just submit it and be like, "Okay, whatever. It runs." + +\[12:04\] But there's definitely cases that that could be an issue, or it's definitely a bad idea to have something where you've only tested it in one order on your computer, and then when you go to push it to production, it's gonna be running in a different order and you don't actually know for sure that's gonna work. + +**Vojtech Vitek:** That's exactly right. So this out of order thing - it may have some edge cases, but you're very unlikely to hit them. But once you hit them, it's gonna be a big problem, and you will have a hard time dealing with some production issue. But there's one more thing that I wanted to mention - I think the sequential numbering might have another benefit, which is if you are pushing a single change, a single deployment consisting of five different migration files, maybe one of them is adding a new table, a second one is adding a new column somewhere else, the third one is doing indexes or some data manipulation... What happens if this doesn't go well? How do you roll back to the very first migration that you wanted to push out? + +If for example the fourth one failed, with the timestamp migrations the tool would have to somehow track it in the database, so you can roll back the right ones, because you cannot really rely on the number anymore. So you need to rely on some other mechanism. With sequential numbers, it's quite easy. If you apply four, then you'll just roll back the recent four, and that's it. + +**Jon Calhoun:** Yeah, I think that's a good thing to keep in mind, because while most of the time migrations hopefully go well, the few times that they don't, especially if you're trying to deploy, or something, that's not a time when you want to be trying to figure out your tooling or how it works... I know I've had experiences - not in production, but I've had other experiences where something goes wrong with the migration and it's always frustrating to have to sit there and figure out "How do I fix this and get back to a working state again?" + +**Vojtech Vitek:** I call this "Panic mode." + +**Jon Calhoun:** It is panic mode sometimes. + +**Vojtech Vitek:** Because you cannot think straight when there are people calling you that "Production is down. What am I gonna do?" + +**Jon Calhoun:** So talking about that - when you talked about migrations that go bad, I think different tools sort of handled this differently... So I know one of the approaches is to run the entire migration in a transaction, and then other tools I guess sometimes don't... What's the difference, for people out there who -- what's the difference between running a migration, maybe what are some of the pros or cons to not doing that? + +**Vojtech Vitek:** So Goose is actually giving you the option to run a migration within the transaction, or not. Which is good. Because not all the SQL queries can be run within the transaction. By default, it is run in transaction, and it's good, because if something fails, you know that it will roll back to the previous state, and the database engine will handle the rest. You can know for sure that the state of the database will be in good shape. + +There's also one more thing, which is -- yeah, that's what I wanted to mention about Golang Migrate. It is using a database locking mechanism. So for example, when you run this, it will lock the database first, and then it will start to apply all the migrations, and then it will commit the transaction, and then it will unlock the database again. + +So that's one of the differences between Golang Migrate and Pressly Goose. Pressly Goose, on the other hand, doesn't use any locking mechanism, because it supports lots of different database drivers, and not all of them can do locking... And it actually defers this problem to the executor. So you as a developer or ops guy are responsible to run the migration as a singleton process, and that can be done in multiple ways. In Kubernetes you would spin up a job or a pod, which has to succeed, and Kubernetes will handle it because of the unique name of the pod... Or if you have some other tooling, you need to make sure that you run this first before deploying the new version of the application. + +\[16:05\] If you do the locking, that's fine too, but I actually ran into some issues with that when I locked the whole database, and then I had to ask a DBA team to go into the database to fix it for me, because the database was in a locked state and I couldn't fix it myself. + +**Jon Calhoun:** Okay. So when we're talking about these things running in migrations, and you guys have also mentioned that it's possible to undo a migration, but in reality there are migrations that you can't truthfully undo, if that makes sense... So can you give some examples of what those types of migrations might be for people out there wondering what those are? + +**Vojtech Vitek:** I guess you cannot really undo stuff like deleting tables, deleting columns, right? I think a good practice in general is don't delete stuff prematurely. You are better off leaving the table for a week or two, and then delete it, when you know for a fact that it's not being used anywhere. Because if you deploy a version of your schema and maybe your application that deletes some data, you are losing a way to roll back. That's kind of the problem. + +**Jon Calhoun:** Okay. I was gonna say, deletes are the ones that I've seen most common. There's sometimes some alters in data and things like that that are also hard to revert, but it kind of depends on what the alter is. + +**Vojtech Vitek:** I think any type of data manipulation would be another example, too. So if you're changing the data, let's say you're using some JSON data structure in MySQL or JSON postgres and you're manipulating the data, you're better off to, again, do some backups first, because there's no way back. + +**Jon Calhoun:** It makes sense. + +**Break:** \[17:48\] + +**Jon Calhoun:** So when we're talking about migrations, do you have advice on how to setup that process? Because at times we do need to eventually delete data, or maybe we need to add a new column, then we need to deploy some code that uses that column... So do you have advice to some people, what process they should be using for that type of release, if that makes sense? + +**Mike Fridman:** I could probably describe what we did, and I think that's what a lot of folks do... You typically wanna decouple your migration steps from your application. So this is the one thing that a lot of folks try to do, and it may or may not work out well - in your runtime, for example, you're trying to add your migration steps within the same application that you're running; so it's literally the same binary that you're compiling. And that's gonna potentially get you into troubles, and we can elaborate on that... But what we've seen works best is when you have your application and then you have your migration step. + +In our case, for example, we have Kubernetes, so that's doing a rolling upgrade where we sequentially apply the migration steps first, and then we do a rolling upgrade of the nodes. Now, the thing to remember there is that those nodes are going to have an old version and a new version of your application running at one time. So they're gonna co-exist. So whatever changes you're making to your database have to be forwards and backwards compatible in all the new versions of the application. + +\[19:56\] So some migrations you actually have to split into two steps. So you can't do it all in one step. The way that usually works is write your migration, update your code, deploy that changeset to production, and then write another migration and/or more code changes, and then deploy that again. The main reason you wanna do that nowadays is to accomplish zero downtime deployment. + +It's easy when you can just say "Oh, I'm gonna turn the application into maintenance mode", and then you don't have to worry about writes happening to your database and you can go nuts. But if you have a high-available system, you have to be careful with how you structure that. And the way I explain it is I think the way most folks do it. + +**Jon Calhoun:** That makes sense. It's one of those things that it's hard, and sometimes it's more painful to do, because you have to take work and basically split it up over several deploys, and that can sort of feel like a drag... But at the same time, I've seen it save my butt several times, that's how I'd put it, where you can actually make these data changes and things still continue to run, and you have a little bit of time to make sure it's what you want before you go through. + +Speaking about that, can you think of any other mistakes or pitfalls that people might make whenever they're setting up migrations? I know one example that we got from Twitter was that Nate had -- uh I forget who specifically, but I have Nate in my notes... Nate had mentioned that his team likes to use just 01-migration.up, or basically they pick a specific name... So it's 01-migration, and 02-migration, and they actually use the word "migration" so that if they end up with out-of-order or two conflicting -- like, two developers have 02-migration, when they go to commit the code, they immediately see "Oh, this one's already taken. I need to go update it." Are there any other tips like that, or pitfalls that you've seen people fall into? + +**Mike Fridman:** I think if you're using tooling, you can typically catch those things. For example Goose offers a way to just create those files without you having to think about it. So it'll look in your Migrations folder, see that 76 was the last migration, and it'll create 77, does it like 007-77, with a name so that you don't even have to think about naming at that point. Yeah, I don't really have much in the way, because the tooling takes care of so much of that. + +**Jon Calhoun:** Okay. So when you talk about the tooling, would you suggest that people who want to learn a little bit more about this process actually spend some time maybe building a really simple migration tool on their own? + +**Mike Fridman:** Oh, 100%. And this is one of those things I wanted to point out. The real heroes here are the folks writing the database drivers. Thank you to those people for writing those drivers, because none of this would be possible. These migration tools that we mentioned - Goose, Golang Migrate, Dbmate and a plethora of other ones, they're really just thin wrappers with some functionality in terms of like reading some files, parsing the SQL statements, and then applying that through sqlDB or a SQL transaction, and then just exec with context, or something like that. + +So definitely worth spending some time thinking through like how would you write a migration tooling... You'll come to the consensus that it's actually not that bad. + +**Jon Calhoun:** It's actually something I did once on my own, and I did it in Go as well, and I've found the version in Go was a lot cleaner than I expected it to be. I think you can basically get something working in like a hundred lines of code, or something like that. I think you had mentioned earlier that Goose supports Go files as well... Is that correct? + +**Vojtech Vitek:** Yes, it does. + +**Jon Calhoun:** Okay. So even supporting your own custom functions as migrations wasn't that bad, in the sense that you could basically just write a function that reads a SQL file and turns that into a Go function. And then you can also just pass Go functions into like "Here's my list of migration steps, here's the ID for each one", and let it run them. + +I've showed it to some people who have had some confusions around migrations in the past, and it's kind of crazy how enlightening that is, just to see "Oh, there's nothing too magical going on. It's really just following some series of steps." And not to say there's not more going on with your tooling, because there's a lot of things it can do to help save you, but it's nice to see what exactly the core of it is, and that it's really nothing too magical going on there. + +\[24:08\] So the next question I have for you is - talking about going to production... So we all deploy in different ways. For instance, Mike, you had talked about building this into your application, or some people like to build it in their application, so that when they deploy, it just kind of automatically does that... And you've talked about also having it as a separate step. So what are some of the ways you've seen people push migrations to production? What does that process look like? + +**Mike Fridman:** Yeah, so I think there's three main ways. The first one is doing it ad-hoc, like running SQL against your database, creating tables and so forth... So we call that, let's say, the manual way. Then there's the semi-manual way, where you have a tool, let's say on your local host, like Goose or Golang Migrate, and you set up the connection strings, and then you're just reading directly from your local file system, and then applying that to production... But then, eventually, when you get to larger organizations where you truly wanna have auditability, versioning, being able to track what happened in production, then you set up a continuous deployment environment where that step of physically taking some files and applying them in production is carried out in your CI/CD pipeline. + +**Jon Calhoun:** And I think you and I talked at one point - or maybe it was Vojtech and I, I'm not sure... But I know one of the things that sometimes comes up is that when we're running on one server, things feel simple, because you can deploy, you can wait for the migrations to run, and then you can deploy the new code... But a lot of times, people end up getting in the case where they're running in multiple servers... So do they need to change their process I guess is what I'm asking, to support that sort of setup, where their application is running in multiple places? Or what does that look like? + +**Vojtech Vitek:** I think it may depend on the tool itself that you're using. As I mentioned previously, some of those migration tools use a locking mechanism, so even if you have hundreds of nodes running at the same time, and you're deploying them via some rolling update mechanism, they will eventually hang on and wait for the first instance to finish the migration first, because of the lock-in mechanism. It has some pitfalls too, because the mechanism can hang, and then you need to resolve it by hand... + +I would say the other approach is to really run a singleton process. So you just need to make sure that your CD pipeline has either a manual step or it has some dependency between your migration, and then deployment of the rest of the application. I saw people using Init Containers in Kubernetes to do this, that relies on the locking mechanism, I saw people (including me) that are running a single process instead and waiting for the success or failure on that one... So I guess it depends on the tool. + +**Jon Calhoun:** Another one I've seen people ask about is GitHub Actions. Have you seen people use something like that? Like, if they're using that for deployment and continuous integration stuff, have you seen people successfully use a step there that runs migrations before doing a deploy? Or are there some reasons why that would be a bad idea? + +**Vojtech Vitek:** I can't really think of a reason. It really depends on the infrastructure and the CI/CD pipeline of that specific organization. I mean, GitHub Actions is at the end a CI/CD pipeline anyway, and that's what most of the folks out there are using for deployments as well. So as long as you can depend on one step to another, and make sure that the migration was successful first before deploying the rest of the application, you're good to go. + +**Jon Calhoun:** That makes sense. I know it's one that people are more familiar with a lot of times, especially when they're getting started. I feel like GitHub Actions are a little bit more approachable. Obviously, when you get to larger-scale corporations there's usually some process involved. And I know these questions are hard to answer, because everybody's deployment process and everybody's production environment is so different. I think that's something it takes a while to realize when you're new to development, until you actually go work places and realize that everybody has a million custom tools built, because they've kind of evolved this production environment... So you take everything you're learning and you kind of learn to adapt it to where you're working and what you're doing. + +**Mike Fridman:** \[28:22\] I think GitHub Actions are perfectly fine, because in GitHub Actions you have workfiles and you have jobs, and then within jobs you have steps... And as long as one of those steps -- and steps are run sequentially, if I remember correctly, within a single job. So as long as one of those steps is like "Apply your migrations" and then the next step is "Roll out your application", I think that's a perfectly valid way to do it. It's actually how I do it for the little toy website that I have, bestofgo.dev. That's exactly how it works, with GitHub Actions. + +**Jon Calhoun:** Awesome. So the next thing I have is about testing. People always wanna know "How do you test different parts of your code?" and when it comes to touching data, obviously that's a very important thing to make sure you don't mess up. We've all heard horror stories about people accidentally dropping databases and things like that... So how do you test migrations, and what are some different techniques people can use to avoid having big mistakes? + +**Mike Fridman:** Yeah, so testing migrations can usually be done in your continuous (CI) environment. The best way I've found doing that is you kind of apply your migration, add some data, apply a migration, add some data, apply a migration, add some data, and keep doing that going all the way up. Then once you're all the way at the top, run migrate down all the way down to zero. So you're kind of testing both the up and down, and you do that literally on every single merge to your main branch, or on every single PR push... So that's one way to test it, and it's usually a good way to test it. And that's actually why we have these tools, is because a lot of times you wanna mimic what production looks like, or what it currently is now, and what to change that would look like... And you don't always have access to your production database. So you can't just drop the schema and then try to test against that. You wanna decouple that, especially at larger organizations. So your migrations is a source of truth, and you're kind of migrating up to whatever the production is, and then testing your change site as you go up and down. + +**Jon Calhoun:** So when you talk about testing this, what does the test actually look like? Is it like a Go test file, is it something else you're doing? How would you typically do that? + +**Mike Fridman:** Right. For example, now I work at Buf, and the way we do it is for every migration that we apply, we also do a few insert and update statements. And then the next person that comes in, they write a few insert or update statements, and you're kind of just constantly checking to make sure that the things that you migrated, you can actually insert data into there. + +And then once at the end of the test, it'll run all the down migrations and make sure that you have nothing left in your database. So literally doing that inside something like testing.t, and just a big test file, with some helpers around that. + +**Vojtech Vitek:** What I also like to do on top of that is to check the final schema of the database. So both Postgres, MySQL and modern databases let you explore the schema, drop the schema from the database. And what I've found really useful is to check out the desired schema into your GitHub repository, and then check your staging or even production schemas against that and compare it from time to time, just to make sure that the schema is valid, and in the same state everywhere. + +**Jon Calhoun:** Is that something that people are checking into their tooling? Or sorry, is that something they're checking into Git, the schema that they're gonna be like "Okay, everybody should compare to this"? Or how does that check process work in your mind? + +**Vojtech Vitek:** Right. So you can check the final schema after you run all the migrations into your Git repository... And then this is also applicable to your local host. So even when you're developing your own new migrations, you see the changes against the file, and you can compare if this is exactly what you meant to do. So Git actually helps you see the changes in the final schema, which is very useful. + +**Jon Calhoun:** \[32:12\] Okay. + +**Mike Fridman:** Yeah, so maybe I can expand on that one. I also found that useful, where you kind of apply all of your migrations, and at the end of that, let's say in CI, you ask CI to drop the current schema. So you're running, let's say, a Postgres database in a container, you're applying all your migrations and you're dropping a schema, and now the developer also checked in a schema file, and then you can do a diff against those... And if that (let's say) developer accidentally locally added an extra index, that'll show up in the diff, and it'll fail CI, saying "You're trying to commit something that you probably applied that isn't part of your migrations." + +And when you go to auditing, you can then ask, let's say once a quarter or once a month, or maybe even automated in some way, a DBA to say "Give me the current schema in production", and then compare that against the schema that you have checked in on your main branch. And you'll oftentimes find that -- let's say you go on vacation and something exploded and some DBAs logged into the database and added some indexes, and all of a sudden that's not reflected in your migrations, so you've kind of lost that source of truth... But checking in that schema and then comparing it from time to time allows you to catch any discrepancies that might have happened because people did some manual operations. + +**Jon Calhoun:** So when that type of thing happens, does it make sense to try to go through the effort of somehow getting a migration file -- like, basically, you want a migration that's not actually gonna run when it goes to production, because it's already been run in production. So is it worth the effort of trying to do that, so local development environments sort of reflect production better, or how do you handle those situations? + +**Vojtech Vitek:** I think this type of changes that let's say some performance engineers or DBAs would do is adding the indexes, as Mike mentioned. And just to make sure that, for example the CPU of the database server goes down immediately when there's a high traffic spike, or whatever... And for those SQL queries, you can always write an absurd query. For example, you can check if the index already exists, so create index if it doesn't exist, and do the operation without failure. + +But yes, I believe that it's worth it to keep the schema the same everywhere... For example, I'm working with one of my customers who has production in five different regions around the globe, and I would be very surprised if there was a different database schema just in Europe, compared to North America, and then something would blow up because of it. + +So I like to be in charge of our database schema, because schemas is our state, the most important thing of our application. + +**Jon Calhoun:** That makes sense. And it also makes sense what you were saying earlier, Mike, about developers can sometimes make changes locally when they're sort of tinkering with things and it's easy to forget, especially if you have a long weekend or anything; you can forget that you even tried some different change to your database... And we don't always just drop our database and rebuild it before we submit things. So it's definitely a good check to have. + +**Break:** \[35:19\] + +**Jon Calhoun:** So I guess we're getting close to the end, so we're gonna start talking about unpopular opinions soon... But before that, are there any big takeaways, or words of wisdom, or anything like that that you'd like to share with people before we move on? + +**Vojtech Vitek:** Write documentation. Disaster recovery documentation. Be ready. Be ready to have backups and be able to actually apply them if you need to... Because if you are in panic mode, if the company is down just because of some production issue, you will not think straight. So be ready, that's my advice. + +**Jon Calhoun:** Vojtech, is that something you suggest people actually go through the process of trying it once before they get into actual panic mode? + +**Vojtech Vitek:** Yeah, that's definitely useful. When you're writing this disaster recovery document, you should be also trying it by hand. That's kind of the point. + +**Jon Calhoun:** Okay. So basically have a playbook to go by, so that you aren't making it up as you go and you're also freaking out at the same time. + +**Vojtech Vitek:** Yeah, that's right. And same goes for rollbacks if there's a problem, same goes to potentially think about if it's worth it to do the rollback or if you should go and release a hotfix instead. It's all about trade-offs, but you've gotta be ready. + +**Jon Calhoun:** Mike? + +**Mike Fridman:** Yeah, I was just gonna piggyback on that and say please, please, please have backups. You don't realize it until you need it, and if it's not there when you need it, you're screwed. So make sure you have backups and make sure you understand how to actually apply those backups and do disaster recovery. It's one of those things everyone talks about, everyone documents... But if you don't actually go through the steps and do it, when there's panic and you don't know what you're doing, then you're gonna be in trouble. + +**Jon Calhoun:** I think I've also heard stories of people where they thought they had backups, and then when it came time to actually use them, they realized that the backup process wasn't working the way they thought it was... So all of a sudden, they don't have the backups they thought they did. Like, yes, you wanna write this stuff and test it, but I think it's also good to periodically verify that it's still working the way you expected with your backups, because the last thing you wanna do is find out your backups are six months old when you have a major outage. + +**Vojtech Vitek:** Yeah, that's for sure. And if you're using a database from one of those big providers, then use their own backups. There's no need to write your own tooling for that. + +**Jon Calhoun:** It's one of those things where -- I do a lot of smaller projects myself, or myself and one or two other developers, and sometimes it can feel like you're getting cut by a thousand papercuts when you have all these different bills stacking up... And you aren't a big company, so you're paying for all these things, like backups and database hosting... But at the same time, there are so many easy ways to just break everything or screw everything up if you don't just rely on them. They are incredibly helpful; I know that sometimes people like to avoid them or like to try to roll something custom because they don't wanna spend that extra couple dollars... But at the end of the day, if you're planning on doing something using real data and people are paying for, it's worth getting it right. Okay, we're going to move on to Unpopular Opinions... + +**Jingle:** \[39:38\] to \[39:54\] + +**Jon Calhoun:** Okay, Vojtech, I think you said you have an unpopular opinion for us. Would you like to share? + +**Vojtech Vitek:** Yeah. I mean, generics - I'm not against it, I'm very excited about the feature; I can't wait for it, and I'm already playing with it... But I'm also scared for the Go community, like what's gonna happen in the next two years. I'm sure that people will use it and abuse it to the levels that we will eventually realize that generics don't really help there. + +\[40:24\] I think this kind of happened with the Go channels back in the day... People were very excited about the channels, and they used it everywhere, prematurely; they didn't really make sense. And it settled, and now people only use it for specific use cases. And I think the same applies to Go generics. So people, please don't go crazy. + +**Jon Calhoun:** I think my one hope around that is that anybody coming to Go from a language that already has generics isn't gonna be like "Oh, I need to use these generics", because they've already had enough experience with it... Whereas the concurrency stuff, if you're coming from a language that doesn't have great concurrent primitives, it's really tempting to use them all the time, because you just haven't ever seen them before. + +An example of this is when I first learned about meta programming in Ruby, I used it way, way too much, because it was just so cool. And then you start debugging code that's using meta programming and you're like "Oh. I don't wanna use this ever again if I can avoid it." + +**Vojtech Vitek:** Yeah, I think the reason why we all love Go - it's simplicity. So I'm just hoping that we will not lose it any time soon. + +**Jon Calhoun:** Mike, do you have an unpopular opinion you'd like to share? + +**Mike Fridman:** Yeah, I do. Hindsight is always 20/20, but I think one of the biggest mistakes in Go is the logger struct. It should have been a logging interface. And because there's no unified interface to marry all of these third-party packages, you have everyone reinventing what a logger interface is... Which creates a huge problem, because if let's say you're in an organization and you wanna create a structured logger that 10, 12, 15 teams are gonna use, you're constantly gonna have them coming up to you and saying "Well, this doesn't satisfy this package that I imported", and "This one doesn't satisfy this package." All of a sudden, your structured logger that you wrote for your company is gonna have 50 methods on it because it's trying to satisfy all of these different community logging interfaces. + +**Jon Calhoun:** That's one that in my mind I think would have been hard to predict upfront. It's almost like the standard library - it's hard for them to enforce "This is what a logger should look like all the time", because like you're saying, everybody has different wants and needs... I don't disagree with you fully, but I'm also wondering... If I was the person in charge of designing that interface, I would have struggled to be like "What is going to satisfy everybody's needs?" + +But you might also be right, in the sense that had there been an interface, maybe all those third-party tools would have been more likely to adapt, and make their tool work for it... Because I think about HTTP handlers, and stuff like that, and almost every router out there tries to make it work well with the standard HTTP handler... I mean, there are a couple that don't, for various reasons, but for the most part, a lot of them do. + +**Vojtech Vitek:** Maybe a good start would be to have this interface implementing the methods that are currently available in the standard logger struct. That's a good start. + +**Jon Calhoun:** I'm trying to even think about which ones are there, because I don't think I ever used the standard logging struct... \[laughs\] I'm sure I've used it, but it's not common, is how I'd put it. + +**Mike Fridman:** So I would have thought everybody would have agreed that we needed a logging interface, but it seems to be slightly more unpopular than I would have expected, so I'm actually genuinely curious what the feedback and the reaction would be... Because I've struggled with that quite a bit. + +**Jon Calhoun:** I'm assuming that you're in a place -- like, when you talk about a company that's gonna be using it in 15 projects, to me that sounds like a reasonably-sized company, with a decent amount of engineers... Whereas I think part of the reason that -- it's not that I disagree with that, it's just more it's never been as much of an issue... But as I said earlier, a lot of my projects are very small, so it's easy to just be like "Company-wide, this is what we're using. Everybody get on board." And I think at a big company, that's much harder to do. + +**Mike Fridman:** Yeah... But it also creates a little bit of fragmentation, because if you have everyone redefining what an interface is within their project, I feel like that's something that should inevitably be solved at the standard library level... But who knows. + +**Jon Calhoun:** \[44:18\] It's a tough one. I guess we'll see. I suppose it's something that could potentially introduce a new package in the future for, and at that point third-party libraries would hopefully start to move towards that... But who knows. + +So I have a follow-up, maybe unpopular type of question for you... What do you guys think of the any alias that they're adding in Go 1.18? For anybody listening that doesn't know what this is, basically they're adding a type alias where the word "any" is going to map to the empty interface. So interface, then left curly bracket, right curly bracket, where there's no methods defined. + +**Vojtech Vitek:** I don't have a strong opinion on this one. I'm quite used to any from TypeScript, even though if you're using TypeScript, you shouldn't be using any... Right? So I think it's fine; but yes, I get your point, because suddenly we will see interfaces and anys everywhere, and it will get mixed up. + +**Mike Fridman:** Yeah... I also don't have much of a strong opinion on that one. I installed 1.18, saw a whole bunch of errors with empty interfaces and any, I'm like "I'm going back to 1.17.5 and call it a day." + +**Jon Calhoun:** I used it briefly, and I will say that getting used to typing "any" was weird. I'd have to use it more to really get it... And I do agree that -- hopefully, a lot of projects just switch everything over to any, because without doing that... You're still gonna have that confusion of somebody new to Go eventually runs into this empty interface and they're like "What is this thing? Why is it here?" Because I think that's kind of an inevitable step for anybody learning Go, is they see this empty interface and they're like "What is this thing? This makes no sense." For some reason, just at first, it made no sense to me. Then when you stop to think about it, you're like "Oh, that makes more sense." It just wasn't as intuitive as I would have thought. + +**Mike Fridman:** That's such a good point, because one of the beautiful things about Go is I could go type something in a search engine, find an article from 2015 which would be 6-7 years old now, and it would be exactly the same. Not much has changed in Go. So there's so much resources out there that are gonna reference empty interfaces. And then now we have "any". So for new folks that don't know that, it might cause confusion. And the more things you add to a language to bloat it, the harder it becomes, and you lose what Vojtech said, the simplicity of the language. + +**Vojtech Vitek:** I wonder if the go fmt will eventually replaces interfaces with anys. + +**Jon Calhoun:** It would be a nice way to clean it up, but I don't know... I guess we'll have to wait and see. I know I've seen people show some -- I think it was Brad Fitzpatrick... I think that's who it was; I'm trying to remember for sure... But I think he had posted something on Twitter that was like a short little snippet of how you could go replace it in all of your code, if you wanted to. + +So things like that... But still, like Mike said, with tutorials it's not gonna be that easy, which is gonna be kind of challenging. And that's gonna make it different, I suppose. I know that was one of the things that always frustrated me with JavaScript, is it felt like every few years there was kind of a new way to do certain things, and you'd still see tutorials or code that used the old way, because nobody's gonna go update everything that exists... And that could be challenging to somebody who's learning. + +So I'm curious to see how it goes... I'm hopeful, but at the same time I suspect it might also lead to more confusion, for at least like the first year, where people are coming and they're like "Why are both of these things doing the same thing?" + +Okay, thank you, Mike, thank you, Vojtech for joining me. I will play us out. + +**Vojtech Vitek:** Thanks for inviting me. diff --git a/Observability in the wild strategies that work_transcript.txt b/Observability in the wild strategies that work_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..5b27dfef0b456f555673e769ce5e5861954f8803 --- /dev/null +++ b/Observability in the wild strategies that work_transcript.txt @@ -0,0 +1,347 @@ +**Mat Ryer:** Hello, and welcome to Grafana's Big Tent, the podcast all about the people, community, tools and tech around observability. I'm joined today by - it's only Matt Toback. Hello, Matt. + +**Matt Toback:** Hi, Matt. What do you mean "only"? + +**Mat Ryer:** Yeah, no, it's just understated; it's just like "I can't believe it's you, really, in a lot of ways..." + +**Matt Toback:** It's how you say it, it's not what you say. "I can't believe it's you! You're here!" + +**Mat Ryer:** Oh, right. Okay. I can't believe it's you...! + +**Matt Toback:** You're like, "You're here... You're here." + +**Mat Ryer:** Yeah. Well, don't worry, it's not just me and you. That would be obviously tiresome for both of us. We have a special guest joining us today. + +**Matt Toback:** That would be tiresome for both of us! + +**Mat Ryer:** That's how you say it. You're right. \[laughter\] Today we're joined by Nayana Shetty from Lego. Hello, Nayana. + +**Nayana Shetty:** Hi, Mat, and Matt. \[laughter\] + +**Mat Ryer:** Yeah, just one would suffice. We'll share it. Happy to share. And you are the principal engineer who loves talking about SRE and microservices, right? + +**Nayana Shetty:** Yes. I think over the years I've been in teams where we've built microservices, and when you scale up and have hundreds of microservices, how do you then make them reliable and keep them reliable? That's what I'm interested in. + +I was working at the Financial Times, where we had all of these hundreds of microservices, and how do we manage it there. And now I've moved to the Lego Group, where we're going through massive digital transformation. And here it's like we want to build these hundreds of microservices, so should we care about reliability now, or can we think about it in ten years' time, when we have these microservices? So yeah, that is my context to how site reliability, microservices, all of these work together. + +**Mat Ryer:** Yeah, because it did use to be kind of an afterthought really, didn't it? Which is why I think SRE is short for "Sorry", right? + +**Nayana Shetty:** That's one way of looking at it. I always think of it like "Sorry, I don't understand why people don't think about site reliability in the first instance", or "Sorry, I don't understand why people would build this in such a way that it is half broken, or you don't think about the future of this product, and you are very close to reinventing the wheel every few months if you went in that direction." So yeah, sorry is probably one of the ways of looking at it. + +**Matt Toback:** Just even on a personal note, I'm excited that Nayana is here and joining us because we met in 2018 for the first time in an attic in Amsterdam... Which - when said that way, it doesn't feel weird at all, right? \[laughter\] + +**Nayana Shetty:** I mean, we were talking loads of monitoring, Grafana, and Graphite, and all of those things... So yeah, attic didn't make a difference then, so... + +**Matt Toback:** But it was wild, because we planned GrafanaCon EU in 2018, which was technically our third Grafana Con, but was kind of our biggest up to that point, and the most what felt well-produced. It was in the middle of an arctic chill... Do you remember that, that the canals had frozen over? + +**Nayana Shetty:** Yeah, I had struggles getting back home... It took me two hours, where it should have probably taken me only half an hour to get home. + +**Matt Toback:** Oh, man... + +**Nayana Shetty:** So yeah, I completely remember that. + +**Matt Toback:** And we were in this stage where Tom had just joined the company, we announced it on stage, the acquisition of Kausal; the entire company fit on stage, it was 25 people when we said kind of goodbye at the end; there's still this photo that circulates here where everyone is just kind of shoulder to shoulder, the whole company. + +But I do remember - you stand out to me, us being up there, kind of in that breakout room, talking about what you were trying to do with the Financial Times... And it feels like you kind of continued in this natural progression, in this natural journey. When you think back to you then, how did you see the world? + +**Nayana Shetty:** \[06:15\] At that point we were investigating -- we had quite a lot of monitoring tools at the Financial Times, and I was working on the team that provided monitoring as a service to other teams... And my head was going mad thinking "Okay, how do I as a team with 4-5 engineers be able to support these 20-30-odd engineering teams who all want monitoring?" I'm there using from Nagios, to Zabbix, some Graphite, some -- I think there were very few installations of Prometheus at that point... And it was like "How do we get all of these different use cases together, and how do we get them on a platform which could work together?" And it made me -- I was worried at that point, and three years later or four years later, looking at it, it's like... + +**Matt Toback:** You're still worried. No... \[laughs\] + +**Nayana Shetty:** I mean, I have moved on from the Financial Times, so I'm less worried about the Financial Times monitoring systems, but I still worry about the same use case. I see it here as well in the Lego Group, where there's different monitoring tools that we've got across the organization, and it's "How do we get them all together?" and "How do we say a single story that everyone could understand?" rather than every single team trying to solve the same problem. So it's still very similar, but probably we have better tools and processes in place that can help us... So that's how I see it. + +**Mat Ryer:** Yeah, something you said earlier stood out... This idea that you're like "Why did you build it like this?! If only you'd built it differently, we'd be in a much better position now." + +**Matt Toback:** \[laughs\] "If you only did it right", is that what you're saying? + +**Mat Ryer:** Yeah, but that's the question... When should we start caring about this stuff? When should we start worrying about how are we gonna operate this? + +**Nayana Shetty:** I think this kind of relates to how I've moved in the journey in my career, and stuff. I started off as a test engineer, just doing some manual testing, then moved on to doing more QA, more quality-related things rather than just testing... And over the years I've seen the transition in a lot of organizations as well, where they've moved to this shift left, and test early, release as small as possible, and continuous iterations, and stuff. So all of this I think kind of leads to that point of "How do you make your future better?" + +And one of the quotes I've often used is being kind to your future self. How can you make your life easy in the future? Think about that today, when you're building whatever you're building. And that comes with -- if you're building a new product, think about "Do you even have to build it? Can you just look at what's in the market and re-use it?" If it's a non-differentiating thing, then why build it? If it is a differentiating thing - yes, put your heart and soul into it. But then when you're doing it, make sure you think about the sustainability aspects of your product, and not just today what the customer would get. + +I've often used this carrot and stick kind of approach in teams, to show the benefits of what you could get out of thinking about monitoring, observability upfront. Usually, the carrots are like "You build it in the right way, then you can actually forget about your systems, because they will take care of themselves." And the stick approach is often "If you didn't do it, then you have to go into the routers, or all of those other things that comes with making a system more observable and keeping it sustained once it's up and running. So yeah, I think that's what I've used in the past to actually help teams nudge in that direction. + +**Mat Ryer:** Yeah. I mean, you know, I would be kind to my future self, but I'm too busy dealing with all the stuff that my past self left me to do... \[laughter\] So I don't know. + +\[10:15\] So that's the thing - if you think about how it's gonna be, where it's gonna be running, the realities of that... If you think about that, the earlier, the better, almost, isn't it? + +**Nayana Shetty:** It is that. And it's also that - yes, you are fixing things from yesterday; but if you don't fix it and leave some goodies along with it, then tomorrow you're fixing today's problem. So you're still in that vicious cycle. To get away from that vicious cycle I think you need to actually step back sometimes and put that extra effort. + +I remember one of the tech principles we had in the Financial Times for ft.com was -- I've forgotten this... \[laughs\] + +**Matt Toback:** It's okay. As we're talking about the past self and future self - is it okay that I've completely forgotten about observability and now I'm just on a personal journey and I'm thinking about all the decisions made, and yet to make, and how to provide goodies for everyone? \[laughter\] Or for me? I don't know. I forgot that part. + +But Nayana, the carrot and stick, can I ask you - has there been a stick that you've seen people try to use, that just didn't work? Or not didn't work, but either was too harsh, or just kind of like -- not ill-intentioned, but ill-executed? + +**Nayana Shetty:** I think it's about the motivation factor behind doing something. That's how I saw the carrot and stick. So the carrots were the motivation factors that we were providing to teams, saying "If you did something right and if you thought about how do you monitor something, how do you add alerts in place, and how do you make sure it's auto-recoverable where possible, and stuff, then you don't have to worry about it." That is more of the motivation for the team to be like "I can be really proud about what I build, and I don't have to think about it again." + +But at the same time, we know that every team has these deadlines to meet, and there are product owners who would have their own feature set to build, so it's that kind of scenarios where you actually still need the stick to help the teams be like "Look - I mean, yes, we understand your pressures, but this is more important as well." And I did remember the quote, and it was "Slow down to speed up." That was one of the tech principles we worked with. + +Basically, yes, you can go at 100 miles per hour today, but then if you don't build it in such a way that you have put those measures in place, then tomorrow you have to break and stop. But if you slowed down and went at, say, 60 miles per hour, you're there for the long run, and you would go on longer. So that's how I would see some of this stuff. + +**Mat Ryer:** Yeah, that is so true. We actually built a little project before Grafana, we built a project management tool called Pace. And it was trying to get across that exact thing, which is that you feel great going at a thousand miles an hour, but there's important things to do along the way... And it's hard to retro-fit a lot of this stuff, so thinking about it upfront sometimes can save you. + +It's a bit like how you design for failure as well. In the perfect world all the messages flow perfectly and there's no problem, but in reality it's way more messy, things fail... So idempotency and things come into play, where you may design expecting this is gonna fail. + +I do write Go code, and Go has error handling as a kind of explicit feature. There are values that are just returned as the second argument to functions, and things like this. And that frustrates a lot of people, because they're used to exceptions or something, that's just sort of automatic... But it forces you to think about what's gonna happen if this thing fails, and that's a great discipline to get into. + +**Nayana Shetty:** \[14:14\] And I think it's a myth to think that your system won't fail... Like, always build your system in such a way that it will fail. If it doesn't, then you have a problem. \[laughs\] So make sure you add those checks in place, so when it fails, it can smoothly recover. + +**Mat Ryer:** Yeah. I know some companies that have that as part of the proper testing approach - literally, things will break on purpose; it's a first-class concern that they have. And it is that thing of -- I don't know, is it just ego, that people think "I'm so good. I'll write this, it's gonna be great"? What's going on? + +**Matt Toback:** It can't be... Honestly, it can't be. We've all known and experienced it enough... Do you think? + +**Mat Ryer:** Yeah, I don't know... But the thing is, when I'm writing code and it doesn't work, it's shocking how quickly I'm like "There's something wrong with the processor." \[laughter\] "The processor is not working. Or physics has changed." I'll go to "Physics has changed" before it's my fault. But it turns out I just did a capital letter where I shouldn't have. + +**Nayana Shetty:** I've been in teams where they do pairing, and those mobbing sessions and stuff... They have kind of helped in sense-checking people's egos a bit, and being like "I'm not the best." And when two people talk about it, I think it does help think "Yeah, that is the reality that we live in, and this is what you need to consider." + +**Matt Toback:** Is there anything that you would -- even like you were saying before, that progression between being in manual testing, and then Q&A, and then moving to SRE... Was there like a moment where it clicked, where people just started incorporating testing into the code? Do you see the same progression happening in observability, to where there'll be some moment and some kind of click where it just becomes part of it, as opposed to the separate thing that happens afterwards? + +**Nayana Shetty:** I have seen it work in some teams, and... A lot of teams I've worked in are all autonomous teams, so they can basically build how they want, using whatever technologies they want. What has often helped teams like that is having some sort of guardrails which actually says -- and also being aware that not all applications need the same level of checks and monitoring and all of these... So being aware that "Okay, there is a level of criticality of my app, and if it is a highly critical app, then let me put all of the things in." And if it is a less critical app, in that case you would just have maybe just a simple health check. That would be good enough. You don't need to go all board. And one thing that we had when I was working for FT was you always have at least a basic check on all your apps, otherwise -- like, we used to get this service operability score for our applications, and the score used to go down when we didn't have some of these things in place. That was a nice measure, where people thought about it from the beginning, rather than as an afterthought about some of these. + +But what could happen in these kinds of scenarios is people go all-in and they just say "Oh, I'm gonna monitor everything, have all my logs..." You don't need to go all board on this. There's a limit to how much you need to monitor as well. And understanding the criticality of your app and then building your observability around that is probably something that teams should think about. + +**Matt Toback:** How would you -- if a team was listening to this and they were trying to understand the criticality of the app and then make decisions around it, if you had them sitting in a room, how would you explain it and say "Here. Start here. Do this"? + +**Nayana Shetty:** \[17:54\] I think it depends on the business criticality. If it is a highly business-critical application, which means if it went down for, say, more than 15 minutes, then we wouldn't be in business - if it's that kind of app, then you need to have your alerting in place, the right level of logging in place, which actually gives us any of the audit records that actually us what's happened with the applications. And then any sort of health checks. + +So there's probably like two levels of monitoring that we should think about. One is the application-level monitoring, and then there is the system-level monitoring. Being able to figure out where the problem is soon enough is something very critical when it's a 15-minute recovery thing. But if it is an application that's less critical, then maybe just having the application-level monitoring is good enough, where you could take longer to actually investigate, look into the logs and actually figure out where the problem area is, and stuff. + +So I would suggest teams to think about how critical their app is, and that is something the business should help them with, not something that the team just decides "Oh, this is the most critical thing." + +And once you know the business criticality of something, then it is coming up with some sort of checks saying "If it is a highly critical system, then we do both application, as well as system-monitoring." Otherwise just one of them, based on your use cases, and stuff. + +In the past I've spoken about the USE method, and RED method that we could use for these kinds of things... I prefer using RED method or Google's four signals; it depends on what your team's needs are and what fits into your use cases. So you would use the RED method, which is rate, error and duration for every single application that you build. And it's very easy to see that in a microservice world where you have different kinds of applications, you have the same three parameters that you're measuring across all of them, so it actually helps the team analyze -- irrespective if that belongs to your team or any other team, you just know where the problem is. + +And the same with systems side of things - you would go with the USE method, which is utilization, saturation and errors, and you would do this for the CPU disk or network and all of those different areas, and you basically know where the problem is, and it's easy to find out. + +I would say it is hard, it takes time, so invest based on how much returns you would get on these when you put these checks in. That is something the teams should be mindful about when they are investing in monitoring and alerting. + +**Matt Toback:** Is the primary counter-balance in your mind the effort that it takes to keep this monitored well, or is it also cost? Do you think about the cost to operate, or the backend? + +**Nayana Shetty:** It is the cost, and at the end of the day it should be the cost to the business, as in "How much does having the system down cost us?" and you basically work backwards from there, saying "If this was down for 15 minutes, it would cost the business so much", which means we as a team should be investing more time in actually getting the right amount of measures so we can solve the problem or narrow down the problem quickly. + +I would always focus on the business value rather than the team's individual product value, and stuff... But yeah, it depends on -- like, if it were an internal system... Again, one of the teams I was in, we were building monitoring tools for other teams. So we don't have real business value as such as our team, but we were supporting teams that had really high value systems. So that kind of meant that we had to think about the application-level, as well as system-level monitoring on our systems, and stuff. + +**Break:** \[21:50\] + +**Mat Ryer:** I love that advice of "Pay attention to the value you're gonna get from the effort that you put in." I like monorepos - I'd just like to put that out there; I love monorepos. And the reason I like them is because you can have a pull request that has a unit test, some backend code, maybe some API changes, frontend code in there too, hopefully with some frontend tests maybe... And it's nice that that all gets applied to the system in one go. Does that also apply to this sort of field or the instrumentation of that? Should we be having those kinds of conversations at that point, so that we kind of think about it as we go? + +**Nayana Shetty:** \[23:59\] I would love to say yes, but I've not seen a team do it really well. I can see the challenges -- like, when you have this monorepo and everyone is contributing to the same central repository, there is a challenge that the parameters that you would think about for your product and your monitoring systems might be different to what another team would be looking at. So there is a challenge with how do you then look at this as a product that we sell to customers. You have to think about capability monitoring maybe, rather than your individual product monitoring, where you're thinking about "What is a capability that I'm providing to the customer?" And those should probably be things that we have at a central level, and we do it as in when we add new features, we make sure we don't break the monitoring that we've got across the capability. + +But on a single individual product team's perspective - yeah, I don't know how much value it would add, so it depends on that, I guess. I'm not sure. Have you seen it work in your teams, or something? + +**Mat Ryer:** Well, we have at least the conversation... When there's a PR for like a big feature, we will chat about it, and say "What do we need from this? What's going on here that later we're gonna need?" And it is that thing about "Be kind to your future selves." But I don't know that we've got that right yet, or anything... Because in a way, we don't really know what's important upfront necessarily. But sometimes you do, and I like that there are guidelines that we can follow to give us a good foundation, and then of course, we're gonna have to fine-tune it depending on our particular case. + +**Nayana Shetty:** I agree, yeah. + +**Matt Toback:** All this to me starts to distill down into -- it's some amount of like if you are doing the centralized monitoring, or there's a level of that, and then you have to communicate this down to these teams and you have to get them to buy in... Where do you do that? Or even how would you suggest someone else do that well? + +**Nayana Shetty:** Central teams pushing things is -- like, irrespective of it being monitoring or anything in general is really hard. It should always be driven by -- like, what I've seen work really well is the ones that are driven by value-add to the individual team itself. + +As an example, when we were building this Amazon Linux, like a baseplate image that everyone could apply, and they can run their own EC2 instances. When they had this, what we said we will do as part of it is we said "You're going to get monitoring --" I think we were pushing logs to Splunk in that case. So you would get that feature for free. You would get authentication for free. Have those kinds of things that you will get for free as part of whatever feature set you would give, and that has often been a nice way to drive teams to be like "Oh, yes, I like that, and I will do it." + +**Matt Toback:** Right. Make it so easy that they would rather adopt it, rather than trying to do it themselves. + +**Nayana Shetty:** Yeah, exactly. Another example that came to mind was we had this central repository for -- like a CRM system which we had to enter all of our system information in. Basically, it was like a -- because we had so many microservices, we had a central system where you could go in and query for any particular system with something called a system code and we would know if that system was live, was active, who was working on it, all of that information. + +And what we did when we built this - we said "If you put the right information in this, then you would automatically have a dashboard that would show up only your team's monitoring in it." That was an incentive for teams to be like "Oh, if I did this, then I get my own dashboard. Let me do that." + +\[27:47\] So I think it's that showing intrinsic value beyond just what you want them to achieve out of it. That's how I've seen it work really well in teams. So yeah, you need to have some sort of carrots to actually get people to move towards your solutions and stuff. + +**Matt Toback:** Yeah, yeah. + +**Mat Ryer:** Yeah, that's great. I think that applies to... Everything! \[laughter\] Yeah, if you make it easy and it's sort of a no-brainer... And one example of that is where we can -- like, if we've got APIs, we can just instrument on the endpoints very easily, in a simple way, usually with some middleware or something in the code. There's lots of packages that do this. I think there's probably space for more things like that, more of that for devs. + +**Nayana Shetty:** Yeah. And I think it's the -- I would really want to see central teams be more mindful about this... Because as a central team, you're building these amazing tools, and at the end of it you kind of think "Oh, if I put a documentation together and self-service, everyone's gonna come and use it." But then each individual product teams have their own little agendas to work towards, and the product initiators do their own OKRs... All of those things. So this is like an extra bit of cognitive load onto those teams, which they can avoid if you were to do a lot more promotion within teams, saying "If you did this, you would get a lot of benefit", and it will take some of the risks that you have taken on yourself. + +And also, it's that education piece of "You care about your product. We will help you care about your product." That's something to think about, I guess. + +**Matt Toback:** Yeah. + +**Mat Ryer:** So what are some common mistakes that we make when we're trying to do this? With the best intentions in the world, we want to do this properly, but are there any things you see that people misunderstand, or common mistakes, common gotchas that you've seen? + +**Nayana Shetty:** I think knowing how much is enough is one of the things I've often seen, where there are teams who just put the basic thing available because it's there in a checklist somewhere, and then they move on... Which is probably not the best for your product. So it's being aware of the value of your product and what is the lifecycle or the journey that your product is going on. That is probably one of the things. + +The other thing that I have seen and I've struggled a lot with is, like I mentioned, about this USE method and RED method to actually build your dashboards. It's very hard to get your network-related monitoring right, and the saturation for networks... Like, how do you do that? And get the wrong set of -- like, I've seen myself having a wrong set of dashboards and alerting, and wondering why this is going off everytime something happens, when it shouldn't happen. + +So I think it's just being okay to experiment and continuously tinker your monitoring and alerting as you go along is probably something that teams should be conscious that -- like, it's not that you build it once and then it's there forever, but there is a continuous evolution that happens with your monitoring... Like how your feature sets go through that. You have to do the same with your observability side of things as well. + +**Mat Ryer:** Yeah, yeah, as -- + +**Matt Toback:** Mat, can I answer, too? + +**Mat Ryer:** Let me just check... No. \[laughter\] + +**Matt Toback:** Oh, come on...! + +**Mat Ryer:** Brutal, but... \[laughs\] Yeah, please, I'd love to hear what you think, of course. + +**Matt Toback:** Okay. So Nayana, you were thinking about the value derived, and focusing on that for the customers... I do think that's a common gotcha, where you build all these tools and you're like "We did it! We did it! It's all there. All you have to do is this." + +\[31:46\] I think the common gotcha is forgetting that you need to deliver something that someone can just adopt easily, like you said. It is a version of -- I was thinking like car parts, or legos, I guess... Dropping off like a collection of car parts and being like "There you go!" And you're like "I wanna drive. I get that I can get there, but you haven't helped me really at all." And you call left, and that's where the metaphor I think breaks, but... + +But I do think there's some version of that too, stopping short of actually delivering the value to the person consuming it, as opposed to just dropping a collection of pieces that can work, but they have to do the last mile. + +**Mat Ryer:** Well, in a way, what helps that definitely is gonna be this "You build it and you run it." We're not throwing this thing over the wall for someone else to operate... Which I know that actually lots of people do still do that. And there's a disconnect. When you are yourselves kind of running it, you're the customer of that data. So a bit like when you're dogfooding software, if you're building dev tools like we do at Grafana - we dogfood a lot; we'll use our tools a lot internally. That's how they're so good, frankly. It's not like we're imagining the user of this; we are the user, and I think that makes a big difference, doesn't it? + +**Nayana Shetty:** Yeah. And also, I think one of the comments I've heard a few people say about this - build your code in such a way that you can debug it at three in the morning. It doesn't mean that you have to do it every day, but if it breaks at a time that you're not fully in focus, you still can get to it easily... And that is something that I think people should be thinking about when building their products, and stuff. + +**Mat Ryer:** That's such a great point, I think... And that leads me to our next question, which is around drills. Should we be doing drills at 3 AM, and living that experience to see what it's like? + +**Nayana Shetty:** Three o'clock is probably taking the Mickey out of people if you were doing drills... \[laughter\] + +**Matt Toback:** Do people do drills? I guess they do, but it's probably not common, is it? + +**Nayana Shetty:** I have seen it done, and I think it's a very artificial environment where the drills happen. One of the things we did when I was at FT was we had these incident drills. Basically, you emulate an incident and then you go about with the team "How do you go about actually figuring out where the problem is?" So you start with which alert it was, and then look at the traces, and then look at what the logs were... You go through the whole cycle of it. It was a way to ease the whole awful hour support that we had within the organization, but at the same time, there were a lot of people who were not very keen of this. Because it's an artificial environment, people felt like that is not reality, so why do it? + +**Mat Ryer:** It's because you didn't at 3 AM. \[laughter\] + +**Nayana Shetty:** Yeah, maybe that. + +**Matt Toback:** I think there's like a touch of maturity in actually embracing drills, whether or not it's artificial. It's that idea "Oh, this is artificial. This is dumb. We don't wanna do this. It's not gonna be like this in real life." And then you think about any kind of -- I don't know, any team environment or any kind of practice that you need to do, because it's more than just debugging the code. Everything is interconnected, and you want to be able to do some of these things more than once, so that way every time doesn't feel like you're the first time on stage. And it feels like you just wanna be like "Hey, what's the right analogy to make if you want to convince someone to actually practice?" I don't know... + +**Mat Ryer:** Yeah, it's not the same... It's not the same, because you know it's a drill. Unless you're doing something where you literally break something and it's not really broken... Or maybe it is, and you're doing something kind of "That seems a bit extreme..." It is gonna feel different, but that still doesn't mean that there's not plenty of stuff to practice... Like practicing when you practice driving. There's an instructor next to you watching everything. That is a very strange situation to be in... But you still move the steering wheel and do the -- I don't drive, but there's a gear stick, I know that... And the horn... + +**Matt Toback:** \[36:03\] You press the horn to go... \[laughter\] + +**Mat Ryer:** Yeah, horn to go, and then you leave that on, so that everyone knows you're there. I do, because they need to get out of the way... \[laughs\] + +**Nayana Shetty:** I think there's also value in -- the other way of looking at drills is like shadowing, and when there's an actual incident not having just one or two people involved in it. Yes, it might be the most critical thing, but having more people just listening and see what's happening, and just be there sometimes helps them understand "Oh, this is how I would go about solving this", looking at those people. So yeah, I think it's a mixture of drills and shadowing maybe that could work in teams. + +**Matt Toback:** I even think yesterday - and I realize why I'm all fired up about this... I parsed through it. Yesterday I visited the sales team, and they were doing these workshops, and they were doing Radical Candor, which is all about feedback, and giving feedback, and getting feedback, and being able to do it well. And then you break out and you're like "Okay, yeah. Check, check, check. I get it. I could totally do that." And then you break out into these triads, and then you practice it. And there's a part of you that goes like "Oh, I don't need to do this. I get it. I get the concepts." And then you try and do it, and you can kind of feel yourself, like, places that are a little bit creaky, or maybe you don't quite get it as much. So I think it's actually where I'm fired up to where even if it is artificial, some of those joints might be either rusty, or creaky, or don't articulate well, and you don't realize that until you do it. + +**Mat Ryer:** I think that was a rubbish point. + +**Matt Toback:** We can cut that. Cut, cut, cut. \[laughter\] + +**Nayana Shetty:** I think it's also a good exercise to do, just to test your documentation, if your documentation is up to scratch. When you've written something, you've written with good intent, but when someone's actually following it, does it make sense? It's something that the drills can actually capture. There's more than one benefit of having drills, I think. + +**Break:** \[38:04\] + +**Mat Ryer:** Nayana, you mentioned earlier this idea that if you do too much, you can overdo it and end up with basically alert fatigue... Just alerts going off. What do we mean really by alert fatigue? + +**Nayana Shetty:** I'm going to give an example so people can relate to it. I was in one of the teams where we used to get close to 1,500 alerts on a weekly basis... We had around 80-odd microservices. So it wasn't like just one microservice, or anything. But then my team was three people looking at this, and it's at that point which you realize "Are they actually looking at this thing, or is it all just being ignored as just noise? Let's just ignore it." And I think it's that point where you start ignoring your alerts, is where you've gone to that stage where you can't take anymore alerts. You're fatigued with the whole alerting itself. And I think it's better to have less alerts, for the most important things, rather than have too many and try to filter it out. + +One of the exercises we did when we have these alert fatigues and like a thousand-odd alerts is we consciously stopped some of the alerts to see who will start shouting. And it happened that more than 50% of these alerts when we turned off, no one actually shouted at us. Like, "Was that even important?" Going through that exercise on a regular basis, where you see if you're ignoring more than at least 10% of your alerts, then go and do something about it. Maybe turn them off, and no one will care. + +\[43:16\] I think teams need to be conscious that it's okay to miss a faulty alert, compared to missing out on a real alert which would have cost us millions of pounds, or whatever it is. So I think it's being careful to put the right alerts in, and stopping at that, and not just going overboard with it. Let's take an example. We have Grafana in our systems, and we have alerting with Grafana. I have this tool, so I can put as many alerts as possible. Not going wild with it, but actually knowing where to stop. That's how I would describe this whole alert fatigue. And with time, it does happen with teams. So it's worth going back and auditing them and making sure you keep them clean as much as possible. + +**Matt Toback:** I wonder if you could do a -- what would be the equivalent of a bug bounty for alerts? How do incent people to go and clean those up and celebrate being like "They're gone!" + +**Nayana Shetty:** Oh, that's hard. What I've done in this is actually gone and turned them off myself, and being like "Let's see who's going to shout." + +**Matt Toback:** Yeah. \[laughs\] + +**Nayana Shetty:** And when no one shouts, you know that they're not important enough. So that is something that I've done, but I don't know how you would -- hm, interesting... Do you have any ideas? + +**Mat Ryer:** We need some analytics, don't we, on the usage of it, really... Then we all can say "No one's looked at these alerts for ages." Or you could put a specific time on it if you want. I don't wanna design the application now... But yeah, something like that. + +I like the idea that you should go back and look at them and pay attention to whether you still need them, and things. This is a little bit like how in GitHub or in your project management tool - if you have loads of stuff in there, most of it is just getting ignored. And in a way, it creates also this idea that you're so far away from being done... Which we of course are, but it sort of reinforces that. + +So it is that thing of it's just so much there, it stops being useful. I like the idea that -- does it take experience though, do you think, to know what's useful and what's not? + +**Nayana Shetty:** It could do... I mean, the more you see these things, you realize where it's useful and where not... And this is where I've seen some of the junior engineers and teams struggle, which is like - they start worrying about every single alert that comes on Slack, or whichever is your preferred tool, and they're like "Oh, what do I do? I've got this alert. It might be my change." But it might not be related to your change at all. + +So I think it's something that a team should do on a regular basis as like a team activity or something like that, where they sanity-check their alerts. + +One way we used to do it is any alert that we actually did anything with, we started putting some -- all of our alerts used to come to Slack, so we started putting some emojis on it, so we know which of them were actually things that mattered. And on a weekly basis we were like "Oh, there were N of these, which we did nothing with, so maybe we can get rid of it." It's very hard to get that feedback cycle on alerts, I've found. + +**Matt Toback:** Mat, you need to collect the emojis and then feed it back. Does the IP go in the other way on emojis? + +**Mat Ryer:** We do that in the Grafana Incident tool, but I need to tell the on-call team about that idea. That's such a good idea, because... + +**Matt Toback:** Yeah, that is kind of fun. + +**Mat Ryer:** ...you could collect that data and literally you then -- "Oh, hello!" + +**Matt Toback:** You're like, "What does dancing penguin mean?" \[laughter\] + +**Mat Ryer:** Yeah. It means it's cool. Who owns that idea, legally? I don't know... + +**Nayana Shetty:** I do, because I-- \[laughter\] + +**Mat Ryer:** It's a good idea. + +**Matt Toback:** \[46:59\] Solved! \[laughter\] I wanted to quickly touch on - as you progress in your career, often you're going to walk into new organizations. And you're new to Lego... And everything today is just reinforced this thought that I was kind of \[unintelligible 00:47:13.13\] on this morning, which is all these things are true, all these methods are proven, and in some ways, it has nothing to do with the technology and everything to do with the landscape that you're walking into. And then you have to figure out how and what you introduce. + +I guess I'm curious - how much about open source tooling makes it easier to transfer into a new organization, and even just how much -- how do you approach going into a new org having this experience, but then also not understanding how everything fits together? + +**Nayana Shetty:** I mean, this is so relatable... I'm going through this now, given I've been in the Lego Group for only three months... And I care about monitoring and in general sustainability of products quite a lot. I've been looking at different teams doing this and thinking "Okay, this team has this Grafana dashboard to do it. This other team has New Relic and they're doing something with it." + +I think for me, I was lucky that I was in a team that was building monitoring tools as a service, providing monitoring tools as a service to other teams, so for me it was easier to catch on to what is happening in different areas within the Lego Group. But I think what I fall back to is always think about what are the core aspects of monitoring. So it's things like logging, metrics, alerting, tracing, so notifications... Some of those core things. And it is looking at those aspects and thinking "How is the team solving these problems?" And where the team have used their tool, I have often just endorsed what they've got and looked into it. But where they haven't, I have often suggested open source tools in those use cases, because of two reasons. One, it's easy to get started and get going with it. You don't need any licensing and all of those kind of challenges that come with a proprietary tool. + +On the other side, there's a lot of community that can help you getting started with the tool as well. So I think those are reasons why I would prefer -- like, when suggesting to teams, I would prefer open source technologies when it comes to this space. + +As an example, when I was doing some experimentation for my own personal project, I could have gone with one of the tools that was already available in the organization where I was working, but then I was like "I mean, if I left this organization, I can't take that tool with me, so it's better to have it on more open source tools." In that case I used -- I think it was Graphite and Grafana that I used in that use case. + +**Matt Toback:** Good choice. + +**Nayana Shetty:** But while there is transferable skills within the organization -- like, as an example, the Lego technology, I think we have around 200 or 250-odd teams in it, and if these people within the teams have to move between each other, speaking a common language is quite important... And having that community outside of the Lego Group who can help us with this is quite important. I feel like that is where the power of using open source technologies comes from. + +I have come from an organization where we were a very big advocate of open source technologies, so I probably would be singing the song of "Let's go all-in on open source." \[laughter\] + +**Mat Ryer:** I'm interested then what's next, what's coming up, and how do you keep your finger on the pulse of what's going on? + +**Nayana Shetty:** I often think less about tools and more about the capabilities that we really need in an organization. It could be anything from like what do we need in terms of system infrastructure side of things, or like the topic for today, more around observability. Around observability I often think the capabilities that we need are logging metrics. An organization can invest in having multiple tools for the same thing, or it could be one tool that does all of it. It depends on the kind of organization you are. + +\[51:22\] I have often leaned towards what's happening in the DevOps communities or like in the monitoring communities to actually get insights from them, saying "Oh, there's--" I think two years ago was when I was introduced to Loki, which is the logging tool, and I got super-excited about this, mainly because we were using another logging tool within the organization which was super-expensive. And like "Do we use this super-expensive tool which has some bells and whistles which we don't use, or can we go with something like Loki?" And it is finding out capabilities that you care about, and looking at what is happening in that particular market. + +Within the monitoring space, I think my preference within logging would be -- like, if you are in the AWS line, then something like AWS CloudWatch, or Loki, Splunk... These are a couple of tools that I have used in the logging space. You can use the same kind of tools for metrics as well, but there are better tools for metrics. Prometheus is really good. Or Graphite, which again, I have spent quite a lot of my career in Graphite, so I probably have a preference in this space, and see what innovation is happening in the Graphite space... But yeah, Prometheus is just probably another one which is really good in the time-series database side of things. + +And then it's also to do with your metrics aggregation. So you have all of these different metrics and logs and everything that you're collecting, but how do you visualize them together? So you need something around the visualization layer, which is where Grafana or Kibana, one of these kind of tools would come in handy. + +And finally, I think it's the whole -- you're doing all of this because you want alerting. So what do we have in the alerting space? Thinking about "Do I use Slack notifications? Do I use email notifications? SMSes? Pager duty?", whatever. And it's just making sure you understand the capabilities of what you're trying to solve, and finding core products in each of those areas. And it could be the same product that solves everything, or it could be different products. And yeah, I tend to lean towards communities and conferences to actually figure out what is hot in the market in places. + +**Matt Toback:** Yeah. And what I really like about this is -- that's all kind of like through the lens -- or at least the way you think about it as like what are the problems that you're trying to solve for the customer, and what is the value that you're trying to provide. So even at that point, all of these could become sort of like -- it's not interchangeable, but you can solve the problem in 400 different ways. And I just really like that you started -- it feels like you started with that approach, to say "What do you actually need to do? What do you need to protect?" and then figure out how to do it. And I like that. That feels like the most transferable skill between company to company. + +**Nayana Shetty:** Yeah. And also within teams. When there are so many different teams, and every team is autonomous to use their own tools... I think you need the core principles to be the same, irrespective of what tools they're using, and that's where I find having the capabilities and the principle layer set right would help teams figure out what is the best tool for their use cases and stuff. + +**Mat Ryer:** That's amazing. Loads of great practical advice there. And you yourself have spoken at conferences on this subject as well. We'll put one in the show notes for people interested, but you can also google - or use any search engine; other search engines are available. You can duckduckgo and find -- just for your name, and you'll... Yeah, Nayana Shetty. + +Well, unfortunately that's all the time we have, I'm afraid... Thank you so much - Matt Toback was here... Weren't you, Matt? + +**Matt Toback:** I was. Is there anything that you wanna say to your future self, just before we go? + +**Nayana Shetty:** Oh, I love that. I think it is "You don't have to solve all problems today. There are things that you can leave for the future to solve." \[laughs\] That is what I would say to my future self. + +**Mat Ryer:** \[55:33\] Hm. It'd be more useful if we could send messages to our past selves though, because we could tell them what the stocks are gonna do... + +**Matt Toback:** No... Because we know what happens, right? The space-time continuum... It's just not good. + +**Mat Ryer:** Biff gets it doesn't he? Okay. We won't do it then. We won't do it then, Matt. Okay. Well, yes, thank you so much. + +**Nayana Shetty:** I've enjoyed it. This is a topic that I enjoy speaking in general, so... + +**Mat Ryer:** Are you speaking at any other conferences coming up? Or will you? + +**Nayana Shetty:** No. Because I'm new to the organization, I'm just stepping away from speaking, so that I can gather some insights from the organization before I start speaking... So yeah, none. + +**Mat Ryer:** But the videos of your past self are still available, of course, so I do recommend people check them out... + +**Nayana Shetty:** Yeah. + +**Mat Ryer:** Thank you so much to Matt Toback and our special guest, Nayana Shetty. I've been Mat Ryer (still am), and thank you very much for listening. See you next time on Grafana's Big Tent. + +**Outro:** \[56:37\] + +**Mat Ryer:** Have you learned how to play the riff? + +**Matt Toback:** I have not yet. But I have time. + +**Mat Ryer:** Okay. You should learn it. + +**Matt Toback:** On the airplane to Whistler I'm gonna take the piece out. My seatmate won't be terribly happy with that, but... + +**Mat Ryer:** "Please put your seatbacks and basses away. We're coming in to land." + +**Matt Toback:** You could do upright in the seat, maybe... + +**Mat Ryer:** You'd definitely get a ticket of like a double-bass. But you could also do Seinfeld links with it. You could make a little joke, a zinger to the staff... + +**Matt Toback:** Ba dom ba dowm bum bum + +**Mat Ryer:** Just do a bit of bass... That would be good, wouldn't it? \[laughter\] + +**Nayana Shetty:** Yeah, that would be nice. \[laughter\] diff --git "a/Observability in the wild\357\274\232 strategies that work_transcript.txt" "b/Observability in the wild\357\274\232 strategies that work_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..2037ab0d63fbf94a7a83f3b60780755f83a85817 --- /dev/null +++ "b/Observability in the wild\357\274\232 strategies that work_transcript.txt" @@ -0,0 +1,848 @@ +[0.00 --> 2.78] Hey everyone, Jared Santo here, GoTime's producer. +[3.24 --> 8.40] This episode is a bit different than what you're used to, but hopefully it feels familiar as well. +[8.74 --> 13.44] We are helping our friends at Grafana Labs produce a podcast all about observability. +[13.76 --> 19.56] It's called Big Tent and it's hosted by Matt Toback, Tom Wilkie, and GoTime's very own Matt Reier. +[20.16 --> 24.64] So, today on GoTime we're featuring episode 6 of Grafana's Big Tent, +[24.64 --> 27.64] where both the Matts talk with Nayana Shetty from Lego Group +[27.64 --> 30.18] all about observability strategies that work. +[30.78 --> 33.94] GoTime returns to its regularly scheduled programming next week +[33.94 --> 39.36] when Ron Evans travels back from 2053 to warn Matt and Natalie all about Go's future. +[39.98 --> 42.18] Okay, Grafana's Big Tent. Here we go. +[46.76 --> 49.50] This episode is brought to you by Chronosphere. +[49.50 --> 54.86] When it comes to observability, teams need a reliable, scalable, and efficient solution +[54.86 --> 57.46] so they can know about issues well before their customers do. +[57.64 --> 61.08] They need a solution that helps them move faster than the competition. +[61.64 --> 65.14] And companies born in the cloud-native era often start with Prometheus for monitoring, +[65.42 --> 67.16] which is obviously an amazing piece of software, +[67.54 --> 70.30] but they quickly push it to its limits and often outgrow it. +[70.50 --> 74.08] They run into issues with siloed data, missing long-term storage, +[74.38 --> 77.34] and wasted engineering time firefighting the monitoring system +[77.34 --> 79.64] versus delivering their application with confidence. +[79.64 --> 82.62] They describe the system as a house of cards, +[82.62 --> 87.50] where a single developer's seemingly benign change can overload the whole monitoring system, +[87.60 --> 91.80] or they say they're flying blind because they pride themselves on making data-driven decisions, +[92.24 --> 95.46] but losing visibility means they lose this competitive edge. +[95.78 --> 99.40] Ryan Sokol, VP of Engineering at DoorDash, has this to say about Chronosphere. +[99.76 --> 100.00] Quote, +[100.00 --> 108.20] Quote, +[108.20 --> 109.36] End quote. +[109.74 --> 114.06] Chronosphere is the observability platform for cloud-native teams operating at scale. +[114.50 --> 117.06] Learn more and get a demo at chronosphere.io. +[117.06 --> 118.06] Again, +[118.40 --> 119.86] chronosphere.io. +[128.26 --> 129.42] Let's do it. +[129.92 --> 130.72] It's go time. +[142.14 --> 142.86] Hello, +[143.34 --> 146.00] and welcome to Grafana's Big Tent, +[146.00 --> 151.92] the podcast all about the people, community, tools, and tech around observability. +[152.68 --> 156.18] I'm joined today by, it's only Matt Toback. +[156.24 --> 156.76] Hello, Matt. +[157.12 --> 157.60] Hi, Matt. +[157.82 --> 158.46] What do you mean, only? +[158.80 --> 160.50] Yeah, no, it's just understated, isn't it? +[160.60 --> 163.16] Just, I can't believe it's you, really, in a lot of ways. +[163.38 --> 165.00] You could say, it's how you say it. +[165.02 --> 165.64] It's not what you say it. +[165.68 --> 166.76] You say, I can't believe it's you. +[166.96 --> 167.58] You're here. +[167.72 --> 168.08] Oh, right. +[168.14 --> 168.34] Okay. +[168.40 --> 170.26] I can't believe it's, I can't believe it's you. +[170.40 --> 171.34] And you're like, you're here. +[171.80 --> 172.46] You're here. +[173.02 --> 173.24] Yeah. +[173.24 --> 173.92] Well, don't worry. +[174.00 --> 175.06] It's not just me and you. +[175.06 --> 177.44] That would be obviously tiresome for both of us. +[177.58 --> 180.94] We have a special guest joining us today. +[181.12 --> 182.92] That would be tiresome for both of us. +[183.12 --> 183.46] Thank you. +[183.52 --> 184.14] It's how you say it. +[184.14 --> 184.48] You're right. +[186.56 --> 190.30] Today, we're joined by Nayana Shetty from Lego. +[190.44 --> 191.00] Hello, Nayana. +[191.38 --> 192.00] Hi, Matt. +[192.18 --> 192.64] I'm mad. +[194.24 --> 196.06] Yes, just one would suffice. +[196.14 --> 196.66] We'll share it. +[196.74 --> 197.38] Happy to share. +[197.38 --> 203.64] And you are a principal engineer who loves talking about SRE in microservices, right? +[204.42 --> 204.68] Yes. +[204.78 --> 205.02] Yeah. +[205.44 --> 209.36] I think over the years, I've been in teams where we've built microservices. +[210.02 --> 218.22] And it's when you scale up and have hundreds of microservices, how do you then make them reliable and keep them reliable? +[218.22 --> 220.54] That's what I'm interested in. +[220.54 --> 228.02] And yeah, it's been so I was working in the financial times where we had all of these hundreds of microservices. +[228.48 --> 229.78] And how do we manage it there? +[229.78 --> 233.54] And now I've moved to the Lego group where we're going through massive digital transformation. +[233.54 --> 238.02] And here it's like, we want to build these hundreds of microservices. +[238.40 --> 241.00] So should we care about reliability now? +[241.08 --> 245.46] Or can we think about it in like 10 years time when we have these microservices? +[245.90 --> 251.26] So yeah, that is my context to how site reliability, microservices, all of this fit together. +[251.78 --> 254.38] Yeah, because it did used to be a kind of afterthought, really, didn't it? +[254.42 --> 257.66] Which is why I think SRE, I think it's short for sorry, right? +[257.66 --> 260.32] That's one way of looking at it. +[260.38 --> 267.42] Or like, I always think of it like, sorry, I don't understand why people don't think about site reliability in the first instance. +[267.42 --> 273.66] Or sorry, I don't understand why people would build this in such a way that it is like half broken. +[273.66 --> 276.96] Or like, you don't think about the future of this product. +[277.10 --> 283.22] And you are very close to reinventing the wheel every few months if you went in that direction. +[283.34 --> 286.56] So yeah, SRE is probably one of the ways of looking at it. +[287.66 --> 297.32] Just even on a personal note, I'm excited that Naina is here and joining us because we met in 2018 for the first time in an attic in Amsterdam. +[297.90 --> 300.12] Which when said that way, doesn't feel weird at all, right? +[301.00 --> 307.76] I mean, we were talking loads of monitoring Grafana and graphite and all of those things. +[307.88 --> 309.74] So yeah, attic didn't make a difference then. +[309.74 --> 315.70] But it was wild because we did, we planned GrafanaCon EU in 2018, +[315.70 --> 325.06] which was technically our third GrafanaCon, but was kind of our biggest up to that point and the most what felt well produced. +[325.20 --> 326.56] And it was the middle of an Arctic chill. +[326.70 --> 327.30] Do you remember that? +[327.52 --> 328.84] That the canals had frozen over? +[329.50 --> 329.72] Yes. +[329.82 --> 330.00] Yeah. +[330.06 --> 332.68] I had struggles getting back home. +[332.94 --> 337.78] And like, it took me two hours where it should have probably taken me only like half an hour to get home. +[338.02 --> 339.70] So yeah, I completely remember that. +[339.70 --> 344.26] And we were in this stage where Tom had just joined the company. +[344.42 --> 346.42] We announced it on stage, the acquisition of Causal. +[346.60 --> 348.00] The entire company fit on stage. +[348.08 --> 348.92] It was 25 people. +[349.34 --> 352.56] Like when we said kind of goodbye at the end, there's still this photo that circulates here, +[352.64 --> 354.56] where everyone is just kind of shoulder to shoulder. +[354.94 --> 357.30] And the stage is the whole company. +[357.78 --> 362.66] But I do remember, you stand out to me, us being up there kind of in that breakout room, +[362.90 --> 365.14] talking about what you were trying to do at the Financial Times. +[365.14 --> 370.72] And it does, it feels like you kind of continue in this natural progression, in this natural journey. +[371.42 --> 374.82] And like when you think back to you then, how did you see the world? +[375.64 --> 382.42] At that point, we were investigating, like we had quite a lot of monitoring tools at the Financial Times. +[382.68 --> 387.86] And I was like, and I was working in the team that provided monitoring as a service to other teams. +[387.86 --> 400.58] And my head was going mad thinking, okay, how do I as a team with like four or five engineers be able to support like these 20, 30 odd engineering teams who all want monitoring? +[401.36 --> 412.40] And they're using from Nagios to Zabig, some Graphite, some started, I think there were very few installations of Permitius at that point. +[412.40 --> 416.06] And I was like, how do we get all of these different use cases together? +[416.06 --> 419.44] And how do we get them on a platform which could work together? +[419.60 --> 422.64] And it made me, like I was worried at that point. +[422.64 --> 426.62] And like three years later or four years later, looking at it, it's like- +[426.62 --> 427.48] You're still worried. +[427.64 --> 427.74] No. +[428.06 --> 430.60] I mean, I have moved on from the Financial Times. +[430.70 --> 434.58] So I'm less worried about the Financial Times monitoring systems. +[434.58 --> 436.86] But I still worry about like the same use case. +[436.86 --> 442.74] I see it here as well in the Lego group where there's different monitoring tools that we've got across the organization. +[442.74 --> 444.62] And it's how do we get them all together? +[444.62 --> 451.86] And like, how do we say a single story that everyone could understand rather than every single team trying to solve the same problem? +[451.98 --> 458.92] So it's still very similar, but probably we have better tools and like processes in place that can help us. +[459.00 --> 459.78] So that's how I see it. +[460.12 --> 460.60] Yeah. +[460.78 --> 465.80] Something you said earlier stood out, this idea that you're like, why did you build it like this? +[465.80 --> 469.98] If only you'd built it differently, we'd be in a much better position now. +[470.26 --> 471.84] So it kind of like- +[471.84 --> 473.64] If you only did it right, is that what you're saying? +[474.90 --> 475.34] Yeah. +[475.38 --> 476.34] But that's the question. +[476.34 --> 479.50] Like, when should we start caring about this stuff? +[479.64 --> 482.98] When should we start worrying about how are we going to operate this? +[484.02 --> 489.82] I think this kind of relates to like how I've moved in the journey in my career and stuff. +[489.82 --> 500.64] So I started off as a test engineer, just doing some manual testing, then moved on to doing more QA, like more quality related things rather than just testing. +[500.64 --> 517.02] And over the years, I've seen the transition in a lot of organization as well, where they've moved to this shift left and test early, like release as small as possible and continuous iterations and stuff. +[517.02 --> 523.44] So all of this, I think, kind of leads to that point of like, how do you make your future better? +[523.84 --> 528.42] And one of the quotes I have often used is being kind to your future self. +[528.56 --> 531.96] Like, how can you make your life easy in the future? +[532.08 --> 535.32] So think about that today when you're building whatever you're building. +[535.32 --> 546.22] And that comes with if you're building a new product, think about do you even have to build it or like can you just look at what's in the market and reuse it? +[546.30 --> 549.82] If it's a non-differentiating thing, then why build it? +[549.94 --> 553.14] If it is a differentiating thing, yes, put your heart and soul into it. +[553.14 --> 562.68] But then when you're doing it, make sure you think about the sustainability aspects of your product and not just today what the customer would get. +[562.68 --> 575.80] Yeah. And it's like I've often used this carrot and stick kind of approach in teams to show the benefits of what you could get out of thinking about monitoring observability from the front. +[575.80 --> 579.46] And usually the carrots are like you build it in the right way. +[579.56 --> 584.30] Then you don't have you can actually forget about your systems because they will take care of themselves. +[584.30 --> 598.96] And the stick approach is often if you didn't do it, then you have to go into the rotors or like all of those other things that comes with like making your systems more observable and keeping it sustained once it's up and running and stuff. +[598.96 --> 604.82] So, yeah, I think that's what I've used in the past to actually help teams nudge in that direction and stuff. +[605.00 --> 605.24] So, yeah. +[605.24 --> 613.16] Yeah. Yeah. I mean, you know, I would be kind to my future self, but I'm too busy dealing with all the stuff that my past self left me to do. +[613.68 --> 617.50] So I don't know. But yeah, I think so that's the thing. +[617.58 --> 626.54] If you think about how it's going to be, where it's going to be running, like the realities of that, if you think about that, the earlier, the better almost, isn't it? +[626.54 --> 639.68] So, yeah, it is that. And it's also that, yes, you are fixing things from yesterday, but if you don't fix it and leave some goodies along with it, then you're fixing tomorrow. +[639.68 --> 642.90] You're fixing today's problem. So you're still in that vicious cycle. +[643.34 --> 649.90] So to get away from that vicious cycle, I think you need to actually step back sometimes and put that extra effort. +[649.90 --> 657.90] I remember one of the tech principles we had in the Financial Times for FD.com was... +[657.90 --> 659.98] I've forgotten this. +[661.04 --> 662.32] It's okay. No, it's okay. +[662.70 --> 670.10] As we're talking about, you know, like the past self and future self, is it okay that I've completely forgotten about observability and now I'm just on a personal journey? +[671.46 --> 677.86] I'm thinking about all the decisions made and yet to make and how to provide goodies for everyone. +[677.86 --> 681.44] So, Naina, or for me, I don't know. I forget that part. +[681.66 --> 690.86] But, Naina, the carrot and stick, can I ask you, right, is there been a stick that you've seen people try to use that just didn't work, right? +[690.92 --> 700.10] Or not didn't work, but either was too harsh or like just kind of like ill, you know, not ill-intentioned, but ill-executed? +[700.10 --> 706.48] I think it's about the motivations and the motivation factor behind doing something. +[707.10 --> 708.84] And that's how I saw the carrot and stick. +[709.22 --> 717.90] So, the carrots were the motivation factors that we were providing to teams saying, if you did something right and if you thought about like how do you monitor something, +[717.90 --> 726.70] how do you add alerts in place and how do you make sure it's auto-recoverable where possible and stuff, then you don't have to worry about it, right? +[726.80 --> 734.80] So, that is more of the motivation for the team to be like, I don't, like I can be really proud about what I build and I don't have to think about it again. +[735.08 --> 744.92] But at the same time, we know that every team has these deadlines to meet and like there are product owners who would have their own feature sets to build. +[744.92 --> 756.50] So, it's that kind of scenarios where you actually still need the stick to help the teams be like, look, I mean, yes, we understand your pressures, but this is more important as well. +[756.88 --> 763.76] And I did remember the quote and it was slow down to speed up, which actually like that was one of the tech principles we worked with. +[763.76 --> 776.74] So, basically, yes, you can go at 100 miles per hour today, but then if you don't build it in such a way that you have put those measures in place, then tomorrow you have to break and stop. +[776.92 --> 783.46] But if you slow down and went at, say, 60 miles per hour, you're there for the long run and you would go on longer. +[783.58 --> 786.60] So, that's how I would see some of this now. +[786.60 --> 788.32] Yeah, that is so true. +[788.58 --> 790.90] We actually built a little project before Grafana. +[791.04 --> 803.74] We built a project management tool called Pace and it was trying to get across that exact thing, which is that you feel great going at a thousand miles an hour, but you do have, you know, there's important things to do along the way. +[803.90 --> 806.42] And it's hard to retrofit a lot of this stuff. +[806.72 --> 809.52] So, it's, you know, thinking about it up front sometimes can save you. +[809.52 --> 812.98] It's a bit like how you design for failure as well. +[813.28 --> 819.20] Like, you know, in the perfect world, your system, all the messages flow perfectly and there's no problems. +[819.38 --> 821.94] But in reality, it's way more messy. +[822.10 --> 822.86] Things fail. +[823.32 --> 829.78] And so, that idempotency and things come into play where you may design expecting this is going to fail. +[830.88 --> 836.70] I write Go code and Go has error handling as a kind of explicit feature. +[836.70 --> 840.54] They're values that are just returned as the second argument to functions and things like this. +[841.14 --> 845.96] And that frustrates a lot of people because, you know, they're used to exceptions or something that's just sort of automatic. +[846.40 --> 851.24] But it forces you to think about what's going to happen if this thing fails. +[851.70 --> 853.98] And that's great discipline to get into. +[854.68 --> 858.10] And I think it's a myth to think that your system won't fail. +[858.54 --> 858.76] Yeah. +[858.92 --> 862.42] Like, always build your system in such a way that it will fail. +[862.72 --> 864.66] If it doesn't, then you have a problem. +[864.66 --> 865.30] Yeah. +[866.10 --> 869.32] So, make sure you add those checks in place. +[869.52 --> 872.98] So, when it fails, it can smoothly recover and all of those. +[873.80 --> 874.04] Yeah. +[874.36 --> 874.56] Yeah. +[874.56 --> 882.40] And I know some companies that have that as part of the proper sort of testing approach is they'll literally things will break on purpose. +[882.60 --> 885.32] And, you know, it's a first class concern that they have. +[885.32 --> 890.20] And it is that thing of, yeah, don't, you know, of course, like, I don't know. +[890.26 --> 894.40] Is it just ego that people think I'm so good, I'll write this, it's going to be great? +[894.60 --> 895.22] What's going on? +[895.32 --> 896.20] It can't be. +[896.52 --> 897.68] Honestly, it can't be, right? +[897.70 --> 901.10] Like, we've all known and experienced it enough. +[901.34 --> 901.60] Yeah. +[901.72 --> 902.16] Do you think? +[902.16 --> 902.68] I don't know. +[902.72 --> 911.40] But the thing is, when I'm writing code and it doesn't work, it's shocking how quickly I'm like, there's something wrong with the processor. +[912.32 --> 914.92] The processor is not working. +[915.06 --> 916.12] Or physics has changed. +[916.56 --> 919.58] That's why I'll go to physics has changed before it's my fault. +[920.06 --> 922.68] But it turns out I just did a capital letter where I shouldn't have. +[922.68 --> 929.64] I think, like, I've been in teams where they do pairing and, like, those mobbing sessions and stuff. +[929.88 --> 938.20] And, like, they have kind of helped in, like, sense-checking people's egos a bit and be like, I'm not the best. +[939.00 --> 944.70] And when two people talk about it, I think it does help think, okay, there is a reality that we live in. +[944.86 --> 946.84] And this is what you need to consider. +[946.84 --> 962.72] Is there anything that you would, like, even in that, like you were saying before, that progression between, like, being in, you know, manual testing and then QA and then moving to SRE, was there any, was there, like, a moment where it clicked, where people started, I mean, they just started incorporating testing into the code, right? +[962.76 --> 971.88] Do you see the same progression happening in observability to where there'll be some moment and some kind of click where it just becomes part of it as opposed to the separate thing that happens afterwards? +[971.88 --> 974.34] I have seen it work in some teams. +[974.54 --> 978.02] And, like, a lot of teams I've worked in are all autonomous teams. +[978.16 --> 982.68] So, they can basically build however they want using whatever technologies they want. +[983.12 --> 995.58] What is often have teams like that is having some sort of guardrails, which actually says, and, like, also being aware that not all applications need the same level of checks and monitoring and all of these. +[995.58 --> 999.36] So, being aware that, okay, there is a level of criticality of my app. +[999.54 --> 1003.44] And if it is a highly critical app, then let me put all of the things in. +[1003.84 --> 1010.60] And if it is a less critical app, in that case, you would just have, like, maybe just a simple health check. +[1010.70 --> 1011.60] That would be good enough. +[1011.68 --> 1013.50] You don't need to go all board. +[1013.50 --> 1022.90] And one thing that we had when I was working for FT was you always have at least a basic check on all your apps. +[1023.24 --> 1028.56] Otherwise, like, we used to get this service operability score for our applications. +[1028.88 --> 1033.72] And the score used to go down as in when we didn't have some of these things in place. +[1034.08 --> 1040.96] And that was, like, a nice measure where people thought about it from the beginning rather than as an afterthought about some of these. +[1040.96 --> 1049.94] But what could happen in this kind of scenarios is people go all in and they just say, oh, I'm going to monitor everything, have all my logs go in. +[1050.10 --> 1052.46] Like, you don't need to go all board on this. +[1052.56 --> 1055.14] There's a limit to how much you need to monitor as well. +[1055.30 --> 1063.00] And understanding the criticality of your app and then building your observability around that is probably something that teams should think about. +[1063.00 --> 1073.86] How would you, if a team was listening to this, right, and they were trying to understand the criticality of the app and make decisions around it, like, how would you, if you had them sitting in the room, how would you explain it and say, here, start here, do this? +[1073.86 --> 1076.56] I think it depends on the business criticality. +[1076.84 --> 1086.24] And if it is a highly business critical application, which means if it went down for, say, more than 15 minutes, then we wouldn't be in business. +[1086.24 --> 1100.46] If it's that kind of app, then you need to have your alerting in place, monitoring, like, the right level of logging in place, which actually gives us any of the audit records that actually show us what's happened with the applications. +[1101.32 --> 1104.10] And then any sort of health check. +[1104.16 --> 1108.08] So there's probably, like, two levels of monitoring that we should think about. +[1108.18 --> 1109.76] One is the application level monitoring. +[1109.76 --> 1111.96] And then there is the system level monitoring. +[1112.16 --> 1120.04] So being able to figure out where the problem is soon enough is something very critical when it's a 15 minutes recovery thing. +[1120.30 --> 1132.34] But if it is an application that's less critical, then maybe just having the application level monitoring is good enough where you could take longer to actually investigate, look into the locks and actually figure out where the problem area is and stuff. +[1132.34 --> 1137.66] So I would suggest teams to think about, like, how critical their app is. +[1137.76 --> 1143.50] And that is something the business should, like, help them with, not something that the team just decides, oh, this is the most critical thing. +[1143.92 --> 1155.60] And once you know the business criticality of something, then it is coming up with some sort of check saying if it is a highly critical system, then we do both application as well as system monitoring. +[1156.08 --> 1159.12] Otherwise, just one of them based on your use cases and stuff. +[1159.12 --> 1166.40] And, like, in the past, I've spoken about, like, the use method and red method that we could use for these kind of things. +[1166.68 --> 1171.02] Like, I prefer use and red method over the Google's four signals. +[1171.18 --> 1176.20] It depends on what your team's needs are and what fits into your use cases. +[1176.84 --> 1183.22] So you would use a red method, which is rate, error, and duration for every single application that you build. +[1183.22 --> 1188.72] And it's very easy to see that in a microservice world where you have different kinds of applications. +[1189.38 --> 1192.84] You have the same three parameters that you're measuring across all of them. +[1192.96 --> 1199.62] So it actually helps the team analyze irrespective of if that belongs to your team or any other team. +[1199.70 --> 1201.10] You just know where the problem is. +[1201.46 --> 1203.82] And the same with systems side of things. +[1203.82 --> 1208.40] You would go with, like, the use method, which is utilization, saturation, and errors. +[1208.86 --> 1213.86] And you would do this for the CPU, disk, or network, and all of those different areas. +[1214.06 --> 1216.82] And you basically know where the problem is. +[1216.86 --> 1218.06] And it's easy to find out. +[1218.40 --> 1220.24] I would say it is hard. +[1220.44 --> 1221.68] It takes time. +[1221.98 --> 1226.96] So invest based on how much returns you would get on these when you put these checks in. +[1226.96 --> 1233.96] So that is something the team should be mindful about when they are investing in monitoring or, like, learning and stuff. +[1234.88 --> 1239.90] Is the primary counterbalance, in your mind, the effort that it takes to keep this monitored well? +[1240.00 --> 1240.88] Or is it also cost? +[1241.06 --> 1243.88] Do you think about the cost to operate or the back end? +[1244.52 --> 1245.74] It is the cost. +[1246.00 --> 1248.72] And at the end of the day, it should be the cost to the business. +[1248.72 --> 1253.22] As in, how much does having the system down cost us? +[1253.70 --> 1260.80] And you basically work backwards from there, saying, if this was down for 15 minutes, it would cost the business so much. +[1260.88 --> 1271.34] Which means we, as a team, should be investing more time in actually getting the right amount of measures so we can solve the problem or narrow down the problem quickly. +[1271.34 --> 1278.98] And, like, I would always focus on the business value rather than the team's individual product value and stuff. +[1279.44 --> 1287.52] But, yeah, it depends on, like, if you were an internal system, like, in one of the teams I was in, we were building monitoring tools for other teams. +[1287.72 --> 1291.68] So we don't have real business value as such as our team. +[1291.88 --> 1296.76] But we were supporting teams that had, like, really high value systems. +[1296.76 --> 1304.68] So that kind of meant that we had to think about the application level as well as system level monitoring on our systems and stuff. +[1308.98 --> 1312.98] This episode is brought to you by our friends at FireHydrant. +[1313.26 --> 1316.04] FireHydrant is the reliability platform for every developer. +[1316.46 --> 1320.24] Incidents, they impact everyone, not just SREs. +[1320.40 --> 1325.98] They give teams the tools to maintain service catalogs, respond to incidents, communicate through status pages, +[1325.98 --> 1328.14] and learn with retrospectives. +[1328.50 --> 1333.90] What would normally be manual error-prone tasks across the entire spectrum are responding to an incident. +[1334.18 --> 1337.36] They can all be automated in every way with FireHydrant. +[1337.66 --> 1342.74] They have incident tooling to manage incidents of any type with any severity with consistency. +[1343.28 --> 1346.42] Declare and mitigate incidents all from inside Slack. +[1346.80 --> 1349.82] Service catalogs allow service owners to improve operational maturity +[1349.82 --> 1353.12] and document all your deploys in your service catalog. +[1353.12 --> 1359.36] Incident analytics allow you to extract meaningful insights about your reliability over any facet of your incident +[1359.36 --> 1361.08] or the people who respond to them. +[1361.48 --> 1365.40] And at the heart of it all, incident runbooks, they let you create custom automation rules, +[1365.64 --> 1370.86] convert manual tasks into automated, reliable, repeatable sequences that run when you want. +[1371.22 --> 1375.24] You can create Slack channels, Jira tickets, Zoom bridges instantly after declaring an incident. +[1375.74 --> 1378.32] Now your processes can be consistent and automatic. +[1378.32 --> 1380.46] The next step is to try it free. +[1380.60 --> 1384.96] Small teams, up to 10 people, can get started for free with all FireHydrant features included. +[1385.30 --> 1386.70] No credit card is required. +[1387.16 --> 1389.30] Get started at firehydrant.io. +[1389.68 --> 1391.62] Again, firehydrant.io. +[1391.62 --> 1406.90] I love that advice of pay attention to the value you're going to get from the effort that you put in. +[1407.34 --> 1409.20] When I, I like mono repos. +[1409.34 --> 1410.76] I just like to put that out there. +[1410.82 --> 1411.70] I love mono repos. +[1411.70 --> 1417.60] And the reason I like them is because you can have a pull request that has a unit test, some back-end code, +[1417.72 --> 1423.36] maybe some API changes, front-end code in there too, hopefully with some front-end tests maybe. +[1423.74 --> 1427.26] And it's nice that that all gets applied to the system in one go. +[1427.60 --> 1432.08] Does that also apply to like this sort of field or the instrumentation of that? +[1432.20 --> 1438.20] Should we be having those kinds of conversations at that point so that we kind of think about it as we go? +[1438.20 --> 1442.96] I would love to say yes, but I've not seen a team do it really well. +[1443.22 --> 1450.84] So I can see the challenges of like when you have this mono repo and everyone's contributing to the same central repository, +[1451.22 --> 1458.54] there is a challenge that the parameters that you would think about for your product and your monitoring systems +[1458.54 --> 1462.00] might be different to what another team would be looking at. +[1462.00 --> 1469.38] So there is a challenge with how do you then look at this as a product that we sell to customers? +[1469.72 --> 1475.14] Like you have to think about capability monitoring maybe rather than your individual product monitoring +[1475.14 --> 1479.52] where you're thinking about what is the capability that I'm providing to the customer? +[1480.18 --> 1484.14] And those should probably be things that we have at a central level. +[1484.28 --> 1487.24] And we do it as in when we add new features, +[1487.24 --> 1491.34] we make sure we don't break the monitoring that we've got across the capability. +[1491.34 --> 1497.80] But on a single individual product team's perspective, yeah, I don't know how much value it would add. +[1497.90 --> 1499.12] So it depends on that, I guess. +[1499.40 --> 1500.76] So yeah, I'm sure. +[1501.14 --> 1504.26] Have you seen it work in your teams or something? +[1504.90 --> 1509.30] Well, we have at least the conversation when there's a PR for like a big feature, +[1509.92 --> 1514.56] we will chat about it and say like, what do we need? +[1514.64 --> 1515.64] What do we need from this? +[1515.64 --> 1519.18] It's like, what's going on here that later we're going to need? +[1519.38 --> 1521.56] And it's that thing about becoming your future selves. +[1522.20 --> 1527.12] So yeah, but I don't know that we've got that right yet or anything, you know, +[1527.20 --> 1531.54] because in a way we don't really know what's important upfront necessarily. +[1531.82 --> 1534.32] So it's, you know, but sometimes you do. +[1534.38 --> 1538.64] And I like that there are guidelines that we can follow to give us a good foundation. +[1538.64 --> 1541.96] And then, of course, we're going to have to fine tune it depending on our particular case. +[1542.36 --> 1542.70] I agree. +[1543.02 --> 1543.18] Yeah. +[1543.92 --> 1547.44] All this, to me, starts to distill down into, right? +[1547.50 --> 1551.66] Like it's some amount of like, if you are doing the centralized monitoring, right? +[1551.66 --> 1553.06] Or there's a level of that, right? +[1553.10 --> 1557.72] And then you have to communicate this down to these teams and you have to get them to buy in, right? +[1558.56 --> 1559.88] You know, what do you do that? +[1559.96 --> 1562.58] Or even how would you suggest someone else do that well? +[1562.58 --> 1570.34] So central teams pushing things is like, irrespective of it being monitoring or anything in general is really hard. +[1570.84 --> 1579.58] And it should always be driven by, like what I've seen work really well is the ones that are driven by like value add to the individual teams itself. +[1580.14 --> 1590.14] So as an example, when we were building this Amazon Linux, like a base plate image that everyone could apply and they can run their own EC2 instances. +[1590.14 --> 1599.66] When they had this, what we said we will do as part of it is we said you're going to get monitoring to like, I think we were pushing logs to Splunk in that case. +[1599.74 --> 1601.72] So you would get that feature for free. +[1601.86 --> 1603.60] You would get authentication for free. +[1603.76 --> 1610.74] You would get like, have those kind of things that you will get for free as part of whatever features that you would give. +[1610.96 --> 1616.34] And that has often been a nice way to drive teams to be like, oh, yes, I like that. +[1616.52 --> 1617.32] And I will do it. +[1617.48 --> 1617.52] Right. +[1617.52 --> 1621.68] Make it so easy that they can't, like they would rather adopt it rather than try and do it themselves. +[1621.94 --> 1622.78] Yeah, exactly. +[1622.98 --> 1632.98] So like another example that came to mind was we had this central repository for like a CRM system, which we had to enter all of our system information. +[1632.98 --> 1642.88] And basically it was like a, because we had so many microservices, we had like a central system where we could go and query for any particular system with something called as a system code. +[1643.08 --> 1649.46] And we would know if that system was live, was active, who was working on it, all of those information. +[1649.46 --> 1660.58] And what we did when we built this was we said, if you put the right information in this, then you would automatically have a dashboard that would show up only your teams monitoring in it. +[1660.74 --> 1666.56] That was like an incentive for teams to be like, oh, if I did this, then I get my own dashboard. +[1666.76 --> 1667.30] Let me do that. +[1667.30 --> 1673.82] So I think it's that showing intensive value beyond just what you want them to achieve out of it. +[1673.94 --> 1676.64] That's how I've seen it work really well in teams. +[1676.94 --> 1683.70] So yeah, you need to have some sort of courage to actually get people to move towards your solutions and stuff. +[1683.92 --> 1684.02] Yeah. +[1684.20 --> 1684.40] Yeah. +[1684.40 --> 1684.44] Yeah. +[1685.06 --> 1685.34] Yeah. +[1685.34 --> 1685.82] That's great. +[1685.92 --> 1687.94] I think that applies to everything. +[1690.02 --> 1690.42] Yeah. +[1690.48 --> 1693.56] If you make it easy and it's sort of a no brainer. +[1693.56 --> 1705.72] And like one example of that is where we can, like, if we've got APIs, we can just instrument on those, on the endpoints very easily in a simple way, usually with some middleware or something in the code. +[1705.88 --> 1708.02] And there's lots of packages that do this. +[1708.32 --> 1714.28] So I do quite like, yeah, I think, and I think there's probably space for more things like that. +[1715.02 --> 1717.28] More of that for devs. +[1717.48 --> 1718.04] Yeah. +[1718.18 --> 1730.16] And I think it's the, like, I would really, like, want to see central teams be more mindful about this because as a central team, you're building these amazing tools. +[1730.76 --> 1737.66] And at the end of it, like, you kind of think, oh, if I put a documentation together and self-service, everyone's going to come and use it. +[1737.66 --> 1748.26] But then each individual product teams have their own little agendas to work towards and their own, like, the product initiatives to their own OKRs, all of those things. +[1748.26 --> 1761.72] So this is like an extra bit of, like, cognitive load onto those teams, which they can avoid if you were to do a lot more promotion within teams saying, if you did this, you would get a lot of benefits. +[1761.72 --> 1765.72] And it will take some of the risks that you have taken on yourself. +[1765.94 --> 1769.84] And also it's that education piece of you care about your product. +[1770.04 --> 1772.08] We will help you care about your product. +[1772.20 --> 1774.06] That's something to think about. +[1774.06 --> 1775.24] Yeah. +[1776.96 --> 1783.84] So what are some common mistakes that we make when we're trying to do this with the best intentions in the world? +[1783.96 --> 1792.18] We want to do this properly, but are there any things you see that people misunderstand or common mistakes, common gotchas that you've seen? +[1792.78 --> 1803.16] I think it's knowing how much is enough is, like, one of the things that I've often seen where there are teams who just put the basic thing available because it's there in a checklist somewhere. +[1803.16 --> 1807.30] And then they move on, which is probably not the best for your product. +[1807.30 --> 1817.56] So it's being aware of the value of your product and, like, what is the life cycle that your product or the journey that your product is going on? +[1817.66 --> 1819.30] That is probably one of the things. +[1819.64 --> 1828.08] The other thing that I have seen and I've struggled a lot with is, like I mentioned about this use method and red method to actually build your dashboards. +[1828.34 --> 1832.94] It's very hard to get, like, your network-related monitoring right. +[1832.94 --> 1836.24] And your saturation for networks. +[1836.40 --> 1837.72] Like, how do you do that? +[1838.12 --> 1852.84] And get the wrong set of, like, I've seen myself having a wrong set of dashboards and alerting and wondering why this is going off every time something happens when it shouldn't have and stuff. +[1852.84 --> 1869.62] So I think it's just, like, being okay to experiment and, like, continuously tinker your monitoring and alerting as you go along is probably something that teams should be conscious that, like, it's not that you build it once and then it's there. +[1869.70 --> 1870.34] It's there forever. +[1870.72 --> 1873.94] But there is a continuous evolution that happens with your monitoring. +[1873.94 --> 1877.18] Like how your feature sets go through that cycle. +[1877.52 --> 1881.76] You have to do the same with your observability side of things as well. +[1882.34 --> 1882.72] Yeah, yeah. +[1883.02 --> 1884.52] Matt, can I answer too? +[1885.16 --> 1886.04] Let me just check. +[1886.74 --> 1886.96] No. +[1887.66 --> 1888.54] Oh, come on. +[1888.84 --> 1891.70] Brutal, but please, I'd love to hear what you think. +[1891.70 --> 1897.54] So, but, Niana, you were talking about the value, like, derived, right, and focusing on that for the customers. +[1897.76 --> 1903.44] That I do think that's a common gotcha where you build all these tools and you're like, we did it. +[1903.52 --> 1904.00] We did it. +[1904.04 --> 1904.78] Like, it's all there. +[1904.92 --> 1906.12] All you have to do is this, right? +[1906.16 --> 1911.82] And I think the common gotcha is forgetting that you need to deliver something that someone could just adopt easily. +[1911.90 --> 1916.54] Like you said, like, it is a version of, I was thinking, like, car parts, right? +[1916.54 --> 1921.86] And then, like, or Legos, I guess, but, like, dropping off, like, a collection of car parts and being like, there you go. +[1922.20 --> 1925.30] And you're like, you know, like, I want to drive, right? +[1925.30 --> 1929.02] Like, I don't, like, I get that I can get there, but you haven't helped me really at all. +[1929.38 --> 1932.58] And there's, you know, and you call a Lyft and that's where the metaphor, I think, breaks. +[1932.88 --> 1936.26] But I do think there's some version of that too, right? +[1936.28 --> 1945.94] Like stopping short of actually delivering the value to the person consuming it as opposed to just dropping a collection of pieces that can work, but they have to do the last mile. +[1945.94 --> 1951.30] Yeah, well, in a way, what helps that definitely is going to be this, you build it, you run it. +[1951.44 --> 1958.92] You know, we're not throwing this thing over the wall to someone, for someone else to operate, which I know that actually lots of, lots of people do still do that. +[1959.14 --> 1960.58] And there's a disconnect. +[1961.02 --> 1965.98] When you are yourselves kind of running it, you're the customer of that data. +[1966.20 --> 1973.22] So a bit like when you're dogfooding software, if you write, if you're building dev tools, like we do at Grafana, we dogfood a lot. +[1973.22 --> 1975.46] Like we'll use our tools a lot internally. +[1975.46 --> 1981.96] That's how they're so good, frankly, like, because they've been, you know, it's not like we're imagining the user of this. +[1982.32 --> 1983.46] We are the user of it. +[1983.50 --> 1985.40] And I think that makes a big difference, doesn't it? +[1985.80 --> 1986.00] Yeah. +[1986.22 --> 1994.30] And also, I think like one of the comments I've heard a few people say about is build your code in such a way that you can debug it at three in the morning. +[1994.30 --> 1997.16] I mean, it doesn't mean that you have to do it every day. +[1997.36 --> 2003.78] But if it breaks at a time that you're not fully in focus, you still can get to it easily. +[2004.06 --> 2009.58] And that is something which I think, yeah, people should be thinking about while building the products and stuff. +[2009.58 --> 2012.72] That's such a great point, I think. +[2012.96 --> 2017.02] And that leads me on to this next question, which is around like drills. +[2017.40 --> 2024.38] Do we should we be doing like drills at 3 a.m. and living that experience to see what it's like? +[2024.52 --> 2028.08] Three o'clock is probably taking the mickey out of people if you were doing drills. +[2028.08 --> 2031.24] Do people do drills? +[2031.72 --> 2032.68] I guess they do. +[2033.10 --> 2033.24] Right. +[2033.32 --> 2034.04] But it's probably not. +[2034.12 --> 2035.36] It's not common, is it? +[2035.82 --> 2037.00] I have seen it done. +[2037.96 --> 2041.96] And I think it's a very artificial environment where the drills happen. +[2041.96 --> 2047.10] So one of the things that we did when I was at the FT was we had this incident drills. +[2047.64 --> 2051.80] So basically, you emulate an incident and then you go about with the team. +[2052.18 --> 2055.28] How do you go about actually figuring out where the problem is? +[2055.28 --> 2062.38] So you start with like which alert it was and then look at the traces and then look at what the logs were. +[2062.52 --> 2064.22] And you go through the whole cycle of it. +[2064.54 --> 2070.80] It was a way to like ease the whole out of our support that we had within the organization. +[2071.34 --> 2071.70] Yeah. +[2071.76 --> 2075.24] But at the same time, there were a lot of people who were not very keen of this. +[2075.82 --> 2080.26] Because it's an artificial environment, people felt like that is not reality. +[2080.44 --> 2081.64] So why do it? +[2081.70 --> 2083.02] That's because you didn't do it 3 a.m. +[2083.02 --> 2085.34] Yeah, maybe that. +[2085.86 --> 2089.14] I think there's like a touch of maturity in actually embracing drills. +[2089.24 --> 2090.84] Like whether or not it's artificial, right? +[2090.86 --> 2093.12] Like it's that idea like, oh, this is artificial. +[2093.32 --> 2093.84] This is dumb. +[2094.10 --> 2094.82] You know, we don't want to do this. +[2094.84 --> 2095.92] It's not going to be like this in real life. +[2096.04 --> 2103.20] And then I think like you think about any kind of, I don't know, like either team environment or any kind of like practice that you need to do. +[2103.20 --> 2105.64] Because it's more than just debugging the code, right? +[2105.66 --> 2107.56] It's like everything is interconnected, right? +[2107.56 --> 2111.74] And you want to be able to do some of these things more than once. +[2111.80 --> 2114.06] So that way every time doesn't feel like you're the first time on stage. +[2114.70 --> 2121.76] And it does, it feels like you just want to be like, hey, like what's the right analogy to make if you want to convince someone to actually practice? +[2122.52 --> 2123.18] I don't know. +[2123.18 --> 2124.22] Yeah, definitely. +[2124.46 --> 2125.64] It's not the same. +[2125.84 --> 2128.26] It's not the same because you know it's a drill. +[2128.60 --> 2135.46] Like unless you're doing something where you literally, you break something and it's not really broken or maybe it is. +[2135.54 --> 2138.26] And you're, you know, doing something kind of, that seems a bit extreme. +[2138.46 --> 2140.86] But you are, it is going to feel different. +[2141.04 --> 2144.68] But that still doesn't mean that there's not plenty of stuff to practice. +[2144.68 --> 2151.50] And, you know, like practicing your, when you practice driving, you know, there's an instructor next to you watching everything. +[2151.64 --> 2155.10] That's a very strange situation to be in. +[2155.24 --> 2161.04] But, but you still are like, you still move the steering wheel and do the, I don't drive, but there's a gear stick. +[2161.18 --> 2161.52] I know that. +[2161.74 --> 2162.68] And the horn. +[2164.02 --> 2165.28] You press the horn to go. +[2165.54 --> 2165.94] Yeah. +[2166.06 --> 2166.76] Horn to go. +[2166.82 --> 2167.92] And then you leave that on. +[2168.00 --> 2169.02] So everyone knows you're there. +[2169.10 --> 2171.88] I do because they need to get out of the way. +[2171.88 --> 2176.58] I think there's also like value in like this. +[2176.84 --> 2179.56] The other way of looking at drills is like shadowing. +[2179.76 --> 2185.36] And when there's an actual incident, not, not having just one or two people involved in it. +[2185.44 --> 2192.36] Yes, it might be the most critical thing, but having more people just listen in and see what's happening and like just be there. +[2192.70 --> 2198.02] Sometimes helps them understand, oh, this is how I would go about solving this, looking at those people. +[2198.02 --> 2203.58] So, yeah, I think it's, it's a mixture of drills and a shadowing maybe that could work in teams. +[2204.06 --> 2207.94] I even think yesterday, and I realized why I'm all fired up about this. +[2208.14 --> 2209.10] I parsed through it. +[2210.02 --> 2216.36] Yesterday we did, I visited the sales team and they were doing these workshops and they were doing radical candor, right? +[2216.36 --> 2221.40] Which is like all about like feedback and giving feedback and getting feedback and, and being able to, to do it well. +[2221.40 --> 2224.06] And then you break out and you're like, okay, yeah, check, check, check. +[2224.12 --> 2224.42] I get it. +[2224.42 --> 2224.78] I get it. +[2224.84 --> 2225.10] I get it. +[2225.12 --> 2226.06] Like I could totally do that. +[2226.22 --> 2229.88] And then you break out into these triads and then you practice it, right? +[2229.92 --> 2233.16] And there's a part of you that goes like, oh, I don't, I don't need to do this. +[2233.16 --> 2233.92] Like I get it. +[2233.92 --> 2234.52] I get the concepts. +[2234.58 --> 2242.16] And then you try and do it and you're like, and you kind of feel yourself like places that, you know, are a little bit creaky or, you know, maybe you don't quite get it as much. +[2242.16 --> 2246.30] So I think, I think it's actually where I'm fired up to where, even if it is artificial, right? +[2246.34 --> 2253.08] Some of those, some of those joints might be, you know, either rusty or creaky or, or don't articulate well until, and you don't realize that until you do it. +[2253.36 --> 2255.34] I think that was a rubbish point. +[2256.32 --> 2257.50] We can cut that. +[2257.64 --> 2258.28] Cut, cut, cut. +[2259.22 --> 2264.82] I think it's also like a good exercise to do just to test your like documentation and stuff. +[2264.82 --> 2276.28] And like, if it's, if your documentation is up to scratch and like, when you've written something you've written with, with good intent, but when someone's actually following it, does it make sense? +[2276.28 --> 2279.52] Is something that the drills can actually capture and stuff. +[2279.80 --> 2283.14] So yeah, there's more than one benefit of having drills. +[2288.50 --> 2291.34] This episode is brought to you by Honeycomb. +[2291.50 --> 2293.28] Find your most perplexing application issues. +[2293.28 --> 2301.06] Honeycomb is a fast analysis tool that reveals the truth about every aspect of your application in production. +[2301.52 --> 2305.52] Find out how users experience your code in complex and unpredictable environments. +[2305.80 --> 2310.72] Find patterns and outliers across billions of rows of data and definitively solve your problems. +[2311.16 --> 2312.64] And we use Honeycomb here at Change. +[2312.68 --> 2316.50] Well, that's why we welcome the opportunity to add them as one of our infrastructure partners. +[2316.50 --> 2324.34] In particular, we use Honeycomb to track down CDN issues recently, which we talked about at length on the Kaizen edition of the Ship It podcast. +[2324.58 --> 2325.28] So check that out. +[2325.50 --> 2325.98] Here's the thing. +[2326.22 --> 2329.48] Teams who don't use Honeycomb are forced to find the needle in the haystack. +[2329.60 --> 2332.76] They scroll through endless dashboards playing whack-a-mole. +[2332.98 --> 2336.02] They deal with alert floods, trying to guess which one matters. +[2336.02 --> 2341.62] And they go from tool to tool to tool playing sleuth, trying to figure out how all the puzzle pieces fit together. +[2342.02 --> 2348.28] It's this context switching and tool sprawl that are slowly killing teams' effectiveness and ultimately hindering their business. +[2348.68 --> 2355.44] With Honeycomb, you get a fast, unified, and clear understanding of the one thing driving your business. +[2355.70 --> 2356.12] Production. +[2356.66 --> 2359.10] With Honeycomb, you guess less and you know more. +[2359.10 --> 2364.70] Join the swarm and try Honeycomb free today at honeycomb.io slash changelog. +[2364.86 --> 2368.34] Again, honeycomb.io slash changelog. +[2368.78 --> 2375.30] And by Acuity, a new platform that brings fully managed Argo CD and enterprise services to the cloud or on-premise. +[2375.62 --> 2380.48] The platform is a versatile Kubernetes operator for handling cluster deployments the GitOps way. +[2380.80 --> 2384.32] And I'm here with Kelsey Hightower, angel investor and advisor to Acuity. +[2384.82 --> 2388.54] Kelsey, why are you excited about Argo CD and what's happening here with Acuity? +[2388.54 --> 2395.62] When I think about Argo CD, it represents the transition from traditional CICD. +[2395.78 --> 2399.82] You know, you have a big server with a built-in workflow engine. +[2400.28 --> 2406.36] And you can only do what that system can do, whether it's Jenkins, whether it's Spinnaker, you name it. +[2406.62 --> 2408.76] Those things tend to be all-in solutions. +[2409.08 --> 2414.78] And they're all predicated on having like their own built-in workflows, UIs, and ways of doing things. +[2414.78 --> 2427.06] And then when I think about kind of the Argo CD, that whole open source movement kind of backed by the ideas we saw in the Kubernetes world, which was each of those steps is nothing more than just a step in a workflow. +[2427.36 --> 2431.54] And after 10, 20 years of doing CICD, how best to represent those steps? +[2431.54 --> 2439.54] And it turns out this whole container thing is probably the best way to have little snippets of logic sit at each of those steps in the workflow. +[2439.54 --> 2443.46] And then you can kind of exchange them and share them to build any pipeline you want. +[2443.72 --> 2447.98] So the way to look at this is Kubernetes has never had a workflow engine or tool. +[2447.98 --> 2457.94] And so when you think about kind of Argo workflow or Argo CD, which is kind of a specialized workflow, kind of attacking the how do you roll out software problem, that's the way I would think about it. +[2457.98 --> 2463.82] So if you're all in on Kube and you like the Kubernetes ecosystem, then you kind of have a choice of workload types. +[2463.96 --> 2467.18] And I would probably just say it's another workload type you can put in your toolbox. +[2467.18 --> 2474.58] So if you've got something that can benefit from a workflow engine and reuse the logic that you already have in containers, it kind of feels like the perfect fit. +[2475.00 --> 2475.56] The perfect fit. +[2475.64 --> 2475.92] All right. +[2475.98 --> 2476.46] Thanks, Kelsey. +[2476.46 --> 2480.60] Well, the next step is to head to Acuity.io slash changelog. +[2480.68 --> 2483.30] They are inviting all of our listeners to join the closed beta. +[2483.84 --> 2486.32] Again, Acuity.io slash changelog. +[2486.46 --> 2487.94] Links are in the show notes. +[2499.38 --> 2506.24] Niana, you mentioned earlier, like this idea that, you know, if you do too much, you can overdo it. +[2506.24 --> 2510.14] And end up with basically alert fatigue, just alerts going off. +[2510.20 --> 2512.26] What do we mean really by alert fatigue? +[2512.46 --> 2515.18] I'm going to give an example so people can relate to it. +[2515.18 --> 2522.22] I was in one of the teams where we used to get close to 1500 alerts on a weekly basis. +[2522.22 --> 2528.44] And this was like we had around 80 odd microservices. +[2528.44 --> 2530.62] So it wasn't like just one microservice or anything. +[2531.12 --> 2534.30] But then my team was like three people looking at this. +[2534.78 --> 2543.58] And it's at that point which you realize that are they actually looking at this thing or is it all just being ignored as like just noise? +[2543.58 --> 2544.72] Let's just ignore it. +[2544.86 --> 2553.20] And I think it's that point where you start ignoring your alerts is where you've gone to that stage where you can't take any more alerts. +[2553.32 --> 2555.92] So you're fatigued with the whole alerting itself. +[2555.92 --> 2565.32] And I think it's better to have less alerts for the most important things rather than have too many and try to just like filter it out. +[2565.46 --> 2575.14] One of the exercises we did when we had these alert fatigues and like thousand odd alerts is we consciously stopped some of the alerts to see who will start shouting. +[2575.14 --> 2581.30] And it happened that more than 50% of these alerts when we turned off, no one actually shouted at us. +[2581.38 --> 2583.32] So it was like, was that even important? +[2583.68 --> 2593.44] Going through that exercise on a regular basis where you see if you're ignoring more than at least 10% of your alerts, then go and do something about it. +[2593.64 --> 2595.74] Maybe turn them off and no one will care. +[2595.74 --> 2610.20] And I think teams need to be conscious that it's okay to miss a faulty alert compared to missing out on a real alert, which would have cost us like millions of pounds or whatever it is. +[2610.40 --> 2622.94] So I think it's being careful to put the right alerts in and stopping at that and not just going overboard with, oh, let's take an example that we have Grafana in our systems and we have alerting with Grafana. +[2622.94 --> 2630.60] I have this tool so I can put as many alerts as possible, not going wild with it, but actually like knowing where to stop. +[2630.84 --> 2633.30] That's how I would describe this whole alert fatigue. +[2633.52 --> 2636.20] And it's with time, it does happen with teams. +[2636.20 --> 2644.50] So it's worth going back and auditing them and making sure you keep them clean as much as possible. +[2645.44 --> 2650.26] I wonder if you could do a, like what would be like the equivalent of a bug bounty for alerts? +[2650.26 --> 2655.76] Like how do you incent people to go and clean those up and celebrate being like, they're gone? +[2656.24 --> 2657.32] Oh, that's hard. +[2657.54 --> 2662.36] What I've done in this is actually gone and turned them off myself and be like, let's see who's going to shout. +[2663.74 --> 2667.30] And like when no one shouts, you know that they're not important enough. +[2667.40 --> 2672.10] So that is something that I've done, but I don't know how you would, interesting. +[2672.24 --> 2673.16] Do you have any ideas? +[2673.54 --> 2677.46] We need some analytics, don't we, on the usage of it really? +[2677.46 --> 2684.74] Then we can say, you know, no one's looked at these alerts for ages, you know, or you could put a specific time on it if you want. +[2684.84 --> 2686.82] I don't know, I don't want to design the application now. +[2686.94 --> 2688.16] But yeah, something like that. +[2688.22 --> 2695.18] I mean, I like the idea that you should go back and look at them and pay attention to whether you still need them and things. +[2695.24 --> 2703.62] This is a little bit like how in GitHub, like, or in your project management tool, if you have loads of stuff in there, most of it's just getting ignored. +[2703.62 --> 2714.76] And in a way, it creates this also this idea that you're so far away from being done, which we of course are, but you don't, you know, it sort of reinforces that. +[2714.90 --> 2718.88] So it is that thing of if there's just so much there, it stops being useful. +[2719.20 --> 2725.18] I like the idea that, does it take experience though, do you think, to know what's useful, what's not? +[2725.18 --> 2725.90] Does that? +[2726.14 --> 2727.00] It could do. +[2727.24 --> 2732.88] I mean, the more you see these things, you will realize where it's important, where it's useful and where not. +[2733.18 --> 2742.88] And this is where I've seen some of the junior engineers and teams struggle, which is like, they start worrying about every single alert that comes on Slack or whichever is your preferred tool. +[2742.98 --> 2744.36] And they're like, oh, what do I do? +[2744.44 --> 2745.14] I've got this alert. +[2745.26 --> 2749.30] It might be my change, but it might not be related to your change at all. +[2749.30 --> 2758.54] So I think it's something that the team should do on a regular basis as like a team activity or something like that, where they sanity check their alerts. +[2758.98 --> 2765.24] One way we used to do is any alert that we actually did anything with, we started putting some everything. +[2765.68 --> 2767.00] All of our alerts used to come to Slack. +[2767.10 --> 2769.14] So we used to start putting some emojis on it. +[2769.26 --> 2772.68] So we know which of them were actually things that mattered. +[2772.68 --> 2778.14] And like on a weekly basis, we were like, oh, there were 10 of these, which we did nothing with. +[2778.54 --> 2779.72] So maybe we can get rid of it. +[2779.88 --> 2785.18] So that like, it's very hard to get that feedback cycle on alerts I've found. +[2785.56 --> 2788.30] Matt, you need to collect the emojis and then feed it back. +[2788.42 --> 2790.52] Is the API going the other way on emojis? +[2790.78 --> 2796.44] We do that for in the Grafana incident tool, but I need to tell the on-call team about that idea. +[2796.52 --> 2797.22] That's such a good idea. +[2797.38 --> 2798.18] That is kind of fun. +[2798.28 --> 2802.16] You could collect that data and literally, yeah, you then, oh, hello. +[2802.68 --> 2804.24] You're like, what does dancing penguin mean? +[2804.56 --> 2804.86] Yeah. +[2806.24 --> 2806.64] Yeah. +[2806.90 --> 2807.82] It means it's cool. +[2809.76 --> 2811.90] Who owns that idea legally? +[2812.04 --> 2812.52] I don't know. +[2812.98 --> 2813.62] I do. +[2813.62 --> 2814.00] I feel like. +[2814.32 --> 2814.42] Yeah. +[2817.28 --> 2818.14] That's a good idea. +[2818.26 --> 2818.66] Solved. +[2822.34 --> 2827.02] I want to, I wanted to quickly touch on as you progress in your career, right? +[2827.04 --> 2829.76] Often you're going to walk into new organizations, right? +[2829.78 --> 2832.62] And you're, you're new to Lego and everything today is just reinvigorating. +[2832.68 --> 2835.86] And for us, the thought that we had this, or that I was kind of noodling on this morning, +[2835.86 --> 2837.90] which is all these things are true. +[2837.90 --> 2841.34] All these methods are like, are kind of proven. +[2841.76 --> 2846.42] And in, you know, in some ways, like it has nothing to do with the technology and everything +[2846.42 --> 2848.30] to do with the landscape that you're walking into. +[2848.38 --> 2851.86] And then, then you have to figure out how, how and what you introduce. +[2851.86 --> 2857.82] And I guess I'm curious, like how much about open source tooling makes it easier to transfer +[2857.82 --> 2858.78] into a new organization? +[2859.48 --> 2864.72] And even just how much, like, how do you approach going into a new org, having this experience, +[2864.72 --> 2867.84] but then also not understanding how everything fits together? +[2867.84 --> 2868.56] Hmm. +[2869.06 --> 2870.46] I mean, this is so relatable. +[2871.02 --> 2872.42] I'm going through this now. +[2872.52 --> 2872.66] Yeah. +[2872.66 --> 2877.26] Like, given I've been in the Lego group for only three months and I care about monitoring +[2877.26 --> 2880.42] and like in general sustainability of products quite a lot. +[2880.42 --> 2886.70] But I've been looking at different teams doing this and thinking, okay, this team has this +[2886.70 --> 2888.28] Grafana dashboard to do it. +[2888.36 --> 2891.06] This other team has Neuralic and they're doing something with it. +[2891.22 --> 2896.86] I think for me, I was lucky that I was in a team that was building monitoring tools as +[2896.86 --> 2901.24] a service, like providing monitoring tools as a service to other teams. +[2901.24 --> 2907.92] So for me, it was like easier to catch on to what is happening in different areas within +[2907.92 --> 2908.64] the Lego group. +[2908.64 --> 2915.54] But I think what I fall back to is always think about what are the core aspects of monitoring. +[2916.00 --> 2920.42] So it's things like logging, metrics, alerting, tracing. +[2920.76 --> 2923.20] So notifications, some of those core things. +[2923.48 --> 2928.68] And it is looking at those aspects and thinking, how is the team solving these problems? +[2929.28 --> 2936.08] And where the team have done, used a tool, I have often just endorsed what they've got and +[2936.08 --> 2941.86] looked into it, but where they haven't, I have often suggested open source tools in those +[2941.86 --> 2944.28] use cases because two reasons. +[2944.42 --> 2946.98] One, it's easy to get started and get going with it. +[2947.08 --> 2951.28] You don't need any licensing and all of those kinds of challenges that come with a proprietary +[2951.28 --> 2951.76] tool. +[2951.76 --> 2958.18] And on the other side, there's a lot of community that can help you getting started with the +[2958.18 --> 2958.76] tool as well. +[2958.76 --> 2965.34] So I think those are reasons why I would prefer, like when suggesting to teams, I would prefer +[2965.34 --> 2968.16] open source technologies when it comes to this space. +[2968.16 --> 2975.36] I mean, as an example, when I was doing some experimentation for my own personal project, I could have gone +[2975.36 --> 2978.84] with one of the tools that was already available in the organization when I was working. +[2979.00 --> 2983.76] But then I was like, I mean, if I left this organization, I can't take that tool with me. +[2984.14 --> 2987.64] So it's better to have it on more open source tools. +[2987.92 --> 2991.84] I mean, in that case, I used, I think it was Graphite and Grafana that I used in that case. +[2991.84 --> 2998.58] But it is that while there is transferable skills within the organization, like as an +[2998.58 --> 3006.06] example, the Lego technology, I think we have around 200 or 250 odd teams in it. +[3006.70 --> 3013.82] And if these people within the teams have to move between each other, speaking the common +[3013.82 --> 3015.26] language is quite important. +[3015.68 --> 3020.64] And having that community outside of the Lego group who can help us with this is quite +[3020.64 --> 3021.08] important. +[3021.08 --> 3025.74] And I feel like that is where the power of using open source technologies comes from. +[3026.10 --> 3030.34] And I mean, I have come from an organization where we were a very big advocate of open source +[3030.34 --> 3030.84] technology. +[3031.08 --> 3035.96] So I probably would be singing the song of let's go all in on open source. +[3037.84 --> 3040.76] So I'm interested then what's next? +[3041.02 --> 3044.90] What's coming up and how do you keep your finger on the pulse of what's going on? +[3044.90 --> 3051.94] I often think about like less about tools and more about the capabilities that we really +[3051.94 --> 3053.30] need within an organization. +[3054.16 --> 3059.14] And like it could be anything from like, what do we need in terms of system infrastructure +[3059.14 --> 3060.00] side of things? +[3060.00 --> 3063.56] Or like the topic for today, more around observability. +[3063.56 --> 3070.30] So around observability, like I often think the capabilities that we need are logging metrics. +[3070.30 --> 3076.40] And like an organization can invest in having multiple tools for the same thing. +[3076.56 --> 3079.94] Or it could be one tool that does all of it. +[3080.06 --> 3082.60] It depends on the kind of organization you are. +[3082.60 --> 3091.42] So I have often leaned towards like what's happening in like the DevOps communities or like in the +[3091.42 --> 3098.56] monitoring communities to actually get insights from them saying, oh, there's I think two years +[3098.56 --> 3103.50] ago was when I was introduced to Loki, which is the logging tool. +[3103.68 --> 3110.58] And I got super excited about this mainly because we were using another logging tool within the +[3110.58 --> 3112.50] organization, which was super expensive. +[3112.60 --> 3118.74] And like, do we use this super expensive tool, which has some belts and visits, which we don't +[3118.74 --> 3119.12] use? +[3119.22 --> 3120.92] Or can we go with something like Loki? +[3121.42 --> 3127.28] And it is finding out capabilities that you care about and looking at what is happening in +[3127.28 --> 3129.48] that particular market and stuff. +[3129.74 --> 3136.34] Within the monitoring space, logging, I think my preference within logging would be like if +[3136.34 --> 3142.36] you are in the AWS land, then something like AWS CloudWatch or like Loki, Splunk. +[3142.36 --> 3146.66] These are a couple of tools that I have used in the logging space. +[3146.82 --> 3149.70] You can use the same kind of tools for metrics as well. +[3150.04 --> 3156.92] But there are better tools for metrics like Pometheus is really good or Graphite, which again, like I have +[3156.92 --> 3159.74] spent quite a lot of my career in Graphite. +[3159.74 --> 3164.76] So I probably have a preference in this space and see what innovation is happening in the +[3164.76 --> 3165.70] Graphite space. +[3165.88 --> 3170.44] But yeah, Pometheus is probably another one which is really good in the time series database +[3170.44 --> 3171.52] side of things. +[3172.02 --> 3175.36] And then it's also to do with like your metrics aggregation. +[3175.44 --> 3179.60] So you have all of these different metrics and logs and everything that you're collecting, +[3179.60 --> 3181.78] but how do you visualize them together? +[3181.78 --> 3186.76] So you need something around the visualization layer, which is where like Grafana or like +[3186.76 --> 3189.66] Kibana, one of these kinds of tools would come in handy and stuff. +[3190.14 --> 3194.20] And finally, I think it's the whole, you're doing all of this because you want to alert +[3194.20 --> 3194.48] things. +[3194.64 --> 3196.72] So what do we have in the alerting space? +[3196.80 --> 3200.22] Thinking about like, do I use like Slack notifications? +[3200.22 --> 3204.92] Do I use email notifications, SMSs, page of duty, whatever. +[3205.52 --> 3210.46] And it's just making sure you understand the capabilities of what you're trying to solve +[3210.46 --> 3213.24] and finding core products in each of those areas. +[3213.34 --> 3216.92] And it could be the same product that solves everything or it could be different products. +[3217.12 --> 3223.08] And yeah, I tend to lean towards communities and conferences to actually figure out what is +[3223.08 --> 3225.14] hot in the market and in places. +[3225.98 --> 3226.10] Yeah. +[3226.10 --> 3230.52] And what I really like about this is like, that's all kind of like through the lens of +[3230.52 --> 3233.50] release the way that you think about it as like, what are the problems that you're trying to solve +[3233.50 --> 3235.22] for the, for the customer? +[3235.36 --> 3235.50] Right. +[3235.52 --> 3237.12] And then what is the value that you're trying to provide? +[3237.16 --> 3241.88] So even at that point, like all of these could become sort of like, it's not interchangeable, +[3241.88 --> 3244.58] but you can solve the problem in 400 different ways. +[3244.66 --> 3244.96] Right. +[3245.04 --> 3248.52] And, and I just really liked that you start with, it feels like you start with that approach +[3248.52 --> 3250.34] to say like, what do you actually need to do? +[3250.64 --> 3251.80] What do you need to protect? +[3252.28 --> 3254.12] And then figure out how to do it. +[3254.30 --> 3255.48] And I like that. +[3255.48 --> 3259.66] That feels like the most transferable skill between company to company. +[3259.90 --> 3260.02] Yeah. +[3260.14 --> 3265.22] And also like within teams, when there are so many different teams and every team's autonomous +[3265.22 --> 3270.20] to use their own tools and stuff, then like, I think you need the core principles to be +[3270.20 --> 3272.94] the same irrespective of what tools they're using. +[3272.94 --> 3278.94] And that's where I find having the capabilities and the principle layer set, right, would help +[3278.94 --> 3282.70] teams figure out what is the best tool for their use cases and stuff. +[3282.70 --> 3283.94] Well, that's amazing. +[3284.12 --> 3286.16] Loads of great practical advice there. +[3286.24 --> 3289.80] And you yourself have spoken at conferences on this subject as well. +[3289.94 --> 3293.78] I noticed that people can, we'll put, put one in the show notes for people interested, +[3294.14 --> 3296.58] but you can also Google or use any search engine. +[3296.68 --> 3297.78] Other search engines are available. +[3298.50 --> 3304.16] You can duck, duck, go and find just for your name and you'll, yeah. +[3304.16 --> 3304.94] Nayana Shetty. +[3305.44 --> 3308.20] Well, unfortunately that is the time. +[3308.38 --> 3310.60] That's all the time we have, I'm afraid. +[3311.16 --> 3312.32] Thank you so much. +[3312.94 --> 3314.24] Matt Toback was here, weren't you, Matt? +[3314.64 --> 3315.32] I was. +[3315.36 --> 3318.54] Is there anything that you want to say to your future self just before we go? +[3319.06 --> 3320.12] Oh, I love that. +[3320.12 --> 3325.00] I think it is, you don't have to solve all problems today. +[3325.18 --> 3329.32] There are things that you can leave for the future to solve. +[3330.32 --> 3333.12] That is what I would say to my future self. +[3334.94 --> 3339.14] It'd be more useful if we could send messages to our past selves though, because we could +[3339.14 --> 3341.44] like tell them what the stocks are going to do in that. +[3341.66 --> 3344.18] No, because we, we know what happens, right? +[3344.32 --> 3345.42] The space-time continuum. +[3346.18 --> 3347.02] It's just not good. +[3347.16 --> 3348.02] Biff gets it, doesn't he? +[3348.30 --> 3349.04] Biff, he does. +[3350.12 --> 3350.56] Okay. +[3351.70 --> 3352.40] We won't do it then. +[3353.12 --> 3354.22] We won't do it then, Matt. +[3354.38 --> 3355.40] You've, yeah. +[3356.08 --> 3356.44] Okay. +[3356.96 --> 3358.40] Well, yes. +[3358.40 --> 3359.20] Thank you so much. +[3360.26 --> 3361.14] I've enjoyed it. +[3361.24 --> 3364.60] I mean, this is a topic that I enjoy speaking in general. +[3364.84 --> 3365.28] So yeah. +[3365.60 --> 3368.50] Are you speaking at any other conferences coming up or will you? +[3369.06 --> 3371.30] No, because I'm new to the organization. +[3371.30 --> 3376.88] I'm just like stepping away from speaking so that I can gather some insights from the organization +[3376.88 --> 3378.02] before I start speaking. +[3378.64 --> 3379.30] So yeah, none. +[3379.30 --> 3382.78] But the videos of your past self are still available, of course. +[3383.02 --> 3384.76] I do recommend people check them out. +[3385.36 --> 3385.50] Yeah. +[3385.88 --> 3389.62] Thank you so much to Matt Toback and our special guest, Nayana Shetty. +[3389.94 --> 3391.30] I've been Matt Raya, still am. +[3391.74 --> 3393.40] And thank you very much for listening. +[3393.54 --> 3396.24] We'll see you next time on Grafana's Big Tent. +[3396.24 --> 3410.30] Have you learned that? +[3410.42 --> 3411.96] Have you learned how to play the riff? +[3412.20 --> 3413.18] I have not yet. +[3413.32 --> 3413.90] But I have time. +[3414.24 --> 3415.56] You should let it. +[3415.56 --> 3417.66] I'm going to on the airplane to Whistler. +[3417.80 --> 3418.92] I'm going to take the bass out. +[3420.18 --> 3422.58] My seatmate won't be terribly happy with that. +[3422.94 --> 3425.14] Please put your seat backs and basses away. +[3425.58 --> 3426.58] We are coming into land. +[3427.28 --> 3429.00] You could do upright in a seat, maybe. +[3429.40 --> 3429.84] Oh, yeah. +[3429.84 --> 3432.34] You'd have to get a ticket for like a double bass. +[3432.34 --> 3437.40] But you could also do like Seinfeld links with it. +[3437.48 --> 3441.86] It could be like make a little joke, a zinger to the staff. +[3443.80 --> 3445.08] Just do a bit of bass. +[3445.66 --> 3446.48] That would be good, wouldn't it? +[3446.58 --> 3447.52] Yeah, that would be nice. +[3452.62 --> 3458.22] If you enjoyed this conversation, maybe subscribe at BigTent.fm. +[3458.36 --> 3461.60] And if you're a long-time listener of GoTime, share the show with a friend. +[3461.60 --> 3464.52] It's the best way to pay it forward and spread the love. +[3464.90 --> 3467.72] Also, don't forget to check out Changelog++. +[3468.14 --> 3470.42] That's our membership program with a bunch of perks. +[3470.84 --> 3471.78] Support our work. +[3472.00 --> 3473.04] Make the ads disappear. +[3473.50 --> 3476.56] And hey, we just added a free pack of stickers to everyone who joins. +[3476.92 --> 3479.84] Learn more at changelog.com slash plus plus. +[3480.08 --> 3482.18] Thanks again to Fastly for CD-ing for us. +[3482.42 --> 3485.40] To the mysterious Breakmaster Cylinder for the always fresh beats. +[3485.58 --> 3486.50] And to you for listening. +[3486.80 --> 3487.52] We appreciate you. +[3487.80 --> 3491.22] Stay tuned for Dead Program's time travel-inspired guest appearance. +[3491.22 --> 3493.94] It's coming up next time on GoTime. +[3493.94 --> 3497.26] Game on. diff --git a/Revisiting Caddy_transcript.txt b/Revisiting Caddy_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..d2a52eac07c672b10cb723f098e745110099daf6 --- /dev/null +++ b/Revisiting Caddy_transcript.txt @@ -0,0 +1,317 @@ +**Natalie Pistunovich:** It's a great evening here in Berlin, it's a finally nice, warm weather, and I'm very happy to welcome my co-host, Jon. Hi! + +**Jon Calhoun:** Hi, Natalie. How are you? + +**Natalie Pistunovich:** Good! I hope you're also doing well. + +**Jon Calhoun:** A little tired, but I think that's to be expected with a newborn in the house. + +**Natalie Pistunovich:** Yeah, congratulations for that. It's crazy. + +**Jon Calhoun:** Oh, thanks. + +**Natalie Pistunovich:** We are joined today by a crew of Caddy people, Matt and Mohammed. So Matt is the author of the Caddy web server, and you're working on it full-time, relying completely on sponsorships. And Mohammed is a product manager by day and a Gopher student by night, and also the creator of Caddy SSH. Hi! Where are you joining us from? + +**Mohammed S. Al Sahaf:** I am joining from Saudi Arabia. + +**Natalie Pistunovich:** It must be super-late for you. + +**Mohammed S. Al Sahaf:** Yup! + +**Matt Holt:** And I'm joining you from Utah. + +**Natalie Pistunovich:** \[03:56\] So we're kind of like across so many different timezones... Always fun to have those shows. The benefit of the internet. So tell us a little bit about yourselves, and your intro question will be not an animal that starts with the first letter of your name, but WHEN did you start using Go? + +**Matt Holt:** Mohammed, you go first. + +**Mohammed S. Al Sahaf:** Okay. I started playing around with Go sometime in 2014. In 2013 I was doing a program of networks and system administration and I was chatting with a friend who's a software engineer, and he kept telling me there is a new language called Go by Google... And it returns the errors, so you're forced to deal with it, and you never forget it. + +In 2014 I was doing a web dev job as my full-time. We were a Java shop, and I was so much annoyed by Java that I decided to look into this new language my friend has been bugging me about. I got hooked since then. + +**Matt Holt:** Very cool. I started writing Go probably about ten years ago now. Ten or eleven years ago, right around Go 1. Because our company was a .NET shop, and we were looking at easier to deploy, faster alternatives, and just kind of a more fun language overall. So then I started using it, and we wrote a couple products in it, and then I started using it for personal projects, including a web server... So that's how that all started. + +**Natalie Pistunovich:** What was the web server that you were using at work back at that time? + +**Matt Holt:** It was a mix of NGINX or Apache, whichever had better copy-paste examples... I could never figure Apache out at least. + +**Natalie Pistunovich:** Yeah. But also, NGINX was not very straightforward without those examples, I would say. At least for me. + +**Jon Calhoun:** My experience with both of those has been you figure out enough to get whatever you need done done, and then you forget about it in six months, so then you have to figure it all out again. So nobody really understands it, they just learn enough to do what they need to do. + +**Matt Holt:** Yeah. There's a lot of moving parts, too. I've found myself setting up four or five different components just to get a basic website up and running... So I wanted to kind of simplify things. + +**Mohammed S. Al Sahaf:** So it's pretty much like the XKDC comic of Git, except it's now web servers. + +**Matt Holt:** Yeah, I suppose so. + +**Mohammed S. Al Sahaf:** You call up the friend who knows how to \[unintelligible 00:06:08.13\] + +**Matt Holt:** Yeah. + +**Jon Calhoun:** I wonder if these days, Matt, you're making it worse... Because more of the people I talk with are using Caddy. So finding people who know how to use NGINX and Apache is gonna become a rarity, and when you've got a company relying on it, that could be a problem. + +**Matt Holt:** I don't see that as a problem. + +**Jon Calhoun:** Well... It could be a problem for that company, for a short period at least. + +**Matt Holt:** It could be. We are encouraging all companies to switch to Caddy. + +**Jon Calhoun:** So I guess we can jump into the Caddy stuff then... Because I think - Natalie, correct me if I'm wrong; I think the goal is to talk a little bit about Caddy and then start talking about this project Mohammed is working on as well, and going in a little more depth there. + +**Natalie Pistunovich:** Yeah, definitely. And for those who are less familiar with the -- for the two people who are listening that are slightly less familiar with Caddy... So far, those people figured out that we're talking about a web server. What is special about that web server? Unlike NGINX and friends... + +**Matt Holt:** Well, you'll find that it comes with all the benefits of a Go program. So it's easy to deploy, and it has no external dependencies, and it has memory safety, which you don't get from C servers like Apache, NGINX and Envoy... And with Caddy 2 we have made it kind of a very powerful platform on which you can deploy your Go services, and you can extend it -- it's pretty much infinitely extensible. It's not like writing some Lua code, or some little scripting things that's interpreted. You're compiling your Go code natively, so you're getting native CPU instructions for your extensions. So there's a ton of benefits there. + +And probably the thing that people mostly talk about when they think about what's special about Caddy is that it uses HTTPS by default. So it's the only server that does that automatically and by default, without needing any config; it will just try and make TLS work for your site, and manage that all automatically. + +**Jon Calhoun:** \[08:06\] See, I can say for myself that was pretty much the original draw to Caddy, was - you wanna set up some sort of HTTPS, and setting it up in other places has always historically been a pain in the butt. I remember the first time I had to get a certificate, and I was probably a teenager at the time... And it was one of those things where trying to do that while not having the money to really wanna do it for a side project was a nightmare for the longest time. So seeing things like Letsencrypt, or - is it ZeroSSL, is that one of the other ones? There's a bunch of them out there. + +**Matt Holt:** Yup. There's a few others. + +**Jon Calhoun:** Seeing different companies like that that are offering much easier options out of the box has been pretty awesome. Alright, so Matt - Caddy is written in Go entirely; is that correct? + +**Matt Holt:** Yup. Pure Go. + +**Jon Calhoun:** Okay. And Mohammed... You, I believe, wrote an extension. So Matt was saying that it's extensible... Can you verify that that's easy to do? Can you talk a little bit about that process? + +**Mohammed S. Al Sahaf:** Oh, it made my life a lot simpler. I only had to focus on the logic of the SSH server as in authentication, session management, and PTY, and all of that... And that was a breeze, because I didn't need to write any of the logic for the listeners, and the reload management and the config management loading and all of that stuff - it was all handled by Caddy. I only had to focus on my own project. + +**Natalie Pistunovich:** You said that you're a Gopher in the evenings, and by day you are a project manager... So why going on a web server as a fun project? Was that a need, was that just something fun that you've found? + +**Mohammed S. Al Sahaf:** Yeah... My day job is not into the software development side of things. It's still related to tech; I'm a product manager of the payment gateway at a local bank. So I don't get to practice programming day by day. And programming at night, in the evening is basically my stress relief. + +**Natalie Pistunovich:** And from all the projects, a web server is the interesting one? + +**Mohammed S. Al Sahaf:** Well, Caddy was fun enough, and interesting, and easy to contribute to. There are a lot of sides and angles that you can nibble at, and at the time - whenever I started getting involved, the Caddy versioning tool was still being written, it was still in the beta phases... So there were a lot of angles that I managed to look at and work on. So it was kind of, let's say, a ramp to get on. + +**Natalie Pistunovich:** So it's interesting that you said that it was easy to get involved in and easy to contribute to. That sounds like a fun place for starters. Can you tell a little bit more about this for people who want to get involved and are not sure how to start, what to start, what makes starting easy...? + +**Mohammed S. Al Sahaf:** Yeah, sure. So there are a lot of sides that can be looked at... And the architecture of Caddy is engineered well enough that if you're looking at a certain handler, for example, most likely you don't need to look anywhere else besides that particular package, or one or two files that you need to look at. And depending on what area you're trying to get at, you will find something to work on. + +If you wanna work on the CLI, it's there. If you wanna work on the config loading and reloading, and the config structure itself, the configuration files - it's all there. If you wanna write a handler that simplifies the lives of a particular process on the web server, you can look at some of those. There's a ton of things that are in there. If you wanna look at PKI, it's in there. + +And you'll always -- it's the nature of projects; there's a fractal nature. There's so many things that yo can work on, and you will always find something that's either a good first issue, or something that requires more depth and knowledge about the project itself, its architecture, or the Go runtime; if you wanna look at performance, you can go begin there. That will be perhaps a bit challenging. If I wanna pick up some easy fruits it's all in there. + +\[12:13\] For example, one of the first things I worked on was basically looking at the warnings or the messages golint was giving (golint CI). It was basically saying "Change the order of these fields and that struct to make it more compact", and stuff like that. + +So it was easy to get into the project and learn more about it as I worked on it more and more, picking up stuff like that. + +I think there's still some low-hanging fruit for other beginners or fresh contributors wanna look at. You can basically fire up the project on VS Code and run the linter, or any of the available linters, whether it's Golang CI Lint, or whatever, and find some areas to fix. And we have a few in the issue list that can be looked at. + +**Matt Holt:** Yeah. I might also jump in and say you can learn a lot from hacking on Caddy, but it might not be the best project for absolute beginners, if you're absolutely new to Go, or absolutely new to open source... Just because its sheer size - there's a lot of packages and different pieces. So it might be a little overwhelming if you're brand new. But if you're experienced with Go and/or experienced with open source, it's a lot easier to work through the feedback cycle. + +Mohammed from the beginning has been awesome about optimizations and little nuanced things that I had never thought to figure out. So that was really cool. + +**Natalie Pistunovich:** Is it complicated to write such an extension? How much time would you say such a project will take? Or it depends if the extension would be complicated, or simple... + +**Matt Holt:** I guess Mohammed can answer, too. He's probably written more third-party extensions than I have. But I think it depends. If it's a simple extension, some extensions are just a couple lines of code. Others are thousands of lines. It just depends on what they do. I managed in Mohammed's SSH server. I think it took two years to put that together. That's a huge app module. + +**Mohammed S. Al Sahaf:** Well, the two years is because my time is divided between work, school and this side project, plus my personal life... And this is why it took such a stretch. It shouldn't have taken two years, honestly. + +**Jon Calhoun:** So when you're talking about these extensions, I assume with Caddy - based on my knowledge of it - it's a server where you started out and you provided a config file, and it sets up a server in some way based on that... So whether it's grabbing a certificate for your domain and reverse-proxying it to something on a local host, or something like that. When you build an extension, what exactly is that doing? Is that giving it new config variables that it can then take and do stuff with, or is it something else aside from that? + +**Matt Holt:** So at its core, all Caddy knows how to do is - there's like four keys in its config. There's admin, logs, storage, and apps. And all it knows how to do is set up admin and logs (logging) so you can see output, and then storage and apps are all modular; they're all extensible. So Caddy expects apps to fulfill an interface that have a start and stop method. That's it. So it calls starting all the apps when a config is loaded, and it calls stop when it's unloaded. Beyond that, it's up to the app module to do its job. + +So the HTTP server has a start function and when it's called, that's when it starts its engine, so to speak, and starts serving your site. And then when it's stopped, it shuts down gracefully. So Caddy doesn't know anything about HTTP, really. So that's why I use C now... I don't know if this is clear to the listeners yet, but the implication with Mohammed's work with Caddy SSH is that you can now deploy a Caddy instance that does all that you need to do with one unified configuration. So you need to run an HTTPS server, you put that in your config; you need a memory-safe SSH server - you put that in your config. And you just deploy this one binary that's static and has no dependencies, and is memory-safe, and such, and it takes care of all the TLS for you. Obviously, SSH doesn't necessarily use TLS, but the idea is that it's kind of your one-stop-shop for memory-safe static deployments. + +**Break:** \[16:39\] + +**Jon Calhoun:** Mohammed, if I were to install your Caddy SSH extension and to run it, this would allow me to SSH into the server that's running Caddy, correct? + +**Mohammed S. Al Sahaf:** Yeah, exactly. + +**Jon Calhoun:** Okay. And the idea was to give it useful defaults and things like that, so that people setting this up are less likely to make mistakes that are, I assume, pretty common when setting things up like that? + +**Mohammed S. Al Sahaf:** Yeah, especially with the SSH, because over the years - SSH has been around since the '90s, and the defaults have been changing and improving over time, but we have so many blogs and tutorials written that were probably written for the early 2000's, while we are here in 2022 and SHA 1 isn't safe anymore, and RSA - you would need to have a minimum of 2048 bits, and so on. + +\[20:01\] At the same time, you'll find those tutorials, those blog posts saying "Well, generate an RSA 1024-bit key and it should be alright to use with your SSH server." The goal is just like Caddy now is working with safe defaults for TLS \[unintelligible 00:20:14.10\] and have more than representation in defaults for the certificates will do the same for the SSH server. So one of the things that I made sure to implement is the keys that are generated automatically, they follow the modern recommendations. For example, if there is no RSA key available minimum, one will be generated automatically for you, and it will be 4096 bit by default. + +The other key that's automatically generated is the ECDSA, but the regular DSA isn't generated, and so on. All of that is written in the comments, in the docs, in the code docs itself. + +**Jon Calhoun:** So getting into this, I assume that this means you had to actually learn about all of the proper ways to do things now, versus the past... And you had mentioned that reading blogs - sometimes you find blogs that were written in 2000, when security practices might be a little bit different. Would you have any advice to people? How did you go about managing that process of trying to figure out if what you're reading is actually up to date and accurate, versus understanding what current practices are, versus like "Oh, this is an old practice. I shouldn't be doing that one." + +**Mohammed S. Al Sahaf:** That's a tough one. One of the things that I had to look at, that's an easy one to look at, is look at when was it published. If it was published in the '90s, cross it out, look for something more recent. I'm not a cryptographer, I'm gonna admit this, but I know from a few readings that elliptic curves is a bit controversial, probably... And if you know that, then you might correct me. But at the same time, there is a strong recommendation for it. And this is why one of the defaults that I opted for is to have elliptic curve DSA. + +The same thing for RSA. We know from recent findings that RSA 1024 bits isn't safe anymore... So higher bits, and so on. It's more following the recommendations of well-known cryptographers, for example. + +I remember reading one of -- some of the stuff I came across, and I used a few blog posts by Matthew Green; I don't remember exactly the titles, but it was part of my research that I went through and I've found which algorithms should I use and which not. + +**Jon Calhoun:** So when you're going through these algorithms, has Go been a language where you've found a lot of them were already implemented for you, or have you had to go and try to implement some of these on your own? + +**Mohammed S. Al Sahaf:** No, I never needed to implement any of those on my own, which is delightful. Go had everything battery-included. Again, one of the things that's readily available is the 825519. That was already in there, and it's part of the standard lib. I just had to only grab it from there and generate the keys with it. Same for all of them. The crypto package of the Go standard library is pretty comprehensive for my needs. + +**Natalie Pistunovich:** One of the skills you picked up in this project is cryptography. Is there any other interesting skill that you picked up along the way? + +**Mohammed S. Al Sahaf:** Oh, I had to learn a lot about pseudo-teletype. That has a ton of stuff that I had to go through, whether it was for Unix systems or for Windows. Especially on the Windows side, because it didn't have the PTY interface, and then they developed the \[unintelligible 00:23:45.04\] interface, but it had its complexity, so I had to wade through all the Windows-provided dll's, and so on... So definitely PTY was a good, interesting area to work on and learn about. + +**Matt Holt:** I'm glad you did that, because I don't even really wanna go there... It sounds kind of gnarly. + +**Mohammed S. Al Sahaf:** \[24:06\] Oh, it is. It's funny, because here we are, we have the basic kind of system we work on almost always virtual machines, yet we're still SSH-ing into a server. The other side - and the name still implies it; it says teletype, pseudo-teletype, and you will tell the server "Here are my screen dimensions. Give me the text or the output based on those screen dimensions." And when you change the window of your shell, your SSH client will tell the server "Okay, these are the new dimensions of width and height. Now resize everything and give me back the response." Which is strange, because the remote machine isn't really connected to a screen. It's emulated all the way down, and it all goes back to, I think, the '60s or the '50s, whenever they had actual teletypes connected to the actual hardware, and it actually had to tell it "Yes, I have a screen of this particular size, and you have to redraw everything to that size." But now, here we are in 2022 and we still have to do all of that, because hey, we are emulating everything down to the pixel, because we have legacy. + +**Matt Holt:** I was just looking at the list of modules... If you go to the Caddy download page, or the modules page, and you look at the SSH app that you wrote - it registers a lot of plugins or modules. I see SSH actors, actor matchers, ask, lots of ask and authentication, different providers and flows, config loaders and matchers, session authorizers, signers... Even an SFTP subsystem, it looks like. This is pretty thorough. What can you do with this exactly? + +**Mohammed S. Al Sahaf:** Okay, so the actors are equivalent or analogous to handlers in HTTP. And the actor matchers are analogous to the matchers in the HTTP app. So just like you can define multiple possible actions or handlers for a particular request based on certain criteria, now you can use the matchers to match - if it's user Jon, then run this actor. And if it's user Natalie, run this other actor. And if it's Matt, just - sorry. Shut it out. + +**Matt Holt:** This is cool. So you can kind of compose your own SSH logic with JSON in a memory-safe SSH server... And it looks like you have two actors so far - a shell and a static response. So you can start a shell for the user who logged in, or just write a static - like a screen, basically, to the user... And it's cool - so you can match on various SSH connection properties, like user, their IP address, their group, their protocol extensions, the critical options in the certificate of the user... This is really neat. + +**Mohammed S. Al Sahaf:** For the shell actor there are a few enhancements that I've been working on locally. I know there are certain areas or gaps in the implementation that I have picked up on -- that I've fixed locally, and I haven't pushed yet... Because C is in the way. So the shell needs a lot of enhancements. There are a lot of actors that could be implemented. Perhaps proxy... I don't know. It takes a bit of creativity. Shell and static response were the least creative things, and they were probably the mostly used actors or expected functionalities out there... And this is why I went with them. + +**Matt Holt:** \[27:54\] I could see this being extended to implement custom SSH apps. I've seen kind of a resurgence lately in graphical, SSH-based applications, and I could see this maybe being a good platform to launch and deploy those. So you can deploy a website, but you can also deploy an SSH app. So if a user wants to SSH into it, you can have this nice tty. But that's really cool. It's retro and it's cool. + +**Jon Calhoun:** I think one of the ones I saw recently was by Charm, I think was the company... They have like a self-hosted Git server that made it look all pretty and stuff when you SSH-ed in. But every time I see one of those, I just think that like that's its own style of programming, making something look good in the terminal, and doing all that stuff. + +Mohammed, you had said you had to figure all that out, and I feel like that's a skill that if somebody asked your average developer to go build that right now, they would need some sort of reference material, because it's just not something people do as much anymore. + +**Mohammed S. Al Sahaf:** Making anything look pretty is difficult, and it takes skill. + +**Jon Calhoun:** That's true. So Mohammed, you've spend, it sounds like, two years - granted, it was a part-time project - working on this. And it's a massive extension. If somebody was gonna get started building an extension for Caddy, or getting involved in that sort of thing, do you have any recommendations for them that you've learned throughout the whole process, or have there been things that you wish were different? + +**Mohammed S. Al Sahaf:** As Matt said, there are four basic concepts that you need to look at. If you're implementing a module for Caddy, you're 99% of the time - or probably 100% of the time - implementing an interface, and you're registering your module as part of a namespace where Caddy can find it and load it up. + +So I would say find exactly where within Caddy do you want to add your module, which particular functionality or namespace you're looking at. Is it a listener wrapper, is it a handler, is it a connection policy? And then find the interface that you need to implement for that area and start from there. You can start small, and then add more features into it. + +I admit this is pretty much a generic advice for all projects. But this is what it is like for Caddy. And you need to look at how you're config structure is gonna look like. So consider how you would like your JSON configuration to look like, and move from there. And this was the only complex part, or the difficult part for me to work on - which shape of JSON structure of the configuration is the nicest to work with, or the one that makes more sense? So once you figure your JSON out, everything is easy from there. + +**Matt Holt:** Yeah, I wrestled with that, too. I think when I first started writing Caddy 2, the first 3-4 months was to me just going back and forth on what the JSON should look like... So Caddy's native config structure is JSON, but most people use the Caddy file for ease of use... But I've actually got quite a few users who do use JSON because you can program it, you can automate things a little easier. Everything compiles down to JSON, and you can always output JSON, you can ingest JSON everywhere... It's very ubiquitous. + +We actually do have a lot of business users in JSON... So that might be helpful if you wanna contribute - knowing your native config, what that looks like, and kind of getting that right. Because that'll impact a lot of future capabilities as well. + +**Jon Calhoun:** Matt, I think you had said that pretty much everything inside of Caddy is essentially a module that's running... Am I recalling that correctly? + +**Matt Holt:** Yeah. Everything except for its core, like, module loading and logging; it's core API, and such. Yeah. + +**Jon Calhoun:** So are any of those built-in extensions that would be worth checking out for somebody who's just getting started, versus the ones that are a lot more complex? + +**Matt Holt:** \[31:58\] Yeah, I would check out extensions or modules that are like what you want to build. So if you want to build an HTTP handler that handles requests, then you should look at HTTP handlers. A simple one is the static response handler, where you just hardcode a response. That's a pretty good one. If you wanna write an app, the HTTP app is pretty complex. The TLS app might be a little better. There's also the PKI app, that's even simpler. So yeah, just look at the kind of module that you want to implement, and we have documentation pages explaining what the different kinds of modules are, and kind of how that works. + +**Natalie Pistunovich:** Jon you asked Mohammed what is your tip for people who want to start, and Mohammed, you gave a very good answer of "Start something small, and build on top of that. Don't make that too complex." Do you feel that your experience as a product manager helped you start in an organized way? And if yes, if you've found something that works well in your team at work and you took with you to building this, what are some tips that you can share? + +**Mohammed S. Al Sahaf:** Yeah, it's been really a great feedback loop where at my job as a product manager I have to fix something or develop something in a certain manner for customers and I know the product is gonna be large, so I have to structure it in a way where it is extensible and easier to implement new functionalities, that will not make it a breaking change for the customers. At the same time, in my project I have something similar where I need to develop something, but the only difference is at my job the product I have is something I inherited. But here I have something that I have to develop from scratch... + +**Natalie Pistunovich:** Which do you prefer? \[laughs\] + +**Mohammed S. Al Sahaf:** Develop from scratch. It's a hundred times easier than inheriting something and I have to fix it. \[unintelligible 00:33:52.04\] You have a legacy and you have thousands of customers already using it, and there's this edge case that's not an edge case anymore, and it's really a painful corner that you need to fix, and it cannot be fixed in a way that's backwards-compatible... And now either I have to live with it and add bandaids on top of bandaids as we go on, or I have to do the breaking change and go through the pain of migrating or teaching thousands of customers how to follow the new, correct way. + +And you know, if it's a breaking change, you're gonna have all of those customers yelling at you over \[unintelligible 00:34:37.03\] phone calls. And with that, I find developing something from scratch is way better. The way it feeds back is - for me as a product manager, one of the things I took upon myself, because I was frontline support at one point in time, and then I shifted to product management... And what I do all the time is I take an hour or two every day and I sit with the frontline support team and I ask them for feedback - what do the customers complain about? And when I started working on the project, the SSH app, I had to take a similar perspective to that, like "I'm gonna structure this app in a way that will be used in a critical function, in a critical area. Now, I need it to be done in a way where support is gonna be easy, the customer onboarding is gonna be nice and simple and straightforward, and it should be intuitive." And this is where I had to take on the frontline support hat, like "What is gonna be a stress point for them, and how do I work around that?" I don't know if that makes sense or not... + +**Natalie Pistunovich:** \[35:57\] This is an interesting answer. It makes a lot of sense, and it also makes me think about people who take the path from project or product management into developer... Because you hear a lot about people going from software into product management, but I have to say, I personally know less people who took this path. But it sounds like this gives you such a toolbox that you wouldn't have otherwise. So that's an interesting to think about. + +**Matt Holt:** I think more software should be written like that, with that approach. + +**Jon Calhoun:** I think in general there's a lot of people who get into software development and -- I don't know how to describe this... It's almost like they view things like this is the way it should be for good software, and they ignore the fact that there are real business needs that need to be achieved, and things that need to be maintained. And I think it takes - especially newgrads who are going into software, it takes them a little bit to realize that there is a business, and business actually matters more than the actual software being the prettiest thing in the world... So being a product manager probably makes that very clear, like "The only thing that matters is my project moving forward." Whereas a software developer -- I know some software developers that could probably sit there, rewriting the same application for like six years, trying to make it perfect. + +**Break:** \[37:07\] + +**Jingle:** \[39:24\] + +**Natalie Pistunovich:** So, gentlemen... What is the unpopular opinion that you brought onboard? + +**Matt Holt:** Mine is that vanilla JS is enough for anyone. + +**Natalie Pistunovich:** Okay... + +**Jon Calhoun:** Do you mean like modern vanilla JS? + +**Matt Holt:** Yeah, of course. I'm writing kind of a frontend app in my spare time right now, and there definitely are a few little pain points, but it's kind of like those pain points in Go, where you just write a less function, or something like that. So you do that in vanilla JavaScript, it's such a -- like, the traditional way of building web applications, it just runs so fast, compared to... And it's way less clunky than a lot of modern ones, with just kind of a plain stack, nothing fancy. + +**Jon Calhoun:** \[40:24\] I kind of wonder if half that issue stems from the fact that for the longest time vanilla JS was very hard to use, by itself at least... And I agree with you that it's gotten way, way better, to the point that if somebody started over, I'd be like "You can start with vanilla JS." But I don't know if I'd use it to build an entire frontend or not... I haven't tried recently, so I can't really speak to what that feels like... But I can definitely say that I understand why people are in the mindset of "That's the way we do it." + +**Matt Holt:** Yeah. I mean, maybe there's a place for frameworks, but I'm not even using jQuery at this time, and it's really been a breeze putting this together. I understand what's going on -- the DOM is a weird place, and there are definitely quirks, JavaScriptisms that are just a little strange; you've gotta read some documentation carefully or you get bit pretty hard... But it's just -- I have full control over things that are going on. I know exactly the performance of something... I have a very good grasp of it, just because it's bare-browser, so to speak. + +**Natalie Pistunovich:** I'll just say that already one disagreement showed up on Slack, so... Kudos. There will be a survey afterwards on Twitter... We'll see how that predicts the outcome. + +**Matt Holt:** Yeah. + +**Natalie Pistunovich:** Mohammed, how about you? + +**Mohammed S. Al Sahaf:** Alright... Mine is Microsoft Excel is a net negative in this world. + +**Jon Calhoun:** No caveats at all. + +**Natalie Pistunovich:** \[laughs\] + +**Jon Calhoun:** Can you elaborate? + +**Mohammed S. Al Sahaf:** I've seen coworkers spending countless hours trying to figure out some weird issue Excel is doing, and it's just a waste of time. When I come around to help them, what I do is I take their Excel file, convert it to CSV somehow, run it through SQLite and do whatever they need to do, generate whatever information or value they need to generate, and then take that output, put it back into Excel. And most of the time, whatever you're doing in Excel, it would be a lot easier if you just learned some SQL; and it's easy to learn. I can teach some businesspeople how to use SQL. Learn that, use SQLite... There are some tools, there's DB Browser, there's tons of other stuff out there. Just ditch Excel, throw it in the bin, have no regrets. + +**Matt Holt:** So you're talking about spreadsheets in general? + +**Mohammed S. Al Sahaf:** Yeah. Spreadsheets in general. + +**Matt Holt:** So like Google Sheets, yeah? + +**Mohammed S. Al Sahaf:** \[unintelligible 00:42:55.03\] please, reach out for SharePoint, use whatever. There are tons of alternatives. + +**Matt Holt:** I mostly agree, I think, so... + +**Jon Calhoun:** I mostly agree, but then... I guess the hard part there - and like Mohammed is saying, we need to educate people on using SQL, which I think would be great if people understood basic ways to do queries, or just simple ways to automate things... It'd be great to see a generation of kids growing up knowing how to at least write a little script... + +I've definitely seen people sit on a spreadsheet, doing things that -- manually typing in, that you should not be manually typing in that stuff, or trying to do what you're doing. But on the other side, I think there's just a whole generation of people working that are just not gonna learn something new. And that puts us in a weird spot of like "How do you keep that still going, while the others need something better?" + +**Natalie Pistunovich:** It feels like Copilot just writes SQL for you. + +**Jon Calhoun:** I don't know if I'd trust -- you're talking about the AI code stuff? + +**Natalie Pistunovich:** The GitHub autocomplete thing, yeah. + +**Jon Calhoun:** \[44:00\] I don't know if I'd trust that fully to write all my queries, but... That'd be interesting to try though. It's hard for me too, because I use spreadsheets for simple stuff all the time, where I just don't wanna code, and it's just real quick to throw something -- now, granted, most of my spreadsheets are throw-away spreadsheets of like throwing a couple things in here real quickly, doing some calculations, and... It's like a calculator almost. + +**Matt Holt:** I'm kind of the same way. I use it like a disposable calculator, simple stuff... If I have to start looking up formulas, I'm switching to SQLite. + +**Natalie Pistunovich:** Well, we also have a rank of the top three unpopular opinions, and top three unpopular unpopular opinions, which is kind of popular opinions... So it looks like we have a candidate for each from this episode. + +**Matt Holt:** It's fun. + +**Jon Calhoun:** So Matt, I have another question for you... + +**Matt Holt:** Oh, no... + +**Jon Calhoun:** How often do you get grief over using the init function for your modules, or for your extensions? + +**Matt Holt:** More than once... + +**Jon Calhoun:** More than once? \[laughs\] + +**Matt Holt:** The counter-argument from those people is that the importing package, so the Caddy main should be calling not only importing, but also calling the register module function. I don't really see the point in that though. I feel like if you're importing it, you want to plug it in. + +**Jon Calhoun:** I'm not saying one thing one way or the other, I just -- I laugh that you go to the documentation and the third line in \[unintelligible 00:45:15.24\] and I'm like "I feel like this right here has gotta lead to a lot of hatemail of some sort..." Although -- I mean, given your use case, I personally don't know a much better option for what you're doing, where you need a bunch of things imported and built with it. + +**Matt Holt:** I mean, it's either the two lines of code that you add to - well, more than that; you'd have to import the package and then you'd have to know all the module types in that package and call register module yourself. I don't think that's necessary. I think if you're using a package that has N number of modules, just plug them in. They're already there. + +I think Caddy has some other unpopular design decisions, but I still stand by for the most part, some of them especially. Some people still don't love the JSON thing, but the JSON thing is wonderful, and you don't have to use it. + +**Jon Calhoun:** So we can just make a list of unpopular Caddy opinions...? + +**Natalie Pistunovich:** \[laughs\] + +**Matt Holt:** Yeah - to be fair, the number of unpopular Caddy opinions has come down recently. Since Caddy 2, I think Caddy has been a little more on the popular side overall. There's definitely still some debates... And some things I don't think there are good answers to, but come at me. We'll see. We can talk about things. + +**Jon Calhoun:** I think what you're saying is -- to me at least, I view what you're doing as you're trying to choose the lesser of two evils... Essentially, of ways to approach a problem, and it is a tough one, with the extension. Because making an application -- I don't think many people have written software that needs to be extensible, especially built with other code... And that's a very unique problem to solve, I think. + +**Matt Holt:** Yeah. + +**Jon Calhoun:** I think Mark Bates has done it with Buffalo stuff before, but that's one of the few people I know that's even gone into that realm at all. Most projects I just don't think support it. They'd just be like "These are the modules we have. If you want others, good luck." + +**Matt Holt:** \[47:07\] Yeah. I mean, sometimes there is still some debate over Caddy's module design in terms of the fact that they have to be compiled in... Some people kind of hate that. But a lot of people love it, and it has a lot of advantages. + +**Jon Calhoun:** What would the alternative be? Having like a second server running that it communicates with? + +**Matt Holt:** Yeah, so there's a couple options... You could do some sort of RPC thing, or interprocess communication, so IPC... But that has performance penalties, because you're going through the kernel. You can embed a scripting language, and have some sort of interpreter. There's a few Go interpreters out there, or interpreters written in Go for various languages, like Lua or Go actually, and probably some other languages as well, like Starlark and such... But then you don't know your code is broken until it comes to your runtime. And so it's like -- I don't know, it just made more sense; just compile the code, get native performance, ship a static binary... Don't worry about it. You don't wanna be messing with things in production anyway, so... + +**Jon Calhoun:** Yeah, I don't think that's an easy one to come up with a perfect solution for, I guess. + +**Matt Holt:** There isn't one, but it's a pretty good one. + +**Jon Calhoun:** Because I'm thinking of like VS Code, where every extension is JavaScript, but then there's the fact that -- it's not the JavaScript that's terribly inefficient or anything, but it's definitely not going to be as performant as other languages, in most cases. So you have that limitation when you're setting up things. + +VS Code doesn't seem to have it quite as bad, but I know with Atom that was one of the issues, with that editor - because it was JavaScript, there were definitely times where coming from like Sublime Text it felt slower... + +**Matt Holt:** Yeah. + +**Jon Calhoun:** And I think part of that was because they wanted something extensible, and Sublime Text - it was extensible, but it was Python, and it was a little bit harder to do, if I recall correctly. + +**Matt Holt:** There is one more really maybe unpopular decision that I'm lukewarm about, and that is that you can't set a global config in the Caddy file. You can't get one line of configuration there and have it apply to the whole config, to all your sites automatically. We have snippets, so you can put it in a snippet and then import it into each of your sites, but you still have to put that import line. Some people want the Caddy file to work like cascading stylesheets, but if you've ever written CSS, you'll know why I don't love that idea. Anyway, that's all I can think of... + +**Natalie Pistunovich:** Cool. To finish, I'll say that the show notes have the links to extending Caddy, and also to the Caddy SSH extension by Mohammed... And we will also include the XKCD that you mentioned in the very beginning. + +Thanks everyone who joined, and listened, and responded in the chat, and thank you, Mohammed and Matt for joining. We wish everyone a good rest of your day, whatever time it is. diff --git "a/Spooky stories to scare devs \357\243\277\303\274\303\253\302\252_transcript.txt" "b/Spooky stories to scare devs \357\243\277\303\274\303\253\302\252_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..71dc829930a22e71792526ffaffe960aa09550d6 --- /dev/null +++ "b/Spooky stories to scare devs \357\243\277\303\274\303\253\302\252_transcript.txt" @@ -0,0 +1,663 @@ +**Mat Ryer:** Hello, and welcome to Ghost Time. I'm Matt Ryer. Today we're talking about tech horror stories. I'm joined, as ever... Johnny Boo! Johnny Boursiquot is here. Hello, Johnny. + +**Johnny Boursiquot:** Hello, Matt... + +**Mat Ryer:** Welcome to the spooky Go Time episode. Again, in the spirit of it -- you've really got to get in the spirit of it... No? Yeah... \[laughter\] We're also joined here by Kris Brandow. Spooky ghosts... Hello, Kris. + +**Kris Brandow:** Hello! I'm back, again. Finally. + +**Mat Ryer:** Yes, welcome back again. We're also joined by your friend and mine, Natalie PistunoWitch. Hello, Natalie. + +**Natalie Pistunovich:** Hello! + +**Mat Ryer:** Yeah, getting into the spirit... \[laughter\] We have a special guest joining us... You're not gonna believe this. It's spoopy Dee. Dee Kitchen. Welcome, Dee. + +**Dee Kitchen:** Thank you. I'm enjoying being here. I even got the backdrop right... + +**Mat Ryer:** Good. Well, that's a good start, because we've literally just started. I mean, really, the only way is down now, in a lot of ways... But hopefully we don't go there. But we are talking about scary things today. How are you generally with scary things, Dee? + +**Dee Kitchen:** That's my career. All of it. + +**Mat Ryer:** \[laughs\] Yeah. Okay. Anyone else? Anyone scared of ghosts? + +**Natalie Pistunovich:** I'm scared of Ghost Time. + +**Mat Ryer:** Yeah, you're scared of Ghost Time. + +**Dee Kitchen:** I'm actually scared of horror movies. I don't really watch them. + +**Mat Ryer:** Oh, yeah? + +**Natalie Pistunovich:** Same. + +**Dee Kitchen:** Heebie-jeebies for me. + +**Mat Ryer:** Hm... + +**Kris Brandow:** I just find them boring. + +**Mat Ryer:** Yeah... + +**Natalie Pistunovich:** Did you come from the industry? I remember you saying, Kris, that you see a movie, the first few minutes, and you know exactly how it's gonna be laid out. + +**Kris Brandow:** Yeah... It's the curse of having a creative writing degree and specializing in screenwriting... All movies are just kind of ruined. + +**Natalie Pistunovich:** It's all generics. + +**Mat Ryer:** That's what I said, you'd be happier if you're just an idiot. I've always said that. + +**Kris Brandow:** Is this something you know from personal experience, Mat? \[laughter\] + +**Johnny Boursiquot:** Ooh, shots fired... + +**Mat Ryer:** I also don't really like horror films, especially if there's any kind of contradiction in it. I can't deal with that. Like, if there's an invisible thing that can grab you, first of all, it's invisible, it would be blind; we've covered this. But also, if it can grab you, you can grab it, you can hurt it... Like, it's not fair. It's like when the physics don't apply generally; then I'm just out. And I just tell everyone in the cinema, I'm like, "Sorry, everyone. I can't stay. I've got to go because of the inconsistencies of the physics." And I just go and get some popcorn and go. + +**Natalie Pistunovich:** \[06:10\] Do you get sweet or salty popcorn? + +**Mat Ryer:** Salty. + +**Natalie Pistunovich:** Do you have a choice? Or is it always salty? + +**Mat Ryer:** No, you have a choice. \[laughter\] Don't you have a choice? Yeah... What do you mean? Like, the police are going around saying "Hey, are you only having salty? What are you doing?" + +**Natalie Pistunovich:** I only discovered in my late 20s that some other countries sell popcorn that is not just salty in the cinema. + +**Mat Ryer:** Oh, right? Yeah... + +**Natalie Pistunovich:** And then I came to the US, and then it's like not just two flavors, but 15. + +**Mat Ryer:** Yeah, of course. Yeah. Yeah. + +**Dee Kitchen:** That's a horror story there... + +**Mat Ryer:** Yeah. You can choose individual bits of corn, and have different flavors, and just have as many as you want. You just program it. You do it as an app, and then it pops it on demand. + +**Kris Brandow:** You say that, but we do have soda machines where you can choose your own flavor. + +**Mat Ryer:** Yeah, I've seen that. + +**Kris Brandow:** Those freestyle Coke things... + +**Mat Ryer:** Has anyone come up with a good one yet? Because I imagine they're all terrible. Someone's like, "You know what? I've accidentally pressed these three, and I've made a brand new flavor that never existed before." + +**Kris Brandow:** Well, no... I think they make it so you can't make any truly terrible-tasting ones, because that would be perhaps bad for them, so... + +**Mat Ryer:** Oh, really? Clever. How did they do that? Well, we'll never know... Well, speaking of horror stories - let's get into this, shall we? Who Wants to kick us off with a spooky story? Oh, by the way, we should actually introduce Dee, because Dee, you wrote a package that I think a lot of people here will be familiar with. Can you tell us about bluemonday? + +**Dee Kitchen:** Oh, bluemonday... It's named because there was a package we'll call Black Friday, which has all the best markdowns, and it's a markdown package. And after you've generated markdown, markdown can include HTML, which makes it dangerous. It's probably you're using this because you've got user-generated content, and you want to sanitize it. So bluemonday is named after the New Order song, but follows Black Friday... And it basically sanitizers HTML; it's the only Go package that sanitizers HTML, which is a foolish and reckless thing to attempt to take on. But that's what I did. + +**Mat Ryer:** Amazing. And what do you like about it, and what don't you like about it? + +**Dee Kitchen:** I like that it works. \[laughter\] It's a streaming parser, it's got fixed memory, so you can use it quite comfortably in a lot of situations, and \[unintelligible 00:08:20.10\] I don't like when people tell me there's security issues with it, and then I have to go "Oh, I'm supposed to take this open source thing seriously?" I do appreciate it. I should actually say, I do appreciate security reports. But at the same time, you never can predict when they're going to turn up, and you never know what kind of words you're going to open to try and actually figure it out. + +**Mat Ryer:** Yeah, there must be a lot of responsibility, actually, because it is a package that is used, and quite trusted. + +**Dee Kitchen:** Yeah, it's used. I don't know how many stars it's got, but the stars don't portray the amount of times it's used. Like, it's used in Hugo, and everyone uses Hugo. And this is the HTML sanitizer that keeps Hugo safe. And it's used in so many things. It's got literally thousands of dependencies. Do I take it seriously and stressfully? No. No, I don't. I've figured that if someone is brave enough to take an open source project with an MIT license or a BSD 3-Clause, or whatever it is, and incorporate into their production software, that's on them. + +**Mat Ryer:** Okay, fair enough. Well, I have done that, but... Good to know. I genuinely have used it though, quite a few times, so... I like it, because it's like you opt into what you want to support. You explicitly say the things that you want to allow. + +**Dee Kitchen:** Yeah, there's no way of defining what makes a good HTML sanitizer. Everyone's got a different rule, depending on their use case. But the Java OWASP, Open Web Application Security thing - There sanitize it to find this really beautiful interface for sort of going, "I want to allow images, but I don't want to allow these images that end in .gif." And I copied their API, and then extended it for my own use. So yeah, it's a really good way of doing it. + +**Mat Ryer:** \[09:58\] Nice. Okay, well, I'm gonna tell you about a horror story in tech of mine, that happened quite recently... I have this project which interacts with Twitter, and it interacts with the Twitter API... And so it polls results and then compares them, and stuff. And that's just one of the things it does at a regular interval. And then what happened recently was something happened where the API key changed, and that request failed. And because of the way I was doing it in GCP, it meant essentially that it would retry. And because it was scheduled, it kept compounding. And this ran up a $1,000 bill for me, for yours truly... $1,000 given, paid, gone... So that's a bit of a tech horror story. And the advice for me -- + +**Dee Kitchen:** Is it tax-deductible? \[laughter\] + +**Mat Ryer:** Probably... + +**Dee Kitchen:** AWS famously refunds you if you get something wrong. Did GCP not do that? + +**Mat Ryer:** I don't know. It's quite recent. I haven't yet tried that. Do you think I should get in touch with support and see if they'll -- + +**Dee Kitchen:** $1,000 would motivate me. + +**Mat Ryer:** Yeah. There you go, $1,000. Okay, well, I'll try it, and I'll let the listeners know how we get on... + +**Johnny Boursiquot:** It could have been worse, right? + +**Mat Ryer:** It could have been $2000... + +**Johnny Boursiquot:** Yeah. + +**Mat Ryer:** Or just $1,001. It would be worse, wouldn't it? + +**Johnny Boursiquot:** It would. + +**Mat Ryer:** What would you do, Johnny, if you saw that...? + +**Johnny Boursiquot:** I'd call you and say "Hey, you've got a grand? I hear you're loaded... And just wasting $1,000 here, $1,000 there, on your bugs, and stuff..." + +**Mat Ryer:** Honestly, when I found out about it, I wanted to just karate-chop the air. That was the kind of spooky reaction I had to it. Just like \[Whew\] in the air. Angry. But yeah, it's a good lesson though. Like, set budgets and stuff on your things. + +**Johnny Boursiquot:** Do set an alarm, yeah. Budget alarms. + +**Mat Ryer:** Yeah. Observability. + +**Johnny Boursiquot:** Yes, yes. And you'd know a thing or two about that, yeah? + +**Mat Ryer:** Yeah... Okay, who can beat my $1,000 bill? Not $1,000 bill? Oh yeah, it was a $1,000 bill; but that makes it sound like it was one thing, doesn't it? Like a single bill that had $1,000 on it; so it's not that. It was just paid through a bank transfer. Okay, who's got another one? + +**Johnny Boursiquot:** I have one that could have cost many 1000s of dollars... + +**Mat Ryer:** Oh, Johnny... + +**Johnny Boursiquot:** ...if it wasn't spotted. So one of the things you can do with function as a service things, like AWS Lambda, for example, is that you can trigger a Lambda when you write an object to an S3 bucket. Word of advice - do not have your Lambdas write to a bucket that they are themselves responding to. \[laughter\] + +**Mat Ryer:** Ohh...! + +**Johnny Boursiquot:** Because that's gonna give you a very nasty bill. Yeah, and you will not like what you see. So yeah, thankfully, budget alarms came to the rescue... \[laughs\] + +**Mat Ryer:** Yeah, there you go. That's the lesson there. So what happens is, an object goes in the first time, that triggers the Lambda, the Lambda then writes something into that same bucket, which then triggers another Lambda, which then writes something. + +**Johnny Boursiquot:** Right. + +**Mat Ryer:** And how quickly does that get out of hand? + +**Johnny Boursiquot:** Very quickly. \[laughs\] Like, if you want to see how well a Lambda scales on your own dime, you can do that. And, yeah, it'll cost you money very quickly. + +**Mat Ryer:** Wow. Yeah. Okay. Pretty good... But yeah, that's -- the alerts came to the rescue. Nice one. Okay, anyone else got one for us? + +**Dee Kitchen:** I've got another infinite loop one... Are we allowed to name company names? I don't know, maybe it's internal and I shouldn't... + +**Mat Ryer:** Yeah, I don't know. + +**Dee Kitchen:** I worked for a certain company which has an orange logo that has a bit of a light shining behind it, and they man-in-the-middle the entire internet... Now, with that in mind, when I was working for said company, in their DDoS team... We didn't DDoS people; we were protecting against DDoSes. + +**Mat Ryer:** Yeah, I've wondered... The DDoS team! + +**Dee Kitchen:** \[unintelligible 00:13:56.02\] that's the opposite of what we're doing. No, we were trying to protect, and they have a system... They've got all these 200 POPs (points of presence), and thousands and thousands of servers... And every single one of these is protecting some of the traffic; each machine can do like 20,000 requests per second. + +\[14:15\] And yet, they need to be able to actually show the value back to the customer, and make these sort of decisions centrally. So you send all the logs somewhere, and they're all been sent to one data center. So what you end up with is like, if you're doing globally 10 million requests per second, you get 10 million log lines per second in one place. + +**Johnny Boursiquot:** Nice...! + +**Dee Kitchen:** A certain customer, on a certain point in time - industry and type to be non-disclosed - wrote an infinite loop in their client, and it basically spikes 8 million requests per second on top of our normal load. + +**Mat Ryer:** Oh, wow. + +**Dee Kitchen:** And they basically broke our logging, the entire visibility. So they were effectively under attack, but now flying blind, because we couldn't see anything because they'd broken all the logging. That was not a good day. + +**Mat Ryer:** Yeah... That one doesn't sound fun? What happened? + +**Dee Kitchen:** We figured out which customer it was, but we couldn't figure out the rest. But we asked them what they've done, and they figured out that bit and stopped it. + +**Mat Ryer:** Oh, wow. + +**Johnny Boursiquot:** They fessed up to it, they owned up to it? Somebody wrote an infinite loop? + +**Dee Kitchen:** Yes. \[laughs\] I think they realized... They must have seen what was happening on their side. + +**Johnny Boursiquot:** So they didn't pull a well -- well, I will not name names, but they didn't blame it on on of the interns? + +**Dee Kitchen:** Oh, we've got a certain thing where actually an intern did that. I had that intern, and he's actually really good tie. He's become a full-time engineer in that team. He's really good. Learned a lesson that none of us will replicate. + +**Johnny Boursiquot:** Oh, yeah. I love interns. I just don't like to throw them under the bus when something goes wrong with my company... \[laughs\] + +**Dee Kitchen:** No, that one was interesting... Also, at said man-in-the-middle company, we had a system -- the system was brilliant, right? You could send an instruction to any machine in the world in under 10 seconds, and every machine received the same instruction. And that's great when you want to say there's a new domain name, because you just have the whole world at the same time. But it's really bad when you say there's a new way to stop traffic. And we've made a greedy regex. And the greedy regex was the problem. Now, frankly, the system shouldn't have allowed it, but the system did allow it. And we were all at lunch, it was an all-hands lunch, and the next thing we know, we just get people running and going "The internet's down." Because we used our own systems. And we lost everything internally at the same time. That was hard, too. + +**Johnny Boursiquot:** I feel like there are many lessons there... + +**Dee Kitchen:** There was a lot of lessons. + +**Mat Ryer:** What we're having for lunch... \[laughter\] + +**Dee Kitchen:** Lunch went cold... + +**Mat Ryer:** That's scary, ain't it? But hang on... So can you explain - someone that doesn't know what a greedy regex is... What do you mean by that? + +**Dee Kitchen:** Yeah, a greedy regex... I mean, if you do something like .\*, what you're saying is match any character, any number of times. So if you .\*.\*, you've now exploded this any character, any number of times, followed by any character, any number of times. And what you're doing is you're increasing the CPU computation. You're still putting the same fixed input in, but what it can match is now -- you've doubled the possibility in just that one go. And essentially, that's what happened. But some of the inputs were web pages and web traffic, so they weren't small. They were quite large inputs. And under that condition, they consumed all the CPU. + +So wherever this rule was applied - and we had shipped it globally to every single website, every single bit of traffic - we fried every single machine instantly. So it was about four hours for us to recover from that. And the teams I saw, they did interesting things. We were connecting directly to machines, and looking at the Prometheus on them, because we had no other observability. + +**Mat Ryer:** Wow... And what was the impact of that? How many people were affected? + +**Dee Kitchen:** Everything was affected. We knocked out a lot. DNS, TLS, HTTP, everything. It was one of those nightmare scenarios. And you sit there as a company, you sit there and you sort of go, "What are these meteorites?" the dinosaurs went extinct by a meteorite. "As a company or product service offering, what's the meteorite that's going to hit us?" At that company, we were hit by every meteorite we predicted. It survived, but still, on the days when they hit, it lays waste to everything. + +\[18:15\] And everyone has them. The thing that you've got to realize is when you're there, you've got to sympathize with -- you can accidentally see another company go through this... They're having a bad day, and you've got to sympathize, because one of those meteorites is gonna hit you one day. + +**Mat Ryer:** Yeah, we see the HugOps goes around often on social media, and things, of people sending their support in those difficult times... + +**Dee Kitchen:** Yeah, there's a good tradition of sending cakes to each other to sort of go "Thinking of you..." + +**Mat Ryer:** Yeah. + +**Dee Kitchen:** And try and not have your salespeople ambulance-chase. + +**Johnny Boursiquot:** How well was your break-glass procedures documented? + +**Dee Kitchen:** It was pretty good. We were lucky that this happened during a London lunch hour when all of the SRE team - the original SRE team - were there. It's possible for a few people to break glass; they could do so in about five minutes once we actually understood what we actually had to do. It took about 20 minutes for us to gain any visibility and to sort of understand, "Hey, it's this feature. Go turn that off." + +**Johnny Boursiquot:** Regular expressions, huh... + +**Dee Kitchen:** That's still hard. They're hard for everyone. Easy to write, hard to understand what they're doing. + +**Mat Ryer:** Why are they called that? What is regular about them? + +**Dee Kitchen:** That's out of my domain. I don't know if anyone's got the answer for that... + +**Mat Ryer:** No, genuinely... + +**Kris Brandow:** I've just watched a talk from Strange Loop about regular expressions, and the speaker did go into this, and I've completely forgotten what she said. But we can probably put that talk in the show notes. It was a really good one, about just like the history, like Ken Thompson came up; it was pretty cool. So I'm like, "Oh, I know that dude..." Yeah, but no, it has to do with like mathy things, and finite automata, and all of that... + +**Mat Ryer:** Yeah. + +**Natalie Pistunovich:** When I was a junior, which is the closest to intern I was, I was working with a team lead, and when we deployed something together, we looked at it... And I forget what it was exactly, but this is a company that receives a lot of pings from the SDK of the many clients, and it's all real time. And if that is not logged, then the entire transaction flow, user flow is gone forever. And we deployed something we worked on together, we worked on it for half a day, we tested it, we did all the good practices - because that's how you do with a junior, you want to show that you're very thorough, and you go through all the tests, deploy, look at all the metrics, and see that it behaves as expected. And then he went to lunch, and then I stayed. Ta-da...! \[laughter\] And then I proceeded to do something else, and then suddenly, a weird behavior started pinging Slack, all the monitoring channels, that "Something's wrong, something's weird." + +And then some of the colleagues that were there tried to see where it comes from, and we couldn't figure this out in 15 minutes, and then, bravely, I came to the head of the DevOps team - there was no SRE team at the time - and I said, "I think it's this thing that we did. Can I revert it?" Nobody else from all the other senior people that was - I don't have another better word than brave, but I don't use the word brave... Nobody else wanted to do anything about that, because nobody was sure. And then I was like, "Let's do this. Let's try. At worst -- it cannot get much worse than that", and reverting that indeed succeeded, and then we were all very happy. And then I was like, "I think I know how to fix it. Can I try?" and then he looked at me and said "No. Stay away." \[laughter\] Yeah, I to next-level brave. + +**Mat Ryer:** Yeah. Well, what a great way to learn stuff though, isn't it? + +**Johnny Boursiquot:** Break them, yeah. + +**Mat Ryer:** How often does that memory come back to haunt you, Natalie? + +**Natalie Pistunovich:** Every time I'm asked, at least... \[laughs\] Every time I get to speak with other junior people. To give the good example of it, obviously, we'll break something, so... + +**Mat Ryer:** Right. + +**Natalie Pistunovich:** ...be reasonable about your expectations. + +**Mat Ryer:** \[22:01\] Nice. + +**Johnny Boursiquot:** I'm curious if you have a way -- like, now that you're older and wiser, and you've been through the experience, and it was a great teacher... I'm wondering, do you have strategies now for doing things that are scary, that could break things? Like, do you have a strategy for tackling that now? + +**Natalie Pistunovich:** The thing is we did everything right at the time, right? So we did all the tests we could think of, we thought what do we expect in the logs, in the monitoring, in the dashboard, and we observed. So the only thing that I would do different is not deploy before lunch. \[laughs\] So you can, I guess, observe for longer. Speaking of spooky things, right? + +**Mat Ryer:** What were you having for lunch? + +**Natalie Pistunovich:** I don't think I had lunch that day. I mean, I was like nudging stuff, but I don't think I had a lunch-lunch. + +**Mat Ryer:** Yeah, there's another theme emerging here... One of the main reasons to write good code is so you can just have lunch... For a good reason. + +**Natalie Pistunovich:** That's another thing I would the differently; I would write good code. Yeah. \[laughter\] + +**Mat Ryer:** That's optional. If you've got good tests, you don't need good code. Controversial. + +**Johnny Boursiquot:** So to me, what I usually tell junior members of staff is to be like, "Look, we expect you to break things. It's just part of sort of maturing as an engineer. What is helpful, and even if you follow the playbooks and you do the right things and everything else, sometimes things will go wrong; whoever writes these things, how many people have touched the documentation you're looking at, there's a chance that they might have overlooked something, or they took something for granted that you as a junior haven't encountered yet, so you don't take it for granted. So there's some steps in between. So there's some unwritten things in between the lines, that are sort of being conveyed, that you have not yet matured enough to kind of pick up on. So just document every step you take." + +It's much easier for the team to go back and say "Okay, what --" Because the first thing they're going to do is ask you, "What did you do?!" Right? \[laughter\] So after everybody calms down, you can say, "Well, these are the steps I took." Even to this day, I do this, right? If I'm working with a system that I haven't come across before, and I don't know what the side effects are of the things that I'm going to do, I'll literally, in a document somewhere, literally be copying the commands that I'm issuing in the command line - I'm literally gonna copy them into this doc, and I'm basically I'm capturing the output as I go. + +Now, one could say that's sort of extreme... I mean, again, if there was a playbook for it, if there was some automation that I could just click the button, or issue the command and let it do its thing... But if I have to do this step-by-step thing, that means there's no playbook for it. That means there's no automation, there's no script, whatever it is. So I'm just gonna be literally documenting what I'm doing step by step by step, and if something breaks, I know exactly what broke. Or if I can't tell what broke, I can ask my team and say, "Hey, these are the steps I was following." Right? And then nine times out of ten - maybe I'm just lucky - they'll be like, "Oh, yeah, this thing - you should have done this command before you do this", whatever, and then we find out that there's a gap in the documentation, or a gap in the process, or something like that. + +**Mat Ryer:** Yeah. And you can update it. + +**Johnny Boursiquot:** Yeah. So literally, just track what you're doing; that may actually end up helping you. Hey, guess what - that might even turn into a playbook or an opportunity for automation for whatever it is that you're working on. + +**Mat Ryer:** Yeah, it's like step one, SSH in. Step two, check the Go version. Step three, drop all the database tables. \[laughter\] Spot the problem. + +**Natalie Pistunovich:** It's just pen testing. It's fine. + +**Mat Ryer:** Exactly. Yeah. You shouldn't be able to do that, really. If you can do that... + +**Dee Kitchen:** I think that's important though, that they're taking steps, because it helps with something else... It helps people admit that they've possibly done something. Who in their early career has got the courage to say, "I've mucked up", right? "I potentially have lost you money, or time." Most people are terrified; and you're terrified legitimately, because you've got no experience in the industry, you're brand new, you're finally being paid to do something, and you think you're not very good. We're long in our career, and we probably think we're not very good. So early career, you're crushed. And that ability to turn around and go, "That might have been me. I think I did that. I pressed this button, and then it broke." That's tough. + +**Mat Ryer:** \[26:01\] Yeah. Well, I think that speaks to like the blameless culture that's important. It's important to reach the point where your people aren't punished for these mistakes... Because the last thing you want is people - like you say, they bury it, they try and hide it, or just don't tell anybody, which could make the problem much worse. So yeah, I think that culture plays a big part, doesn't it? + +**Kris Brandow:** Yeah. You should always blame systems and not people. If something went wrong, it's not the person's fault, it's why did the system allow the person to do that? + +**Mat Ryer:** Right. So when Johnny says something that's mean to me, it's not Johnny that I have to complain about. It's the system that lets Johnny get on the podcast and be horrible to me. \[laughter\] + +**Kris Brandow:** Exactly. It's part of the system to be mean to you, Mat, don't you know? + +**Johnny Boursiquot:** Indeed. + +**Mat Ryer:** Yeah, it feels like it sometimes... + +**Break:** \[26:51\] + +**Mat Ryer:** Have we got any more horror stories? Oh, by the way, this campfire's warm isn't it? We can probably put an effect of a campfire over the top; let's pretend we're all gathered around a campfire. Oh, what do you think of the campfire, Johnny? + +**Johnny Boursiquot:** Sure. Yeah, yeah. \[unintelligible 00:28:44.21\] + +**Mat Ryer:** I'm convinced by that performance, Johnny... Have you done actual theater? What about you, Kris? What do you think of the fire? It's cozy, isn't it? + +**Kris Brandow:** Sure. + +**Mat Ryer:** Okay... + +**Kris Brandow:** Crackly, warm fire... We don't have any marshmallows, so it's not as good... + +**Mat Ryer:** Don't we? It's imaginary land. It's podcast land. You can have anything you want. Check this out... What's this? Look, look at your face... Look, it's marshmallows. Natalie, what do you think of the fire? + +**Natalie Pistunovich:** It shouldn't be burning servers... + +**Mat Ryer:** No, it shouldn't be burning servers though... No, this is a fire that doesn't actually release any carbon. It's a good fire. + +**Dee Kitchen:** It's basically my GPU overheating. \[laughter\] + +**Natalie Pistunovich:** It's the sound of my computer + +**Dee Kitchen:** The money fire... My electric bill... + +**Kris Brandow:** It's some old Intel Macs, you know... We just turned them on, opened Slack, and now they've made us a nice fire. It's good. + +**Johnny Boursiquot:** \[laughs\] Just have Slack and a regular expression running; it'll generate enough heat to cook your marshmallow. + +**Kris Brandow:** And those fans can definitely fly us somewhere. We could all go visit Mat in the UK. + +1:Yeah. I mean, make sure you do go through proper passport control. Don't just fly in at any point, because that's illegal. But yeah, otherwise do, please visit; we'd love to have you. + +\[29:57\] Yeah, I remember talking about hot CPUs... The CPU Hot program that I used to have on an Amiga, and basically run it, and it made the CPU hot. And that was a program that you could have -- it was on like a front of a magazine, for some reason... "What's that doing...?" + +**Dee Kitchen:** Someone wrote another infinite loop... + +**Mat Ryer:** Yeah, there you go. They've turned their horror story into a big success story, because they got on a magazine cover with a floppy disk. So... + +**Natalie Pistunovich:** Interesting. Now with the energy costs going up here in Europe, all the heaters are becoming more expensive, because people assume they will not have gas to heat their house. Many apartment buildings have these systems with gas. So you buy like electrical heaters to warm the place in case you might need that. And they've become really expensive. So really, what you're saying is that all you need is an old computer... Which is probably cheaper at this point. + +**Mat Ryer:** I bet we see a spike in the downloads of Slack in that area... \[laughter\] + +**Natalie Pistunovich:** Or that CPU Hot... + +**Kris Brandow:** How many Electron apps can I install on one machine? + +**Mat Ryer:** Okay, has anybody got any other horror-horror stories? + +**Dee Kitchen:** I've got more... I've got one which is something that's kind of triggering. I don't know if anyone else has got sort of triggers from being horrified. One of my old bosses used to come to me, and if he started the sentence with, "What do you know about...", then I knew immediately it was downhill. It's like, "What do you know about Perl?" It was like, "Uh-oh... Where is this going?" Or "What do you know about directory services in exchange?" It's like, "Um... That they exist?" "Great! You'll do!" and off you'll be shipped to a client site. \[laughter\] + +And I ended up at one of these clients sites, and there was a customer... And it was a big, big company. And they were basically doing a split. Merger and acquisition is the normal thing you hear about; they're doing the other thing, they're splitting in two. And they basically said to the contractor, the company I worked for, to split their Active Directory, to clone it, and then rename it to the other company name, so they had a perfect copy renamed. And I just turn up on site, I know a bit of Visual Basic, a little bit of C\# at that point... How do I approach this problem? I did not know at all. + +So I get in touch with Microsoft Professional Services and go "How would you do this?" And they're like, "You don't. That's impossible. Don't do that. That's reckless and foolish, and it's not supported." And I was like, "Okay, but I'm being paid for this, and I know no better, because I'm barely in my mid-20s, and I'm gonna take a stab at this." + +And I wrote a script for the registry on one of the cloned Exchange Server machines, and I basically renamed everything. And I fired it up afterwards, and it worked. + +**Johnny Boursiquot:** Wow... + +**Dee Kitchen:** That was my day's work done. I left. I have no idea whether that worked... \[laughter\] It appeared to work. + +**Johnny Boursiquot:** "Why is your two-week notice exactly after this project ended?" "Oh, I don't know. I don't wanna know." \[laughter\] + +**Dee Kitchen:** Anyway, contractors, that's what you get. + +**Mat Ryer:** Wow... That could have just worked though. But you know, I never trust code that works first time. That's why I like a failing test before... + +**Dee Kitchen:** The only thing that I was really, really scared about -- the name of the company in the Active Directory was six characters long. And I was reasonably sure that that was a magic value. So I asked them for a new name to be the Active Directory name that was also six characters long. That's the only thing that I think was intelligent about what I did that day. The rest of was luck... + +**Natalie Pistunovich:** And lots of regexes... \[laughs\] Again. + +**Johnny Boursiquot:** Yeah... Good call. + +**Mat Ryer:** Spooky... Spoopy. + +**Kris Brandow:** Spoopy... + +**Dee Kitchen:** But yeah, that's the scary question... No one's asked me that for years. "Hey, what do you know about...?" + +**Johnny Boursiquot:** Now you'd be like "Nothing... Nothing. Nothing at all. I don't want to know anything about it. I know nothing." + +**Dee Kitchen:** Maybe that's the thing though... When you see people late in their career, and you're just like "What do you know about this?" and you think they're doing the "I've forgotten more than you'll ever learn", but no, they're actually going "I don't want to do this..." \[laughter\] + +**Mat Ryer:** Active Directory. Never heard of it, mate. Never heard of him. Thinks it's someone man's name... Active Directory. That's a weird name for a man. I think that's a man, just really selling that you know it, just to get the job... Oh, just a tip there for people that want to get into that. Like I said, I'm jet-lagged. And this is a spooky Halloween party special... How are those marshmallows looking, Kris? + +**Kris Brandow:** \[34:15\] They're toasty, and brown, and delicious. + +**Mat Ryer:** Oh, perfect. Well done. Are you gonna share, or...? + +**Kris Brandow:** No... + +**Mat Ryer:** Nah... + +**Kris Brandow:** These are my marshmallows now. + +**Mat Ryer:** Yeah. Although they're imaginary, but... But you know, I still want one. + +**Kris Brandow:** We have an infinite supply. Everybody can make their own marshmallows. + +**Mat Ryer:** Oh, yeah. Okay. + +**Natalie Pistunovich:** All you need is an infinite loop. + +**Johnny Boursiquot:** Speaking of loops, I have a scary story... But it's actually the story that I think helped me gain a new appreciation for sort of how to integrate systems, how distributed systems have sort of pitfalls, there's a trade-off for everything, and all the things that we value as best practices for dealing with integrating systems. + +So I was working for an organization, a very awesome organization, a nonprofit, which basically helps students, especially in underserved communities, sort of prepare to take sort of standardized testing, and that kind of thing. So for months, leading up to a major sort of testing day - students are going to come, sit down into their classrooms, they're going to be logging in and taking an assessment to help them with the real things. And this is like a coordination across multiple schools and everything, so that the whole county is doing this thing. So we're talking maybe like 3,000 to 4,000 students that are gonna all sit down and do this thing. + +**Mat Ryer:** Wow. + +**Johnny Boursiquot:** And basically, I'm part of the team that basically has been working on this sort of integration for months now. We had different systems talking to each other, and everything else. And development, and even in staging, everything works perfectly. \[laughs\] Systems can talk to each other, we're sending lots of traffic, keeping an eye on things, and we are observing to the best of our ability with the tools that we have on production day, basically, which is when students actually sit down to do the thing... + +What we didn't test against is basically having roughly 4,000 students trying to log into the system at the same time. \[laughs\] And because you have these systems that are talking to each other for authentication, and pulling things, whatever it is, basically, we just had a thundering herd kind of situation happening, and we didn't account for that, because all of our tests, even our integration tests and everything else - they didn't factor in that kind of scale. And I take responsibility for that, because I was one of the team leads, and basically, one of my questions was supposed to be, "What is the expected number of users and clients on this system?" And we touched on these things, but there were bottlenecks in the system that we should have better accounted for. And thankfully, we had enough of an understanding of what was happening, we had enough observability to be like, "Oh, crap, we know where the bottleneck is. We need to go do this", whatever. + +So within a matter of about an hour and a half or so, while students are waiting there - because they can't really dismiss everybody and send everybody home; we're talking like countywide 4,000+ students, all this coordination across months... To me, this remains the best and worst moment of my career, because I'm like, "Here I am, I'm supposed to be serving these kids." Often, we are so far removed from the consequences of our code, right? Good or bad. We're so far removed from it. But here I am, I knew exactly what the impact that my mistake was having on these kids, right? And they already have been given a short straw in life, and here I am just making that worse, right? + +So after that incident, I was never again, like, "What do I need to do? What do I need to learn? Who do I need to talk to?" Like, it was -- I had to level up. And any point in my career - I can't remember a single incident that has driven me to level up as much as this one, because the impact was so real. It was so in my face. Just undeniable. I thought that was scary... + +**Natalie Pistunovich:** \[38:11\] That is the $1,001 bill. + +**Mat Ryer:** Yeah. + +**Johnny Boursiquot:** Listen, I would have gladly handed over $1,000 out of my pocket like to be like, "Look, whatever this is, make it go away right now." + +**Mat Ryer:** Like, to the kids. Just give it to the kids. + +**Johnny Boursiquot:** To the kids. \[laughs\] + +**Mat Ryer:** "Uncle Johnny has messed up again. Come and collect your $20 bills, everyone... No college for you, but here's some..." Yeah, that's horrific. Yeah, but when the stakes are that high, Johnny, it's like... That is scary. + +**Johnny Boursiquot:** Yeah. And you feel awful. Like awful, awful, awful for being the responsible for that. + +**Dee Kitchen:** What do you think about engineers sort of having that sort of sense of consequence? Because it comes in multiple directions; you get salespeople going, "If you don't do this, we're gonna lose a million-dollar deal." It's not that the engineer's gonna receive no million dollars round. They're just gonna get their normal salary; they're getting the normal pay. That's a little bit abstract. It's a ton of stress. + +And likewise, when you're working in incidents, someone will turn and go, "It's affected this air traffic control signal" or "It's affected this sort of kids, or hospital, or the traffic lights are down" or whatever. Kind of unhealthy, isn't it? It's like, we've got to keep it abstract enough that -- because I get that it changes us. Probably all of those incidents changed all of us. They were all horrifying moments. But I'm also like, "That's the path to burnout, sort of accepting the consequences for all of these things." + +**Mat Ryer:** Yeah. Some value in being somewhat abstracted from the consequences. + +**Dee Kitchen:** Yeah. If you consider it too much, it just weighs so heavily. It's too serious. It's too much of a... And the things you need to do to actually get out of those situations - they become even more horrifying and scary. "What if I prolong this? What if I make it worse?" And sometimes you've just got to be a bit fearless, and you can't if you've got that sort of burden on you that we put on ourselves. + +**Kris Brandow:** I feel like this is where systems can be helpful, though. I think we as an industry are pretty bad at understanding that there can be bad consequences. It's like, something terrible happens and we're like, "Oh no, this terrible thing happened", but so much of the time, that terrible thing was completely predictable, and we just didn't predict it, because we thought it'd go fine. + +Like, I worked for a lifeguard -- when I was young, I worked as a lifeguard, and I remember one of the things that we always did is we trained a lot, but also did a lot to make sure that the environment was always a safe one. So we did a lot of that. That's why lifeguards yell at people so much, and they're like, "Don't run! Don't do these things that might result in you getting injured." I kind of feel like in software engineering we just let people do whatever, and then people slip and bash their head, and they're bleeding all over the place, and we're like, "Oh no, how did this happen?" And it's like, "Well, not only did we not tell people not to run, but we also left giant puddles of water on the floor, because we didn't put down the proper mats to make sure that, if they are running, they can do it safely..." Like, there's all these other things that you have to set up as precautions. I feel like in software engineering we just kind of don't do... And part of me wonders if we don't do that because there aren't enough consequences flowing down to the engineers. + +I feel like the number of times I've been at companies that -- I've worked at banks, and people have been like, "Well, this isn't like life or death." And I'm like, "This is affecting people's money and their livelihood. What do you mean?!" And like, that's always the thing that gets rolled out, is if it's like, "Oh, well, we're not doing things that could kill people." It's like, "Well, but we're doing things that can substantially affect people's lives", and I feel like we have to take that into account. Because when we don't, we do lots of immoral things, like run psychological experiments on people without their knowledge, and other terrible things, because we're like, "Ah, what's the harm?" + +**Mat Ryer:** I haven't done that for weeks. I don't know why you're bringing it up. \[laughter\] It's spooky, ain't it? It's a spooky show. + +**Break:** \[41:59\] + +**Kris Brandow:** I have a spooky story, I think... It feels spooky. So I'd recently joined this company, and of course, because it's the modern day, they're using Kubernetes. + +**Mat Ryer:** Classic. + +**Kris Brandow:** And of course, they're using all the shiny things with Kubernetes, so that's using Istio. No one actually knows how any of this works. It's just like, "Oh, this is what we're supposed to be using." So we have this big ol' cluster, and it's running, and our DevOps people are pulling their hair, because they hate all of this... And I start like reading through the codebase and looking at things, and I'm like, "Okay, these auth policies look a little funky." And then I go talk to people, and they're like, "We don't really have any auth policies. Everything's just kind of open right now. Like, everything in the backend can just talk to each other. There's no auth policies." And I'm like, "Are you sure? Because I see these auth policies in the codebase." And they're like, "Yeah, but we don't think they're being used for anything." I'm like, "Okay..." So I just kind of like let it go, and go about my business... And then I have like a few more of these conversations, and I'm like, "This feels weird. But you know, all these people know more than I do." + +And then months and months later, someone stumbles across this one auth policy that has no labels, and no access rules, which in Istio language means that it applies to literally everything, and allows all traffic in. So this one policy had just opened our entire API, including the public API, to the entire internet, for anybody to do anything without needing any authorization. You just needed like a JSON web token that you could easily get from anywhere. And I was just like -- so it caused all these problems, everybody's freaking out, and then they just rip that policy out and they're like, "Okay, well, without this policy, we'll be fine." But then all of those auth policies that had been sitting there, that I was like, "These look funky", all those took over, and all of those were broken. So then it broke all the APIs. So then they had to put the other policy back, and then go through, and like, go find every single auth policy within Istio, and then fix all of those auth policies, and then they could finally remove that one policy that was opening everything to the world. + +I think the total amount of time that the door was just open was about nine months, and to the knowledge that people had, nothing bad happened... But yeah, it was quite horrifying. + +**Johnny Boursiquot:** That is still a security incident requiring disclosure, I'm afraid. + +**Kris Brandow:** Yeah. + +**Dee Kitchen:** I know. I'm just like, "Is this disclosure?" \[laughter\] + +**Kris Brandow:** Yeah, it was like, "Oh... Oh, no." It taught me a lesson that like when I see funky things, I should probably bring them up a little bit. Like, "No, no, that policy is there", and that policy definitely doesn't work. And some of the broken things were like some YAML white spacing thing, where it's like something was tabbed in a little too far, and then people were just copying and pasting these policies, and then not testing them, which was like another thing that we had to go back and be like, "Please test the things that you put into the codebase... Pretty please, thank you..." + +**Mat Ryer:** Yeah, I think that also is a bit of a lesson, is if there are bits of code and you're like, "No, yeah, but that doesn't do anything now. Like, that used to be doing something and now it doesn't." It's like, either take it out... If it's really not doing something, get rid of it." + +**Johnny Boursiquot:** Prove it. + +**Mat Ryer:** Exactly. It probably is doing something... And if it isn't, maybe it should be, like in your case, Kris. + +**Kris Brandow:** Yeah. + +**Mat Ryer:** Spoopy... + +**Natalie Pistunovich:** The zombie apocalypse. + +**Mat Ryer:** Yeah. + +**Johnny Boursiquot:** ...will be caused by Istio... \[laughter\] + +**Natalie Pistunovich:** Zombie code just haunting us. + +**Dee Kitchen:** I'm never sure what's worse than those things... Like, the insecure environment, where it just like nothing applies, or the extremely secure environment. I've seen somewhere, they've been really locked down, and like everything; you've got an IP firewall rule for everything, and you're like, "I have total confidence." And then mysteriously, API calls between machines just didn't work. And it's like, "Why now?" + +And what we realized - the debugging for this went wild; it went really low. And we were down at Wireshark and we're watching what's going on, and we're watching what's going on inside the kernel, but we were turning on contract connection tracking. And this is in TCP, it's got little tables, state tables in there to keep track of the sort of TCP connections. But you can overflow these tables; we were turning them off, and every time we turned them on and off, we were toggling which IP firewall rules were actually matching or not. + +\[46:16\] So we were taking existing connections and then just randomly dropping them every time we flipped these things... But we could never observe it, and we were just there the whole time, just going "We've lost it again. There goes the connection." + +And it took us weeks of just poking around, going "What's going on? I can't see it. Ghost in the machine." Yeah, too secure is a problem. Honestly. + +**Mat Ryer:** Well, in that spirit, Dee, what's your pin number? Just give us three of the numbers. Let's play mastermind. + +**Dee Kitchen:** It's just like -- what was it, Spaceballs? 1234... \[laughter\] + +**Mat Ryer:** Yeah. No one's gonna suspect that, I think. No one's going to try that, are they? + +**Kris Brandow:** 0000... + +**Dee Kitchen:** Does it allow you? I don't think systems allow you to do that. + +**Mat Ryer:** Why? + +**Dee Kitchen:** I don't know. I'm gonna try and change my pin now... \[laughter\] + +**Mat Ryer:** Oh yeah, I don't know... Maybe because it's too easy. But I don't know. Any other horror stories before we throw a -- what do you put on fire, water? You don't do that, do you? You just let it die out on its own. No, we've got to be responsible. How are we going to sort this fire out? + +**Johnny Boursiquot:** By throwing water in the electrical equipment? + +**Kris Brandow:** Foam. We'll use foam. + +**Dee Kitchen:** Turn it off. + +**Kris Brandow:** We'll use found close Slack. Yeah, close Slack... + +**Mat Ryer:** Yeah, and the fire will die down. So before we put the fire out, has anyone got any other final horror stories? + +**Dee Kitchen:** Well, you know a few of mine... Any one you want to hear? + +**Mat Ryer:** What, do you mean in real life? \[laughter\] + +**Dee Kitchen:** I wrote one down, which actually I wrote in advance, about doing a SQL statement and accidentally double -- putting in the semicolon after the from table. So an update. So the WHERE statement didn't apply... And that was to a production system... + +**Mat Ryer:** So what was the effect of that then? So normally, you would be updating something and specifying the WHERE, which will limit what gets changed, right? + +**Dee Kitchen:** Yeah, exactly. I tweeted this when you actually asked about it, and I think no one really appreciated what it does and how it happened. But I executed the query; I was just tidying up some debt that was leftover, and it should have been really trivial... And I practiced it, and then I copy and pasted it into the console. But after I copied and pasted the first one, for whatever reason, I fingered a semicolon, and then put in the WHERE line. But that makes it two commands. So it successfully did the update column set value equals on table, and it didn't apply the WHERE. So it updated -- I think was like 90 million rows, or something... + +**Kris Brandow:** Oh, God... + +**Dee Kitchen:** And the machine was very fast; faster than I was at finding Ctrl+Z. + +**Mat Ryer:** Oh, no...! + +**Johnny Boursiquot:** And that's why you work inside of transaction blocks, kids. \[laughs\] + +**Dee Kitchen:** That's why that's advisable, but it was not what I was doing that day. The real mess though was actually sort of going "How can we restore this when our database backup was like 12 hours ago, and there's 12 hours of changes in other tables since then?" + +**Johnny Boursiquot:** Ouch... + +**Dee Kitchen:** So you're not going to sort of do anything there. So it's a pull the old sort of thing, extract that table, and then go and update all the necessary rows to the right things. It takes time... + +**Mat Ryer:** Yeah. I liked that you couldn't keep up because the machine was so fast. Is that why you now insist only on running on Raspberry Pi's? + +**Dee Kitchen:** Intel Max. Run on Intel Max. It's the solution for everything. \[laughter\] + +**Mat Ryer:** Okay, well that sound - we all heard that sound, didn't we? How would you describe that sound that we just heard? + +**Dee Kitchen:** Spooky... + +**Mat Ryer:** Natalie? Spooky, yes... Natalie, how would you describe that sound we've just heard? + +**Natalie Pistunovich:** Did I miss the sound? + +**Mat Ryer:** Yeah, we heard that sound... Kris, how would you describe it? + +**Natalie Pistunovich:** Was it the marshmallows? + +**Kris Brandow:** I think it's the sound of us closing Slack so our fire's going away. + +**Mat Ryer:** Was it marshmallows? Yeah, it's kind of a spooky sound... Wasn't it, Johnny? How would you describe that, Johnny? + +**Johnny Boursiquot:** As silent as your hairline? \[laughter\] + +**Mat Ryer:** \[50:13\] I mean, it's getting very poetic, and slightly unusual banter there... + +**Johnny Boursiquot:** You asked for it. I mean, you did ask for it. + +**Mat Ryer:** Amazing. Yeah. Is it silent because it's so far away? Like, it's in the distance... That's what we mean. Like, it's screaming, but you can't hear it... + +**Johnny Boursiquot:** It's too far back? + +**Mat Ryer:** Yeah. It's just screaming "Why, dad?! Why?!" You know what I mean? If you know you're going to look like this, don't have kids. If you know you're gonna make kids that look like me, don't have them. That's my advice... But dad didn't listen. + +**Johnny Boursiquot:** We're all talking about recessions nowadays, but you've been in one for quite some time, right Matt? \[laughter\] + +**Mat Ryer:** Oh, there we go... That was a good one... That was a good one, yeah. Good point. Okay, well... That sound of Johnny talking tells us it's time for... Unpopular Opinions! + +**Jingle:** \[51:16\] + +**Mat Ryer:** Okay... Who's gonna kick us off with the first popular opinion? Dee! You've been chosen. + +**Johnny Boursiquot:** You scared him. He was like, "What?! What happened?! What happened?!" \[laughs\] + +**Mat Ryer:** It's just sounds. I promise no one's pushing their hand around the Ouija board to make it spell out about my hairline. I don't want to hear anything from the ghosts about my hairline, Johnny... Right? So don't make it do that. Everyone put your hand on it. Let it be natural, and we'll see. Yeah, it's floated over to Dee. Okay, Dee... + +**Dee Kitchen:** It was a fix... \[laughs\] + +**Mat Ryer:** ...what's your unpopular opinion? It was a fix, yeah. + +**Dee Kitchen:** Mine comes from the last Go Time. One of the guests said that Go is brilliant, you don't have to worry about security; it does everything for you. You don't have to worry about the memory management. Everything's super cool. And I think it actually has made Go more insecure, because people are so -- they put so much trust and safety in the actual sort of language, that a lot of the basics are dropping by the wayside. + +For me, the number one thing people should do is sanitize inputs. And it's not because I wrote bluemonday but they should do it on everything. And I just don't think I see anyone doing it anywhere. There's just great, big holes in everything, but people are still there going "But we've got memory safety." Memory safety saves you from yourself, not from others. + +**Mat Ryer:** Hm... What do we think? Is that popular, or unpopular...? + +**Kris Brandow:** Well, it should be popular... Sanitize all your inputs? + +**Dee Kitchen:** It should be. But if you look at all of the open source stuff that's out there, very few people actually sanitize, check, validate their inputs. They're just like "I've mapped this input directly to a struct, and I'm going to use it." They take their form fill, and they're on their way. + +**Mat Ryer:** Yeah. One simple version of that is just a limit reader when you're reading a body, or like of a request; you can error if it's too big, and things like that. There's bits like that. But you end up doing quite a lot of that heavylifting yourself every time. + +**Dee Kitchen:** There are libraries out there where you can add tags and say "This should be a text field. It should be no longer than this length. There should be a number, and it should be no longer than this." But far too few projects do it. + +**Mat Ryer:** Do they use reflection? I think I've avoided them if they -- but it's not a great reason to avoid it. I just tend to not... + +**Dee Kitchen:** Why are you afraid of your reflection? Are you a vampire? + +**Mat Ryer:** Well, it's because I am DraCool -- Yeah because you know, that's my... Russian accent there. + +**Kris Brandow:** Because he doesn't want to see that hairline, obviously... + +**Johnny Boursiquot:** Ohh...! + +**Mat Ryer:** That's why I don't have mirrors... + +**Johnny Boursiquot:** Kris is in on it... Alright, everybody take a turn. + +**Mat Ryer:** Yeah, Kris is in on it. + +**Johnny Boursiquot:** Take a stab at Matt... \[laughs\] + +**Mat Ryer:** \[54:10\] I'm like a pinata. I'm like a really rubbish pinata. Imagine buying a pinata for kids and it's me. \[laughter\] You'd take it back, wouldn't you? You'd be like "No, we'll probably go for the unicon instead, on second thought... I should have guessed that, really..." Okay, yeah. Fine. Thanks, Kris. It's a Halloween special, you're allowed to do that... + +**Natalie Pistunovich:** Spooky pinata. + +**Mat Ryer:** Yeah, exactly. Okay, and other unpopular opinions? + +**Natalie Pistunovich:** I have one. + +**Mat Ryer:** Natalie PistunoWitch...! \[wolf howl\] They don't know I've just done that, so it just sounded like a sound effect in the background. Come on. + +**Natalie Pistunovich:** Some of the training that you should be taking occasionally throughout your career, even annually, should not be about things that are in the future, like new things, like new technologies coming and so on, but also a little bit about the back. A little bit of Assembly every now and then. It might be useful accidentally at some point in your life, because you did it, and even if not, it might -- like, you'll see patterns, because it's all the same things, with just more and more abstractions. But it's still the same things. Just seeing how it's done, and how things were solved, and how problems worked... It might help you figure out the future when you rely on the past. + +And it's unpopular -- I know, you all will not agree with me, it's unpopular. I also did not do that, and don't allocate time or budget for that. + +**Mat Ryer:** Hmm... I do know what you mean. I actually have this book called, "But how do we know?" I got it off the Amazon website. And basically, it talks about computing from the very bare beginnings, like literally logic gates, and then how you make a bit out of two NAND gates, just showing how the logic works... And then building up everything in a computer like that. And it is amazing. But yeah, that was like, not something I need. + +And actually, something else that occurred to me when you were saying that one is having like training or paying attention to things that you already think you're good at. So not just new things that are new to you. Things that you already think, "Yeah, I've got that nailed." You might be surprised; like, plenty of other things to learn. Yeah, I like that one. We'll test that one, and see if that's unpopular or not. + +**Natalie Pistunovich:** Even the way you did that is interesting, because you were already a couple of years a software developer, and then you look into gates, and so on, and that kind of helped you put this in place. When I started my degree, the first course I did was those logic gates, and everything was scary; calculus was scary. And then those gates - like, what? I don't know, just let me pass that test and leave me alone. And if it could be the other way around, start with the programming courses, and then you go about semiconductors, and then you go about circuits, and then you speak about gates - it might have been more interesting, and actually it would make more sense. Maybe for me, at least. + +**Dee Kitchen:** I think that's the path that people take accidentally. Like, those who do bootcamps, and then they gradually, eventually, over like 5-10 years become like systems engineers and are working on like the Kernel, or TLS, or something - they don't know they're doing that, but that's kind of what we're doing. They're looking back at the fundamentals. I agree it's unpopular. I don't know anyone who does that, but I think it's genius, and we should. + +**Natalie Pistunovich:** We should open a university teaches that way. + +**Kris Brandow:** I feel like that's how my career has been. I just have gotten like down the stack further and further... + +**Mat Ryer:** That's interesting, yeah. + +**Kris Brandow:** It's been fun. I've written like a few operating system kernels, like toy kernels. It was infuriating. Modern processors are terrible... But mostly because they're so old. Like, "Oh, maybe I have this code from the '70s I might need to run on my Intel 12-gen chip", or something. You never know. It's like, "No, Intel, I don't need to run code from 1980 on my new processor. Thank you." + +**Mat Ryer:** I like that too though, that people can start higher up in the stack, and can still be doing things without understanding everything. Because I've seen it, people held themselves back because they think "Well, I just don't know how all of this stuff, and I need to--" They don't know that they don't need to know it necessarily, which is why I say you sometimes don't need to know a lot of this stuff. Just sort of get on with it, and try things... But it doesn't work for everybody. I think there's so many different styles and things that people appreciate, and things that work... So I wouldn't like to force it on everybody; as we like to say, it depends. + +And I'm afraid - put down your goblets of red wine. Yes, it was definitely wine, yeah... And put also away those sandwiches of miscellaneous... No, there weren't any sandwiches. No, let's just not do the sandwiches, but we'll do the wine but, and then keep the wine, because they think it's blood. + +Okay, and that's all the time we have on today's Ghost Time. Thanks for joining us, everybody. See you next time, and stay spoopy! diff --git "a/Spooky stories to scare devs \360\237\221\273_transcript.txt" "b/Spooky stories to scare devs \360\237\221\273_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..1d44f6cbd4c6f5c786fc70e9cb1f0dc54bf8346c --- /dev/null +++ "b/Spooky stories to scare devs \360\237\221\273_transcript.txt" @@ -0,0 +1,1474 @@ +[0.00 --> 2.52] I've got one which is something that's kind of triggering. +[2.74 --> 5.40] I don't know if anyone else has got sort of triggers from being horrified. +[5.78 --> 7.46] One of my old bosses used to come to me, +[7.64 --> 9.68] and if he started the sentence with, +[10.10 --> 11.04] what do you know about? +[11.54 --> 13.62] Then I knew immediately it was downhill. +[14.10 --> 16.32] It's like, what do you know about Perl? +[16.48 --> 18.20] It's like, uh-oh, where's this going? +[18.66 --> 21.66] Or, what do you know about directory services and exchange? +[21.80 --> 23.98] It's like, um, that they exist? +[24.14 --> 24.92] Great, you'll do. +[25.34 --> 27.32] And off you'll be shipped to a client's site. +[30.00 --> 32.98] What's up, friends? +[33.06 --> 35.40] This episode is brought to you by Sourcegraph. +[35.86 --> 38.66] With the release of Sourcegraph 4.0 +[38.66 --> 41.32] and the Starship event just a few weeks behind us, +[41.54 --> 44.62] it is super clear that Sourcegraph is becoming not just CodeSearch, +[44.62 --> 47.76] but a full-on code intelligence platform. +[48.18 --> 48.96] And I'm here with Joel Cortler, +[49.06 --> 51.30] Product Manager of Code Insights for Sourcegraph. +[51.58 --> 54.78] Joel, this move from CodeSearch to Code Intelligence +[54.78 --> 56.22] is a really big deal. +[56.48 --> 59.04] How would you explain this feature, Code Insights, +[59.04 --> 61.38] if you're just talking to folks in the hallway track +[61.38 --> 62.58] of your favorite conference? +[63.16 --> 64.98] I would really start with the technical +[64.98 --> 66.40] because before I was a product manager, +[66.50 --> 67.42] I used to be an engineer as well. +[67.70 --> 70.38] And it's really cool and exciting just to be able to say, +[70.54 --> 72.58] we're going to turn your code base into a database. +[73.04 --> 75.38] And the structured language that you need to interact +[75.38 --> 77.44] is just the ability to write a code search. +[77.74 --> 79.66] You know, literal search, that's totally fine. +[79.82 --> 80.86] Regular expression, you know, +[80.88 --> 82.30] that'll give you a few more advanced options, +[82.56 --> 83.46] even a structural search. +[83.46 --> 86.60] But the number of long-tail possibilities it unlocks, +[86.76 --> 89.02] truly the journey of building this product +[89.02 --> 90.98] was just saying, well, we've just unlocked, +[91.20 --> 93.20] you know, an infinite number of possibilities. +[93.64 --> 95.54] We got to figure out some immediate use cases +[95.54 --> 97.48] so we can start to, you know, invest in this product, +[97.60 --> 98.30] build it and sell it. +[98.74 --> 100.12] But we're only getting started +[100.12 --> 101.42] in terms of the number of uses +[101.42 --> 102.72] that we're uncovering for it. +[103.06 --> 104.88] The story I told you about discovering, +[104.96 --> 106.32] like, version tracking turned out +[106.32 --> 107.40] to be a really important use case +[107.40 --> 108.96] that wasn't even on our roadmap six months +[108.96 --> 110.44] prior to discovering that +[110.44 --> 111.76] as we were already planning to launch this product +[111.76 --> 112.92] until we talked to enough folks, +[113.22 --> 114.12] realized this was a problem, +[114.12 --> 115.26] and then found, well, oh, +[115.26 --> 117.38] that's like a simple regular expression capture group +[117.38 --> 118.46] that you can just plug right in +[118.46 --> 119.94] because we really built this system +[119.94 --> 122.10] to not limit the power of what we built. +[122.18 --> 123.08] We don't want to give you, like, +[123.20 --> 124.08] three out-of-the-box templates +[124.08 --> 125.26] and you can only change, like, +[125.28 --> 126.18] one character or something. +[126.30 --> 128.10] It's truly, like, the templates are there +[128.10 --> 129.20] to hold your hand and get you started, +[129.32 --> 131.18] but if you can come up with anything +[131.18 --> 132.40] you want to track in your code base, +[132.46 --> 133.48] you can do that with Code Insights. +[133.74 --> 134.86] I love it. Thank you, Joel. +[134.98 --> 138.52] So right now there is a treasure trove of insights +[138.52 --> 139.60] just waiting for you. +[139.88 --> 141.32] Living inside your code base, +[141.32 --> 144.34] your code base is now a querible database +[144.34 --> 145.36] thanks to Sourcegraph. +[145.78 --> 148.32] This opens up a world of possibilities for your code +[148.32 --> 149.92] and the intelligence you can gain from it. +[150.20 --> 152.04] A good next step is to go to +[152.04 --> 155.46] about.sourcegraph.com slash code dash insights. +[155.76 --> 157.02] The link will be in the show notes. +[157.38 --> 159.86] See how the teams are using this awesome feature. +[160.06 --> 164.94] Again, about.sourcegraph.com slash code dash insights. +[165.30 --> 167.58] Again, this link is in the show notes. +[167.58 --> 183.34] Let's do it. +[183.92 --> 185.00] It's go time. +[185.00 --> 186.92] Welcome to Ghost Time. +[187.64 --> 189.56] Your source for spooky conversations +[189.56 --> 191.24] from around the digital campfire. +[191.70 --> 194.62] Special thanks to our partners Fastly and Fly.io +[194.62 --> 197.42] for helping us bring you Go Time each and every week. +[197.80 --> 198.08] Okay. +[198.52 --> 199.88] Here we go. +[199.88 --> 212.14] Hello. +[212.14 --> 212.44] Hello. +[213.40 --> 216.10] And welcome to Ghost Time. +[216.48 --> 217.28] I'm Matt Ryer. +[217.66 --> 220.68] Today we're talking about tech horror stories. +[220.96 --> 222.76] I'm joined, as ever, +[223.40 --> 224.06] Johnny Boo. +[224.36 --> 225.52] Johnny Borsico's here. +[225.60 --> 226.08] Hello, Johnny. +[226.74 --> 228.02] Hello, Matt. +[228.02 --> 231.18] Welcome to the spooky Go Time episode. +[231.78 --> 232.90] Are you getting in the spirit of it? +[232.98 --> 235.42] You've really got to get in the spirit of it. +[236.30 --> 236.98] Are you? +[237.12 --> 237.32] No. +[237.80 --> 238.02] Yeah. +[239.14 --> 242.80] We're also joined here by Chris Brando. +[244.58 --> 245.48] Spooky ghost. +[245.56 --> 246.30] Hello, Chris. +[247.04 --> 247.34] Hello. +[247.82 --> 248.92] I'm back again. +[249.14 --> 249.48] Finally. +[249.88 --> 250.20] Yes. +[250.28 --> 251.26] Welcome back again. +[251.82 --> 254.74] We're also joined by your friend and mine, +[254.86 --> 256.32] Natalie Pistonowitch. +[256.52 --> 258.00] Hello, Natalie. +[258.56 --> 258.94] Hello. +[260.38 --> 260.78] Yeah. +[260.96 --> 261.98] Getting into the spirit. +[264.20 --> 264.64] Yeah. +[265.04 --> 266.62] We have a special guest joining us. +[266.88 --> 267.88] You're not going to believe this. +[268.32 --> 269.12] It's Spoopy D. +[269.54 --> 270.30] D Kitchen. +[270.50 --> 271.16] Welcome, D. +[271.52 --> 272.18] Thank you. +[272.30 --> 273.66] I'm enjoying being here. +[274.04 --> 275.22] I even got the backdrop for it. +[275.60 --> 275.88] Hmm. +[275.96 --> 276.24] Good. +[276.36 --> 278.60] Well, that's a good start because it literally just started. +[278.76 --> 282.20] So if you, I mean, really the only way is down now in a lot of ways, +[282.24 --> 284.42] but hopefully we don't go there. +[284.42 --> 286.92] But we are talking about scary things today. +[287.00 --> 288.84] How are you generally with scary things, D? +[289.20 --> 290.14] That's my career. +[291.74 --> 292.42] All of it. +[292.78 --> 293.00] Yeah. +[293.66 --> 294.10] Okay. +[294.24 --> 294.74] Anyone else? +[294.86 --> 295.78] Anyone scared of ghosts? +[296.40 --> 297.64] I'm scared of ghost time. +[298.34 --> 298.58] Yeah. +[298.62 --> 299.56] You're scared of ghost time. +[300.06 --> 301.72] I'm actually scared of horror movies. +[301.84 --> 302.70] I don't really watch them. +[302.96 --> 303.26] Oh. +[303.40 --> 303.82] Oh, yeah. +[303.98 --> 304.32] Same. +[304.50 --> 305.68] Heebie-jeebies for me. +[306.04 --> 306.32] Hmm. +[306.48 --> 307.46] I just find them boring. +[307.92 --> 308.18] Yeah. +[308.80 --> 309.90] But you come from the industry. +[310.12 --> 313.06] I remember you saying, Chris, that you see a movie and you kind of, +[313.22 --> 315.78] the first three minutes and you know exactly how it's going to be laid out. +[316.24 --> 316.56] Yeah. +[316.56 --> 320.78] It's the curse of having a creative writing degree and specializing in screenwriting. +[321.08 --> 323.92] It's just, all movies are just kind of ruined. +[325.12 --> 325.56] Yeah. +[325.66 --> 326.58] It's all generics. +[326.72 --> 327.32] That's what I said. +[327.34 --> 329.00] You'd be happier if you're just an idiot. +[329.56 --> 330.44] I've always said that. +[333.18 --> 335.48] Is this something you know from personal experience, Matt? +[337.72 --> 338.52] Shots fired. +[338.52 --> 344.10] I also don't really like horror films, especially if there's any kind of contradiction in it. +[344.18 --> 345.14] I can't deal with that. +[345.42 --> 350.44] Like, if there's an invisible thing that can grab you, first of all, it's invisible. +[350.62 --> 351.34] It would be blind. +[351.50 --> 352.24] We covered this. +[352.54 --> 355.32] But also, if it can grab you, you can grab it. +[355.42 --> 356.24] You can hurt it. +[356.34 --> 357.66] Like, it's not fair. +[357.96 --> 361.62] It's like when the physics don't apply generally and I'm just out. +[361.76 --> 364.12] And I just tell everyone in the cinema, I'm like, sorry, everyone. +[364.22 --> 365.06] I can't stay. +[365.60 --> 368.02] I've got to go because of the inconsistencies of the physics. +[368.02 --> 369.86] No, just go and get some popcorn and go. +[370.18 --> 371.66] Do you get sweet or salty popcorn? +[372.66 --> 373.02] Salty. +[373.42 --> 374.14] Do you have a choice? +[374.38 --> 375.26] Or is it always salty? +[375.48 --> 376.12] No, you have a choice. +[379.00 --> 379.88] You have a choice. +[380.24 --> 380.48] Yeah. +[381.08 --> 381.68] What do you mean? +[381.74 --> 384.70] Like, there's the police are going around saying, hey, are you only having salty? +[385.06 --> 385.54] What are you doing? +[386.38 --> 392.42] I only discovered in my late 20s that some other countries sell popcorn that is not just salty in the cinema. +[392.90 --> 393.58] Oh, right. +[393.82 --> 394.12] Yeah. +[394.12 --> 397.50] And then I came to the US and then it's like not just two flavors, but 15. +[398.02 --> 398.56] Yeah, of course. +[398.68 --> 398.78] Yeah. +[399.68 --> 400.04] Yeah. +[400.44 --> 401.70] That's a horror story there. +[402.06 --> 402.22] Yeah. +[402.22 --> 409.10] You can choose individual bits of corn and have them different flavors and just have as many as you want. +[409.24 --> 409.86] You just program it. +[409.90 --> 412.18] You do it as an app and then it pops it on demand. +[412.86 --> 415.96] You say that, but we do have soda machines where you can choose your own flavor. +[416.34 --> 416.70] Yeah. +[416.98 --> 417.74] I've seen that. +[417.82 --> 419.48] Those freestyle Coke things. +[419.60 --> 419.68] Yeah. +[419.68 --> 421.04] Has anyone come up with a good one yet? +[421.10 --> 422.80] Because I imagine they're all terrible. +[423.34 --> 425.36] But you reckon someone's like, do you know what? +[425.38 --> 429.64] I've accidentally pressed these three and I've made a brand new flavor that never existed before. +[430.00 --> 430.34] Well, no. +[430.46 --> 435.94] I think they make it so you can't make any truly terrible tasting ones because that would be perhaps bad for them. +[436.28 --> 436.30] So. +[437.26 --> 437.82] Oh, really? +[438.06 --> 438.42] Clever. +[438.42 --> 439.28] How do they do that? +[440.00 --> 440.92] Oh, well, we'll never know. +[441.52 --> 444.78] Well, speaking of horror stories, let's get into this, shall we? +[444.82 --> 447.16] Who wants to kick us off with a spooky story? +[447.30 --> 454.08] Oh, by the way, we should actually introduce Dee because Dee wrote a package that I think a lot of people here will be familiar with. +[454.42 --> 455.56] Can you tell us about Blue Monday? +[456.58 --> 457.40] Ah, Blue Monday. +[457.90 --> 462.34] It's named because there was a package called Black Friday, which is all the best markdowns. +[462.34 --> 463.52] And it's a markdown package. +[464.10 --> 468.76] And after you've generated markdown, markdown can include HTML, which makes it dangerous. +[469.20 --> 474.72] It's probably you're using this because you've got user-generated content and you want to sanitize it. +[474.72 --> 478.88] So Blue Monday is named after the New Order song, but follows Black Friday. +[479.46 --> 481.52] And it basically sanitizes HTML. +[481.78 --> 488.54] It's the only Go package that sanitizes HTML, which is a foolish and reckless thing to attempt to take on. +[488.92 --> 489.66] But that's what I did. +[490.90 --> 491.40] Amazing. +[491.40 --> 494.68] And what do you like about it and what don't you like about it? +[495.00 --> 496.12] I like that it works. +[496.54 --> 498.38] It's all a streaming parser. +[498.62 --> 499.60] It's got fixed memory. +[499.78 --> 503.22] So you can use it quite comfortably in a lot of situations and throw a lot through it. +[503.52 --> 507.42] I don't like when people tell me there's security issues with it and then I have to go, +[507.82 --> 509.98] oh, I'm supposed to take this open source thing seriously. +[510.64 --> 511.86] I do appreciate it. +[511.90 --> 514.00] I should actually say I do appreciate security reports. +[514.20 --> 520.54] But at the same time, you never can predict when they're going to turn up and you never know what kind of worms you're going to open +[520.54 --> 521.96] to try and actually figure it out. +[521.96 --> 527.82] Yeah, there must be a lot of responsibility, actually, because it is a package that is used and quite trusted. +[527.82 --> 529.70] Yeah, it's used. +[530.28 --> 534.90] I don't know how many stars it's got, but the stars don't betray the number of times it's used. +[534.90 --> 537.28] Like it's used in Hugo and everyone uses Hugo. +[537.54 --> 540.20] And this is the HTML sanitizer that keeps Hugo safe. +[540.36 --> 542.04] And it's used in so many things. +[542.12 --> 544.22] It's got literally thousands of dependencies. +[544.68 --> 546.42] Do I take it seriously and stressfully? +[546.62 --> 548.02] No, no, I don't. +[548.26 --> 553.96] I figure that if someone is brave enough to take an open source project with a MIT license or BSD3 clause, +[554.04 --> 557.90] whatever it is, and incorporate into their production software, that's on them. +[558.66 --> 559.44] Okay, fair enough. +[559.82 --> 562.44] Well, I have done that, but no, good to know. +[562.44 --> 565.44] I genuinely have used it, though, quite a few times. +[565.86 --> 570.72] So I like it because it's like you opt in to what you want to support, don't you? +[570.78 --> 573.22] Like you explicitly say the things that you want to allow. +[573.82 --> 577.68] Yeah, there's no way of defining what makes a good HTML sanitizer. +[577.96 --> 580.06] Everyone's got a different rule, depending on their use case. +[580.44 --> 584.52] But the Java OWASP, Open Web Application Security thing, +[585.04 --> 588.42] their sanitizer defined this really beautiful interface for sort of going, +[588.42 --> 592.74] I want to allow images, but I don't want to allow this images that end in .gif. +[593.16 --> 596.32] And I copied their API and then extended it for my own use. +[596.78 --> 598.66] So, yeah, it's a really good way of doing it. +[599.96 --> 600.40] Nice. +[600.64 --> 606.62] Okay, well, I'm going to tell you about a horror story in tech of mine that happened quite recently. +[606.62 --> 612.74] I have this project which interacts with Twitter and interacts with the Twitter API. +[613.14 --> 616.26] And so it poll results and then compares them and stuff. +[616.88 --> 619.14] And that's just one of the things it does at a regular interval. +[619.70 --> 624.18] And then what happened recently was something happened where like the API key changed +[624.18 --> 625.84] and that request failed. +[626.20 --> 631.38] And because of the way I was doing it in GCP, it meant essentially that it would retry. +[631.56 --> 634.46] And because it was scheduled, it kept compounding. +[634.46 --> 640.14] And this ran up a $1,000 bill for me, for yours truly. +[640.58 --> 644.22] $1,000 given, paid, gone. +[644.76 --> 646.90] So that's a bit of a tech horror story. +[648.06 --> 649.28] Any advice for me? +[649.74 --> 650.68] Is it tax deductible? +[653.50 --> 653.94] Probably. +[654.34 --> 656.60] AWS famously was funded if you get something wrong. +[656.68 --> 657.70] Did GCP not do that? +[657.96 --> 658.46] I don't know. +[658.50 --> 659.42] It's quite recent. +[659.54 --> 661.10] I haven't yet tried that. +[661.24 --> 663.50] Do you think I should get in touch with support and see if they'll... +[663.50 --> 665.08] $1,000 would motivate me. +[665.66 --> 666.02] Yeah. +[666.58 --> 667.00] There you go. +[667.06 --> 667.44] $1,000. +[668.40 --> 668.76] Okay. +[668.84 --> 671.54] Well, I'll try it and I'll let the listeners know how we get on. +[671.82 --> 672.52] Could have been worse. +[672.94 --> 673.18] Right? +[673.62 --> 674.66] Could have been $2,000. +[674.98 --> 675.30] Yeah. +[675.68 --> 676.86] Well, just $1,001. +[677.50 --> 678.66] Would be worse, wouldn't it? +[678.90 --> 679.20] It would. +[679.58 --> 681.58] What would you do, Johnny, if you saw that? +[682.00 --> 683.14] I'd call you and say, hey. +[683.14 --> 685.20] You got a grand. +[685.48 --> 685.96] I hear you. +[686.08 --> 686.66] You're loaded. +[687.28 --> 691.00] You're just wasting $1,000 here, $1,000 there on your bugs and stuff. +[691.30 --> 694.34] Honestly, when I found out about it, I wanted to just karate chop the air. +[694.80 --> 698.52] That was the kind of spooky reaction I had to it. +[698.60 --> 700.16] Just like, whew, in the air. +[700.64 --> 700.92] Angry. +[701.78 --> 703.42] But yeah, it's a good lesson though, isn't it? +[703.48 --> 706.34] Like set budgets and stuff on your things. +[706.70 --> 707.32] Do set an alarm. +[707.44 --> 707.60] Yeah. +[707.94 --> 708.16] Yeah. +[708.34 --> 709.06] Budget alarms. +[709.52 --> 709.88] Observability. +[709.88 --> 711.32] Yes, yes. +[711.38 --> 712.68] And you know a thing or two about that, yeah? +[714.68 --> 715.04] Yeah. +[715.88 --> 716.24] Okay. +[716.46 --> 718.34] Who can beat my $1,000 bill? +[718.68 --> 719.74] Not a $1,000 bill. +[719.92 --> 721.06] Oh yeah, it was a $1,000 bill. +[721.14 --> 724.06] But that makes it sound like it was one thing, doesn't it? +[724.06 --> 725.46] Like a single bill. +[725.90 --> 727.28] It had $1,000 on it. +[727.58 --> 728.40] So it's not that. +[728.54 --> 730.68] It was just paid through bank transfer. +[731.40 --> 732.66] Okay, who's got another one? +[733.14 --> 736.72] I have one that could have cost many thousands of dollars. +[737.00 --> 737.86] Oh, Johnny. +[737.86 --> 738.94] It wasn't spotted. +[739.36 --> 739.44] Okay. +[739.44 --> 745.86] So one of the things you can do with function as a service things like AWS Lambda, for example, +[746.16 --> 751.08] is that you can trigger a Lambda when you write an object to an S3 bucket. +[751.36 --> 759.42] Word of advice, do not have your Lambdas write to a bucket that they are themselves responding to. +[761.32 --> 761.88] Oh. +[761.88 --> 765.94] Because that's going to give you a very nasty bill. +[766.46 --> 766.66] Yeah. +[766.72 --> 768.96] And you will not like what you see. +[769.60 --> 772.36] So yeah, thankfully, Budget Alarms came to the rescue. +[772.94 --> 773.20] Uh-huh. +[773.52 --> 774.04] There you go. +[774.08 --> 774.80] That's the lesson there. +[774.80 --> 777.98] So what happens is an object goes in the first time. +[778.06 --> 779.10] That triggers the Lambda. +[779.30 --> 784.14] The Lambda then writes something into that same bucket, which then triggers another Lambda. +[784.40 --> 784.64] Right. +[784.72 --> 785.96] Which then writes something. +[786.32 --> 788.88] And like, how quickly does that get out of hand? +[789.74 --> 790.26] Very quickly. +[791.02 --> 791.10] Yeah. +[791.10 --> 799.30] Like, if you want to see how well Lambda scales on your own dime, you can do that. +[799.30 --> 802.92] And yeah, it'll cost you money very quickly. +[803.32 --> 803.66] Wow. +[803.90 --> 804.10] Yeah. +[804.58 --> 804.90] Yeah. +[805.02 --> 805.32] Okay. +[805.68 --> 806.18] Pretty good. +[806.52 --> 808.78] But yeah, the alerts came to the rescue. +[809.22 --> 809.56] Nice one. +[809.66 --> 809.86] Mm-hmm. +[810.20 --> 810.50] Mm-hmm. +[810.86 --> 811.42] Okay. +[811.76 --> 812.86] Anyone else got one for us? +[813.28 --> 814.78] I've got another infinite loop one. +[815.08 --> 816.60] Are we allowed to name company names? +[816.76 --> 817.22] I don't know. +[817.36 --> 818.82] Maybe it's internal and I shouldn't. +[818.94 --> 819.10] Yeah. +[819.22 --> 819.58] I don't know. +[819.58 --> 823.68] I worked for a certain company which has an orange logo that has a bit of a light flying +[823.68 --> 826.04] shining behind it, and they man in the middle of the entire internet. +[826.04 --> 832.58] Now, with that in mind, when I was working for said company and their DDoS team, we didn't +[832.58 --> 833.16] DDoS people. +[833.30 --> 834.70] We were protecting against DDoSers. +[834.80 --> 834.88] Yeah. +[835.00 --> 835.40] I've wounded. +[835.74 --> 836.04] I don't know. +[836.10 --> 836.82] The DDoS team. +[837.02 --> 839.22] I just suddenly realized, I was like, that's the opposite of what we're doing. +[839.66 --> 843.22] Now, we were trying to protect, and they have a system, right? +[843.48 --> 848.06] They've got all these 200 pops or points of presence and thousands and thousands of servers. +[848.62 --> 851.66] And every single one of these is protecting some of the traffic. +[851.84 --> 854.06] Each machine can do like 20,000 requests per second. +[854.06 --> 859.72] And yet they need to be able to actually show the value back to the customer and make the +[859.72 --> 860.82] sort of decision centrally. +[861.02 --> 864.52] So you send all the logs somewhere, and they're all being sent to one data center. +[865.34 --> 869.64] So what you end up with is like, if you're doing globally 10 million requests per second, +[869.86 --> 872.58] you get 10 million log lines per second in one place. +[872.98 --> 874.10] Ah, nice. +[874.10 --> 879.72] Certain customer on a certain point in time, industry and type to be non-disclosed, +[880.06 --> 884.50] wrote an infinite loop in their client and basically spiked 8 million requests per second +[884.50 --> 885.88] on top of our normal load. +[886.00 --> 886.40] Oh, wow. +[886.58 --> 889.62] And they basically broke our logging, entire visibility. +[890.00 --> 895.06] So they were effectively under attack, but now flying blind because we couldn't see anything +[895.06 --> 896.14] because they'd broken all the login. +[896.66 --> 897.24] Oh, no. +[897.84 --> 899.02] That was not a good day. +[899.68 --> 899.96] Yeah. +[900.24 --> 901.40] That one doesn't sound fun. +[901.52 --> 902.08] What happened? +[902.72 --> 905.56] We figured out which customer it was, but we couldn't figure out the rest. +[905.68 --> 908.58] But we asked them what they'd done, and they figured out that bit and stopped it. +[908.72 --> 909.14] Oh, wow. +[909.32 --> 910.06] They fessed up to it. +[910.10 --> 910.84] They owned up to it. +[911.18 --> 912.26] Somebody run infinite loop. +[912.58 --> 912.92] Yes. +[913.58 --> 914.58] I think they realized. +[915.32 --> 917.24] They must have seen what was happening on their side. +[917.24 --> 922.54] So they didn't pull a, well, I will not name names, but they didn't blame any intern? +[923.64 --> 927.52] Oh, we've got a certain thing where we actually, an intern did that. +[927.66 --> 928.48] I heard that intern. +[928.60 --> 929.68] He's actually a really good guy. +[930.00 --> 931.98] He's become a full-time engineer in that team, you know. +[932.32 --> 932.98] He's really good. +[933.22 --> 934.74] Learned a lesson that none of us will replicate. +[935.08 --> 935.44] Oh, yeah. +[935.52 --> 936.48] I love interns. +[936.48 --> 941.34] I just don't like to throw them under the bus when something goes wrong with my company. +[942.62 --> 944.26] No, that one was interesting. +[944.26 --> 946.32] Also, that said, a man-in-the-middle company. +[946.68 --> 947.86] We had a system. +[948.12 --> 949.22] The system was brilliant, right? +[949.26 --> 952.82] You could send an instruction to any machine in the world in under 10 seconds, and every +[952.82 --> 954.36] machine received the same instruction. +[955.36 --> 959.08] And that's great when you want to say there's a new domain name, because you tell the whole +[959.08 --> 959.92] world at the same time. +[960.20 --> 965.04] But it's really bad when you say there's a new way to stop traffic, and we've made a +[965.04 --> 965.94] greedy regex. +[966.32 --> 967.90] And the greedy regex was the problem. +[968.32 --> 972.54] Now, frankly, the system shouldn't have allowed it, but the system did allow it. +[972.54 --> 975.36] And we were all at lunch. +[975.42 --> 976.54] It was an all-hands lunch. +[976.86 --> 981.98] And the next thing we know, we just get people running in going, the internet's down. +[982.56 --> 987.28] Because we used our own systems, and we lost everything internally at the same time. +[988.42 --> 989.82] It was a... +[989.82 --> 990.64] That was hard, too. +[991.12 --> 992.78] I feel like there are many lessons there. +[993.20 --> 994.24] There was a lot of lessons. +[994.60 --> 995.62] What were we having for lunch? +[998.18 --> 999.20] Lunch went cold. +[999.20 --> 1001.40] That's scary, isn't it? +[1001.52 --> 1001.96] But hang on. +[1002.06 --> 1005.18] So just can you explain to someone who doesn't know what a greedy regex is? +[1005.34 --> 1006.14] What do you mean by that? +[1006.68 --> 1007.00] Yeah. +[1007.42 --> 1011.16] A greedy regex, I mean, if you do something like .star, what you're saying is match any +[1011.16 --> 1012.28] character any number of times. +[1012.36 --> 1017.36] But if you do .star, .star, you've now exploded this any character any number of times, followed +[1017.36 --> 1019.26] by any character any number of times. +[1019.36 --> 1021.76] And what you're doing is you're increasing the CPU computation. +[1021.76 --> 1028.24] You still put the same fixed input in, but what it can match is now you've doubled the +[1028.24 --> 1029.84] possibility in just that one go. +[1030.42 --> 1031.80] And essentially, that's what happened. +[1031.98 --> 1034.70] But some of the inputs were web pages and web traffic. +[1035.22 --> 1037.20] So they weren't small. +[1037.32 --> 1038.30] They were quite large inputs. +[1038.68 --> 1040.74] And under that condition, they consumed all the CPU. +[1041.52 --> 1045.74] So wherever this rule was applied, and we had shipped it globally to every single website, +[1045.90 --> 1049.20] every single bit of traffic, we fried every single machine instantly. +[1049.20 --> 1053.24] So it was about four hours for us to recover from that. +[1053.88 --> 1056.62] And the teams I saw, they did interesting things. +[1056.74 --> 1059.98] We were connecting directly to machines and looking at the Prometheus on them because we +[1059.98 --> 1061.26] had no other observability. +[1061.92 --> 1062.14] Wow. +[1062.64 --> 1064.40] And what was the impact of that? +[1064.54 --> 1065.66] How many people were affected? +[1066.24 --> 1067.30] Everything was affected. +[1067.70 --> 1069.04] We knocked out a lot. +[1069.88 --> 1073.30] DNS, TLS, HTTP, everything. +[1073.54 --> 1075.22] It was one of those nightmare scenarios. +[1075.72 --> 1078.70] And you sit there as a company, you sit there and you sort of go, what are these meteorites? +[1078.70 --> 1082.76] The dinosaurs were made extinct by a meteorite as a company or product service offering. +[1083.06 --> 1084.82] What's a meteorite that's going to hit us? +[1085.22 --> 1087.96] At that company, we were hit by every meteorite we predicted. +[1088.50 --> 1089.34] It survived. +[1089.96 --> 1093.98] But still, on the days when they hit, it lays waste to everything. +[1094.50 --> 1095.44] And everyone has them. +[1095.64 --> 1098.74] The thing that you've got to realize is when you're there, you've got to sympathize with, +[1099.08 --> 1101.38] you know, you can externally see another company go through this. +[1101.76 --> 1102.90] They're having a bad day. +[1102.98 --> 1106.04] And you've got to sympathize because one of those meteorites is going to hit you one day. +[1106.04 --> 1110.56] Yeah, we see the hug-ups goes around often on social media and things. +[1110.66 --> 1114.60] People sending their support in those difficult times. +[1115.04 --> 1119.08] Yeah, there's a good tradition of sending cakes to each other to sort of go thinking of you. +[1119.20 --> 1119.38] Yeah. +[1119.54 --> 1122.22] And trying not to have your salespeople the ambulance chase. +[1122.58 --> 1123.34] Mm-hmm. +[1123.34 --> 1128.02] How well was your break glass procedures documented? +[1128.62 --> 1129.42] It was pretty good. +[1129.56 --> 1134.28] We were lucky that this happened during a London lunch hour where all of the SR team, +[1134.66 --> 1136.46] the original SRE team were there. +[1136.78 --> 1138.56] It's possible for a few people to break glass. +[1138.66 --> 1142.86] They could do so in about five minutes once we actually understood what we actually had to do. +[1143.30 --> 1147.00] It took about 20 minutes for us to gain any visibility and sort of understand, +[1147.54 --> 1149.12] hey, it's this feature. +[1149.34 --> 1150.12] Go turn that off. +[1150.56 --> 1151.66] Regular expressions, huh? +[1151.66 --> 1152.90] They're still hard. +[1153.22 --> 1154.06] They're hard for everyone. +[1155.20 --> 1156.04] Easy to write. +[1156.20 --> 1157.32] Hard to understand what they're doing. +[1157.50 --> 1158.58] Why are they called that? +[1158.76 --> 1160.90] Because what is regular about them? +[1161.58 --> 1162.70] That's out of my domain. +[1162.90 --> 1164.54] I don't know if anyone's got the answer for that. +[1165.06 --> 1165.90] No, genuinely have. +[1166.36 --> 1170.12] I just watched a talk from Strangeloop about regular expressions. +[1170.84 --> 1172.68] And the speaker did go into this. +[1172.70 --> 1175.10] And I've completely forgotten what she said. +[1175.40 --> 1177.66] But we can probably put that talk in the show notes. +[1177.66 --> 1180.22] It was a really good one about just the history. +[1180.46 --> 1182.26] Ken Thompson came up, which is pretty cool. +[1182.42 --> 1183.60] I'm like, oh, I know that dude. +[1184.52 --> 1184.70] Yeah. +[1184.78 --> 1189.34] But no, it has to do with mathy things and finite automata and all of that. +[1190.18 --> 1190.34] Yeah. +[1190.34 --> 1196.98] When I was a junior, which is the closest to intern I was, I was working with a team lead. +[1197.20 --> 1199.62] And when we deployed something together, we looked at it. +[1200.08 --> 1201.88] And I forget what it was exactly. +[1201.98 --> 1209.46] But this is a company that receives a lot of pings from the SDK of the many clients. +[1209.46 --> 1210.52] And it's all real time. +[1210.52 --> 1216.74] And if that is not logged, then the entire transaction, like flow, user flow is gone forever. +[1217.78 --> 1220.66] And we deployed something we worked on together. +[1220.80 --> 1222.00] We worked on it for half a day. +[1222.10 --> 1222.80] We tested it. +[1222.86 --> 1225.66] We did all the good practices because that's how you do with a junior, right? +[1225.72 --> 1227.12] You want to show that you're very thorough. +[1227.26 --> 1233.92] You go for all the tests, deploy, look at all the metrics and see that it behaves as expected. +[1234.54 --> 1236.50] And then he went to lunch and then I stayed. +[1236.94 --> 1237.78] Da-da-da. +[1237.78 --> 1238.78] Da-da-da. +[1239.82 --> 1242.34] And then I proceeded to do something else. +[1242.38 --> 1246.58] And then suddenly a weird behavior started pinging Slack, all the monitoring channels. +[1246.74 --> 1247.36] It's like something wrong. +[1247.44 --> 1247.94] Something's weird. +[1248.44 --> 1254.20] And then some of the colleagues that were there tried to see where it comes from. +[1254.90 --> 1257.80] And we couldn't figure this out in 15 minutes. +[1257.80 --> 1263.40] And then bravely came to the head of the DevOps team. +[1263.54 --> 1265.60] There was no SRE team at the time. +[1265.60 --> 1267.80] And I said, I think it's this thing that we did. +[1267.90 --> 1268.76] Can I revert it? +[1269.18 --> 1275.08] Nobody else, from all the other senior people that don't have another better word than brave, +[1275.14 --> 1276.20] but I don't want to use the word brave. +[1276.52 --> 1280.44] Nobody else wanted to do anything about that because nobody was sure. +[1280.54 --> 1282.44] And then I was like, let's do this. +[1282.52 --> 1283.46] Let's try it worse. +[1283.68 --> 1285.34] It cannot get much worse than that, right? +[1285.50 --> 1287.38] And reverting that indeed succeeded. +[1287.54 --> 1288.66] And then we were all very happy. +[1288.66 --> 1290.42] And then I was like, I think I know how to fix it. +[1290.48 --> 1290.94] Can I try? +[1291.04 --> 1292.54] And then he looked at me and said, no. +[1293.66 --> 1294.44] Stay away. +[1297.36 --> 1299.10] I went to the next level of brave. +[1299.44 --> 1299.76] Yeah. +[1300.02 --> 1302.54] Well, what a great way to learn stuff though, isn't it? +[1303.66 --> 1304.20] Break them. +[1304.42 --> 1304.54] Yeah. +[1305.20 --> 1308.32] How often does that memory come back to haunt you, Natalie? +[1309.14 --> 1310.62] Every time I'm asked to please. +[1311.02 --> 1311.24] Yeah. +[1312.36 --> 1315.16] Every time I get to speak with other junior people. +[1315.16 --> 1318.06] So to give the good example of it, obviously you will break something. +[1318.42 --> 1320.90] So be reasonable about your expectations. +[1322.52 --> 1322.88] Nice. +[1323.06 --> 1327.18] I'm curious if you have a way, like now that you're older and wiser and, you know, +[1327.20 --> 1329.08] you've been through the experience, which is a great teacher. +[1329.44 --> 1334.72] I'm wondering, do you have strategies now for doing things that are scary, +[1334.90 --> 1335.92] that could break things? +[1335.92 --> 1337.94] Like, do you have a strategy for tackling that now? +[1338.34 --> 1341.00] The thing is we did everything right at the time, right? +[1341.00 --> 1343.66] So we did all the tests we could think of. +[1343.66 --> 1348.86] We thought, what do we expect in the logs and the monitoring and the dashboard? +[1349.26 --> 1350.02] And we observed. +[1350.62 --> 1353.36] So the only thing that I would do different is not to plot it for lunch. +[1356.42 --> 1358.58] So you can, I guess, observe for longer. +[1359.56 --> 1360.94] Speaking of spooky things, right? +[1361.16 --> 1362.22] What are we having for lunch? +[1362.96 --> 1363.18] Yeah. +[1363.18 --> 1364.46] I don't think I had lunch that day. +[1364.96 --> 1368.12] I mean, I was like gnashing stuff, but I don't think I had a lunch lunch. +[1368.30 --> 1368.48] Yeah. +[1368.50 --> 1370.98] There's another theme emerging here. +[1370.98 --> 1375.54] One of the main reasons to write good code is so you can just have lunch. +[1376.56 --> 1377.48] Quite a good reason. +[1377.78 --> 1379.00] That's another thing I would do different. +[1379.14 --> 1379.90] I would write good code. +[1380.18 --> 1380.28] Yeah. +[1382.16 --> 1382.84] That's optional. +[1383.20 --> 1385.02] If you've got good tests, you don't need good code. +[1386.36 --> 1386.72] Controversial. +[1386.72 --> 1393.14] So to me, what I usually tell junior members of staff is to be like, look, we expect you +[1393.14 --> 1393.84] to break things. +[1393.98 --> 1396.68] It's just part of sort of maturing as an engineer. +[1397.04 --> 1398.76] What is helpful, right? +[1398.88 --> 1402.62] And even if you follow the, you know, the playbooks and you do the right things and everything +[1402.62 --> 1404.60] else, you know, sometimes things will go wrong. +[1404.70 --> 1408.20] Whoever writes these things, how many people have touched the documentation you're looking +[1408.20 --> 1408.40] at? +[1408.40 --> 1412.12] There's a chance that they might have overlooked something or they take something for granted +[1412.12 --> 1414.34] that you as a junior haven't encountered yet. +[1414.34 --> 1415.40] So you don't take it for granted. +[1415.54 --> 1416.56] So there's some steps in between. +[1416.64 --> 1418.94] So there's some unwritten things in between the lines, right? +[1418.98 --> 1423.82] That are sort of being conveyed that you have not yet matured enough to kind of pick up on. +[1424.08 --> 1425.94] So just document every step you take. +[1426.44 --> 1429.34] It's much easier for the team to go back and say, okay, what? +[1429.54 --> 1432.48] Because the first thing they're going to say is ask you, what did you do? +[1433.90 --> 1438.12] You know, so after everybody comes down and you can say that, well, these are the steps +[1438.12 --> 1438.38] I took. +[1438.50 --> 1441.96] I mean, and that means, you know, literally like even to this day, I do this, right? +[1441.96 --> 1445.52] If I'm working with a system that I haven't come across before and I don't know what the +[1445.52 --> 1449.12] side effects are of the things that I'm going to do, I'll literally like in a document +[1449.12 --> 1453.88] somewhere, literally be copying the commands that I'm issuing the command line, right? +[1453.88 --> 1457.50] I'm literally going to copy them into this doc and I'm basically, I'm capturing the output, +[1457.58 --> 1457.72] right? +[1457.74 --> 1458.30] As I go. +[1458.88 --> 1460.90] Now, one could say that's sort of extreme. +[1461.38 --> 1464.48] I mean, again, if there was a playbook for it, if there was some automation, I could just +[1464.48 --> 1467.02] click the button or it should have come in and let it do its thing. +[1467.02 --> 1471.54] But if I have to, you know, do this step-by-step thing, that means there's no playbook for it. +[1471.58 --> 1472.62] That means there's no automation. +[1472.76 --> 1473.58] There's no script, whatever it is. +[1473.62 --> 1477.06] So I'm just going to be literally documented what I'm doing step-by-step-by-step. +[1477.12 --> 1479.14] And if something breaks, I know exactly what broke. +[1479.62 --> 1483.52] Or if I can't tell what broke, I can ask my team and say, hey, these are the steps I was +[1483.52 --> 1483.84] following. +[1484.42 --> 1484.52] Right? +[1484.58 --> 1487.38] And then nine times out of 10, maybe I'm just lucky. +[1487.78 --> 1491.02] They'll be like, oh, yeah, this thing you should have done, this command before, you know, +[1491.04 --> 1491.74] you do this, whatever. +[1491.86 --> 1493.22] And then we find out, right? +[1493.74 --> 1497.32] There's a gap in the documentation or a gap in the process or something like that, right? +[1497.60 --> 1497.78] Yeah. +[1497.78 --> 1498.74] And you can update it. +[1499.06 --> 1499.22] Yeah. +[1499.30 --> 1503.38] So by literally just track what you're doing, that may actually end up helping you. +[1503.44 --> 1504.28] Hey, guess what? +[1504.34 --> 1507.82] That might even turn into a playbook or an opportunity for automation for whatever it is that you're +[1507.82 --> 1508.20] working on. +[1508.56 --> 1508.76] Yeah. +[1508.80 --> 1508.98] Yeah. +[1508.98 --> 1510.68] It's like step one, SSH in. +[1510.98 --> 1512.84] Step two, check the Go version. +[1512.84 --> 1515.24] Step three, drop all the database tables. +[1516.34 --> 1516.66] Whip. +[1517.16 --> 1518.12] Spot the problem. +[1519.10 --> 1519.96] It's just pen testing. +[1520.16 --> 1520.36] It's fine. +[1520.42 --> 1520.70] Exactly. +[1520.80 --> 1520.94] Yeah. +[1521.02 --> 1523.18] You shouldn't be able to do that, really, if you can do that. +[1523.18 --> 1526.52] I think that's important, though, that they've taken steps because it helps with something +[1526.52 --> 1526.80] else. +[1526.90 --> 1529.72] It helps people admit that they've possibly done something. +[1530.20 --> 1534.76] Who in that early career has got the courage to say, I've mucked up, right? +[1535.18 --> 1537.26] I potentially have lost you money or time. +[1537.60 --> 1538.64] Most people are terrified. +[1539.08 --> 1542.04] And you're terrified legitimately because you've got no experience in the industry. +[1542.12 --> 1542.68] You're brand new. +[1543.12 --> 1545.04] You're finally being paid to do something. +[1545.16 --> 1546.44] And you think you're not very good. +[1546.62 --> 1546.78] How? +[1546.78 --> 1549.50] Well, we're long in our career and we probably think we're not very good. +[1549.86 --> 1551.20] So early career, you're crushed. +[1551.52 --> 1554.66] And that ability to turn around and go, that might have been me. +[1555.24 --> 1556.48] I think I did that. +[1556.78 --> 1558.18] I pressed this button and then it broke. +[1558.90 --> 1559.72] That's tough. +[1560.26 --> 1560.40] Yeah. +[1560.62 --> 1564.28] Well, I think that speaks to the blameless culture that's important. +[1564.28 --> 1571.14] It's important to reach the point where people aren't punished for these mistakes because the +[1571.14 --> 1576.46] last thing you want is people, like you say, they bury it, they try and hide it or just +[1576.46 --> 1579.66] don't tell anybody, which could make the problem much worse. +[1579.90 --> 1582.52] So yeah, I think that culture plays a big part, doesn't it? +[1583.06 --> 1583.24] Yeah. +[1583.38 --> 1585.44] Should always blame systems and not people. +[1585.80 --> 1585.98] Yeah. +[1586.32 --> 1587.16] Something went wrong. +[1587.24 --> 1588.28] It's not the person's fault. +[1588.44 --> 1590.98] It's why did the system allow the person to do that? +[1591.12 --> 1591.38] Right. +[1591.38 --> 1597.66] So like if when Johnny says something that's mean to me, it's not Johnny that I've complained +[1597.66 --> 1597.94] about. +[1598.06 --> 1601.56] It's the system that lets Johnny get on a podcast and be horrible to me. +[1603.00 --> 1603.44] Exactly. +[1604.04 --> 1606.36] It's part of the system to be mean to you, Matt. +[1606.48 --> 1607.02] Don't you know? +[1607.38 --> 1607.70] Indeed. +[1609.16 --> 1609.80] Oh yeah. +[1609.94 --> 1611.30] It feels like it sometimes. +[1621.38 --> 1628.80] Hey friends, this episode is brought to you by my friends and potentially your friends +[1628.80 --> 1630.02] too at FireHydrant. +[1630.22 --> 1634.14] And I'm here with Robert Ross, founder and CEO of FireHydrant. +[1634.28 --> 1639.32] And Robert, there are several options out there for incident management, but what is it that +[1639.32 --> 1640.82] makes FireHydrant different? +[1640.82 --> 1645.64] The reason that we think that FireHydrant is on to something is because we're meeting +[1645.64 --> 1647.16] companies really where they are. +[1647.44 --> 1652.56] We face the same problems that every company in the industry that is building and releasing +[1652.56 --> 1654.42] software is also facing. +[1654.80 --> 1659.68] So where you want people to be able to sign up for FireHydrant and immediately be able +[1659.68 --> 1665.60] to kick off an incident using the best practices that we've built and we've experienced and have +[1665.60 --> 1668.16] gathered through the other amazing customers that use our tool. +[1668.16 --> 1673.26] It really is a very quick time to value and we want people to have a long jump from where +[1673.26 --> 1676.72] they are to where they want to be in incident management. +[1677.30 --> 1677.60] I love it. +[1677.66 --> 1678.20] Thank you, Robert. +[1678.48 --> 1683.60] Small teams up to 10 people can get started for free with all FireHydrant features included. +[1683.74 --> 1685.46] There's no credit card required to sign up. +[1685.80 --> 1687.48] They are making it too easy to get started. +[1687.68 --> 1690.38] So check them out at FireHydrant.com. +[1690.38 --> 1693.00] Again, FireHydrant.com. +[1698.16 --> 1714.06] Have we got any more horror stories? +[1714.30 --> 1717.26] Oh, by the way, this campfire's warm, isn't it? +[1717.26 --> 1720.50] So we'll probably put an effect of a campfire over the top. +[1720.78 --> 1723.02] So just pretend we're all gathered around a campfire. +[1723.12 --> 1724.46] Oh, what do you think of the campfire, Johnny? +[1725.12 --> 1725.42] Sure. +[1725.42 --> 1725.86] Yeah. +[1726.08 --> 1726.34] Yeah. +[1726.56 --> 1727.04] That's how about it. +[1727.76 --> 1729.24] I'm convinced by that performance, Johnny. +[1730.08 --> 1731.84] Have you done actual theater? +[1732.00 --> 1732.10] Have you? +[1732.20 --> 1733.88] Because what about you, Chris? +[1734.14 --> 1735.06] What do you think of the fire? +[1735.20 --> 1735.98] It's cozy, isn't it? +[1736.54 --> 1736.86] Sure. +[1737.82 --> 1738.14] Okay. +[1739.38 --> 1740.36] Crackly warm fire. +[1740.48 --> 1742.14] We don't have any marshmallows, so it's not as good. +[1742.20 --> 1742.50] Don't we? +[1742.62 --> 1743.48] It's imaginary land. +[1743.68 --> 1744.44] It's podcast land. +[1744.78 --> 1745.56] You can have anything you want. +[1745.62 --> 1746.16] Check this out. +[1746.24 --> 1746.72] What's this? +[1746.84 --> 1747.10] Look. +[1747.34 --> 1748.12] Look at your face. +[1748.74 --> 1749.04] Look. +[1749.14 --> 1749.80] It's marshmallows. +[1750.74 --> 1751.98] Natalie, what do you think of the fire? +[1753.34 --> 1754.58] Shouldn't be burning servers. +[1754.58 --> 1756.24] No, it shouldn't be burning servers. +[1756.54 --> 1759.42] No, this is a fire that doesn't actually release any carbon. +[1759.80 --> 1760.40] It's a good fire. +[1760.94 --> 1762.64] It's basically my GPU over here. +[1762.64 --> 1763.02] He's like, oh, my God. +[1765.04 --> 1766.58] It's the sound of my cooking. +[1766.86 --> 1767.70] It's a money fire. +[1767.86 --> 1768.60] My electric bill. +[1769.24 --> 1770.70] It's some old Intel Macs. +[1770.76 --> 1774.34] You know, we just turned them on, open Slack, and now they've made us a nice fire. +[1774.50 --> 1774.92] It's good. +[1774.92 --> 1778.96] Just have Slack and a regular expression running. +[1779.34 --> 1781.34] That'll generate enough heat to cook your marshmallow. +[1782.38 --> 1784.42] And those fans can definitely fly us somewhere. +[1784.64 --> 1786.80] We could all go visit Matt in the UK. +[1787.34 --> 1787.66] Yeah. +[1787.94 --> 1790.72] I mean, make sure you do go through proper passport control. +[1790.88 --> 1793.76] Don't just fly in at any point because that's illegal. +[1793.76 --> 1796.00] But if you, yeah, otherwise do, please visit. +[1796.08 --> 1796.90] We'd love to have you. +[1797.52 --> 1803.22] Yeah, I remember talking about hot CPUs, the CPU hot program that I used to have on an Amiga. +[1803.52 --> 1805.80] And basically run it and it just made your CPU hot. +[1806.44 --> 1808.14] And that was a program that you could have. +[1808.18 --> 1810.64] It was on like a front of a magazine for some reason. +[1811.22 --> 1811.84] What's that doing? +[1812.58 --> 1814.36] Someone wrote another infinite loop. +[1814.96 --> 1815.66] Yeah, there you go. +[1815.66 --> 1823.82] They've turned their horror story into a big success story because they got on a magazine cover with a floppy disk. +[1825.08 --> 1825.62] Interesting. +[1825.62 --> 1834.48] Now with the energy costs going up here in Europe, all the heaters are becoming more expensive because people assume they will not have gas to hit their house. +[1835.24 --> 1839.36] Many houses have this, apartment buildings have this systems with gas, right? +[1839.68 --> 1844.82] So you buy like electrical heaters to warm the place in case you might need that. +[1845.08 --> 1846.52] So they become really expensive. +[1846.66 --> 1849.18] So really what you're saying is that all you need is an old computer. +[1849.36 --> 1849.60] Yeah. +[1849.72 --> 1851.18] Which is probably cheaper at this point. +[1851.18 --> 1854.52] I bet we see a spike in the downloads of Slack in that area. +[1854.52 --> 1857.22] Or that CPU hot. +[1857.62 --> 1859.90] How many Electron apps can I install on one machine? +[1861.52 --> 1861.92] Okay. +[1862.04 --> 1865.44] Has anybody got any other horror stories? +[1866.84 --> 1867.32] I've got more. +[1867.52 --> 1870.02] I've got one which is something that's kind of triggering. +[1870.26 --> 1872.92] I don't know if anyone else has got sort of triggers from being horrified. +[1873.30 --> 1877.28] One of my old bosses used to come to me and if he started the sentence with, +[1877.28 --> 1878.64] what do you know about? +[1879.16 --> 1881.24] Then I knew immediately it was downhill. +[1881.70 --> 1883.92] It's like, what do you know about Perl? +[1884.12 --> 1885.84] It's like, oh, where's this going? +[1886.32 --> 1889.32] Or what do you know about directory services and exchange? +[1889.44 --> 1891.62] It's like, um, that they exist? +[1891.82 --> 1892.56] Great, you'll do. +[1893.00 --> 1894.92] And off you'll be shipped to a client site. +[1895.38 --> 1899.50] And I ended up at one of these client sites and there was a customer and it was a big, +[1899.68 --> 1900.44] big company. +[1900.80 --> 1902.28] And they were basically doing a split. +[1902.92 --> 1904.84] Merger and acquisition is the normal thing you hear about. +[1904.92 --> 1905.62] They're doing the other thing. +[1905.72 --> 1906.58] They're splitting in two. +[1907.42 --> 1912.00] And they basically said, they contracted the company I worked for to split their active +[1912.00 --> 1916.20] directory, to clone it and then rename it to the other company name. +[1916.28 --> 1917.96] So they had a perfect copy renamed. +[1918.28 --> 1920.20] And I'm just like, turn up on site. +[1920.32 --> 1922.92] I know a bit of visual basic, a little bit of C sharp at that point. +[1923.24 --> 1924.72] How do I approach this problem? +[1925.16 --> 1926.54] I did not know at all. +[1926.70 --> 1930.34] So I get in touch with Microsoft Professional Services and go, how would you do this? +[1930.34 --> 1931.44] And they're like, you don't. +[1931.48 --> 1932.06] That's impossible. +[1932.86 --> 1933.54] Don't do that. +[1933.82 --> 1937.04] That's reckless and foolish and it's not supported. +[1937.48 --> 1939.80] And it's like, okay, but I'm being paid for this. +[1939.88 --> 1944.00] And I know no better because I'm barely in my mid twenties and I'm going to take a stab +[1944.00 --> 1944.36] at this. +[1944.60 --> 1949.24] And I wrote a script going through the registry on the, uh, one of the cloned exchange server +[1949.24 --> 1949.64] machines. +[1949.80 --> 1954.26] And I basically renamed everything and I fired it up afterwards and it worked. +[1954.86 --> 1955.26] Wow. +[1955.48 --> 1956.46] That was my day's work done. +[1956.46 --> 1957.14] I left. +[1957.14 --> 1958.94] I have no idea whether that worked. +[1961.04 --> 1961.92] It appeared to work. +[1962.12 --> 1965.16] Oh, well, I just took notice exactly after this project ended. +[1965.52 --> 1966.08] Oh, I don't know. +[1966.14 --> 1966.70] I don't want to know. +[1968.60 --> 1969.64] They rate contractors. +[1969.78 --> 1970.40] That's what you get. +[1971.50 --> 1971.86] Wow. +[1972.56 --> 1973.96] That could have just worked though. +[1974.30 --> 1976.94] But you know, I never trust code that works first time. +[1977.02 --> 1978.92] That's why I like a failing test before. +[1979.26 --> 1981.54] The only thing that I was really, really scared about it. +[1981.60 --> 1984.20] The name of the company in the Active Directory was six characters long. +[1984.20 --> 1987.66] And I was reasonably sure that that was a magic value. +[1988.16 --> 1991.52] So I asked them for a new name to be the Active Directory name. +[1991.64 --> 1993.18] That was also six characters long. +[1993.42 --> 1996.94] That's the only thing that I think was intelligent about what I did that day. +[1997.72 --> 1998.98] The rest is luck. +[1999.54 --> 2000.46] And lots of regexes. +[2001.60 --> 2002.04] Again. +[2002.68 --> 2003.08] Yeah. +[2003.60 --> 2004.30] Good call. +[2005.56 --> 2006.00] Spooky. +[2006.58 --> 2007.06] Spoopy. +[2007.56 --> 2008.04] Spoopy. +[2008.28 --> 2010.70] But yeah, that's the scary question though. +[2010.74 --> 2011.76] No one's asked me that for years. +[2012.10 --> 2013.06] Hey, what do you know about? +[2015.06 --> 2015.38] Yeah. +[2015.42 --> 2016.04] What do you know about this? +[2016.08 --> 2017.16] You're like, no, you're like, nothing. +[2017.32 --> 2017.56] Nothing. +[2017.68 --> 2018.16] Nothing at all. +[2018.40 --> 2019.70] I don't want to know anything about it. +[2019.78 --> 2019.98] Yeah. +[2020.62 --> 2021.40] I know nothing. +[2021.64 --> 2022.62] Maybe that's the thing, right? +[2022.62 --> 2024.90] When you see people late in their career and you're just like, +[2024.92 --> 2025.64] what do you know about this? +[2025.70 --> 2028.72] And you think they're doing the, I've forgotten more than you'll ever learn. +[2028.84 --> 2030.80] But no, they're actually going, I don't want to do this. +[2032.48 --> 2033.20] Active Directory. +[2033.84 --> 2034.66] Never heard of it, mate. +[2034.66 --> 2035.52] Never heard of him. +[2035.86 --> 2037.42] Thinks it's someone man's name. +[2038.34 --> 2039.12] Active Directory. +[2039.24 --> 2040.46] That's a weird name for a man. +[2040.66 --> 2041.58] I think that's a man. +[2042.18 --> 2045.00] Just really selling, you know, that you don't know it just to get out of the job. +[2045.26 --> 2047.60] Well, just a tip there for people that want to get into that. +[2047.96 --> 2049.46] Like I said, I'm jet lagged. +[2050.24 --> 2052.42] And this is spooky Halloween party special. +[2052.98 --> 2054.78] How are those marshmallows looking, Chris? +[2055.38 --> 2057.64] They are toasty and brown and delicious. +[2058.34 --> 2058.90] Oh, perfect. +[2059.10 --> 2059.50] Well done. +[2060.18 --> 2061.98] Are you going to share or? +[2062.82 --> 2063.22] No. +[2063.42 --> 2063.68] No. +[2063.68 --> 2066.84] These are my marshmallows now. +[2066.84 --> 2067.02] Good point, yeah. +[2067.72 --> 2068.00] Yeah. +[2068.14 --> 2073.64] I know they're imaginary, but, you know, but I still want one. +[2074.12 --> 2075.14] We have an infinite supply. +[2075.26 --> 2076.58] Everybody can make their own marshmallows. +[2077.40 --> 2077.76] Oh, yeah. +[2077.88 --> 2078.14] Okay. +[2078.14 --> 2080.32] All you need is an infinite loop. +[2080.76 --> 2082.46] Speaking of loops, I have a scary story. +[2082.46 --> 2097.22] But if that's really a story that I think helped me gain a new appreciation for sort of how to integrate systems, how distributed systems have sort of pitfalls. +[2097.22 --> 2104.84] It's a tradeoff for everything and all the things that we value, right, as best practices for dealing with, you know, integrating systems. +[2104.84 --> 2117.28] So I was working for an organization, that very awesome organization, a nonprofit, which basically helps students, right, especially in underserved communities, sort of prepare to take sort of standardized testing and that kind of thing. +[2117.28 --> 2126.46] So for months leading up to a major sort of testing day, right, students are going to come, sit down into their classrooms. +[2126.78 --> 2132.16] They're going to be logging in and taking an assessment, right, to help them, right, with the real things. +[2132.40 --> 2136.94] And this is like a coordination across multiple schools and everything. +[2137.02 --> 2138.84] So the whole county, right, is doing this thing. +[2138.88 --> 2144.62] So we're talking like, you know, maybe like 3,000 to 4,000 students, right, that are going to all sit down and do this thing. +[2144.92 --> 2145.02] Wow. +[2145.02 --> 2152.06] And basically I'm part of the team that's basically has been working on this sort of integration, right, for months now, right? +[2152.12 --> 2154.24] We have different systems talking to each other and everything else. +[2154.82 --> 2157.98] In development and even in staging, everything works perfectly. +[2159.32 --> 2163.04] Systems can talk to each other, you know, like we're sending, you know, lots of traffic. +[2163.18 --> 2167.90] We're keeping an eye on things and we're observing to the best availability, you know, with the tools that we have. +[2168.50 --> 2174.00] And on production day, basically, which is when students actually sit down to do the thing, +[2174.00 --> 2183.12] what we didn't test against is basically having roughly 4,000 students trying to log in to the system at the same time. +[2185.66 --> 2190.86] You know, and because you have these systems that are talking to each other for authentication and pooling things, whatever it is, +[2190.94 --> 2194.22] basically we just had a thundering herd kind of situation happening. +[2194.22 --> 2201.04] And we didn't account for that because all of our tests, even our integration tests and everything else, they didn't factor in that kind of scale. +[2201.38 --> 2204.74] And I take responsibility for that because I was one of the team leads. +[2204.88 --> 2212.20] And basically one of my questions was supposed to be, what is the expected number of users and clients, right, on the system? +[2212.20 --> 2219.30] And basically we touched on these things, but there were bottlenecks in the system, right, that we should have better accounted for. +[2219.74 --> 2226.68] And thankfully we had enough of an understanding of what was happening with enough observability to be like, oh, crap, we know where the bottleneck is. +[2226.72 --> 2227.64] We need to go do this, whatever. +[2227.82 --> 2233.14] So within a matter of, you know, about an hour and a half or so, while students are waiting there, +[2233.14 --> 2236.32] because they can't really dismiss everybody and send everybody home, right? +[2236.36 --> 2242.34] We're talking like countywide, 4,000 plus students, all this coordination across months, you know. +[2242.64 --> 2248.18] To me, this remains the best and worst moment of my career because I'm like, here I am. +[2248.30 --> 2249.96] I'm supposed to be serving these kids. +[2249.96 --> 2255.60] Like as an, often we are so far removed from the consequences of our code, right, good or bad, right? +[2255.62 --> 2256.68] We're so far removed from it. +[2256.76 --> 2263.00] But here I am, I knew exactly, right, what the impact that my mistake was having on these kids. +[2263.14 --> 2263.34] Right. +[2263.36 --> 2266.56] And they already, right, have been given a short straw in life. +[2266.60 --> 2268.58] And here I am just making that worse, right? +[2268.94 --> 2271.68] So after that incident, I was like, never again. +[2271.72 --> 2272.86] Like, what do I need to do? +[2272.96 --> 2273.98] What do I need to learn? +[2274.16 --> 2275.30] Who do I need to talk to? +[2275.70 --> 2277.76] Like it was, I had to level up. +[2277.76 --> 2283.64] At any point in my career, I can't remember a single incident that has driven me to level up as much as this one. +[2283.74 --> 2286.00] Because the impact was so real. +[2286.18 --> 2287.84] It was so in my face. +[2287.96 --> 2288.88] It's just undeniable. +[2290.18 --> 2290.72] That was scary. +[2290.72 --> 2294.68] That is the $1,001 bill. +[2296.00 --> 2296.78] Yeah, that's a face. +[2296.98 --> 2299.60] Listen, I would have gladly like handed over $1,000. +[2300.10 --> 2301.26] Like out of my pocket. +[2301.50 --> 2303.30] Like to be like, look, whatever this is. +[2303.34 --> 2303.88] To the kids. +[2304.02 --> 2305.26] Make it go away right now. +[2305.28 --> 2306.02] Just give it to the kids. +[2306.12 --> 2306.44] To the kids. +[2307.64 --> 2308.00] Right. +[2308.56 --> 2310.40] Uncle Johnny's messed up again. +[2310.54 --> 2312.94] Come and collect your $20 bills, everyone. +[2313.54 --> 2316.68] No college for you, but here's some, I don't know. +[2316.78 --> 2317.60] Yeah, that's horrific. +[2317.60 --> 2324.04] Well, yeah, but when the stakes are that high, Johnny, like that is like, ooh, that is scary. +[2324.44 --> 2324.62] Yeah. +[2324.86 --> 2325.08] Yeah. +[2325.18 --> 2326.50] And you feel awful. +[2326.74 --> 2331.04] Like awful, awful, awful for being responsible for that. +[2331.26 --> 2331.48] Oof. +[2331.48 --> 2332.02] Mm-hmm. +[2332.90 --> 2337.12] What do you think about sort of engineers sort of being, having that sort of sense of consequence? +[2337.42 --> 2338.88] Because it comes in multiple directions. +[2339.04 --> 2343.34] You get salespeople going, if you don't do this, we're going to lose a million dollar deal. +[2343.52 --> 2345.62] It's not like the engineer is going to receive a million dollars, right? +[2345.64 --> 2346.96] They're just going to get their normal salary. +[2347.12 --> 2348.18] They're getting the normal pay. +[2348.72 --> 2350.12] That's a little bit abstract. +[2350.28 --> 2351.04] It's a ton of stress. +[2351.04 --> 2351.60] Mm-hmm. +[2351.60 --> 2355.60] And likewise, when you're working incidents, you know, someone will turn and go, it's affected +[2355.60 --> 2360.74] this air traffic control signal, or it's affected this sort of kids or hospital, or traffic +[2360.74 --> 2361.92] lights are down, or whatever. +[2362.66 --> 2364.04] Kind of unhealthy, isn't it? +[2364.52 --> 2370.02] It's like, we've got to keep it abstract enough that, because I get that it changes us. +[2370.38 --> 2372.32] All of those incidents changed all of us. +[2372.64 --> 2373.82] They're all horrifying moments. +[2373.82 --> 2378.52] But I'm also just like, that's the path to burnout, to sort of accepting the consequence +[2378.52 --> 2379.66] for all of these things. +[2379.92 --> 2380.02] Yeah. +[2380.16 --> 2384.90] So some value in being somewhat abstracted from the consequences. +[2385.18 --> 2385.58] Is that what you mean? +[2385.88 --> 2386.20] Yeah. +[2386.48 --> 2389.66] If you consider it too much, it just weighs so heavily. +[2389.88 --> 2390.80] It's too serious. +[2390.96 --> 2392.04] It's too much of a... +[2392.04 --> 2395.26] And the things you need to do to actually get out of those situations, they become even +[2395.26 --> 2396.46] more horrifying and scary. +[2396.70 --> 2398.02] What if I prolong this? +[2398.10 --> 2399.28] What if I make it worse? +[2399.28 --> 2402.08] And sometimes you've just got to be a bit fearless. +[2402.08 --> 2406.48] And you can't if you've got that sort of burden on you that we put on ourselves. +[2406.98 --> 2409.64] I feel like this is where systems can be helpful, though. +[2409.72 --> 2414.94] Because I think we, as an industry, are pretty bad at understanding that there can be bad +[2414.94 --> 2415.44] consequences. +[2415.44 --> 2416.94] It's like something terrible happens. +[2417.02 --> 2419.28] And we're like, oh no, this terrible thing happened. +[2419.38 --> 2422.08] But so much of the time, that terrible thing was completely predictable. +[2422.74 --> 2426.50] And we just didn't predict it because we thought it'd go fine. +[2427.08 --> 2428.18] Like I worked for a lifeguard. +[2428.36 --> 2429.70] When I was younger, I worked as a lifeguard. +[2429.70 --> 2434.60] I remember one of the things that we always did is we trained a lot, but also did a lot +[2434.60 --> 2437.50] to make sure that the environment was always a safe one. +[2438.08 --> 2439.12] So it was like, we did a lot of... +[2439.12 --> 2441.14] That's why lifeguards yell at people so much. +[2441.26 --> 2442.04] And they're like, don't run. +[2442.22 --> 2445.02] Don't do these things that might result in you getting injured. +[2445.50 --> 2448.40] I kind of feel like in software engineering, we just let people do whatever. +[2448.74 --> 2452.02] And then people slip and bash their head and they're bleeding all over the place. +[2452.08 --> 2453.44] And we're like, oh no, how did this happen? +[2453.56 --> 2458.78] And it's like, well, not only did we not tell people not to run, but we also left giant puddles +[2458.78 --> 2462.40] of water on the floor because we didn't put down the proper mats to make sure that even +[2462.40 --> 2463.92] if they are running, they can do it safely. +[2464.16 --> 2467.82] There's all these other things that you have to set up as precautions that I feel like in +[2467.82 --> 2469.96] software engineering, we just kind of don't do. +[2470.04 --> 2474.96] And part of me wonders if we don't do that because there aren't enough consequences flowing +[2474.96 --> 2476.16] down to the engineers. +[2476.16 --> 2481.78] I feel like the number of times I've been at companies that I've worked at banks and +[2481.78 --> 2484.46] people have been like, well, you know, this isn't like life or death. +[2484.62 --> 2488.12] And I'm like, this is affecting people's money and their livelihood. +[2488.56 --> 2489.72] Like, what do you mean? +[2490.50 --> 2494.04] And that's always a thing that gets rolled out is if it's like, oh, well, we're not doing +[2494.04 --> 2495.14] things that could kill people. +[2495.28 --> 2498.02] It's like, well, we're doing things that can substantially affect people's lives. +[2498.24 --> 2501.00] And I feel like we have to take that into account. +[2501.00 --> 2505.32] Because when we don't, we do lots of immoral things like run psychological experiments +[2505.32 --> 2508.22] on people without their knowledge and other terrible things. +[2508.22 --> 2510.48] Because we're like, ah, what's the harm? +[2510.64 --> 2511.84] I haven't done that for weeks. +[2512.00 --> 2513.04] I don't know why you're bringing it up. +[2515.08 --> 2515.92] It's spooky. +[2516.84 --> 2518.08] It's a spooky show. +[2525.02 --> 2526.78] I have a spooky story, I think. +[2527.02 --> 2527.98] It feels spooky. +[2527.98 --> 2530.92] So I'd recently joined this company. +[2531.94 --> 2535.28] And of course, because it's the modern day, they're using Kubernetes. +[2536.24 --> 2538.44] And of course, they're using all the shiny things of Kubernetes. +[2538.60 --> 2539.46] They're also using Istio. +[2540.10 --> 2541.56] No one actually knows how any of this works. +[2541.62 --> 2543.92] It's just like, oh, this is what we're supposed to be using. +[2544.08 --> 2547.98] So we have this big old cluster and it's running and our DevOps people are pulling their hair +[2547.98 --> 2549.04] out because I hate all of this. +[2549.52 --> 2552.08] And I start reading through the code base and looking at things. +[2552.18 --> 2555.22] And I'm like, okay, these auth policies look a little funky. +[2555.22 --> 2559.70] And then I go talk to people and they're like, we don't really have any auth policies. +[2560.12 --> 2561.90] Everything's just kind of open right now. +[2562.38 --> 2564.08] Everything in the back end, you just talk to each other. +[2564.16 --> 2564.98] There's no auth policies. +[2565.14 --> 2566.96] And I'm like, are you sure? +[2567.04 --> 2568.92] Because I see these auth policies in the code base. +[2569.00 --> 2571.32] They're like, yeah, but we don't think they're being used for anything. +[2571.46 --> 2572.58] And I was like, okay. +[2572.86 --> 2575.12] So I just kind of let it go and go about my business. +[2575.26 --> 2576.92] And then I have a few more of these conversations. +[2577.16 --> 2578.84] And I'm like, this feels weird. +[2578.96 --> 2580.86] But all these people know more than I do. +[2580.86 --> 2590.56] And then months and months later, someone stumbles across this one auth policy that has no labels and no access rules. +[2591.24 --> 2596.34] Which in Istio language means that it applies to literally everything and allows all traffic in. +[2597.06 --> 2605.86] So this one policy had just opened our entire API, including the public API, to the entire internet for anybody to do anything without needing any authorization. +[2605.86 --> 2610.76] You just needed a JSON web token that you could easily get from anywhere. +[2611.72 --> 2614.22] And I was just like, so it caused all these problems, right? +[2614.68 --> 2615.64] Everybody's freaking out. +[2616.14 --> 2617.98] And then they just ripped that policy out. +[2618.06 --> 2619.98] And they're like, okay, well, without this policy would be fine. +[2620.18 --> 2624.48] But then all of those auth policies that had been sitting there that I was like, these look funky. +[2624.70 --> 2625.84] All those took over. +[2626.00 --> 2627.22] And all of those were broken. +[2627.62 --> 2629.36] So then it broke all the APIs. +[2629.36 --> 2637.98] So then they had to put the other policy back and then go through and go find every single auth policy within Istio and then fix all of those auth policies. +[2638.66 --> 2643.38] And then they could finally remove that one policy that was opening everything to the world. +[2643.50 --> 2646.98] And I think the total amount of time that the door was just open was about nine months. +[2648.02 --> 2651.32] And to the knowledge that people have, nothing bad happened. +[2651.98 --> 2654.70] But yeah, it was quite horrifying. +[2655.42 --> 2658.76] That is still a security incident requiring disclosure, I'm afraid. +[2658.76 --> 2659.12] Yeah. +[2659.72 --> 2660.24] I know. +[2660.34 --> 2661.86] I'm just like, is this disclosure? +[2664.00 --> 2665.24] No, it was, yeah. +[2665.32 --> 2666.80] I was like, oh, oh no. +[2667.28 --> 2671.44] It taught me a lesson that like when I see funky things, I should probably bring them up a little bit soon. +[2671.52 --> 2672.90] It's like, no, that policy is there. +[2672.94 --> 2675.30] And that policy definitely doesn't work. +[2675.88 --> 2681.12] And some of the broken things were like some YAML white spacing thing where it's like something was tabbed in a little too far. +[2681.16 --> 2684.84] And then people were just copying and pasting these policies and then not testing them, +[2684.84 --> 2690.88] which was like another thing that we had to go back and be like, please test the things that you put into the code base. +[2691.20 --> 2691.76] Pretty please. +[2691.86 --> 2692.20] Thank you. +[2692.40 --> 2692.48] Yeah. +[2692.54 --> 2699.54] I think that also is a bit of a lesson is if there are bits of code and you're like, no, yeah, but that doesn't do anything now. +[2699.62 --> 2702.20] Like that used to be doing something and now it doesn't. +[2702.20 --> 2704.68] It's like either take it out. +[2705.00 --> 2707.18] If it's really not doing something, get rid of it. +[2707.34 --> 2707.60] Prove it. +[2708.00 --> 2708.32] Exactly. +[2708.48 --> 2709.78] It probably is doing something. +[2710.28 --> 2713.56] And if it isn't, maybe it should be like in your case, Chris. +[2713.80 --> 2714.08] Yeah. +[2714.72 --> 2715.36] Spoopy. +[2715.78 --> 2716.92] The zombie apocalypse. +[2717.30 --> 2717.56] Yeah. +[2718.94 --> 2720.62] Will be caused by Istio. +[2722.84 --> 2725.82] Zombie code just haunting us. +[2726.18 --> 2728.04] I'm never sure what's worse than those things. +[2728.04 --> 2733.52] Like the insecure environment where we're just like nothing applies or the extremely secure environment. +[2733.74 --> 2736.72] I've seen some where they've been really locked down and like everything. +[2736.88 --> 2741.02] You've got an IP firewall for everything and you're like, I have total confidence. +[2741.62 --> 2744.54] And then mysteriously API calls between machines just didn't work. +[2744.80 --> 2746.54] We're just like, why not? +[2746.98 --> 2750.30] And what we realized, the debugging for this went wild. +[2750.38 --> 2751.16] It went really low. +[2751.60 --> 2755.40] And we were down at Wireshark and we're watching what's going on and we're watching what's going on inside the kernel. +[2755.40 --> 2759.66] But we were turning on contract connection tracking. +[2760.00 --> 2760.96] And this is in TCP. +[2761.28 --> 2766.80] It's got little tables, state tables and network there to keep track of the sort of TCP connections. +[2766.94 --> 2768.14] But you can overflow these tables. +[2768.28 --> 2769.32] We were turning them on and off. +[2769.76 --> 2775.78] And every time we turned them on and off, we were toggling which IP firewall rules were actually matching or not. +[2776.12 --> 2781.10] So we were taking existing connections and then just randomly dropping them every time we flipped these things. +[2781.42 --> 2782.66] But we could never observe it. +[2782.66 --> 2785.74] And we were just there the whole time just going, we lost it again. +[2786.04 --> 2786.92] There goes the connection. +[2787.44 --> 2791.24] And it took us weeks of just poking around going, what's going on? +[2791.74 --> 2792.44] Can't see it. +[2792.62 --> 2793.40] Ghost in the machine. +[2794.84 --> 2795.24] Wow. +[2796.00 --> 2797.20] Yeah, too secure is a problem. +[2798.26 --> 2798.66] Honestly. +[2799.12 --> 2801.42] Well, in that spirit, D, what's your pin number? +[2802.80 --> 2804.28] Just give us three of the numbers. +[2805.32 --> 2806.38] Let's play Mastermind. +[2806.96 --> 2808.76] Is this like, what's it, Spaceballs? +[2808.98 --> 2809.88] One, two, three, four? +[2810.34 --> 2810.58] Yeah. +[2811.44 --> 2813.06] No one's going to suspect that, I think. +[2813.54 --> 2814.78] No one's going to try that, are they? +[2815.20 --> 2816.54] Zero, zero, zero, zero. +[2816.92 --> 2817.80] Does it allow you? +[2817.88 --> 2819.58] I don't think systems allow you to do that. +[2819.80 --> 2820.08] Why? +[2820.72 --> 2821.00] I don't know. +[2821.06 --> 2822.28] I'm going to try and change my pin now. +[2822.34 --> 2822.46] Yeah. +[2823.04 --> 2823.38] You might. +[2823.38 --> 2825.52] Oh, yeah. +[2825.62 --> 2826.10] I don't know. +[2826.46 --> 2826.70] Maybe. +[2826.90 --> 2827.60] Because it's too easy. +[2828.24 --> 2828.74] I don't know. +[2829.92 --> 2834.96] Any other horror stories before we throw a, what do you put on fire? +[2835.06 --> 2835.42] Water? +[2835.78 --> 2836.60] Don't do that, do you? +[2836.66 --> 2838.10] Just let it die out on its own? +[2838.26 --> 2839.24] No, we've got to be responsible. +[2839.76 --> 2841.46] How are we going to sort this fire out? +[2841.92 --> 2843.78] By throwing water and the electrical equipment? +[2845.72 --> 2846.48] Use foam. +[2846.72 --> 2847.28] We'll use foam. +[2847.36 --> 2847.70] Turn it off. +[2847.72 --> 2848.60] We'll close slack. +[2848.60 --> 2849.32] What's on fire? +[2849.50 --> 2849.64] Yeah. +[2849.80 --> 2851.96] Close slack, and then the fire will die down. +[2851.96 --> 2856.06] So before we put the fire out, has anyone got any other final horror stories? +[2856.70 --> 2859.18] Well, you know, if you're mine, anyone you want to hear? +[2859.60 --> 2861.14] Well, what do you mean in real life? +[2863.06 --> 2867.18] I wrote one down, which actually I wrote in advance about doing a sequel statement and +[2867.18 --> 2874.08] accidentally double putting in the semicolon after the from table, so an update. +[2874.58 --> 2878.50] So the where statement didn't apply, and that was to a production system. +[2878.92 --> 2881.86] Oh, so what's the effect of that then? +[2881.86 --> 2886.38] So normally you would be updating something and specifying the where, which will limit +[2886.38 --> 2887.92] what gets changed, right? +[2888.32 --> 2889.06] Yeah, exactly. +[2889.38 --> 2892.48] I tweeted this when you actually asked about it, and I think no one really appreciated +[2892.48 --> 2896.00] what it does and how it happened. +[2896.16 --> 2897.46] But I executed a query. +[2897.60 --> 2902.22] I was just tidying up some debt that was left over, and it should have been really trivial. +[2902.22 --> 2905.68] And I practiced it, and then I copy and pasted it into the console. +[2905.84 --> 2910.78] But after I copied and pasted the first one, for whatever reason, I fingered a semicolon +[2910.78 --> 2912.78] and then put in the where line. +[2912.92 --> 2914.20] But that makes it two commands. +[2914.50 --> 2918.58] So it successfully did the update column set value equals on table. +[2919.06 --> 2920.46] And it didn't apply the where. +[2920.62 --> 2923.62] So it updated, I think it was like 90 million rows or something. +[2923.90 --> 2924.68] Oh, God. +[2924.68 --> 2927.08] And the machine was very fast. +[2927.84 --> 2929.70] Faster than I was at finding Control Z. +[2929.96 --> 2930.90] Oh, no. +[2931.48 --> 2934.94] And that's why you work inside of transaction blocks, kids. +[2936.22 --> 2940.36] That's why that's advisable, but was not what I was doing that day. +[2940.36 --> 2946.20] The real mess there is actually sort of going, how can we restore this when our database backup +[2946.20 --> 2950.84] was like 12 hours ago, and there's 12 hours of changes in other tables since then? +[2951.16 --> 2951.26] Ouch. +[2951.68 --> 2953.34] So you're not going to sort of do anything there. +[2953.44 --> 2959.06] So it's a pull the old sort of thing and extract that table and then go and update all the necessary +[2959.06 --> 2960.04] rows for the right things. +[2960.88 --> 2961.86] Takes time. +[2963.58 --> 2964.06] Yeah. +[2964.30 --> 2967.48] I like that you couldn't keep up because the machine was so fast. +[2967.48 --> 2970.88] Is that why you now insist only on running on Raspberry Pis? +[2971.48 --> 2972.24] Intel Max. +[2972.70 --> 2973.64] Run on Intel Max. +[2973.74 --> 2974.80] It's the solution for everything. +[2977.34 --> 2977.78] Okay. +[2977.88 --> 2978.96] Well, that sound. +[2979.24 --> 2981.26] We all heard that sound, didn't we? +[2982.26 --> 2984.22] How would you describe that sound that we just heard? +[2984.50 --> 2985.10] That sound. +[2985.28 --> 2985.66] Spooky. +[2985.82 --> 2986.08] Natalie. +[2986.76 --> 2987.20] Spooky. +[2987.36 --> 2987.66] Yes. +[2988.04 --> 2990.06] Natalie, how would you describe that sound that we just heard? +[2990.84 --> 2991.80] Did I miss the sound? +[2992.46 --> 2993.46] Yeah, we heard that sound. +[2995.68 --> 2997.20] Chris, how would you describe it? +[2997.20 --> 2998.04] Was it the marshmallows? +[2999.24 --> 3002.56] I think it's the sound of us closing slack so our fire is going away. +[3002.70 --> 3003.28] Was it marshmallows? +[3003.70 --> 3005.36] Yeah, it's kind of a spooky sound, wasn't it, Johnny? +[3005.48 --> 3006.82] How would you describe that, Johnny? +[3007.06 --> 3008.02] If you had to, what'd you do? +[3008.60 --> 3010.28] As silent as your hairline? +[3011.60 --> 3012.50] Oh, wow. +[3014.10 --> 3019.80] I mean, it's getting very poetic and slightly unusual. +[3020.20 --> 3021.08] Banter that. +[3021.08 --> 3023.94] You asked for it. +[3024.06 --> 3025.38] I mean, you did ask for it. +[3025.44 --> 3025.64] Amazing. +[3026.20 --> 3026.36] Yeah. +[3026.64 --> 3028.64] Is it silent because it's so far away? +[3028.82 --> 3029.64] Like, distant. +[3029.90 --> 3030.80] It's in the distance. +[3031.62 --> 3032.28] Does that mean? +[3033.08 --> 3035.08] Like, it's screaming, but you can't hear it. +[3035.48 --> 3036.46] It's too far back. +[3037.20 --> 3037.44] Yeah. +[3037.44 --> 3040.20] It's just screaming, why, Dad? +[3040.98 --> 3041.34] Why? +[3042.42 --> 3043.14] Do you know what I mean? +[3043.26 --> 3045.62] If you know you're going to look like this, don't have kids. +[3045.98 --> 3048.20] If you know you're going to make kids that look like me, don't have them. +[3048.38 --> 3049.28] That's my advice. +[3049.80 --> 3050.78] But my dad doesn't listen. +[3050.78 --> 3053.96] We're all talking about recessions nowadays, but you've been in one for quite some time, +[3054.02 --> 3054.34] right, Matt? +[3056.06 --> 3057.18] Oh, there we go. +[3058.06 --> 3058.86] That was a good one. +[3062.64 --> 3063.42] That was a good one. +[3063.42 --> 3063.62] Yeah. +[3064.00 --> 3064.26] Yeah. +[3065.02 --> 3065.64] Good point. +[3066.86 --> 3067.22] Okay. +[3067.42 --> 3067.64] Well. +[3070.54 --> 3075.52] That sound of Johnny talking tells us it's time for Unpopular Opinions. +[3075.52 --> 3079.52] Unpopular Opinions. +[3086.52 --> 3089.44] Unpopular Opinions. +[3091.68 --> 3092.68] Okay. +[3093.42 --> 3100.00] Who's going to kick us off with the first unpopular opinion? +[3100.92 --> 3101.20] D. +[3103.48 --> 3104.28] You've been chosen. +[3104.48 --> 3104.86] You scared him. +[3105.62 --> 3106.32] He was like, what? +[3106.54 --> 3106.82] What? +[3106.88 --> 3107.24] What happened? +[3107.32 --> 3107.60] What happened? +[3108.40 --> 3112.38] Let's just say, I promise no one's pushing their hand around the Ouija board to make it +[3112.38 --> 3115.20] spell out that I'm about my hairline. +[3115.36 --> 3118.50] I don't want to hear anything from the ghosts about my hairline, Johnny. +[3118.92 --> 3119.18] Right? +[3119.28 --> 3121.44] So don't make it do that. +[3121.90 --> 3122.80] Everyone put your hand on it. +[3122.80 --> 3123.64] Let it be natural. +[3123.90 --> 3124.36] And we'll see. +[3124.54 --> 3124.76] Yeah. +[3124.80 --> 3126.10] It's floated over to D. +[3126.72 --> 3127.32] Okay, D. +[3127.70 --> 3128.26] It's a fix. +[3128.26 --> 3128.76] It's a fix. +[3129.56 --> 3130.44] Unpopular Opinions. +[3130.58 --> 3131.16] It was a fix. +[3131.30 --> 3131.44] Yeah. +[3132.26 --> 3133.52] Mine comes from the last go. +[3133.74 --> 3134.70] The last go time. +[3134.94 --> 3136.44] One of the guests said that Go's brilliant. +[3136.56 --> 3137.56] You don't have to worry about security. +[3137.72 --> 3138.64] It does everything for you. +[3138.76 --> 3140.10] You don't have to worry about the memory management. +[3140.34 --> 3140.88] It's everything. +[3140.88 --> 3141.78] It's super cool. +[3141.88 --> 3148.16] And I think it actually has made Go more insecure because people are so, they've put so much +[3148.16 --> 3153.20] trust and safety in the actual sort of language that a lot of the basics are dropping by +[3153.20 --> 3153.62] the wayside. +[3154.12 --> 3156.82] For me, the number one thing that people should do is sanitize inputs. +[3156.82 --> 3160.10] And it's not because I wrote Blue Band-Aid, but they should do it on everything. +[3160.42 --> 3162.56] And I just don't think see anyone doing it anywhere. +[3163.12 --> 3165.16] And there's just great big holes in everything. +[3165.34 --> 3167.86] But people are still there going, but we've got memory safety. +[3168.72 --> 3171.28] Memory safety saves you from yourself, not from others. +[3171.28 --> 3175.90] What do we think? +[3175.98 --> 3177.54] Is that popular or unpopular? +[3177.82 --> 3178.74] It should be popular. +[3180.04 --> 3181.18] Sanitize all your inputs? +[3181.48 --> 3182.28] It should be. +[3182.66 --> 3186.98] But if you look at all of the open source stuff that's out there, very few people actually +[3186.98 --> 3189.42] sanitize, check, validate their inputs. +[3189.88 --> 3193.98] They're just like, I mapped this input directly to a struct and I'm going to use it. +[3194.44 --> 3196.72] You know, they take their form fill and they're on their way. +[3197.06 --> 3197.26] Yeah. +[3197.26 --> 3201.14] One simple version of that is just a limit reader when you're reading a body. +[3201.48 --> 3202.98] Or like of a request. +[3203.26 --> 3206.28] Like you can error if it's too big and things like that. +[3206.40 --> 3207.18] There's bits like that. +[3207.28 --> 3210.94] But you end up doing quite a lot of that heavy lifting yourself every time. +[3211.32 --> 3214.78] There are libraries out there where you can add tacks and say, this should be a text field. +[3214.84 --> 3216.10] It should be no longer than this length. +[3216.22 --> 3216.92] There should be a number. +[3217.06 --> 3217.94] It should be no longer than this. +[3218.36 --> 3220.50] But far too few projects do it. +[3220.74 --> 3221.90] Do they use reflection? +[3222.06 --> 3223.90] I think I've avoided them if they... +[3223.90 --> 3226.36] But although that's not a great reason to avoid it. +[3226.36 --> 3228.36] I just tend to not... +[3228.36 --> 3230.02] Why are you afraid of your reflection? +[3230.48 --> 3231.08] Are you a vampire? +[3231.28 --> 3232.28] Well, it's... +[3232.28 --> 3233.88] Because I am a Dracula. +[3234.60 --> 3235.38] Yeah, the casino. +[3236.78 --> 3238.90] Because he doesn't want to see that hairline. +[3239.14 --> 3239.38] Obviously. +[3239.84 --> 3239.94] Oh! +[3240.58 --> 3242.28] That's why I don't have mirrors. +[3243.48 --> 3244.30] Chris is in on it. +[3244.40 --> 3244.68] All right. +[3244.74 --> 3245.54] Everybody take a turn. +[3245.70 --> 3246.56] Yeah, Chris is in on it. +[3246.84 --> 3247.84] Take a stab at Matt. +[3248.98 --> 3249.92] I'm like a pinata. +[3250.84 --> 3252.48] Like a really rubbish pinata. +[3252.62 --> 3254.96] Imagine you buying a pinata for kids and it's me. +[3255.48 --> 3256.12] You know what I mean? +[3256.12 --> 3257.20] You'd take it back, wouldn't you? +[3257.82 --> 3260.66] You'd be like, no, we'd probably go for the unicorn instead in second thought. +[3260.66 --> 3262.20] Should have guessed that, really. +[3262.54 --> 3263.16] Okay, yeah. +[3263.88 --> 3264.28] Fine. +[3264.60 --> 3265.06] Thanks, Chris. +[3265.16 --> 3265.88] It's Halloween special. +[3266.02 --> 3267.76] You're allowed to do that. +[3268.18 --> 3268.82] Spooky opinion. +[3269.86 --> 3270.46] Yeah, exactly. +[3271.08 --> 3271.40] Okay. +[3271.46 --> 3272.80] Any other unpopular opinions? +[3273.18 --> 3273.80] I have one. +[3274.48 --> 3276.42] Natalie Pistano, which? +[3277.26 --> 3277.50] How? +[3277.50 --> 3277.82] Ooh! +[3278.96 --> 3279.56] I don't assume. +[3280.34 --> 3281.76] They don't know I've just done that. +[3282.00 --> 3283.88] So it just sounds like a sound effect in the background. +[3284.10 --> 3284.36] Come on. +[3284.66 --> 3289.58] Some of the training that you should be taking occasionally throughout your career, even annually, +[3289.58 --> 3294.70] should not be about things that are in the future, like new things, like new technologies +[3294.70 --> 3295.52] coming and so on. +[3295.52 --> 3297.58] But also a little bit about the back. +[3298.64 --> 3300.84] A little bit of assembly every now and then. +[3301.56 --> 3305.96] Might be useful accidentally at some point in your life because you need it. +[3306.08 --> 3310.10] And even if not, it might, like you'll see patterns because it's all the same things, +[3310.14 --> 3312.74] it's just more and more abstractions, but it's still the same things. +[3312.74 --> 3320.36] So seeing how it's done, how things were solved, how problems were, might help you figure future +[3320.36 --> 3322.02] when you rely on the past. +[3322.74 --> 3323.48] And it's unpopular. +[3323.76 --> 3325.18] I know you all haven't been agree with me. +[3325.28 --> 3325.82] It's unpopular. +[3326.02 --> 3326.44] None of us. +[3326.60 --> 3330.72] I also did not do that and don't allocate time or budget for that. +[3332.02 --> 3333.26] I do know what you mean. +[3333.42 --> 3336.30] I actually have this book called But How Do It Know? +[3336.60 --> 3339.24] It's just on, I got it off the Amazon website. +[3339.24 --> 3344.92] And basically it talks about computing from the very bare beginnings, like literally logic +[3344.92 --> 3350.00] gates and then how you make a bit out of two NAND gates and like just showing how the logic +[3350.00 --> 3352.80] works and then building up everything in a computer like that. +[3352.94 --> 3354.02] And it is amazing. +[3354.24 --> 3357.58] But yeah, it was like not something I need. +[3357.86 --> 3362.06] And actually something else that occurred to me when you were saying that one is having like +[3362.06 --> 3365.42] training or paying attention to things that you already think you're good at. +[3365.56 --> 3369.14] So not just new things that are new to you, things that you already think +[3369.14 --> 3370.06] yeah, I've got that nailed. +[3370.58 --> 3371.70] You might be surprised. +[3372.00 --> 3373.68] Like plenty of other things to learn. +[3373.84 --> 3374.08] Yeah. +[3374.56 --> 3375.34] I like that one. +[3375.44 --> 3376.34] We'll test that one. +[3376.66 --> 3377.90] See if that's unpopular or not. +[3378.08 --> 3381.46] Even the way you did that is interesting because you were already a couple of years +[3381.46 --> 3384.10] a software developer and then you looked into gates and so on. +[3384.12 --> 3386.54] And that kind of helped you put this in place. +[3386.68 --> 3391.24] When I started my degree, the first course I did was those logic gates and everything was +[3391.24 --> 3391.46] scary. +[3391.58 --> 3392.34] Calculus was scary. +[3392.46 --> 3394.00] And then those gates, like what? +[3394.42 --> 3395.32] I don't know that. +[3395.48 --> 3396.70] Just let me pass that test. +[3396.76 --> 3397.24] Leave me alone. +[3397.24 --> 3401.98] And if it would be the other way around, start with the programming courses and then +[3401.98 --> 3406.60] you go about semiconductors and then you go about circuits and then you speak about gates. +[3406.76 --> 3410.62] It might have been more interesting and actually would make more sense. +[3410.70 --> 3411.60] Maybe for me, at least. +[3412.40 --> 3415.04] I think that's the path that people take accidentally. +[3415.52 --> 3419.44] Like those who do boot camps and then they actually eventually over like five, 10 years +[3419.44 --> 3423.34] become like systems engineers and are working on like kernel or TLS or something. +[3423.34 --> 3426.76] They don't know they're doing that, but that's kind of what they're doing, right? +[3426.80 --> 3428.08] They're looking back at the fundamentals. +[3428.54 --> 3428.62] Yeah. +[3428.72 --> 3429.18] I agree. +[3429.24 --> 3429.76] It's unpopular. +[3429.98 --> 3432.70] I don't know anyone who does that, but I think it's genius and we should. +[3432.90 --> 3435.02] We shouldn't open a university that teaches that way. +[3435.44 --> 3436.94] I feel like that's how my career has been. +[3437.10 --> 3439.74] I just have gone like down the stack further and further. +[3440.14 --> 3440.72] That's interesting. +[3440.88 --> 3441.04] Yeah. +[3441.22 --> 3441.82] It's been fun. +[3442.26 --> 3445.56] I've written like a few like operating system kernels, like toy kernels. +[3445.56 --> 3447.14] It was infuriating. +[3447.34 --> 3451.64] Modern processors are terrible, but mostly because they're so old. +[3452.18 --> 3454.60] Like, oh, maybe I have this code from the 70s. +[3454.64 --> 3458.82] I might need to run on my Intel 12th gen chip or something. +[3458.94 --> 3459.62] You never know. +[3459.92 --> 3463.68] It's like, no, Intel, I don't need to run code from 1980 on my new processor. +[3463.84 --> 3464.20] Thank you. +[3464.20 --> 3471.58] I like that too, though, that people can start higher up in the stack and still be doing things +[3471.58 --> 3473.18] without understanding everything. +[3473.62 --> 3475.08] Because I've seen it. +[3475.14 --> 3480.24] People help themselves back because they think, well, I just don't know all of this stuff. +[3480.34 --> 3484.86] And they don't know that they don't need to know it necessarily, which is why I say you +[3484.86 --> 3486.58] sometimes don't need to know a lot of the stuff. +[3486.68 --> 3488.42] Just sort of get on with it and try things. +[3488.82 --> 3490.24] But it doesn't work for everybody. +[3490.24 --> 3493.80] I think there's so many different styles and things that people appreciate. +[3493.80 --> 3494.84] Things that work. +[3495.02 --> 3498.16] And so wouldn't like to force it on everybody. +[3498.30 --> 3500.26] As we like to say, it depends. +[3501.54 --> 3506.66] And I'm afraid put down your goblets of red wine. +[3506.78 --> 3507.70] Yes, it was what? +[3507.78 --> 3509.02] It was definitely wine. +[3509.36 --> 3509.74] Yeah. +[3510.48 --> 3514.48] And put also away those sandwiches of miscellaneous. +[3515.84 --> 3517.20] No, there weren't any sandwiches. +[3517.58 --> 3520.28] No, let's just not do the sandwiches, but we'll do the wine bit. +[3520.28 --> 3523.38] And then keep the wine because I think it's blood. +[3523.80 --> 3524.56] Okay. +[3524.72 --> 3529.32] And that's all the time we have on today's Ghost Time. +[3529.74 --> 3531.30] Thanks for joining us, everybody. +[3531.98 --> 3534.20] See you next time and stay spoopy. +[3534.20 --> 3544.24] Do you have a spooky story that'll scare your fellow devs? +[3544.42 --> 3546.22] Let us know in the comments. +[3546.78 --> 3549.56] The link to discuss this episode is in your show notes. +[3550.14 --> 3552.58] And if you enjoyed this Halloween-themed edition of Go Time, +[3552.82 --> 3554.86] please do share the pod with your friends. +[3555.10 --> 3556.96] We appreciate you helping spread the word. +[3557.32 --> 3559.44] Thanks once again to our partners at Fastly. +[3559.44 --> 3562.88] They ship all of our pods super fast to wherever you listen. +[3563.18 --> 3565.00] Check them out at Fastly.com. +[3565.36 --> 3566.54] And to Fly.io. +[3567.12 --> 3569.78] Post your app servers and database closer to your users. +[3570.24 --> 3571.10] No ops required. +[3571.56 --> 3573.44] Learn more at Fly.io. +[3573.90 --> 3577.66] Next time on Go Time, Ian and Natalie are joined by Tim Smith +[3577.66 --> 3580.38] to discuss Go in medicine and biology. +[3581.08 --> 3583.44] Subscribe now if you haven't yet so you don't miss it. +[3583.44 --> 3585.84] We'll have that episode ready for you next week. diff --git a/Stay agile out there_transcript.txt b/Stay agile out there_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..865ad9c8b166b3979948dc5506b51670133b0e3c --- /dev/null +++ b/Stay agile out there_transcript.txt @@ -0,0 +1,255 @@ +**Natalie Pistunovich:** Good time of the day, everyone, wherever you join us from. Today Angelica and myself are going to be talking about a topic that is very close to Angelica's heart, and I am very excited to finally learn and kind of clarify everything that has to do with Agile. We are being joined by Inbal Cohen, who is also based in Berlin, just like myself. Hi, Inbal. + +**Inbal Cohen:** Hello. + +**Natalie Pistunovich:** How are you doing? + +**Inbal Cohen:** Very well. Thank you very much for having me today. I'm excited. + +**Natalie Pistunovich:** It's our pleasure. We're excited to have you. Inbal, could you tell us a little bit about yourself? + +**Inbal Cohen:** So hello, I'm Inbal. I've been working in the field of product management for the past about ten years. I've had all sorts of different types of positions, mostly in startup companies in mobility, and fintech, and gaming. I've been CPO, CTO, director of product, and at the moment, I work at a company called Amazon, which you might have heard of, as a Technical Senior Product Manager. Maybe important to mention - everything that I talk about in this show are my personal opinions, and they don't represent the company. What else...? I've been living in Berlin for nearly six years, and really loving it here. + +**Natalie Pistunovich:** That's great. Thank you for a very thorough intro. It's cool to see how many parts of the industry you have touched, both different industries, but also different roles. Definitely cool. And let's start with a point. What is agile? + +**Inbal Cohen:** What is agile? + +**Natalie Pistunovich:** What is the agile methodology? + +**Angelica Hill:** And is it a methodology? + +**Natalie Pistunovich:** What is the methodology? + +**Inbal Cohen:** And is it a methodology? \[laughter\] Maybe just to start - it's not a cult, it's not a religion, and although it seems that anyone that is working in Agile, or has anything to do with it would be talking about it as they've discovered the truth of the world - which, honestly, I do the same thing; I swear, it's totally legit. That's a fantastic question... And maybe a little bit more to the point - when we talk about the different types of ways that people think about agile, most of the times they're very stuck on the thought of story points, or it has to be like lots of meetings, absolute chaos. Or I've heard things like you don't have any certainty about things. And I think that maybe the best thing to do when we're starting to talk about agile is actually strip everything down and go way back to the basics. And when I talk about the basics, that means going through the very basic principles of the manifesto. And when you read through those, you realize that Agile is actually a methodology. There are people that say that it's not a methodology; it might be just a way of thinking, in order to create progress in an iterative and incremental way. Just progress. Is it for a team of software developers? Is it for a team of people that are working on HR? Is it for the team that's working on marketing? That doesn't really matter at this point. + +**Natalie Pistunovich:** Okay, so agile methodology is a list of steps to make incremental progress. + +**Inbal Cohen:** \[05:40\] It's more like a mindset that helps you understand what are the steps that you will need to take in order to get your job done well. So here's an example... One of the things that came a bit before agile was the whole concept of lean. So there's this company - you've probably already heard this story 10,000 times, so shorten it down; there's a company you might have heard of, Toyota... They were looking at the manufacturing lines that they had, and they decided to cut the waste, cutting down all the different types of wasteful processes that were happening, that were wasting a lot of time and energy and money for the company, and they decided to call it that methodology lean. It's a way of thinking. How do I strip down everything that isn't creating value for me at this moment? One step forward would be agile. So we're taking into consideration everything that lean has taught us on reducing waste, and now we're saying, "Okay, we don't have waste, but what do I do with the people that I do have? There's a goal in mind, there is something that I want to achieve. How do I get there?" It's a very big and large project. + +Let's say it's a project I have a very large project to accomplish; there is a piece of software that I want to develop. What do I need to do? I have 14 people. How do I arrange these people in a kind of way that they would make the most out of their time and their efforts? + +And another thing to take into consideration - agile isn't just about what to do. It's also how to nurture the people that are part of that team; how to create motivation, how to make sure that people are working in their highest capacity, and also working together in a very, very close way. That's one of the things that we also learned about agile; it's all about the people. + +**Angelica Hill:** And then how does agile relate to the practice of like Scrum, and the kind of day-to-day practices that I'm sure many of us are familiar with - retro, standup, sprint planning, grooming, etc. How do those two play together? + +**Inbal Cohen:** I love that. So Scrum - and you've probably also heard about Kanban; there's Scrumban, there's Safe, there is Less, there's all sorts of different types of frameworks that we could take to our advantage. And I think that frameworks are absolutely cool. They're awesome. There is a person that sat down and said, "Okay, I understood the principles of agile, but there are people that are coming and asking me, "What should I actually get done?" And there were people that sat down and said, "You know what - this is a way to get it done." For example with Scrum, they're saying, "You need to have a defined group of people. You need to have a list of tasks. You need to prioritize that list of tasks. You need to go through different types of ceremonies in order to arrive to a point where you're creating increments of valuable software. How do we get that done with ceremonies that repeat on regular intervals?" + +There's all sorts of things that happen in Scrum that you would also see in other frameworks, like Kanban, for example. Still, we're breaking down larger tasks into smaller tasks; we're organizing, we're prioritizing them. We're taking them through swim lanes of development. But each and every methodology, or each and every framework, sorry, would fit differently, depending on the team that you're working with. And what I've found interesting is the better you know the different types of frameworks that you have available, the smarter you are to be able to pick and choose which pieces of those frameworks are actually suitable for your team. Don't follow frameworks blindly, because that's not what they're there for. They're there to encourage you to be curious about the different types of ways that you can implement agile in your work. + +**Angelica Hill:** Okay. And then, I think - well, this is a personal preference... I feel like you can't talk about agile without talking about waterfall. I feel like whenever I'm talking about process, or whenever I'm talking to my team about agile practices, it's always like, "Oh, but this is waterfall. What is the difference?" Like, there's a lot of talk about companies who do agile, but actually do waterfall. So it'd be great if you could give a little like TL;DR, like what is waterfall, and what is the difference between waterfall and agile? + +**Inbal Cohen:** \[10:03\] I love that. And I think that waterfall is extremely important for understanding agile. One of the examples that I like giving about waterfall is - so I worked in automotive; I learned about it a little bit, and that's why my example comes from automotive. So let's think about a factory that's, say, producing cars. If someone would come into a factory and say, "Hey, let's produce a really cool new car", and someone would say, "Okay, let me think. Let me first give you" - and if you know agile, you know that the perfect example, if someone asks you for a car, give them like a skateboard. And if you don't know this example, you'll have to read up on it, or we'll talk about it later. But in a car factory, that doesn't actually work that way. + +When we're trying to produce a car, what we have to go through is a process of maybe two or three years of planning, trying out all sorts of different options. And then you take about two or three years sourcing all the materials that you need in order to produce - the steel, and the rubber, and the tires, and whatever it is. And then it takes you a few years to produce a few different types of prototypes, what you would choose one from, and only then you would get to the point, maybe ten years later, where you can start mass-producing this vehicle that you chose to produce. That's waterfall. Waterfall says, "Plan everything, work on everything, test everything, and then ship everything." It's a big bang. It's all or nothing. Okay? + +**Natalie Pistunovich:** So it's a waterfall because it's like stage by stage... + +**Inbal Cohen:** Exactly. It's step by step. Every time you finish a step, you're done. The process isn't planned to go back to the drawing board. You're already ahead. The train has already moved out of the station. And in agile, we're saying it's not step by step. It's like a type of loop-de-loop. And why is it always drawn as a loop-de-loop is because we're saying, "If I as a software company will try to produce the most perfect software", and you as a product team come up to your CEO and you say, "We have it. This is it. We know what we're building. 10 years, and you have the most amazing software", your CEOs will say like, "Dude, I don't have the resources for 10 years from now. And the software that you're going to produce is not going to be relevant in 10 years." Right? Like, what piece of software do we even know that is still relevant... This was part of our pre-conversation. But there are very few things that last so long, right? And in the industry that we have going on right now, especially if we're looking at apps, if we're looking at web services, you're expecting to get updates on a regular basis. We need to be able to iterate very, very quickly, with small increments, and produce, produce, produce, produce, and not just wait until the end. + +**Natalie Pistunovich:** So agile is this feedback loop. Waterfall, we say that's step by step, not going back. How about Scrum? + +**Inbal Cohen:** So that's agile. + +**Natalie Pistunovich:** So Scrum is part of agile. + +**Natalie Pistunovich:** Scrum is a framework of agile. And you can have different ones. You have Scrum, you have Kanban, you have Scrumban... You have all sorts of different frameworks that are doing those at scale. There's Nexus, there's all sorts of different types of frameworks to do iterative work. And it really depends on the type of things that you're getting done. So just imagine, there are companies that are doing solely backend products; there's nothing to see. Even if there is a UI, that's not the part that we're focusing on. You wouldn't expect their development cycles to match an app gaming company. You're not going to expect the same level of interest, let's say, coming up from the different updates, and you would also not expect them to be so regular. + +\[14:08\] The amount of complexity might be different, the different types of languages that are being used are different, there might be the differences of the sizes of the teams... So according to all of these different parameters, a person that is in product, probably leadership in product, would need to decide on the right framework or the combination of frameworks that might be working for that team, and then on a deeper level, the product manager themselves of that team need to be able to have the feedback loops with their team, "Does this fit us? Does it not?" + +**Angelica Hill:** So is it like pick your own adventure? I feel going in you have all these frameworks, all these many different ways of doing things, and it really is kind of pick and choose as and when relevant. It may well be the same team has one project they do very strict agile, but then the next is waterfall... It really is kind of as and when applicable, whatever is the best way to get your end result in the most efficient, effective manner. Am I right? You don't really have to choose, it's kind of the benefit here is there's all these wonderful frameworks, learn about all of them, and then see which one fits your team and your needs best. + +**Inbal Cohen:** Absolutely. And doing it differently -- in my personal opinion, doing it differently would actually be missing out. If you're just trying to copy whatever a guidebook is giving you... You know, you just bought a book, or you went through one training, and you're like, "This is it. This is the only way of doing things", you're really limiting yourself, because agile is a way of thinking, and it's not so limited. + +So being able to open yourself up and know more about all the different frameworks, and being able to do that, as you're saying, go on your own adventure - that's part of the fun, isn't it? Like, getting to know your team, getting to know what works for you, trying different things out... Super-important. + +**Natalie Pistunovich:** Okay, so as a developer, I'll ask you, what tips do you have for me to make the most of agile? + +**Inbal Cohen:** That's fantastic. First of all, I would say educate yourself. One of the things that a lot of times we do is that we depend on other people within our team to be the experts in whatever their subject matter expertise is... Which is fantastic. That's their job; they should be the expert. But having some type of basic knowledge about the things that you should expect different people of different professions is extremely important. The more educated you are on the types of expectations that you can have from your product manager, from your product owner, from your Scrum master, whatever type of a title that you have going on in your team - that's the first, most important step. + +The second thing is to really have in mind, agile isn't about output. It's not only about what you're delivering. It's about the process that happens in between. It's about creating high-performance teams. And let's say the aim, the goal of the person that is managing the team in an agile way - which is most probably, again, a product leader - should be to have a high-performance team. And they need to be able - and you could help them that way, to tell them and to express how to make you more effective at your job. Are you feeling comfortable? Do you feel that you're stressed? Do you feel that the work environment that you have - does it suit you or not? Do you understand the purpose of what you're doing? Understanding purpose is a major role in agile. You have to have a team that knows "Why are we doing what we're doing?" + +\[17:56\] And I'm going to quote someone that, again, I hope I'm not too much of a cliché, but Martin Kagan wrote a book, it's called Inspired; it's pretty much -- can I say that it's the Bible of PMs? Like, not to have any religious affinities, but it's a book that PMs very much look to, to get obviously inspired. And one of the things that he says there is that if you're using your developers only to develop code, you're only using half of their skills. And that's something that's really important for you as a developer to know. If you're sitting in a team and all you're doing is coding, they're not getting the most out of you. You should be a person that needs to be advised. You need to know the big picture; you should know the vision of the company, the vision of the product that you're working on, and you should be able to pitch in and also give advice on new features or new things that are coming up, to be able to say what are the trade-offs, what can be done differently; is there one thing that should happen before the other? You should also be able to say, "I think that's a bad idea." So these might be good initial steps. + +**Natalie Pistunovich:** So as a developer, you say if I only kind of execute, what some people might refer to as a code monkey, you're only basically using half of the resources of this team of developers. And then you should be able to chime in and provide the feedback. And what stage would that make sense? So it can be very early on, when you've just planned the features, and prioritized and so on, but then, on the one hand, you lose the speed, let's say, if you're going to be asking all your developers. But if it actually comes down to "Here's the stage where I am as a product manager giving this to the developers", and then we might have to shelf back some ideas that were already panned out, and so on. So... When? + +**Inbal Cohen:** Well, it really depends on your style. As a product, I think it also really depends on the level of confidence the product manager or owner has in themselves and in the team. I would strongly recommend, just put aside ego, put aside any mistrust, anything that happened in the past, in your career, or with the team, come clean to the table, and understand, for any profession within development - and probably this also applies for other professions in general - if you're able to come and have an open and honest conversation with the professionals around you, and each and every person around the table sits there and says, "I'm just going to have an open, honest conversation about what's going on right now", the effectiveness rate of that team would increase at least twice as much. And it's hard for me to give numbers on it. It's just -- it creates a special type of magic. + +And then specifically for your question, take these opportunities to understand that it's not just chiming in. You are part of the thought process. I would expect, as a product manager -- for myself, I have a rule. I mention anything that I want to have developed four times before it gets developed. One time - and now we're talking a little bit about the different ceremonies, but these are ceremonies that I like having anyway; so one time, I'll just mention it in a stand up. "Hey, I had a meeting with someone about this thing. It might be interesting." Once, right? A second time might be bringing it up for like a grooming, or at the retrospective, or somewhere in one of the different ceremonies that I have with the team. "Hey, remember that thing I talked about? It's actually proven to have really good numbers." Or "This is proving to have a really strong business case. I'm going to move ahead and start writing a story about it, or a ticket about it", whatever kind of method you have. Is there anyone around the table that would be interested in working on this with me? + +\[21:58\] And then a third time would probably already be in a grooming or a refining session. I already have my ticket; this is me sitting down with the team, already opening it up, breaking it down into smaller pieces... And then maybe a fourth time, if we have an extra estimation meeting, if we have an extra meeting for putting things into our product backlog. + +So by the time you as a developer, when you saw that ticket, you're like, "I already know what this is about. I know who to approach. I know who was interested in it. I know which people were involved in the conversations. I know how important this is. I know that my PM did her homework. I'm in, I'm involved." So right at the beginning, I guess. + +**Angelica Hill:** So would it be a fair assumption to say that one of the tangential benefits of working in agile style as opposed to others, like waterfall, is that there is more space for input? And as you say, we keep on saying "Iterative approach. Iteration." Is kind of a side note benefit that developers have more of an opportunity, because we're working on this, whether it's a two-week sprint or whatever cadence cycle, where you're regularly being asked, "Are we working on the right thing? Do you understand?" Is there input? Is there more important, is there tech debt that we need to address? Is it a right assumption to say agile is also great because there is more developer input, as opposed to, "Hey, Product Manager coming in, we're gonna work on this thing for six months. Go do the work." + +**Inbal Cohen:** Absolutely. And it's not just the developer, it's the entire team. Because just imagine - in waterfall, we have people that are sitting in these very, very tight compartments, that have to do with the steps of development that they're in. So there is a planner, he's sitting alone in a room with a group of planners, and the developers are sitting in a group of developers, and the testers are sitting with a group of testers... And there's very minimal friction, because there's only handovers. + +And then agile, by doing things in an iterative way - that's one thing - and two, by taking advantage of cross-platform teams, having everybody sit together through, which again, I think is an extremely important point... The table is flat when we're looking at a team; the table is flat. No one has the higher chair. Developers are sitting the same with the project manager, with the product manager, with the designer, with the tester, with the backend, the frontend... Everybody. Everybody's sitting together around the same table. It's absolutely flat. Everybody gets to chime in. And not only do they get to chime in, they get to chime in every week, or two weeks, or three weeks, or worst case, once a month, which is pretty much the slowest cycle, right? Slowest sprint that I've heard of... But at least once a month. And it's not like you disappear in the background. As a product manager, I would expect you to be present, and I would expect you to pitch in and give your opinion. So absolutely yes. + +**Break:** \[25:05\] + +**Angelica Hill:** When we're thinking about like the day-to-day practices of agile, working in a team, pushing projects forward, I'd love to hear a little bit of how you've seen the culture change as we've become a more remote-first industry. How has remote working, especially during COVID, and now with kind of -- I think people going into the office every day is going to be very much the exception, and remote is going to be the rule... How does that impact agile practices, and what have you seen in terms of how it's changed agile culture? + +**Inbal Cohen:** It changed a lot. It shook me, absolutely... Because one of the things that I really insisted on is seeing people face to face. Even when I had remote teams in different countries. So one, I always had an office for my remote team. They would sit together in the same room. And the second thing was, I would visit them on regular intervals. + +I think the teams that I visited the least was four times a year; once in a quarter, I would go and sit in those offices for a week or two weeks together with the team. + +For my teams that are local, next to me, for example in Berlin, we would sit in one room, and I would be the person with the smallest room for my team, because I wanted everybody to sit close together as well, and not just in a big, open space. Because there's a very special moment, especially as a product person - you're sitting down, you're typing something, you have a thought, you just lift your head and you can say, "Hey, what do you think about this?" And you get that instant reaction. You get an instant reply. That's priceless. + +And now, when remote, all of that is as if gone. I've tried to weirdest things. I can talk about it now... I had sessions, I had days with my teams, camera open, everybody's just sitting at home - and this is way before we had like a proper setup; this is like super-uncomfortable... Just to try to create this type of environment. + +So what I would say, biggest change - people that are in charge of the agile within the team, whoever they are, really have to step up their game with everything that has to do with social interactions between the different team members. You have to be present, you need to find a way to actively communicate with everybody, on a daily basis. And a teeny tiny advice that I can give - if you're doing agile, any type, if you're having stand ups, and even if you're not doing agile, make sure that everybody turns on their camera for stand up. Have that 10-15 minutes in a day where you know for sure that everybody saw everybody else's faces. It kind of connects you for a moment. + +And then a second thing that I would give a piece of advice would be have an extra meeting that doesn't have to be part of your agile framework, whatever it is that you're doing. Maybe it's a coffee hour, maybe it's an afternoon drinks. Something. Give your team the opportunity to show up as they would want to, without having to be there. So it's not a mandatory time. It's not part of official work. It's just something that we can do just to have a chat, and gossip. Everybody likes to gossip. \[laughter\] + +**Natalie Pistunovich:** \[30:17\] And then what would be your advice for reaching high-performance from the office? Not remotes. Or advices. + +**Inbal Cohen:** I was so tempted to say, "Have plenty of coffee breaks", but that's not very inspirational... + +**Natalie Pistunovich:** It will make you code faster. + +**Inbal Cohen:** For sure. And that is the most important thing about developers, as you know - how fast do you code. + +**Natalie Pistunovich:** You convert the coffee into tests. + +**Inbal Cohen:** \[laughs\] Exactly. I think in the workspace -- there's a whole thing, you can read about it. There's even a book, The Five Dysfunctions of a Team. One of the most common things that I see in teams - and that's the one thing that breaks teams, and that's the one thing to work on first... If there's one thing, if you join a team and there's one thing that you would be able to get done, is check the pulse on trust. And being able to nurture and create trust between human beings is a lot easier in-person. + +So I'd say even if you're working in a remote or a hybrid setting, when you come to the office, don't come to the office when no one's there. Especially if you're making a special effort. Maybe if it's only once a month or once a week, try to make sure that there are other people there, one. Two, don't set up any meetings; have it as a day that you're trying to connect with people around you, trying to get to know them better, try to make them get to know you better... Because what happens is that one of the things that a lot of times in development - efficiency, being effective, and everything that has to do with that... People, again, forget we are human beings. And at the end of the day, when you have a person that you've managed to connect on on a personal level - it doesn't have to be too deep, right? We can just have a coffee together, and talk about where you came from, or what's your favorite food. But having that very basic human connection would increase the chances of you being able to trust that person in the workplace, and then be able to have fun. Ah, did I say fun? But yes, you would be able to have fun together, working, and be able to trust them. And high-performance teams have fun. They're not just like work-work-work... + +**Natalie Pistunovich:** Yeah, it's true, it is a harder thing to do these days, in this post-pandemic setup that many companies are, indeed, like you say, hybrid. And there's an office, you can show up, you can not show up, or maybe some companies have some days in the week where you can come, or maybe it's just like a hot chair, and you can come whenever it's available. So you say coordinating with your teammates would be a good thing to do. I guess it's especially nice if you're new in the team, or if your team kind of got a refreshment over the last few years, to get to know each other again. + +And so would you recommend kind of getting to know and making friends mostly with your teammates, or just with random people you meet in the office and you're not sure who they are or what they do? How to go about that, other than making everybody put their picture on Slack and saying "I'm gonna find you"? + +**Inbal Cohen:** Well, I can tell you that it's a very difficult question for me to answer. And the reason is that I'm thinking about the people that are listening to this podcast, right? I am pretty high in any test that I took on the scale of being an extrovert. My friends have this joke that I can't even go to the supermarket without finding a new friend. So if you asked me, honestly, for me, I make friends with everybody. I have a coffee -- I would go to other floors. I just joined recently the job that I have right now. I go on other floors to make coffee, just to kind of see who's there and like spark some type of conversation to see who I can talk with. + +**Natalie Pistunovich:** \[34:13\] That is interesting. Visiting at their coffee machines. That is a strategy. + +**Inbal Cohen:** You know that you find out that not all floors were made the same...? + +**Natalie Pistunovich:** \[laughs\] So do the coffee machines get better as you go higher up? + +**Inbal Cohen:** That's a tricky question. I actually found that the one that is on the ground floor is the fanciest one. But more of the point is - with your team, super-important. Go together on lunch; have a lunch date. That's something that's relatively easy for you to organize. But then meeting people from other teams also gives you a bit of a refresher, because you might be meeting people that are, first of all, obviously on a different team, so they don't know what you're working on exactly, and it doesn't really matter... And then just imagine you might meet someone that has a completely different profession than you; they don't even know what you do. They don't know what Go is, and it doesn't matter. You can connect with another person on a personal level. And just imagine that you would be able to go on breaks, at work, not talking about work, the entire break. And that's very refreshing. I think that it's a fun feeling, just being able to have something that is completely out of it to be able to really disconnect. + +**Natalie Pistunovich:** Mm-hm. I guess that's why some people at work join interest groups. Like, we are people who like Asian food on Tuesdays, or sports clubs, or whatnot. Yeah, that makes sense. + +**Angelica Hill:** So I'm gonna jump in and ask a question that's been on my mind, that has not to do with this, it has to do with agile again. But I think we've talked a lot about like the benefits of agile, but I'd love to hear a little bit about the cons. Like, why is agile not always the best choice? When is agile not right? What are the pain points around the fact that agile - and this is a bit of a leading question, giving away my opinion... Agile sometimes as a product manager, when there's like -- you have loose deadlines, things change on like a sprint-to-sprint basis; if you have tight deadlines, it can be a bit more difficult to stick to them, and therefore sometimes - putting my hand up - you kind of fall into waterfall to meet the deadline. So I'd love to hear a little bit about where agile kind of might be difficult, may introduce cons, may not be the right fit, may need to be adjusted to not be strict agile practice. + +**Inbal Cohen:** I love that. And I love also the way that you lead into that question, because it points back full-circle to the beginning of our conversation... And the reason being is, if you strip all the frameworks away, strip it all off, go back to your principles, go back to the values, then you will see that if you try to stand just by the -- we are working together as a team. Are we doing that? Yes. Does everybody understand what we're doing? Yes. Does everybody understand the importance of it? Yes. Are we boxing a period of time which we can try, to the best of our availability, to provide an increment of value at the end? Not just doing something for doing it. Are we trying to achieve an increment of value within a time box? Yes. + +So you're doing agile, right? You're doing the very basics of it. You might not be following your framework perfectly. Okay, you broke your sprints, you did whatever; you did a little bit of waterfall in the middle; you're actually a Scrum person and you did a bit of Kanban - okay, but there is no agile police. It's not about making anyone happy except yourself with your team. + +\[37:59\] So I think that the biggest disadvantages of agile is pretty -- and this is too, again, culty, but it's more about the misrepresentation of agile; it's more about the things that we put on top of an agile thinking. Just have an agile thinking. If you start from that and you're doing that, you're already good. And being flexible. We're thinking about these iterations, we're thinking about the frameworks, but agility - it's being flexible, it's moving, it's being able to change according to the things that are happening around me. And this if that's something that you feel a need to do, I would trust you as a professional saying "This is this is it. This is the pivot that we need." + +**Angelica Hill:** And who in your opinion is the driver of the methodologies, the frameworks that a team should use? This is just my personal question and something I've been kind of thinking through... Like, a lot of teams have project managers, product managers, engineering managers... Who is the driver of the practices? Is it truly collaborative as always it should be, or is it that like it is really part of the project discipline to establish what is the right framework? I'm just interested in your personal view; there's no right answer, I'm sure, but it's something I've kind of been thinking through. Like, can an engineer be like, "Hey, Angelica, this agile thing is just not working. We need to try this other thing"? + +**Inbal Cohen:** Absolutely. And one of the things that we have -- we can call people that way; I have been called that way for a while... It's called an agile evangelist. Going back to like, religious, "This is not a cult." What is it being an agile evangelist, and why do I think that each and every person that is working with agile should be an evangelist? I think that one of the principles or the things that guide me when I mention this is I am here to provide you, as a team or as a company, with information on how we can get things done in an effective way, where everybody's super-happy. Okay? This is what I do. That's my profession. I'm going to tell you the different types of things that you can get done. You can choose if you want to drink the Kool Aid, if you want to participate or not. + +As a team, as each and every developer sitting in that team, you need to be able to say "I accept this, I understand it. And actually, this is helping my work. And I'm going to stick to it. I want to make sure that everybody sticks to it." And I think it's a lot like -- if you've ever been in a good meeting, with really active participants, you would notice that there is a person that's leading the conversation, but you would always have the additional people in the conversation being able to say, "Oh, sorry, off-topic..." Or, "Hm, let's take it offline." Not only the moderator would be the one driving the conversation; all the people around the table would be actively participating. Because they're a part of it; they get the point. They want to arrive to a conclusion in this meeting. And working with agile is exactly the same. + +So you as the product leader can provide the information to the audience, to the team, to the company, but it's each and every person's role to be a small evangelist of themselves. Each and every person within that team needs to be able to say "That sounds waterfallish." Or "What about the deadline?" "I don't understand that increment of value." Or "Hey, what happened to our retrospective? I really need to pour my heart out." + +Each and every person within the team should be accountable, where we have a shared responsibility and accountability, and an individual one. And in that shared one, we're all maintaining that agile atmosphere thinking, because it benefits us, and we want to. + +**Angelica Hill:** \[42:03\] So is there a world in which we would maybe have almost like a agile interview stage? Like, it feels to me, as we talk more, that everyone on the team, regardless of discipline, engineer, designer, product, project, should know - not in great detail; they don't have to read every single documentation or every book, but the basics of what agile is, what waterfall is etc. Is there a world in which perhaps it would bode well to have like almost like a project methodology interview stage, where people just talk through the kind of working styles... Have you done waterfall before? Have you done agile? Both to make sure that everyone has that information, i.e. like if you were going into an interview and you saw you had an interview about agile, you'd probably do a quick YouTube video on it. But also if there is an overarching, like certain companies tend to err on the waterfall, versus the agile; it almost feels like it could be like a culture fit. Just spitballing here... What are your thoughts? + +**Inbal Cohen:** Yes. And I actually do that. + +**Angelica Hill:** Oh, awesome. Tell us more. + +**Inbal Cohen:** I love it that you're mentioning that. There was one piece of an interview - and this is me giving away a bit of my secrets, but there's one piece of an interview that I do for any person that I interview, for any job. It doesn't even matter. It could be sales, it could be any job. And there's one thing that I do, and I don't give them prior preparation, like "Hm, we're going to test you on agile now", because I don't want people to prepare in advance. I want to see your mindset. I want to get an understanding, is this something that you could put in your mind? And what I do is I have the websites - there's a website for the principles - and I just put on the principles on the screen, and I asked them to look through them and tell me, if you had to choose one... And I give them time to read them through. If you had to choose one, the most important one, which one would you choose? And why? And then we have a conversation about it. + +And it's super-insightful. Super-insightful to see what do you choose, why did you choose that one? Where are you coming from? What's the job, what's the role that you're aiming for? And then involving that into an entire conversation about how much prior knowledge they had. + +**Natalie Pistunovich:** That is a really interesting way of going about that. I do see in a lot of job descriptions for developers that in addition to mentioning the technology, they also mention the methodology that you're using. So yeah, it makes sense. But I can't say I was ever asked about that. But I guess if I don't have to know them in advance, it will be interesting for me to kind of read through and see which one resonates with me. So I like this. If you would ask me to read them in advance and have that prepared, it probably would be a different thing... So I'm glad to hear that this is not that. + +**Break:** \[45:05\] + +**Natalie Pistunovich:** Last question for today would be, as a Go developer that is working as part of a cross-platform team, what advice, or what exactly would I be doing? + +**Inbal Cohen:** I think that that's great. It's a great summary. I would say, take this entire conversation, and if I had to crunch it up into a few sentences, right? So one, get to know your team. Who's on your team? What did they do? Why are they doing that? Why do they think that it's cool? What's their favorite food? Right? And then a second thing is, who is the product leader on your team? What can you learn from them? A product leader on your team should be a source of information, the same like anyone else. The same like you would go in and get advice from your backend developer, from your frontend developer, from your designer, from whoever about this next thing that you might be doing - go and get advice from your product leader. Like, what are you doing? What does that actually mean? Right? + +Then another thing would be think, try to have a thought with yourself, "What motivates me? What do I don't like so much about my job right now, and what do I really enjoy?" And then try to initiate a conversation at the end of whatever cycle that you have - if it's a sprint, if it's a week, if it's two weeks... Usually, what we have at the end of these cycles is a retrospective, which is a type of meeting where we can sit and have a conversation about the week that we had. That would be a fantastic opportunity to say, "Hey, I actually thought about it this week, and one of the things that bothered me are that we don't meet as a team. That we never have lunch together. Is that something that we can arrange?" So being able to be open and honest. + +And then the last one would be think about trust. Do you trust your teammates? Do they trust you? Is there something that you would be able to do to up-one the trust in your team? + +**Natalie Pistunovich:** That's a great checklist, and definitely something to bookmark. + +**Angelica Hill:** Yeah. And I would double-click just quickly, and have a comment, slash also request, which I'm going to regret saying, but I'm gonna say it anyway... Please, software engineers, ask your product manager what that incremental improvement you are providing is. Like, if we're following agile, we should be providing value on a sprint-by-sprint basis. Also your product manager, "Why are we doing this work in this sprint? What are we providing? We're agile. What value are you getting us to do? Why are we working on this feature versus this other one?" Continue to ask those questions, one, because it helps your product manager think in that agile fashion, but also it holds them accountable, and ultimately, to your last point, it builds trust, you as engineers knowing very clearly "Okay, my product manager has a very clear reason why this ticket has to be done today." But also, we as product managers should be held accountable. We should be thinking through like, "Oh, I just personally want this tech debt done." No. Why is this more important than this other feature? + +So I think it's about building that trust both ways, but also holding both the engineers and the product managers accountable to truly be agile. And if I'm like, "Oh, we have to work on this thing for six months", I would like my engineers (this is the bit I'm gonna regret) to be like "Angelica, why are we working in waterfall? I thought we were an agile team." Final point... \[laughs\] + +**Inbal Cohen:** Amen to that. + +**Natalie Pistunovich:** Definitely, yeah. Having a why is something I can absolutely stand behind. And I think everybody will agree with you, Angelica, that this is a popular opinion. But now it's time for the unpopular opinions. + +**Jingle:** \[50:42\] to \[50:59\] + +**Natalie Pistunovich:** For anybody who's not joining the live stream, Inbal was dancing to all the tunes like she's heard them forever... So either you've been listening secretly to this podcast, or I can understand why you make friends in the supermarket. You're are super-friendly. + +**Inbal Cohen:** Yay! + +**Natalie Pistunovich:** \[laughs\] + +**Inbal Cohen:** I love the tunes. + +**Natalie Pistunovich:** Thank you, thank you. Okay, so what is your unpopular opinion today, Inbal? + +**Inbal Cohen:** My unpopular opinion has to do with roadmaps. Roadmaps are a very, very dangerous tool, that a lot of times is used against development teams, instead of by development teams for the business. + +**Natalie Pistunovich:** Interesting... + +**Inbal Cohen:** And my unpopular opinion would be - and this is funny, because it's an unpopular opinion in many companies that I've been in, but it's actually how they're supposed to be done... Which is roadmaps should be a combination of two directions meeting in the middle. From the top, we have the Why; the business telling us "These are the numbers that are going to make our company super-successful." We have the middle part, which is the product saying What. What are we building. I understand the need, and now I'm going to think about this fantastic software that's going to solve all our problems. And on the bottom, we have the How. These are the development teams - how is this going to get done? + +And roadmaps are supposed to take those two edges of the Why and the How, and bring them in the middle, which also is the house of When. So when we're trying to drop on development teams when things are done, we're limiting how things are done. There's all sorts of different types of metrics that we can look at that have to do with effort, and time, and what comes out in the end... But I can tell you that when we do roadmaps that start from the bottom, on how the different things can be done according to those business objectives, they are much more accurate. That's one thing. And the second thing is development teams, regardless of who is representing them, are much more confident on giving red flags, or notifying whoever needs to know if there are any delays. Because they are accountable for the dates that they provided themselves, and not resentful of the dates that were brought upon them. That is my unpopular opinion. + +**Natalie Pistunovich:** I'm like that meme with all the numbers... I know you two understand each other, because you both are product, but I was lost in here... Can you say that in one line for me? I'm like that meme where you see all the math formulas... Like, I understand, it makes all sense... It does need to meet in the middle... But what is that unpopular part of the opinion? + +**Inbal Cohen:** \[53:59\] That roadmaps need to come from development up, and not from business down. So roadmaps aren't something that you are given as a development team. It's something that you build. + +**Natalie Pistunovich:** I agree with you. I think you have my voice. \[laughs\] + +**Angelica Hill:** I want to be clear, is the proposition that software engineers work on the roadmap and bring it to product managers? I want to be very like logistical here. And are they then accountable for saying why they think that it's the correct course of action? In the same way, if you do it the other way around, product is accountable for convincing that and evangelizing why that's the correct roadmap for the engineers? Would it just be the opposite? + +**Inbal Cohen:** So how it would work in like in a really nice setup... Business comes in and says "We have an X amount of new clients that we have to onboard. We have a deadline. It's a hard deadline. We have to have this by the end of 2023, or the company goes belly up." Right? So these are the types of deadlines that we can't argue with. And we come down to the development team; as product, we already have a What in mind, we already know what we want to build, we have an idea of what we want to get done in order to achieve that business goal... And we come to them and we say, "This is kind of what we want to build. It's a rough idea. It's not particular; it's not to the tee. I have a general vague initiative that I have in mind. What of this could I get done in 12 months?" And then we would sit together and tear it apart, take down the pieces... This is quite literally the exact same thing like we're planning a sprint; it's exactly the same like taking a story, breaking it apart into tasks, and then putting it up and seeing what fits inside my backlog without overdoing it, or without making it too small. Exactly the same thing, just at scale, on a roadmap. + +And what happens a lot of times is that companies kind of forget that it's a sprint at scale. They forget that, and that's one of the issues that we have a lot of times with companies that think that they're doing agile, is that a lot of times when you have stress, any stress - it could be stress on the system, it could be stress on people, it could be a business stress... Any type of stress that comes on top of people that don't have an agile mindset, they defer to either it's waterfall, or micro-managing, or all sorts of different types of behaviors that are absolutely not agile, and they kind of forget that they already have a way to actually have a prediction, or have some type of understanding of what they can expect in a period of time in the future. And it doesn't really matter if it's two weeks or a year. + +**Angelica Hill:** So can you even use roadmaps effectively in agile? Or is there an argument that any roadmap you create today will be different tomorrow, so what's the point? + +**Inbal Cohen:** There is a point. And the reason that there is a point is, one, I don't really like roadmaps; I like more the idea of goal maps. I have goals I would like to achieve. It's not about having a specific amount of lines of code, it's about an increment of values that I want to achieve in a certain amount of time. And there are all sorts of shortcuts that I'm willing to take along the way if things don't work out; that's being agile. + +But on the other hand, it is very important for me to have, one, a Northern Star. Where am I aiming? What is the direction that I'm aiming my ship? And then having these milestones on the way as these buoys to make sure that I'm still on course to get to that direction that I want to arrive to. If I don't put any of these insights, I might even not have a ship. You might find out that everybody's in like their lifeboats, all over the place. You need to have a direction, and you need to have these milestones to guide you through. And also to be able to even say, "We didn't hit a milestone." That's also okay, as long as you know where the milestone is, and where are you, you can still calculate where are you on route to give some type of prediction on what's going to happen. + +**Natalie Pistunovich:** \[58:16\] Yeah, that makes a lot of sense. It also connects to a conversation we had a couple episodes ago about velocity. It makes a lot of sense, everything, so thank you for clarifying everything and sharing all that. I'll use the last minute to share my unpopular opinion that I've been planning for the last week... It's on the topic of food. My unpopular opinion is that you should only get nice, fresh ingredients, fancy everything, if you're planning to be easy on the spices. It doesn't make sense to take -- I don't want to say a nice steak, because I'm vegan, but it's a very easy to explain example... Don't take a piece of meat if you're going to drench it with spices. Take tofu. It's going to taste the same, with the same spices. + +The interesting thing is that we will be asking our Twitter followers for how unpopular is this unpopular opinion, and I think, at least in the field, in this little crowd of developers, Inbal, what you say, you might end up in the wall of popular unpopular opinions, because it does make sense for developers to give that feedback. Thanks for sharing that, very much. + +**Angelica Hill:** But the developers want to be accountable for that cognitive extra load... + +**Natalie Pistunovich:** This was not part of the unpopular opinion... + +**Angelica Hill:** Do they want to? \[laughs\] + +**Natalie Pistunovich:** I can't hear you... The connection breaks... \[laughs\] + +**Angelica Hill:** Do developers want to take on that extra cognitive load? Do you want to be accountable for putting together the roadmap? + +**Natalie Pistunovich:** Yeah, that is a good point. I'm thinking whether this is a popular or an unpopular opinion. That's a fair one. + +**Angelica Hill:** I mean, it's probably two sides of the same coin, and part of me is like, I really want to continue doing it. I really enjoy doing it. But also, another part of me is like, developers already have so much they need to do in their lives. We put so much accountability on them. I feel like this is something that almost like we as product or project can, with heavy input from them, drive the boat on. + +**Natalie Pistunovich:** So here's a question for you to think about, just kind of going forward, nothing related to this specific episode... So you mentioned developers have a lot of responsibility, and a way to go about let's say respecting this responsibility while writing code is actually writing a lot of tests to make sure the behavior is as expected. What is the test equivalent of roadmaps, sprint planning, and so on? + +**Angelica Hill:** Oh, I love that... That was a brilliant question, that we could have a whole other episode on. + +**Natalie Pistunovich:** Absolutely. We will. + +**Inbal Cohen:** I should have used that as my unpopular opinion, talking about testing... \[laughs\] + +**Angelica Hill:** Next time. + +**Natalie Pistunovich:** Well, we have many more episodes, and we hope you'll join us again. + +**Inbal Cohen:** Thank you for having me. I had fun! + +**Natalie Pistunovich:** Thanks a lot for joining, and thanks everyone who listened. diff --git a/The art of the PR Part 1_transcript.txt b/The art of the PR Part 1_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..336b59a53899b5b3d67057b1b8b49db7e176a3c9 --- /dev/null +++ b/The art of the PR Part 1_transcript.txt @@ -0,0 +1,363 @@ +**Natalie Pistunovich:** Hello, and welcome to Go Time. Today we're going to be talking about PRs. What makes a good PR, how do you do the best PR review, is there such thing as a PR that is too small, too big, too filled with emojis...? We'll be debating all the details and trying to help our fellow Gophers master the art of the PR. + +Today I'm joined by three wonderful PR pros. First, we have the wonderful Jeff Hernandez, who is an associate software engineer at the New York Time. Hello, Jeff. How are you today? + +**Jeff Hernandez:** Hi. Doing well, thanks for having me back. + +**Natalie Pistunovich:** Thank you for joining us again. Next up we have Sarah Duncan, who's a staff software engineer at the New York Times. She also teaches an introductory programming course at a high school. Thank you for joining us. I know you're a first-time Go Time guest, so - lovely to have you. + +**Sarah Duncan:** Yeah, thanks for having me. + +**Natalie Pistunovich:** And last but certainly not least we have Natasha Dykes, who is a senior software engineer at the New York Times, and happens to be a cycling enthusiast. + +**Natasha Dykes:** Hi. Thanks for having me. + +**Natalie Pistunovich:** Hi. Thank you for being here. And we have the beautiful, the wonderful, the incomparable Natalie, who is my co-host. Hello! + +**Angelica Hill:** Hi, Angelica! I think it's very smooth you skip pronouncing my last name. I would probably do the same thing. It's so complicated. + +**Natalie Pistunovich:** I haven't been on in a whole second, so I'm kind of trying to minimize the amount of babbling and mistakes I get myself into. \[laughter\] + +**Angelica Hill:** Great strategy. + +**Natalie Pistunovich:** Yeah. We'll see if that persists throughout the episode... I might just get over-excited and fumble over my words. But thank you all for joining me today... I'm extremely excited to talk about PRs. + +So we're gonna start with the very basics - what is a PR, and why do we even do them? I'm gonna pass it over to you, Sarah... When you're talking to your wonderful high school students and they go "What is a PR? What is this thing?", how do you explain it to them? + +**Sarah Duncan:** Well, PR stands for pull request, and it is typically used to refer to somebody who's been making changes to a shared codebase, making a request to add those changes back into the common, main - typically it's like a branching situation, so the main branch of that codebase... And that also typically corresponds to moving that code into production, if it's a production system... So it's a way to get a review on your work and basically ask to add code to the main system. + +**Natalie Pistunovich:** And why is it useful? Why is this something that we want to be doing? Is that something that maybe -- I don't know, Jeff, do you find PRs useful? + +**Jeff Hernandez:** I mean, coming from, at the Times at least, I'm an associate software level, so I'm kind of like the entry-level... So it's a great way to get feedback from my senior engineers, and basically get a lot of kind of feedback from them in terms of I can be doing better, or code structure, how that could be improved... Or even tiny, little optimizations... Or maybe there's a certain way that, for instance, Go likes to do things, because as we all know, Go is very opinionated... So it's like a way to tap into that resource from our senior engineers. + +\[08:01\] I feel like it's a great tool, especially for someone that's coming into a new team, kind of getting the lay of the land, getting the norms that you typically might not get in through other forms of documentation. + +**Natalie Pistunovich:** For sure. And in terms of assessing PRs, is there such thing as a good PR review? Is there such thing as a bad PR review? I would love to hear how you assess going about either putting in a PR, or reviewing a PR. Maybe Natasha - when you're putting in a PR, how do you decide whether it's time to go, or whether you need to wait and do some more work... At what point do you feel like "No, this is ready to be reviewed by my peers"? + +**Natasha Dykes:** I think for me it's helpful when I review a ticket - usually, it's for a feature or something that I'm working towards... I review it to see if the work that I've done actually meets the requirements, and at that point I can either say "Okay, it's ready to go." I've cleaned up any notes for myself, or made sure I did `go fmt`, all of those small things... And then I'll open up a PR. But I would say you can even open up the PR before you're ready to merge. Sometimes it's good to just get that early feedback... So it really depends on the work that I'm doing. + +**Natalie Pistunovich:** And Natalie, I see your intake of breath... Do you have something you'd like to add? + +**Angelica Hill:** Yeah. You asked earlier what is a PR... And it's interesting also to compare PR and CR, and why is it even "pull". So PR, as Sarah said, stands for "pull request." So let's split that into two questions. Why "pull"? Why not "push", why not "merge" or some other thing? And what is the difference, or what do you prefer - CR, which stands for "code review", or PR as "pull request"? It's an interesting question to discuss, it's also interesting even a little bit in the context of like what does that represent. I'm curious to hear everyone's thoughts. + +**Natasha Dykes:** I think code review is more semantically correct for the work that I'm typically doing... But I think the changes that you could potentially make don't necessarily need to touch code. It could be like a readme update, it could be some other stuff that's kind of supporting the repository versus just the code. I don't know why it's called "pull request" though, I've never really thought of that, so I'm curious to see what others think. + +**Jeff Hernandez:** Now, this gets me kind of wanting to go to Wikipedia and learn about where "pull request" came from, and that whole background. That's one of the things I'm interested about, just like where everything came from in terms of software engineering best practices, and naming conventions... Because once you're in the industry, you know, it's just something we know; but it's not something -- at least it's a term that's used across the whole industry, but it's not something that we as a team choose to use. It's kind of like a shared language amongst software engineers. + +**Natalie Pistunovich:** So I was super-cheeky and I did google it just now... And it says that the name "pull request" comes from the idea that you're requesting the project to pull changes from your fork. That might not encompass all ways that we now use it in our language, but that's what Google's telling me on the top line... + +**Angelica Hill:** Yeah, it is interesting, right? You have this project -- I think most of us use Git in some way; GitHub, GitLab, or... I don't know too many other personal variations, but I'm sure that exists as well. So we all kind of eventually are used to the concept of having a main branch, and then branching your changes from that, and then asking to merge that back, whether frequent, or not frequent... And yeah, we all say PR. I've been always saying PR, but then recently I had to work with this ticketing system that is called "Click up." And there they said that the label was code as in CR. And that was kind of interesting. Probably the first time I remember, let's say, encountering this style. So I wanted to dive a little into the semantics of that. Then there's also merge request, which kind of makes sense, but actually also not really in use. + +**Natasha Dykes:** \[12:12\] I've also heard it called a changelog, or CL... So there's a lot of different terms for it. + +**Angelica Hill:** What are you familiar -- what do you know changelog to be? + +**Natasha Dykes:** It's the same thing as like a pull request, but certain companies call it a changelog. + +**Angelica Hill:** So you're submitting a changelog? + +**Natasha Dykes:** Yeah, that's what I've understood it to be. It's just like something I heard in passing. + +**Jeff Hernandez:** Yeah, I've seen that in the Go repository, references to actually the GitHub Go repo, and issues, people referring to changelogs. + +**Natasha Dykes:** Yeah. It's a Google thing. + +**Jeff Hernandez:** It's a Google thing, yeah. I'd totally be interested in learning more about that. + +**Natalie Pistunovich:** And that was me thinking that the most basic of questions, "What is a PR?", was just me doing my due diligence for the newbies, but now I see it's a whole debate. There's so many different words to use. This is great, we're opening up a Pandora's box of PR words, and ways to think about code... + +**Angelica Hill:** I mean, software is all about naming things... It starts this early. + +**Natalie Pistunovich:** Okay. So when you're thinking about a PR, I've heard many people complain about "Oh, this PR is too long", "Oh, this PR is so short. Why didn't you put it in one big PR?" Is there such thing as a too long or a too short PR, too big/too small? Or is it really just down to dealer's choice, whatever that team is happy to review. If you're happy to review a 1,000-line PR, or a 5-line PR. I'd love to hear anyone's view on the length of the PR and how big it should be. + +**Sarah Duncan:** Yeah, I definitely think that pull requests - or merge requests, or code reviews - can be too long. Sometimes a pull request is too large, and my benchmark for helping set that norm on my team for what is an appropriate-sized pull request is really around how well the pull request can be reviewed. So if you do have a thousand-line pull request... Another common adage about engineers - we're lazy, and we're probably not gonna review a thousand-line pull request as in-depth as we would a 20-line pull request. + +So I think it's about thinking about how you can best set up your reviewers to give you a quality, thorough review. I have some thoughts around how to set that norm, but I'd love to hear from others what you think about a length of a PR, and whether a pull request can be too big or too small. + +**Natasha Dykes:** Yeah, I don't think that pull requests can necessarily be too small, because it just takes one character to make a bug... So you're gonna have to make that change... But I do think that -- I agree with you, it could be too long to have a reviewer to actually sit down and understand all the changes, especially if they're not as familiar with the code as you are... So I don't necessarily have the best strategy, but I think just encapsulating certain functionalities helps, instead of just an entire feature... Because it might be broken down into different core parts. So that's typically how I lean on it. I don't know, Jeff, do you have any thoughts? + +**Jeff Hernandez:** Yeah, I think it's totally depending on what you're working on at the current moment. So if you're building a new API from the ground up, setting up the handling for the JSON, the payload, maybe that can be a PR. And then the actual business logic can be a separate thing. You're building up as you go. That's something that I've learned from other people on the team... Because I have been guilty of the extremely large PRs, where the ticket is done, but it's all in one PR, and no one wants to review it, so you're gonna maybe get a review by next week if you keep pinging people and bothering them... But otherwise, no one's gonna wanna touch that. + +\[16:04\] And then you don't ever wanna get in a situation where you have to offer to go step by step with the reviewers... Like, "I've made this change because of this reason", on a call. I feel like that's the worst-case scenario, where you have to actually walk them through it. Sometimes it's necessary, but it's something I would avoid. It's kind of like a walk of shame. You have to go through the whole thing with them, but... They can definitely be too big. + +**Natalie Pistunovich:** A walk of shame where you can walk through the glory of your coding... \[laughter\] + +**Jeff Hernandez:** I guess... \[laughs\] + +**Natalie Pistunovich:** I'm just teasing. + +**Jeff Hernandez:** It's all how you think about it, right? + +**Natalie Pistunovich:** "Look at this beautiful PR that I've constructed over many months, that you now have to review with me... Look at this clever naming convention I did. Look at this great function. Don't you love this goroutine?" + +Awesome! So we talked a little bit about -- well, actually you alluded to this, Sarah, like being cognizant of the people who are reviewing your PR. So I'd love to hear a little bit about how many people should review your PR before you merge it; is it one person, is it two people, is it the entire team? What are the thoughts around reviewing the PR and how many people need to know about this great work you've done before it's in the world? Natalie, you're smiling... + +**Angelica Hill:** I'm sorry I'm doing this again, I want to bring back to the previous question... \[laughter\] + +**Natalie Pistunovich:** Do it. + +**Angelica Hill:** So when you say too long or too short, what do we measure? Number of lines? Number of commits? Number of files? It can be -- depending on many things, the answers can vary in the same PR. It can be one PR where you removed one file, but that's like many lines... Is that long, is that short? This is my thought now. But whoever listens doesn't know that Angelica can see me thinking out loud like that, so... I'm guessing there's a question coming up, so you have to trust her psychic abilities as a product manager to kind of look at the engineers in front of her and like "You want to say something. Speak." That's the secret superpower of product managers. + +**Natalie Pistunovich:** Yeah, I can see it. Also, I'm learning, now, actually just as a Go Time host, every time I ask a question I should pause and defer to Natalie, and be like "Do you have a follow-up question?" \[laughs\] That is a great question, I'm very glad that you answered-- asked it... That definitely said asked it... I'm already babbling. Do you have a view on that? + +**Angelica Hill:** Yes... + +**Natalie Pistunovich:** Wonderful. I'd love to hear it. + +**Angelica Hill:** But we're here to hear other people's opinion, so I'm curious to hear the crowd's opinions more than I am interested in sharing. I mean, I'm also happy to share mine, but I'm more curious about others. Is there a convention in your team, for example? Would you say this is too much - what would look at? + +**Jeff Hernandez:** Yeah, I'm really curious about hearing more about Sarah's - the team's standard... But at least from me, if I'm reviewing something, the first thing I look at is the number of files that were changed. That's just like the easiest thing to look at. If it's a long list, I am gonna go to lunch maybe, and then come back to it later... It's definitely easier to review something that's a couple files, versus something that's touching multiple directories at a time... Because then I have to go back and forth, referencing things and seeing how it might affect other things... So definitely files would be my number one thing. And then, typically the norm on our team is one commit per PR. So it's not something that gives much information for us. + +**Sarah Duncan:** Yeah, I definitely think files are a good initial indicator... And sometimes that can be misleading, because you could have a bunch of files only have one line change, or you moved a folder into a subfolder, and that changed a bunch of files, and you can just check all those off as okay... But I think that it's a good question, because I think it's a hard thing to actually measure well... + +\[19:57\] The analogy I try to use and that I recently used at the New York Times to kind of explain how to break down your PRs - it ties into your question, Angelica, about the number of reviewers... My analogy is that if my ticket is to bake a cake, and let's say it's chocolate cake, we've got some butter cream frosting, a little raspberry layer in there, it's a nice cake - if I go and bake the cake and come back... And Angelica, you're a frosting expert, and Jeff, you're a cake expert, and Natasha, you are the absolute queen of fillings - it's harder, I have to get all of you in a room and cut a slice of cake, and you have to pick it apart and be an expert on your piece, and it takes more work to give feedback... And it takes more work for me to go fix something. + +So if I'm trying to perfect the chocolate cake, and Jeff, you're like "This cake is too dry.", You have to go back and make it less dry", that is so much more work for me to reconstruct that cake all over again... But if I break that down and first I make my chocolate sponge, Jeff, you taste that and you give me feedback on it, I make my filling that I'm gonna use, and Natasha, I get your input on that... Angelica, I ask you for your input on the frosting, and kind of perfect those individual pieces, and then I assemble - that final assembled cake is gonna be a lot more successful. + +So I like to think about it as like breaking it down so that ideally one person could review it, and maybe one subject matter expert could review that part. Then when we get to the assembly piece, you're judging me on the assembly, you already know that the individual components are -- we're all on the same page about what those components are. + +So I think it's a hard thing to measure, because even just that "Oh, one subject matter expert is not a hard and fast rule." This is one of the things I think is something that comes more with experience and intuition. It's a skill that you're able to hone as you gain more experience around what the right size to ask for feedback on is... But that's my analogy, is that if you're baking a cake and you're asking for a review on it, it's easier to get the individual components reviewed first, and then bake your cake... Or assemble your cake. \[laughter\] + +**Natalie Pistunovich:** That was so great... + +**Sarah Duncan:** The teacher in me... You know, we've gotta bring in the fun things. + +**Natalie Pistunovich:** That was wonderful. I feel like I've learned so much, genuinely. I will never think of PRs and constructing of a feature again in the same way. Like, "Can I see the frosting, please?" + +**Sarah Duncan:** Please. \[laughs\] + +**Natalie Pistunovich:** "Let me check the topping." + +**Sarah Duncan:** Yeah. How moist is the cake? + +**Natalie Pistunovich:** "How moist is this cake?" I'm gonna have to be very careful whom I say that to, because some people don't really like the word "moist". I don't particularly, but it's fine. + +**Sarah Duncan:** Judge your audience, yeah. + +**Natalie Pistunovich:** Judge my audience. Good perception. So flipping over kind of to the other side of a PR, how do you go about reviewing a PR? I'm gonna pause. Natalie, would you like to talk about putting in a PR in any more detail before we switch over? + +**Angelica Hill:** \[laughs\] No, no, sorry. I'll go back to being German and sticking to the schedule. + +**Natalie Pistunovich:** No, no, please, don't. I came across like I was being a little bit sassy... + +**Angelica Hill:** No, no, no... + +**Natalie Pistunovich:** ...but it was a genuine question. + +**Angelica Hill:** I love making German jokes, any day. + +**Sarah Duncan:** I should have made it a German chocolate cake. I'm so sorry that I missed that opportunity... \[laughter\] + +**Natalie Pistunovich:** Oh, my gosh... + +**Angelica Hill:** A Strudel... This is Austrian. + +**Natalie Pistunovich:** Natalie will be the final -- like, once the beautiful cake has been constructed, she is the ultimate SMA + +**Angelica Hill:** Wait, did you use sugar or salt? + +**Sarah Duncan:** The ultimate taste testing... \[laughter\] + +**Break:** \[24:00\] + +**Natalie Pistunovich:** Great! So you are going to review a PR... What are the almost unspoken rules of PR review? Are there actual rules of PR review? I'd love to hear how do you approach reviewing your colleagues' PRs. + +**Natasha Dykes:** I think it goes a long way if you lead with empathy. Go through the PR, address anything that you think could be updated, but not in a way that you're talking down to people... Just having a willingness to learn, a willingness to teach... I think these are all core factors of doing a good PR review. And then you can get into the nitty-gritties of like your team-specific strategy, and patterns and all of that, and the correctness of the code. But I think the first part is really key. + +**Sarah Duncan:** Yeah, I totally agree. I think there are a lot of small behaviors that we as engineers can adopt to be more empathetic in our code reviews. I really enjoy some of the resources that other people have already written and shared around this. I particularly like Alex Hill's "The Art of Giving and Receiving Code Reviews Gracefully" (something like that). That idea of giving a code review gracefully I think ties into the empathy thing you were talking about, Natasha... Just like putting yourself in the position of the person who's receiving the feedback. Ultimately, a pull request - this code review is feedback, and is collaboration in our every day as engineers... So some of the simple things, like instead of saying "you", saying "we". Instead of making a statement, asking a question. + +For example, instead of saying "Oh, you should use this other function. It already does what you're doing here", asking "Oh, can we use this other function here? Is there something that we can reuse?" And that gives the opportunity -- because you could be wrong. I'm a staff engineer, and I'm wrong all the time... So if I'm just coming into a code review and saying "Oh, you should have done this" and "You should have done that", that not only is not giving feedback in a way that will be easily received, but it's really assuming that I always know the answer, and I always know what's best. + +\[28:13\] I think remembering that the person who's coming in with the pull request has spent so much time, presumably, on this problem, that - yes, the fresh pair of eyes is really helpful, and you might see something that they didn't, but also giving acknowledgment to the work that they've put in on this pull request, and that they might have thought through that problem, and that there might be something that you're missing, because you haven't spent as much time thinking about the solution. So that empathy is really important. + +**Jeff Hernandez:** Yeah. + +**Sarah Duncan:** Check your ego... + +**Jeff Hernandez:** Totally, plus a thousand what they've said so far. It's not a one-sided street, it's two sides. You can take something back from the review, and then they can take something forward... It's always an opportunity to learn something new, especially from people who have way more experience than you do. And this is just an opportunity to ask questions, especially like "Oh, why did you do it this way, versus some other way?" It's definitely a great opportunity to kind of -- you know, just learn more about what they're working on, and learn more about the system. There always has to be an eye for "Oh, is this what we're trying to achieve with this particular ticket" But definitely leading with empathy is something that's great. I also wanna look up that article now, because I've never read that... The Alex Hill one. + +**Sarah Duncan:** It's really great. + +**Natalie Pistunovich:** I'll make sure I put it in the episode notes. And it is a great article... So when you're thinking about reviewing, are you predominantly reviewing for the functionality, and like "Does this thing work?" or are you also commenting what are the I guess rules around commenting on style? Like, how the code is being written, the stylistic choices that have been made. + +**Natasha Dykes:** I think a bit of both... + +**Sarah Duncan:** Yeah, definitely. And not to over-reference this article, but one of the things I think is one of my key takeaways from this was that there are those kinds, like code style things that can be automated. So having a team norm of "Oh, we're always using this linter for this codebase, we're using this formatter. Pre-commit hook - those have to pass before you can make this pull request." It just automates away a lot of the things that we can be kind of nitpicky over... But sometimes there are more code patterns, and those are harder to automate, but they can be a source of contention, because there can be a lot of very strong opinions around what patterns we're following and how code should be structured. + +That's where I think having a set of norms on your team that you regularly revisit when somebody new joins, when you have a new repo you're working in... Those norms will help smooth that conversation, because if you have all already agreed, like "Oh, we're gonna make sure we follow DRY practices", or whatever - those kinds of agreements make it a lot easier to have that code review conversation, because that's like a shared expectation that you have. And it's the same thing for giving interpersonal feedback. If you have a shared expectation and a shared goal, then you can easily use that as a reference point and be like "Hey, since this is something that we've agreed on as a team, I'm noticing this here. Do you think that we could reshape this so that it follows this practice that we have agreed to use for this repo?" That's a much easier conversation to have than "Oh, I don't like how this is styled. I think we should do it this way instead", and kind of like bringing your perspective into it. It's a lot easier to bring a team norm that you've already all agreed on, that's like the team's perspective on this, and the team's perspective on how to move forward. + +**Jeff Hernandez:** \[31:58\] It's really important having that team understanding. At least in my previous company we had shared standards as to how things -- we had pillars in everything, we had a standards committee that we were trying to... For code style, and stuff like that. + +But on the other side of the coin, I'm kind of -- sometimes if I see a spelling mistake, I will point that out in the PR, just because it's... Like, if it's already committed and I see it, I wish someone had called it out in the PR. In my PR I will fix it... + +So it's just like those little tiny things that kind of add to your personality as a reviewer. You might be known as that person that calls out your grammar mistakes, which I feel like I used to be, and I'm not as much these days... But when I see it, I will sometimes call it out. + +**Natasha Dykes:** I think I've gotten that feedback on my review from you, so... \[laughter\] So I can attest. + +**Natalie Pistunovich:** I'm intrigued - do you review differently depending on the level of the person that put in the PR? I.e. if you're going in and you're reviewing a staff engineer's PR, versus an associate engineer's PR - do you approach it differently? + +The reason I ask is that - maybe this is a leading question... Say you're a staff engineer, and you're reviewing the PR of someone you know is preferably new... Would you approach that slightly differently, in that you might add more comments, and maybe more detail as to why you've suggested it, because you know this person's still learning? Or, agnostic of level, you approach every PR in the same way. To summarize, does it matter who put the PR in? + +**Natasha Dykes:** I think it could help. If I know that someone is new to the company or to the team, I usually kind of pepper my review comments with links, or context, or even ask if they want to jump on a call and we can talk through certain things, if they have more follow-up questions, just so it's a little bit more synchronous conversation-wise, instead of all over the place... But I also try to leave good feedback, even if there's a person who's above me and run laps on me with the work that they do... I like to just say "This is great. I learned a lot" or praise some of the work that they've done... Because I think it could be easy to just not get that type of feedback as often, because we're kind of expected to do that kind of work. + +**Jeff Hernandez:** Being the least senior person on the team, it's difficult not to feel intimidated by other engineers when they ask for reviews... Because you're like "Oh, I'm early in my career. What do I have to offer to someone else who's been working for 10-15 years?" But I try not to think of that as much anymore; I try to think of it with my current understanding of things, I try to give the best feedback I can to that person... Because it's always nice to have a second pair of eyes, even if they're less experienced. It's always nice to have a fresh pair; they might see something that you've missed... And I take it as a learning opportunity at the same time for myself. + +**Natasha Dykes:** Definitely. + +**Natalie Pistunovich:** It's a fresh pair of eyes, but also that might be some very, very senior engineer who's been doing the same thing for many, many years, and therefore has gotten into a habit of doing things a certain way... And you, fresh bunny rabbit that you are, coming in with all the new technological lingo, and new open source -- like, you might be way more engaged and be much more on top of the new technologies and ways of doing things, so you might be able to come in and be like "Hey, have you considered this new style that you haven't done in 20 years?" Maybe it will be useful, and you can teach them something. + +**Jeff Hernandez:** Exactly. Usually, when I do "Oh, you should try it this way", I usually have links to support what I'm saying, just to be like "See, these other people are doing it this way as well. It's not just me." + +**Sarah Duncan:** \[35:52\] Yeah, and I think the value of a fresh perspective also is in challenging assumptions. I'm a staff engineer, but there are definitely things that, to your point, Angelica, I've gotten used to, or I've gotten into the habit of, and sometimes that does lead me to make assumptions, and I work to check myself. But having teammates being able to ask questions and check assumptions I think leads to really valuable conversation, because maybe that will lead us in a different direction, maybe I can explain more about something that I've done in my work... + +But I actually think when I'm approaching reviews and when I'm asking to be reviewed, I think of it more as like the subject matter expertise in a codebase, because I recently switched teams and I'm coming onto my new team and I have a lot of wealth of knowledge in terms of architecture, and some of the things that I'm bringing from my past team... But I'm actually learning some of these languages for the first time. So even though I have a big-picture idea around our architecture, and I'm doing a lot of things as the tech lead for my team... So my teammates actually know these languages, and the language patterns that come with these repos better than I do... So I actively look for them to be critical of the code that I'm putting forth... Because that allows me to learn the same way that if they were putting in a PR in Python, and I'm more experienced in Python, I would wanna teach them some of the things that I have picked up about Python along the way. + +So I actually think this is an area where I would hope leveling doesn't matter. It's more about the subject matter expertise in a repo, and helping that person level up their expertise a little bit further. And anybody can help anybody else level up their expertise. + +**Angelica Hill:** It's interesting to hear for me all the answers. I guess you're all U.S-based, and you mostly work with American colleagues... Is my assumption correct, or would you say your teams are also -- or you get to work, let's say, with people who are not Americans? + +**Sarah Duncan:** Yeah, I've worked with non-native English speakers before. On my current team I'm working with mostly native English -- all the engineers are native English speakers, but one of my teammates even was recently working with some of our engineering teams that are working out of other countries as well. So we have some international work... + +**Angelica Hill:** Would you say your experience of feeling comfortable to correct and to be corrected is the same as with people from the same background as you? Or is it different in any way when you review the code, or get reviewed by people, wherever they're based? Like, who just grew up in other places. + +**Sarah Duncan:** Yeah, that's a really great question. I definitely think that I have an easier time giving grace to somebody who's giving me a review if English isn't their first language... Because tone is a hard thing to pick up. I've worked with -- whether English is not their first language, or whether somebody has issues with tone for some other reason, like some kind of mental illness - or disability is probably the better term - for thinking through how something is gonna be perceived. I've worked with colleagues who are on the spectrum, and that's not always something that they have an easy time interpreting and figuring out how tone is gonna come across... + +So in those situations it's easier for me to give empathy to that person when they're giving me a review... Being like, "Oh, they didn't mean to hurt my feelings. They're just giving me feedback." But sometimes, because we are a big company, sometimes you are getting a review from somebody you don't know, so you don't always know if that's the case... + +I think it's easier when you're working with somebody you know. If English isn't their first language, I always have an easier time being like "Oh, they didn't mean it in a way that hurts my feelings." \[laughs\] Sometimes that's harder for me to get to that point; if it's somebody who isn't a native English speaker and it's one of those things I think is easy to make assumptions around, which is why I brought up working with somebody who is neurodivergent, it's just like... That's not always something you can tell right away. It's easier to tell English isn't somebody's first language, but I try to take that empathy that I learned working with those colleagues and bring it to all my other colleagues as well. + +**Angelica Hill:** \[40:25\] It's also interesting to think of -- I mean, yes, English is not the native language, so that's a very good differentiator... And also, different cultures have different relationships with feedback, or even saying to somebody "You did something wrong." + +**Sarah Duncan:** Yeah. That was a good point. + +**Natalie Pistunovich:** Being the only person who is not in the U.S, I would actually love to hear your perception on this. I know you keep on saying "We need to hear from the guests", and I agree, these wonderful guests... But I would love to hear your experience. I mean, I got culture shock moving to the U.S, but I wasn't reviewing PRs in London. + +**Angelica Hill:** And you are a native speaker. + +**Natalie Pistunovich:** Yes, and I am a native speaker. So I have those two things which make it slightly easier coming in... But I would love to hear your perception, and kind of what you've heard from colleagues, friends, in your experience. + +**Angelica Hill:** Almost all the teams that I worked in are quite mixed, because Europe is a lot easier for work relocations. The Green Card and so on exists in the U.S, but it's quite harder than just the work visa in Germany or other European countries... So I don't remember the last time I worked in a team where everybody comes from the same country, or we had more than 10% native English speakers. Although English is always the main language, because you need to have something that is not the programming language... + +And I think in the beginning I used to have some kind of misunderstandings exactly, because what was said, understanding empathy, understanding they did not mean that, they meant this... Coming across myself sometimes as a bit more "Why don't you say that in a nicer way?" and so on. + +So definitely, there's all sorts of balance there to strike, and it's very interesting to hear how people cope with that, and it's obviously written is even less easy to understand, like written communication in comparison to spoken communication. And even in spoken you can easily get lost, and so on, and intonations mean different things... + +The thing about the culture is let's say that sometimes it's kind of impolite either to correct, or even to just say no to someone. So I had to learn and actively practice that when I propose some idea, I have to start with something like "Feel free to reject that", or "There might be better alternatives, but..." And sometimes still, depending on the context, it sounds sort of foreign, but I think it's fine that it sounds foreign, because it's kind of like a standard across all the different cultures you'll get to work with, that this is "Be nicer" over being not nice enough. + +**Sarah Duncan:** Yeah, definitely. + +**Natalie Pistunovich:** My comment is gonna be slightly less serious, but it came to mind and it sounded like something that I would love to do... Can you attach voice notes to PRs? + +**Angelica Hill:** You have Looms in Git pull requests... + +**Natalie Pistunovich:** Yeah. Because I feel like I would love it if someone made a suggestion and then I could have a little voice note where they could say "This is why..." Because then you'd hear their tone, and you'd hear them explaining it, and then you could do like a verbal readout of your very long PR... That might be cool. + +Do emojis help with that? I find it really difficult to communicate through my words without emojis. I use emojis to help me more accurately portray the intention behind my comments. I.e. when I say something like "Oh, I'm not sure about this", I'll do like a thinking emoji, and like a funny, like a tongue emoji to show that it's not like a "What is this?" Because I feel like words are not enough, which is why - and those of you who interact with me regularly, I think it'll be very hard for you to find any space where I haven't put an emoji after my message, or some kind of imagery, whether it be a meme or a GIF, to try and level up my communication from just being words, to having that extra layer of emotion. + +Is it appropriate to add many emojis to your PR? I know you can add the emoji reactions, but within it, is that a useful tool that in fact can be used to try and alleviate that risk of being misconstrued via just written word? + +**Angelica Hill:** \[44:37\] You mean the pull request description, or actually each commit? + +**Natalie Pistunovich:** I'm open to views on either. + +**Sarah Duncan:** Yeah, I recommend using emojis, and setting up norms around emojis in code review norms. It's one of the suggestions that some of the other staff engineers and I put together in the norms template that we shared out internally at the times... And I think the way that we use emojis in that template is around communicating the intention behind a review comment. + +One of the things that's really hard to tell based on just the written word in terms of tone is whether that piece of feedback is blocking or not. So is this just a stylistic thing that you think will help me level up my skills? Is this something that I actually need to do before merging in this code? That kind of communication can sometimes be hard, and especially I think within different levels. I know when I was more entry-level I had a hard time just coming out and asking "Do I actually have to do that though?" So we use emojis as a way of kind of categorizing the comment. So if it is blocking, you can communicate that, being like "Oh, this will cause implications in this way", and it's a good prompt for the reviewer to think about why it's blocking, and communicate that. But it's also ways of sharing stylistic feedback in a way that's not blocking, and saying "Oh, I see that you did it this way. Typically, when I approach this problem, I do it this other way." It's just different styles. Sharing my style... + +So I think it helps to categorize the intention, to your point, Angelica, around what the comment is supposed to be communicating. It adds a little bit of color, I think, to just the plain text otherwise. But I'm curious if others have different opinions on it. Maybe you find them annoying... \[laughs\] + +**Jeff Hernandez:** Follow-up to that... So I'm 100% pro emoji, as people on my team can attest to... I'm always using emojis in Slack, and sometimes in GitHub. But I guess if it's kind of meant to convey blocking versus non-blocking, as per your example - is that like a way to kind of not have to use the actual supported GitHub features? Like, you're requesting changes and kind of blocking the merge, like full-stop? Because I don't know how people perceive that, but maybe it could come off a little harsh actually using that feature, and maybe the emojis is a nicer way to say that... But if it's really blocking, shouldn't we just be using that specific feature to prevent it from actually going through? + +**Sarah Duncan:** I think in that scenario you still do request changes, and kind of formally block the pull request... But I know as a reviewer if I get -- like, a change is requested on my PR, and I go in and there are like 20 comments, it's hard for me to maybe initially sift through those comments to figure out which ones are the ones that are causing the PR to be blocked. So it helps streamline the communication in that way, where if I'm blocking Angelica's pull request and I leave a bunch of comments, and some of them are like "Great job! This is really cool!" and some of them are style, and some of them are actually questions, and then actually some of them are actual blocking requests for changes, it helps narrow to those, and create a clearer checklist, in a way, of what you're expecting to be changed before being asked to re-review the code. + +So I just think it's helpful, even if you are formally blocking the pull request, to communicate what things you expect to be changed before you think that it could be merged. + +**Jeff Hernandez:** Gotcha. I don't know if I was just projecting my own fears saying that... \[laughter\] Like the email where it says "Changes have been requested", and "Oh no, what did I do wrong?" I need to do it. + +**Sarah Duncan:** Don't worry, we all do stuff wrong in our pull requests all the time. It is part of being an engineer. \[laughs\] + +**Break:** \[48:49\] + +**Natalie Pistunovich:** So I just have one more question, because we are running out of time, regrettably... But my question is "Can you teach people how to do a good PR, how to review PRs?" Is there truly an art to reviewing a PR, or is it really just us all bumbling through, learning as we go, trying to do the best we can, and really, you're never the kind of PR pro; you're always gonna have more to learn about how to create a PR better, review a PR better. Are you ever gonna be like "Right, I'm done. I am the PR pro. Amazing"? + +**Jeff Hernandez:** I think there is an art to it... You're kind of balancing the empathy of giving this person feedback, and acknowledging how much work they've put into it, but at the same time you have to -- it feels like a critical service; you're going to be the one that might be paged during the middle of the night if something goes wrong, and "Oh, this commit is what's causing the issue." + +But also, you kind of have to balance that with the time. You still have to get things done. It's a big balancing act of how much -- like, is this critical that this needs to get changed, even though it's like a stylistic thing that goes against our team norms, but it's imperative that this gets through?" Or maybe they've already spent a lot of time on it, so it's definitely the art of juggling or balancing, so... + +**Natasha Dykes:** \[54:14\] I think just experience goes a long way. Learn by doing. You can kind of experience a lot of things that you wanna emulate, or things that you're like "Wow, that didn't leave me feeling really good about myself, so I know not to do this in the future." But also, just getting more experienced in whatever language that you're working on to then make those suggestions that can make the code better, and more performant. That's like another layer of doing a PR review... But it's still something that comes with more time. + +**Sarah Duncan:** Yeah, and I think from the perspective of authoring a PR and asking for a review, that's also something that is like a skill that can be honed and taught and improved at any level... Because figuring out when a PR is done, how big it is, but also in some situations maybe you're introducing a change, and maybe you have to write up why you think that change is the right path to move forward, and there's always room to improve on making a concise argument. I think that's something that we can always improve on, is how to make a concise, clear argument for the change that you're introducing, or how to best walk your reviewer through the pull request just based on your PR ask. There's a lot of skills that can be honed in this whole process, on both sides. + +**Natalie Pistunovich:** Well, thank you. Regrettably, we've babbled all our time away, so we are going to move into our Unpopular Opinion time. + +**Jingle:** \[55:49\] to \[56:04\] + +**Natalie Pistunovich:** Right, so get ready, guests, we're gonna jump right on in. Jeff, what is your unpopular opinion? + +**Jeff Hernandez:** I don't know how I'm gonna top my last unpopular opinion, but... I don't know, I feel like I'll just spark a -- it's not a big unpopular opinion, but it'll start like a war... Kind of like Playstation vs. Xbox. I'm gonna go with "Dogs are the better pet than cats." I know there's lots of cat people... + +**Sarah Duncan:** Whow... \[laughter\] + +**Jeff Hernandez:** So it'll be the unpopular opinion in this group, but... I have three dogs, and... + +**Natasha Dykes:** Okay... Those are fighting words. + +**Sarah Duncan:** \[unintelligible 00:56:38.03\] + +**Natasha Dykes:** Just put it in your next PR. + +**Jeff Hernandez:** \[laughs\] Exactly. They're fighting words, you know? + +**Natalie Pistunovich:** I've got your back, Jeff. I'm team dog. + +**Jeff Hernandez:** Team dog, all the way. Yeah. + +**Natalie Pistunovich:** Sarah, what is your unpopular opinion? + +**Sarah Duncan:** Yeah, I'll be interested to see how unpopular this is or not, but I think aspiring software engineers would be better off taking more writing and philosophy courses, and fewer computer science theory courses. + +**Natalie Pistunovich:** Yes... + +**Natasha Dykes:** Interesting. + +**Natalie Pistunovich:** No, that's intriguing. Why is that? Can we dig one level deeper? + +**Sarah Duncan:** Yeah, I think a lot about being a successful software engineer is kind of this ability to be able to make a concise argument, be able to understand other arguments and perspectives, and incorporate it, and use that to kind of revise your opinion and put forth another concise argument. We see this in architecture documents all the time; I was on the architecture review board at the Times for a long time, and I chaired it for a while... I know Natasha has been on the contributor board as well, and we see so many long, long documents that could be half the size they are, and I think a lot of engineers don't use the theory classes that they took in college, or -- assuming that if you are studying computer science in college, I think a lot of those classes are not put to use as a practical software engineer... But the area where I see a lot of software engineers kind of having to improve on the job is in their ability to make a concise argument. + +\[58:16\] My sister is getting her Ph.D. in English, so I'm maybe a little biased, because I see how much work she has put into this skill... And I myself was -- my major was in the philosophy department, even though it was a logic and computer science interdisciplinary degree... So I use the skills that I got from my humanities classes more than a number of the computer science theory classes that I had to take in college. + +**Natalie Pistunovich:** Hm. Okay. Jeff... I didn't know you renamed yourself Natasha, Jeff... \[laughter\] + +**Jeff Hernandez:** I was gonna say, it sounds like you did a Go Time episode debating the value of a computer science degree vs. a dedicated software engineering... Because they're two different fields, essentially, right? + +**Sarah Duncan:** Invite me back for that one... + +**Natalie Pistunovich:** Shameless plug - I think we did one that was beating around the bush of that. I'd say a few months - maybe it was years ago - ago. Me and Kris Brandow, who's another Go Time host. We did one around English literature and its value to software engineering... But I think that was a while ago, so absolutely, let's do another one. But again, certainly not least - Natasha, what is your unpopular opinion? + +**Natasha Dykes:** My unpopular opinion I think might be accepted in a computer science, like this kind of circle, but who knows...? I feel like the world doesn't need another superhero movie. There's too many of them already. I'm done with it. There's so many other stories we could tell... So yeah, \[unintelligible 00:59:39.06\] At least take a break for a couple of years. + +**Natalie Pistunovich:** There are so many more bugs that need to be superheroes. We've gotta have Beetleman, Ladybird Lady, Catterpillar Kazam! There's so many opportunities. + +**Natasha Dykes:** They could do something with the gopher... Like Gopherman, or something... + +**Natalie Pistunovich:** Gopherwoman. + +**Natasha Dykes:** Okay, yes. + +**Angelica Hill:** Sarah, I loved what you said. You gave me an idea to see my -- next time I have some free time to hack, I'm gonna use something like GPT-3 to just create a plugin for GitHub that will just run whatever you wanna say through that, to sound more empathetic and more... Something. I don't know. + +**Sarah Duncan:** I love that. + +**Angelica Hill:** By the time this episode is out, maybe it will be on the marketplace. + +**Sarah Duncan:** Incredible. + +**Angelica Hill:** This will be really, really, really useful for everyone. + +**Sarah Duncan:** Yeah, send it my way. + +**Angelica Hill:** \[laughs\] + +**Natalie Pistunovich:** Action item for everyone listening... Make it happen. C'mon. Put our heads together, we can make that happen. + +Well, it has been an absolute joy talking to you all, genuinely. I can't wait to get you all back; I have so many more episode ideas... I'm sure you do too, Natalie... I hope you have a wonderful rest of your days, but for now, adios, and goodbye. diff --git a/The art of the PR Part 2_transcript.txt b/The art of the PR Part 2_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..7575f2bc304271eb6f8048c4ec193dd30edfc50d --- /dev/null +++ b/The art of the PR Part 2_transcript.txt @@ -0,0 +1,627 @@ +**Natalie Pistunovich:** Good part of the day to everyone, wherever yu're joining from. Here's Angelica and I, we are back to talk about pull requests. Hi, Angelica. How are you doing? + +**Angelica Hill:** I'm very, very well, thank you. I didn't think that PRs would take two episodes, but I'm surprised and excited that they will. + +**Natalie Pistunovich:** As everything about pull requests, it always takes a bit longer than expected, huh? \[laughter\] + +**Anderson Queiroz:** That's true. + +**Natalie Pistunovich:** Today we are joined by Anderson. + +**Anderson Queiroz:** Hello. + +**Natalie Pistunovich:** Hi. How are you doing, Anderson? + +**Anderson Queiroz:** I am really good, and really happy to be here. Thank you. + +**Natalie Pistunovich:** And you're joining us from the U.K. + +**Anderson Queiroz:** Exactly. + +**Natalie Pistunovich:** What are you doing in the U.K.? + +**Anderson Queiroz:** I'm at the GopherCon U.K. I'm a Brazilian that's based in Berlin, but now I'm here in London, directly from a hotel room. + +**Angelica Hill:** How are you liking it? + +**Anderson Queiroz:** Oh, it's really good. + +**Angelica Hill:** Yeah. And you're not just saying that because I'm on the call... + +**Anderson Queiroz:** No, no. + +**Angelica Hill:** Okay. + +**Anderson Queiroz:** I love the U.K. I did an exchange program here for one year and a half in Glasgow. + +**Angelica Hill:** Oh, awesome. + +**Anderson Queiroz:** So yeah... I really like it. + +**Angelica Hill:** I mean, I prefer Edinburgh to Glasgow - unpopular opinion... + +**Anderson Queiroz:** \[03:59\] Yeah, true... + +**Angelica Hill:** I mean, Edinburgh Castle is so incredible. + +**Natalie Pistunovich:** Wait, no, that's the end of the show! + +**Angelica Hill:** Oh, I'm sorry. I'm getting ahead of myself, Natalie. I'm sorry, I'm sorry! \[laughter\] + +**Natalie Pistunovich:** It's fun to go back to in-person conferences, hm? + +**Anderson Queiroz:** Yeah, that's true. It's really good. The last time I was here in London, my brother was here. We were like, "Okay, the place here is gonna be small to GopherCon U.K. at some point", but then the pandemic, and now I think it's smaller... So yeah, let's see. + +**Angelica Hill:** Okay... + +**Natalie Pistunovich:** So Anderson, tell us about yourself. You're doing Go. + +**Anderson Queiroz:** Yeah, I'm doing Go since a bit before going to Berlin, so I think about five years now... I did before many Java; it was a lot of Java. I started with C, actually, some Python, JavaScript in the backend, and now Go. I love it, and I chose it as my language to specialize. + +**Natalie Pistunovich:** And you're working at Elastic. + +**Anderson Queiroz:** Exactly. I work with Elastic Agent. + +**Natalie Pistunovich:** Which is the head product for all the products that we know and love, like Elasticsearch. + +**Anderson Queiroz:** Yeah, exactly. Yeah. + +**Natalie Pistunovich:** Formerly known as the ELK Stack. + +**Anderson Queiroz:** Yeah. Now the Elastic Stack, yeah. + +**Natalie Pistunovich:** Lots of Go there. I love this stack + +**Anderson Queiroz:** There is, yeah. + +**Natalie Pistunovich:** Fun. So as part of the job you do pull requests. + +**Anderson Queiroz:** Of course. + +**Natalie Pistunovich:** Do you also do merge requests? + +**Anderson Queiroz:** We call everything pull request, yeah. + +**Natalie Pistunovich:** \[laughs\] + +**Angelica Hill:** Are you gonna do the same thing you did last time, when you caught us all off-guard with the many different -- + +**Natalie Pistunovich:** No, no, no... Just for those... + +**Angelica Hill:** ...many different ways to refer to it... \[laughs\] + +**Natalie Pistunovich:** It's crazy how many names are there for this... Yeah. For those who are listening now and have not listened to the previous episode, "The Art of the PR: Part 1", Angelica and I were discussing, among other things, also the many different names and concepts that represents. So Anderson, we asked you if you're doing lots of pull requests, but then we started talking about other things... Do you? + +**Anderson Queiroz:** I do. + +**Natalie Pistunovich:** Do you review more, or do you write more? + +**Anderson Queiroz:** Now I write more. In my past job I read a lot more code. A lot of pull requests. And funny enough, I've never worked in a company that didn't have pull requests. To me, software development as a professional means pull requests and code review. That's my standard. + +**Angelica Hill:** Why did you do more reading in your past job? Is it just a very different area you were working in? + +**Anderson Queiroz:** No, my past company I joined among other things to help to lead the transition to Go. So they were pretty much a Ruby shop, and they decided to migrate to Go. They were migrating, but then needed someone with expertise in Go, to bring best practices, how to do... So I did that. + +I did a lot of workshops, and teaching and mentoring, and it involved -- a lot of the teams would come to me with pull requests, so I'd do a really extensive review, not only code \[unintelligible 00:06:49.25\] functionality, but also as an opportunity to teach Go and \[unintelligible 00:06:52.19\] the conventions, the best practices. + +**Angelica Hill:** And then now you've moved to somewhere where Go is bigger, so therefore you don't have to play such a big kind of a reviewer role. + +**Anderson Queiroz:** Exactly. Now I am a software engineer, as a senior, but there I was a tech lead. I think when you go above senior in the tech lead and stuff you start coding less, reading more, writing more... Right? + +**Angelica Hill:** Okay... + +**Anderson Queiroz:** I mean, writing more specifications and documentations. + +**Angelica Hill:** So do you feel like there's a direct correlation between seniority and how much reading of PR versus writing you do? Like, if we were to plot that on a graph, how much you read PRs and how much you write... Could we, to generalize the industry, say that the more senior you are, the less you're gonna be writing PRs and the more you're gonna be reviewing them? + +**Anderson Queiroz:** I think the more you are senior, you're gonna look for more things. So because you're gonna be able to evaluate -- if you're good in the language, you can evaluate the language and the conventions, you can evaluate the general software architecture. When you're in a role that as a senior you are also a mentor, there's a lot of meeting juniors in the team; I think you are reading a lot, because you have this responsibility of the team, of the product. But if you're in a company where everyone's senior-senior... In Berlin definitely the titles are inflated. So you know, perhaps everyone doesn't need this level of attention in mentoring or teaching; that happens a lot through PRs. + +**Natalie Pistunovich:** \[08:21\] That's an interesting question, yeah... Because sometimes in Go they have this graph of contributions that is no longer just the squares, but also what type to bring more - to you make more issues, do you review, and so on. + +**Anderson Queiroz:** Oh, yeah...! + +**Natalie Pistunovich:** So if you can look in that, then you kind of know what people's roles are, or you see how it changes over time, or something. + +**Angelica Hill:** If I showed you a load of graphs, would you be able to guess the seniority of that engineer? + +**Anderson Queiroz:** Perhaps... I think I looked quickly at mine and now it's 50%. + +**Natalie Pistunovich:** Nice. + +**Anderson Queiroz:** Pretty much. + +**Angelica Hill:** Do you think that that's the balance that most engineers would like, or do you think that there is such thing as someone who prefers to be reviewing more than writing? + +**Anderson Queiroz:** When you think as a software engineer that the patch is about writing, writing... Right? They write code... I think at least either you want or you need to pass knowledge forward. So you're gonna need to read. I really like to review code, for both reasons. One... Because, I mean, I have a passion for Go, and I am a quite methodic person, so I like to ensure that the code is good, the conventions are there... I'm super-picky about proper error handling... So this is something that -- if you're not handling the errors properly, or you're really bad at the conventions, I'm gonna be commenting that. But I have special commenting to ensure that I'm not overwhelming the other person, or to feel like I'm just complaining, saying that your job is bad. + +**Angelica Hill:** So you say you try to give feedback, you don't wanna overwhelm the person... How do you do that? Is it that you limit yourself to "Okay, I'm only gonna put six comments"? Is it the way in which you phrase your review? How do you make sure that you're not overwhelming? + +**Anderson Queiroz:** Yeah, I learned to experience in feedback. I put tags. So I start a PR with suggestions, or sometimes I put questions... But a question is a question, right? So it feels a bit redundant, but I put. And then I have blocker... And sometimes the suggestion is less Go convention. Or depending on the wrapper, blocker \[unintelligible 00:10:25.29\] Go convention. So I try to categorize in suggestion, question, and blocker. I put blocker \[unintelligible 00:10:32.20\] So basically, the blocker is either I see there is a problem in the code, so it needs a change, or I believe there is a problem, or I don't believe that this implementation is good... So at least you need to answer that. You can say \[unintelligible 00:10:47.28\] that's fine, but I need an explanation. So these are the blockers. + +Suggestion is exactly that - I believe it can be better. If I were writing this code, I would do it different, but you don't have to take it. Sometimes I put like init. It could be just "Erase this blank line between the function call and the error handling." This is init. Or a typo... So these things. And then if I've never reviewed a PR from this person, or perhaps I see there is a lot of comments, I go to the person, or even in the PR, I put "Look, there is these three categories. The only thing that's really important are the blockers. So the blockers, please comment or change. The suggestions are suggestions, and as always, you are free to \[unintelligible 00:11:33.11\] 100%. Just answer the blockers and talk to me." That's how I do. + +**Natalie Pistunovich:** So you say that in some situations it kind of makes sense to just speak in person instead of writing a lot, for example. + +**Anderson Queiroz:** Sometimes it's easier. I think sometimes speaking is better, sometimes writing is better... And on pull requests, writing on the pull request is a lot slower. Sometimes it's worth jumping on a Slack chat; it's already enough. Sometimes we talk. + +I think nowadays at Elastic, because we are distributed, it's a lot more through Slack, rather than really a call... But sometimes - yeah, I've jumped on a call, for small things. + +**Natalie Pistunovich:** \[12:18\] And how do you decide when it's better to do this and when it's better to not? Do you have a thumb rule, or is it all just feeling? + +**Anderson Queiroz:** If there is a lot of back-and-forth, it's easier to jump on a call. Or if it's something that I really want to understand... But most of the times I always \[unintelligible 00:12:32.13\] on the chat. Now, Elastic is a distributed company, so we write a lot more. So it's a lot more common just to write than jump on a call. + +I think jumping on a call is more personal, so if I'm close to someone, it's easier. It feels more natural to jump on a call. And if you don't know the person so well, or something, you kind of end up just chatting. + +**Natalie Pistunovich:** So you say there's actually three types of giving feedback. One, writing in the pull request, two, writing on Slack, and three, hopping on a call. + +**Anderson Queiroz:** Yeah. On Slack - I use it more to clarify. Because if it's specific about the code, I'd rather have it in the pull request. I kind of document it anyway, right? I try also to explain why what I'm saying... As I said, in my past company I was really with this job to teach, to mentor in Go; I would link a lot of references. Also because I learned in another company -- I remember my colleague and I, we were reviewing a pull request, of a new joiner, and I would say "Oh yeah, this function is too long. This function is too complex." This happens. And the guy was like "Why?!" + +**Natalie Pistunovich:** "Why did you butcher my pull request?" + +**Anderson Queiroz:** "You are guessing. This is your opinion." And I was like, "That's true. We don't have like a metric to say that." But also it was common sense. It was like "This is too big", you know, in the company. But this guy was a new-joiner, so everything was different for him. + +So again, clarifications - I can do also in the pull request. But I think it's important to document what's happening in the code, there. + +**Natalie Pistunovich:** In the code itself, to kind of document, to make it self-explanatory. + +**Anderson Queiroz:** Exactly. + +**Angelica Hill:** So in terms of -- you talked about giving feedback, not overwhelming... Do you feel like PRs are also a good place to kind of -- especially for more junior engineers, give them like props on things they're doing well? Like, "Oh, I really like the way you did this thing." Or "Oh, this is great." Or "This function is structured really well." Do you feel like PRs are also a way to give positive feedback? + +**Anderson Queiroz:** Yeah. I think it is. It is important, and it's something that I would like to do more as well. I mean, I never go to a pull request looking for errors or to try to diminish someone's job. But at the same time, I go to a pull request looking for errors, so they don't go to production. + +At the end of the day, you're looking for problems in issues to prevent things bad from going to production. But I think it's super-important to do the praise, to assert when someone does something nice. Sometimes someone just fixes something a bit random, but it's the same function. It's not really going away from the aim of the pull request, and this is really good, to praise. At least a thumbs up. + +**Natalie Pistunovich:** So we said that there's some stages that you escalate the communication through some way, and definitely you want to include more positive feedback, that is not "Correct this", and so on. Or explain yourself. What other changes would you make to the PR process based on pain points you have with the flow? + +**Anderson Queiroz:** It's a really slow process, because you write the code and then you submit the PR. And then someone else - sometimes more than one person - they have to stop and read it. And how do you synchronize that? And then I think it's gonna depend a lot how companies do... I've been working for different processes. Some processes, they ensure that the pull request will be reviewed, some not so often... + +\[16:04\] Now I'm definitely overwhelmed by GitHub notifications, so sometimes it just slips through, some PRs, and days later someone tags me, "Anderson, can you give a review?" Like, "Oh my God, sorry." + +**Natalie Pistunovich:** Because it's something that -- kind of a bit of a ping pong, or does it happen more with new pull request? + +**Anderson Queiroz:** It happens when you have to re-review. You do the first review, and then you have to look again. + +**Natalie Pistunovich:** Yeah. + +**Anderson Queiroz:** So to get the proper ping, it's hard. And the other thing that -- at least on GitHub... I've used a bit of BitBucket as well. I've submitted one, perhaps two PRs to the Go wrapper, with Gerrit. But on GitHub it's really hard, when you do the first review, you're asking for changes, and then people do changes, to see exactly what changed. To be like "Okay, I'm gonna review just the changes." + +**Natalie Pistunovich:** Isn't this a new commit? + +**Anderson Queiroz:** If someone just forced -- + +**Natalie Pistunovich:** Oh, forced push. Okay. + +**Anderson Queiroz:** If someone's a forced push... Right? So sometimes it's hard to get just the bit that I have to re-review. That's a pain point. + +**Natalie Pistunovich:** Yeah, I had a Twitter poll the other day on "What do you do when the pull request stretches so much, and there's one million comments that lead to one million commits? Do you squash that or not?" And I think the answer that I liked most was "During the pull request have as many commits as needed, and ones that's good to go merge, then squash it all into readable, one or two commits, or however many logically needed..." But exactly to allow the proper review, like you say, as many commits as needed. + +**Anderson Queiroz:** Yeah. + +**Natalie Pistunovich:** Yeah, definitely a good practice. + +**Angelica Hill:** So you kind of chatted a little bit about your love of Go... And this is a Go podcast, but I would love to hear a little bit about your approach to PRs in other languages, i.e. does the way that you review PRs differ depending on the language you're reviewing? + +**Anderson Queiroz:** I think they do differ, most because of the expertise, right? I'm quite comfortable in Go to understand the language and understand what's happening, to know some caveats, some catches, and also to talk about conventions. In other languages, not so much. So I guess in other languages I am gonna focus more on general software architecture, because this is \[unintelligible 00:18:16.16\] and functionality... And of course, try to use as much as I know from the language convention. + +I think that's another thing... Pull requests are great, great to learn language conventions. I've learned so much about conventions in Go and the other languages through pull requests... And then if I'm not so expert, I'm gonna try to put and point the best that I know... But I know that I'm not the expert. And a lot of the times it's probably not really my wrapper, so I'm not there to enforce anything. So they're gonna be more on the suggestion side. + +**Angelica Hill:** I mean, taking kind of one step more granular, are there things that are more important for Go when you're reviewing, i.e. stylistic choices, almost principles that you might adhere to more closely than you would in other languages? + +**Anderson Queiroz:** I mean, Go is opinionated. You have to `go fmt`. You have to format the code properly. Even though we still have some space to discuss about how to format... But I think that's the first thing. I usually would like to be super-strict about how the imports are sorted, but I am not. If the wrapper is consistent, I think it's a lot easier to enforce this thing. If it isn't, not so much. And in other languages they don't have so much, so it's gonna be more about team conventions, rather than the language. And in Go you get a lot from the language. + +I haven't seen so much, but because Go is so focused, let's say, in concurrency, sometimes people try to either sneak in concurrency when they shouldn't, or they are not using the right tools... Also because, you know, "Oh, concurrency! Let's use channels!" No. Channels are for something. A mutex - they are for other things. Waitgroups are for other things, right? + +\[20:08\] So this is another thing that I would say, "Okay, no. Perhaps we can do different, or we can do better." Or "This is too complex to understand if you use \[unintelligible 00:20:16.17\] If you remove this channel and put a waitgroup, it's a lot easier. \[unintelligible 00:20:20.02\] Channels - probably not. So I think of these things. + +**Angelica Hill:** When there are new features and new things released in Go, do we see an uptick in people using those in PRs? Like, you just get over-excited, like "Oh, generics...! Every PR now has generics." \[laughs\] + +**Anderson Queiroz:** I always wanted to push the new things and use the new, as soon as possible. I think in general if you're working with microservices, it's a lot easier, because you can just update the version and redeploy, and even if something breaks, you can roll it back a lot easier. Now at Elastic \[unintelligible 00:20:59.00\] distributed bineries that is gonna go to I don't know how many clients in the whole world... So we have a release cycle, so we have to choose "Okay, let's change the version." + +We have several wrappers that use Go, and you try to keep everyone in the same version. So it's a slower process. But as much as I can, and as much as I know what's coming up, I try to incorporate, if I can. + +On that topic, have you folks started to use any, instead of the empty interface? + +**Natalie Pistunovich:** \[laughs\] It's an interesting poll to write. I don't know, I think I'd phrase that as an unpopular opinion. "Don't use that", or something. It can be one more unpopular opinion for your stash... + +**Angelica Hill:** Your library that we're soon gonna have... + +**Anderson Queiroz:** I think I saw the first use of it today, in the workshop with Bill Kennedy. His code had any, and I was like "Oh, yay! We can use any now." + +**Angelica Hill:** What was the use case that he used? + +**Anderson Queiroz:** It was a map for a logger... + +**Angelica Hill:** Okay. + +**Anderson Queiroz:** The map string empty interface. I think it was in a logger, or something. Or no, I think he was parsing at JSON. So instead of map string empty interface, it was map string any. + +**Angelica Hill:** Okay. Well, we can do the poll and then tag Bill and be like "Please tell us. Give us the example so we can all understand how you use this." + +**Anderson Queiroz:** Yeah... He was asking everyone to have Go 1.18 because of that. He was like "Okay, you're gonna use any, so please..." + +**Break:** \[22:27\] + +**Natalie Pistunovich:** What do you do when you have very large pull requests? Lots of files, lots of commits, lots of lines. + +**Anderson Queiroz:** I sit and cry. + +**Natalie Pistunovich:** How do you get on top of that? + +**Anderson Queiroz:** No, that's a tough one... I try to review at once. Sometimes it's not possible. I think there's no magic, you just have to go through it. + +**Natalie Pistunovich:** Do you review everything on a high level? You know, see the list of the commits, if they tell some story, or maybe look at the list of files, or do you just dive into the first one, and one by one, until it starts making sense? + +**Anderson Queiroz:** I never look at the commits. I don't know if because when I'm coding and committing -- like, I'm gonna squash everything before morning. First things first. So to me, the commits themselves - they don't matter so much. I try to put in a way - if I need to revert something, I do. But at the end of the day, there's a good chance that I'm gonna just do one commit. So I never look at the -- neither the wrapper commit history. Only if I need to understand why it happened. But in a pull request, I never look at the commit history. I just look at the diff, and it's always on GitHub. + +**Natalie Pistunovich:** And look at the files by the name, basically. + +**Anderson Queiroz:** The big ones I just go click, and "I've seen this file. I've seen this file." + +**Natalie Pistunovich:** Hm... So just by the order of appearance. Because sometimes it's not always the correct flow, kind of. + +**Anderson Queiroz:** Yeah. If it's hard to understand, I get the code, I check out the feature branch, and I go to see... Because I also sometimes... You know, I'm also going on, you wanna jump, you wanna understand how it was called to do something, and then it sits there on an IDE and you have the code. And also if I wanna suggest a change... + +So either it's something really simple and I'm 100% sure that it works, or I'm gonna probably write it in the code itself, and I might test it, to not suggest something that is broken. + +**Natalie Pistunovich:** Maybe even a few steps back... When you go to review a pull request, do you start by reading the issue? So the first thing you do is read the issue. + +**Anderson Queiroz:** \[26:13\] Yeah. I have to understand what's happening there. + +**Natalie Pistunovich:** Then do you review the diff on GitHub, or in your IDE? + +**Anderson Queiroz:** No, on GitHub. + +**Natalie Pistunovich:** So you go kind of file by file on GitHub. + +**Anderson Queiroz:** Well you know... + +**Natalie Pistunovich:** \[laughs\] Philosophical questions... + +**Anderson Queiroz:** Good point. Because to me, that review -- actually, going back to the other episode... We're doing a code review. It's not the pull request so much, right? What's happening. + +**Natalie Pistunovich:** \[laughs\] + +**Anderson Queiroz:** You have to comment on it. So it's really hard to comment on it, at least for the tools that I use at GitHub and the IDE. It's hard to comment on the code. If it's something that for some reason -- you know when you do either a greenfield project, or the pull request is a huge refactor... You know, everything changed, so it's pretty much new code... So on that rare occasion, I might open the code itself, because then I can read in like an execution order, let's say... And then my comments - they are probably gonna be comments on the code itself. + +I've done it once or twice, and then I created my branch out of this branch, and then I open up pull requests for this branch, so the person can see my comments in the code, without having to go look for it. + +But this was pretty much in either greenfield projects - you know, when you start something new, or when you're just adding so much new code that the pull request itself, it's hard because it's completely out of order and everything, and there is a lot to comment on. + +**Natalie Pistunovich:** Yeah, that's the hardest ones, when there's so much to handle there. + +**Anderson Queiroz:** Yeah. + +**Natalie Pistunovich:** Do you sometimes find yourself re-reading the whole thing to kind of - once to read it, and second time to make sense? + +**Anderson Queiroz:** Yeah, definitely. Going back to an old Go Time, when you talked about documentation and reading documentation - I have the perseverance; if I'm not understanding, I'm gonna read it over and over and over again, and I'm gonna do my best to understand. If I don't, I'm gonna ask. But I think if something's puzzling me, then I'm probably gonna block and say "Look, I think this is important. I don't get it. Please explain." + +**Natalie Pistunovich:** Yeah. + +**Angelica Hill:** Do you comment as you go, or do you read through, fully digest, and then go through and do all your comments? + +**Anderson Queiroz:** No, no, I comment as I go, as a string, and then when I get down there, I'm like "Oh yeah, that's why!" So I go back and delete the comment. + +**Angelica Hill:** Oh, okay. + +**Anderson Queiroz:** You know, you're like "Dude, it doesn't make sense." Something like "Why?! How does it work?" And then you read it, "Ahh...!" And then I go back, delete and edit the comments. + +**Angelica Hill:** You go back, "Oh, I'm so sorry. Never mind..." + +**Anderson Queiroz:** I mean, on GitHub you don't submit the review, right? + +**Angelica Hill:** Yeah. + +**Anderson Queiroz:** So I just delete. But the feeling is exactly that I'm apologizing for asking something stupid. \[laughter\] + +**Angelica Hill:** When you're going over a PR, do you feel like - or any reviewer - should test the changes? And to what extent should you test the changes, if you think that they should be tested by the reviewed? + +**Anderson Queiroz:** I think that's the one million dollar question. + +**Angelica Hill:** Give us the answer, we're ready! + +**Natalie Pistunovich:** Maybe another unpopular opinion... + +**Anderson Queiroz:** No, I think the answer is it depends... + +**Angelica Hill:** Okay. + +**Anderson Queiroz:** We say that a lot at Elastic. But I think it depends on your role on this pull request. Imagine that I'm a maintainer of open source, and I'm taking something in from a person that's not from the wrapper, I'd be extra careful. So perhaps I would try to run the code. Usually, I don't really test the code. But again, if it's something that I feel that's really critical and I wanna be sure that it's 100% working, I might test it. But it's rare cases. + +**Angelica Hill:** Okay. And does the length of the PR or the scope of the change change that opinion? + +**Anderson Queiroz:** \[30:06\] Yeah... + +**Angelica Hill:** Okay. + +**Anderson Queiroz:** I think the length not so much... Let's be honest, the longer the PR, the less detailed is the review. We are humans, we get tired. Come on. If you're reviewing like 15-20 files, the last one - you're tired. It's just a human thing. And it's something that I've done a few times. I don't like it, but it happens. The longest PRs, I review, and then I submit, and then when I'm doing the re-review, I find new things; I'm like "Ahh... I can't let it pass. I'm gonna have to put a comment where there wasn't a comment, and it was not changed, because now I saw it." That's the thing... + +**Angelica Hill:** So given that, what is a reasonable time to expect a PR review? If you put in a PR today, is it the next hour, by end of day, a week? Does it depend? + +**Anderson Queiroz:** The real answer is "It depends." + +**Angelica Hill:** \[unintelligible 00:31:00.20\] + +**Anderson Queiroz:** Yeah. At my first job we -- did we use JIRA? I don't know, whatever; we used columns. We used columns, so there was the column "PR reviews." So the open PRs were there for review. And then there could be -- I think we were three people, so there could be only two PRs on review. So do you wanna put something for review, we have to take something to review. So this helped to keep the process running, everyone reviewing. Nowadays -- I think when you are really running a \[unintelligible 00:31:33.02\] a microservice you just deploy, you usually expect something in the next day to get an answer. Not Elastic... Within the week, I'd say. + +**Angelica Hill:** Okay. And do you have different commitments when it comes to your internal team PRs, versus people who are maybe contributing to your service? i.e. like in our system we have a lot of external teams that will contribute to our service and ask for PR reviews... What is a reasonable timeline to commit to review those external PRs? + +**Anderson Queiroz:** I think there are two categories. If it's just a normal flow, they can go in the same flow. But if it's something that someone external is doing, because our team doesn't have the capacity, and then it is really important, probably I will try to prioritize this review. + +But also, if someone that's not from the team, who doesn't know the wrapper conventions, it's probably gonna be a more thorough review. I strongly believe that at least your code should be consistent. I'd rather have something that I don't like, but it's consistent and it's always there, than half of the code I like, half of the code I don't like, and another third I don't even have an opinion. + +So in extended reviews I think there's the extra consistency in code conventions from the wrapper that you have to put through. And then it's better to be quicker. + +**Angelica Hill:** For sure. I really like that column policy. I might have to implement that on my team. + +**Anderson Queiroz:** Yeah, right?! + +**Angelica Hill:** You can't put a PRon if there are already two. You have to review them. I love that. + +**Anderson Queiroz:** Yeah, it makes things to move. And I think it's nice. + +**Angelica Hill:** I agree. + +**Natalie Pistunovich:** It's a waitgroup. We've pretty much described this concept now. + +**Anderson Queiroz:** It's a channel with a buffer. + +**Natalie Pistunovich:** \[laughs\] Yeah. + +**Angelica Hill:** And then I'm gonna have to be the bad girl who comes into Slack and someone's like "Oh, I'm ready to put my PR in this big new feature and I'm like "You're not allowed to..." + +**Natalie Pistunovich:** \[unintelligible 00:33:24.15\] + +**Angelica Hill:** "Go review Bob's PR." + +**Natalie Pistunovich:** Throwing error... + +**Angelica Hill:** \[laughs\] + +**Anderson Queiroz:** Yeah... There's like the poking PR review. You only get your PR reviewed when you poke someone. You don't want that. + +**Natalie Pistunovich:** Exception. \[laughs\] It could be a fun way of teaching all sorts of Go concepts, now that -- this gave me some ideas. Thank you. + +**Angelica Hill:** What, doing a PR review? + +**Anderson Queiroz:** By poking people? + +**Natalie Pistunovich:** Yeah. With limiting this. This is a fun way to discuss this. And on the way you discuss error throwing, and correct errors, and also exceptions, and so on. + +**Angelica Hill:** Okay... + +**Natalie Pistunovich:** Goroutines, if you suddenly have to split into that... Yeah. + +**Anderson Queiroz:** \[34:04\] Talking about teaching Go in unusual ways... I was thinking today someone should write a Go program that simulates how the queue for the food works here, and then make a proper Go concurrent, good program for that... Because the queues are necessary here. You have a lot of contingence. I was like "You know, you can make better concurrency here." + +**Natalie Pistunovich:** Because you have lots of food stations that people miss? + +**Anderson Queiroz:** That's the point, you have a lot of food stations. You can have a lot of concurrent access to that. But no, \[unintelligible 00:34:31.24\] A huge queue, and everyone goes to everything that they don't want. \[unintelligible 00:34:35.28\] + +**Natalie Pistunovich:** They didn't read the docs. They don't know what's the food. + +**Anderson Queiroz:** They didn't read the docs, exactly. + +**Natalie Pistunovich:** Everything can be explained with tech + +**Angelica Hill:** Lesson learned. Always read the docs first... And if they're bad docs, then... + +**Natalie Pistunovich:** Improve them. + +**Angelica Hill:** ...improve them! + +**Anderson Queiroz:** Open a pull request for the docs. + +**Angelica Hill:** Yes. \[laughs\] + +**Anderson Queiroz:** That's something that I love... If I'm reading documentation and it's easy to open a pull request, \[unintelligible 00:35:02.03\] fail or inconsistent something, I open a pull request. I think it's such a valuable contribution, and it's so easy most of the times... I love these docs that have the button Edit, and then you go direct to GitHub to create the pull request. That's fantastic. + +**Natalie Pistunovich:** Yeah. + +**Angelica Hill:** I feel like I get into a bit of a rabbit hole, and I have to stop myself editing documentation... Because it went from actually making it correct to actually just implying my personal stylistic choices when writing documentation, and phrasing... Like, "I like this adjective slightly better, actually..." + +**Anderson Queiroz:** Oh, yeah... + +**Angelica Hill:** So I had to pull back to be like "Okay, review for correctness, not for like "I want a comma here." + +**Anderson Queiroz:** I think that's so hard... For me, as a non-native speaker, sometimes I don't think this sentence is correct. I think it's missing a comma, I think it's missing an article. And it was like, "Honestly, you don't know English so much. I don't even know if you could do that in proper Portuguese, like proper grammar." But I think it's important. + +My take is if I believe it's compromising their understanding, I'm gonna probably suggest something. And also, sometimes when the comment is there for a long time, I just make that change and suggest someone is gonna review it + +**Natalie Pistunovich:** I sometimes put into some AI, if I don't understand something, and I read it 2-3 times, and I keep staring at it... I'm like, "Just explain that to me in other words", and that helps. + +**Anderson Queiroz:** Good. AI... \[laughs\] + +**Natalie Pistunovich:** It's like pinging somebody... But yeah. But also make the changes, just like you, because I think if I as a non-native don't understand this, there must be another non-native that gets lost there... And clarity is important. + +**Anderson Queiroz:** Yeah. + +**Natalie Pistunovich:** Let's say you're interviewing, whether you are the candidate or you are the interviewing person. And part of the interview is reviewing a pull request from somebody from your team. What tips do you have for somebody to do this well? + +**Anderson Queiroz:** I've actually never been really in these shoes, on either side. I've been asked - in one of the codebases, any of the codebases, what would you change, or something? I think at the end of the day, a lot of the time, to interview for culture fit, and a person that's nice, and it's good to work with, this is super-important. So I think if you're in an interview, just be sure to be nice, in your comments and everything. Don't go like, "Oh yeah, this is crap. This is bad." Just be nice, be polite, link documentations, and \[unintelligible 00:37:34.26\] bring arguments; don't say "Okay, do that" or "Change that" or "This needs change" without a reason. + +Usually, in an interview you don't have so much time... So I would go for "Oh yeah, this name is not ideal. The effective goal - there is a section on name conventions that explains why it should be like that. So as it's in Go, it's better to be like that." + +\[38:03\] So try always to bring something to support your views. And when it's an opinion - and that's something that I really do in pull requests; when it's my opinion, I say "Look, this is my opinion. I believe that's better, because this, this and this. It's up to you, because I don't see a flaw here. I just think it can be better," but in my opinion. + +**Angelica Hill:** If you were interviewing someone and their task was to review a PR, what would be things that they did that would maybe have you like "Oh, no... I don't know about that..." + +**Anderson Queiroz:** A thing would be to be aggressive, and impolite. To just diminish the code and say the code is bad, or something... An even if they show that they don't know what they're doing. + +I think if you interview people, you know some people don't know what they're doing, or they're just trying to fool you. If it's such \[unintelligible 00:38:53.17\] + +**Natalie Pistunovich:** It's better to say you don't know. + +**Anderson Queiroz:** It's not like that... It's like, "Great! Just go!" And you know, I just go "Yeah. Mm-hm." I just incentivize. "Yeah, go. Go. Please." And at the end, "Ah yeah, thank you very much. We're gonna be in contact." + +**Natalie Pistunovich:** Yeah. And definitely staying honest is a lot better than making things up. + +**Anderson Queiroz:** Yeah. Please say "I don't know." I think if someone to me in an interview says "Look, I don't know. I don't remember." Or "I don't know, and I think in that place I can get the information" - you're always scoring a hundred points with me. If you're trying to just BS me through it... Nah. + +**Natalie Pistunovich:** That's fair. + +**Angelica Hill:** Anderson will not be having that. + +**Break:** \[39:34\] + +**Angelica Hill:** Okay, so we have touched on this a little earlier in the episode, but I wanna dig a little bit deeper... In fact, I will ask you a question first. Are you engaged in any kind of open source projects? I know you said that you contribute to Go a little, is that a world in which you feel like you have engaged and put PRs in? + +**Anderson Queiroz:** So I can cheat my answer, right? + +**Angelica Hill:** Okay. + +**Anderson Queiroz:** Yes, I'm involved in open source, because I work at Elastic, and the majority of the wrappers are open source, right? But as a 100% open source contribution that I'm not working for the company, no. It's something that I always wanted. As I said, I managed to get a commit too in Go, but I haven't fully got to participate in a project. I'm still trying. Oh no, I think I got one in Kubernetes, too. See, it's one of my goals... You know, there's plans, things you wanna do, so try a bit and it goes back and forth... At GopherCon, again, perhaps the flame is just lighting up again. + +**Angelica Hill:** Reignite that passion. + +**Anderson Queiroz:** Let's see if it's gonna happen. But no, I'm not really engaged on, let's say an external open source project, that's not part of my daily job. + +**Angelica Hill:** And when you have done it, is part of the reason why you think it's difficult to engage fully - is anything to do with the difficult or the different process to put in a PR when it is an open source project, as opposed to internal, like within work PR reviews and submission? + +**Anderson Queiroz:** I think to me what's always difficult is to find something meaningful to work. You don't know what you can do... There is a tag, "First good issue" and so on, but I think that's the point - we're lost, and we don't have someone to go "Oh, please help me." Or "I tried that." Or "Shall I do that?" Because my pros and my teammates \[unintelligible 00:43:55.17\] to go and do a refactor. If it's a completely external project that I don't know anyone there, or anything, I'm gonna be afraid. And sometimes you're not even able to run the project. + +\[44:12\] So I think that's a lot of \[unintelligible 00:44:12.29\] if you would have something -- not necessarily a mentor, but perhaps a channel to ask questions. And even say "Oh, I wanna take this issue." Because sometimes the good first issues - they opened one year ago. And it's like, "Dude, I don't know if it's worth to fix that or not." And then you open the pull request everything, and no one reviews it. And then you ping "No one reviews." That thing got stale, you got demotivated. + +**Angelica Hill:** Yeah. Fair enough. And do you think there's a higher bar as to what you're willing to put in as a PR for an open source project? I speak about this from my own personal experience... When I was trying to get into like "Okay, I wanna contribute to open source." "Oh, you should just go in and do like a little change." But I was like, "Yeah, but I feel like I'm contributing it to an open source library where everyone can see." And they're like, "Oh, Angelica made like a one-line change... Or changed that one function name." + +**Anderson Queiroz:** I think -- and I talk for myself -- I always had the feeling to contribute to open source projects the bar is super-high. You have to be an expert developer, and everything... And the reality is no. The issue is there, the problem exists... I think this is the best advice I can do to any junior developer - just go for it. The no, you already have, right? So your change is not there, the bug is not fixed. The documentation is not improved, the feature is not there. So this is not gonna change; if your change didn't get there, you learned something. You played with a new technology. + +I was trying to submit pull requests for the Kubernetes code, just fixing linting issues. I understand a bit how that piece of code worked. And it was like "Oh, how they structured that. And look at the packages", because I had to read through the packages to fix linting issues. I think one PR got merged, the other ones got stale... That's life. It happens. And I learned something. So that's right, go for it. Let the other one say no. It's not your job to say no for you. It's their job. + +**Natalie Pistunovich:** That is a popular opinion, I bet. It applies to many fields in life. + +**Anderson Queiroz:** That's true. + +**Angelica Hill:** Just do it. + +**Anderson Queiroz:** Yes. Try. + +**Natalie Pistunovich:** Yeah. Well then, alright... The fun part. Unpopular opinions. Before we started the recording, Anderson, you mentioned you have several unpopular opinions, and you were wondering whether you should go for the most -- how did you phrase that? I forgot. You used a good word. The most... + +**Anderson Queiroz:** Controversial? + +**Natalie Pistunovich:** Yeah, exactly. Or the least controversial. + +**Anderson Queiroz:** Yeah. I'll pick a controversial one... But I can explain. As I said on the PR, you can explain. You should not write more than 100 columns. Write your code; your code should not really pass 100 columns. + +**Natalie Pistunovich:** From width. + +**Anderson Queiroz:** Yeah. First things - there's no magical number like 100 cut. I would say 110 is okay, when it's really bad to cut. 120 is almost a hard limit; don't really go over that. Why? First things first. Do you read books on landscape? No. Right? Come on, I think everyone had that; we got this email, you're like on your 4K or whatever monitor, and that thing goes from side to side, and you're reading -- for the listeners, I'm moving my head, as, you know, reading from one side to the other. + +**Angelica Hill:** Like you're watching tennis. + +**Anderson Queiroz:** I feel like a typewriter that goes from side to side. So it's hard to read, because we don't read in landscape, we read in portrait. The second thing is not everyone has got a screen as big as yours. There are people coding in 14, in 13 inches. They want to have two tabs open, perhaps. So if you go much more than 100, it's gonna be bad for some people, and I believe for everyone too long is hard to read. So that's my unpopular opinion. + +**Natalie Pistunovich:** \[48:14\] I want to disappoint you that I think I agree with you, and I even take this into writing emails. I try to keep it -- I don't know how many characters that is, but I sometimes break lines. Like, one sentence into 3-4 lines, just so it stays, so you don't have to scroll in case images, logos, whatever happens in somebody's signature that it suddenly gets stretched. + +**Angelica Hill:** I think you're unfortunately preaching to the choir with me and Natalie. We're both like, "Yes. Please! My neck will hurt." + +**Anderson Queiroz:** Perhaps in Go -- + +**Natalie Pistunovich:** A Chrome plugin that just truncates things for you. + +**Angelica Hill:** Yeah. + +**Anderson Queiroz:** Perhaps in Go. But I think if you go to Java, the things are long there. + +**Natalie Pistunovich:** Yeah, 100 characters is just the function name. + +**Angelica Hill:** Oh, my gosh... \[laughter\] I feel like that one was a good one... But if we have time, Natalie, I wanna hear another one. + +**Natalie Pistunovich:** Yeah. + +**Angelica Hill:** I want us to get an unpopular one from you. + +**Anderson Queiroz:** Oh, my God... + +**Angelica Hill:** What was the other one that you were thinking about saying, that you cheekily in your mind were like "No, I'm not gonna say that." + +**Anderson Queiroz:** No, I think this one's like the unpopular/kind of popular... + +**Angelica Hill:** Okay... + +**Anderson Queiroz:** I mean, I see a lot happening, so perhaps an unpopular... Return new is wrong. Period. You have to wrap the errors and add more context. Always. I cannot count how many times I had to go through the code and dig deep and deep and deep to discover where this error came from. You know, it's like when you try to write to the disk and you get an error. You get something like... I got one. "Too many colons in the address." You're like, "This is a valid address. How there is too many colons?" And then you have to understand where this address was gonna be used, which method it was, and then "Oh yeah, on this context there are too many colons." But they didn't have this information. + +So return new is wrong, you have to wrap your errors. Now you don't have an excuse. You have error wrapping in the standard library. + +**Natalie Pistunovich:** So you think wrapping errors will be an unpopular opinion. + +**Anderson Queiroz:** Yeah. + +**Angelica Hill:** I feel like this second one is gonna probably be more unpopular. + +**Anderson Queiroz:** Yeah. Because a lot of people just return the error. + +**Angelica Hill:** Yeah. + +**Natalie Pistunovich:** Interesting. + +**Anderson Queiroz:** That's one of the things that I look in pull requests. Everyone's like "Dude, could you wrap that?" And then it goes back on that. If it's my repo, my code, I might say "No." Now, when I say "my", please, my team's, right? I don't have this possessive. I think code must be owned by a group, and it must be a consensus in the group... But yeah, this is a thing that I'm gonna point. And usually if it's an external adding code, even more important. Like, your repo, your rules, our repo, our rules. Here we wrap till died. If I'm on your repo, I play by your rules. + +**Angelica Hill:** I like that one. I'm also having so many more ideas. This always happens when me and you have an episode, Natalie. I'm like "So many more episode ideas..." + +**Natalie Pistunovich:** Write them all down! + +**Angelica Hill:** "Does your code belong to you, or to the world...?" + +**Anderson Queiroz:** I mean, in open source that's definitely a question, right? + +**Natalie Pistunovich:** Yeah, also with AI tools that are writing code, that's a question. + +**Angelica Hill:** Yeah. + +**Anderson Queiroz:** Oh God, yeah. + +**Angelica Hill:** Who is the true owner of code...? + +**Anderson Queiroz:** Yeah, that's a good one. + +**Natalie Pistunovich:** Yeah, licensing is interesting, for sure, within Copilot and friends. + +**Anderson Queiroz:** When the AI put the bug in production, who do you blame? Who ran the AI, who wrote the AI... + +**Natalie Pistunovich:** ...or who reviewed the PR? + +**Anderson Queiroz:** That's a good one, right? Can I review a PR? + +**Natalie Pistunovich:** For sure, for sure. + +**Anderson Queiroz:** Would you trust that? + +**Natalie Pistunovich:** I might have used that in the past, yes... \[laughs\] + +**Anderson Queiroz:** Oh, that's nice. Okay. + +**Natalie Pistunovich:** I'm a big fan of AI in coding. I think it's a fun combination. I'm very happy to automate myself out of a job. + +**Anderson Queiroz:** That's good. \[unintelligible 00:51:45.22\] I know almost zero about AI in coding, but I'm super-interested. It looks really interesting to see where is it going. + +**Natalie Pistunovich:** Yeah, definitely... Both from the side of writing and from the side that's relevant to this episode, which is the reviewing and the explaining, and so on. + +**Anderson Queiroz:** I think in general, AIs can see the context, and they carry so much information. Part of the things that we just can't. And sometimes someone really experienced can, but they cannot teach it... So there's definitely a lot of value. + +**Natalie Pistunovich:** Yeah, exactly. About the context in particular. + +**Angelica Hill:** I sense a Twitter poll, Natalie... "Would you let an AI review your code?" + +**Natalie Pistunovich:** Hm... Maybe this can be my unpopular opinion for this episode. I would not have it just review "Good/Bad", but I would use it as something like "Here's the code. What does it do?" Or "List the problems, and then read the output, and give it a secondary review." I think this might end up being a popular opinion. + +**Anderson Queiroz:** True. + +**Angelica Hill:** Now you've explained it and won us over to your side. I feel like it will be popular. + +**Anderson Queiroz:** Yeah... + +**Natalie Pistunovich:** Okay, my next Chrome plugin. Each episode is like 15 other ideas, yeah... Summarize this PR for me. Well, this has been fun and this has been inspiring. Anderson, thanks a lot for joining us and sparing some of your time at GopherCon U.K. + +**Anderson Queiroz:** Thank you very much for having me. It was really good. I'm really happy. + +**Angelica Hill:** Thank you so much. We're gonna have to get you back on for the "Who owns our code?" episode. \[laughter\] + +**Anderson Queiroz:** Yeah, I would love it. + +**Natalie Pistunovich:** It's a plan. And everybody I hope you will also join us, and thanks for joining us this time. + +**Angelica Hill:** Bwye! + +**Anderson Queiroz:** Thank you very much. Bye-bye! diff --git "a/The art of the PR\357\274\232 Part 1_transcript.txt" "b/The art of the PR\357\274\232 Part 1_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..9cac60a73d9fd73e2adb5f43f9109919d0e942ce --- /dev/null +++ "b/The art of the PR\357\274\232 Part 1_transcript.txt" @@ -0,0 +1,1732 @@ +[0.00 --> 7.00] If I'm blocking Angelica's pull request and I leave a bunch of comments and some of them are like, great job, this is really cool. +[7.00 --> 10.22] And some of them are style and some of them are actually questions. +[10.22 --> 13.82] And then some of them are actual blocking like requests for changes. +[13.82 --> 24.48] It helps kind of narrow to those and like create more of a clear checklist in a way of what you're expecting to be changed before being asked to re-review the code. +[24.48 --> 35.54] So I just think it's helpful, even if you are formally like blocking the pull request to communicate like what things you expect to be changed before you think that it could be merged. +[35.92 --> 44.88] Gotcha. I don't know if I was just projecting my own fear, like seeing the email where it says this has been like changes have been requested and oh no, what did I do wrong? +[45.22 --> 46.18] I need to do that. +[46.80 --> 49.76] Don't worry, we all do stuff wrong in our pull requests all the time. +[50.14 --> 51.46] That's part of being an engineer. +[54.48 --> 60.84] This episode is brought to you by Sourcegraph with the launch of their Code Insights product. +[60.84 --> 63.84] Teams can now track what really matters in their code base. +[64.16 --> 70.46] Code Insights instantly transforms our code base into a queryable database to create visual dashboards in seconds. +[70.94 --> 73.90] And I'm here with Joel Cortler, the product manager of Code Insights for Sourcegraph. +[74.36 --> 84.06] Joel, the way teams can use Code Insights seems to pretty much be limitless, but a particular problem every engineering team has is tracking versions of languages or packages. +[84.58 --> 87.26] How big of a deal is it actually to track versions for teams? +[87.72 --> 89.46] Yeah, it's a big deal for a couple of reasons. +[89.64 --> 91.46] The first is, of course, just compatibility. +[91.46 --> 95.58] You don't want things to break when you're testing locally or to break on your CI systems or test systems. +[96.06 --> 102.24] You need to have some sort of level of like version unification, minimum version support, and all of that needs to be compatible forward. +[102.24 --> 116.38] But the other thing we learned was that for a lot of customers, especially, you know, engineering organizations that are pretty established, they have older versions of things or even older versions of like SaaS tools they don't use anymore that they haven't fully removed because they're like not sure if it's still in use or they, you know, lost focus on that. +[116.38 --> 119.50] And they're spinning up old virtual machines that they're still paying for. +[119.62 --> 123.62] They're using, you know, old SaaS subscriptions they're afraid to cancel because they're not sure if anyone's actually using it. +[123.74 --> 137.00] And so getting off of those versions not just like saves you the headaches and the risks and the vulnerabilities of being on old versions, but also literally the money of, you know, older systems running more slowly or the build times or, you know, virtual machines and SaaS tools that you're no longer using. +[137.00 --> 139.22] Before you had this ability, we talked to teams. +[139.52 --> 140.90] There are basically three ways you could do this. +[141.18 --> 144.62] You could slack a million people and ask for just like an update point in time. +[144.88 --> 153.88] You could have sort of one human in one spreadsheet where like it's somebody's job every Friday or every two weeks to just like search all the code and find all the versions and write it down in a Google sheet. +[154.12 --> 158.40] Or there were a couple of companies I came across with in-house systems that were sort of complicated. +[158.60 --> 161.52] You had to know, you know, maybe Kotlin, but you didn't know Kotlin. +[161.52 --> 168.80] But if you want to use this system, you had to learn Kotlin and you'd have to sort of build the whole world from scratch and run basically a tool like this with a pretty steep learning curve. +[169.18 --> 176.72] And now for all three of those, you could replace it with a single line source graph search, which is basically just the name of the thing you're trying to track and the version string in the right format. +[176.98 --> 180.24] And then we have templates that will help you get started if you're not sure what that format is. +[180.36 --> 183.50] And then it'll automatically track all the different versions for you, both historically. +[183.68 --> 185.84] So even if you start using it today, you can see your historical patterns. +[185.96 --> 187.30] And then, of course, going forward. +[187.90 --> 188.10] Very cool. +[188.16 --> 188.56] Thank you, Joel. +[188.56 --> 193.10] So right now there is a treasure trove of insights just waiting for you. +[193.44 --> 199.96] Living inside your code base right now, teams are tracking migrations, adoption, deprecations. +[200.28 --> 203.14] They're detecting and tracking versions of languages and packages. +[203.14 --> 207.10] They're removing or ensuring the removal of security vulnerabilities. +[207.48 --> 209.06] They understand their code by team. +[209.14 --> 210.84] They can track their code smells and health. +[210.84 --> 215.40] And they can visualize configurations and services and so much more with code insights. +[215.40 --> 222.02] A good next step is to go to about.sourcegraph.com slash code dash insights. +[222.30 --> 224.84] See how other teams are using this awesome feature. +[225.08 --> 229.90] Again, about.sourcegraph.com slash code dash insights. +[230.16 --> 231.92] This link is in the show notes. +[231.92 --> 248.04] Let's do it. +[248.80 --> 249.70] It's go time. +[250.44 --> 255.52] Welcome to Go Time, your source for diverse discussions from all around the Go community. +[256.04 --> 257.14] Subscribe to the pod. +[257.14 --> 260.82] If you haven't yet, head to go time.fm for all the ways. +[261.16 --> 263.54] And if you dig the show, please do tell your friends. +[263.70 --> 264.48] That'd be pretty cool. +[264.92 --> 269.40] Special thanks to our partners at Fastly for shipping all of our pods super fast to wherever +[269.40 --> 270.00] you listen. +[270.22 --> 271.94] Check them out at fastly.com. +[272.04 --> 273.90] And to our friends at fly.io. +[274.40 --> 276.50] Post your app servers close to your users. +[276.72 --> 277.66] No offs required. +[278.10 --> 279.40] Learn more at fly.io. +[279.70 --> 280.60] Okay, here we go. +[282.36 --> 286.32] Hello and welcome to Go Time. +[286.32 --> 290.08] Today we're going to be talking about PRs. +[290.36 --> 291.50] What makes a good PR? +[292.06 --> 294.22] How do you do the best PR review? +[294.80 --> 299.72] Is there such thing as a PR that is too small, too big, too filled with emojis? +[300.20 --> 305.22] We'll be debating all the details and trying to help our fellow gophers master the art of +[305.22 --> 305.62] the PR. +[306.18 --> 309.40] Today I'm joined by three wonderful PR pros. +[309.40 --> 316.78] First, we have the wonderful Jeff Hernandez, who is a associate software engineer at the +[316.78 --> 317.44] New York Times. +[317.88 --> 318.52] Hello, Jeff. +[318.62 --> 319.34] How are you today? +[319.80 --> 321.26] Hi, doing well. +[321.36 --> 322.28] Thanks for having me back. +[322.50 --> 323.74] Thank you for joining us again. +[324.58 --> 329.04] Next up, we have Sarah Duncan, who's a staff software engineer at the New York Times. +[329.04 --> 333.62] She also teaches an introductory programming course at a high school. +[333.86 --> 335.00] So thank you for joining us. +[335.04 --> 336.78] I know you're a first time Go Time guest. +[336.94 --> 338.22] So lovely to have you. +[338.72 --> 339.58] Yeah, thanks for having me. +[339.94 --> 345.56] And last, but certainly not least, we have Natasha Dykes, who is a senior software engineer +[345.56 --> 349.10] at the New York Times and happens to be a cycling enthusiast. +[349.46 --> 349.64] Hi. +[349.92 --> 350.70] Thanks for having me. +[351.04 --> 351.44] Hiya. +[351.76 --> 352.98] Thank you for being here. +[352.98 --> 360.58] And we have the beautiful, the wonderful, the incomparable Natalie, who is my co-host. +[361.02 --> 361.28] Hello. +[361.84 --> 362.68] Hi, Angelica. +[362.74 --> 365.16] I think it's very smooth you skip pronouncing my last name. +[365.22 --> 366.50] I would probably do the same thing. +[366.58 --> 368.10] It's so complicated. +[368.84 --> 370.56] I haven't been on in a hot second. +[370.56 --> 375.48] So I'm kind of trying to minimize the amount of babbling and mistakes I get myself into. +[376.74 --> 377.60] Great strategy. +[378.06 --> 380.34] Yeah, we'll see if that persists throughout the episode. +[380.34 --> 383.62] I might just get overexcited and fumble over my words. +[384.12 --> 385.94] But thank you all for joining me today. +[386.14 --> 388.90] I'm extremely excited to talk about PRs. +[389.12 --> 391.02] So we're going to start with the very basics. +[391.66 --> 394.40] What is a PR and why do we even do them? +[394.86 --> 398.76] So I'm going to pass over to you, Sarah, when you're talking to your wonderful high school +[398.76 --> 400.92] students and they go, what is a PR? +[401.06 --> 402.00] Like, what is this thing? +[402.42 --> 403.46] How do you explain it to them? +[403.82 --> 404.04] Sure. +[404.04 --> 411.88] Well, PR stands for a pull request and it is typically used to refer to somebody who's +[411.88 --> 416.90] been making changes to a shared code base, making a request to add those changes back +[416.90 --> 419.94] into the common main. +[420.50 --> 422.22] Typically, it's like a branching situation. +[422.36 --> 424.26] So it's the main branch of that code base. +[424.70 --> 429.90] And that also typically corresponds to moving that code into production if it's a production +[429.90 --> 430.38] system. +[430.38 --> 438.02] So it's a way to get a review on your work and basically ask to add code to the main +[438.02 --> 438.42] system. +[439.08 --> 439.90] And why is it useful? +[440.14 --> 441.78] Why is this something that we want to be doing? +[442.14 --> 446.08] Is that something that maybe, I don't know, like Jeff, do you find PRs useful? +[446.56 --> 450.18] I mean, coming from like, at the times, at least I'm an associate software level. +[450.24 --> 451.42] So I'm kind of like the entry level. +[451.90 --> 457.46] So it's a great way to get feedback from my senior engineers and basically get a lot of +[457.46 --> 462.42] kind of feedback from them in terms of like what I can be doing better or like kind of +[462.42 --> 464.56] like code structure, how that could be improved. +[464.90 --> 468.26] Or even like tiny little like optimizations. +[468.66 --> 473.74] Or maybe there's a certain way that, for instance, Go likes to do things because as we all know, +[473.78 --> 474.70] Go is very opinionated. +[475.48 --> 480.42] So it's just it's like a way to tap into that resource from our senior engineers. +[480.42 --> 485.78] I feel like it's a great tool, especially for someone that's coming into a new team, +[486.16 --> 490.54] kind of getting the lay of the land, getting the norms that you typically might not get +[490.54 --> 492.52] like in through other forms of documentation. +[493.72 --> 494.16] No, for sure. +[494.66 --> 500.38] And in terms of assessing PRs, is there such thing as a good PR review? +[500.48 --> 502.86] Is there such thing as a bad PR review? +[502.86 --> 509.14] I would love to hear how you assess going about either putting in a PR or reviewing a PR. +[509.76 --> 515.52] Maybe Natasha, like when you're putting in a PR, how do you decide whether it's time to go +[515.52 --> 518.18] or whether you need to wait and do some more work? +[518.26 --> 521.34] Like at what point do you feel like, no, this is ready to be reviewed by my peers? +[521.84 --> 526.16] I think for me, it's helpful when I review like a ticket. +[526.30 --> 529.40] Usually it's for a feature or something that I'm working towards. +[529.40 --> 534.22] Sometimes I review to see if like the work that I've done actually meets the requirements. +[534.68 --> 537.80] And at that point, I can either say like, okay, it's ready to go. +[538.12 --> 543.84] I've cleaned up any notes for myself or made sure I did like go format, all of those small +[543.84 --> 544.28] things. +[544.40 --> 545.46] And then I'll open up a PR. +[545.80 --> 551.02] But I would say like, you can even open up a PR before you're ready, like ready to merge. +[551.12 --> 553.18] Sometimes it's good to just get that early feedback. +[553.60 --> 556.08] So it really depends on the work that I'm doing. +[556.08 --> 559.72] And Natalie, I see your intake of breath. +[560.02 --> 561.32] Do you have something you'd like to add? +[561.70 --> 561.96] Yeah. +[562.12 --> 564.72] You asked earlier, what is a PR? +[564.98 --> 567.80] And it's interesting also to compare PR and CR. +[568.10 --> 569.80] And why is it even pull, right? +[570.12 --> 573.22] So PR, as Sarah said, stands for pull request. +[573.74 --> 576.28] So let's split that into two questions. +[576.76 --> 577.36] Why pull? +[577.52 --> 578.20] Why not push? +[578.30 --> 579.10] Why not merge? +[579.14 --> 580.68] Or some other thing? +[580.68 --> 584.70] And what is the difference or what do you prefer? +[585.12 --> 589.20] CR, which stands for code review or PR as pull request. +[589.64 --> 591.02] It's an interesting question to discuss. +[591.12 --> 595.08] It's also interesting, even a little bit in the concept of like, what does that represent? +[595.56 --> 596.94] Curious to hear everyone's thoughts. +[597.52 --> 602.74] I think code review is more semantically correct for like the work that I'm typically doing. +[602.74 --> 608.32] But I think the changes that you could potentially make don't necessarily need to touch code. +[608.88 --> 610.50] It could be like a readme update. +[610.62 --> 616.88] It could be some other stuff that's like kind of supporting the repository versus just the code. +[617.54 --> 619.72] I don't know why it's called pull request though. +[619.98 --> 620.84] Never really thought of that. +[620.98 --> 623.46] So curious to hear what others think. +[623.88 --> 631.64] Now this gets me kind of wanting to go to Wikipedia and kind of learn about where pull request came from and like that whole background. +[631.64 --> 639.20] That's one of the things I'm like interested about, like just like where everything came from in terms of like software engineering best practices and naming conventions. +[639.20 --> 648.54] Because once you're in the industry, you know, it's just something we know, you know, but it's not something that like, at least it's like, it's a term that's used across the whole industry. +[648.54 --> 653.12] But it's not something that we as a team or something that we choose to use. +[653.14 --> 656.42] It's just kind of like a shared language among software engineers. +[656.58 --> 656.70] Right. +[656.70 --> 659.96] So I was super cheeky and I did Google it just now. +[659.96 --> 669.10] And it says that the name pull request comes from the idea that you're requesting the project to pull changes from your fork. +[669.70 --> 677.60] That might not encompass all ways that we now use it in our language, but that's what Google's telling me on the top line. +[678.52 --> 679.52] Yeah, it is interesting, right? +[679.56 --> 680.78] You have this project. +[680.88 --> 683.26] I mean, most of us use Git in some way. +[683.46 --> 688.66] GitHub, GitLab, or I don't know too many other personal variations, but I'm sure that exists as well. +[688.66 --> 702.32] So we're all kind of eventually are used to the concept of like having a main branch and then making branching kind of your changes from that and then asking to merge that back, whether frequent or not frequent. +[702.32 --> 705.52] And yeah, like we all say PR. +[705.80 --> 711.76] I've been always saying PR, but then recently I had to work with a ticketing system that is called ClickUp. +[711.76 --> 716.04] And there they said that the label was code is in CR. +[716.62 --> 718.22] And that was kind of interesting. +[718.76 --> 721.86] Probably the first time I remember, let's say, encountering this. +[722.00 --> 726.34] So I also went to dive a little into the semantics of that. +[726.86 --> 731.94] Then there's also merge request, which kind of makes sense, but actually also not really in use. +[732.30 --> 736.20] I've also heard it called like a changelog or CL. +[736.20 --> 738.78] So there's a lot of different terms for it. +[739.18 --> 741.76] What are you familiar with, you know, changelog to be? +[742.14 --> 747.94] It's the same thing as like a pull request, but that's the certain companies call it a changelog. +[748.16 --> 749.78] So you're submitting a changelog? +[750.00 --> 752.00] Yeah, that's what I've understood it to be. +[752.16 --> 753.94] It's just like something I heard in passing. +[754.70 --> 764.58] Yeah, I've seen that in like in the Go repository, like in references to like the actual like the GitHub Go repo and like issues, people referring to changelogs. +[764.58 --> 765.90] Yeah, it's a Google thing. +[766.12 --> 766.84] It's a Google thing. +[766.92 --> 769.70] Yeah, we'd totally be interested in learning more about that. +[770.20 --> 776.68] That was me thinking that like the most basic of questions, what is a PR was just me doing my due diligence for the newbies. +[776.70 --> 778.30] But now I see it's a whole debate. +[779.34 --> 781.22] So many different words to use. +[781.22 --> 782.12] This is great. +[782.24 --> 787.82] We're opening up a Pandora's box of PR words and ways to think about code. +[788.04 --> 788.62] I mean, right. +[788.74 --> 790.12] Software is all about naming things. +[790.70 --> 791.78] It starts this early. +[791.78 --> 800.84] Okay, so when you're thinking about a PR, like I've heard many people complain about, oh, this PR is too long or this PR is so short. +[800.92 --> 802.36] Why didn't you put it in one big PR? +[802.60 --> 805.86] Is there such thing as a too long or a too short PR? +[805.98 --> 806.86] Too big, too small? +[807.00 --> 812.42] Or is it really just down to like dealer's choice, whatever that team is happy to review? +[812.42 --> 822.44] If you're happy to review a thousand line PR or a five line PR, I'd love to hear anyone's view on kind of the length of the PR and how big it should be. +[822.44 --> 832.18] Yeah, I definitely think that pull requests or merge requests or code reviews, these can be too long. +[832.58 --> 834.54] Sometimes a pull request is too large. +[834.78 --> 847.36] And my benchmark for helping set that norm on my team for what is an appropriate size pull request is really around like how well the pull request can be reviewed. +[847.36 --> 861.14] So if you do have like a thousand line pull request, another common adage about engineers, like we're lazy and we're not probably not going to review a thousand line pull request as in depth as we would a 20 line pull request. +[861.56 --> 868.46] So I think it's about thinking about how you can best set up your reviewers to give you a quality, thorough review. +[868.46 --> 881.84] So I have some thoughts around like how to set that norm, but I'd love to hear from others like what you think about a length of a PR and whether a pull request can be too big or too small. +[882.36 --> 888.96] Yeah, I don't think a pull request can necessarily be too small because it just takes one character to make a bug. +[888.96 --> 891.28] So you're going to have to make that change. +[891.66 --> 894.50] But I do think that I agree with you. +[894.62 --> 906.18] It could be too long to kind of have a reviewer to actually sit down and understand all the changes, especially if they're not as familiar with you or with the code as you are. +[906.40 --> 916.32] So I don't necessarily have the best strategy, but I think just like encapsulating certain functionalities, I think helps instead of just like an entire feature. +[916.32 --> 919.90] Because it might be broken down into different kind of core parts. +[920.32 --> 922.62] So that's typically how I lean on it. +[923.62 --> 924.86] Now, Jeff, you have any thoughts? +[925.50 --> 929.98] Yeah, I think that's it's totally like depending on what you're working on at the current moment. +[930.16 --> 930.38] Right. +[930.46 --> 939.06] So if you're building a new API from the ground up, like setting up the handling for the JSON, the payload, maybe that can be a PR. +[939.24 --> 941.38] And then the actual business logic can be a separate thing. +[941.44 --> 942.68] You're building up as you go. +[942.74 --> 942.90] Right. +[942.90 --> 955.74] And that's kind of something that I've learned from other people on the team because I have been guilty of this extremely large PR is where the ticket is done, but it's all in one PR and no one wants to review it. +[955.78 --> 960.92] So you're going to maybe get a review if my next week, if you keep pinging people and bothering them. +[961.04 --> 963.46] But otherwise, no one's going to want to touch that. +[963.46 --> 972.06] And then you don't ever want to get in a situation where you have to offer to kind of go step by step with the reviewers. +[972.70 --> 975.28] Like I've made this change because of this reason, like on a call. +[975.40 --> 980.22] I feel like that's the worst case scenario where you have to actually walk them through it. +[980.56 --> 984.62] Sometimes it's necessary, but it's something I would avoid. +[984.98 --> 986.10] And I don't know. +[986.16 --> 988.10] It's just it's kind of like a walk of shame. +[988.10 --> 992.28] I mean, like you have to go through the whole thing with them, but they can definitely be too big. +[992.38 --> 996.02] Walk of shame or you can like walk through the glory of your coding. +[997.38 --> 998.12] I guess. +[999.76 --> 1000.70] I'm just teasing. +[1000.90 --> 1002.08] It's all how you think about it, right? +[1002.44 --> 1008.70] Look at this beautiful PR that I constructed over many months that you now have to review with me. +[1009.22 --> 1011.84] Look at this clever naming convention I did. +[1011.96 --> 1013.38] Look at this great function. +[1014.04 --> 1015.62] Don't you love this go routine? +[1018.52 --> 1024.94] Also, so we talked a little bit about what you alluded to, Sarah, like being cognizant of the people who are reviewing your PR. +[1025.40 --> 1030.90] So I'd love to hear a little bit about kind of how many people should review your PR before you merge it. +[1031.24 --> 1032.08] Is it one person? +[1032.18 --> 1032.86] Is it two people? +[1032.94 --> 1034.36] Is it the entire team? +[1034.36 --> 1042.90] What are the thoughts around like reviewing the PR and how many people need to know about this great work you've done before it's in the world? +[1043.38 --> 1044.30] Natalie, you're smiling. +[1045.18 --> 1046.50] I'm sorry I'm doing this again. +[1046.58 --> 1048.20] I want to bring back to the previous question. +[1049.92 --> 1050.70] Do it. +[1051.10 --> 1053.66] So when you say too long or too short, what do we measure? +[1054.34 --> 1058.94] Number of lines, number of commits, number of files. +[1058.94 --> 1062.30] It can be depending on many things. +[1062.64 --> 1064.40] The answers can vary in the same PR. +[1065.00 --> 1069.44] Like it can be one PR where you removed one file, but that's like many lines. +[1069.90 --> 1070.40] Is that long? +[1070.46 --> 1070.84] Is that short? +[1071.44 --> 1072.58] This is my thought now. +[1072.76 --> 1076.98] But whoever listens doesn't know that Angelica can see me thinking out loud like that. +[1077.24 --> 1079.36] So I'm guessing there's a question coming up. +[1079.36 --> 1085.62] So you have to trust her psychic abilities as a product manager to kind of look at the engineers in front of her and like, you want to say something? +[1085.84 --> 1086.02] Speak. +[1086.54 --> 1088.44] That's the secret superpower of product managers. +[1088.66 --> 1089.42] I can see it. +[1089.56 --> 1093.16] Also, I'm learning like now, actually, just as a go time host. +[1093.28 --> 1097.94] Every time I ask a question, I should pause and defer to Natalie and be like, do you have a follow up question? +[1099.88 --> 1101.06] That is a great question. +[1101.14 --> 1102.42] I'm very glad that you asked it. +[1103.22 --> 1104.40] That definitely makes sense. +[1104.50 --> 1105.30] Asked it. +[1106.06 --> 1106.78] Really babbling. +[1107.32 --> 1108.36] Do you have a view on that? +[1108.94 --> 1109.30] Yes. +[1110.12 --> 1110.48] Wonderful. +[1110.66 --> 1111.66] I'd love to hear it. +[1112.16 --> 1114.72] But we're here to hear other people's opinions. +[1114.94 --> 1118.94] So I am curious to hear the crowd's opinions more than I am interested in sharing. +[1119.14 --> 1121.76] I mean, I'm also happy to share mine, but I'm more curious about others. +[1122.44 --> 1124.80] Is there like a convention in your team, for example? +[1124.96 --> 1127.14] How do you, would you say this is too much? +[1127.20 --> 1128.04] Like, what would you look at? +[1128.20 --> 1131.94] Yeah, I'm really curious about hearing more about Sarah's like the team standard. +[1131.94 --> 1139.26] But at least from me, if I'm reviewing something, the first thing I look at is the number of files that were changed. +[1139.80 --> 1141.22] That's just like the easiest thing to look at. +[1141.28 --> 1145.86] If it's like a long list, I am going to go to lunch maybe and then come back to it later. +[1146.18 --> 1153.84] It's definitely easier to just see, to review something that's like a couple files versus something that's touching multiple directories at a time. +[1153.84 --> 1159.08] And then because I have to go back and forth referencing things and seeing how it might affect other things. +[1159.54 --> 1162.72] It's just definitely files would be my number thing, number one thing. +[1162.80 --> 1165.24] And then typically the norm on our team is one commit. +[1165.38 --> 1167.72] So it's not something like one commit per PR. +[1167.90 --> 1171.30] So it's not something that gives much information for us. +[1171.78 --> 1174.76] Yeah, I definitely think files are a good initial indicator. +[1174.76 --> 1190.76] And sometimes that can be misleading because you could have a bunch of files only have one line change or you moved a folder into a subfolder and that changed like a bunch of files and you can just check all those off as okay. +[1191.24 --> 1197.12] But I think that it's a good question, Ali, because I think it's a hard thing to actually measure well. +[1197.12 --> 1210.02] And the analogy I try to use and that I recently used at the New York Times to kind of explain like how to break down your PRs ties into your question, Angelica, about the number of reviewers. +[1210.44 --> 1222.32] But my analogy is that if my ticket is to bake a cake and let's say it's chocolate cake, we've got some buttercream frosting, little raspberry layer in there. +[1222.38 --> 1223.12] It's a nice cake. +[1223.12 --> 1235.10] Like if I go and bake the cake and come back and Angelica, you're a frosting expert and Jeff, you are a cake expert and Natasha, you are the absolute queen of fillings. +[1235.64 --> 1236.42] It's harder. +[1236.66 --> 1243.78] I have to get all of you in a room and cut a slice of cake and like you have to pick it apart and like be an expert on your piece. +[1244.04 --> 1249.18] And it takes more work to give feedback and it takes more work for me to go fix something. +[1249.18 --> 1256.18] So if I'm trying to perfect the chocolate cake and Jeff, you're like, this cake is too dry. +[1256.42 --> 1260.14] Like you have to go back and make it make it less dry. +[1260.52 --> 1265.12] That is so much more work for me to reconstruct that cake all over again. +[1265.78 --> 1273.98] But if I break that down and first I make my my chocolate sponge, Jeff, you taste that you give me feedback on it. +[1273.98 --> 1279.08] I make my filling that I'm going to use Natasha, I get your input on that Angelica. +[1279.28 --> 1284.08] I ask you for your input on the frosting and kind of perfect those individual pieces. +[1284.08 --> 1290.08] And then I assemble that final assembled cake is going to be a lot more successful. +[1290.08 --> 1302.04] So I like to think about it as like breaking it down so that I ideally one person could review it and maybe like one subject matter expert could review that part. +[1302.98 --> 1306.42] And then when we get to the assembly piece, you're judging me on the assembly. +[1306.56 --> 1311.56] You already know that the individual components are what we're on the same page about what those components are. +[1311.56 --> 1319.30] So I think I think it's a hard thing to measure because even just that like, oh, one subject matter expert is not like a hard and fast rule. +[1319.52 --> 1324.60] This is one of the things I think is something that comes more with experience and intuition. +[1325.00 --> 1332.98] It's like a skill that you're able to hone as you gain more experience around like what the right size to ask for feedback on is. +[1332.98 --> 1344.78] But that's my my analogy is that if you're baking a cake and you're asking for a review on it, it's easier to get the individual components reviewed first and then bake your cake or assemble your cake. +[1347.68 --> 1349.28] That was so great. +[1350.30 --> 1353.58] It's the teacher in me, you know, we got to bring in the fun things. +[1355.96 --> 1356.98] That was wonderful. +[1357.26 --> 1358.62] I feel like I've learned so much. +[1358.72 --> 1362.78] Genuinely, I will never think of could be ours and construction of a. +[1362.78 --> 1363.42] Now they're cake. +[1363.42 --> 1364.74] Feature again in the same way. +[1365.20 --> 1367.14] Like, can I see the frosting, please? +[1367.24 --> 1367.46] Please. +[1368.12 --> 1369.38] Let me check the topping. +[1369.58 --> 1369.74] Yeah. +[1369.90 --> 1371.10] How moist is the cake? +[1371.86 --> 1373.30] How moist is this cake? +[1373.60 --> 1378.02] I'm going to be very careful who I say that to because some people don't really like the word moist. +[1378.74 --> 1381.00] I don't particularly, but it's fine. +[1381.74 --> 1382.48] Judge your audience. +[1382.68 --> 1382.84] Yeah. +[1383.64 --> 1384.48] Judge your audience. +[1384.48 --> 1385.58] Good perception. +[1386.54 --> 1391.46] So flipping over kind of to the other side of a PR. +[1391.46 --> 1395.44] How do you go about reviewing a PR? +[1396.00 --> 1396.76] I'm going to pause. +[1397.12 --> 1403.68] Natalie, would you like to talk about putting in a PR in any more detail before we switch over? +[1403.90 --> 1404.30] No, no. +[1404.38 --> 1404.62] Sorry. +[1404.72 --> 1406.76] I'll go back to being German and sticking to the schedule. +[1408.30 --> 1408.88] No, no, no. +[1408.88 --> 1409.66] Please don't. +[1409.96 --> 1410.24] I was. +[1410.52 --> 1413.10] That came across like I was being a little bit sassy. +[1413.24 --> 1413.76] No, no, no, no. +[1413.78 --> 1415.24] But it was a genuine question. +[1415.24 --> 1417.40] I love making German jokes. +[1417.52 --> 1417.74] Anything. +[1419.18 --> 1421.06] I should have made it a German chocolate cake. +[1421.14 --> 1421.90] I'm so sorry, Natalie. +[1421.98 --> 1422.96] I missed that opportunity. +[1423.14 --> 1423.86] Oh, my gosh. +[1424.22 --> 1424.62] Eskule. +[1425.50 --> 1426.46] This is Austrian. +[1426.66 --> 1430.52] Natalie will be the final, like once the beautiful cake has been constructed. +[1430.64 --> 1432.34] She is the ultimate SMA. +[1432.40 --> 1433.72] Wait, did you use sugar or salt? +[1433.88 --> 1435.12] The ultimate taste tester. +[1435.12 --> 1456.70] This episode is brought to you by our friends at Fire Hydrant. +[1457.10 --> 1459.74] Fire Hydrant is a reliability platform for every developer. +[1460.24 --> 1462.72] Incidents are a win, not an if situation. +[1462.72 --> 1467.24] And they impact everyone in the organization, not just SREs. +[1467.34 --> 1470.86] And I'm here with Robert Ross, founder and CEO of Fire Hydrant. +[1471.26 --> 1473.52] Robert, what is it about teams getting distracted by incidents +[1473.52 --> 1476.84] and not being able to focus on the core product that upsets you? +[1477.22 --> 1481.16] I think that incidents bring a lot of anxiety and sometimes fear +[1481.16 --> 1485.36] and maybe even a level of shame that can cause this paralysis +[1485.36 --> 1488.02] in an organization from progress. +[1488.02 --> 1494.44] And when you have the confidence to manage incidents at any scale of any variety, +[1494.76 --> 1496.72] everyone just has this breath of fresh air +[1496.72 --> 1499.50] that they can go build the core product even more. +[1499.86 --> 1502.88] I don't know if anyone's had the opportunity, maybe is the word, +[1502.98 --> 1504.40] to call the fire department. +[1504.60 --> 1507.24] But no matter what, when the fire department shows up, +[1507.34 --> 1510.26] it doesn't matter if the building is hugely on fire. +[1510.40 --> 1511.88] They are calm, cool, and collected +[1511.88 --> 1513.94] because they know exactly what they're going to do. +[1513.94 --> 1517.14] And that's what Fire Hydrant is built to help people achieve. +[1517.14 --> 1518.52] Very cool. Thank you, Robert. +[1518.64 --> 1522.90] If you want to operate as a calm, cool, collected team +[1522.90 --> 1525.28] when incidents happen, you got to check out Fire Hydrant. +[1525.60 --> 1528.22] Small teams, up to 10 people can get started for free +[1528.22 --> 1531.20] with all the features, no credit card required to sign up. +[1531.50 --> 1533.18] Get started at firehydrant.com. +[1533.52 --> 1535.52] Again, firehydrant.com. +[1535.52 --> 1548.40] Great. So you are going to review a PR. +[1548.82 --> 1553.50] What are the almost like unspoken rules of PR review? +[1553.50 --> 1556.60] Are there actual rules of PR review? +[1557.18 --> 1561.22] Love to hear like, how do you approach reviewing your colleagues' PRs? +[1561.22 --> 1565.28] I think it goes a long way if you lead with empathy, +[1565.72 --> 1567.32] kind of like go through the PR, +[1568.02 --> 1571.02] address anything that you think could be updated, +[1571.24 --> 1573.76] but not in a way that you're talking down to people, +[1574.36 --> 1578.34] having a willingness to learn, a willingness to teach. +[1578.92 --> 1583.74] I mean, these are all like core factors of doing a good PR review. +[1584.22 --> 1586.00] And then you can get into the nitty gritties +[1586.00 --> 1590.82] of like your team's specific strategy and patterns and all of that. +[1590.82 --> 1592.06] And like the correctness of the code. +[1592.12 --> 1594.10] But I think that first part is really key. +[1594.88 --> 1595.90] Yeah, I totally agree. +[1596.00 --> 1600.26] And I think there are a lot of small behaviors +[1600.26 --> 1601.82] that we as engineers can adopt +[1601.82 --> 1604.98] to be more empathetic in our code reviews. +[1605.22 --> 1607.50] I really enjoy some of the resources +[1607.50 --> 1611.06] that other people have already written and shared around this. +[1611.18 --> 1613.84] I particularly like Alex Hill's The Art of Giving +[1613.84 --> 1617.22] and Receiving Code Reviews Gracefully. +[1617.34 --> 1618.06] It's something like that. +[1618.06 --> 1624.00] And that idea of giving a code review gracefully, +[1624.08 --> 1626.08] I think ties into the empathy thing +[1626.08 --> 1627.06] that you're talking about, Natasha. +[1627.34 --> 1629.94] And just like putting yourself in the position +[1629.94 --> 1631.86] of the person who's receiving the feedback. +[1632.24 --> 1634.56] Ultimately, a pull request, +[1634.70 --> 1637.18] like this code review is feedback +[1637.18 --> 1641.94] and is collaboration in like our everyday as engineers. +[1641.94 --> 1644.80] And so some of the simple things, +[1645.26 --> 1648.32] like instead of saying you, saying we. +[1648.56 --> 1651.64] And instead of making a statement, asking a question. +[1652.22 --> 1653.92] So for example, instead of saying like, +[1654.78 --> 1657.86] oh, like you should use this other function, +[1657.96 --> 1659.52] it already does what you're doing here. +[1660.62 --> 1664.08] Asking, oh, like can we use this other function here? +[1664.18 --> 1665.72] Like is there something that we can reuse? +[1665.72 --> 1670.64] And that gives the opportunity because you could be wrong. +[1670.72 --> 1672.34] I'm a staff engineer and I'm wrong all the time. +[1672.66 --> 1676.68] So if I'm just coming into a code review and saying like, +[1676.74 --> 1678.60] oh, you should have done this and you should have done that. +[1678.68 --> 1681.54] Like that not only is not giving feedback +[1681.54 --> 1683.38] in a way that will be easily received, +[1683.38 --> 1689.08] but it's really like assuming that I always know the answer +[1689.08 --> 1690.52] and I always know what's best. +[1690.52 --> 1694.50] And I think remembering that the person who's coming in +[1694.50 --> 1698.62] with the pull request has spent so much time, +[1698.78 --> 1701.26] presumably on this problem that yes, +[1701.30 --> 1702.88] the fresh pair of eyes is really helpful +[1702.88 --> 1704.98] and you might see something that they didn't, +[1705.28 --> 1708.22] but also giving acknowledgement to the work +[1708.22 --> 1710.30] that they've put in on this pull request +[1710.30 --> 1713.04] and that they might have like thought through that problem +[1713.04 --> 1715.50] and that there might be something that you're missing +[1715.50 --> 1717.16] because you haven't spent as much time +[1717.16 --> 1718.56] thinking about the solution. +[1719.32 --> 1720.50] So that empathy is really important. +[1720.52 --> 1721.20] Yeah. +[1721.50 --> 1722.28] Check your ego. +[1724.02 --> 1727.06] I would totally plus a thousand they've said so far. +[1727.24 --> 1728.14] It's definitely like for, +[1728.56 --> 1730.76] it's not just like it's a one side street, right? +[1730.78 --> 1731.24] It's two sides. +[1731.32 --> 1734.76] You can take something back from the review +[1734.76 --> 1736.10] and then they can take something forward. +[1736.46 --> 1738.48] It's always an opportunity to learn something new, +[1738.80 --> 1742.10] especially from people who have way more experience +[1742.10 --> 1742.60] than you do. +[1743.14 --> 1745.46] And it's just an opportunity to ask questions, +[1745.60 --> 1747.98] especially like, oh, why did you do this way +[1747.98 --> 1749.62] versus some other way, right? +[1749.62 --> 1752.72] And it's definitely a great opportunity +[1752.72 --> 1754.86] to kind of, you know, +[1754.86 --> 1757.52] just learn more about what they're working on +[1757.52 --> 1758.94] and learn more about the system. +[1759.48 --> 1760.92] There always has to be an eye for, +[1761.16 --> 1763.08] oh, is this like what we're trying to achieve +[1763.08 --> 1764.34] with this particular ticket? +[1764.88 --> 1766.60] But definitely leading with empathy +[1766.60 --> 1768.70] is something that's great. +[1768.70 --> 1770.46] I also want to look up that article now +[1770.46 --> 1771.56] because I've never read that, +[1772.26 --> 1773.48] the Alex Hill one, so. +[1774.12 --> 1774.84] It's really great. +[1775.02 --> 1776.84] I'll make sure I put it in the episode notes. +[1777.80 --> 1779.12] And it is a great article. +[1780.94 --> 1782.68] So when you're thinking about reviewing, +[1782.80 --> 1785.08] are you predominantly reviewing for functionality +[1785.08 --> 1787.06] and like, does this thing work? +[1787.32 --> 1789.86] Or are you also commenting slash what are the, +[1790.04 --> 1792.64] I guess, rules around commenting on like style? +[1793.04 --> 1794.66] Like how the code has been written, +[1794.82 --> 1796.82] the stylistic choices that have been made? +[1796.82 --> 1798.16] I think a bit of both. +[1798.80 --> 1800.00] Yeah, definitely. +[1800.44 --> 1802.84] And not to over-reference this article, +[1803.30 --> 1805.80] but one of the things I think is like +[1805.80 --> 1807.68] one of my key takeaways from this was +[1807.68 --> 1811.62] that there are those kinds of like code style things +[1811.62 --> 1813.64] that can be automated. +[1814.30 --> 1816.08] So having a team norm of like, +[1816.18 --> 1818.80] oh, we're all using this linter for this code base. +[1818.82 --> 1820.10] We're using this formatter. +[1820.66 --> 1821.76] Pre-commit hook, +[1821.84 --> 1824.44] those have to pass before you can make this pull request. +[1824.44 --> 1827.46] It just automates away a lot of the things +[1827.46 --> 1829.30] we can be kind of nitpicky over. +[1830.14 --> 1834.42] But sometimes there are more like code patterns +[1834.42 --> 1836.46] and those are harder to automate, +[1836.86 --> 1839.94] but they can be like a source of contention +[1839.94 --> 1842.18] because there can be a lot of very strong opinions +[1842.18 --> 1845.04] around how, what patterns we're following +[1845.04 --> 1846.42] and how code should be structured. +[1846.42 --> 1851.00] That's where I think having a set of norms on your team +[1851.00 --> 1855.42] that you regularly revisit when somebody new joins, +[1855.64 --> 1858.36] when you have a new, like a new repo you're working in, +[1858.72 --> 1862.28] those norms will help smooth that conversation +[1862.28 --> 1864.94] because if you have all already agreed, +[1865.10 --> 1866.64] like, oh, we're going to make sure +[1866.64 --> 1869.00] we follow dry practices or whatever, +[1869.20 --> 1871.76] like those kinds of agreements +[1871.76 --> 1876.60] make it a lot easier to have that code review conversation +[1876.60 --> 1880.22] because that's like a shared expectation that you have. +[1880.32 --> 1882.68] And it's the same thing for giving interpersonal feedback. +[1882.80 --> 1884.88] If you have a shared expectation and a shared goal, +[1885.10 --> 1887.30] then you can easily use that as a reference point +[1887.30 --> 1888.10] and be like, hey, like, +[1888.28 --> 1890.30] since this is something that we've agreed on as a team, +[1890.48 --> 1892.04] I'm noticing this here. +[1892.12 --> 1894.36] Do you think we could reshape this +[1894.36 --> 1895.56] so that it follows this practice +[1895.56 --> 1897.76] that we have agreed to use for this repo? +[1897.76 --> 1901.62] And that's a much easier conversation to have +[1901.62 --> 1904.48] than like, oh, I don't like how this is styled. +[1904.60 --> 1906.24] I think we should do it this way instead. +[1906.54 --> 1909.14] And kind of like bringing your perspective into it. +[1909.42 --> 1911.42] It's a lot easier to bring like a team norm +[1911.42 --> 1912.66] that you've already all agreed on +[1912.66 --> 1914.68] that's like the team's perspective on this +[1914.68 --> 1917.72] and the team's perspective on how to move forward. +[1918.46 --> 1919.28] It's really important, +[1919.34 --> 1920.82] like having that team understanding, right? +[1920.94 --> 1922.10] At least in my previous company, +[1922.10 --> 1924.88] we had like shared standards as to how things, +[1924.88 --> 1926.52] we had like pillars and everything, +[1926.52 --> 1928.30] that we had like a standards committee +[1928.30 --> 1929.80] that we were trying to do +[1929.80 --> 1932.14] for like code style and stuff like that. +[1932.70 --> 1934.86] But just like on the other side of the coin, +[1935.04 --> 1936.00] I'm kind of like, +[1936.44 --> 1938.72] sometimes if I see a spelling mistake, +[1938.72 --> 1940.14] I will point that out in the PR +[1940.14 --> 1941.98] just because it's like, +[1942.10 --> 1943.92] if it's already committed and I see it, +[1944.08 --> 1946.30] I wish someone had called it out in the PR +[1946.30 --> 1947.42] and I will put in like, +[1947.88 --> 1949.20] in my PR, I will fix it. +[1949.48 --> 1951.06] So it's just like those little tiny things +[1951.06 --> 1955.08] that kind of add to your personality as a reviewer. +[1955.08 --> 1956.92] Like you might be known as that person +[1956.92 --> 1959.72] that calls out your grammar mistakes, +[1959.72 --> 1962.54] which I feel like I used to be +[1962.54 --> 1965.02] and I'm not as much these days, +[1965.16 --> 1966.62] but when I see it, +[1966.70 --> 1968.24] I will sometimes call it out. +[1968.46 --> 1969.90] I think I've gotten that feedback +[1969.90 --> 1971.00] on my review from you. +[1971.36 --> 1974.68] So my PR can attest. +[1976.76 --> 1977.50] I'm intrigued. +[1977.62 --> 1978.98] Do you review differently +[1978.98 --> 1982.08] depending on the level of the person +[1982.08 --> 1983.74] that put in the PR? +[1984.06 --> 1985.36] I like, if you're going in +[1985.36 --> 1987.64] and you're reviewing like a staff engineer's PR +[1987.64 --> 1990.46] versus a associate engineer's PR, +[1990.66 --> 1991.68] do you approach it differently? +[1992.30 --> 1994.12] The reason I ask is that, +[1994.40 --> 1995.70] maybe this is a leading question, +[1996.22 --> 1997.58] say you're a staff engineer +[1997.58 --> 1998.94] and you're reviewing the PR +[1998.94 --> 2001.16] of someone you know is like preferably new. +[2001.54 --> 2004.14] Would you approach that slightly differently +[2004.14 --> 2005.98] in that you might add more comments, +[2005.98 --> 2008.56] maybe more detail as to why you've suggested it +[2008.56 --> 2010.20] because you know this person's still learning +[2010.20 --> 2012.88] or agnostic of level, +[2013.08 --> 2015.10] you approach every PR in the same way. +[2015.74 --> 2016.86] I like to summarize, +[2017.08 --> 2018.66] does it matter who put the PR in? +[2019.48 --> 2020.36] I think it could help. +[2020.56 --> 2022.16] Like if I know that someone +[2022.16 --> 2024.62] is more new to the company +[2024.62 --> 2025.54] or to the team, +[2025.58 --> 2028.98] I usually kind of pepper my review comments +[2028.98 --> 2031.42] with links or like context +[2031.42 --> 2034.24] or even ask like if they want to kind of like +[2034.24 --> 2035.10] jump on a call +[2035.10 --> 2036.74] and we can talk through certain things +[2036.74 --> 2038.92] if they have more like follow-up questions +[2038.92 --> 2040.70] just so it's a little bit more synchronous +[2040.70 --> 2041.98] conversation wise +[2041.98 --> 2044.08] instead of like kind of all over the place. +[2044.52 --> 2047.36] But I also try to leave like good feedback +[2047.36 --> 2050.02] even if like there's a person who's above me +[2050.02 --> 2052.58] and you know run laps on me +[2052.58 --> 2054.16] with like the work that they do. +[2054.48 --> 2056.56] I like to just say like this is great, +[2056.56 --> 2058.44] I learned a lot or you know praise +[2058.44 --> 2059.64] some of the work that they've done +[2059.64 --> 2061.32] because I think it can be easy +[2061.32 --> 2065.26] to just not get that type of feedback as often +[2065.26 --> 2066.42] because you're kind of expected +[2066.42 --> 2067.96] to do that kind of work. +[2068.58 --> 2072.40] So being the least senior person on the team, +[2072.72 --> 2075.40] it's difficult not to like feel intimidated +[2075.40 --> 2078.80] by other engineers when they ask for reviews +[2078.80 --> 2079.66] because you're like, +[2079.76 --> 2082.00] oh, I just I'm early in my career. +[2082.10 --> 2084.20] What do I have to offer to someone else +[2084.20 --> 2085.70] who's been working for like 10, 15 years? +[2085.70 --> 2089.36] But I try not to think of it as that as much anymore. +[2089.92 --> 2092.60] Try to think of it with my current understanding of things. +[2092.86 --> 2095.10] I try to give the best feedback I can to that person +[2095.10 --> 2097.60] because it's always nice to have a second pair of eyes +[2097.60 --> 2099.44] even if they're less experienced. +[2099.58 --> 2101.26] It's always nice to have like fresh pair +[2101.26 --> 2103.10] and then you might see something that you've missed +[2103.10 --> 2105.34] and take it as a learning opportunity +[2105.34 --> 2106.50] at the same time for myself. +[2106.98 --> 2107.34] Definitely. +[2107.96 --> 2109.04] It's a fresh pair of eyes, +[2109.16 --> 2112.02] but also like there might be some very, +[2112.12 --> 2113.04] very senior engineer +[2113.04 --> 2115.46] who's been doing the same thing for many, many years +[2115.46 --> 2117.24] and therefore has got into a habit +[2117.24 --> 2118.68] of doing things a certain way +[2118.68 --> 2122.42] and you fresh bunny rabbit that you are +[2122.42 --> 2125.38] coming in with all the new technological lingo +[2125.38 --> 2127.26] and new open source. +[2127.44 --> 2129.32] Like you might be way more engaged +[2129.32 --> 2131.56] and be way much more on top of the new technologies +[2131.56 --> 2132.82] and ways of doing things. +[2132.82 --> 2134.82] So you might be able to come in and be like, +[2134.98 --> 2136.38] hey, have you considered this new style +[2136.38 --> 2138.12] that you haven't done in 20 years, +[2138.12 --> 2139.30] but maybe it'll be useful +[2139.30 --> 2141.00] and you can teach them something. +[2141.56 --> 2141.78] Exactly. +[2141.92 --> 2143.64] And I usually, when I do make like, +[2143.82 --> 2144.86] oh, you should try it this way. +[2144.92 --> 2147.26] I usually have links to support what I'm saying +[2147.26 --> 2148.64] just to be like, +[2148.74 --> 2151.00] see, like these other people are doing it this way as well. +[2151.24 --> 2151.86] It's not just me. +[2152.70 --> 2152.90] Yeah. +[2152.94 --> 2156.06] And I think the value of a fresh perspective +[2156.06 --> 2158.60] also is in like challenging assumptions. +[2158.84 --> 2160.44] So I know as a, +[2160.88 --> 2161.86] like I'm a staff engineer, +[2161.86 --> 2163.34] but there are definitely things that, +[2163.82 --> 2164.44] to your point, Angelica, +[2164.54 --> 2165.54] like I've gotten used to +[2165.54 --> 2166.96] or I've gotten into the habit of. +[2166.96 --> 2169.92] And sometimes that does lead me to make assumptions. +[2169.92 --> 2171.98] And I, I work to check myself, +[2172.06 --> 2175.00] but having teammates being able to ask questions +[2175.00 --> 2175.90] and check assumptions, +[2175.90 --> 2179.30] I think leads to really valuable conversation +[2179.30 --> 2182.34] because maybe that will lead us in a different direction. +[2182.34 --> 2185.28] Maybe I can explain more about something +[2185.28 --> 2186.96] that I've done in my work. +[2187.32 --> 2189.62] But I actually think when I'm approaching reviews +[2189.62 --> 2192.04] and when I'm asking to be reviewed, +[2192.38 --> 2196.12] I think of it more as like the subject matter expertise +[2196.12 --> 2197.24] in a code base +[2197.24 --> 2199.96] because I recently switched teams +[2199.96 --> 2201.84] and I'm coming onto my new team +[2201.84 --> 2202.64] and I have a lot, +[2203.00 --> 2205.10] I have a lot of wealth of knowledge +[2205.10 --> 2206.12] in terms of architecture +[2206.12 --> 2208.32] and some of the things that I'm bringing from my past team, +[2208.50 --> 2211.40] but I'm actually learning some of these languages +[2211.40 --> 2213.14] for the first time. +[2213.44 --> 2217.36] And so even though I have a big picture idea +[2217.36 --> 2218.30] around our architecture +[2218.30 --> 2220.18] and I'm doing a lot of things +[2220.18 --> 2221.88] as the tech lead for my team, +[2222.12 --> 2224.14] some of my teammates actually know these languages +[2224.14 --> 2226.04] and the language patterns +[2226.04 --> 2229.42] that come with these repos better than I do. +[2230.12 --> 2232.64] And so I actively look for them +[2232.64 --> 2235.98] to be critical of the code that I'm putting forth +[2235.98 --> 2238.00] because that allows me to learn +[2238.00 --> 2241.82] the same way that if they were putting in a PR in Python +[2241.82 --> 2243.40] and I'm more expert in Python, +[2243.40 --> 2245.60] I would want to teach them some of the things +[2245.60 --> 2247.84] that I have picked up about Python along the way. +[2247.84 --> 2249.74] So I think it's, +[2249.88 --> 2251.52] I actually think this is an area +[2251.52 --> 2252.76] where I would, +[2253.10 --> 2254.42] I hope leveling doesn't matter. +[2254.58 --> 2256.84] It's more about like the subject matter expertise +[2256.84 --> 2257.38] in a repo +[2257.38 --> 2259.74] and like helping that person +[2259.74 --> 2261.66] level up their expertise +[2261.66 --> 2263.14] like a little bit further +[2263.14 --> 2266.30] and anybody can help anybody else +[2266.30 --> 2268.06] level up their expertise more. +[2268.62 --> 2270.72] It's interesting to hear from me all the answers. +[2270.94 --> 2273.00] I guess you're all US-based +[2273.00 --> 2276.64] and you all mostly work with American colleagues. +[2277.00 --> 2278.38] Is my assumption correct? +[2278.46 --> 2281.60] Or would you say your teams are kind of with also, +[2281.76 --> 2282.54] or you get to work, +[2282.68 --> 2284.36] let's say with people from other people +[2284.36 --> 2285.26] who are not Americans? +[2285.88 --> 2289.84] Yeah, I've worked with like non-native English speakers +[2289.84 --> 2291.38] before on my current team. +[2291.46 --> 2293.42] I'm working with mostly native English, +[2293.50 --> 2295.22] all the engineers are native English speakers, +[2295.22 --> 2296.98] but one of my teammates even +[2296.98 --> 2300.16] was recently working with some of our engineering teams +[2300.16 --> 2302.70] that are working out of other countries as well. +[2303.14 --> 2305.36] So we have some international work. +[2306.16 --> 2307.38] Would you say your experience +[2307.38 --> 2310.00] of feeling comfortable to correct +[2310.00 --> 2311.00] and to be corrected +[2311.00 --> 2313.08] is the same as with people +[2313.08 --> 2314.38] from the same background as you? +[2314.80 --> 2317.44] Or is it different in any way +[2317.44 --> 2319.64] when you like review the code +[2319.64 --> 2320.92] or get reviewed by people +[2320.92 --> 2322.84] who are wherever they're based, +[2322.92 --> 2324.80] like just grew up in other places? +[2325.56 --> 2326.72] Yeah, that's a really great question. +[2327.42 --> 2328.62] I definitely think that +[2328.62 --> 2331.66] I have an easier time giving grace +[2331.66 --> 2333.10] to somebody who's giving me a review +[2333.10 --> 2334.84] if English isn't their first language +[2334.84 --> 2339.24] because tone is a hard thing to pick up. +[2339.76 --> 2341.68] So both I've worked with +[2341.68 --> 2344.08] whether like English is not their first language +[2344.08 --> 2348.22] or whether somebody has issues with tone +[2348.22 --> 2349.18] for some other reason, +[2349.36 --> 2351.86] like some kind of like mental like illness +[2351.86 --> 2354.70] or disability is probably the better term +[2354.70 --> 2357.18] for thinking through +[2357.18 --> 2358.94] like how something is going to be perceived. +[2359.06 --> 2360.08] I've worked with colleagues +[2360.08 --> 2361.88] who are like on the spectrum +[2361.88 --> 2363.24] and that's not always like something +[2363.24 --> 2365.92] that they have an easy time interpreting +[2365.92 --> 2368.34] and figuring out how tone is going to come across. +[2368.84 --> 2370.10] So in those situations, +[2370.10 --> 2370.74] I have a lot, +[2371.54 --> 2373.14] it's easier for me to give empathy +[2373.14 --> 2374.06] to that person +[2374.06 --> 2375.06] when they're giving me a review, +[2375.18 --> 2375.46] being like, +[2375.56 --> 2377.02] oh, they didn't mean that to hurt my feelings. +[2377.02 --> 2378.24] They're just giving me feedback. +[2379.24 --> 2381.74] But sometimes because we are a big company, +[2382.38 --> 2383.38] like sometimes you're getting a review +[2383.38 --> 2384.80] from somebody you don't know. +[2384.94 --> 2386.80] And so you don't always know if that's the case. +[2387.16 --> 2388.58] So I think it's easier +[2388.58 --> 2389.76] when you're working with somebody you know. +[2390.38 --> 2391.96] If English isn't their first language, +[2392.16 --> 2393.60] I always have an easier time being like, +[2393.66 --> 2394.58] oh, they didn't mean it. +[2394.64 --> 2395.94] Like in a way that hurts my feelings. +[2397.62 --> 2398.88] Sometimes that's harder for me +[2398.88 --> 2400.34] to like get to that point +[2400.34 --> 2402.64] if it's somebody who isn't +[2402.64 --> 2403.86] like a native English speaker. +[2403.86 --> 2405.80] And it's one of those things +[2405.80 --> 2408.12] I think is easy to make assumptions around, +[2408.12 --> 2408.94] which is why I brought up +[2408.94 --> 2410.56] like working with somebody +[2410.56 --> 2411.90] who is like neurodivergent. +[2412.06 --> 2412.54] It's just like, +[2413.22 --> 2414.04] that's not always something +[2414.04 --> 2415.36] you can tell right away. +[2415.50 --> 2416.88] And so it's easier to tell +[2416.88 --> 2418.98] when English isn't somebody's first language. +[2419.42 --> 2421.64] But I try to like take that empathy +[2421.64 --> 2423.08] that I learned working with those colleagues +[2423.08 --> 2424.92] and bring it to all my other colleagues as well. +[2425.32 --> 2426.94] It's also interesting to think of that. +[2427.22 --> 2429.48] I mean, yes, English is not native language. +[2429.48 --> 2431.32] That's like a very good differentiator. +[2431.70 --> 2433.44] And also different cultures +[2433.44 --> 2436.68] have different relationship with feedback +[2436.68 --> 2438.02] or even saying somebody, +[2438.22 --> 2438.96] you did something wrong. +[2440.34 --> 2441.90] Yeah, that's a good point. +[2442.38 --> 2444.32] Being the only person who is not in the US, +[2444.72 --> 2446.06] I would actually love to hear +[2446.06 --> 2447.86] like your perception on this. +[2447.94 --> 2448.64] I know you keep on saying +[2448.64 --> 2449.76] we need to hear from the guests. +[2449.82 --> 2451.44] And I agree, these wonderful guests. +[2451.92 --> 2453.38] But I would love to hear your experience. +[2453.48 --> 2454.56] I mean, I got culture shock +[2454.56 --> 2455.52] moving to the US, +[2455.54 --> 2457.36] but I wasn't reviewing PRs in London. +[2457.36 --> 2458.50] And you are a native speaker. +[2458.98 --> 2460.12] Yes, and I am a native speaker. +[2460.36 --> 2461.66] So I have those two things +[2461.66 --> 2463.90] which make it slightly easier coming in. +[2463.90 --> 2466.18] But I would love to hear your perception +[2466.18 --> 2467.06] and kind of what you've heard +[2467.06 --> 2468.04] from colleagues, friends, +[2468.54 --> 2469.32] and your experience. +[2469.86 --> 2471.78] Almost all the teams that I worked in +[2471.78 --> 2472.88] are quite mixed +[2472.88 --> 2474.48] because Europe is a lot easier +[2474.48 --> 2476.86] for work relocations. +[2477.68 --> 2478.90] The green card and so on +[2478.90 --> 2480.10] exists in the US, +[2480.22 --> 2481.54] but it's quite harder +[2481.54 --> 2483.32] than just the work visa in Germany +[2483.32 --> 2484.52] or other European countries. +[2484.70 --> 2486.36] So I don't remember the last time +[2486.36 --> 2488.18] I worked in a team +[2488.18 --> 2489.66] where everybody comes from the same country. +[2489.86 --> 2491.84] Or we had like more than 10% +[2491.84 --> 2493.06] native English speakers. +[2493.06 --> 2495.16] Although English is always the main language +[2495.16 --> 2497.24] because you need to have something +[2497.24 --> 2498.58] that is not the programming language. +[2499.32 --> 2500.58] And I think in the beginning, +[2500.82 --> 2503.06] I used to have some kind of misunderstandings +[2503.06 --> 2505.68] exactly because what was said, +[2506.08 --> 2508.96] like understanding empathy, +[2509.16 --> 2510.70] understanding they did not mean that, +[2510.80 --> 2511.60] they meant this, +[2512.06 --> 2514.12] coming across myself +[2514.12 --> 2515.82] sometimes as a bit more, +[2516.32 --> 2517.92] why don't you say that in a nicer way? +[2518.04 --> 2518.54] And so on. +[2518.54 --> 2522.36] And so definitely there is all sorts of bounds +[2522.36 --> 2523.16] there to strike. +[2523.32 --> 2526.04] And it's very interesting to hear +[2526.04 --> 2527.90] how people cope with that. +[2527.96 --> 2529.76] And it's also obviously written +[2529.76 --> 2533.10] is even less easy to understand, +[2533.26 --> 2534.32] like written communication +[2534.32 --> 2535.46] in comparison to speaking, +[2535.94 --> 2536.92] to spoken communication. +[2537.10 --> 2537.86] And even in spoken, +[2537.98 --> 2540.00] you can easily get lost and so on. +[2540.58 --> 2542.72] And intonations mean different things. +[2542.72 --> 2545.08] It's the thing about cultures that say +[2545.08 --> 2547.92] that sometimes it's kind of impolite +[2547.92 --> 2548.64] either to correct +[2548.64 --> 2550.30] or even to just say no to someone. +[2550.80 --> 2553.84] So I had to learn and actively practice +[2553.84 --> 2556.26] that when I propose some idea, +[2556.36 --> 2557.90] I have to start with something like, +[2558.10 --> 2559.94] feel free to reject that +[2559.94 --> 2562.24] or there might be better alternatives, +[2562.44 --> 2562.64] but, +[2563.20 --> 2565.16] and sometimes it's still, +[2565.38 --> 2566.30] depending on the context, +[2566.30 --> 2568.26] it sounds sort of foreign, +[2568.26 --> 2569.30] but I think it's fine +[2569.30 --> 2570.74] that it sounds foreign +[2570.74 --> 2572.74] because it's kind of like a standard +[2572.74 --> 2575.06] across all the different cultures +[2575.06 --> 2576.20] you'll get to work with +[2576.20 --> 2577.62] that this is a, +[2578.12 --> 2581.16] be nicer over being not nice enough. +[2581.68 --> 2582.28] Yeah, definitely. +[2582.74 --> 2584.58] My comment is going to be very like, +[2585.12 --> 2586.20] less, slightly less serious, +[2586.32 --> 2587.00] but it came to mind +[2587.00 --> 2587.90] and it sounded like something +[2587.90 --> 2588.76] that I would love to do. +[2589.30 --> 2592.50] Can you like attach voice notes to PRs? +[2592.50 --> 2594.42] You have looms in Git pull requests. +[2594.90 --> 2595.22] Yeah, +[2595.22 --> 2597.08] because I feel like I would love it +[2597.08 --> 2598.74] if someone made like a suggestion +[2598.74 --> 2601.36] and then I could like have a little voice note +[2601.36 --> 2602.06] where they could say, +[2602.14 --> 2602.68] this is why, +[2602.80 --> 2604.00] because then you hear the tone +[2604.00 --> 2606.42] and you hear them explaining it +[2606.42 --> 2609.50] and then you could do like a verbal readout +[2609.50 --> 2610.74] of your very long PR. +[2611.42 --> 2612.44] That might be cool. +[2614.06 --> 2615.86] Do emojis help with that? +[2616.16 --> 2617.58] Like I find it really difficult +[2617.58 --> 2620.80] to communicate through my words +[2620.80 --> 2622.38] without emojis. +[2622.50 --> 2624.84] Like I use emojis to help me +[2624.84 --> 2626.44] more accurately portray +[2626.44 --> 2629.20] the intention behind my comments. +[2629.36 --> 2630.66] I, when I say something like, +[2631.04 --> 2632.54] oh, I'm not sure about this. +[2632.64 --> 2634.12] I'll do like a thinky emoji +[2634.12 --> 2636.40] and like a funny like tongue emoji +[2636.40 --> 2637.56] to show that it's like, +[2637.94 --> 2639.34] not like a, what is this? +[2639.88 --> 2641.98] Because I feel like words are not enough, +[2642.04 --> 2642.78] which is why, +[2642.92 --> 2644.52] and those of you who interact with me regularly, +[2645.10 --> 2645.76] I think you'll be, +[2645.88 --> 2647.90] it'll be very hard for you to find any space +[2647.90 --> 2650.18] where I haven't put an emoji after my message +[2650.18 --> 2652.38] or some kind of imagery, +[2652.38 --> 2653.80] whether it be a meme or a GIF +[2653.80 --> 2657.50] to try and kind of level up my communication +[2657.50 --> 2658.58] from just being words +[2658.58 --> 2660.90] to having that extra layer of like emotion. +[2661.40 --> 2666.46] Is it appropriate to add many emojis to your PR? +[2666.58 --> 2668.74] I know that you can like add the emoji reactions, +[2668.98 --> 2669.64] but within it, +[2669.92 --> 2672.42] is that a useful tool that in fact could be used +[2672.42 --> 2674.64] to try and alleviate that risk +[2674.64 --> 2677.20] of being misconstrued via just written word? +[2677.20 --> 2679.52] So you mean the pull request description +[2679.52 --> 2680.70] or actually each commit? +[2682.32 --> 2684.42] I mean, open to views on either. +[2685.00 --> 2686.72] Yeah, I recommend using emojis +[2686.72 --> 2690.34] and like setting up norms around emoji use +[2690.34 --> 2692.06] and like code review norms. +[2692.34 --> 2693.58] It's like one of the suggestions +[2693.58 --> 2696.12] that some of the other staff engineers +[2696.12 --> 2700.06] and I put together in the norms template +[2700.06 --> 2702.54] that we shared out with internally at the times. +[2702.54 --> 2706.24] And I think it's the way that we use emojis +[2706.24 --> 2709.02] in that template is around communicating +[2709.02 --> 2711.16] the intention behind a review comment. +[2711.96 --> 2715.40] So one of the things that is really hard to tell +[2715.40 --> 2718.78] based on just the written word in terms of tone +[2718.78 --> 2721.28] is like whether that piece of feedback +[2721.28 --> 2722.62] is blocking or not. +[2723.28 --> 2726.26] So is this just a stylistic thing +[2726.26 --> 2729.48] that you think will help me level up my skills? +[2729.48 --> 2732.50] Is this something that I actually need to do +[2732.50 --> 2734.42] before merging in this code? +[2735.16 --> 2737.78] That kind of communication can sometimes be hard +[2737.78 --> 2742.20] and especially I think within different levels. +[2742.34 --> 2743.78] Like I know when I was more entry level +[2743.78 --> 2746.20] I had a hard time just like coming out +[2746.20 --> 2746.76] and asking like, +[2746.86 --> 2747.98] do I actually have to do that though? +[2749.00 --> 2752.28] So we use emojis as a way +[2752.28 --> 2754.26] of kind of categorizing the comment. +[2754.64 --> 2756.14] So if it is blocking, +[2756.68 --> 2758.16] you can communicate that being like, +[2758.16 --> 2760.90] oh, like this will cause implications in this way. +[2761.46 --> 2762.06] And you kind of, +[2762.34 --> 2764.10] it's a good prompt for the reviewer +[2764.10 --> 2765.46] to think about why it's blocking +[2765.46 --> 2766.84] and communicate that. +[2767.22 --> 2770.12] But it's also ways of sharing +[2770.12 --> 2771.52] like stylistic feedback +[2771.52 --> 2772.72] in a way that's not blocking +[2772.72 --> 2773.18] and say like, +[2773.26 --> 2775.78] oh, like I see that you did it this way. +[2775.86 --> 2777.34] Like typically when I approach this problem, +[2777.38 --> 2779.04] I do it this other way. +[2779.70 --> 2782.20] Just different styles, sharing my style. +[2783.04 --> 2784.20] So I think it helps +[2784.20 --> 2785.98] to kind of categorize the intention +[2785.98 --> 2786.98] to your point Angelica +[2786.98 --> 2789.94] around what the comment +[2789.94 --> 2791.88] is supposed to be communicating. +[2792.06 --> 2794.00] It like adds a little bit of color, +[2794.10 --> 2796.48] I think to just the plain text otherwise. +[2797.24 --> 2798.56] But curious if others +[2798.56 --> 2799.66] have different opinions on it. +[2799.76 --> 2801.16] Maybe you find them annoying. +[2802.34 --> 2803.74] Follow up to that. +[2803.92 --> 2806.30] So I'm 100% pro emoji +[2806.30 --> 2808.94] as people on my team can attest to. +[2809.06 --> 2811.26] I am always using emojis in Slack +[2811.26 --> 2813.02] sometimes in GitHub. +[2813.02 --> 2815.42] But I guess for that, +[2815.96 --> 2817.34] if it's kind of meant to convey +[2817.34 --> 2818.60] blocking versus not blocking +[2818.60 --> 2819.96] as your example, +[2820.76 --> 2822.08] is that like a way to kind of +[2822.08 --> 2824.70] not have to use the actual +[2824.70 --> 2825.92] like supported GitHub feature +[2825.92 --> 2827.96] where it's like you want, +[2828.04 --> 2829.02] you're requesting changes +[2829.02 --> 2831.44] and kind of blocking the merge +[2831.44 --> 2832.76] like full stop? +[2832.88 --> 2833.90] Because I don't know +[2833.90 --> 2834.92] how people perceive that, +[2834.96 --> 2835.94] but maybe it could come off +[2835.94 --> 2836.54] a little harsh, +[2836.78 --> 2838.70] like actually using that feature +[2838.70 --> 2839.76] and maybe like emojis +[2839.76 --> 2841.56] isn't a nicer way to say that. +[2841.64 --> 2843.50] But like if it's really blocking, +[2843.84 --> 2844.98] shouldn't we just be using +[2844.98 --> 2846.34] that specific feature +[2846.34 --> 2847.32] to prevent it +[2847.32 --> 2848.18] from actually going through? +[2848.48 --> 2849.92] I think in that scenario, +[2849.92 --> 2852.22] like you still do request changes +[2852.22 --> 2854.16] and kind of formally block +[2854.16 --> 2854.98] the pull request. +[2855.20 --> 2856.72] But I know as a reviewer, +[2856.78 --> 2858.36] if I get like changes requested +[2858.36 --> 2858.98] on my PR +[2858.98 --> 2861.08] and I go in and there are like 20 comments, +[2861.34 --> 2863.86] it's hard for me to maybe initially +[2863.86 --> 2865.34] like sift through those comments +[2865.34 --> 2866.40] to figure out which ones +[2866.40 --> 2867.66] are the ones that are causing +[2867.66 --> 2869.20] the PR to be blocked. +[2869.54 --> 2870.74] So it helps like streamline +[2870.74 --> 2872.08] the communication in that way +[2872.08 --> 2873.90] where if I'm blocking +[2873.90 --> 2875.36] Angelica's pull request +[2875.36 --> 2877.28] and I leave a bunch of comments +[2877.28 --> 2878.16] and some of them are like, +[2878.34 --> 2879.16] great job, you, +[2879.32 --> 2880.12] this is really cool. +[2880.28 --> 2881.86] And some of them are style +[2881.86 --> 2884.14] and some of them are actually questions +[2884.14 --> 2886.50] and then some of them are actual blocking +[2886.50 --> 2887.80] like requests for changes. +[2887.80 --> 2890.38] It helps kind of narrow to those +[2890.38 --> 2891.98] and like create more of, +[2892.14 --> 2894.92] I think like a clear checklist +[2894.92 --> 2896.14] in a way of like what, +[2896.40 --> 2897.74] you're expecting to be changed +[2897.74 --> 2899.10] before being asked +[2899.10 --> 2900.12] to re-review the code. +[2900.50 --> 2901.58] So I just think it's helpful +[2901.58 --> 2903.46] even if you are formally +[2903.46 --> 2904.90] like blocking the pull request +[2904.90 --> 2908.38] to communicate like what things +[2908.38 --> 2909.34] you expect to be changed +[2909.34 --> 2911.76] before you think that it could be merged. +[2912.42 --> 2912.52] Gotcha. +[2912.62 --> 2913.68] I don't know if I was just projecting +[2913.68 --> 2914.54] my own fear, +[2914.64 --> 2915.28] like seeing the, +[2915.38 --> 2916.78] like the email where it says +[2916.78 --> 2917.86] this has been like, +[2918.06 --> 2918.98] changes have been requested +[2918.98 --> 2920.10] and oh no, +[2920.16 --> 2921.10] what did I do wrong? +[2921.42 --> 2922.40] I need to do that. +[2922.98 --> 2923.54] Don't worry. +[2923.60 --> 2924.68] We all do stuff wrong +[2924.68 --> 2925.94] in our pull requests all the time. +[2926.40 --> 2927.60] That's part of being an engineer. +[2927.60 --> 2956.08] This episode is brought to you by Honeycomb. +[2956.08 --> 2957.48] Don't find your most perplexing +[2957.48 --> 2958.56] application issues. +[2958.86 --> 2961.68] Honeycomb is a fast analysis tool +[2961.68 --> 2962.58] that reveals the truth +[2962.58 --> 2963.88] about every aspect +[2963.88 --> 2965.78] of your application in production. +[2966.20 --> 2967.26] Find out how users experience +[2967.26 --> 2968.54] your code in complex +[2968.54 --> 2970.24] and unpredictable environments. +[2970.50 --> 2971.28] Find patterns +[2971.28 --> 2972.28] and outliers +[2972.28 --> 2973.90] across billions of rows of data +[2973.90 --> 2975.50] and definitively solve your problems. +[2975.86 --> 2976.62] And we use Honeycomb +[2976.62 --> 2977.36] here at Change. +[2977.40 --> 2977.48] Well, +[2977.48 --> 2978.02] that's why we welcome +[2978.02 --> 2979.42] the opportunity to add them +[2979.42 --> 2981.22] as one of our infrastructure partners. +[2981.22 --> 2982.22] In particular, +[2982.22 --> 2983.12] we use Honeycomb +[2983.12 --> 2984.86] to track down CDN issues recently, +[2984.86 --> 2986.60] which we talked about at length +[2986.60 --> 2987.84] on the Kaizen edition +[2987.84 --> 2989.06] of the Ship It podcast. +[2989.32 --> 2989.98] So check that out. +[2990.24 --> 2990.72] Here's the thing. +[2990.96 --> 2992.18] Teams who don't use Honeycomb +[2992.18 --> 2993.32] are forced to find +[2993.32 --> 2994.20] the needle in the haystack. +[2994.32 --> 2995.34] They scroll through +[2995.34 --> 2996.44] endless dashboards +[2996.44 --> 2997.48] playing whack-a-mole. +[2997.70 --> 2998.88] They deal with alert floods, +[2999.10 --> 2999.70] trying to guess +[2999.70 --> 3000.74] which one matters. +[3001.12 --> 3002.06] And they go from tool +[3002.06 --> 3002.92] to tool to tool +[3002.92 --> 3003.74] playing sleuth, +[3004.00 --> 3004.60] trying to figure out +[3004.60 --> 3005.72] how all the puzzle pieces +[3005.72 --> 3006.34] fit together. +[3006.68 --> 3007.94] It's this context switching +[3007.94 --> 3009.04] and tool sprawl +[3009.04 --> 3009.94] that are slowly killing +[3009.94 --> 3011.00] teams effectiveness +[3011.00 --> 3012.36] and ultimately hindering +[3012.36 --> 3013.00] their business. +[3013.40 --> 3013.90] With Honeycomb, +[3014.00 --> 3014.98] you get a fast, +[3015.30 --> 3015.84] unified, +[3016.12 --> 3017.62] and clear understanding +[3017.62 --> 3019.02] of the one thing +[3019.02 --> 3020.16] driving your business. +[3020.42 --> 3020.84] Production. +[3021.36 --> 3021.94] With Honeycomb, +[3022.04 --> 3022.90] you guess less +[3022.90 --> 3023.84] and you know more. +[3024.22 --> 3024.90] Join the swarm +[3024.90 --> 3026.70] and try Honeycomb free today +[3026.70 --> 3028.30] at honeycomb.io +[3028.30 --> 3029.44] slash changelog. +[3029.74 --> 3030.14] Again, +[3030.28 --> 3031.38] honeycomb.io +[3031.38 --> 3033.06] slash changelog. +[3033.28 --> 3034.26] And by our friends +[3034.26 --> 3034.98] at Chronosphere, +[3035.46 --> 3036.18] scaling cloud native +[3036.18 --> 3037.08] is complicated +[3037.08 --> 3038.42] and Chronosphere helps teams +[3038.42 --> 3039.20] take back control +[3039.20 --> 3040.36] of observability, +[3040.76 --> 3041.74] tame rampant data growth, +[3042.04 --> 3043.18] reduce cloud native complexity, +[3043.64 --> 3044.48] and increase confidence +[3044.48 --> 3045.36] of the business. +[3045.82 --> 3046.68] And I'm here with Mark Mow, +[3046.76 --> 3047.86] co-founder and CEO +[3047.86 --> 3048.38] of Chronosphere. +[3048.86 --> 3049.38] Mark, when it comes +[3049.38 --> 3050.62] to cloud native observability, +[3051.02 --> 3052.10] what are the pain points +[3052.10 --> 3052.76] of Kubernetes +[3052.76 --> 3053.96] and making sure +[3053.96 --> 3054.68] it's reliable? +[3055.20 --> 3055.34] You know, +[3055.36 --> 3056.02] I think the shift +[3056.02 --> 3056.52] to Kubernetes +[3056.52 --> 3057.78] has really changed +[3057.78 --> 3059.94] the way we design applications. +[3060.26 --> 3061.80] It's changed the way we, +[3062.04 --> 3063.26] it's changed our infrastructure +[3063.26 --> 3063.74] as well. +[3063.78 --> 3064.22] So it's introduced +[3064.22 --> 3065.02] a lot of change, +[3065.08 --> 3065.48] I would say, +[3065.54 --> 3066.10] and that's probably +[3066.10 --> 3067.02] why it's causing +[3067.02 --> 3068.28] a lot of issues +[3068.28 --> 3069.64] in the observability space. +[3069.96 --> 3070.56] I think one thing +[3070.56 --> 3071.10] we're finding +[3071.10 --> 3073.46] is that a lot of companies +[3073.46 --> 3073.90] out there +[3073.90 --> 3074.98] are focused on +[3074.98 --> 3076.70] producing a lot more data +[3076.70 --> 3077.52] and there's a lot of focus +[3077.52 --> 3078.78] on more metrics, +[3078.94 --> 3079.48] more traces, +[3079.64 --> 3080.28] more logs, +[3080.40 --> 3081.52] because these environments +[3081.52 --> 3082.70] we're trying to monitor +[3082.70 --> 3083.80] are far more complex +[3083.80 --> 3084.44] these days. +[3084.56 --> 3085.38] I think that's maybe +[3085.38 --> 3086.24] one of the mistakes +[3086.24 --> 3087.58] the industry is running into +[3087.58 --> 3088.42] and it's interesting +[3088.42 --> 3089.54] because obviously +[3089.54 --> 3090.68] for all the solutions +[3090.68 --> 3091.00] out there, +[3091.02 --> 3091.80] the vendors out there, +[3091.90 --> 3092.74] the more data +[3092.74 --> 3093.48] that gets produced, +[3093.56 --> 3094.36] the better it is +[3094.36 --> 3095.38] for all the vendors +[3095.38 --> 3095.88] out there. +[3095.88 --> 3097.26] But what's interesting +[3097.26 --> 3098.28] is that along with +[3098.28 --> 3099.64] that increased volume +[3099.64 --> 3100.10] of data, +[3100.30 --> 3101.02] people aren't actually +[3101.02 --> 3102.50] getting better outcomes +[3102.50 --> 3103.08] out of it. +[3103.40 --> 3104.50] People's number +[3104.50 --> 3104.98] of incidents +[3104.98 --> 3105.70] that people are running +[3105.70 --> 3106.76] to are still rising, +[3107.28 --> 3108.30] people's MTTRs, +[3108.36 --> 3109.12] MTTDs, +[3109.18 --> 3110.26] meantime to detection +[3110.26 --> 3111.00] and resolutions +[3111.00 --> 3112.04] actually getting higher +[3112.04 --> 3113.08] as opposed to lower. +[3113.24 --> 3114.06] So I think this is +[3114.06 --> 3114.82] the common state +[3114.82 --> 3115.88] that a lot of companies +[3115.88 --> 3116.76] find themselves in +[3116.76 --> 3117.42] and of course +[3117.42 --> 3118.06] with the increased +[3118.06 --> 3119.20] volume of data, +[3119.52 --> 3120.68] folks' bills increase +[3120.68 --> 3121.36] and the problem +[3121.36 --> 3122.28] actually gets harder. +[3122.48 --> 3122.72] So I think +[3122.72 --> 3123.96] that's a common state +[3123.96 --> 3124.62] we find a lot +[3124.62 --> 3125.26] of companies into +[3125.26 --> 3125.84] and this is probably +[3125.84 --> 3126.88] why it's top of mind +[3126.88 --> 3127.58] for a lot of companies +[3127.58 --> 3128.00] out there. +[3128.44 --> 3128.74] Very cool. +[3128.82 --> 3129.34] Thank you, Martin. +[3129.46 --> 3129.72] All right, +[3129.74 --> 3130.32] the next step +[3130.32 --> 3131.12] is to head to +[3131.12 --> 3132.10] chronosphere.io +[3132.10 --> 3133.28] to explore the platform +[3133.28 --> 3134.32] and get a demo. +[3134.54 --> 3134.88] Again, +[3135.00 --> 3136.10] chronosphere.io. +[3136.10 --> 3164.30] So I just have +[3164.30 --> 3165.28] one more question. +[3165.28 --> 3166.82] because we are running +[3166.82 --> 3168.04] out of time regrettably +[3168.04 --> 3170.60] but my question is +[3170.60 --> 3172.42] like can you teach people +[3172.42 --> 3174.72] how to do a good PR, +[3174.94 --> 3176.34] how to review PRs? +[3176.88 --> 3178.40] Like is there truly an art +[3178.40 --> 3180.36] to reviewing a PR +[3180.36 --> 3181.54] or is it really just us +[3181.54 --> 3183.08] all like bumbling through, +[3183.72 --> 3184.88] learning as we go, +[3185.28 --> 3186.92] trying to do the best we can +[3186.92 --> 3188.62] and really you're never +[3188.62 --> 3190.48] the kind of PR pro. +[3190.82 --> 3191.70] You're always going to have +[3191.70 --> 3192.42] more to learn about +[3192.42 --> 3194.34] how to create a PR better, +[3194.48 --> 3195.46] review a PR better. +[3195.86 --> 3196.38] Like are you ever +[3196.38 --> 3196.88] going to be like, +[3196.96 --> 3197.54] right, I'm done. +[3197.66 --> 3198.76] I am the PR pro. +[3199.38 --> 3199.74] Amazing. +[3199.74 --> 3202.82] I think there is an art to it. +[3203.00 --> 3203.96] You're kind of balancing +[3203.96 --> 3205.52] the empathy +[3205.52 --> 3207.62] of giving this person feedback +[3207.62 --> 3209.24] and acknowledging +[3209.24 --> 3209.90] how much work +[3209.90 --> 3211.00] they've put into it +[3211.00 --> 3211.92] but at the same time +[3211.92 --> 3212.50] you have to, +[3212.74 --> 3212.90] you know, +[3212.94 --> 3213.50] you kind of, +[3213.94 --> 3214.50] if you're like, +[3215.12 --> 3216.30] if this is like a critical service, +[3216.42 --> 3217.42] you're going to be the one +[3217.42 --> 3217.80] that's, +[3218.28 --> 3219.04] that might be paged +[3219.04 --> 3219.94] during the middle of the night +[3219.94 --> 3221.06] if something goes wrong +[3221.06 --> 3221.88] and all you see that, +[3222.10 --> 3222.52] this commit +[3222.52 --> 3224.66] is what's causing the issue. +[3225.18 --> 3226.20] But also you, +[3226.26 --> 3227.86] you kind of have to balance +[3227.86 --> 3229.60] that with the time, +[3229.72 --> 3229.90] you know, +[3229.94 --> 3231.56] you still have to get things done. +[3232.00 --> 3233.20] It's a big balancing act +[3233.20 --> 3234.12] of how much, +[3234.46 --> 3234.68] like, +[3235.12 --> 3235.98] is something like, +[3236.16 --> 3236.86] is this critical +[3236.86 --> 3238.20] that this needs to get changed +[3238.20 --> 3238.86] even though it's like +[3238.86 --> 3239.82] a stylistic thing +[3239.82 --> 3241.02] that goes against +[3241.02 --> 3242.70] like our team norms +[3242.70 --> 3244.32] but it's imperative +[3244.32 --> 3245.70] that this gets through +[3245.70 --> 3247.34] or maybe they've already +[3247.34 --> 3248.34] spent a lot of time on it. +[3248.66 --> 3249.68] So it's definitely like +[3249.68 --> 3251.80] the art of like juggling +[3251.80 --> 3252.94] or balancing, +[3253.16 --> 3253.30] so. +[3253.90 --> 3255.22] I think just experience +[3255.22 --> 3256.32] goes a long way. +[3256.62 --> 3257.60] Like learn by doing. +[3257.84 --> 3258.66] You can kind of +[3258.66 --> 3260.32] experience a lot of things +[3260.32 --> 3261.12] that you want to emulate +[3261.12 --> 3262.64] or things that you're like, +[3262.70 --> 3262.82] wow, +[3262.88 --> 3263.46] that didn't make, +[3263.84 --> 3264.54] leave me feeling +[3264.54 --> 3265.76] really good about myself +[3265.76 --> 3267.36] so I know not to do this +[3267.36 --> 3268.02] in the future. +[3268.28 --> 3269.56] But also like +[3269.56 --> 3271.46] just getting more experienced +[3271.46 --> 3273.34] in whatever language +[3273.34 --> 3274.54] that you're working on +[3274.54 --> 3276.36] to then make those +[3276.36 --> 3277.34] suggestions +[3277.34 --> 3278.98] that can make the code +[3278.98 --> 3279.58] better +[3279.58 --> 3281.08] and more performant. +[3281.18 --> 3282.18] That's like another layer +[3282.18 --> 3283.88] of doing a PR review +[3283.88 --> 3285.30] but still something +[3285.30 --> 3286.58] that comes with more time. +[3287.12 --> 3287.22] Yeah, +[3287.30 --> 3287.76] and I think +[3287.76 --> 3288.60] from the perspective +[3288.60 --> 3289.96] like authoring a PR +[3289.96 --> 3291.08] and asking for a review +[3291.08 --> 3291.78] that's also +[3291.78 --> 3293.46] something that +[3293.46 --> 3294.52] is like a skill +[3294.52 --> 3295.90] that can be honed +[3295.90 --> 3296.38] and taught +[3296.38 --> 3297.62] and improved +[3297.62 --> 3298.50] at any level +[3298.50 --> 3299.06] because +[3299.06 --> 3300.58] figuring out like +[3300.58 --> 3302.12] when a PR is done, +[3302.26 --> 3303.26] how big it is, +[3303.62 --> 3305.20] but also in some situations +[3305.20 --> 3306.96] maybe you're introducing +[3306.96 --> 3307.94] a change +[3307.94 --> 3309.80] and maybe you have +[3309.80 --> 3310.44] to like write up +[3310.44 --> 3311.10] like why +[3311.10 --> 3312.48] you think that change +[3312.48 --> 3313.18] is the right path +[3313.18 --> 3313.80] to move forward +[3313.80 --> 3314.56] and there's always +[3314.56 --> 3316.16] room to improve on +[3316.16 --> 3317.10] like making +[3317.10 --> 3318.86] a concise argument. +[3318.86 --> 3319.86] I think that's +[3319.86 --> 3320.84] something that +[3320.84 --> 3321.98] we can always improve on +[3321.98 --> 3322.52] is how to make +[3322.52 --> 3323.16] a concise, +[3323.32 --> 3323.80] clear argument +[3323.80 --> 3324.78] for the change +[3324.78 --> 3326.12] that you're introducing +[3326.12 --> 3327.62] or how to best +[3327.62 --> 3328.12] like walk +[3328.12 --> 3329.00] your reviewer +[3329.00 --> 3330.40] through the pull request +[3330.40 --> 3332.04] and just based +[3332.04 --> 3333.84] on your PR ask. +[3333.84 --> 3335.16] There's a lot of skills +[3335.16 --> 3336.20] that can be honed +[3336.20 --> 3337.24] in this whole process +[3337.24 --> 3337.98] on both sides. +[3338.38 --> 3339.20] Well, thank you. +[3339.86 --> 3340.34] Regrettably, +[3340.48 --> 3341.08] we've babbled +[3341.08 --> 3342.22] all our time away +[3342.22 --> 3343.50] so we are going +[3343.50 --> 3344.50] to move into +[3344.50 --> 3346.06] our unpopular +[3346.06 --> 3347.42] opinion time. +[3347.42 --> 3365.16] Right, so +[3365.16 --> 3366.72] get ready guests +[3366.72 --> 3367.34] we're going to jump +[3367.34 --> 3368.16] right on in. +[3368.66 --> 3369.04] Jeff, +[3369.44 --> 3370.38] what is your +[3370.38 --> 3371.46] unpopular opinion? +[3372.10 --> 3372.48] I don't know +[3372.48 --> 3373.12] how I'm going to top +[3373.12 --> 3374.64] my last unpopular +[3374.64 --> 3375.06] opinion +[3375.06 --> 3375.88] but +[3375.88 --> 3376.84] I don't know +[3376.84 --> 3377.20] I feel like +[3377.20 --> 3378.14] I'll just spark +[3378.14 --> 3378.46] a, +[3378.72 --> 3379.52] it's not a big +[3379.52 --> 3380.22] unpopular opinion +[3380.22 --> 3380.64] but it's like +[3380.64 --> 3381.44] it'll start +[3381.44 --> 3382.16] like a war +[3382.16 --> 3383.12] kind of like +[3383.12 --> 3383.48] you know +[3383.48 --> 3384.22] PS3 +[3384.22 --> 3384.56] or like +[3384.56 --> 3385.06] PlayStation +[3385.06 --> 3385.74] versus Xbox +[3385.74 --> 3386.44] I'm going to go +[3386.44 --> 3387.28] with like +[3387.28 --> 3388.18] dogs are the +[3388.18 --> 3388.68] better pet +[3388.68 --> 3389.54] than cats. +[3389.86 --> 3390.16] I know there's +[3390.16 --> 3391.06] lots of +[3391.06 --> 3391.58] cat people. +[3392.18 --> 3393.70] I mean +[3393.70 --> 3394.42] so it'll be +[3394.42 --> 3395.02] the unpopular +[3395.02 --> 3396.00] opinion in this +[3396.00 --> 3396.58] group but +[3396.58 --> 3397.72] I have three +[3397.72 --> 3398.10] dogs. +[3398.58 --> 3398.80] Okay. +[3399.60 --> 3400.14] So fighting +[3400.14 --> 3400.46] words. +[3401.40 --> 3401.94] Just wait +[3401.94 --> 3402.38] for your next +[3402.38 --> 3402.70] PR. +[3403.06 --> 3403.26] Yeah, +[3403.32 --> 3405.30] they're fighting +[3405.30 --> 3405.62] words, +[3405.70 --> 3405.88] you know, +[3405.92 --> 3406.22] I know. +[3406.22 --> 3407.58] I got your +[3407.58 --> 3407.84] back, +[3407.96 --> 3408.12] Jeff. +[3408.22 --> 3408.58] I'm team +[3408.58 --> 3408.88] dog. +[3409.24 --> 3409.66] Team dog +[3409.66 --> 3410.06] all the way, +[3410.14 --> 3410.32] yeah. +[3412.42 --> 3412.86] Sarah, +[3413.12 --> 3413.76] what is your +[3413.76 --> 3414.28] unpopular +[3414.28 --> 3414.76] opinion? +[3415.46 --> 3415.70] Yeah, +[3415.72 --> 3416.00] I'll be +[3416.00 --> 3416.54] interested to +[3416.54 --> 3417.32] see how +[3417.32 --> 3417.80] unpopular +[3417.80 --> 3418.48] this is or +[3418.48 --> 3419.02] not but +[3419.02 --> 3419.78] I think +[3419.78 --> 3420.92] aspiring +[3420.92 --> 3421.36] software +[3421.36 --> 3421.82] engineers +[3421.82 --> 3422.20] would be +[3422.20 --> 3422.64] better off +[3422.64 --> 3422.98] taking +[3422.98 --> 3424.04] more +[3424.04 --> 3425.52] writing and +[3425.52 --> 3425.94] philosophy +[3425.94 --> 3426.78] courses and +[3426.78 --> 3427.40] fewer +[3427.40 --> 3428.52] computer science +[3428.52 --> 3429.26] theory courses. +[3429.26 --> 3430.60] yes. +[3432.16 --> 3432.72] Interesting. +[3433.22 --> 3433.58] No, +[3433.60 --> 3433.78] that's +[3433.78 --> 3434.16] intriguing. +[3434.94 --> 3435.42] Why is +[3435.42 --> 3435.62] that? +[3435.68 --> 3435.92] Can we +[3435.92 --> 3436.36] dig one +[3436.36 --> 3437.00] level deeper? +[3437.74 --> 3437.92] Yeah, +[3437.94 --> 3438.58] I think a +[3438.58 --> 3439.20] lot about +[3439.20 --> 3439.64] being a +[3439.64 --> 3439.98] successful +[3439.98 --> 3440.46] software +[3440.46 --> 3441.28] engineer is +[3441.28 --> 3441.84] kind of this +[3441.84 --> 3442.76] ability to +[3442.76 --> 3443.66] be able to +[3443.66 --> 3444.88] make a +[3444.88 --> 3445.26] concise +[3445.26 --> 3445.66] argument, +[3446.04 --> 3446.46] be able to +[3446.46 --> 3447.10] understand other +[3447.10 --> 3447.74] arguments and +[3447.74 --> 3448.44] perspectives and +[3448.44 --> 3449.30] incorporate it +[3449.30 --> 3451.18] and use that +[3451.18 --> 3452.58] to kind of +[3452.58 --> 3453.32] revise your +[3453.32 --> 3454.18] opinion and put +[3454.18 --> 3454.74] forth another +[3454.74 --> 3455.42] concise argument. +[3455.54 --> 3456.06] We see this in +[3456.06 --> 3457.00] architecture documents +[3457.00 --> 3457.60] all the time. +[3458.14 --> 3458.90] I was on the +[3458.90 --> 3459.58] architecture review +[3459.58 --> 3460.00] board at the +[3460.00 --> 3460.90] Times for a +[3460.90 --> 3461.40] long time and +[3461.40 --> 3461.86] I chaired it +[3461.86 --> 3462.36] for a while. +[3462.52 --> 3463.24] I know Natasha +[3463.24 --> 3463.82] has been on the +[3463.82 --> 3464.40] architecture review +[3464.40 --> 3465.20] board as well and +[3465.20 --> 3466.60] we see so many +[3466.60 --> 3468.36] long, long +[3468.36 --> 3469.40] documents that +[3469.40 --> 3470.32] could be half the +[3470.32 --> 3470.90] size that they +[3470.90 --> 3472.08] are and I +[3472.08 --> 3472.88] think a lot of +[3472.88 --> 3474.70] engineers don't +[3474.70 --> 3476.18] use the +[3476.18 --> 3476.94] theory classes +[3476.94 --> 3477.78] that they took +[3477.78 --> 3479.10] in college or +[3479.10 --> 3480.52] assuming that +[3480.52 --> 3481.80] if you are +[3481.80 --> 3482.78] studying computer +[3482.78 --> 3483.16] science in +[3483.16 --> 3483.56] college, I +[3483.56 --> 3484.00] think a lot of +[3484.00 --> 3484.60] those classes are +[3484.60 --> 3485.80] not put to +[3485.80 --> 3486.26] use as a +[3486.26 --> 3486.98] practical software +[3486.98 --> 3488.00] engineer but +[3488.00 --> 3489.86] the area where +[3489.86 --> 3490.58] I see a lot of +[3490.58 --> 3491.24] software engineers +[3491.24 --> 3491.98] kind of having to +[3491.98 --> 3492.74] improve on the +[3492.74 --> 3493.84] job is in their +[3493.84 --> 3494.72] ability to make a +[3494.72 --> 3496.32] concise argument and +[3496.32 --> 3497.74] my sister is +[3497.74 --> 3499.12] getting her PhD in +[3499.12 --> 3499.80] English so I'm +[3499.80 --> 3500.58] maybe a little +[3500.58 --> 3501.56] biased because I +[3501.56 --> 3502.34] see how much +[3502.34 --> 3503.06] work she has +[3503.06 --> 3503.80] like put into +[3503.80 --> 3504.90] this skill but +[3504.90 --> 3506.40] and I myself +[3506.40 --> 3508.06] was my major +[3508.06 --> 3508.86] was in the +[3508.86 --> 3509.62] philosophy department +[3509.62 --> 3510.26] even though it +[3510.26 --> 3511.24] was a logic and +[3511.24 --> 3512.08] computer science +[3512.08 --> 3513.28] interdisciplinary degree +[3513.28 --> 3515.32] so I used the +[3515.32 --> 3515.92] skills that I +[3515.92 --> 3516.96] got from +[3516.96 --> 3517.34] from my +[3517.34 --> 3518.22] humanities classes +[3518.22 --> 3520.24] more than a +[3520.24 --> 3520.82] number of the +[3520.82 --> 3521.42] computer science +[3521.42 --> 3522.28] theory classes that +[3522.28 --> 3522.94] I had to take in +[3522.94 --> 3523.30] college +[3523.30 --> 3526.06] okay Jeff I +[3526.06 --> 3526.54] didn't know you +[3526.54 --> 3527.42] renamed yourself +[3527.42 --> 3528.28] Natasha Jeff +[3528.28 --> 3529.42] I was just gonna say +[3529.42 --> 3530.02] that sounds like a +[3530.02 --> 3531.10] good go time +[3531.10 --> 3532.66] episode debating the +[3532.66 --> 3533.40] value of like a +[3533.40 --> 3534.04] computer science +[3534.04 --> 3534.96] degree versus like a +[3534.96 --> 3536.26] dedicated software +[3536.26 --> 3536.92] engineering because +[3536.92 --> 3537.54] they're like two +[3537.54 --> 3538.66] different fields +[3538.66 --> 3539.46] essentially right +[3539.46 --> 3540.40] invite me back for +[3540.40 --> 3540.86] that one +[3540.86 --> 3542.34] shameless plug I +[3542.34 --> 3543.00] think we did +[3543.00 --> 3543.80] one that was +[3543.80 --> 3544.58] beating around the +[3544.58 --> 3545.38] bush of that I +[3545.38 --> 3546.56] say a few maybe +[3546.56 --> 3547.30] it was years ago +[3547.30 --> 3548.36] a few months ago +[3548.36 --> 3549.40] me and Chris +[3549.40 --> 3550.06] Brando who's +[3550.06 --> 3550.74] another go time +[3550.74 --> 3551.60] hosted one around +[3551.60 --> 3552.08] like English +[3552.08 --> 3552.80] literature and its +[3552.80 --> 3553.50] value to software +[3553.50 --> 3555.48] engineering but I +[3555.48 --> 3555.86] think that was a +[3555.86 --> 3556.72] while ago so +[3556.72 --> 3557.80] absolutely let's do +[3557.80 --> 3558.82] another one but +[3558.82 --> 3560.20] again certainly not +[3560.20 --> 3561.96] least Natasha what +[3561.96 --> 3562.60] is your unpopular +[3562.60 --> 3564.16] opinion my unpopular +[3564.16 --> 3565.66] opinion I think +[3565.66 --> 3568.14] might be accepted in +[3568.14 --> 3569.16] computer science like +[3569.16 --> 3569.86] this kind of circle +[3569.86 --> 3571.22] but who knows I feel +[3571.22 --> 3572.10] like the world doesn't +[3572.10 --> 3573.56] need another superhero +[3573.56 --> 3574.90] movie there's too +[3574.90 --> 3575.94] many of them already +[3575.94 --> 3578.10] I'm done with it +[3578.10 --> 3578.90] like there's so many +[3578.90 --> 3579.68] other stories we +[3579.68 --> 3581.44] could tell so yeah +[3581.44 --> 3582.98] just let's end that +[3582.98 --> 3584.52] at least take a break +[3584.52 --> 3585.54] for a couple years +[3585.54 --> 3586.62] there are so many +[3586.62 --> 3587.80] more bugs that need +[3587.80 --> 3588.76] to be superheroes +[3588.76 --> 3590.52] we gotta have like +[3590.52 --> 3592.02] beetle man lady +[3592.02 --> 3593.04] bird lady +[3593.04 --> 3595.24] caterpillar +[3595.24 --> 3596.66] kazam +[3596.66 --> 3597.58] there's so many +[3597.58 --> 3598.48] opportunities +[3598.48 --> 3599.08] you could do +[3599.08 --> 3599.64] something with the +[3599.64 --> 3601.20] gopher like gopher +[3601.20 --> 3601.98] man or something +[3601.98 --> 3602.50] that could be +[3602.50 --> 3604.30] go for woman +[3604.30 --> 3605.22] okay yes +[3605.22 --> 3609.16] Sarah I loved what +[3609.16 --> 3610.04] you said you gave +[3610.04 --> 3611.12] me an idea to see +[3611.12 --> 3612.42] my next time I have +[3612.42 --> 3613.34] some free time to +[3613.34 --> 3614.22] hack I'm gonna use +[3614.22 --> 3615.58] something like GPT3 +[3615.58 --> 3616.58] to just create a +[3616.58 --> 3618.36] plugin for github +[3618.36 --> 3619.48] that will just run +[3619.48 --> 3620.26] whatever you want to +[3620.26 --> 3621.76] say through that to +[3621.76 --> 3622.94] sound more empathetic +[3622.94 --> 3624.82] and more something I +[3624.82 --> 3625.82] don't know but I +[3625.82 --> 3626.50] love that by the time +[3626.50 --> 3627.38] this episode is out +[3627.38 --> 3628.04] maybe it will be on +[3628.04 --> 3628.52] the marketplace +[3628.52 --> 3629.72] incredible this will +[3629.72 --> 3630.42] be really really +[3630.42 --> 3632.18] useful for everyone +[3632.18 --> 3633.04] yeah send it my +[3633.04 --> 3633.42] way +[3633.42 --> 3636.18] action item for +[3636.18 --> 3636.94] everyone listening +[3636.94 --> 3638.70] make it happen +[3638.70 --> 3640.02] come on put our +[3640.02 --> 3640.66] heads together we +[3640.66 --> 3641.44] can make it happen +[3641.44 --> 3643.08] well it has been an +[3643.08 --> 3645.18] absolute joy talking +[3645.18 --> 3646.22] to you all genuinely +[3646.22 --> 3646.96] I can't wait to get +[3646.96 --> 3647.96] you all back I have +[3647.96 --> 3649.26] so many more episode +[3649.26 --> 3650.64] ideas I'm sure you +[3650.64 --> 3652.50] do too Natalie I +[3652.50 --> 3652.94] hope you have a +[3652.94 --> 3653.76] wonderful rest of +[3653.76 --> 3654.82] your days but for +[3654.82 --> 3657.00] now adios and +[3657.00 --> 3657.52] goodbye +[3657.52 --> 3663.66] all right that is +[3663.66 --> 3664.32] go time for this +[3664.32 --> 3665.24] week thanks for +[3665.24 --> 3667.00] listening now is the +[3667.00 --> 3667.68] best time to +[3667.68 --> 3668.62] subscribe if you +[3668.62 --> 3669.88] haven't yet head to +[3669.88 --> 3671.18] go time dot fm for +[3671.18 --> 3672.70] all the ways and if +[3672.70 --> 3673.96] you are a regular go +[3673.96 --> 3675.10] time listener check out +[3675.10 --> 3676.10] our membership program +[3676.10 --> 3677.42] directly support our +[3677.42 --> 3679.00] work save yourself some +[3679.00 --> 3679.92] time by ditching the +[3679.92 --> 3681.34] ads and get bonuses +[3681.34 --> 3682.86] like exclusive content +[3682.86 --> 3684.82] and free stickers check it +[3684.82 --> 3686.50] out at changelog.com +[3686.50 --> 3687.58] slash plus plus +[3687.58 --> 3689.12] thanks again to our +[3689.12 --> 3690.58] partners at fastly for +[3690.58 --> 3692.32] CD ending for us to +[3692.32 --> 3694.02] fly.io for serving up +[3694.02 --> 3694.96] our app to the +[3694.96 --> 3696.00] mysterious breakmaster +[3696.00 --> 3696.98] cylinder for these +[3696.98 --> 3698.44] dope beats and to you +[3698.44 --> 3699.44] for being part of the +[3699.44 --> 3700.74] go time community we +[3700.74 --> 3702.34] appreciate you that is +[3702.34 --> 3703.54] all for this week we'll +[3703.54 --> 3705.24] talk to you next time on +[3705.24 --> 3705.90] go time +[3705.90 --> 3724.42] hello diff --git "a/The art of the PR\357\274\232 Part 2_transcript.txt" "b/The art of the PR\357\274\232 Part 2_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..032bb4aa22c502b86326bc2bd7443d0a1d7c1860 --- /dev/null +++ "b/The art of the PR\357\274\232 Part 2_transcript.txt" @@ -0,0 +1,978 @@ +[0.00 --> 5.80] The answer is it depends, but I think it depends on your role on this pull request, right? +[5.88 --> 11.92] So imagine that I maintain an open source and I'm taking something in from a person that's not from the repo. +[12.18 --> 16.20] I would be extra careful, so perhaps I would try to run the code. +[16.20 --> 26.70] Usually, I don't really test the code, but again, if it's something that I feel that's really critical and I want to be sure that's 100% working, I might test it. +[30.00 --> 34.38] This episode is brought to you by Sourcegraph. +[34.86 --> 39.30] Sourcegraph is universal code search to let you move fast, even in big code bases. +[39.38 --> 46.54] Here's CTO and co-founder, Biong Lu, explaining how Sourcegraph helps you to get into that ideal state of flow in coding. +[46.72 --> 51.70] The ideal state of software development is really being in that state of flow. +[51.70 --> 62.04] It's that state where all the relevant context information that you need to build whatever feature or bug that you're focused on building or fixing at the moment, that's all readily available. +[62.20 --> 67.58] Now, the question is, how do you get into that state where you don't know anything about the code necessarily that you're going to modify? +[67.96 --> 70.24] That's where Sourcegraph comes in. +[70.50 --> 73.60] And so what you do with Sourcegraph is you jump into Sourcegraph. +[73.70 --> 77.04] It provides a single portal into that universal code. +[77.04 --> 80.68] You search for the string literal, the pattern, whatever it is you're looking for. +[80.80 --> 83.86] You dive right into the specific part of code that you want to understand. +[84.24 --> 97.00] And then you have all these code navigation capabilities, jump to definition, find references that work across repository boundaries that work without having to clone the code to your local machine and set up and mess around with editor config and all that. +[97.12 --> 102.46] Everything is just designed to be seamless and to aid in that task of, you know, code spelunking or source diving. +[102.46 --> 110.64] And once you've acquired that understanding, then you can hop back in your editor, dive right back into that flow state of, hey, all information I need is readily accessible. +[110.86 --> 115.32] Let me just focus on writing the code that influenced the feature or fixes the bug that I'm working on. +[115.68 --> 117.82] All right. Learn more at Sourcegraph.com. +[117.92 --> 125.94] And also check out their bi-monthly virtual series called DevToolTime covering all things DevTools at Sourcegraph.com slash DevToolTime. +[132.46 --> 142.08] Let's do it. +[142.74 --> 143.72] It's Go Time. +[144.42 --> 149.56] Welcome to Go Time, your source for diverse discussions from all around the Go community. +[150.08 --> 152.04] Subscribe to the pod if you haven't yet. +[152.18 --> 154.86] Head to GoTime.fm for all the ways. +[155.12 --> 157.56] And if you dig the show, please do tell your friends. +[157.74 --> 158.50] That'd be pretty cool. +[158.50 --> 164.04] Special thanks to our partners at Fastly for shipping all of our pods super fast to wherever you listen. +[164.26 --> 165.98] Check them out at Fastly.com. +[166.20 --> 167.96] And to our friends at Fly.io. +[168.42 --> 170.54] Host your app servers close to your users. +[170.76 --> 171.68] No ops required. +[172.12 --> 173.42] Learn more at Fly.io. +[173.74 --> 174.62] Okay, here we go. +[176.10 --> 179.42] Good part of the day to everyone wherever you're joining from. +[179.98 --> 181.36] And here's Angelica and I. +[181.44 --> 183.78] We are back to talk about pull requests. +[184.12 --> 184.86] Hi, Angelica. +[184.90 --> 185.36] How are you doing? +[186.20 --> 187.58] I'm very, very well. +[187.58 --> 188.20] Thank you. +[188.20 --> 194.26] I didn't think the PRs would take two episodes, but I'm surprised and excited that they will. +[194.68 --> 197.94] As everything about pull requests always takes a bit longer than expected, huh? +[198.48 --> 199.06] That's true. +[200.32 --> 202.56] Today we are joined by Anderson. +[202.98 --> 203.32] Hello. +[203.60 --> 204.74] Hi, how are you doing, Anderson? +[205.18 --> 206.26] I'm really, really good. +[206.34 --> 207.68] I'm really happy to be here. +[208.04 --> 208.58] Thank you. +[208.86 --> 211.46] And you are joining us from the UK. +[212.24 --> 212.76] Exactly. +[213.26 --> 214.28] What are you doing in the UK? +[214.76 --> 216.00] I'm at the GopherCon UK. +[216.00 --> 218.66] So I'm a Brazilian that's based in Berlin. +[219.16 --> 222.62] But now I'm here in London, directly from an hotel room. +[222.82 --> 223.76] How are you liking it? +[224.22 --> 224.98] Oh, it's really good. +[225.20 --> 225.42] Yeah. +[225.80 --> 227.66] And you're not just saying that because I'm on the call. +[228.36 --> 228.88] No, no. +[229.10 --> 229.62] Okay, good. +[229.74 --> 230.48] I love the UK. +[230.80 --> 233.48] I did an exchange program here for one year and a half in Glasgow. +[233.72 --> 234.12] Oh, awesome. +[234.22 --> 234.86] So, yeah. +[235.28 --> 235.84] I really like it. +[236.18 --> 238.12] I mean, I prefer Edinburgh to Glasgow. +[238.64 --> 239.44] Unpopular opinion. +[239.68 --> 240.46] Yeah, true. +[240.70 --> 242.26] I mean, Edinburgh Castle is so incredible. +[242.40 --> 243.86] Wait, no, that's the end of the show. +[244.26 --> 244.80] Oh, I'm sorry. +[244.86 --> 246.06] Getting ahead of myself, Natalie. +[246.14 --> 246.48] I'm sorry. +[246.58 --> 246.96] I'm sorry. +[249.46 --> 251.82] It's fun to go back to in-person conferences, huh? +[252.52 --> 253.54] Yeah, that's true. +[253.60 --> 254.40] It's really, really good. +[254.76 --> 257.18] Last time that I was here in London, my brother was here. +[257.18 --> 262.12] We were like, okay, the place here is going to be small to go for call UK at some point, +[262.32 --> 264.76] but then the pandemic, and now I think it's smaller. +[265.40 --> 266.10] So, yeah, let's see. +[267.26 --> 267.70] Okay. +[268.18 --> 270.16] So, Anderson, tell us about yourself. +[270.24 --> 270.88] You're doing Go. +[271.50 --> 276.20] Yeah, I'm doing Go since a bit before going to Berlin. +[276.44 --> 278.66] So, I think about five years now. +[279.20 --> 282.34] I did before many Java was a lot of Java. +[282.34 --> 289.02] I started with C, actually, some Python, JavaScript in the back end, and now Go. +[289.08 --> 289.56] I love it. +[289.68 --> 292.06] And then it chooses my language to specialize. +[292.74 --> 294.26] And you're working at Elastic. +[294.70 --> 295.04] Exactly. +[295.58 --> 296.96] I work with the Elastic agent. +[297.32 --> 301.80] Which is the head product for all the products that we know and love, like Elasticsearch. +[302.18 --> 302.82] Yeah, exactly. +[303.20 --> 303.38] Yeah. +[303.86 --> 305.64] Formerly known as the ilk stack. +[306.32 --> 306.62] Yeah. +[306.94 --> 308.12] Now the Elastic stack. +[308.26 --> 308.40] Yeah. +[308.78 --> 309.50] Lots of go there. +[309.50 --> 311.50] I love this stack. +[312.34 --> 313.02] Fun. +[313.38 --> 315.96] And so, as part of the jobs, you do pull requests. +[316.46 --> 316.86] Of course. +[317.22 --> 319.00] Do you also do merge requests? +[319.64 --> 321.06] We call everything pull requests. +[321.94 --> 322.34] Yeah. +[322.52 --> 325.22] Are you going to do the same thing you did last time, where you caught us all off guard, +[325.34 --> 326.40] with the many different... +[326.40 --> 327.08] Just for those. +[327.34 --> 329.94] The many different ways to refer to it. +[330.06 --> 333.40] It's crazy how many ways there, how many names are there for this. +[333.54 --> 333.70] Yeah. +[334.00 --> 336.92] For those who are listening now and have not listened to the previous episode, +[336.92 --> 340.84] to the art of pull requests part one, Angelica and I were discussing, +[340.84 --> 345.72] among other things, also the many different names and what concepts that represents. +[346.36 --> 350.46] So, Anderson, we asked you if you're doing lots of pull requests, but then we started talking about other things. +[350.88 --> 351.34] Yeah. +[351.52 --> 351.78] Do you? +[352.12 --> 352.66] I do. +[353.24 --> 355.70] Do you review more or do you write more? +[355.70 --> 357.14] Now, I write more. +[357.28 --> 359.92] My past job, I read a lot more code. +[360.52 --> 361.84] A lot pull requests. +[362.58 --> 365.76] And funnily enough, I've never worked in a company that didn't have pull requests. +[365.86 --> 371.04] To me, software development as a professional means pull requests and code review. +[371.86 --> 372.80] That's my standard. +[373.40 --> 376.74] Why did you do more reading in your past job? +[376.80 --> 378.78] Is it just a very different area you were working in? +[378.78 --> 379.18] No. +[379.40 --> 385.42] My past company, I joined, among other things, to help to lead the transition to Go. +[385.56 --> 385.88] Right. +[385.98 --> 390.12] So, they were pretty much a ruby shop and they decided to migrate to Go. +[390.56 --> 395.40] They were migrating, but they needed someone with expertise in Go, you know, to bring back +[395.40 --> 396.54] practice, how to do. +[396.54 --> 403.20] So, as that, I did a lot of workshops and, you know, kind of teaching mentoring and involved +[403.20 --> 406.20] a lot of the teams would come to me with pull requests. +[406.78 --> 412.22] So, I do like do like a really extensive review, not only code practice functionality, but also +[412.22 --> 417.30] as an opportunity to teach Go and the standards, the conventions, the best practices. +[418.10 --> 420.36] And then now you've moved to somewhere where Go is bigger. +[420.58 --> 423.50] So, therefore, you don't have to play such a big kind of reviewer role. +[423.70 --> 424.04] Exactly. +[424.04 --> 428.78] And now I am as a soft engineer, right, as a senior, but there I was like a tech lead. +[429.40 --> 433.82] Yeah, I think when you go above senior and in tech lead and stuff, you start coding less, +[433.92 --> 435.58] reading more, writing more, right? +[435.78 --> 436.16] Okay. +[436.64 --> 439.36] I mean, writing more specifications and documentations. +[440.04 --> 446.88] So, do you feel like there's a direct correlation between seniority and how much reading of PR +[446.88 --> 449.04] versus writing you do? +[449.04 --> 454.86] Like if we were to plot it on a graph, how much you read PRs, how much you write, could +[454.86 --> 459.56] we like to generalize the industry say that the more senior you are, the less you're going +[459.56 --> 461.62] to be writing PRs and the more you're going to be reviewing them? +[462.00 --> 467.22] I think the more you're a senior, you're going to look for more things, right? +[467.22 --> 471.54] So, because you're going to be able to evaluate, if you're good in the language, you can evaluate +[471.54 --> 473.24] the language, the conventions, right? +[473.28 --> 476.04] You can evaluate the general soft architecture, right? +[476.38 --> 480.36] When you're in a role that as a senior, you are also a mentor, there is a lot of like +[480.36 --> 481.68] mid and juniors in the team. +[481.76 --> 485.52] I think we're reading a lot because you have this responsibility with the team, with the +[485.52 --> 485.94] product. +[485.94 --> 492.68] But if you're in a company where everyone's senior, senior, in Berlin, definitely the titles +[492.68 --> 493.26] are inflated. +[493.78 --> 499.10] So, you know, perhaps everyone doesn't need this level of attention and mentoring or teaching +[499.10 --> 501.66] that happens a lot through PRs. +[502.12 --> 503.12] That's an interesting question. +[503.22 --> 503.32] Yeah. +[503.36 --> 507.34] Because some time ago, they added this like graph of contributions that is no longer just +[507.34 --> 510.90] the squares, but also what type do you read more? +[510.98 --> 511.90] Do you make more issues? +[512.06 --> 513.16] Do you review and so on? +[513.16 --> 517.22] So if you can look at that, then you kind of know what people's roles are or see how +[517.22 --> 518.68] it changes over time or something. +[519.00 --> 522.82] If I showed you a load of graphs, would you be able to guess the seniority of that engineer? +[523.30 --> 523.74] Perhaps. +[524.12 --> 527.42] I think I looked quickly at mine and now it was 50-50. +[527.76 --> 528.16] Nice. +[528.38 --> 528.64] Okay. +[528.70 --> 529.16] Pretty much. +[529.48 --> 532.56] Do you think that that's the balance that most engineers would like? +[533.16 --> 537.36] Or do you think that there is such thing as someone who prefers to be reviewing more than +[537.36 --> 537.74] writing? +[538.40 --> 542.86] I think when you think as a soft engineer that the passion is about writing, writing, +[543.16 --> 545.12] right, they rather write code. +[545.64 --> 551.38] I think at least either you want or you need to pass knowledge forward, right? +[551.42 --> 554.10] So you're going to need to write, to read. +[554.46 --> 557.06] I really like to review code for both reasons, right? +[557.14 --> 562.02] One, because I have a passion for code and I am a quite methodic person. +[562.02 --> 565.44] So I like to ensure that the code is good. +[566.20 --> 567.84] The conventions are there. +[568.22 --> 570.60] I'm super picky about proper error handling. +[571.32 --> 575.72] So these are some things that if you're not handling the errors properly or you're really +[575.72 --> 578.64] bad at the conventions, I'm going to be commenting there. +[579.08 --> 584.58] But I have a special of commenting to ensure that I'm not overwhelming the other person or +[584.58 --> 587.58] to feel like just complaining, saying that your job's bad. +[587.58 --> 589.92] So you say you try to give feedback. +[590.26 --> 592.04] You don't want to overwhelm the person. +[592.62 --> 593.58] How do you do that? +[593.66 --> 597.80] Like, is it that you limit yourself to, OK, I'm only going to put six comments? +[598.04 --> 601.00] Is it the way in which you phrase your review? +[601.12 --> 603.00] How do you make sure that you're not overwhelming? +[603.96 --> 607.14] Yeah, I learn to experience any feedback. +[607.32 --> 608.48] I put tags, right? +[608.52 --> 613.74] So I start with PR with like suggestion or sometimes I put question, but a question is +[613.74 --> 614.28] a question, right? +[614.28 --> 619.62] You know, so it feels a bit redundant, but I put, and then I have blocker, right? +[619.66 --> 622.56] And sometimes like suggestion slash go convention, right? +[622.70 --> 628.56] Or depending on the repo, blocker slash go convention, you know? +[628.64 --> 633.00] So I try to categorize suggestion, question, and blocker. +[633.12 --> 634.34] I put blocker-ish. +[634.48 --> 638.84] So basically blocker is either I see there is a problem in the code, right? +[638.84 --> 644.96] So it needs a change or I believe there is a problem or I don't believe that this implementation +[644.96 --> 645.64] is good. +[645.76 --> 648.16] So at least you need to answer that, right? +[648.20 --> 649.44] You can say like, it's agree with you. +[649.82 --> 650.28] That's fine. +[650.36 --> 652.46] But I need an explanation, right? +[652.48 --> 653.66] So these are the blockers. +[654.32 --> 655.48] Suggestion is exactly that. +[655.54 --> 657.06] Like, I believe it can be better. +[657.32 --> 662.14] If I were writing this code, I would do different, but you don't have to take it. +[662.14 --> 664.60] Sometimes I put like a nit. +[665.00 --> 669.78] It can be, you know, just erase this blank line between the function call and the error +[669.78 --> 670.84] handling, right? +[670.88 --> 671.78] This is a nit, right? +[671.98 --> 672.82] Or typo. +[673.28 --> 674.04] So these things. +[674.24 --> 680.60] And then if I've never reviewed a PR from this person, or perhaps I see there's a lot +[680.60 --> 685.78] of comments, I go to the person or even the PR, I put, look, there is this three categories. +[685.90 --> 687.54] The only thing that's really important are the blockers. +[687.54 --> 692.98] So the blockers, please comment or change the suggestions are suggestions. +[693.30 --> 696.46] And as always, you're free to disagree a hundred percent, right? +[697.02 --> 698.98] Just answer the blockers and talk to me. +[699.28 --> 700.32] That's how I do. +[700.76 --> 705.28] So you said that in some situations, it kind of makes sense to just speak in person instead +[705.28 --> 708.14] of writing a lot, for example. +[708.72 --> 709.56] Sometimes it's easier. +[710.18 --> 712.36] I think sometimes speaking is better. +[712.50 --> 714.42] Sometimes writing is better. +[714.42 --> 719.04] And on pull requests, writing the pull requests is a lot slower, right? +[719.08 --> 722.62] So sometimes it's worth jumping on a Slack chat. +[722.78 --> 723.54] Is that right enough? +[723.86 --> 724.82] Sometimes we talk. +[725.46 --> 731.06] I think nowadays at Elastic, because we are distributed, it's a lot more through Slack +[731.06 --> 733.00] rather than really a call. +[733.66 --> 737.88] But sometimes I have jumped on a call for small things. +[738.36 --> 741.48] And how do you decide when it's better to do this and when it's better to not? +[741.48 --> 744.26] Do you have a thumb rule or is it all just feeling? +[744.90 --> 748.98] If there is a lot of back and forth, it's easier to jump on a call or something that +[748.98 --> 750.18] I really want to understand. +[751.06 --> 755.72] But most of the times, I guess, I always could handle the chat. +[756.18 --> 758.32] Now, Elastic is a distributed company, right? +[758.36 --> 760.42] So we write a lot more. +[760.96 --> 764.48] So it's a lot more common just to write than jump on a call. +[764.48 --> 766.98] I think jumping on a call is more personal. +[767.86 --> 771.58] So I feel if you're close to someone, it's easier. +[771.88 --> 773.92] You know, it feels more natural to jump on a call. +[774.12 --> 779.00] And if you don't know the person so well or something, you kind of end up just chatting. +[779.50 --> 783.10] So you say there's actually three types of kind of giving feedback. +[783.26 --> 785.28] One, writing in the pull request. +[785.72 --> 787.58] Two, writing on Slack. +[787.72 --> 789.28] And three, hopping on a call. +[789.90 --> 790.16] Yeah. +[790.66 --> 793.62] But on Slack, I use more to clarify. +[793.62 --> 794.74] Right? +[795.38 --> 798.16] Because it's specific about the code I'd rather have on the pull request. +[798.64 --> 800.60] It's kind of documented anyway, right? +[801.28 --> 801.64] It's right. +[801.70 --> 804.44] I was sure to explain why, what I'm saying. +[805.16 --> 811.08] Like, as I said, in my past company, I was really with this job to teach, to mentor and go. +[811.22 --> 813.52] I would link a lot of reference, right? +[814.08 --> 815.94] Also, because I learned in another company. +[816.26 --> 820.34] I remember my colleague and I would review a pull request of your new joiner. +[820.72 --> 822.48] You know, and say like, oh, yeah, this function is too long. +[822.48 --> 823.64] This function is too complex. +[824.04 --> 824.14] Right? +[824.16 --> 824.58] This happens. +[824.80 --> 827.58] And the guy was like, why? +[827.90 --> 829.56] Why did you butcher my pull request? +[829.74 --> 830.44] Are you guessing? +[830.58 --> 831.44] This is your opinion. +[832.00 --> 834.02] And I was like, that's true, right? +[834.04 --> 836.66] We don't have like a metric to say that. +[837.08 --> 839.58] But also it was a common sense between us. +[839.58 --> 841.62] So yeah, this is too big, you know, in the company. +[841.74 --> 843.08] But this guy was a new joiner. +[843.08 --> 844.28] So everything was different from him. +[845.06 --> 849.00] So I guess, yeah, clarifications, I can't do outside the pull request. +[849.14 --> 853.20] But I think it's important to document what's happening in the code there. +[853.20 --> 853.92] Mm-hmm. +[854.22 --> 857.30] In the code itself to kind of document, to make it self-explanatory. +[857.70 --> 858.10] Exactly. +[858.66 --> 862.38] So in terms of kind of, you talked about kind of giving feedback, not overwhelming. +[863.26 --> 868.44] Do you feel like PRs are also a good place to kind of, especially for more junior engineers, +[869.00 --> 871.24] give them like props on things they're doing well? +[871.44 --> 873.68] Like, oh, I really like the way you did this thing. +[873.84 --> 875.14] Or, oh, this is great. +[875.24 --> 877.36] Or this function is structured really well. +[877.84 --> 881.08] Do you feel like PRs are also a way to give positive feedback? +[881.08 --> 882.96] Yeah, I think it is. +[883.14 --> 884.10] It is important. +[884.90 --> 888.24] And it's something that I would like to do more as well. +[888.74 --> 893.22] I mean, I never go to a pull request to looking for errors or to try to diminish someone's job. +[893.54 --> 897.62] But at the same time, I go to a pull request looking for errors so they don't go to production. +[898.22 --> 898.40] Right? +[898.86 --> 904.94] At the end of the day, we are looking for problems and issues to prevent things bad going to production. +[905.04 --> 905.20] Right? +[905.24 --> 910.22] But I think it's super important to do the praise, to assert when someone does something nice. +[910.22 --> 914.26] Sometimes someone just fix something a bit random, but, you know, it's the same function. +[914.46 --> 917.24] It's not really going away from the aim of the pull request. +[917.38 --> 920.32] And then this is really good to praise, at least a thumbs up. +[920.66 --> 926.04] So we said that there are some stages that you escalate the communication to some way. +[926.42 --> 933.98] And definitely you want to include more like positive feedback that is not the correctness and so on or explain yourself. +[933.98 --> 941.14] What other changes would you make to the PR process based on pain points you have with the flow? +[941.48 --> 945.92] It's a really slow process because you write the code and then you submit the PR. +[946.76 --> 946.78] Right? +[946.88 --> 951.50] And then someone else's, sometimes more than WordPress, they have to stop and read it. +[951.80 --> 953.38] And how do you synchronize that? +[953.38 --> 956.54] And then I think it's going to depend a lot how companies do. +[956.76 --> 958.58] I've worked with a different process. +[959.22 --> 959.34] Right? +[959.38 --> 962.34] Some process, they ensure that the pull request would be reviewed. +[962.50 --> 963.44] Some not so often. +[964.12 --> 964.30] Right? +[964.82 --> 968.08] Now I'm definitely overwhelmed by GitHub notifications. +[969.20 --> 973.36] So sometimes it just, it slips through some PRs. +[973.42 --> 975.40] And then I get like days later, someone tag me. +[975.52 --> 977.58] And so can you give me a review? +[977.70 --> 978.68] Like, oh my God, sorry. +[979.24 --> 981.64] Because it's something that's kind of a bit of a ping pong? +[982.02 --> 984.38] Or does it happen more with new pull requests? +[984.70 --> 986.20] And that is when you have to re-review, right? +[986.24 --> 988.36] You do the first review and then you have to look again. +[988.64 --> 988.90] Yeah. +[989.24 --> 989.44] Right? +[989.48 --> 992.32] So to get the proper ping, it's hard. +[992.76 --> 995.40] And the other thing that, at least on GitHub, right? +[995.54 --> 998.32] I have used a bit of Bitbuck as well. +[998.78 --> 1002.52] I submit one, perhaps two PRs should go to the GoRepper. +[1002.52 --> 1003.80] So we've got it. +[1004.80 --> 1006.80] But on GitHub, it's really hard. +[1006.80 --> 1009.54] But when you do the first review, we ask you for changes. +[1009.78 --> 1014.48] And then if people do changes to see exactly what changed, right? +[1014.50 --> 1017.22] To get an MBA book, okay, I'm going to review just the changes. +[1017.48 --> 1018.34] Isn't this a new comment? +[1018.36 --> 1019.68] And if someone just forced. +[1019.70 --> 1020.50] Oh, forced push. +[1020.58 --> 1020.78] Okay. +[1021.26 --> 1023.48] If someone's a forced push, right? +[1023.98 --> 1029.00] So sometimes it's hard to get like just the bit that they have to re-review. +[1029.48 --> 1030.40] That's our main point. +[1030.40 --> 1030.80] Yeah. +[1030.94 --> 1034.34] I had a Twitter poll the other day on the, what do you do when the pull request stretches +[1034.34 --> 1037.54] so much and there's 1 million comments that lead to 1 million commits? +[1037.68 --> 1038.88] Do you squash that or not? +[1038.98 --> 1043.64] And I think the answer that I liked most was during the pull request, have as many commits +[1043.64 --> 1044.02] as needed. +[1044.14 --> 1048.92] And once it's good to go merged, then squash it all into a readable one or two commits or +[1048.92 --> 1050.74] however many logically needed. +[1051.00 --> 1055.08] But exactly to allow the proper review, like you say, as many commits as needed. +[1055.34 --> 1055.58] Yeah. +[1056.22 --> 1056.44] Yeah. +[1056.44 --> 1057.60] Definitely good practice. +[1058.04 --> 1060.48] So you kind of chatted a little bit about your love of Go. +[1060.56 --> 1066.00] This is a Go podcast, but I would love to hear a little bit about your approach to PRs +[1066.00 --> 1066.82] in other languages. +[1067.12 --> 1067.50] I.e. +[1067.64 --> 1071.72] does the way that you review PRs differ depending on the language you're reviewing? +[1072.30 --> 1077.64] I think they do differ most because of the expertise, right? +[1077.64 --> 1082.80] I'm quite comfortable in Go to understand the language and understand what's happening, +[1083.06 --> 1083.28] right? +[1083.74 --> 1088.06] To know some caveats, some catches, and also to talk about conventions. +[1089.00 --> 1090.58] In other languages, not so much. +[1091.34 --> 1096.46] So I guess in other languages, I am going to focus more on general software architecture +[1096.46 --> 1099.74] because this is a general frame thing and functionality. +[1100.48 --> 1105.98] And of course, try to use as much as I know from the language convention, right? +[1105.98 --> 1107.56] I think that's another thing. +[1107.66 --> 1110.30] Prue requests are great, great to learn language conventions. +[1110.54 --> 1115.72] So I've learned so much about conventions in Go and the other language through prue requests. +[1116.34 --> 1117.16] And then that's it. +[1117.24 --> 1122.74] If I'm not so expert, I'm going to try to put in point the best that I know. +[1123.18 --> 1124.50] But I know that I'm not the expert. +[1124.70 --> 1128.26] And a lot of the times, it's probably not really my repo, right? +[1128.32 --> 1130.56] So I'm not there to enforce anything. +[1130.96 --> 1133.44] So they're going to be more on the suggestion side. +[1133.44 --> 1139.64] I mean, taking it kind of one step more granular, are there things that are more important +[1139.64 --> 1146.32] for Go when you're reviewing, i.e. like stylistic choices, almost principles that you might adhere +[1146.32 --> 1149.90] to more closely than you would in other languages? +[1150.72 --> 1152.14] I mean, Go is opinionated. +[1152.40 --> 1153.60] You have to use GoFund. +[1153.88 --> 1159.04] You have to format the code properly, even though we still have some space to discuss about +[1159.04 --> 1159.58] how to format. +[1159.58 --> 1161.48] But I think that's the first thing. +[1161.92 --> 1167.06] I usually would like to be super strict about how the imports are sorted, but I am not. +[1167.54 --> 1172.12] Wait, if the rep is consistent, I think it's a lot easier to enforce this thing. +[1172.20 --> 1174.04] If it isn't, not so much. +[1174.24 --> 1176.72] And all the language, they don't have so much. +[1176.78 --> 1180.34] So it's going to be more about team convention rather than the language. +[1180.48 --> 1183.38] And in Go, you get a lot from the language. +[1183.38 --> 1190.88] I haven't seen so much, but because Go, they're so focused, let's say, right, in concurrence. +[1191.34 --> 1196.60] Sometimes people try to either sneak in concurrence when they shouldn't, or they are not using +[1196.60 --> 1198.60] the right tools. +[1199.18 --> 1201.76] Also because, you know, ah, concurrence, let's use channels. +[1201.92 --> 1203.74] No, channels, they're for something. +[1203.98 --> 1205.34] In Mutex, they're for other things. +[1205.74 --> 1207.24] Weight groups, they're for other things, right? +[1207.24 --> 1212.78] So this is another thing that I would say, okay, no, perhaps we can do different or we +[1212.78 --> 1213.92] can do better, right? +[1213.98 --> 1217.14] Or this is too complex to understand. +[1217.52 --> 1220.86] If you use, I don't know, if you remove this channel, you put in a weight group, it's a +[1220.86 --> 1221.24] lot easier. +[1221.76 --> 1225.06] And you will, do you want to barrier weight group, right? +[1225.64 --> 1227.08] Channels, probably not. +[1227.24 --> 1228.48] So I think these things. +[1229.14 --> 1235.68] When there are new features and new things released in Go, do we see an uptick in people +[1235.68 --> 1237.22] using those and PRs? +[1238.24 --> 1241.18] I feel like you just get like, you just get overexcited, like, oh, generics. +[1241.64 --> 1243.36] Every PR now has generics. +[1243.68 --> 1248.90] I always wanted to push the new things and use the new as soon as possible. +[1249.48 --> 1253.50] I think in general, if you're working with microservices, it's a lot easier, right? +[1253.50 --> 1257.70] Because you can just update the version and redeploy and even something breaks, you can +[1257.70 --> 1258.56] roll back a lot easier. +[1258.92 --> 1263.32] Now at Elastic that you're distributing binaries that you're going to go to, I don't know how +[1263.32 --> 1265.32] many clients in the whole world. +[1265.32 --> 1268.82] So our, we have a release cycle, right? +[1268.88 --> 1272.46] So we have to choose, okay, let's change the version. +[1272.70 --> 1277.02] We have several repls that use Go and you try to keep everyone in the same version. +[1277.12 --> 1278.82] So it's a slower process. +[1279.44 --> 1285.60] But as much as I can and as much as I know, all right, what's coming up, I try to incorporate +[1285.60 --> 1286.80] if I can. +[1287.78 --> 1292.34] Oh, on that topic, have you folks started to use any instead of the empty interface? +[1292.34 --> 1295.56] It could be an interesting poll to write. +[1296.18 --> 1298.84] I don't know if you can phrase that as an unpopular opinion somehow. +[1299.02 --> 1300.32] Don't use that or something. +[1300.58 --> 1304.10] It can be one more unpopular opinion for your stash. +[1304.16 --> 1305.90] Your library that we're soon going to have. +[1305.90 --> 1310.88] I think I saw the first use of it today in the workshop with Bill Kennedy. +[1311.24 --> 1313.12] So he, his code had an any. +[1313.20 --> 1314.38] I was like, oh yeah, right. +[1314.46 --> 1315.42] We can use any now. +[1316.20 --> 1317.62] What was the use case that he used? +[1318.16 --> 1320.78] Was a map for logger? +[1321.24 --> 1321.50] Okay. +[1321.76 --> 1323.54] The map string empty interface. +[1323.54 --> 1326.02] I think it was in a log or something. +[1326.34 --> 1326.52] I know. +[1326.60 --> 1328.06] I think he was parsing a JSON. +[1328.72 --> 1333.10] And so instead of map string empty interface was map string any. +[1333.52 --> 1333.80] Okay. +[1334.08 --> 1337.58] Well, we can do the poll and then tag Bill and be like, please tell us. +[1338.70 --> 1341.58] Give us the example so we can all understand how you use this. +[1341.74 --> 1341.98] Yeah. +[1342.14 --> 1344.52] He was asking everyone to have Go 18 because of that. +[1344.64 --> 1346.72] He was like, okay, you're going to use any, so please. +[1346.72 --> 1363.26] This episode is brought to you by our friends at FireHydrant. +[1363.68 --> 1366.32] FireHydrant is a reliability platform for every developer. +[1366.82 --> 1369.32] Incidents are a win, not an if situation. +[1369.78 --> 1373.80] And they impact everyone in the organization, not just SREs. +[1373.80 --> 1377.44] And I'm here with Robert Ross, founder and CEO of FireHydrant. +[1377.78 --> 1380.10] Robert, what is it about teams getting distracted by incidents +[1380.10 --> 1383.42] and not being able to focus on the core product that upsets you? +[1383.72 --> 1387.72] I think that incidents bring a lot of anxiety and sometimes fear +[1387.72 --> 1391.94] and maybe even a level of shame that can cause this paralysis +[1391.94 --> 1394.60] in an organization from progress. +[1395.18 --> 1398.10] And when you have the confidence to manage incidents +[1398.10 --> 1401.02] at any scale of any variety, +[1401.02 --> 1403.30] everyone just has this breath of fresh air +[1403.30 --> 1406.08] that they can go build the core product even more. +[1406.44 --> 1408.42] I don't know if anyone's had the opportunity, +[1408.72 --> 1410.98] maybe is the word, to call the fire department. +[1411.18 --> 1413.70] But no matter what, when the fire department shows up, +[1413.90 --> 1416.84] it doesn't matter if the building is hugely on fire. +[1417.00 --> 1418.46] They are calm, cool, and collected +[1418.46 --> 1420.50] because they know exactly what they're going to do. +[1420.78 --> 1423.70] And that's what FireHydrant is built to help people achieve. +[1424.18 --> 1424.48] Very cool. +[1424.56 --> 1425.10] Thank you, Robert. +[1425.10 --> 1429.48] If you want to operate as a calm, cool, collected team +[1429.48 --> 1431.86] when incidents happen, you got to check out FireHydrant. +[1432.18 --> 1434.82] Small teams, up to 10 people can get started for free +[1434.82 --> 1437.78] with all the features, no credit card required to sign up. +[1438.08 --> 1439.76] Get started at firehydrant.com. +[1440.12 --> 1442.12] Again, firehydrant.com. +[1452.42 --> 1456.32] What do you do when you have a very large pull request? +[1456.68 --> 1458.96] Lots of files, lots of comments, lots of lines. +[1458.96 --> 1460.18] I sit and cry. +[1460.38 --> 1461.84] How do you get on top of that? +[1462.64 --> 1463.30] No, I... +[1463.96 --> 1464.78] That's a tough one. +[1465.42 --> 1467.42] I try to review at once. +[1468.14 --> 1469.36] Sometimes it's not possible. +[1470.68 --> 1472.90] I think there's no magic. +[1473.08 --> 1474.70] You just have to go through it, right? +[1475.12 --> 1477.28] Do you review everything on a high level? +[1477.42 --> 1479.16] Kind of, you know, see the list of the commits +[1479.16 --> 1481.62] if they tell some story or maybe look at the list of the files? +[1481.76 --> 1483.24] Or do you just dive to the first one +[1483.24 --> 1486.02] and one by one until it starts making sense? +[1486.02 --> 1487.98] I never look at the commits. +[1488.16 --> 1491.74] I don't know if because when I'm coding and committing, +[1491.94 --> 1493.76] I'm going to squash everything before morning. +[1493.94 --> 1494.52] Like, first things first. +[1494.62 --> 1497.30] So to me, the commits itself, they don't matter so much. +[1497.38 --> 1500.62] I try to put in a way if I need to revert something, right? +[1500.70 --> 1501.24] I do. +[1501.50 --> 1503.46] But at the end of the day, there's a good chance +[1503.46 --> 1505.04] that I'm going to just do one commit. +[1505.04 --> 1506.72] So I never look at the... +[1506.72 --> 1508.64] Neither the rep commit history, +[1508.82 --> 1511.02] only if I need to understand why it happened. +[1511.52 --> 1513.72] But in a pull request, I never look at the commit history. +[1513.90 --> 1514.78] Just look at the diff. +[1515.36 --> 1516.76] And it's always on GitHub. +[1517.22 --> 1519.54] And look at the files by the name, basically. +[1519.82 --> 1521.64] The big ones are just, you know, go clicking. +[1521.74 --> 1522.46] I've seen this file. +[1522.60 --> 1523.24] I've seen this file. +[1524.18 --> 1526.22] So just by the order of appearance. +[1526.22 --> 1529.92] Because sometimes it's not always the correct flow, kind of. +[1529.92 --> 1530.36] Yeah. +[1530.86 --> 1533.56] If it's hard to understand, I get the code. +[1533.68 --> 1536.32] I check out the feature branch. +[1537.00 --> 1537.76] And I go to see. +[1537.92 --> 1541.50] Because also sometimes you'll see, you know, +[1541.54 --> 1543.12] there are more stuff going on. +[1543.20 --> 1543.84] You want to jump. +[1543.98 --> 1546.34] You want to understand how it was called or something. +[1546.66 --> 1549.56] And then it's easier on an IDE when you have the code. +[1550.06 --> 1552.24] And also if you want to suggest a change. +[1553.02 --> 1555.12] So either it's something really simple. +[1555.20 --> 1556.56] I'm 100% sure that it works. +[1556.56 --> 1560.38] Or I'm going to probably write it in the code itself. +[1560.52 --> 1561.50] And it might test it. +[1561.62 --> 1562.02] Right. +[1562.10 --> 1564.28] To not suggest something that it's broken. +[1564.88 --> 1566.36] Maybe even a few steps, Mac. +[1566.48 --> 1569.40] When you go to read the review of pull request. +[1569.48 --> 1570.88] Do you start by reading the issue? +[1571.42 --> 1573.18] So the first thing you do is read the issue. +[1573.60 --> 1573.80] Yeah. +[1573.96 --> 1575.48] I have to understand what's happened there. +[1575.88 --> 1581.00] Then do you review it, the diff, you know, on GitHub or in your IDE? +[1581.72 --> 1582.30] No, on GitHub. +[1583.08 --> 1585.32] So you go kind of file by file on GitHub. +[1585.32 --> 1586.06] I never. +[1587.34 --> 1588.20] No, you know, because. +[1588.62 --> 1589.40] Philosophical questions. +[1589.80 --> 1590.22] Good point. +[1590.54 --> 1592.64] Because to me, the review. +[1593.38 --> 1595.32] Actually, going back to the other episode. +[1595.66 --> 1596.64] We are doing code review. +[1596.76 --> 1598.12] It's not the pull request so much, right? +[1598.58 --> 1599.14] What's happening? +[1599.80 --> 1601.72] You have to comment on it, right? +[1601.76 --> 1603.64] So it's really hard to comment on it. +[1604.02 --> 1605.64] At least for the tools that I use. +[1605.72 --> 1607.44] That's GitHub and the IDE. +[1608.10 --> 1609.62] It's hard to comment on the code. +[1609.62 --> 1612.98] If it's something that for some reason. +[1613.54 --> 1620.14] You know, when you do either a Greenfield project or the pull request is a huge refactor. +[1620.46 --> 1621.70] You know, everything changed. +[1621.76 --> 1623.20] So pretty much new code. +[1623.20 --> 1629.02] So on that rare occasions, I might open the code itself, right? +[1629.38 --> 1632.94] Because then I can read in like an execution order, let's say. +[1633.68 --> 1637.54] And then my comments, they are probably going to be comments on the code itself. +[1638.22 --> 1641.40] And then I've done it, I don't know, once or twice. +[1641.40 --> 1644.90] And then I create my branch out of this branch. +[1645.12 --> 1647.62] And then I open a pull request for this branch. +[1647.76 --> 1651.96] So the person can see my comments in the code without having to look for it. +[1652.34 --> 1657.68] But this is pretty much in either Greenfield projects, you know, when you start something new. +[1658.20 --> 1661.70] Or when you're just adding so much new code. +[1662.64 --> 1666.16] That the pull request itself, it's hard because it's completely out of road and everything. +[1666.34 --> 1667.44] And there's a lot to comment on. +[1667.80 --> 1669.10] Yeah, that's the hardest ones, right? +[1669.10 --> 1671.64] When there's so much to handle there. +[1672.22 --> 1672.46] Yeah. +[1672.96 --> 1678.14] Do you sometimes find yourself rereading the whole thing to kind of once to read it and second time to make sense? +[1678.94 --> 1679.24] Oh, yeah. +[1679.30 --> 1680.00] No, yeah, definitely. +[1680.18 --> 1680.42] Definitely. +[1680.52 --> 1686.26] I think, yeah, going back to like old goal time when they talk about documentation and reading documentation. +[1686.56 --> 1687.76] I have the perseverance. +[1688.14 --> 1692.06] If I'm not understanding, I'm going to read it over and over and over again. +[1692.06 --> 1694.08] And I'm going to do my best to understand. +[1694.40 --> 1696.64] If I don't, I'm going to ask, right? +[1696.64 --> 1703.36] But I think if someone thinks puzzling me, then I'm probably going to blog say, look, I think this is important. +[1703.48 --> 1704.06] I don't get it. +[1704.12 --> 1704.88] Please explain. +[1705.34 --> 1705.48] Yeah. +[1705.82 --> 1707.14] Do you comment as you go? +[1707.24 --> 1712.86] Or do you read through fully, digest and then go through and do all your comments? +[1713.62 --> 1714.06] No, no. +[1714.18 --> 1716.36] I comment as I go, like as I stream. +[1716.48 --> 1716.68] Okay. +[1716.68 --> 1720.98] And then when I get down there, I was like, oh, yeah, that's the why. +[1721.10 --> 1722.40] So I go back and delete the comments. +[1722.54 --> 1722.96] Oh, okay. +[1723.16 --> 1725.12] You know, it's like, dude, does it make sense? +[1725.24 --> 1726.52] You know, something like that. +[1727.00 --> 1727.36] Why? +[1727.48 --> 1728.32] How does it work? +[1728.38 --> 1729.08] And then you read it. +[1729.56 --> 1729.64] Oh. +[1730.84 --> 1732.98] And then I go back, delete and edit the comments. +[1733.18 --> 1733.60] Go back. +[1733.66 --> 1734.34] Oh, I'm so sorry. +[1734.42 --> 1734.92] Never mind. +[1735.84 --> 1738.30] I mean, on GitHub, you don't submit the review, right? +[1738.48 --> 1738.70] Yeah. +[1739.02 --> 1740.44] So I just delete. +[1740.58 --> 1741.84] But the feeling is exactly that. +[1741.92 --> 1743.90] I'm apologizing for asking something stupid. +[1743.90 --> 1752.16] When you're going over a PR, do you feel like, or any reviewer should test the changes? +[1753.06 --> 1758.40] And to what extent should you test the changes if you think that they should be tested by the reviewer? +[1758.86 --> 1760.90] I think that's the $1 million question. +[1761.24 --> 1762.06] Give us the answer. +[1762.20 --> 1762.66] We're ready. +[1762.94 --> 1764.32] Maybe another popular opinion. +[1764.32 --> 1768.14] No, I think the answer is it depends, right? +[1768.26 --> 1768.38] Right. +[1768.72 --> 1770.44] We say a lot of that at the last take. +[1770.44 --> 1774.30] But I think it depends on your role on this pull request, right? +[1774.38 --> 1783.40] So if imagine that I maintain an open source and I'm taking something from a person that's not from the repo, I would be extra careful. +[1783.64 --> 1786.42] So perhaps I would try to run the code. +[1786.84 --> 1790.78] Usually I don't really test the code, right? +[1790.78 --> 1798.74] But again, if it's something that I feel that's really critical and I want to be sure that's 100% working, I might test it. +[1799.38 --> 1800.42] But it's rare cases. +[1801.12 --> 1801.26] Okay. +[1801.70 --> 1806.50] And does the length of the PR or the scope of the change change that opinion? +[1807.06 --> 1807.34] Yeah. +[1807.66 --> 1807.92] Okay. +[1808.10 --> 1809.84] I think the length, not so much. +[1810.26 --> 1810.88] Let's be honest. +[1811.12 --> 1815.28] The longer the PR, the less detailed is the review. +[1816.04 --> 1816.62] We are humans. +[1816.70 --> 1817.18] We get tired. +[1817.44 --> 1817.62] Yeah. +[1817.90 --> 1818.24] Come on. +[1818.28 --> 1823.22] Like, if you're reviewing like 15, 20 files, the last one, you're tired. +[1823.34 --> 1824.98] Like, it's just a human thing, right? +[1825.76 --> 1827.82] And then it's something that I've done a few times. +[1828.10 --> 1829.74] I don't like it, but it happens. +[1829.88 --> 1832.56] Like, the longest PRs, I review and then I submit. +[1833.04 --> 1835.48] And then when I'm doing the re-review, I find new things. +[1835.52 --> 1837.86] I'm like, I can't let it pass. +[1837.86 --> 1844.16] I'm going to have to put a comment where there wasn't a comment and it was not changed because now I saw it, right? +[1844.98 --> 1845.74] That's the thing. +[1846.20 --> 1850.94] So, given that, what is a reasonable time to expect a PR review? +[1851.36 --> 1857.40] If you, like, put in a PR today, is it the next hour, like, by end of day, the next day, a week? +[1857.92 --> 1858.78] Does it depend? +[1859.44 --> 1861.04] The real answer is it depends. +[1861.52 --> 1862.84] I got ahead of you there, I knew. +[1862.84 --> 1863.32] Yeah. +[1863.98 --> 1866.28] My first job, we... +[1866.28 --> 1867.52] Did we use Jira? +[1868.02 --> 1868.62] I don't know, whatever. +[1868.70 --> 1869.76] We used columns, right? +[1869.90 --> 1870.44] We used columns. +[1870.60 --> 1872.46] So, there was the column PR review. +[1872.64 --> 1874.76] So, the open PRs were there for review. +[1875.26 --> 1877.66] And then there could be, I think we were three people. +[1877.82 --> 1881.02] So, there could be only two PRs on review. +[1881.36 --> 1882.88] So, do you want to put something for review? +[1882.98 --> 1884.22] We have to take something to review. +[1885.02 --> 1885.20] Right? +[1885.26 --> 1889.56] So, this happened, helped to keep the, you know, the process running, everyone reviewing. +[1890.76 --> 1892.82] Nowadays, like, I think it's a lot of people. +[1892.82 --> 1897.06] I think when you're really running a Jira startup, microservice, you know, just deploy, +[1897.46 --> 1901.30] you usually expect something in the next day to get an answer. +[1902.10 --> 1905.66] Not elastic, within the week, I'd say. +[1906.18 --> 1906.38] Okay. +[1907.16 --> 1910.62] And do you have different commitments when it comes to internally, like, your internal +[1910.62 --> 1914.18] team PRs versus people who are maybe contributing to your service? +[1914.56 --> 1918.04] I.e., like, in our system, we have a lot of external teams that will contribute to our +[1918.04 --> 1919.92] service and ask for PR reviews. +[1919.92 --> 1925.98] Like, what is a reasonable timeline to commit to review those external PRs? +[1926.40 --> 1928.16] I think there are two categories, right? +[1928.20 --> 1931.26] If it's just a normal flow, they go in the same flow. +[1931.82 --> 1937.16] But if it's something that someone external is doing because our team doesn't have capacity +[1937.16 --> 1939.38] and then it is really important, right? +[1939.58 --> 1941.74] Probably, I would try to prioritize this review. +[1942.52 --> 1948.60] But also, if someone that's not from the team or doesn't know the rapid conventions, it's +[1948.60 --> 1950.94] probably going to be a more thorough review. +[1950.94 --> 1955.40] I strongly believe that your code should be consistent, right? +[1955.44 --> 1959.32] I'd rather have something that they don't like, but it's consistent and it's always +[1959.32 --> 1963.54] there, than half of the code they like, half of the code they don't like, and another third +[1963.54 --> 1964.64] they don't even have an opinion. +[1965.90 --> 1971.64] So, in external reviews, I think there's the extra consistent thing in code conventions +[1971.64 --> 1974.14] from the repo that you have to put through. +[1974.86 --> 1977.04] And then it should be, I mean, it's better to be quicker. +[1977.58 --> 1978.16] For sure. +[1978.60 --> 1980.30] I really like that column policy. +[1980.42 --> 1981.78] I might have to implement that on my team. +[1982.10 --> 1982.72] Yeah, right. +[1982.90 --> 1985.08] You can't put a PR on if they're already two. +[1985.16 --> 1986.12] You have to review them. +[1986.70 --> 1987.90] Yeah, it makes things... +[1987.90 --> 1988.50] I love that. +[1988.56 --> 1989.12] ...to move. +[1989.30 --> 1990.26] I think it's nice. +[1990.82 --> 1991.32] I agree. +[1991.60 --> 1992.34] It's a weight group. +[1992.78 --> 1995.00] We pretty much describe this concept now. +[1995.30 --> 1996.70] It's a channel with a buffer. +[1996.70 --> 1997.14] Yeah. +[1998.34 --> 2001.86] And then I'm going to have to be the bad girl who comes into Slack and someone's like, +[2001.90 --> 2004.10] oh, I'm ready to put my PR on this big new feature. +[2004.32 --> 2006.08] And I'm like, you're not allowed to. +[2006.72 --> 2007.12] Counterful. +[2007.44 --> 2008.80] Go review Bob's PR. +[2009.20 --> 2009.80] Throwing air. +[2011.24 --> 2011.52] Yeah. +[2012.02 --> 2014.36] There's the poking PR review. +[2014.60 --> 2017.54] You only get your PR reviewed when you poke someone, right? +[2017.64 --> 2018.36] You don't want to add. +[2019.76 --> 2020.12] Exception. +[2022.64 --> 2026.58] It can be a fun way of teaching all sorts of Go concepts now that... +[2026.70 --> 2028.22] This gave me some ideas. +[2028.32 --> 2028.68] Thank you. +[2028.86 --> 2029.98] We're doing a PR review? +[2030.54 --> 2031.46] By poking people? +[2031.92 --> 2032.24] Yeah. +[2032.66 --> 2033.58] With limiting this. +[2033.70 --> 2035.82] This is a fun way to discuss this. +[2035.96 --> 2039.96] And on the way, you discuss errors, throwing incorrect errors and also exceptions and so on. +[2040.16 --> 2040.56] Okay. +[2041.08 --> 2043.68] Go routines if you suddenly have to split into that. +[2043.84 --> 2043.92] Yeah. +[2044.48 --> 2046.70] Talking about teaching Go in unusual ways. +[2046.70 --> 2057.72] I was thinking today, someone should write a Go program that simulates how the queue for the food works here and then make a proper Go concurrent good program for that. +[2057.98 --> 2060.02] Because the queue is unnecessary here. +[2060.44 --> 2061.84] We have a lot of contingents. +[2062.20 --> 2065.10] I was like, you know, you can make better concurrency here. +[2065.10 --> 2067.80] Because you have lots of food stations that people miss. +[2068.10 --> 2068.64] That's the point. +[2068.70 --> 2069.74] You have a lot of food stations. +[2069.86 --> 2072.56] You can have a lot of concurrent access to that. +[2072.90 --> 2073.94] But no, you get sequential. +[2074.12 --> 2077.00] You get a huge queue and everyone goes through everything that they don't want. +[2077.72 --> 2079.72] I think because we're just out of the... +[2079.72 --> 2080.70] They didn't read the docs. +[2080.94 --> 2081.94] They don't know what's the food. +[2082.04 --> 2082.86] They didn't read the docs. +[2082.96 --> 2083.30] Exactly. +[2084.10 --> 2085.66] Everything can be explained with pick. +[2085.96 --> 2086.78] Lesson learned. +[2086.98 --> 2088.80] Always read the docs first. +[2089.04 --> 2091.10] And if they're bad docs, then... +[2091.84 --> 2092.44] Improve them. +[2093.08 --> 2093.64] Improve them. +[2093.64 --> 2095.00] Open a pull request for the docs. +[2095.16 --> 2095.54] Yes. +[2097.10 --> 2098.40] That's something that I love. +[2098.66 --> 2107.36] If I'm reading documentation that it's easy to open a pull request and to see a failure or inconsistence or something, I open the pull request. +[2107.50 --> 2112.30] I think it's such a valuable contribution and so easy most of the times. +[2112.86 --> 2118.48] I love these docs that they have the button edit and then you go direct to GitHub to create a pull request. +[2118.54 --> 2119.08] That's fantastic. +[2119.54 --> 2119.74] Yeah. +[2119.74 --> 2122.14] I feel like I get into a bit of a rabbit hole. +[2122.14 --> 2134.42] I had to stop myself editing documentation because it went from actually making it correct to actually just implying my personal stylistic choices when writing documentation and phrasing. +[2134.94 --> 2138.00] I like this adjective slightly better actually. +[2138.00 --> 2143.38] So I had to pull back to be like, okay, review for correctness, not for like, I want a comma here. +[2144.40 --> 2145.78] I think that's so hard. +[2146.26 --> 2152.02] And for me, as a known native speaker, sometimes I was like, I don't think this sentence is correct. +[2152.16 --> 2153.04] I think it's missing a comma. +[2153.14 --> 2154.06] I think it's missing an article. +[2154.42 --> 2156.66] And I was like, honestly, you don't know English so much. +[2157.28 --> 2160.72] I don't even know if you could do that in proper Portuguese, like a proper grammar. +[2160.72 --> 2162.16] But I think it's important. +[2162.54 --> 2170.02] My take is if I believe it's compromising their understanding, I'm going to probably suggest something. +[2170.52 --> 2170.64] Right. +[2171.06 --> 2178.20] And also sometimes when the comment is there for a long time, I just, I make the change and suggest someone is going to review that. +[2178.28 --> 2178.44] Is it? +[2178.84 --> 2184.90] I sometimes put into some AI if I don't understand something and I read two, three times and I keep staring at it. +[2184.90 --> 2187.16] I'm like, just explain that to me in other words. +[2187.36 --> 2188.00] And that helps. +[2188.74 --> 2189.26] Good AI. +[2190.22 --> 2191.86] It's like pinging somebody, but yeah. +[2192.54 --> 2198.92] But also make the changes just like you, because I think if I'm as a non-native don't understand this, there must be another non-native that gets lost there. +[2199.30 --> 2199.60] Yeah. +[2199.84 --> 2200.96] And clarity is important. +[2201.90 --> 2208.70] And let's say you're interviewing, whether you are the candidate or you are the interviewing person. +[2209.34 --> 2214.18] And part of the interview is reviewing a pull request from somebody from your team. +[2214.90 --> 2217.98] What tips do you have for somebody to do this well? +[2218.66 --> 2223.36] Actually, never been really on these shoes, neither side. +[2224.04 --> 2229.72] I've been asking, oh yeah, in one of your code bases, any of your code base, what you would change or something. +[2230.32 --> 2237.14] I think at the end of the day, a lot of the time we, to interview for culture fit, right? +[2237.22 --> 2239.62] And a person that's nice, it's good to work with. +[2240.34 --> 2241.26] This is super important. +[2241.26 --> 2247.98] So I think it's, if you're on an interview, just be sure to be nice, right? +[2248.04 --> 2251.28] In your comments and everything, don't go like, oh yeah, this is crap. +[2251.38 --> 2252.10] This is bad. +[2252.36 --> 2253.48] Just be nice. +[2253.58 --> 2254.18] Be polite. +[2254.92 --> 2255.98] Link the commentations. +[2256.28 --> 2260.10] And I advise you to bring arguments, right? +[2260.10 --> 2264.60] Don't say, okay, do that or change that or this need change without a reason, right? +[2264.60 --> 2268.48] If an interview, usually an interview don't have so much time, right? +[2268.56 --> 2273.14] So I would go for, oh yeah, you know, this name is not ideal. +[2273.62 --> 2279.38] You know, the goal, effective goal, there is a section on name convention that explain why it should be like that. +[2279.52 --> 2282.38] So as it's in goal, it's better to be like that. +[2282.68 --> 2286.94] So try always to, to bring something to support your views. +[2286.94 --> 2293.08] And when it's opinion, and that's something that I really do on pre-request, when it's like, it's my opinion, I say like, look, this is my opinion. +[2293.20 --> 2298.74] I believe that's better because this, this, and this is up to you because I don't see a flaw here. +[2299.00 --> 2301.70] I just think it can be better about any opinion. +[2302.56 --> 2311.52] If you were interviewing someone and their task was to review a PR, what would be things that they did that would maybe be like, oh no, I don't know about that? +[2311.52 --> 2315.00] I think it would be to be aggressive, right? +[2315.10 --> 2320.88] And impolite to just diminishing the code and say like the code is bad or something. +[2321.60 --> 2326.48] In interviews, they really, they show that they know, don't know what they're doing. +[2326.90 --> 2333.24] I think if you interview people, you know, some people, they don't know what they're doing or just trying to fool you. +[2333.62 --> 2333.98] Right? +[2334.04 --> 2336.02] If it's hard to see that it's like, oh, no, no. +[2336.42 --> 2337.74] It's better to say you don't know. +[2337.92 --> 2338.74] It's not like that. +[2338.78 --> 2339.30] It's like, great. +[2339.30 --> 2339.86] Just go. +[2339.96 --> 2341.44] And you know, I just, okay. +[2341.52 --> 2341.76] Yeah. +[2342.04 --> 2342.38] Mm-hmm. +[2342.52 --> 2343.76] It's just an incentivize. +[2343.84 --> 2344.20] Yeah, go. +[2344.44 --> 2344.92] Go, please. +[2345.62 --> 2347.18] And then, oh yeah, thank you very much. +[2347.20 --> 2348.04] You're going to be in contact. +[2349.36 --> 2349.72] Yeah. +[2349.86 --> 2352.88] And definitely staying honest is a lot better than making things up. +[2353.30 --> 2353.50] Yeah. +[2353.54 --> 2354.42] Please say, I don't know. +[2354.46 --> 2357.34] I think if someone to me in an interview say, look, I don't know. +[2357.38 --> 2358.00] I don't remember. +[2358.50 --> 2359.52] Oh, I don't know. +[2359.58 --> 2361.72] And I think in that place I can get information. +[2362.14 --> 2365.30] Dude, folk, you're like scoring a hundred points with me. +[2365.54 --> 2368.58] If you're trying to just BS me through. +[2369.06 --> 2369.18] Nah. +[2369.60 --> 2370.04] That's fair. +[2370.04 --> 2371.94] Anderson will not be having that. +[2371.94 --> 2401.92] this episode is brought to you by our friends at chronosphere scaling cloud native is complicated +[2401.92 --> 2407.20] and chronosphere helps teams take back control of observability team rampant data growth reduce +[2407.20 --> 2412.08] cloud native complexity and increase confidence of the business and i'm here with martin mal co-founder +[2412.08 --> 2416.96] and ceo of chronosphere martin when it comes to cloud native observability what are the pain points +[2416.96 --> 2422.06] of kubernetes and making sure it's reliable you know i think the shift to kubernetes has really +[2422.06 --> 2428.60] changed the way we design applications it's changed the way we it's changed our infrastructure as well +[2428.60 --> 2432.56] so it's introduced a lot of change i would say and that's probably why it's causing a lot of +[2432.56 --> 2438.76] issues in the observability space i think one thing we're finding is that a lot of companies out there +[2438.76 --> 2445.12] are focused on producing a lot more data and there's a lot of focus on more metrics more traces more logs +[2445.12 --> 2450.00] because these environments we're trying to monitor are far more complex these days i think that's +[2450.00 --> 2454.68] maybe one of the mistakes the industry is running into and it's interesting because obviously for +[2454.68 --> 2459.20] all the solutions out there the vendors out there the more data that gets produced the better it is +[2459.20 --> 2464.94] for all the vendors out there but what's interesting is that along with that increased volume of data +[2464.94 --> 2470.38] people aren't actually getting better outcomes out of it people's number of incidents that people are +[2470.38 --> 2476.40] running to still rising people's mttrs mttds meantime to detection and resolutions actually getting +[2476.40 --> 2481.34] higher as opposed to lower so i think this is the common state that a lot of companies find themselves +[2481.34 --> 2486.46] in and of course with the increased volume of data folks bills increase and the problem actually +[2486.46 --> 2490.70] gets harder so i think that's a common state we find a lot of companies into and this is probably +[2490.70 --> 2494.94] why it's top of mind for a lot of companies out there very cool thank you martin all right the next +[2494.94 --> 2501.06] step is to head to chronosphere.io to explore the platform and get a demo again chronosphere.io +[2501.06 --> 2530.98] okay so we kind of touched on this a little earlier in the episode +[2530.98 --> 2537.48] but i want to dig a little bit deeper in fact i will ask you a question first are you engaged in +[2537.48 --> 2541.54] any kind of open source projects i know you said that you you know contribute to go a little +[2541.54 --> 2550.02] is that a world in which you feel like you have engaged and put prs in so i can cheat my answer +[2550.02 --> 2555.58] right okay yeah as i'm evolving open source because i work at the last the majority of our +[2555.58 --> 2562.46] repels are open source right but as a 100 open source contribution that i'm not working for the +[2562.46 --> 2568.50] company or not no i read you it's something that i always wanted i just said like i managed to get +[2568.50 --> 2577.10] a committed to in go but i haven't fully got to participate in a project i still try oh no today i think i +[2577.10 --> 2582.32] got one on kubernetes too you see like it's one of my goals you know these plans that things you want +[2582.32 --> 2587.00] to do also have to try a bit you know it goes back and forth i may go for a coin again perhaps +[2587.00 --> 2592.56] you know the flame just like lighting up again reignite that passion that's gonna happen but +[2592.56 --> 2597.96] yeah no i'm not really engaged on let's say an external open source project that's not part of +[2597.96 --> 2603.90] my daily job and when you have done it is part of the reason why you think it's difficult to engage +[2603.90 --> 2609.22] fully is anything to do with that to do with like the difficult or the different process to put in a +[2609.22 --> 2615.62] when it is an open source project as opposed to internal like within work pr reviews and submission +[2615.62 --> 2620.30] i think to me what's always difficult like to find something meaningful to work +[2620.30 --> 2627.50] right sometimes you don't know what you can do there is a tag first good issue and so but +[2627.50 --> 2632.16] i think you're a lot i think yeah that's the point we're lost and you don't have someone to go +[2632.16 --> 2639.06] please help me yeah right or i try that or shall i do that right because my project and my team are +[2639.06 --> 2644.36] comfortable to go and do a refactor yeah it's like complete external project that i don't know +[2644.36 --> 2650.88] anyone there or anything i'm gonna be afraid you know and sometimes you're not even able to run +[2650.88 --> 2656.58] the project so that i think that's a lot of the barrier i think if you would have something +[2656.58 --> 2662.66] not necessarily a mentor but you know perhaps like a channel ask questions right like oh i want +[2662.66 --> 2667.62] to get even that like say oh i want to take this issue because sometimes they're good first issues +[2667.62 --> 2672.92] they open like one year ago i was like dude i don't know if it's worth to fix that or not and then +[2672.92 --> 2677.68] you open the pull request you fix everything no one reviews and then you think no one reviews +[2677.68 --> 2683.42] nothing got stale you get demotivated yeah fair enough and do you think there's a higher bar +[2683.42 --> 2688.08] as to what you're willing to put in as a pr for an open source project i speak about this from my +[2688.08 --> 2692.76] own personal like experience when i was trying to get into like okay i want to contribute to open +[2692.76 --> 2697.68] source oh you should just go in and do like a little change but i was like yeah but i feel like +[2697.68 --> 2703.86] i'm contributing to an open source library where everyone can see and they're like oh angelica made +[2703.86 --> 2710.16] like a one line change or like change that one function name i think and if i talk for myself +[2710.16 --> 2715.02] i always had the feeling to contribute to open source pros the bar is super high yeah +[2715.02 --> 2722.24] are you having to be like an expert developer and everything and the reality is no right and +[2722.24 --> 2730.06] the issue is there the problem exists i think this is the best advice i can do to any junior developer +[2730.06 --> 2736.70] just go for it they know you already have right so your change is not there the bug is not fixed +[2736.70 --> 2743.64] the documentation not proved the feature is not there right so this is not gonna change if your +[2743.64 --> 2748.86] change didn't get there you learn something you play with a new technology i was trying to submit +[2748.86 --> 2755.48] pull requests for the kubernetes code just fixing leading issues i understand a bit how the that piece +[2755.48 --> 2760.56] of code worked and i was like oh how they instructed that and like look at the packages because i had to +[2760.56 --> 2766.20] read through the packages to fix leading issues i think one pr got merged the other ones got stale +[2766.20 --> 2773.56] this life it happens and i learned something yeah so try go for it let the other one say no it's not +[2773.56 --> 2781.48] your job to say no for you right it's their job that is a popular opinion i bet applies to many fields +[2781.48 --> 2792.10] in life that's true just do it yes try yeah well done all right the fun part unpopular opinions +[2792.10 --> 2797.28] before we started the recording anderson you mentioned you have several unpopular opinions +[2797.28 --> 2802.20] and you were wondering whether you should go for the most yeah how did you phrase that i forget you +[2802.20 --> 2809.20] used a good word the most controversial yeah exactly or the least controversial yeah now i'm taking +[2809.20 --> 2816.24] i'm picking a controversial one yeah right but i can explain as i said in the pr you can explain +[2816.24 --> 2823.96] you should not write more than 100 columns right you write your code your code should not really +[2823.96 --> 2832.88] pass 100 columns from which yeah first things there's no magic numbers like 100 cut i would say 110 +[2832.88 --> 2839.32] it's okay when it's really bad to cut right 120 is almost a hard limit don't really go over that +[2839.32 --> 2848.52] why first things first do you read books on landscape no right come on i think everyone had +[2848.52 --> 2854.12] that right you got like this email right you're like in your four key or ever monitor and that thing +[2854.12 --> 2859.92] goes from side to side and you're reading for the listeners i'm moving my head as you know reading +[2859.92 --> 2863.94] from one side to the other like you're watching tennis you feel like you know a typewriter that goes +[2863.94 --> 2873.90] so it's hard to read right because we don't read in landscape we read in portrait the second +[2873.90 --> 2880.92] thing is not everyone has got ice cream as big as yours there are people coding 14 13 inches right +[2880.92 --> 2887.50] they want to have two tabs open perhaps so if you go much more than 100 it's going to be bad for +[2887.50 --> 2894.00] some people and i believe for everyone too long it's hard to read so that's my unpopular opinion +[2894.00 --> 2900.34] i want to disappoint you that i think i agree with you and i even take this into writing emails that +[2900.34 --> 2906.38] i try to keep that i don't know how many characters that is but i sometimes break lines like one sentence +[2906.38 --> 2912.02] into three four lines just so it stays so you don't have to scroll in case yeah images logos i don't know +[2912.02 --> 2916.14] whatever happens in somebody's signature that it suddenly gets stretched i feel like you're +[2916.14 --> 2922.66] unfortunately preaching to the choir with me and natalie we're both like yes please i don't like +[2922.66 --> 2929.50] my neck will hurt perhaps in go a chrome plugin that just truncates things for you yeah perhaps +[2929.50 --> 2934.62] in go but i think if you go to java you know the things are long there yeah 100 characters is just +[2934.62 --> 2942.34] the function name oh my gosh yes i feel like that one was a good one but like if we have time natalie i want +[2942.34 --> 2948.30] to hear another one yeah i want us to get like an unpopular one from you oh my god what was the +[2948.30 --> 2952.24] other one that you were thinking about saying that you cheekily in your mind were like no i'm not gonna +[2952.24 --> 2957.44] say that no i think this one's like the unpopular kind of popular okay i mean i see a lot happening so +[2957.44 --> 2964.08] perhaps in unpopular return new is wrong period right you have to wrap the errors and add more context +[2964.08 --> 2971.30] always i cannot count how many times i had to go to the code and dig deep and deep and deep to discover +[2971.30 --> 2978.04] where this error came from because i you know it's like when you get like a you try to write to the +[2978.04 --> 2982.76] disk and get an error you get something like i got one too many colors in the address it's like +[2982.76 --> 2988.58] this is finally the address how there is too many columns right and then you have to understand where +[2988.58 --> 2993.56] this address was going to be used which method it was and then oh yeah on this context there are too many +[2993.56 --> 2998.68] columns but they didn't have this information so return new is wrong you have to wrap your errors +[2998.68 --> 3003.54] now you don't have an excuse you have ever wrapping when they send the library so i think wrapping errors +[3003.54 --> 3009.52] will be an unpopular opinion yeah i feel like this this second one is gonna probably be more unpopular +[3009.52 --> 3016.04] yeah because a lot of people just return yeah yeah interesting that's one of the things that i'm looking +[3016.04 --> 3021.70] for requests and was like dude could you wrap that and then it goes back on that if it's my repo +[3021.70 --> 3028.44] my code i might say like no now when i say my please like my teams right i don't have this +[3028.44 --> 3036.06] possessive i think code must be owned by a group and it must be a consensus group but yeah this is the +[3036.06 --> 3041.96] thing that i'm gonna point and usually if it's an external adding code even more important like okay +[3041.96 --> 3049.28] like your wrapper your rules our wrapper our rules right so okay here we wrap do that if i'm owning your +[3049.28 --> 3055.02] wrapper i play by your rules i like that one i'm also having so many more ideas this always happens +[3055.02 --> 3059.82] when me and you have episode nathia i'm like so many more episode ideas write them all down does +[3059.82 --> 3065.68] your code belong to you or to the world i mean on an open source that's definitely a question right +[3065.68 --> 3072.48] yeah also with ai tools that are writing code that's a question yeah oh god that's yeah who is the true +[3072.48 --> 3078.64] either code yeah that's a good one yeah licensing is interesting for sure with a copilot and friends +[3078.64 --> 3085.58] when the ai put the bug in production who do you blame who run the ai who wrote the ai +[3085.58 --> 3092.22] or who reviewed the pr that's a good one right can ai review a pr for sure for sure +[3092.22 --> 3100.00] would you trust that i might have used that in the past yes oh that's nice okay i'm a big fan of ai +[3100.00 --> 3106.08] and coding i think it's a fun combination i'm very happy to automate myself out of job that's good +[3106.08 --> 3113.62] but i confess i know almost zero about ai and coding but i'm super interested looks really interesting to +[3113.62 --> 3119.44] see where is it going yeah yeah definitely um both from the side of writing and from the side that's +[3119.44 --> 3125.42] relevant to this episode which is the reviewing and like explaining and so on i think in general +[3125.42 --> 3132.06] ai is they can see the context and it carries so much information see part of the things that we just +[3132.06 --> 3140.04] can't right and sometimes someone read experience can but they cannot teach it so there is definitely a +[3140.04 --> 3145.88] lot of value yeah yeah exactly about the context in particular i sent a twitter poll natalie would you +[3145.88 --> 3151.56] let an ai review your code maybe this can be my unpopular opinion for this episode +[3151.56 --> 3159.00] i would like not have it just review and say good bad but i would use it as something like here is the +[3159.00 --> 3164.90] code what does it do or list the problems and then you know read the output and give it a secondary +[3164.90 --> 3172.58] review i think this might end up being a popular opinion true now you've explained it and want us over +[3172.58 --> 3179.76] to your side i feel like it will be popular yeah okay my next chrome plugin each episode is like 15 other ideas yeah +[3179.76 --> 3187.96] summarize this PR for me yeah well this has been fun and this has been inspiring +[3187.96 --> 3192.76] anderson thanks a lot for joining us and uh sparing some of your time at gopher.com uk +[3192.76 --> 3198.62] thank you very much for having me it was really good i'm really happy thank you so much we're gonna have to +[3198.62 --> 3205.10] get you back on for the who owns our code episode yeah i would love it it's a plan and everybody i hope +[3205.10 --> 3210.54] you will also join us and thanks for joining us this time bye thank you very much bye bye +[3210.54 --> 3221.32] all right that is go time for this week thanks for listening now is the best time to subscribe if you +[3221.32 --> 3227.82] haven't yet head to go time.fm for all the ways and if you are a regular go time listener check out +[3227.82 --> 3233.64] our membership program directly support our work save yourself some time by ditching the ads and get +[3233.64 --> 3241.20] bonuses like exclusive content and free stickers check it out at changelog.com slash plus plus thanks +[3241.20 --> 3248.16] again to our partners at fastly for cdn-ing for us to fly.io for serving up our app to the mysterious +[3248.16 --> 3252.88] breakmaster cylinder for these dope beats and to you for being part of the go time community +[3252.88 --> 3258.62] we appreciate you that is all for this week we'll talk to you next time on go time +[3258.62 --> 3268.62] you diff --git a/The funny bits from 2021_transcript.txt b/The funny bits from 2021_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..51ee7562b61c590d7cfb0a767711eb33d8bbebaa --- /dev/null +++ b/The funny bits from 2021_transcript.txt @@ -0,0 +1,679 @@ +**Jerod Santo:** Whaddup, Gophers? Jerod here, coming at you with a little bonus episode before we get back to your regularly scheduled Go Time. We're calling this "The Funny Bits." It's a compilation of times we cracked up making the show for y'all. If you dig it, let me know. If you don't, email Mat Ryer. + +First up we have episode 161. This was our very first Go Panic game. Everyone teams up on Mat, and - well, here we go. + +**L Körbes:** I was expecting to be able to make fun of Mat, but he wasn't playing, so that didn't work out... + +**Mat Ryer:** It's hard to make fun of me, mate. + +**L Körbes:** No, it's not. + +**Mark Bates:** That's not true at all... + +**Mat Ryer:** Go on, then. Let's do a quick insult. + +**Kat Zień:** It'd be so fun if Mat was playing and we had a question about the book that Mat Ryer wrote, and you got it wrong... That'd be so funny. \[laughter\] + +**Mark Bates:** Mat, next time might I recommend renting some chest hair for under your shirt, if you're gonna open it up that much... + +**Mat Ryer:** Oh, it's not open far, is it? Would you rather like that, like some Puritan? + +**Mark Bates:** Well, far enough to know that you need to rent chest hair next time... + +**Mat Ryer:** \[laughs\] Where do you rent chest hair from? + +**Mark Bates:** I don't know... + +**Kat Zień:** The wig shop? + +**Mark Bates:** ...the Burt Reynolds Museum? \[laughter\] + +**Mat Ryer:** Can you just use a wig? Surely it has to be dedicated. You can't just do that... + +**Kat Zień:** Well, you know how some men just get the patch; you can get a patch of hair for the top and just glue it here. + +**Mat Ryer:** Yeah. If I'm going to a wig shop, I'm gonna wear a wig. + +**Mark Bates:** Why don't you have them do you top to bottom? I mean... \[laughter\] + +**Mat Ryer:** Yeah. \[unintelligible 00:02:03.22\] + +**Mark Bates:** Just go in there and say "I want the works. Top to bottom, I want hair." + +**Mat Ryer:** Yeah. I'd love that. And then they'd just set to work with the Pritt stick, sticking it all on. + +**Mark Bates:** Exactly. It'll be like that scene in Wizard of Oz, where they're buffin' up the Tin Man, and they're stuffin' the Scarecrow with \[unintelligible 00:02:23.05\] and everything. + +**Mat Ryer:** Yeah. I'd love that, mate. I'd love that. I'd be a sort of character... But not, I have a receding -- but I like to say what I like in hair, I more than make up for in forehead. \[laughter\] + +**Mark Bates:** That is true. + +**Jerod Santo:** Up next we have two clips from our Indecent Language Proposals mini-series. These are episodes 166 and 168, featuring a food gun and a centaur. + +**Mat Ryer:** What's the phrase you're using? + +**Daniel Martí:** As my unpopular opinion, you mean... + +**Mat Ryer:** No, no. The foodgun, you were saying, right? + +**Kris Brandow:** Footgun. + +**Mat Ryer:** Yeah. What's that about? Because I wanna explain that for anyone who's never heard it before. + +**Kris Brandow:** Oh... It's basically like if you have a gun - you know, usually you're trying to shoot other things with it, but instead it's gonna hit you in the foot, and that's bad. + +**Mat Ryer:** Oh, footgun. I thought you were saying food-- + +**Johnny Boursiquot:** What did you think we were saying? Foodgun? + +**Mat Ryer:** I thought you were saying foodgun. + +**Johnny Boursiquot:** Like, I'm gonna launch a burger straight at your mouth, or something? + +**Mat Ryer:** I couldn't figure out if it shot food out, or if you used it for shooting food. \[laughter\] + +**Johnny Boursiquot:** Like, you're gonna shoot up some food in the range today? \[laughter\] + +**Mat Ryer:** Oh, man... Yeah. I mean, if you want a language proposal, I propose that our American cousins pronounce their T's a little more, and then we wouldn't get in this mess... \[laughter\] That's a language proposal for me. I'm gonna actually open that. Can you open PRs for America? \[laughter\] + +**Jerod Santo:** \[04:02\] I think we should address why Mat would prefer to replace me with a horse. + +**Mat Ryer:** Oh, you were listening... \[laughter\] Not replace you with one, mate... + +**Jerod Santo:** Oh... Augment? \[laughter\] + +**Mat Ryer:** Yeah, like a centaur. More like a centaur, so you can still do your programming, but you've got a horse's back... + +**Jerod Santo:** \[unintelligible 01:09:08.11\] + +**Johnny Boursiquot:** I hope you're not attached to your legs... + +**Roberto Clapis:** Someone is writing in the chat -- actually, Bill is writing in the chat, "Daniel doesn't have any unpopular opinions." Wanna prove them wrong? + +**Jerod Santo:** Dan, if you wanna share, I'll splice it. + +**Johnny Boursiquot:** Yeah, he'll splice it. Do it. Do it. + +**Jerod Santo:** Or we'll put it in after the outro. Either way, it'll get in. + +**Mat Ryer:** Hang on - if we're splicing, can we have the centaur first? + +**Jerod Santo:** No, not for you. + +**Mat Ryer:** Can we have the centaur first? I'd love you as a centaur. \[laughter\] I don't splice anything for you, Mat. I'll slice things out, but I won't splice things in. Go ahead, Daniel. + +**Jerod Santo:** Here's a good one from episode 171 on Go Embed. No animals were harmed in the making of this clip. + +**Wayne Ashley Berry:** I think it was the bacon comment that did that. + +**Mark Bates:** Well, Mat's a vegan, so he's gonna be on board with that one. + +**Mat Ryer:** Yeah, I don't eat bacon. Don't tell everyone. + +**Carl Johnson:** Well, there's the fake bacon. Do you like fake bacon or no? + +**Mat Ryer:** I don't know why we're spending all this science energy trying to make fake meat... So no, I don't -- I mean, no. I don't get it. + +**Carl Johnson:** The times when I've tried vegan diets, that has mostly been my experience, is that all the fake meat is not worth it... But I do think that some of the fake bacon is okay. + +**Mat Ryer:** To be fair, there are now burgers that are very good. Impossible Burgers, and there's another one (I forget), that are, just as I remember eating burgers. And they're actually terrible for you as well, so... Bonus. \[laughter\] They didn't even bother to make them healthy. + +**Mark Bates:** All of the health benefits of a burger with none of the taste. + +**Mat Ryer:** Exactly. It's actually worse for you. It's less healthy than-- it's better for the animal, you could say... + +**Carl Johnson:** Well, but they make up for it by setting an oil refinery on fire every time... + +**Mat Ryer:** Exactly, yeah. That's what it tastes like, actually. + +**Carl Johnson:** It's like a Carbon offset, but in reverse. \[laughter\] + +**Mark Bates:** Mat has to drive 200 miles just to get one, so... + +**Mat Ryer:** Yeah. It's Carbon onset. + +**Carl Johnson:** Oh, it's a Carbon onset, yeah. + +**Jerod Santo:** Episode 173 featured Carlos Becker and GoReleaser. It also featured this touching moment between Johnny Boursiquot and Mat Ryer. + +**Johnny Boursiquot:** My very critical to you package is 50% covered with tests, but I'm very confident in that 50%. Would you put that in your mission-critical project? + +**Mat Ryer:** Yeah, I would, because I trust you, Johnny. \[laughter\] I would die for you, Johnny. I will release anything you tell me to. + +\[Bryan Adams, "Everything I do" mash-up 00:06:53.25\] + +**Jerod Santo:** Moving on without comment, I am acting like that did not just happen. Here's a question for you... Are frameworks getting an encore? That's actually the title of episode 180 with André Eriksson, which I was fortunate enough to join in a pinch, and needless to say, we had a lot of fun on that one. + +**Johnny Boursiquot:** \[08:04\] Now, here's my take. Jerod, you're not wrong. But. + +**Jerod Santo:** Okay... Is this a big but or a little but? + +**Johnny Boursiquot:** Well, I'll let you know. You can take a look at my but and let me know how big it is. + +**Jerod Santo:** \[laughs\] I'll judge. Okay, go ahead. + +**Johnny Boursiquot:** Okay, you can judge me. \[laughs\] So that whole notion of gophers don't like frameworks, we don't like magical things - yes, there is a lot of truth to that. But - that's where my but \[unintelligible 00:25:40.05\] + +**Jerod Santo:** Let's hear it, let's hear it. You're stalling... \[laughter\] + +**Natalie Pistunovich:** How do you give feedback on open source projects? Do you open an issue and say "I don't like this"? Is this how you do this? \[laughter\] + +**Jerod Santo:** There are plenty of examples of that you can look at... \[laughter\] + +**Natalie Pistunovich:** "It doesn't work on my machine." + +**André Eriksson:** Yeah, that usually goes over well... \[laughter\] + +**Jerod Santo:** "Let me tell you the ten reasons why this is terrible. One." \[laughter\] + +**Natalie Pistunovich:** Well, at least this is detailed, you know? + +**Jerod Santo:** Yeah... \[laughter\] + +**André Eriksson:** Another thing you can do is just open a pull request that deletes the whole project. \[laughter\] + +**Natalie Pistunovich:** That's a serious thing, yeah. + +**Jerod Santo:** That's not unprecedented. It's happened. Maybe not on Encore, but it definitely happened. + +**Johnny Boursiquot:** Awesome. + +**Johnny Boursiquot:** But I think giving folks a choice maybe of a hybrid model where they give the talk and maybe they do live Q&A, or do a pre-recorded talk and then a live Q&A afterwards would be also a nice choice to have. + +**André Eriksson:** Or live talk and pre-recorded Q&A... \[laughter\] + +**Natalie Pistunovich:** We've found it, the perfect formula. + +**Jerod Santo:** There you go. The perfect formula. + +**Jerod Santo:** Episode 184 with Carolyn Van Slyck was all about Porter, but we took a moment during the after-show to swoon on Johnny as well. + +**Johnny Boursiquot:** Now I have on my to-do list to check this out and also see if there are some ways I could help make the project better in some way. I'm looking forward to diving in. + +Carolyn Van Slyck: Oh my goodness, I'd probably swoon if I saw a PR from you... Just saying. \[laughs\] + +**Mat Ryer:** I would. Can you send me a link? \[laughter\] + +**Jerod Santo:** I would also swoon. Three swoons for the price of one, Johnny. + +**Johnny Boursiquot:** That's a lot of swooning. + +**Mat Ryer:** I'll faint. I'll faint like a proper olden days person. + +**Johnny Boursiquot:** Oh, swooning over some text? Okay... \[laughter\] + +**Mat Ryer:** Yeah. Brilliant. + +**Jerod Santo:** Sometimes the funniest bits are merely a result of the impedence mismatch between American English and British English. + +**Mat Ryer:** But now the beta is here, right? It's exciting. + +**Johnny Boursiquot:** And for those who don't speak, you know, Mat - by \[British accent\] beta, he means the beta. + +**Mat Ryer:** Oh, thank you very much. + +**Johnny Boursiquot:** It's just like "Beater." Who's the beater? Why are we beating people up? \[laughs\] \[unintelligible 00:10:42.26\] beating. + +**Mat Ryer:** Yeah, okay. I appreciate it. Thanks for the translation. If you want to correct my English into incorrect English more... + +**Johnny Boursiquot:** Yeah, please call me in. + +**Mat Ryer:** Yeah. It'll be great. + +**Johnny Boursiquot:** \[laughs\] + +**Mat Ryer:** Yeah. Apparently though, Shakespeare would have sounded more American if you heard him now... I don't know if you've heard that before. It's like -- it's not as simple as that. He'd be like \[unintelligible 00:11:06.27\] that kind of thing. + +**Katie Hockman:** Is that your American accent? I really like that, it's good. + +**Mat Ryer:** Yeah, it's Brooklyn. Shakespeare from Brooklyn. "Yo, Romeo!" + +**Jay Conrod:** That would have been great... + +**Mat Ryer:** "Yo, Romeo, where are you?!"He didn't know where Romeo was. Fair enough. + +**Jerod Santo:** Mark Bates was back for episode 190. It's called "How to make mistakes in Go." His presence alongside Johnny and Mat was so disruptively funny that we maybe should have named it "How to make mistakes in Go Time." + +**Mat Ryer:** I never thought I'd say this, Mark, but you should be a little bit louder. Move a little closer to the mic, please... + +**Mark Bates:** I should be a little louder... I don't know why you don't think I'm loud. + +**Mat Ryer:** Well, that's just what I'm hearing... + +**Mark Bates:** Well, maybe turn yourself up, did you ever think of that? + +**Johnny Boursiquot:** \[laughs\] + +**Mat Ryer:** \[11:59\] I don't think that solves the problem, I think that makes it worse, but that's fine... + +**Johnny Boursiquot:** Here we go, here we go... + +**Mark Bates:** I've got a common mistake in Go - working with Mat Ryer. \[laughter\] + +**Mat Ryer:** And that's from the president-- + +**Mark Bates:** I was gonna save that for the unpopular opinions, but I think it's a popular one, so... \[laughter\] So far only David Hernandez seems to be able to stick around and hang in there... + +**Johnny Boursiquot:** Multiple times, too. + +**Mat Ryer:** Don't use his name in vein. + +**Mark Bates:** Multiple times! I don't. I love David. He deserves all that he gets from putting up with you. + +**Mat Ryer:** I love it. I've never had a banter happening and insults concurrently, for both Johnny and Mark at the same time. \[laughter\] + +**Mark Bates:** We try, we try... + +**Mat Ryer:** Speaking of concurrency, here's another one... Teiva, you talk about concurrency like -- we feel like Go is a bit famous for concurrency, it's got some great primitives that makes that so much easier to do than previous languages... Should you always strive to use that to make things concurrent, do you think? + +**Teiva Harsanyi:** That's a nice transition... Thanks for that. \[laughs\] + +**Mat Ryer:** I've been professional. + +**Mark Bates:** I don't think I understand the question. + +**Mat Ryer:** Is concurrency always faster? + +**Teiva Harsanyi:** Yeah. + +**Mark Bates:** That's the question? + +**Mat Ryer:** That's the question. + +**Mark Bates:** No. Next question. \[laughter\] + +**Mat Ryer:** This is not a quiz show, Bates... + +**Mark Bates:** Okay, fine... Continue on with your line of questioning. I mean, it was a yes or no question. Is it faster? No, it's not always faster. + +**Mat Ryer:** Is it a common mistake that people make, that people think concurrency is always faster, Teiva? + +**Mark Bates:** Did you just call-- + +**Mat Ryer:** Teiva, you have to speak before Bates gets in, otherwise \[unintelligible 00:13:30.14\] + +**Johnny Boursiquot:** I keep seeing the setup, and I'm like "Teiva, jump in! Jump in! Jump in!" + +**Mark Bates:** Why am I even on the show...?! \[laughter\] + +**Mark Bates:** What I was gonna say is I grew up -- my first job I was writing Java, and we didn't have generics. Then I went to Ruby and we didn't have generics. + +**Mat Ryer:** How old are you? No, carry on... \[laughter\] + +**Mark Bates:** I'm old. I'm old. My first job out of college I actually was doing ASP, so I kind of lied; I skipped a whole-- + +**Johnny Boursiquot:** Classic ASP? + +**Mark Bates:** Classic ASP. + +**Johnny Boursiquot:** Wow. You go way back. + +**Mat Ryer:** Bates' ASP was classic. Full of banter in the comments... + +**Mark Bates:** It was. It was as classic ASP as it could get. + +**Mat Ryer:** ...funny variable names... You had a blast. Absolutely classic. + +**Johnny Boursiquot:** Good times. Good times. + +**Mark Bates:** Now I can't even remember what I was talking about. + +**Mat Ryer:** Generics. + +**Mark Bates:** Oh, generics. Thank you. + +**Mat Ryer:** Who's got a meaty unpopular opinion? + +**Johnny Boursiquot:** I do. I'll fire the first salvo. + +**Mat Ryer:** Okay... \[laughs\] + +**Johnny Boursiquot:** That came out of nowhere. You look surprised. + +**Mat Ryer:** Yeah, I was surprised. + +**Johnny Boursiquot:** Yeah, I don't often have unpopular opinions. + +**Mat Ryer:** Yeah, I know. Because you're nice. That's what we were saying earlier, Mark, privately, in the text. + +**Mark Bates:** That's what we were saying earlier, yes. We were saying what a nice guy Johnny was. + +**Johnny Boursiquot:** \[laughs\] Yeah, yeah, yeah, yeah.... + +**Mat Ryer:** Don't you think Mark Bates looks like he's just got back from being shipwrecked? + +**Johnny Boursiquot:** Oh, mate. Fire! + +**Mark Bates:** \[Captain Sparrow voice\] "I've been stuck on that island for so long... Thinking about generics and stewing away..." I'm like \[unintelligible 00:15:06.13\] Somehow I gain weight being stranded on a dessert island. I'm not quite sure how that happens. + +**Mat Ryer:** No, you look great though, mate, really. \[unintealligible 00:15:11.19\] jealous. + +**Mark Bates:** Oh, thank you. That's your unpopular opinion, I assume. + +**Mat Ryer:** \[unintelligible 00:15:14.29\] + +**Mark Bates:** Then I look good. Because there's definitely an unpopular opinion in my house. + +**Mat Ryer:** You definitely would make a good Guess Who character, from that game Guess Who. Like, way more interesting. + +**Mark Bates:** Fair enough. + +**Johnny Boursiquot:** Can I opine now? + +**Mat Ryer:** Yes, Johnny, I'm sorry. + +**Mark Bates:** Yeah, go for it. Please, help us. + +**Johnny Boursiquot:** Okay. \[laughs\] I have to break this up. Alright... Yeah. And for those who don't know, this is actually what happens in real-life too, hanging out with these guys. + +**Mat Ryer:** Apart from I began to piggyback off Bates by now... + +**Mark Bates:** That's true. And there's usually a lot more food involved. + +**Johnny Boursiquot:** And alcohol, yeah. Sadly. Anyways, so my -- + +**Mark Bates:** Still waiting on you, Johnny... + +**Johnny Boursiquot:** I know, I'm trying to get it out and y'all keep making me laugh. \[laughter\] + +**Mark Bates:** \[16:01\] Just keep making those mistakes till you find it, and that's okay. + +**Mat Ryer:** Yeah, absolutely. Well, speaking of mistakes, Teiva's book, "100 Mistakes" is -- + +**Mark Bates:** Are you saying that his book is a mistake?! + +**Johnny Boursiquot:** Goodness! We treat our guests better than that. + +**Mark Bates:** Mat, you're doing a terrible job... I think it's a wonderful book, despite what you say... + +**Johnny Boursiquot:** Teiva, don't believe what he says. + +**Mark Bates:** Yeah. Teiva, man, Mat's just -- he's jealous. + +**Johnny Boursiquot:** Yeah. Great book, great work... Your stuff is not a mistake. + +**Teiva Harsanyi:** That was rude. + +**Mark Bates:** He's jealous. His Blueprints book hasn't sold well in years... \[laughter\] + +**Mat Ryer:** \[unintelligible 00:16:31.09\] It's absolutely full of mistakes. + +**Mark Bates:** Yeah, that's why it hasn't sold well in years, mate. + +**Mat Ryer:** Well, you can find out for yourself by buying my book, and also getting a -- Bates, let's try this. + +**Mark Bates:** You can go to my torrent site and just download his book for free. + +**Mat Ryer:** \[laughs\] + +**Johnny Boursiquot:** That's how we treat friends. We just put their stuff-- + +**Mark Bates:** I'll put the link up on Twitter after this show... \[laughter\] + +**Mat Ryer:** Good. Thank you for putting it there, so nobody will see it. + +**Jerod Santo:** Did you know our most popular episode of 2021 on Go Time was episode 196, "Building actually maintainable software"? Here's some of the funniest moments from that awesome conversation. And if you wanna know what our other top episodes were, we posted our top five of the year on Twitter. Follow along at @GoTimeFM. + +**Kris Brandow:** I think there's this myth that exists in our industry that people wouldn't like working on \[unintelligible 00:17:25.13\] It'd be like this miserable thing, where like "Oh, well that's the team of people that doesn't get to do the fun stuff of building features and building new products and doing all of that." But I think -- and I have some friends that are like this; they're just like, "No, just give them \[unintelligible 00:17:39.17\] Let them go and just clean up some certain parts of codebases..." + +**Sam Boyer:** Scratch that itch... + +**Kris Brandow:** Yeah, like the garbage man. Every city -- like, imagine what our cities would be like if there were no garbage people. Not garbage people, but trash collectors. \[laughter\] That's a better -- I mean, it'd be good to have a city that has... + +**Sam Boyer:** I didn't hear it till you said it... + +**Johnny Boursiquot:** I mean, it would be a wonderful world if we didn't have garbage people... \[laughter\] + +**Kris Brandow:** Imagine a world where we didn't have trash collectors. Our streets would be disgusting, our cities would be awful... But there's no one there that's saying -- and there's some people that are trash collectors, and they love their life. They are so happy with what their job is and how they live their life. + +**Kris Brandow:** What would you add, if you have anything, to make Go more... Not just make Go itself more maintainable, but make the code that we write more maintainable? + +**Johnny Boursiquot:** Generics... Hell no. \[laughter\] + +**Sam Boyer:** No, I want generics, but just for me. Everybody else can screw themselves. \[laughter\] Because then I know what mine do, and I don't have to deal with any of your garbage... And then we're fine. Right? I'm fine... That's what's important here. + +Oh, I would add Rust. That's what I would add... \[laughter\] I would really like to have compile-time -- + +**Johnny Boursiquot:** Unpopular... + +**Sam Boyer:** There you go... I would really like to have compile-time guarantees about shared access to global immutable state. Forget this go test race garbage... Come on. Static or nothing, that's my... \[laughter\] Is this a helpful answer? This isn't a helpful answer. I'm sorry. + +**Kris Brandow:** Well, we avoided dependency mentions. It's okay. + +**Sam Boyer:** We did, we did! \[laughter\] Scooted right by it. Well done. + +**Kris Brandow:** Okay... + +**Johnny Boursiquot:** Nice. + +**Sam Boyer:** I mean, my semver one really - that one gets unpopular when you start suggesting alternatives. It sounds fine when you're just like "The world's terrible! Don't do that thing!" and everyone agrees. Then you try to suggest an alternative and everybody's like "Go f\*\*\* yourself!" \[laughter\] So it depends on how far you're walking out, I guess. \[laugher\] + +**Johnny Boursiquot:** \[19:59\] I think I'm gonna fall over... We haven't had a bleep on this show in a while, so this is gonna be a good one. + +**Sam Boyer:** Sorry, sorry. I should have enquired at the beginning. + +**Kris Brandow:** I mean, Peter got his one swear, so I feel like you also can get a couple swears in. It's okay. + +**Sam Boyer:** Okay, alright. + +**Kris Brandow:** We're a mostly family-friendly show. + +**Sam Boyer:** Gotcha. Noted. I'll remember for next time. + +**Jerod Santo:** Our 200th episode special was one of my favorites of the year. I've never seen Mat Ryer sweat like he did hosting all these Go Time regulars. + +**Mat Ryer:** We've also got Johnny Boursiquot here though... Hello, Johnny Boursiquot! + +**Johnny Boursiquot:** Hello! Yes, mate, I'm here, knocking about... + +**Mat Ryer:** You certainly are. I appreciate the accent, because I can understand you, at last... \[laughter\] + +**Angelica Hill:** Really? It just breaks my heart... + +**Mark Bates:** That was pretty bad, Johnny... + +**Johnny Boursiquot:** I know, I know... + +**Angelica Hill:** I've heard you do better, Johnny... + +**Mat Ryer:** Oh no, don't troll Johnny. He's one of the nicest -- + +**Johnny Boursiquot:** Yeah, I came to have fun. + +**Mat Ryer:** Johnny's the nicest one on here. We can't have a go at him. \[laughter\] And last but not least -- what?! It is least... It's Mark Bates. Hello, Mr. Bates. Welcome back. + +**Mark Bates:** Hello, everyone. + +**Mat Ryer:** Yes... + +**Mark Bates:** I'mma do my \[unintelligible 00:21:11.24\] which I can do really, really well, but every other word is a swear, so I don't think that would be appropriate... + +**Mat Ryer:** Angelica, what would you say is the best HTML tag? So your divs, your spans, your p's... Which one would you go with if you had to pick one, which you definitely do? + +**Angelica Hill:** Probably script, just because when I was first learning software engineering I was obsessed with just shoving JavaScript in... I didn't wanna have separate files; just shove it in. + +**Mat Ryer:** Yeah, that's a good answer, I like that... Although it's wrong. The answer is "a", because without that, it's not-- + +**Angelica Hill:** There's no wrong answers, Mat. + +**Mat Ryer:** There are no wrong answers. \[laughter\] + +**Angelica Hill:** How dare you discriminate against my answer? + +**Mat Ryer:** No, you're absolutely right. + +**Angelica Hill:** You get all whiny about people not answering your questions, and I answer it... + +**Erik St. Martin:** And then it's wrong. \[laughter\] + +**Mat Ryer:** \[unintelligible 00:21:59.13\] + +**Angelica Hill:** You're not really encouraging people, are you? + +**Mark Bates:** I love when people call Mat out. This is great. + +**Mat Ryer:** Yeah. I've got different feelings in my tummy. + +**Mark Bates:** Yay, Angelica! I wanna be on her team. + +**Carlisia Thompson:** Angelica, you're my best friend. + +**Angelica Hill:** Oh my gosh, I love this... \[laughter\] + +**Mat Ryer:** Yeah. If I go red, it's just lighting. + +**Carlisia Thompson:** It's your coat. + +**Johnny Boursiquot:** The jacket. + +**Carlisia Thompson:** The jacket, yes. It's a reflection. + +**Mat Ryer:** Yeah, it's just a reflection of this. It's not anger. + +**Carlisia Thompson:** You came well prepared, Mat. You know your crowd. + +**Mat Ryer:** We're way over time, but wasn't it worth it? Thank you so much for joining us on our special 200th episode. Two hundreDTH... Can anyone say that? + +**Kris Brandow:** Two hundreth? + +**Erik St. Martin:** Two hundreth? + +**Carlisia Thompson:** Yes, we can. + +**Erik St. Martin:** I think you leave out the d. + +**Johnny Boursiquot:** Nobody has a problem with that, mate. + +**Kris Brandow:** It's just you. + +**Mat Ryer:** You're not saying the d. + +**Johnny Boursiquot:** It's just you, Mat... + +**Erik St. Martin:** Two hundredth... + +**Mat Ryer:** Two hundreDTH. You've gotta say that d. + +**Kris Brandow:** Two hundreth. + +**Johnny Boursiquot:** Nope. Still wrong, Mat. + +**Carlisia Thompson:** Even I can. + +**Johnny Boursiquot:** There's no d there, what is wrong with you? + +**Mat Ryer:** Is it not? + +**Natalie Pistunovich:** Two hundreth? + +**Mark Bates:** You could say it's ten score, if you really wanted to... \[laughter\] It's an alternate way of getting you the same way, the same math... + +**Mat Ryer:** I can say that. I wish I'd thought of that. + +**Johnny Boursiquot:** He's been in graduate school for too long... + +**Mark Bates:** I'm trying to bring score back. You know, two score and five... I don't mind admitting that. + +**Erik St. Martin:** Two score in four episodes ago... \[laughter\] + +**Johnny Boursiquot:** Fortnightly... Yeah, you'll listen to this episode in a fortnightly time period... \[laughter\] + +**Mark Bates:** Isn't it a fun language...? + +**Mat Ryer:** Oh, isn't it? Yeah, I've had lots of fun... + +**Mark Bates:** \[unintelligible 00:23:24.27\] fortnight, that's for sure. + +**Jerod Santo:** Last but not least, we have episode 202, where Kris, Natalie and Ian Lopshire discussed maintaining ourselves, but also IKEA and Thanos... Wild. + +**Kris Brandow:** Yeah, that lamp didn't fit with the decor because it's like a bad lamp. It's because you didn't buy a lamp that fit with the rest of the stuff. You've gotta buy things that fit together. + +**Natalie Pistunovich:** Yup. Yup. + +**Kris Brandow:** Man, my statements are not ending in questions... \[laughs\] + +**Natalie Pistunovich:** \[23:59\] But they're so wholesome and so true that you just find yourself nodding, like "Yes, absolutely." + +**Kris Brandow:** Yeah, I feel like this is like "Kris just talks to the world, with Ian and Natalie bringing up really good points." + +**Natalie Pistunovich:** \[laughs\] Well, if anybody is listening to this and is gonna build the IKEA equivalent of software - hi. We're happy to inspire you. + +**Kris Brandow:** Yes. + +**Natalie Pistunovich:** Please make it reverse-compatible. I have a very old kitchen in this apartment, and it doesn't fit the new door, so... Please make it reverse-compatible more. + +**Kris Brandow:** So my first question is "Do either of you find meetings to be productive?" + +**Ian Lopshire:** I would say maybe one in ten. One in ten meetings is productive. + +**Natalie Pistunovich:** \[laughs\] Can you predict, Ian, if a meeting is gonna be productive or not? + +**Natalie Pistunovich:** Oh, definitely. Is there more than three people in it? Then no, it's not productive. \[laughter\] It's a good rule, I think. + +**Natalie Pistunovich:** Yeah. + +**Natalie Pistunovich:** You know, we missed this episode -- I just realized, we missed the Thanos joke. All things perfectly balanced, as they should be. + +**Ian Lopshire:** Delete half the code. + +**Natalie Pistunovich:** Yeah. + +**Kris Brandow:** \[laughs\] That is the solution to most of the problems. Just delete half of it. + +**Natalie Pistunovich:** I think I saw somewhere an open source project that is called Thanos that does exactly that, it randomly deletes 50% of your code. \[laughter\] + +**Kris Brandow:** I mean, you could probably on a lot of codebases delete half of the tests and that would make up for half of the code in it. Or delete the tests and it'd make up for most of the code in the codebase. + +**Ian Lopshire:** I would say most of our codebases are more test code than real code. + +**Kris Brandow:** Yeah... + +**Natalie Pistunovich:** Something that deletes randomly 50% of your backlog. + +**Kris Brandow:** Oh, that sounds like fun. + +**Ian Lopshire:** I like that one. + +**Natalie Pistunovich:** \[laughs\] + +**Ian Lopshire:** Have you guys seen GitHub's thing where it automatically writes code for you? You can stub out a function and -- + +**Natalie Pistunovich:** Codex, yeah. + +**Ian Lopshire:** I bet there could be a Thanos snap version of that, where it looks at everything and then simplifies it down to like half(ish). + +**Natalie Pistunovich:** Interesting. + +**Kris Brandow:** It's just like, "Yeah, we're just gonna get rid of most of this. Goodbye... Wipe it away.. Just delete all of your tests..." It's like, "Yeah... Do you really need them?" + +**Natalie Pistunovich:** Or random documentation lines. + +**Kris Brandow:** Or better yet, a tool that just goes around and randomly changes small things within your codebase, but keeps them compilable... Just like, really big annoyances... + +**Ian Lopshire:** That's mean. + +**Natalie Pistunovich:** The real fuzzing... + +**Kris Brandow:** It reminds me - I saw this TikTok once, and it was like a guy, and he was just like, "If you really wanna get back at someone, do these things", and one of them was "Go to some you don't like's house, as to like a party, and take all of their remotes except one", because then they'll be like "Where are all the remotes? I have this one, so clearly no one took them..." Just small things that inconvenience people's lives, \[unintelligible 00:26:43.15\] + +**Ian Lopshire:** Put dead batteries in all the remotes. + +**Kris Brandow:** Or that. It's like, "Why are none of my remotes working?" Just like little inconveniences in people's lives... Don't do that. That's a way to unbalance other people's lives, and this episode was about balancing people's lives. + +**Natalie Pistunovich:** Exactly. Exactly. + +**Kris Brandow:** Although I guess you could balance your life by unbalancing someone else's life... No, that's not good. Don't do that. Don't do that. + +**Ian Lopshire:** \[laughs\] + +**Natalie Pistunovich:** Balance it to the positive, yes. + +**Kris Brandow:** Positive balance. But if you want to write a bot that people can write on their own codebases, that puts small annoyances into their code... + +**Ian Lopshire:** That's just like community contribution, right? Does that count for Hacktoberfest? + +**Kris Brandow:** Yeah... And that could be a tool, like a pull request... Like, make sure you're reviewing your codebase; there's like slight differences... Like, actual bit rot, right? That could be a good name for it, BitRot. Did you run the BitRot bot? + +**Natalie Pistunovich:** \[laughs\] + +**Kris Brandow:** Man, this episode is wild... diff --git a/The myth of incremental progress_transcript.txt b/The myth of incremental progress_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..62f19142051f1b4bfa5bb5750ba92bc31af24ba8 --- /dev/null +++ b/The myth of incremental progress_transcript.txt @@ -0,0 +1,559 @@ +**Kris Brandow:** Welcome, welcome, welcome, gophers! We are here for another episode of Go Time, and this week we're gonna be talking about another interesting little topic in the continuance of our maintenance series... For those of you out there that are counting, this is our eighth episode in the maintenance series. + +The genesis for this episode was a little chat we had in the Gophers Slack, in the Go Time channel. Shout-out to the Go Time channel; if you're not in there, you should definitely join... But it came -- it started with one of our listeners who pointed out that in a lot of our recent episodes we've been talking about how maybe we shouldn't have any structure for things, or maybe we should have all these nice brownfield projects... And we kind of noted that, and our wonderful producer, Jerod, came around with a beautiful analogy of experienced painters liking a blank canvas, and some other folks just liking paint by numbers... And of course, my brain, as a writer, just went immediately to "We're gonna make that into an episode." + +So that is the very episode that you're listening to right now. We've titled it The Myth of Incremental Progress, and I think it's going to be a lot of fun. + +Joining me today I have my usual co-host, Johnny Boursiquot. How are you today, Johnny? + +**Johnny Boursiquot:** Doing alright. + +**Kris Brandow:** And we also have our wonderful producer, Jerod. How are you today, Jerod? + +**Jerod Santo:** Happy to be here. + +**Kris Brandow:** Awesome. And we have someone who you've all heard multiple times on the podcast, but instead of joining us as a guest, this time he is now a panelist, so congratulations on joining us in the co-hosting ranks, Ian... How are you doing today? + +**Ian Lopshire:** I'm doing great! + +**Jerod Santo:** \[04:00\] Yay! Welcome, Ian! Congratulations! + +**Ian Lopshire:** Thanks! + +**Jerod Santo:** \[laughs\] It's kind of self-aggrandizing, isn't it? ...congratulating you for being a panelist on our show...? \[laughter\] + +**Johnny Boursiquot:** "You won a prize." + +**Kris Brandow:** "Congratulations for joining my podcast!" \[laughter\] + +**Jerod Santo:** Yeah. \[laughs\] + +**Kris Brandow:** Alright. So we don't need any introductions, since we're all hosts here... So given Jerod's comment, I am gonna make the analogy - or I guess the translation? I don't know what I'm about to call this, but... That you know, this idea of painting by numbers is really nice. You can paint some really good stuff by numbers, when you go out there into the world and get those nice coloring books... You're like "Okay, I'm making a nice picture." But I'm gonna make the conjecture that you can't paint by numbers your way to being one of those blank canvas artists. These are two separate and distinct skillsets, and one does not lead into the other. And the reason I say that is because I think a lot of the time when we think about these things, we think "Oh, well, I can start by painting by numbers, and I'll eventually wind up being this grand artist, or something like that." So I'd like to hear from my panel, do you agree? Do you disagree ? I'm sure, Johnny, you have tons of nuance, and we're gonna go meta straight away... So yeah, who wants to jump in first? + +**Jerod Santo:** I think it would be useful to attach the analogue to the subject at hand. So if we're talking blank canvas, this was mostly in reference to instructions on what tools I should be using, what practices I should follow, the structure, the architecture of my software project. And I was saying that experienced developers like a blank canvas. They like to start from scratch and draw their picture. + +Paint by numbers - the question is what does that draw across to for you guys? I tend to think of framework, preexisting structure, a scaffold of some kind... Obviously, it doesn't apply one to one, because no one's gonna give you a number and say "Put the code right here, and put the green code right there, and the blue code right there." But is that what you all are thinking when you think of paint by numbers? Because we have to at least agree on what the metaphor applies to, I think, before we can debate its merits. + +**Kris Brandow:** Oh man, Jerod, you're already just ruining the whole fun of this. + +**Jerod Santo:** I'm going meta. \[laughter\] That's what you guys do, isn't it? + +**Johnny Boursiquot:** You fit right in. + +**Kris Brandow:** We don't get to defining what we mean by these things until halfway through the episode. You're just all over the place. But no, I think from my perspective you are right there; I think it is these frameworks oftentimes. Or you know, in the context of the way you write your code, it's all of those projects that get you started and then it's like "Here's your predefined package layout of where all of your packages go, and where you put your readme, and these other document files", and all of that. I think that is a bit -- that's when I think of when I think of paint by numbers for code. But Ian, Johnny... What are your thoughts? + +**Johnny Boursiquot:** I think it would be unfair to say that folks that use frameworks are deliberately choosing sort of a paint by numbers approach... Because frameworks, scaffolding and all these things - these are productivity tools, not really "Strictly beginner, I'm learning how to do things, I need handholding and help getting things done", which is kind of what (I assume) paint by numbers kind of leans towards. You're not an experienced painter yet, you need some structure and some help and some guidance, and you need a starting point to hone your craft until you don't need that anymore, until you can stare at a blank canvas and start creating. + +Frameworks I don't think really fall in that category. They are strictly productivity, I think. Now, can they help somebody who is starting out? Absolutely. But they help you as a beginner to be productive within a journey of building software with other experienced people who are also using this framework for productivity. So I think we should be very careful not to downplay the role of frameworks. They're not really the same thing. + +**Kris Brandow:** \[08:03\] I do feel like there's a slight shift in the analogy there though... Because I think when we're talking about things like paint by numbers, or talking less about the thing that the software actually winds up doing, and the structure that we're giving to the software to start with. So when I say frameworks are paint by numbers, I'm specifically saying for the structuring of your project; not necessarily what it does, or the business thing that it solves, but specifically, you are getting handed the places where you put all of your code. You still have to figure out what code you put there. It's kind of like if you do paint by numbers, but you just pick your own colors, because you understand color theory, so it looks different than other people's paint by numbers. But the structure of what you wind up with is going to be that structure there. + +**Ian Lopshire:** When I read the paint by numbers analogy, I was more thinking like -- you know, in my first years as a software engineer my manager would give me five tasks, and each of those tasks he's already broken down where this change needs to happen, how you should do it, go do it. And the blank canvas for me was saying "Hey, this is a feature that needs added, or this is a thing that needs done. Go figure that all out yourself." + +**Jerod Santo:** I think it's a fascinating way of looking at it, and I see why you did that. The original GoTimeFM chat was around - we had tweeted out a clean architecture template for Go services, which some of the gophers in the chat thought has too much ceremony; it was Bob Martin's Clean Code book, and some of the design patterns laid out therein were implemented directly into this template for you to start from. So it was very much about maybe not -- and to Johnny's point, I agree with you on frameworks; I think maybe we set them aside in this conversation... Because developers build frameworks for themselves to use; so it's not like just for beginners. If you build your own stuff for long enough, you will develop your own framework or tooling. And it is about productivity. + +Now, public consumption frameworks that a new person can come to and become productive faster is one aspect of a framework. But you can use frameworks for many productivity things. I think maybe more design patterns, architecture, structure is what Kris was referring to. Is that fair, Kris? More so than laid out tasks and more so than frameworks. + +**Kris Brandow:** Yeah, I think so. I think I also have another analogy that is similar to this one, of music, and how there are musicians that can either play by ear, or read sheet music, but can't necessarily compose their own songs. Music composition and music playing are two very different skillsets. But I think you'd also think that hey, if you play music for a long time, you might wind up being someone that can compose music, when that is not the same. And maybe that's a little bit of a better analogy than the paint by numbers versus blank canvas, because I think we can all recognize that musicians are very talented people, that even if they are playing songs composed by other people, that doesn't make them any less of a skillful human. + +**Jerod Santo:** Right. + +**Kris Brandow:** But there's a distinction there, and I think that's what I wanted this episode to be about, is like when you have those types of close distinctions, how do we kind of help ourselves not fall into the trap of thinking that we'll wind up becoming a great composer one day by just practicing a lot of music. + +**Johnny Boursiquot:** You become a great composer by composing. You can become a great musician by simply playing music. Maybe yesterday or the day before I saw some tweet, maybe from Mat Ryer or something, maybe he was retweeting somebody else - of a kid, five years old, who's ripping apart a sonata from Bach, or something. The kid's hands are moving fluidly through -- I mean, I'm thinking "Man, this is a prodigy. This kid's genius." I don't know any other five-year-old who can do that. + +And then you obviously have people -- this being on the internet, it's like "Ah, that's nothing. If he can't write music, then he's not a prodigy." I'm thinking "Okay, I can see what you're saying, that in the pantheon of music maybe this particular piece that he happens to be playing, albeit eloquently, could be considered sort of a beginner sonata, versus something that is more complicated in structure, and tempo etc." So while I understood the nuance that people who are critiquing this five-year-old kid, and their music skills and talents - I get the nuance. But to me, at the end of the day I'm thinking "Wow, this kid is pretty good." + +\[12:36\] Now, I wasn't slicing it and saying "Oh, he's good at playing music, but he's bad at composing music." I was saying "Okay, here's somebody performing the art of playing music, and they're pretty good. + +So if I bring this back to software, and I'm like, okay, can I identify somebody who's pretty good at coding, at flying through the keyboard, knowing everything that needs to go in, knows the syntax of the language, knows exactly how to implement an algorithm and do all these things - I can now say "Well, they might be pretty good at those things, but they don't know how to engineer software", or larger-scale software, or distributed systems software. They're good at coding, they're very efficient, they're a great coder, but maybe they're not a programmer, or an engineer... Because these things - you don't just master and memorize syntax and start ripping across the keyboard and doing things. That's a skill, but that's not the complete set of skills you need to build and engineer software that has value to somebody who's willing to pay for it. + +So that's how I see these things. I don't wanna take away from the act of knowing how to code, but at the same time I understand that knowing how to code alone is not sufficient for engineering software. Hopefully, that makes sense. + +**Kris Brandow:** I feel like the YouTube video I sent to you has resonated with you then... + +**Johnny Boursiquot:** Yeah, I think so. Definitely. I could understand it. Basically, we tend to throw around the "What's the difference between a coder and an engineer?" or whatever it is... And this person who actually - I'm ashamed to say, I didn't even know his name, Leslie Lamport, and so much of what he's done impacts my actual day-to-day, like how I make a living, and I didn't even know this person... I feel kind of ashamed of myself; I'm like, "Okay, I have to make the point to go research this person." And he calls out -- and I sat down to, I was just basically, he was effectively having fun, he didn't start out with paint by numbers. He wasn't starting out with somebody else's thing; he sat down, somebody gave him space and pay, he sat down and -- he came from a different background as a mathematician. He came into this field and basically applied what he knew and had mastered in a different field into ours, in the field of computing, and created all these wonderful things that we now take for granted. So that's blank canvas. To me, that is a blank canvas. That is a master of something that comes in and says "Well, how can I apply what I know, what I've been doing for decades? How can I innovate in this space?" + +So I think those who start with a blank canvas are seeking to innovate. Those who start with, for a lack of a better term, the paint by numbers, are seeking to execute. I think there's a difference there. + +**Jerod Santo:** The point that I was trying to make and I'll try to make here again is that while I agree with that, that man is a master who changed the way that software is written. He's like a Picasso. And lots of us are never gonna be Picasso. I'm not saying you all can't be, and I can't be; I'm not trying to limit anybody. And I don't necessarily wanna lift him up to a level that's higher than it deserves. I only know him from the seven-minute video as well. Much to my shame, I didn't know his name prior, so thanks, Kris. We'll link that up in the show notes. I've put it in the chat for those who wanna watch that great little video featuring this man's work, and some of his thoughts. + +\[16:04\] Some of us wanna paint because we love painting. And we're never gonna be Picasso. And we maybe wanna be a working painter, maybe we wanna paint portraits for people who paid us. Isn't even George W. Bush paints portraits now in his retirement? He's average at it, but people still give him respect about it, because he's an ex-president... We're not all gonna be that. I bet he starts with blank canvases now. He probably didn't start there. And so I guess I was trying to say lots of us just wanna be like working software developers who put in an honest day's work, do our best, write some good software, help people out, but aren't necessarily gonna make it to the level of a person who reinvents stuff. So how do we help those people as well, I guess? + +A lot of the conversations on Go Time - which I very much appreciate, but I've also been in the business for a very long time - are very deep, and nuanced, and from an expert perspective, and... Hey, y'all are experts, so of course we want an expert perspective. But sometimes the question is like - well, Ian, when you start a brand new project now, or given that feature today, to build, and you've got your blank canvas, what do you do? Where do you start? You don't have to tell me where I need to start and paint by numbers me; but if I can listen to the way that you go through that process, and get to your end goal, your painted picture, maybe I can learn from that. And so I'm actually setting that up, but I also would love an answer... Like, let's say your boss gives you a feature to build, Ian. You've never built this before, maybe you have... What do you normally do? Where do you get going when you're just staring at a blank screen? + +**Ian Lopshire:** Yeah, it's a great question. I don't know. I don't know the answer to that. \[laughter\] + +**Jerod Santo:** Okay, fair enough. + +**Kris Brandow:** At the beginning, of course... + +**Ian Lopshire:** Yeah, exactly. I mean, I start in one big file and I just go, you know? + +**Jerod Santo:** You just start writing code? I mean, then you're coding. Surely there's steps before that. Okay, let's pass it to Kris. Kris, you write software for a living, you're a blank canvas kind of guy... Where do you start? What do you do? + +**Kris Brandow:** I think because I am who I am, I usually start with a combination of writing some sort of document to make sure I kind of understand what the problem is. Actually, I guess in the beginning-beginning, when someone brings something for me to build, I ask a lot of questions, and make sure that it's like "Okay, well, are you actually asking me for the thing you want, or are you asking me to build the thing you think you need?" That's usually where I start whenever it's a kind of -- actually, it doesn't matter. It doesn't matter if it's a product person or if it's something I'm building for myself. I say "Is the premise that I have assumed here, or someone else has assumed, the correct one?" And once I kind of figure that out, I'll start with writing up a design doc, or a scope doc, and being like "Okay, this is what we're going to build." And then I usually jump into some prototyping, and that's where I start actually throwing some code into a codebase, and then the project usually gets legs from there, and we kind of -- you know, if it's something for work, we develop out that design doc more, we start having discussions, then we actually make the for-real codebase, which might unfortunately be taking that prototype code I built, or it might be starting from fresh, which is definitely a preferable thing, because prototype code is never really written as well as we'd like it to be. + +That's usually the on-ramp that I take... But that's really a lot because I'm a writer, so my comfort zone is an empty document in Google Docs, or an empty Vim window where I can just start typing out, "Okay, what is it that I'm trying to do?" + +**Jerod Santo:** Blank canvas. + +**Kris Brandow:** Yeah. + +**Jerod Santo:** Johnny, does that resonate with you? Is that similar to your process? + +**Johnny Boursiquot:** Yeah, I'd say so. Obviously, there's gonna be basically what am I being asked to do - is it to add a feature in existing software, which hopefully that's an easier task, or it's supposed to be an easier task than basically building something brand new... \[laughter\] + +**Jerod Santo:** It depends on how legacy code that sucker is. + +**Johnny Boursiquot:** \[19:49\] Depending on how old that thing is, right? And whereby Kris jumps into a Google Doc or a Vim window and whatnot and starts writing things out, I tend to be more of the -- I'm a visual thinker, so I tend to fire up one model or build us a chart or something, and I start mapping out flows; what data is coming from where, where does it need to go, what are the boundaries? Basically, I think in systems, so I'm trying to identify what systems am I dealing with here - internal, external, what are the constraints, how much data needs to flow through this pipe, and things like that. So I sort of picture a system in my mind and try to represent that on some sort of canvas, in this case a digital one. + +And then from there, I start to write about the different pieces of that system. Then I take Kris' approach and say "Well, what is this thing? If I were to get a readme for a codebase for this particular repository that represents this component in my system, what should that readme tell me?" So I do readme-driven development for that reason... Because I wanna be able to say "Okay, well if I hand this to an engineer on my team, are they gonna know what to do with it? Are they gonna know where this fits into the bigger picture?" + +So I kind of approach it that way... Basically, I guess you call that top-down, whereby I try to get a "What's the 30,000-foot view of this thing?" and then I sort of slowly descend into the nitty-gritty. + +**Jerod Santo:** I'm similar. I'm an outliner though, so I work in outlines; I think in outlines - you know, top-level, and I star to drill down and start to feather it out as it makes sense... What depends on what, what do I know versus what do I not know, what's easy, what's hard? And then substeps. Trying to take that blank canvas, which to me still after almost 20 years in the industry can be intimidating, and give myself a task list, like Ian received from his boss. Like "Can I produce that, and come up with --" And this is kind of the process of user stories, or whatever, which I've never been good at writing those. As an admin, I want to finish this feature, so I can go home. \[laughter\] + +**Break:** \[21:51\] + +**Jerod Santo:** One thing Dillon Bork said in the chat which I would love to pull into this, which I think is a great point - he says the danger of paint by number in his opinion is that it can lead to a cargo-culting of the various patterns and idioms without any insight or understanding about why the things are done that way... Which, by the way, I think is a great aspect of having conversations like these, is because we hear about some of the why's and the "it depends", and not so much of the blog posts where it says "Here's how you ought to do it because it worked for me." It's harder to cargo-cult a podcast, I think, than it is a blog post or a book. + +\[24:12\] That being said, sometimes you just have to follow the other person's path until you realize when it doesn't actually work for you. I'm totally fine with cargo-culting some sort of rule... I was gonna say the Law of Demeter, but that one's too hard to explain. What's a very simple -- DRY, right? Everyone can remember that one. We all get it wrong, but we can all remember the acronym... And all of us here, I think, would all talk about how DRY is not the best principle in many cases. I think I've heard you guys talk about that. But we didn't realize that until we had tried to DRY the crap out of everything for a while. And then it came back to bite us. + +So I think it's okay to go through that process of, like I was saying in the post, go ahead and paint by numbers for a little while. Don't live there, don't stay there, but until you can start to realize actually blue doesn't look great on the number four. A little bit lighter blue would look nicer there. You start to develop your taste, your experience, your own history of that working well in this context, not in that context.... Then you don't need it quite as much. But I think we do need to start somewhere, and I think that a lot of these idioms, best practices, rules, clean architecture, whatever that means - those are decent starting places. + +**Kris Brandow:** Yeah. And there's something you've said a while back, when we were talking about Leslie Lamport and his path of like "Okay, I'm a mathematician that stumbled into computer science, and then I was studying physics, and then I realized that this distributed database someone's trying to build is actually violating the laws of physics, so we can't possibly build it..." It's interesting that you've kind of said "Okay, we all can't be that", because I think in some ways the way that we've set up our industry is that we idolize people that can do those sorts of things, and we assume two things; and I've seen this over my career, it's happened to me personally, and I've seen it happen to other people, where this assumption that if you're new, you can't possibly be someone like that. Like, having that type of knowledge only comes from years and years and years of practice. And the other side of it is that if you want to stay in this industry, you should want to wind up to be someone like that. + +I think for me personally - it's kind of funny that you say like "Oh, we're not really people like Leslie Lamport." I think my own path through computer science is very similar to the one that Leslie Lamport has taken, and the way that he approached everything. I am - as we know from this podcast - a very heavy analogy person... + +**Jerod Santo:** You're a writer. + +**Kris Brandow:** I'm at my core a writer that happens to write programming languages... And I think that is a very distinct path for me, and that has resulted in me, in my career - like, I was never a junior engineer. I've never been like an SE1, or an SE2. I've never had that experience. I don't quite understand what that experience is like. But I also know that I wouldn't want to be an SE1, or what we usually conceive as an SE1 or an SE2, someone that gets handed tasks and just executes them, and that on-ramp to learning. For me, I always wanna be at that stack+ level that higher level of "Okay, give me the tough problem to solve. I'll sit down and I'll figure it out." + +I think we've mentioned this in an episode a while back, but I think that's actually an important differentiation and distinction that we need to make as an industry, that goes along with this kind of thought process of incremental progress of - there are some people who are those Leslie Lamports, who are the people that can design really great, big systems, or find the next new thing. And there's not a ton of them, but they're not going to take that climbing up the ladder approach to things, so we need a way for them to get onboard. But we also need to recognize that there are plenty of people that are perfectly happy just being handed user stories all day and executing on that. + +**Jerod Santo:** Sure. + +**Kris Brandow:** And there's absolutely nothing wrong with that. There's nothing wrong with being the person that paints in all of the colors to make the beautiful portrait at the end of the day. Because starting with a blank canvas is very intense, and it takes a lot of energy to sit down and figure all the things out... And quite frankly, some people don't need a blank canvas. + +\[28:19\] Your marketing company needs a website; we don't need to spin up a whole new custom Go backend... Just go use a framework. Just go use WordPress, or Drupal, or Wix.com, or something. You don't need to do all of the intense effort. But I think we do, as an industry, have to start realizing that these are not things that stack on top of each other. These are things that are next to each other. + +**Johnny Boursiquot:** But what about resume-driven development? I need to build a brand new framework for this website. \[laughter\] + +**Jerod Santo:** That's true. + +**Kris Brandow:** Kubernetes-driven development. \[laughter\] + +**Jerod Santo:** Yeah, and then we find out in the big tech co's there's now promotion-driven development, which drives all sorts of new projects, because you can't get promoted working on the current project, even though the value is there... And these are incentive structures that people are realizing are misaligned with effectiveness etc. and hard problems to solve internally. I don't envy anybody trying to set those structures up inside of a large company... But yeah, a lot of it is -- not a lot of it; some of it can be exactly what you said, Johnny... It's like, "Yeah, but how do I get that raise? How do I get that new job? How do I get that promotion?" + +Some people - actually, that's the advice they're looking for. It's like, "Hey, I wanna be a software engineer. Y'all have been doing it for all these years... I'm trying to break in. What do I do? Where do I start?" And again, unfortunately, the answer to that still can be "It depends" or "Just pick React." I'm just kidding... \[laughter\] + +**Kris Brandow:** Oh, Lord, Jerod... + +**Jerod Santo:** Sorry, I am from JS Party, so I had to get at something... + +**Kris Brandow:** You've been spending a little too much time on that, on that topic there + +**Jerod Santo:** I think Ian disagrees with something Kris said, because he had a funky look on his face, and then Johnny talked. So Ian, do you wanna say something? + +**Ian Lopshire:** Yeah. Let's go back to that idea that these don't stack, that they're beside each other. I kind of disagree with that on a lot of levels... But let's just take Go, the language, for example. That didn't come out of academics; that came out of a bunch of people's years of experience, knowing how a language should work to be productive, right? I mean, obviously, they're all brilliant people, but that didn't come out of these mathematicians, or that sort of thing. + +I think they can be beside each other, but it definitely does stack. Gaining context and experience does allow you to contribute more later. + +**Kris Brandow:** Yes, but I would say that looking along the dimension of time is likely going to walk you into problem areas... Because it's like, the reason the founders of Go were so great at creating this language that we have is more because of the experiences they have, not so much the length of time of those experiences. So I do agree with you that time is a factor here, but I think it's a factor we need to discount heavily, because I think -- if we were to start from a blank slate, if we start from the beginning, I think we can weigh time pretty well, and it'd be okay. But I think because we have this assumption now that time = experience, we actually have to kind of subtract, or kind of push that down on our levels of understanding... Because I've met so many more people that it's like "Well, you have ten years of the first year of experience or you have twenty years of the first two years of experience." + +So I think we overweight time too much of the time, and that has resulted in us as an industry just saying "Okay, just grind for a few more years." And then you see things on job posts that are like "Okay, we need 70 years of experience in this, or 10 years of experience in that." And I think that's exactly the thing I was talking about, where it's like "You can learn what it takes someone ten years to learn, in one year, if you have one of these brains that just loves learning, or loves exploring." So it has a lot to do with who you are as a human, more than how long that you've been doing something. We don't have really good ways of differentiating that right now, but... + +**Ian Lopshire:** I'll agree that time doesn't mean much there, but by experience I mean like quantity of... I don't even know how to say it. + +**Jerod Santo:** Some other measure that's not time... \[laughter\] + +**Ian Lopshire:** \[32:14\] No, it's not time. It's like -- + +**Jerod Santo:** What he means is like 10,000 hours... \[laughter\] + +**Ian Lopshire:** No... By experience I mean like a base of knowledge. Like, yeah, if you do the same thing every day, you're not growing and learning. But if you're gaining new experiences, that is experience. It's new experience. It's not length of time doing a certain thing. + +**Kris Brandow:** Right. + +**Jerod Santo:** So on our job postings we have to say "Seven years of experiences..." \[laughter\] I'm not disagreeing with you, I'm just saying it's really hard to quantify this thing if it's not time-based. + +**Ian Lopshire:** Yeah, time is like a good analog, but it's not the thing. + +**Jerod Santo:** It's a proxy. + +**Ian Lopshire:** Yeah, exactly. + +**Kris Brandow:** Yeah, I think that's also what I was trying to get at there... Maybe we just shouldn't use time, because it's too easy for us to gamify it. It's too easy for us to sit there and be like... + +**Jerod Santo:** Yeah, you just wait. + +**Kris Brandow:** "Oh, you've been around for ten years, so obviously, you can design systems now." It's like, please, don't... + +**Johnny Boursiquot:** I mean, isn't that a fair assumption to make though? We don't have better heuristics for that, so... There's an assumption that you must make as an employer, or a hiring manager, managing director, whatever, that given enough time, somebody has seen enough kinds of problems that they are experienced enough to help us not repeat certain mistakes. That's a fair assumption to make. I know it happens - people will have seven years of experience and it's just the first year repeated seven times; they don't grow and learn and mature... But I'd argue that those are far and few between. At least in my career, which I've been grateful enough to be doing this for close to 25 years now... I've come across maybe one or two people during that time that basically had a lot of years of experience in the industry, but really that didn't show in the kinds of things they were proposing, the way they communicated... And \[unintelligible 00:33:56.17\] language barrier, and all these things. More like communicating a design, and things like that. + +So it's possible for years to have gone by, for you to not grow as an engineer. And mind you, they were great coders. If I give them a specific tasks, with specifics on "This is what I want you to do. Do this, do this here, go find out what you need to find out there, what you need to fix is over here..." Very specific things - they will get it done quickly and efficiently. But if I said "Okay, design me a web scraper that scrapes this website", they'll be at it for like 3-4 months, until I come back and say "So where are we?" + +So it's one of those things where if you've been experienced enough -- let's use that example... Somebody comes to me and says "Hey, design a web scraper", and they give me the parameters and they say "Hey, this is what it needs to do etc." The first thing I do, I'm like, okay, I go to Google and say-- + +**Jerod Santo:** "npm install a web scraper." \[laughter\] + +**Johnny Boursiquot:** Right? Right?! I mean, why am I gonna reinvent that wheel? I've got better things to do with my time. I need you to be experienced enough to recognize that. And go get a damn package off the internet that does this. I don't need you, as a junior developer, reinventing that wheel. That's not where your time is the most valuable to me. + +So that kind of judgment I think is what comes with time. You've seen a few patterns, you've been coding for a while, so you know how to deal with syntax, and you've been around other people who know what they're doing, so you pick up some of the tricks of the trade, that kind of thing... And then you allow time to allow these things to just gestate and become part of your skillset as an engineer. So I think that's the element, time... I don't think we can discount time here at all. + +And again, I'm biased because I've been doing this for a while, and I can say that "Okay, yeah, time has added so much to my skills", and whatever it is... But I think in my experience you need the time to grow. + +**Kris Brandow:** I wonder if this has to do with where in the industry each of us is positioned... It's interesting to me that you, Johnny and Jerod, had never heard of Leslie Lamport before; I think part of it is before I've spent so much time in distributed systems that it's just like "Oh, yeah, Lamport clocks. Leslie Lamport. Of course." + +\[36:10\] But I think distributed systems specifically is one of those areas where it really isn't the amount of time that matters nearly as much as like the knowledge that you have obtained. And you can obtain that knowledge very quickly, but it takes a -- people's brains are different, so your ability to understand things like the distributed system concept... There is no now, and you can't use time. That's not something that you -- I mean, maybe. Maybe I'm wrong here. Maybe you can sit there and work through these problems long enough and you'll eventually figure out and wrap your brain around the concept that time is not at all what you've thought it is your entire life... But I think that's more of like a way that we have to teach people things, and the way they have to kind of like be curious and be okay enough with being frustrated at not understanding something to kind of dig into... And I think that's a better marker of how experienced someone is and the length of time they've been doing something. Curiosity, at least for this section of things... + +I think part of the problem here ultimately is that -- it's like we're trying to storypoint right now. We're trying to reduce a huge space of things down into a single entity... And it's like, how you gain experience as a frontend engineer is vastly different than how you gain it as a backend engineer, and it's different from a distributed systems engineer, or from an embedded engineer... There's all these different types of engineering, and I think that saying length of time is like an easy thing, because that's something you can trace across all of those different types of engineering... But I think it means a different thing in each space. So I don't think they're as comparable. I have no idea if any of that made sense... Sorry if it didn't. \[laughter\] + +**Jerod Santo:** It seems like what you're arguing for mostly is for diversity of experience... Because the unique perspective is what sometimes - and in the case of Leslie Lamport - actually leads to the innovation. Looking at it from a perspective that nobody else looked at it from, because of his background. Or because of your background as a writer, you look at software differently than I do, for example. So that can be very profitable for all of us, and beneficial, is having diverse perspectives, chipping away at these various problems, which I think we're all for. + +**Kris Brandow:** And this is why you're the producer. You've put that perfectly, you know? \[laughter\] + +**Jerod Santo:** Well, I think when you take the diverse perspectives and then you also give them experience - like, let's all get the experience - I think they all get better. I think he would confess that he's probably better at everything he did when he was in the business for 25 years than when he was in it for five years. That being said, a lot of newbies invent brand new, radical things that nobody else sees, because they don't have the experience to say "Well, that won't work." They just try it. And the rest of us are like "Well, that's never gonna work", because we've been in the business 25 years. + +So a lot of new things, new techniques, new frameworks, thinking about it differently, come out of people who are 21, 22, or three years into the business. Maybe you're not actually young, but young to the industry. + +**Ian Lopshire:** So just kind of riffing on something you said... I think the one piece that you're discounting about time is you might be able to gain all this knowledge very quickly, but what time does provide is opportunity to have decisions that you thought were good, turn out to be bad... And without that time aspect, you don't gain any of that experience. You need the retrospect, you need to look back and evaluate if this was a good or bad decision. + +**Johnny Boursiquot:** I wish I could find some of the people that have written software I've inherited, and be like "Look. Look at what you've done. Look at what you did." \[laughter\] + +**Jerod Santo:** "Look how bad this idea was." That would be a cool reality show... It's like, Johnny just chases down people that used to work on software and shows them the ramifications of their decision-making. That would be a good show. "Look at what you did...!" \[laughs\] + +**Kris Brandow:** \[40:02\] The thing I just thought of when you were saying that, Ian, is like - time is a useful way of comparing yourself, and on your own trajectory. So it's like, you at one year of experience versus you at ten years of experience, or ten years of doing this, hopefully you have acquired a lot more and you've moved a lot more. + +I think the thing I'm pushing against is that that's not comparable between people. I think that's the issue at hand - ten years of my experience is very different from ten years of Johnny's experience, or Jerod, or your experience. That's different, and that has a factor of what is the industry like over the course of ten years... Sometimes we have a lot of change in a small period of time, sometimes there's not much change in a period of time... But yeah, I think Jerod, you put it precisely there, with just like the diversity of experiences that you wind up having. + +I think part of the thing that I disliked about this kind of incremental thinking approach to our industry and the way that we kind of ramp up engineers from SE1 to SE2, to senior, to staff etc. is that a lot of the times when you're in those lower roles, you're kind of indoctrinated into the way of doing things that everybody else is doing, because it's assumed that the people that have more experience know how to get things done, and they've been doing this longer, and they've been successful, so they wanna give you all the tools and they wanna shortcut you from having to do the work that they had to do. But it's like, the work that they had to do is what made them successful. So you can't get around that. + +And I've seen at a number of companies I've been at, this attempted to happen. We've even had someone -- a senior in a company I was talking to, and my philosophy is just hire good people, hire people that are curious and wanna learn, and you'll do well. And he was like shocked at this. "We can't do that. My philosophy on this is I hire you, and then it's like a cook in the kitchen. I teach you how to chop carrots, I teach you how to chop onions, I teach you how to chop this, and then you show me that you can do those things..." And the thing I very much disliked about this is that kind of "I will show you how to do this thing, and I will show you how to do things as I know how to do them, and that's the way that we learn and grow." But that really pulls away a lot of the diversity that we'll wind up having, because once again, we're saying "Well, I'm the one with the experience and knowledge, so I'm gonna give that to you. And you're gonna have it, and you're not gonna have to make the same mistakes that I made." + +**Jerod Santo:** I understand where you're coming from. I don't know if in practice it works out that way all that much, because you yourself buck against that. Maybe you go through it for a little while and you're like "Well, this is BS. I'm not gonna do it that way. This way is better." And I think that people that have those different experiences will start with their boss or their mentors set of rules, and "Here's how you're gonna do it, here's how you chop celery", and then eventually their experience and their knowledge, they're just like "Yeah, I actually learned this other way. I'm gonna start doing it that way." And maybe in certain circumstances you get in trouble for that, or whatever. + +But that brings me to Sandi Metz's rules. Are you guys familiar with these? Sandi Metz is a great OOP teacher/programmer, very experienced. She teaches people how to do object-oriented programming... And she has rules that she just hands out to new developers. And the rules are "Classes can be no longer than 100 lines of code." She has a set of rules, and she's like "No, you're gonna follow these rules. Methods can be no longer than 5 lines of code." + +Most of her work I think is in Java and Ruby, most of her experience and the most people she's teaching. "Pass no more than 4 parameters into a method; hash options are parameters." This to me is paint by numbers, to a certain degree... And I would never follow any of these. But... Okay, a couple of them. 4 parameters is -- I mean, these are good rules, but I probably wouldn't follow them. But if I was just getting started and needed to ramp up, or if I had somebody starting fresh and I was like "How do I do this?", I think starting with a strict set of rules is actually a pretty good place. And then you say "Follow these--" This is Sandi Metz, this is me. She says "You just follow these until you know better. And once you can explain why you're breaking the rule, then go ahead and break the rule. But until then, just follow the rule and your code will be better." That's a pretty useful teaching technique that a lot of people have appreciated. + +**Johnny Boursiquot:** \[44:15\] And you need time for that understanding. + +**Jerod Santo:** Exactly. + +**Kris Brandow:** I feel like the last part of what you said there is the crucial part. Like "Use these until you can basically articulate why they're wrong." And I think there is some form of time component to that, but I think some people will be able to do that much quicker than other people. + +**Jerod Santo:** Yeah. + +**Kris Brandow:** This is just me begin very, like, "I'm glad I never started off as a junior engineer, and I don't want people that are like me to have to slog through that", when they're like "No, but I wanna go design things", and people are like "But you don't know enough. You haven't been around long enough." It's like, "But I'll be around longer." + +**Jerod Santo:** Well, when I was green behind the ears - or wet behind the ears; I don't know how these sayings go. When I was green, I wanted rules. I wanted structure, because I didn't know how to do it otherwise. So I was happy to have them. And that's why in that chat I did say "Some people just wanna paint by numbers for a while." That "for a while" - it was a flippant statement, but that was part of it... Because I don't think anybody should want to paint by numbers their entire career. What kind of joy in life are you gonna get out of that? That being said, there's worse careers, where it's worse off, so... + +**Kris Brandow:** You'd be surprised... I've met a number of people in my career that have just -- they're literally just like "Just give me the task. I will do it. That's all I wanna do." + +**Jerod Santo:** I have, too. I've never understood those people, but I'm happy that they're happy. I totally get it. But I think you should want to advance. I don't know... From my perspective, you should wanna go on from there. If you don't, Godspeed. + +**Kris Brandow:** Once again, diversity of perspectives. + +**Jerod Santo:** Yeah. + +**Kris Brandow:** Like what you've just said, of when you started, you wanted a nice set of rules. I think the thing that actually got me to be in this industry long-term was that in the beginning there weren't any rules, and I had to figure all of this stuff out. And even though I was quite literally screaming at my computer sometimes, because I had the unfortunate start in Drupal, which if you ever use Drupal, it's like the most complex piece of software on the planet; it's absolutely ridiculous. But that screaming is actually what pulled me in, and what made it worth it for me at the end of the day... That I figured this stuff out. I learned how this thing worked. I understand this now. And that is what gave me the energy to continue going through the industry. And I think if I had had someone just kind of giving me the answers, or giving me the rules, it wouldn't have stuck with me. I would have found something else where I could have done that more curious and creative exploration. + +So I think at the end of the day what we're all saying here is diversity matters, and understanding that we are all different, and we have different backgrounds, and there's no single type of engineer at the end of the day. And we should have roles in organizations and promotional paths that allow us to be the diverse individuals that we are. Not thinking that the kind of average engineer is the one that everybody should go after, or aspire to be. + +**Johnny Boursiquot:** But we won't. We're a cog in a machine somewhere, and we just -- \[laughter\] + +**Jerod Santo:** Unfortunately, there's a lot of work to be had in turning spreadsheets into web forms... And that's what a lot of our jobs are. Take this spreadsheet, put it on the web. + +**Kris Brandow:** Spreadsheets though... Let me tell you, spreadsheets are one of the most advanced forms of programming that is out there. + +**Jerod Santo:** I don't disagree. + +**Kris Brandow:** People are like, "Oh, you just write Excel." I'm like, "Have you tried building something good in Excel?" + +**Johnny Boursiquot:** I've seen enterprise-scale software built in Excel, my friend. You have not seen software until you've seen software written in Excel. + +**Jerod Santo:** Yeah. I've seen efforts to replace Excel with custom software written to do the exact same thing, but the people won't adopt it because they're just -- they built that Excel themselves... Like, "No, this is how I do my job. Don't make me do my job some other way." + +**Kris Brandow:** And half the time that custom software doesn't even do everything the Excel does. + +**Jerod Santo:** That's what they're saying. \[laughter\] + +**Johnny Boursiquot:** Exactly. + +**Jerod Santo:** \[48:04\] That's their own point. "This is worse off than what I have now." + +**Kris Brandow:** I feel like Excel is like a peak of programming. It's just like, I don't know, this is like this super-dynamic, instant compilation and -- you just start plugging in some formulas and you can just do some crazy stuff. + +**Johnny Boursiquot:** Yup. + +**Kris Brandow:** If you couldn't tell, I just built out an entire system using -- not Excel, but Google Sheets. So I am very hyped on the spreadsheet train right now. + +**Johnny Boursiquot:** Hey man, spreadsheets will get you a long way. Don't sleep on spreadsheets. Look at Airtable - they built an entire business around spreadsheets... \[laughs\] I mean, come on... You know. + +**Jerod Santo:** Yeah. + +**Kris Brandow:** Yeah, yeah. + +**Break:** \[48:39\] + +**Kris Brandow:** Alright, so we've gotta get to our last segment, but does anybody have anything else they wanna say before we jump into that? + +**Jerod Santo:** I think the coding versus programmer distinction is kind of silly. + +**Kris Brandow:** Oh, come on, Jerod. Now we have to -- + +**Johnny Boursiquot:** Disagree. I disagree. Hard disagree. \[laughter\] + +**Jerod Santo:** Should I save that for unpopular opinions? \[laughter\] No, we've gotta move on. Well, I mean, we should play maybe a little bit of the audio from Leslie Lamport, or at least let people go listen to what he had to say about it... And Johnny described it a little bit, but he does make this distinction between being a coder and being a programmer... + +**Leslie Lamport:** People confuse programming with coding. Coding is to programming what typing is to writing. Writing is something that involves mental effort. You're thinking about what you're going to say. The words have some importance, but in some sense even they are secondary to the ideas. In the same way, programs are built on ideas. They have to do something. And what they're supposed to do - I mean, it's like, what writing is supposed to convey. + +If people are trying to learn programming by being taught to code - well, they're being taught writing by being taught how to type. And that doesn't make much sense. + +**Jerod Santo:** And I think he elucidated it in a way that was okay... I think that these are interchangeable terms that mean different things to different people, and we end up splitting hairs and inventing our own definitions of the words in order to kind of do a weird form of gatekeeping where we're like "You're not a programmer, you're a coder." And I don't know, I just feel like -- come on, guys... We're all trying to do the same stuff here... So that's my hot take on it. + +**Johnny Boursiquot:** People who actually say that though in real life - yeah, those are d-bags. Don't listen to that nonsense. + +**Jerod Santo:** Agreed. + +**Johnny Boursiquot:** This is something that when you're doing some introspection, and you say "Hey, what does someone like Leslie Lamport mean when he says there's a difference between a coder and a programmer", within the context of the things that he was building and actually putting together and innovating on? What does he mean in that context? + +Or if you go into any other company, do they say "Oh, for these people - we consider them coders. We pay them less than those that we consider programmers." I mean, nobody actually does this, right? So it's supposed to be something that helps you -- personally, I think it's introspection. Like, am I coding, am I programming, am I engineering? I don't tend to use coder vs. programmer, I tend to use coder and engineer more readily than I do programmer, because I think -- + +**Jerod Santo:** What's wrong with programmer? + +**Johnny Boursiquot:** Nothing wrong with programmer. I think engineering is -- + +**Jerod Santo:** I agree. + +**Johnny Boursiquot:** \[laughs\] I think engineering is what happens when you add time to the mix, where your decisions were good decisions back when you were coding. + +**Jerod Santo:** But that's your definition of engineer. I think that we all kind of create these definitions, because I've had these conversations with hundreds of people, and they all define it a little bit differently, and a lot of times it's positioning. And I don't think that's what he was doing. But I think what he was doing was putting two words on what I consider busy work, and then hard work. He was like "Here's where I need to be creative and thoughtful, and then the rest of the time I'm coding." It's like, that's not how other people use the terms, so I felt like it was a weird play. But in his headspace if it made sense, I'm not mad at him or anything. I just think that when we tend to use coder, programmer, developer, engineer, architect, we're positioning. And coder tends to be at the bottom of that totem pole, and I agree with you, people do that; they're kind of d-bags. But they too tend to kind of do that, so I just think it's all kind of silliness. + +**Kris Brandow:** \[54:27\] Yeah. If you don't know who Leslie Lamport is - because he's the same person that's like "Unit testing is BS. You should be writing TLA+ instead." \[laughter\] + +**Jerod Santo:** I heard him say that. I was kind of interested in the TLA+ thing. I'm like, "Dang, am I a total loser?" + +**Johnny Boursiquot:** "Brush up on my math here..." + +**Kris Brandow:** Once again, he's a mathematician. He's like, "Why are you taking this -- oh, does addition work? Let me add 2+2, and 4+" It's like, no, just write a proof to make it work. But I think the reason what he said about coders versus programmers resonated with me so much is because it's something I think I've been talking about for like four years now, of like - there are people that love to study the language, and love to write the language well, and do that well. And there are people that are trying to tell a story with it. I feel like the programmers are the people that are trying to tell a story. They're trying to solve a problem. And they might not be the best at writing actual code. + +I think there are plenty of novelists out there who are absolutely crap with the English language, and their vocabulary is super-tiny, but they tell phenomenal stories. And then there's another group of people who are absolutely fantastic with the English language, and know all of the little intricacies, and read the Chicago Manual of Style to themselves to go to sleep at night... And those people are extremely necessary to us. + +**Jerod Santo:** Speaking from experience? + +**Kris Brandow:** I know people who have. Not me. I tried, but... + +**Jerod Santo:** Okay... + +**Kris Brandow:** But I think for us as an industry, we're missing that. Because the writing industry - those people are called editors, and they are in some regards at the top. You think of the editor-in-chief of a newspaper - that's top dog, and that's not a person who's a writer, that's a person who's an editor. And that's a different training skill. + +I remember I was talking to Angelica once and she mentioned how when you're in J-School, you can choose to be a writer or an editor, and you go down these different paths and you learn different things... + +**Jerod Santo:** Sure. + +**Kris Brandow:** ...and I think that's a distinction we need to make as an industry. Because the reason we can swap these words out so easily is because we haven't taken the time to really say "Does the distinction between these two roles matter?" And I think it does. + +I've run into a lot of people who are excellent at taking code that's written and making it phenomenally better, getting better algorithms in there, making it cleaner, doing all that sort of stuff. But if you ask them to write the code themselves, they'll be like "I don't know..." They'll be like, "That's not what I do." And there's also people that are really good at solving the problem, but don't use efficient algorithms, don't design the code well, don't write comments or document anything... And I think you need to have both of those people in the organization, and I think it's valuable to separate those two things, and I feel like that is, in his own special way, the thing that Leslie Lamport was trying to get at. + +**Jerod Santo:** I think that's fair, and I do think he was being introspective, like Johnny says. I don't think he was being obtuse, or damaging in any way. I agree with you, Kris; I think that we need to go through that work... If we could go through that work and have an RFC or a spec that says "Here's what these different roles mean", and we're all going to agree, then I would be okay with it. It's that they mean different things to different people, and I hear people using them as a way of positioning or gatekeeping a lot, and I think that does more harm than good. + +**Kris Brandow:** Yeah, I absolutely agree with that. We should be looking at these as positions on equal footing, not as like "Oh, you're a coder? You're not as good as me." I hate it. I don't like when people do those things. Anyway... + +**Jerod Santo:** Alright, sorry for throwing that bomb in there when you were trying to transition... + +**Johnny Boursiquot:** \[laughs\] Last minute, he's like "Yeah, hey guys..." + +**Kris Brandow:** Are you supposed to be the one that's keeping us on time? \[laughter\] + +**Jerod Santo:** I'll edit this whole part out... + +**Kris Brandow:** Okay, time for unpopular opinions... + +**Jingle:** \[57:51\] + +**Kris Brandow:** Alright... Ian, do you have an unpopular opinion? + +**Ian Lopshire:** I think I do, actually. + +**Johnny Boursiquot:** Hang on... + +**Jerod Santo:** Alright...! + +**Kris Brandow:** Becomes a host, and now all of a sudden he has unpopular opinions... + +**Jerod Santo:** He's official... + +**Ian Lopshire:** Alright, so I've seen a lot of pushback on Twitter and everywhere about this idea of like a take-home project during interviews. I think that's okay. I think a take-home project is fine. I think there should be a time limit; if it's more than 3-4 hours, it should be paid. But I think that take-home projects - great way to interview. Way better than the whiteboarding and the crazy other things that they do in interviewing. + +**Kris Brandow:** I will agree with you, but only if it's actually something that has to do with what the company does, so no LeetCoding, none of that nonsense... And also, if the person doesn't have a crapload of open source code out there. The thing that is annoying is when it's like "Look, there's all of these examples of me being able to write code in the wild", and they're like "But we just need you to write some code for us." That irks me a little bit. But if it's something related to what the company does, I think it's alright. But yeah, definitely make it paid... And also, give people as much time as they want... + +**Ian Lopshire:** Of course. + +**Kris Brandow:** None of this "Get this back to me in a week." If it takes them a month, it takes them a month. + +**Ian Lopshire:** And by take-home project I don't mean any kind of LeetCoding thing, but an actual... I've had it where I've done an issue on a public repo, that kind of stuff. I think that's one of the best ways to do it. + +**Jerod Santo:** So I would agree with you here... So you've got two for two. + +**Ian Lopshire:** So maybe it is popular... + +**Jerod Santo:** There's good ways of doing it and bad ways of doing it. We had Jacob Kaplan-Moss on the Changelog - I'll link that up - talking about principles for hiring engineers... He's put a lot of thought and time into this process; he's been hiring for a very long time... And he has these rules that you can follow, that he thinks are empathetic and fair, and a good way of doing that. + +He also describes how hard of a problem it is, finding out if somebody is a good match. So I'll link that up; it's worth listening to if you haven't yet. Johnny, are you with him? Can we go full-panel-agreement? + +**Johnny Boursiquot:** Yeah, I think especially when you start to rank from worst to best practice around this stuff, I'd definitely pick a take-home instead of a "Let's flip this binary tree together on this whiteboard" kind of thing. + +**Jerod Santo:** On a whiteboard, yeah... Too popular, Ian. + +**Ian Lopshire:** So a popular opinion. My bad, my bad. + +**Kris Brandow:** \[laughs\] I do feel though that people when they go to vote on the Twitter poll, they'll probably be like "No!" We added a lot of nuance to that unpopular opinion, so... + +**Jerod Santo:** Yeah, we'll cut all that out, too. \[laughter\] + +**Ian Lopshire:** None of that nonsense. \[laughs\] + +**Kris Brandow:** Johnny! Do you have an unpopular opinion? + +**Johnny Boursiquot:** Yeah, I don't know if that's gonna be unpopular, but I know for me something I've been struggling with - passion is so overrated, man... It really is. I have a bunch of projects that I'm working on; I was passionate about them when I first started. Now I'm not passionate about them anymore, but I've still gotta get them done; I've still gotta ship them, I've still gotta do the writing, do the recording, do all the things. + +I wanna move on to the new and shiny thing that caught my eye three months ago, but I can't, because I made commitments and I've gotta get these other things that I was passionate about done. So passion fades. The only thing that really matters is doing the work. Do the work, get it done, ship it, and move on. + +**Kris Brandow:** I agree with you. I think I've read a few things that are like, you know, passion is good to get you started, but it's the stubbornness that actually winds up being the thing that gets you through it all. You've gotta be stubborn about something, and be like "I'm getting this thing done, because I said I wanted to get it done, and I like the idea." But yeah, I think passion is overrated. Ian, Jerod? + +**Jerod Santo:** It's a good way to pay somebody less money to do the same job, is for them to be passionate... + +**Johnny Boursiquot:** Oooh...! I like that. + +2:\[01:01:53.29\] So it seems like -- it could be predatory at times, I think, to call on someone's passion, because their excitement, their pure love allows them to say "Well, I don't need as much money" or "I'll work more because I'm passionate." So I think it kind of can be that way... + +**Johnny Boursiquot:** Hm. F U, pay me. \[laughter\] + +**Jerod Santo:** That's your experience talking, Johnny... No, I agree with you, I think it is overrated... And it's the whole 99% perspiration thing. 1% inspiration, 99% perspiration. It goes back -- I don't know who said that originally. Probably Mark Twain, or Abraham Lincoln, or Albert Einstein. One of those three says all the quotes. + +**Ian Lopshire:** I'm along the same lines. I think passion is an internal shortcut to getting things done, but at the end of the day it's the output that matters. So yes, I agree. It's overrated. + +**Jerod Santo:** No unpopular opinions on this episode. + +**Johnny Boursiquot:** Goodness. + +**Kris Brandow:** Okay. Jerod? + +**Jerod Santo:** Alright, I've shared this one elsewhere, but I'll share it here to see if gophers agree or disagree. I think that most of the time that we as developers spend tweaking our configs, customizing our shell, putting our shortcuts into our text editor, writing those one-line shell scripts - most of that time is time not well spent. + +I think we spend hours to save 30 seconds. I think we could just learn the editor as it exists, and I think that we could get a whole lot done in a whole lot less time if we weren't always tweaking stuff. + +**Kris Brandow:** I agree with that. + +**Ian Lopshire:** I agree. + +**Jerod Santo:** Dang it! \[laughter\] Johnny? + +**Johnny Boursiquot:** I will disagree... I will disagree. + +**Jerod Santo:** Yes...! + +**Johnny Boursiquot:** Because there is a productivity gain that materializes once you understand how you work best, how quickly your hands move across the keyboard, for example... It can be as simple as that. Having the right shortcuts in the right place, or being able to write your script, invoke it on your shell in the morning and it does a bunch of things for you that you otherwise would have had to do manually; it saves you time. + +I think all these things have value, and I will agree that there's a threshold. There's an amount where you have diminishing returns on that stuff. So if I'm gonna do something that yes it's tedious in the world of SRE toil is a thing, and sometimes we can't always fix toil to the degree we'd like to... But because doing the toilish thing that takes an hour this week and the next week and the other week - a total of three hours - as opposed to sitting down and taking up the entire week to automate something... You spend like 20 hours to automate something that takes 3 hours - yes, you've automated it, but what was the gain? Especially if it's something that might go away in two months. + +So there's decisions to be made as to where is the value, where does this start diminishing. So I think there's nuance there, on its face, as you're about to cut out all the nuance that I've just said... \[laughter\] I will say "No, I disagree." + +**Kris Brandow:** I feel like though that you just agreed with him... Because he said that -- + +**Jerod Santo:** He did. He agreed with me. \[laughter\] + +**Kris Brandow:** He didn't say it's all useless... + +**Jerod Santo:** But I hedged. I said "Most of the time that we spend..." And the reason I say that is because we have a natural inclination - and I speak from my own personal experience" to overdo it, to over-automate, to over-tweak, to over-customize... Because it's fun, it has dopamine hits, and it's a good way to procrastinate and not do what we're actually supposed to be doing... So we tend to overdo it. So most of the time we spend, it's not time well spent. I'm not saying none of it. Definitely, if you can save 30 minutes every day by spending two hours one day, and you're gonna run that for the next year, you're way in the green. Or in the black? I don't know colors very well. I need to paint by numbers. \[laughter\] + +**Kris Brandow:** It's in the black. + +**Jerod Santo:** You're in the black. + +**Johnny Boursiquot:** Then you just have to change your opinion to "Always" not "Sometimes", and then I'll be right. + +**Jerod Santo:** Okay. That would be more bombastic, probably more unpopular. I'm still learning. I'm still learning how to be as unpopular as I can be. + +**Kris Brandow:** \[01:06:00.10\] Okay, I've got an unpopular opinion... + +**Jerod Santo:** Let's hear it. + +**Kris Brandow:** I was thinking about this episode and I was talking to a friend about it... I will make the assertion that Rust is the Esperanto of programming languages. + +**Johnny Boursiquot:** Oh...! + +**Jerod Santo:** Hm... + +**Johnny Boursiquot:** \[laughs\] + +**Jerod Santo:** Say more. \[laughter\] + +**Johnny Boursiquot:** And for those who don't know, you'd better go watch Encanto. + +**Kris Brandow:** Esperanto is this language that's designed to be like easy to learn... Basically, a better version of all the languages that we have, and to do all those things in a more capable way, by making it easier for people to communicate in all of this... And it has all these aspirations, and in some ways it's kind of mean to replace languages like English, which are difficult to learn, very wordy, really annoying if you don't learn it as your first language... And I feel like that's kind of like what Rust is, where it's kind of targeting, trying to replace C, replace this dominant language in the world. And while it is a nice language - Esperanto is a nice language; I know a fair bit of Esperanto - the aspiration that it'll be the language that everybody will speak, or it will replace something like C in the future is just incorrect, and it will not do that. + +I'm not saying we shouldn't have it, I'm just saying that it's basically Esperanto, but in programming language form. + +**Jerod Santo:** Okay. I don't have a take on that. I don't know enough about Rust. I can't either agree, nor disagree. + +**Johnny Boursiquot:** I just thought you were talking about Encanto. \[laughter\] + +**Kris Brandow:** No, no, no... + +**Johnny Boursiquot:** That's an actual language... And you learn something new every day. + +**Kris Brandow:** Yeah, Esperanto is a real language that people around the world speak and write. + +**Johnny Boursiquot:** Alright. You learn something new every day. + +**Jerod Santo:** Where is it mostly spoken? + +**Kris Brandow:** It's one of those global languages. So it's not centralized anywhere. Literally, someone in the early 1900's sat down and was like "I want to design a good language." So they sat down and they literally created a language, and was like "Okay, this should be easy to learn, it should be very consistent..." + +**Jerod Santo:** Right. + +**Kris Brandow:** It is a very consistent language. It has a nice set of features. It's kind of the opposite of German, or something, where it's just like "Why do I have nine different versions of "the"? It's kind of the opposite of that. + +**Jerod Santo:** Right. I think it's like the XKCD, there's too many specifications. What we need is one more specification that fixes all the problems with the other ones. And then there's one more specification... + +**Kris Brandow:** Yeah, yeah. + +**Jerod Santo:** N+1. + +**Kris Brandow:** That's where Log15 got its name. The popular Go logging package Log15 got its name from that, from the XKCD comic. + +**Jerod Santo:** Wow. + +**Kris Brandow:** That's why it's called Log15, because there are 14 specs; we need to make a new one, and now there are 15. + +**Jerod Santo:** Oh, it's the 15th logger. + +**Kris Brandow:** Yeah. + +**Jerod Santo:** That's cute, I like that. + +**Johnny Boursiquot:** That's a resume-driven project right there. \[laughter\] + +**Kris Brandow:** "I wrote a logger." Okay, well - I guess my opinion is not unpopular... Or I guess it's neither. + +**Jerod Santo:** We'll see. We'll have to vote it out. We'll have people -- I don't know if gophers are gonna... Is it a chance for gophers to dunk on Rust? It might go popular just because of that. + +**Kris Brandow:** Hm, yeah... + +**Ian Lopshire:** If I saw this on Twitter, I'd click "Yes, I agree", but I don't have a reason. \[laughter\] + +**Jerod Santo:** That's how most of the voting goes, I'm sure... + +**Kris Brandow:** I told my friend that I was talking to, I was like "This is gonna be my unpopular opinion in the next episode", so... I'm sticking to that. + +**Jerod Santo:** There you go, friend. + +**Kris Brandow:** Alright... Well, this has been a fun episode, hopefully not too meta for our listeners out there, and you've learned a bit about the myth of incremental progress... I feel like it's buried somewhere in the content of this episode... But yeah, so thank you for joining me, Johnny and Jerod, and welcome, Ian, and thank you for joining as well. + +**Jerod Santo:** Thanks for having us. + +**Ian Lopshire:** Yeah, it was fun. diff --git a/The other features in Go 1.18_transcript.txt b/The other features in Go 1.18_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..f605f6c1de043ca2f532902a76b46e52fe27a6ab --- /dev/null +++ b/The other features in Go 1.18_transcript.txt @@ -0,0 +1,451 @@ +**Mat Ryer:** Hello, and welcome to Go Time. I'm Mat Ryer, and today we're talking about the other features in Go 1.18. Now listen here, you... Come here, come here. Go 1.18 has got two great, big features that everyone's talking about. I mean, everybody. All the popular people are talking about it. Everyone. Well, we're not popular. We're gonna be talking about the other features, not those two. And just for anyone that doesn't know, it's fuzzing and generics. And that's the only time they're gonna be mentioned on this episode. + +In fact, we've got a new rule - if they are mentioned by anyone, even accidentally, unfortunately you will be immediately booted from the podcast. So please bear that in mind. No talk of those two subjects. Pinky promise? That's a pinky promise from me, and we'll find out if we're also gonna get a pinky promise from our special guests today. Joining me is Daniel Martí. Hello, Daniel. + +**Daniel Martí:** Hello. Nice to be back, and nice to bring my technical problems along with me. + +**Mat Ryer:** \[laughs\] Your technical problems are, like you, always welcome, Daniel. Daniel's been using and contributing to Go for quite a few years now, and you've actually written a few tools as well, like the stricter `go fmt`, and what could be described as the opposite, a Go code obfuscator. They're interesting tools, Daniel... How is the `go fmt` more strict? + +**Daniel Martí:** It essentially restricts how you can write and format code in a few extra ways. For example, no empty lines at the start of the function body. Things that I generally do. + +**Mat Ryer:** Cool. Okay, we'll put a link to that in the show notes, if anyone likes -- I like the fact we have `go fmt`, and I like the idea of a more strict one. Oh, Daniel, do you pinky-promise you're not gonna talk about those two other subjects today? + +**Daniel Martí:** Pinky-promise! + +**Mat Ryer:** \[04:03\] Okay, Daniel's pinky-promised. Okay. This is really professional. Okay, we've also -- you're not gonna believe this, Daniel... You will believe it; you already know. But imagine if you didn't... We've also got Michael Matloob with us. Hello, Michael. Welcome to Go Time. + +**Michael Matloob:** Hi! Great to be here. + +**Mat Ryer:** Oh, it's a pleasure to have you. Michael's on the Go Tools team at Google, living in, you know, New York City, no big deal... And he previously worked on go/packages, which is very useful if you're writing code generation tools and things like that, and infrastructure for tooling. And now he works on the Go command, right Michael? + +**Michael Matloob:** Yes. + +**Mat Ryer:** And do you pinky-promise to not mention those two big subjects on this very episode? + +**Michael Matloob:** I won't mention them by name. + +**Mat Ryer:** Okay, that sounds like a good pinky-caveat... + +**Michael Matloob:** It is. It is a pinky-caveat. + +**Mat Ryer:** Okay, fine. We won't mention them by name. We'll see how it goes. So this episode -- obviously, there's a lot of people blogging and talking about the big headline features. A lot of people are very excited, a lot of people are very dismayed about generics in particular; I've just said it, and I can't believe that I have to leave the podcast... But a raft of other things. And this release in particular seems very packed and dense with features. Why is that, do we think? + +**Daniel Martí:** I believe they've been saving a few large features for some time. They've been building up the generics for nearly two years now, and now it's shipping, right? And I did mention the taboo subject, but I think it's also happened with fuzzing, which has been in the works for like a year now. + +**Mat Ryer:** Yeah. I see. By the way, just mentioning another taboo subject doesn't cancel out the other one. You're just compounding your crimes. I asked Daniel and Michael to find a list of the things that they're sort of excited about, or interested in, that we can go through and talk about... And obviously, Michael worked on module workspaces as well, so we'll carve some time out at the end to talk about that in particular. But Daniel, maybe you could kick us off... There's a really interesting one that to me seemed like a silly, unnecessary helped, but it turns out to be actually quite worthy. That was strings.cut. Could you tell us about that? + +**Daniel Martí:** Yeah, so I think anybody who's written any non-trivial amount of code knows that they have to deal with strings, they have to add strings, look at prefixes and suffixes and so on, and one quite common operation is wanting to cut a string in two. For example, maybe you've got a domain name and you want the actual name and the extension. Or maybe you've got a file name and you the file name extension that kind of thing. You can use Go APIs like strings.index, or there's also strings that split N, and you can give it the number 2. So like "split this string in up to two pieces." But these APIs are not super-easy to use. For example, if you use index, it may give you -1, and if you don't check for that, that might panic. And split has the same issue, because it gives you a slice. You could say cut has less sharp edges, so it only gives you two strings for the two sides, and a boolean telling you whether or not it's successfully cut. + +**Mat Ryer:** Yeah, that's nice. So if say you were cutting on a colon and there wasn't a colon in there, it wouldn't be in any way like a panic or a problem. You'd just get a false as the second argument. + +**Daniel Martí:** Exactly. + +**Mat Ryer:** Yeah. What do you think about that, Michael? Have you written code that cuts things up like this? + +**Michael Matloob:** I have, yeah. It would be a nice convenience. I like conveniences. + +**Mat Ryer:** Yeah. I thought this was like an unnecessary helped, because whenever you can already do something, that's usually my preferred way. I looked at some of the commentary on this one, and the number of cases where people are basically doing the same operation over and over again is kind of everywhere. And including some places where we've done it incorrectly, or in a way that would panic if it got some bad input, or something. + +\[08:05\] If there was some testing tool that helped you test out all these different possible ways of responding to input, that'd be great; but not on this episode there isn't. But yeah, okay... So strings.cut, and that's coming in Go 1.18. Okay, Daniel, have you got another one for us? + +**Daniel Martí:** So I've got another one that's significantly more complex than strings.cut, and I believe it was developed by the people at Tailscale over a few years. It's essentially a replacement for the net.ip type. Right now, IP addresses in Go - they're represented as a byte slice. So you can think of a byte slice that can have many lengths. So an IPv4 is gonna be shorter than an IPv6, for example... And they designed a new IP package which they've called netaddr, but now it's being merged as net.ip. So it's /netip, and it's got a bunch of advantages, mostly related around performance, but the two main properties that it has as part of its design, which is "do not use a slice", essentially... One, it's comparable, so you cannot compare slices; you can only compare them to nil. And the other one is that it doesn't allocate. So you can create a new IP without calling make, or new, or anything like that, because I think it's backed by what is essentially a bunch of integers. + +**Mat Ryer:** So will the standard library - bits of it - be rewritten to use this new type, or is this just gonna be something that's available for calling code? + +**Daniel Martí:** I think that's a good question. I think anything that exposes APIs with the old type will have to remain the same because of backwards compatibility. I seem to recall one of the reasons to add this to the standard library is so that, for example, HTTP/2 and /3, which -- I think it's only HTTP/3 which reimplements something like TCP write in user space, and that deals with a lot of IP addresses. So if you can remove a bunch of internal allocations that don't leak into the API, that can be a very large plus. + +**Mat Ryer:** Yeah, that's nice. And do you know if there are gonna be helpers to switch between the two? Do you think we're gonna see code like that flying around for a bit? + +**Daniel Martí:** I believe the package comes with helpers, but my memory is failing me... + +**Mat Ryer:** Okay. I mean, if not, people will probably end up doing that, I imagine... But yeah, it's nice to know that there's a sort of improved data type there. And it's funny - you know, with the Go backwards compatibility promise you can't just break things, and break APIs, and break everyone's code. It's not Python... I shouldn't have a go at other languages, but Python does that a lot... + +But with that promise, of course, your hands get tied. So this is kind of a nice way of releasing almost like more modern implementations, is to sort of release them alongside, and then they kind of co-exist. But does that create confusion? Like, how will people know which one to use? + +**Daniel Martí:** That's a good question. I think the Go standard library has a bunch of cases where there are packages and APIs that everybody knows not to use. There's container/list, that has like a linked list, for example. I don't think everybody's used that outside of an example... I don't think they can deprecate net.IP, as in the existing net.IP type, because it is used in existing APIs. But I think there is gonna be a common understanding that if you want the extra-nice features of the new type, you should just use it. + +**Mat Ryer:** Great. Thank you, nice one. Alright, Michael, maybe you could pick one to talk about next... + +**Michael Matloob:** I'll pick a couple of features that my colleagues Jay and Ryan added to the Go command. So they are vcs build stamping, and a debug buildinfo function to get information about the versions of modules in a Go binary as an API. + +**Mat Ryer:** Hm... + +**Michael Matloob:** \[11:49\] Both of these have a similar core motivation, which is like to improve visibility into binaries and which packages they were built on, so you can determine, say, if binaries were built with certain commits of code, in the case of vcs build stamping, because the main module may not have a version associated with it, and in the case of a buildinfo of dependencies. + +This is shaping up to be a big thing in these days, to know whether your dependencies and the code they are built with have bugs or bad features in them, and if the code that you're running with is safe, and to audit everything properly. We've seen several cases of bad libraries in the wild, and people have to quickly audit if all their code running in production is safe or not. + +**Mat Ryer:** Yeah, and it's a tricky thing. So unpicking that a little bit then... So vcs, version control systems, like the Git hash when you have a certain level that you've committed up to, and everytime you commit, you get a new hash... And so now when we build, that will be incorporated. Is it like it supports all the major vcs systems? So it'll support Git, and... + +**Michael Matloob:** Yeah, yeah. I mean, it definitely supports Git. I don't know other vcs'es we support. Dan, do you know? + +**Mat Ryer:** But they will be coming soon. + +**Daniel Martí:** I think there's Mercurial, Bazaar, Subversion, and that might be it for these days. + +**Mat Ryer:** It's a good selection... Could you name five? \[laughs\] + +**Michael Matloob:** Five vcs systems? + +**Mat Ryer:** Yeah. Just five, dead-quick. + +**Michael Matloob:** Well, Dan named four, so... + +**Mat Ryer:** I know. So it should be easy. + +**Michael Matloob:** I've seen in the Go command a vcs named Fossil that had support somewhere... So that'll round us out. + +**Mat Ryer:** There we go. Fossil. There we go. Five. Brilliant. Learning. Okay, and the other thing is that buildinfo, with all the dependencies, because that is a big thing. We're paying a lot more attention now to reporting vulnerabilities, capturing that data, and then being able to, in the tooling, use that to know whether we are running something that has some known vulnerability. So that is a massive thing... And of course, having this put in there automatically saves us a lot of effort, right? + +**Michael Matloob:** My understanding is the buildinfo is like a function that's like accessible to programs that were just in the Go command before, like go version -m. So it makes it easier for other people to write these auditing programs that can help detect if there are bad versions in your dependencies, and then flag it, or fix it, or whatever. These automated things I hope can solve a lot of manual human work that we've had to do when these issues have come up. + +**Mat Ryer:** Yeah. Well, it's very useful if you have Dependabot running in your continuous integration, or just running in GitHub... Having those tools help. It's all great, so anything in that effort I think is worth having. Very cool. + +And yeah, before we had to use either build tags or do something else funky to get the version. I would always do that, I would have some script that would -- I think I did it with Go Embed as well, successfully, recently. But we just don't have to do that now. And so will we be able to access that version inside the build as well from somewhere? + +**Michael Matloob:** Sorry, access the -- + +**Mat Ryer:** Access the Git hash inside the binary itself? + +**Michael Matloob:** Yeah, it's in the binary. I don't know what the API is... + +**Mat Ryer:** Yeah. But there's some - either method, or... + +**Michael Matloob:** Yeah. + +**Daniel Martí:** It's honestly a bit confusing, because before there was an API to get the module information of yourself, like of your own running binary, but the new API they've added - you can give it a binary path, so you can use it with any binary, without having to shell out to go version -m, blah-blah-blah. It's essentially the same feature. + +**Michael Matloob:** Oh, so the buildinfo includes that, the hash. There we go, okay. + +**Daniel Martí:** \[16:03\] And I think the vcs stamping is also a bit confusing to end users, because you tell them "Go 1.18 now stamps vcs buildinfo", but they might say "If I go install a Go main package that I run Go version -m with Go 1.17, where do you see the module version, right?" But where that doesn't work is if instead of doing a global Git install via a module path, if you Git clone and then go build or go install locally from that Git clone, Go doesn't know what module version that is; it just has a Git clone. It's not resolving that module through the whole proxy system that tells it what version it is. In Go 1.17 it tells you version devl it has no idea. In Go 1.18 it will add some extra, separate metadata that will say "Hey, this was built from Git hash blah-blah-blah, date blah-blah-blah", and so on. + +**Mat Ryer:** Oh, right. Okay. Daniel, your turn to pick one from the list. What else is cool coming in Go 1.18? By the way, do you say one eighteen, one dot one eight, one point eighteen...? How do you say it? + +**Daniel Martí:** Now you're making me doubt myself about how I pronounce these things... + +**Mat Ryer:** Yeah... You just have to forget -- just clear your mind and then just say it. See what happens. + +**Daniel Martí:** One eighteen. + +**Michael Matloob:** Yeah, I say one eighteen. + +**Mat Ryer:** Yeah. Okay. That's good. It is kind of one eighteen; it's not a decimal number, is it? It's semver, so that second number is eighteen. I think we're right. + +**Daniel Martí:** So does that mean that when we reach 1.20 we can go back to one point two? + +**Mat Ryer:** I think for just that release. No, you can't, can you? No, because that's what I mean, it's not decimal. That's it, yeah. Yeah, we got there in the end. + +**Daniel Martí:** So another feature... I mean, it's maybe a bit cheeky that I bring this up, because I worked on this... But `go fmt` without a space now formats files in parallel. So up until we had -- well, you have two tools, which is also confusing. You have `go fmt` without a space, and then you have go fmt. + +**Mat Ryer:** Yeah, it's ironic that the `go fmt` tool can be called in different ways just by changing the formatting... Yeah. + +**Daniel Martí:** Oh, God... The difference between the two tools - and I think it also confuses a bunch of users - is that without a space it takes files and directories, but it doesn't know what packages are... And with a space, it takes the package pattern. So you can give it `./...` for example. And the one that works on packages has always been relatively well parallelized, because what it does is I believe it formats each package in parallel, or something like that, but the one that takes directories and files - it would just do one at a time. And now we've essentially removed the parallelism from the one with the space, and just made both tools use the same kind of parallelism, which is `go fmt` (without a space), when you give it a bunch of files to format, it's just gonna figure out how to format them as fast as possible. + +**Mat Ryer:** That's cool. Does `go fmt` work only within the context of a file at a time then? Like, it doesn't need to know anything else about types, and things, because it's just doing formatting tasks? + +**Daniel Martí:** Yeah, that's correct. + +**Mat Ryer:** So it makes sense. You just do all that at the same time. + +**Daniel Martí:** Yeah. Even though there's a few tricky bits about that, because initially, my naive implementation was just format each file as a separate goroutine as they come in... But some files are really, really tiny. Think like a `dock.go` file that only has like ten lines, with like a package documentation or something like that... And spawning a new goroutine, synchronizing with append, maybe allocating the new parser, the new printer, and stuff like that - it actually consumes quite a lot more CPU just because of the overhead of all those tiny files. So we ended up with something that's kind of like chunking groups of files in groups of similar sizes, so that they're big enough that actually doing that as parallel units it's fast. + +**Mat Ryer:** Oh, that's really cool. That's a surprise. I would not have expected it to be doing that... But it's nice to know that that's measured and done properly. That's very cool. Have you used this then? Did you really notice this in practice, the speed improvements? + +**Daniel Martí:** \[19:59\] I think it depends on what people do. I think many people use the tool that works on packages, and then they just format their packages... But I like using the one with directories. So I go to the root of my repository and I just tell it "format everything", including test files, including everything. And because I did that, it was really slow before. So now, depending on your machine, it's usually about 3-4 times as fast. So for me, for example formatting a large repo might go from like five seconds to two seconds, which is nice. + +**Mat Ryer:** Hm... Matloob, do you format your code? + +**Michael Matloob:** Yes. I mean, we all format our code... + +**Mat Ryer:** It's not a trick question. + +**Michael Matloob:** No, is there anyone who doesn't format their code? + +**Mat Ryer:** Yeah, I don't know... + +**Michael Matloob:** Because I wanna hear about it. It's like a problem we need to solve. + +**Mat Ryer:** Oh, yeah... I don't think so, because you only have to do it a few times, and then when pull requests -- although they improved it in GitHub, where white space was understood better... But it certainly used to be that you'd get just pull requests that every line has changed, because some white space thing. And that got so annoying that it's a very high motivator, I think, to get people formatting... But I don't know, I assume everyone does format their code. Do you do it in the way Daniel described then, or do you do it like me, where just every time you save a file, it does just that file? + +**Michael Matloob:** I don't think I've ever run either of the tools... Or I certainly haven't run either of the tools by hand in years; my editors are just set up to format files as I save them. + +**Mat Ryer:** Oh, yeah. I thought you were saying that you just write it in perfect `go fmt` way the first time. + +**Michael Matloob:** Oh, no, no, I don't. + +**Mat Ryer:** Nailed it. + +**Michael Matloob:** I write it in the wrong way and just let the formatter take care of it, like any good Go developer. + +**Mat Ryer:** Yeah, exactly. To be honest, I'll deliberately make mistakes so that when I hit Save, I get a visual clue that is has formatted. Because if I write it and I get it right, and then I hit Save and nothing happens, I'm like, "The computer is no working." So I genuinely sometimes like to see the little shift into place of things as a clue that it's working... + +**Michael Matloob:** Yeah, that's like a nice way to know that "Oh, the syntax is correct."1 + +**Mat Ryer:** Yeah, that's true. + +**Michael Matloob:** You know it can parse. + +**Mat Ryer:** Yeah, because if it errors, it doesn't complete it. So it actually is a feedback loop thing. There you go, a tip there for everyone. + +**Daniel Martí:** I've actually done that with tests. If you write a ton of software and some tests, and you run the tests and everything is green, you often go "I don't believe that. Let me break one of the tests to see if I'm doing something really dumb right now." + +**Mat Ryer:** Yeah, absolutely. In TDD you talk about that red-green testing for that reason; you have to see the test fail, so you know it's saying something useful; and then when you fix it -- that's true. If I write some code, even if I'm just running it, and I'm gonna run it myself and look at the results in the terminal, without even any tests, if that works the first time, I'm highly suspicious; really suspicious. So yeah, in a way, I'm not happy when it does. + +Okay, we've also got the pacer redesign in the garbage collector. That's interesting... What's going on there then? + +**Daniel Martí:** I brought this up because I think it's a very interesting topic... But I think we should also warn that none of us here are experts in this area. So we can talk about it at a high level, but I'm gonna stop there. If anybody wants to read more about it, we can mention the issue number and then they can go and read the whole doc. And I think that's very reasonable. + +**Mat Ryer:** Yeah, fair enough. Good disclaimer. + +**Daniel Martí:** To give a bit of an intro, the way I understood it - because again, I've just read this... The GC pacer - it's the part of the garbage collector that decides when a new collection should happen. It's sort of the thing that times when the GC should be doing its work. Because if it happens too often, then you're just burning too much CPU, you're wasting time; but if you run it too little, you might be holding on to too much memory, or you might be delaying some things happening in the runtime that you don't want to delay by very long. + +It seems like the GC pacer was designed a while ago. For the purpose that it was designed, it was good, but over time it's accumulated a bunch of debt and a bunch of quirks... And they've sort of sat down and said "Okay, let's redesign it in a way that it does a lot better in these edge cases that we've found in production workloads, that the old one doesn't do very well in." And I think that's where I'm gonna leave it. + +**Mat Ryer:** \[24:16\] Oh, it's very exciting. I'm really interested whenever there are these kind of really low-level -- because it's funny, when you dig into these little subsystems, they're just like other types of programs. They are just doing the same things that we're doing in our programs, but it's such an interesting domain... It always makes it more interesting. I love the fact that as programmers we get this for free; people are doing this work for us to make these improvements... I didn't even know about a pacer, to be honest. So it's very nice to know that that's happening. What do you think about that, Michael? Do you know anything about this? + +**Michael Matloob:** No... I mean, I am not closely acquainted with it, but I think any runtime improvements are well-appreciated. Good work, team. + +**Mat Ryer:** Yup. And there's another Michael who I think was the author of the redesign, is that right? + +**Michael Matloob:** Yeah... + +**Mat Ryer:** Do you know all the other Michaels on the Go team? Have you got together yet with all the rest of the Michaels? + +**Michael Matloob:** There's a Michael -- + +**Mat Ryer:** Or it may just be a contributor, actually... + +**Michael Matloob:** There's two... Are there more than two? Or they're three, I guess, including myself. + +**Mat Ryer:** Okay. + +**Michael Matloob:** I don't wanna be forgetting anyone, so if I forgot -- + +**Mat Ryer:** No, I think we should spend time on this. Don't forget anyone. + +**Michael Matloob:** We should not spend time on this... + +**Mat Ryer:** Mind you, you're just telling me a number; even if you forgot a Michael, they don't know which one -- they don't know that they've been forgotten. I think it's safe. + +**Michael Matloob:** There you go. So everyone can assume that I included them in the list of Michaels. + +**Mat Ryer:** Yeah. So calm down, Michaels; you were counted. All Michaels have been accounted for. + +**Michael Matloob:** Yeah. + +**Mat Ryer:** Excellent. + +**Daniel Martí:** I was thinking, before we go on to the next topic, if anybody wants to read about this, the issue number is \#44167. And at the end of the issue, which is very short, there's a link to the full proposal design, which is very long... And you can read that carefully, and get the full picture. + +**Mat Ryer:** Yeah. It looks very well written. And we'll post the link to all of these in the show notes, so you'll be able to go and actually look at the original issues. And honestly, notice that some of these issues aren't created by members of the Go team, or even popular contributors like Daniel, who've contributed massively. Sometimes these come from just people in the community that have a problem that they wanna solve, or something they care about. So do get stuck in, basically, because you never know, you might get some improvements made, and that'd be great for everyone. + +**Break:** \[26:49\] + +**Mat Ryer:** Does anyone have the M1 chip? Apple's M1. + +**Michael Matloob:** I have it on my personal laptop. + +**Mat Ryer:** Yeah, that accounts. It's fast, isn't it? + +**Michael Matloob:** Oh yeah, it's great. I've been surprised with how fast it is. + +**Mat Ryer:** Me too. I got a new MacBook Pro recently, and it's phenomenal, absolutely. But Go had support for the M1 chip for quite a while, didn't it? What does that look like? How do we support another chip? Could someone just briefly -- and I do mean briefly; we don't need to get into the weeds of it. But what do we have to do? Is it literally we have to add some kind of mapping file for all the instructions so a compiler knows what to compile them into, and it's different if it's a different chip? Because there's also the Rosetta 2 stuff, so that even if a binary on these new architectures hasn't been built for that architecture, this is a translation layer... And to be honest, they're still lighting fast, as far as I can see when I run programs like that. But there are some improvements coming, is that right? + +**Daniel Martí:** I do seem to recall that when the M1 first came out, Go did already support ARM64, so the 64 version of the ARM architecture... But binaries build for Go targeting the architecture didn't work out of the box, for one reason because there wasn't a darwin/ARM64 port yet. So Go did support Mac, and it supported ARM64, but not together yet... So they needed to add some glue code to essentially make those two work together. And I think the other major work they had to do was the whole thing about signing binaries... Because I think the M1 was the first machine that required old binaries to be signed. So they had to teach the linker how to sign binaries locally, something like that. + +**Mat Ryer:** Yeah, yeah. Oh, that's very cool. Well, I just noticed it started working. + +**Michael Matloob:** There's also a lot of work that needs to be done when we're making releases; when Apple makes changes to their operating system, we often have to change the infrastructure we use to produce the Go distributions that people get... And that takes a lot of work. I kind of just wanna mention all the work that the Go release team has done to make our releases smooth... Because sometimes that goes -- it's not explicitly talked about as much. + +**Daniel Martí:** So I imagine every time Apple says a new major version of macOS is coming, I imagine some people start sweating, thinking "Oh no, what is coming..." \[laughter\] + +**Michael Matloob:** I mean, sometimes there's nothing, but sometimes they're destructive. Was it Catalina that they introduced major sign in requirements that caused big problems. + +**Mat Ryer:** Well, yeah, again, we do appreciate all that work. Newer x86-64 machines are also getting improvements, aren't they, Daniel? + +**Daniel Martí:** Yeah, so that's a good segue, because going from, for example, ARM-based machines, there's a lot of versions; if you have an old phone, I believe that's gonna be like ARM version 6. But later phones are gonna be ARM version 8 or 9, which is 64 bits. And if compile a binary that's targeting the lowest-possible denominator, the older version, it's not gonna run as fast as it could on a newer device. So Go has had a flag called -- I think it's called Go ARM64, and you tell if what version of the architecture your target machine supports, and then if you swap a 6 for a 9, it might run 10% faster, depending on what kind of code you're running. And x86-64, i.e. AMD64 desktop CPUs - they don't suffer from as much of the same problem, because they haven't had as many versions, with as many changes in the last decade or two... But you have had some changes. + +\[32:13\] And sort of marrying the same environment variable for ARM64, now we have Go AMD64, and it targets one of four versions. These are sort of standard versions between Intel and AMD, where roughly speaking I believe version 1 is like the common denominator; it's basically every single machine that's a valid AMD64. And then you've got version 2 for things that are starting in 2010 or so, version 3 staring in 2013-2014, and then version 4, which is I think AVX-512, which has mostly server computers or very new desktop computers. + +So if for example you knew you were targeting a cloud machine and you know the cloud machine has all these new instructions, you can swap from the older version 1 to version 3 or 4, and maybe you're gonna save 5% or 10% CPU cost, depending on what kind of code you're running. + +**Mat Ryer:** And presumably, if you choose a higher number and then the architecture is lower, then that's a problem. + +**Daniel Martí:** I believe it's just gonna refuse to run. It's gonna say not supported. + +**Mat Ryer:** Okay, cool. Yeah, it makes sense. There you go, that's good to know. Often I'm so abstracted from the physical hardware in certain environments, to where I would be able to make use of that... But there's certainly some cases where I could probably use that today. I appreciate you telling me about that one. + +**Daniel Martí:** And even if you think "Well, my workload is not that special", I believe in the Go AMD64 version 3 there's an instruction that the runtime garbage collector can use to quickly scan memory for pointers or something like that, in a way that essentially batches the work and makes it a lot faster... So you might get the runtime GCs being like a few percent faster, even if you don't care about new CPUs. + +**Mat Ryer:** Oh, so even if you're not gonna make use of it, maybe the Go tooling and runtime and bits and pieces do. Very interesting. I do wanna speak about one more subject before we get onto workspaces, if we can... And this is something I use a lot, and these are the templates in Go. So we've got text template and HTML template. These sometimes get criticized as being too rudimentary and to low-level, but it sort of has enough of what you need; as long as you mix in Go code, usually in functions that you make available to the templates, you can really do everything you need. But are we getting some new functionality in templates? + +**Daniel Martí:** Yeah. So I added a couple here, which are pretty simple to understand, I think. They both revolve around control flow, or logic, if you wanna think of it that way. So one is about adding break and continuing. So it's the same feature that you have in regular Go loops, but for ranges within a template. + +And the other one is that the and and or operators in boolean expressions now short-circuit in a template, like in Go. Which means that if you do a or b and a is true then b is not evaluated. Whereas right now it evaluates all the expressions and then works out the boolean expression. + +**Mat Ryer:** Yeah. And the result on the expression itself is the same, isn't it? But if you're calling functions within that, then you can save those functions; they won't need to get called. So that short-circuiting sometimes is very important. That's very nice to know... + +So the break and continue - I guess they are quite simple then. So continue is gonna loop back and -- actually, I'm not sure it is that simple... Because the template is kind of declarative, isn't it? What does a continue do then? What happens if there was, within the block, content after the continue? Is that skipped? + +**Daniel Martí:** So you can think of templates as sort of scripts. I don't believe they let you run code forever... At least not that I can remember. But they do have a range statement, where you can say range over for example a slice. And then within that body you can set variables, or you can template them. + +\[36:09\] Like, if you just type something without using the brackets, that's gonna be output as part of the template. If you have two blocks of code within a range, and then between you say continue, then the second block is gonna be omitted, and then you're gonna go back to the top of the range. + +**Mat Ryer:** Okay. So that is how it works in Go, so that should feel quite natural. But that is quite unusual for templating. I don't think I've seen that before. + +**Daniel Martí:** It is a bit unusual, yes. + +**Mat Ryer:** Very cool. Well, we have somebody here, of course, Michael Matloob, who has done a fair bit of work recently on workspaces. This is coming in Go 1.18. Michael, could you just tell us briefly what are Go workspaces? What problem do they solve? + +**Michael Matloob:** So just like at a simple level, the Go command in the module mode allows you to have a single main module that you're working on. That's the module that your current directory is in, and all the files in the module, all the packages in the module are the packages that Go builds by default. And if you have any other code on disk, previously you would have to add replaces, or other ways of getting it in which are kind of annoying if you wanna make changes cross-modules. It was hard to work across two modules at the same time, basically. + +Now, workspaces - you have more than one main module. Those are modules where you are making edits and go builds from, rather than getting it from a specific version. So workspaces allow you to say "These are the modules on disk that I'm working on", and those are the base that the minimal version selection uses when computing its dependency graph. + +We think this is gonna be useful because we've gotten a lot of feedback from people who work across multiple modules. In fact, that was one of the number one complaints we saw in the Go user survey, people working with modules - they had problems when working on multiple modules; they found it cumbersome. So we hope that multi-module workspaces make that workflow a lot easier for them. + +**Mat Ryer:** Yeah, this is definitely something I've encountered. Do you think people were overusing modules? Do you think that we were doing something wrong? It felt like that, because we were kind of fighting with the tools a little bit... + +**Michael Matloob:** What do you mean overusing modules? + +**Mat Ryer:** Well, I mean like sometimes in a project you have multiple packages. Sometimes people will -- each one of those will be a module, instead of just a package inside this bigger module. Things like that. + +**Michael Matloob:** Yeah, I feel like one thing that we learned after some experimentation with modules, after some time using vgo and then modules in the Go command, we learned that multi-module repositories should be rare. They have a lot of surprises when you're working with them. So now our general recommendation is for people to usually have like one module per repository, unless there's a specific, very rare set of use cases where they wanted to have a sub-module in their module. So in that sense, yeah, I guess we were overusing modules, because we were learning how to use modules. Now those modules exist, and we kind of have to work with them. + +**Mat Ryer:** Yeah. I think for packages and things if you're releasing a library that people are gonna use, I think that's kind of great advice. Definitely a time I've seen where multiple modules exist is if you have a monorepo... The way that you would do it at the moment - I use Visual Studio code - you basically open the subfolder just as the root, and that essentially becomes the context of that module, and that's a way to get around it. If you have multiple folders and they have modules in different ones, workspaces I think is gonna enable that now. So you'll be able to operate, right? + +**Michael Matloob:** \[40:18\] Yeah. I mean, one of the driving forces behind us starting to work on modules was the user experience in not just Visual Studio Code, but any editors that user Gopls, which kind of powers the Visual Studio Code Go experience. The team was thinking of different ways of representing multiple modules and providing that information to the Go command... But it had to introduce a new concept that didn't exist in the Go command. The Go command had no concept of people working in multiple modules at the same time. It just had replace directives or requirements. + +So we decided the best thing to do is to make this a first-class feature of the Go command, so not only could Gopls use it, but users who introduce modules can then open up command line and the Go command understands that they're working in the same workspace, in the same set of modules. + +**Mat Ryer:** So how does it actually work in practice then? Do you have to set up a workspace? Is this a new concept of a thing you create? + +**Michael Matloob:** Yes, so we have go.mod files, and now we have go.work files. So you create a go.work file, the syntax is very similar to go.mod; we wanted it to be easy for people to pick up... And the go.work has one new directive as the use directive. So you tell it which directories you want it to use, and all the modules in those directories, if you're under the go.work files same way you're under a go.mod file before, are in your workspace. So you make your go.work file, and CD under it, and now you're using all those modules in any build that you do, or go list, or any such command like that. + +**Mat Ryer:** And it's gonna be aware of those other modules, and you're not gonna be fighting the tools anymore. + +**Michael Matloob:** Yup. + +**Mat Ryer:** And can you do replaces as well in there? + +**Michael Matloob:** You can... We don't think people should need to add replaces, except in very specific circumstances. We actually added replace because if you have multiple modules in your workspaces, they might have conflicting replaces... So replacing the go.work file can override the replaces that are fighting. But if you wanted to use a specific module, then usually the right thing to do is to just use that module in your go.work file. + +**Mat Ryer:** Yeah. So I feel like lots of tools probably got touched by adding this kind of support. Was it a difficult one to get in? + +**Michael Matloob:** Most of the work was in the Go command itself. There's definitely an amount of complexity in the Go command, because their module-loading code - it does more than you'd think. But once we got it to work in the Go command, one of the nice things -- most of our tools call into the Go command. So as long as they're making a call into the Go packages or the Go command, they kind of get all that for free, as long as the go.work file exists on disk. + +We've had to make on the VS Code Go and Gopls teams -- they've made changes to understand go.work files and pass them into the Go command, but... You know, once you pass that in, all the hard work is done by the Go command itself. + +**Mat Ryer:** This is a really nice thing to have. In the way that I work, this is gonna change a lot. There was an experiment for a while where you could try this out with an environment variable, wasn't there? + +**Michael Matloob:** Yeah. I filed a proposal for this, and I made it available, and we had a link for people to easily download a development version that included these changes using the gotip command, and so people could try it in and give feedback. We got some feedback on it, which was super-helpful. And we got a lot of feedback on the issue too, which was very helpful in developing the proposal. Hopefully, we've addressed most of the important issues people have. For anything else, there's 1.19. + +**Mat Ryer:** \[44:27\] Absolutely. Well, honestly, I think these kinds of things make such a difference, so we're so pleased. And honestly, I feel like there's a lot more to talk about with workspaces and modules. Maybe, Michael, you could come back one day and we'll do like a modules and workspaces special. + +**Michael Matloob:** Oh, sure. I'd love to. + +**Mat Ryer:** Okay, cool. Well, we will hold you to that. I do consider this to be legally binding. + +**Michael Matloob:** Like the pinky promises? + +**Mat Ryer:** Pinky promises are, of all the types of promise, I think, they're up there with the most important. You've got local laws, you've got national laws, and then all the way at the top, after the Supreme Court, you've got the little pinky promise there. + +**Michael Matloob:** Pinky promise at the top. Got it. + +**Mat Ryer:** Yeah, I think that's the legal structure of the pinky promise. Well, here's another pinky promise - I promise you're about to hear some unpopular opinions. + +**Jingle:** \[45:25\] to \[45:42\] + +**Mat Ryer:** Okay, who's gonna go first with -- I don't know why I'm speaking in this spooky voice. Who wants to say the first unpopular opinion? + +**Daniel Martí:** Maybe I can start with mine... Mine is that I think code generation should be avoided whenever possible. + +**Mat Ryer:** \[laughs\] + +**Daniel Martí:** I think the main reason for that is because it adds developer friction, it often increases build size and build time, and oftentimes people overestimate how slow reflection is. If you use reflection well, the cost is actually very reasonable. And it's not like you're building your whole program around reflection. You're using it in very careful ways, in small places. + +**Mat Ryer:** What do you think of that, Michael? + +**Michael Matloob:** I guess I don't have a very strong opinion about this either way... + +**Mat Ryer:** I do. + +**Michael Matloob:** Yes? Let's hear your opinion. Is it the popular or unpopular variety? + +**Mat Ryer:** The thing is, I love code gen, because it's like you're doing loads and loads of typing. You just do a bit of typing, and you run a command, and it's just like you've done loads of typing. So that's the thing... Reflection is hard, so it's quite satisfying when you get it right, but editing a template and then running a thing and having 1,200 methods update... You've fixed 1,200 bugs at the same time... \[laughs\] What do you think, of that, Michael? + +**Michael Matloob:** I will say this - I find that working with code generation when using the Go command is not very fun. I don't like using go generate, I don't think it has a good user experience... It happens separate from the build, so it's really easy to have stale files. + +You know, I feel like this ship has sailed, but if you're gonna do a lot of code generation, Basil is very nice for that, but it's not very heavily used in the Go community. Inside of Google we use mostly generated prodos... And it's seamless, because the build just generates some automatically and you don't need to think about them. The tools that take care of all of the annoyances that are caused by code generation. But our tools don't really do that, so there's like a lot of friction when using generated code outside of those build systems. So I get pretty annoyed. If I have to run a make before my go build, I feel like there's a problem. That's kind of answering a different question, but you know... + +**Mat Ryer:** \[48:13\] I think generics -- ugh, I've said it. Yeah, I'll get booted out of here... But I think this rule obviously is very weakly enforced. So much for pinky promises, right? After I was bigging them up and giving them all that legal weight, look at it now. It's been reduced to a silly, childish thing. How sad... + +Well, what I was saying is, I think generics are gonna get rid of a lot of cases for code generation... But reflection is pretty difficult to write, because there's no kind of feedback -- like, you need unit tests really for your feedback to... I mean, you don't really need that; let me rephrase that. I think reflection is quite hard to get right, because it's that sort of meta-programming. But then code generation templates are also meta-programming, and they are often quite difficult to look after and maintain. So maybe you've got some legs. + +I'd be interested to test this one on Twitter @GoTimeFM. We will tweet out a poll and find out if this really is unpopular. It's a candidate for one though, it's a good one. Can you beat him, Michael, is the question? + +**Michael Matloob:** My unpopular opinion is we should bring back the try proposal. + +**Mat Ryer:** Oh, really? + +**Michael Matloob:** And this is where I'm going to not mention other features by name, but I'll say - of all the features that people have proposed as language changes to the Go language, I feel like none have been as potentially impactful as the try proposal was... And I was sad to see it pulled back. Because I think error handling properly is really important to writing good Go code, and I think the language ergonomics should encourage people to handle their errors properly. And so often people will just -- "if err!= nil {return err;}" and just not think about what they're doing with their errors... And I feel like try gave an opportunity to think a little bit harder about wrapping errors properly, and what to do with errors, and kind of nudged people to do the right thing a little bit more. + +Certainly, the proposal as it was needed more work before it should go in, but I really do think we should bring back the try proposal, and keep working on it and make it better. I don't know when we'll have the bandwidth for another big language change like that, but... + +**Daniel Martí:** I have to agree with Michael. I think the reason the try proposal got so much bad feedback is because Go is so opinionated, a lot of its users have gone into this mentality of Go doesn't need features, so sometimes the users can have this knee-jerk reaction of somebody proposes a change to the language and they go "Well, but that wouldn't be Go, right?" And I agree with Michael in the case of try, I think it would have been a very interesting change, and I hope it comes back. + +**Mat Ryer:** So just refresh our memories. What did try do? + +**Michael Matloob:** Yeah, so basically it gave you a mechanism to try with a function that returned an error as its final argument. And then it would allow you to handle that error elsewhere. So you could add -- I think in one of the variations of the proposal there was like a handle for handling a number of tries in a function. I think in another one, if I'm remembering correctly, recover was an option for handling the error... But you could kind of have the errors handled in a single place. Basically, people realize that error handling is awkward in Go, and the awkwardness I think causes people to take shortcuts. + +\[52:10\] So addressing that awkwardness and nudging people towards doing the right thing, especially if try and handle came with helpers... And now we do have functions like `ers`, `is`, and `as` that help people with wrapping errors. Those together would provide a better model for handling errors and for people to think about handling errors. + +**Mat Ryer:** Well, fascinating stuff there... Yeah, I'll tell you what - it's interesting, because I think... See, when I handle errors - and I don't know if I do this different to other people. I may be unusual in this... But when I return the error, I add quite a bit of context there. So each one is different; I'll add and I'll include the thing it's trying to do in that wrapped error. So it's not that I'll have a wrap where I just put the method name into the error, and it's the same every time. So if it was the same every time, having it pulled out and deal with it in one place is quite nice. + +But yeah, the other thing is -- I mean, this doesn't hurt that, but I do like that error handling is at least explicit. The try proposal didn't really interfere with that, but I like the fact that in Go we are kind of handling errors, even if you are just returning it. It's like, as long as you're not forgetting about it, it's nice that they are in the forefront of our minds when we're coding. + +I literally was writing something today, and I literally had to write if err!= nil, and then I had to stop and think "Oh, what do I do if this errors like that?" That actually is not a trivial problem in this particular case. I wasn't able to just return an error. I had to handle that... So I quite like that it's in the forefront, at least of the language, but... Yeah, interesting. We'll definitely find out what other people think on Twitter when we post that one. It's gonna be very interesting, I think. + +**Michael Matloob:** Yeah. I'm interested in seeing how unpopular that is. Do you think that could go for the most unpopular opinion expressed in one of these segments? + +**Mat Ryer:** It could. Or it might surprise us and maybe everyone's like "Yeah, we loved that." Honestly, I think that point of like, we're now so familiar with Go, we have to be careful we don't just become curmudgeons about it and resist any change... Because it should change, it should evolve, it should get better, like all software. Yeah, I'm kind of interested to hear what people think about it. + +**Daniel Martí:** I think Michael also needs to think that if this opinion is gonna be really unpopular, then try is not gonna come back... So you want it to be very popular. + +**Michael Matloob:** I mean, I would like for it to be popular... + +**Mat Ryer:** Yeah. That's not really in the spirit of the segment, but... It's fine. + +**Michael Matloob:** I think it is unpopular, but... + +**Mat Ryer:** Well, we'll find out. + +**Michael Matloob:** ...if it's an opinion that I hold, I would like it to be less unpopular, even though it is unpopular. + +**Mat Ryer:** \[55:07\] Sometimes, when the case is made, in fact it's hard to get unpopular opinions. This is what we've found. Because people make the case so eloquently, like you did. And then people on Twitter... You know, they're easily swayed. They'll believe that now. + +**Michael Matloob:** I mean, if I'm gonna make reference to the g-word again, there was a time in the community where you brought up the g-word, and people are like "No! Not in my Go!" And people were right to be worried about those things. But I think the case was made, people worked really hard to present the case why it would actually be an improvement and really convince people; convince people who use Go that it was actually gonna be a net positive. I think the sentiment now towards the g-word is a lot more positive than it was 5-6 years ago. + +**Mat Ryer:** Yeah. That is definitely true. And then the counter is we don't want it to be too easy to change things because of the backwards-compatibility promise. I do quite like the fact that it's quite a rigorous process before we really get any big changes like this. I think there's value in that, too. So that really only good stuff is gonna get through, hopefully. + +**Michael Matloob:** Every new thing we add is something we have to maintain forever. We do have to be careful about adding new things, for sure. + +**Mat Ryer:** Yeah, because forever is ages, ain't it? + +**Michael Matloob:** It's a pretty long time. + +**Mat Ryer:** Yeah. + +**Daniel Martí:** I also remember somebody recently criticizing Go - I think it was on Hacker News - saying something along the lines of "Go is a popular language that has ignored all the programming language development in the past 15 years." But that's kind of why it works, right? Because it only builds on top of what has been well-tested. And I think the only major exception there is modules, which goes against everything else that has been done in package management. + +**Mat Ryer:** Yeah, that's quite interesting. It is a very stable thing. And yeah, that other thing of having lots of different ways to do the same thing... In JavaScript, you almost have to learn a particular flavor of JavaScript now in order to contribute to a project. Someone using all the latest language features, like the little arrows for functions and things like this... You sort of have to learn all that. + +**Michael Matloob:** Yeah... I mean, I find that sad from the other side. JavaScript and all the other parts of the web ecosystem have become so big that it is impossible for anyone to make a new JavaScript interpreter engine without the resources of a huge multi-national corporation. That sucks. + +**Mat Ryer:** Well, I'm afraid that's all the time we have, on that somber note... You can forget about your dreams of writing your own JavaScript engine. It's not gonna happen... \[laughter\] So just wake up. + +**Michael Matloob:** Sorry... + +**Mat Ryer:** No, it's a harsh wake-up call, Michael, but we needed it. Thank you very much. \[laughter\] Thank you so much to our guests today... Michael Matloob joined us, as did Daniel Martí. It was a pleasure, as always. Thank you very much. Thanks for joining us on Go Time. + +**Daniel Martí:** Thanks! + +**Mat Ryer:** We'll see you next time. diff --git a/The pain of dependency management_transcript.txt b/The pain of dependency management_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..974a2487b5bd96c654c57042aada8546a7656c97 --- /dev/null +++ b/The pain of dependency management_transcript.txt @@ -0,0 +1,231 @@ +**Natalie Pistunovich:** "Good time of the day" is my new favorite greeting, to anybody who is listening any time of the day. Today we are here to talk about dependency managers, and I am here with my co-host, Johnny. Hi, Johnny. + +**Johnny Boursiquot:** Hello. How are you? + +**Natalie Pistunovich:** Good. How are you? + +**Johnny Boursiquot:** I am good, but I am sure I am not as tired as you must be... You just got done finishing hosting GopherCon EU. How did it go this year? + +**Natalie Pistunovich:** Wonderful. + +**Johnny Boursiquot:** Nice. + +**Natalie Pistunovich:** But I am tired, you're right. \[laughter\] And we are here joined today by Baruch Sadogursky. Hi, Baruch. How are you doing? + +**Baruch Sadogursky:** Hello! I am very excited to be on this podcast, because -- well, because it's an awesome podcast, with awesome hosts... And I'm a little bit bummed I missed GopherCon EU this year, but we already set up a date to be there next year, I hope. + +**Natalie Pistunovich:** Absolutely. + +**Baruch Sadogursky:** So yeah, thank you for having me. That's exciting. + +**Natalie Pistunovich:** For anybody who's listening later, or just not watching the recording, Baruch wears a really cool T-shirt that says Yalla. + +**Baruch Sadogursky:** Yalla. + +**Johnny Boursiquot:** What does that mean? + +**Natalie Pistunovich:** For those who don't know, for this probably small part of the population, but - for those who don't know, what is Yalla? + +**Baruch Sadogursky:** So the meaning of the word is "Let's go", because Go Time, Go Podcast, and everything... \[laughter\] + +**Johnny Boursiquot:** Nice. + +**Baruch Sadogursky:** But also, it's the name of our conference, the conference that JFrog organizes; a community DevOps conference in Israel. + +**Natalie Pistunovich:** So Yalla is the name of the conference? + +**Baruch Sadogursky:** Yeah, Yalla DevOps, actually. + +**Natalie Pistunovich:** \[04:12\] Yalla DevOps. + +**Baruch Sadogursky:** Mm-hm. + +**Johnny Boursiquot:** Cool. + +**Baruch Sadogursky:** Which is a very cool name. + +**Natalie Pistunovich:** Now we are here to talk about dependency managers. + +**Baruch Sadogursky:** Yes...! + +**Natalie Pistunovich:** And before we jump into that topic, Baruch, would you like to introduce yourself to the crowd? + +**Baruch Sadogursky:** Yes! So I hate dependency managers with passion, for the last (I would say) at least 15 years. That's kind of the relevant context. I've been a developer advocate with JFrog for (I think) 11 years by now, before that working as a Java developer, senior developer, or whatever it was... Architect... And yeah, so we started in a previous company, in the consultancy company, we were mostly doing the build CI/CD -- there was no CD back then, but building CI parts of a lot of projects, trying new techniques with builds and dependency managers to make it reliable. One of those tools was what later became JFrog Artifactory, and this is how we got into dependency managers in general, and JFrog in particular, and the rest is history. + +My dream is to eventually sit down and write a talk about how terrible every dependency manager on the surface of Earth is. + +**Johnny Boursiquot:** Okay... \[laughs\] + +**Natalie Pistunovich:** I'm looking forward to that talk. Please tell me what conference, I will join. + +**Baruch Sadogursky:** Yup, yup. That'll be everywhere. Once I write it, I will go and preach about how terrible they are to everybody. The most annoying thing is there's no solution, really... And we can talk about that, why it is the issue, we can talk about later. And by the way, Johnny, thank you for wearing this T-shirt. I love it. + +**Johnny Boursiquot:** Yeah, that's actually a JFrog -- + +**Baruch Sadogursky:** Yeah, it's the \[unintelligible 00:05:59.14\] + +**Johnny Boursiquot:** It's completely coincidental. That was not planned at all. \[laughs\] + +**Baruch Sadogursky:** Um, okay... + +**Johnny Boursiquot:** It really wasn't, yeah. + +**Natalie Pistunovich:** It's just because the T-shirts are indeed cool, every year. + +**Baruch Sadogursky:** It's true, yeah. + +**Johnny Boursiquot:** I am interested in -- obviously, this is a space you're very familiar with... I come from basically a long history of using various dependency managers for various languages... And yeah, none of them are really perfect. Nor do I ever expect them to be, but each one has different pain points, each one has things they do well, and some things they could certainly improve... But I'm curious, because you have sort of that broad view of a lot of them, what have you seen that is the common thing that they all don't get quite right? + +**Baruch Sadogursky:** Right. I will quote Sam Boyer a lot. You obviously know Sam; I hope you hosted him on your podcast, and if you didn't, you should. + +**Johnny Boursiquot:** Oh, we have. + +**Baruch Sadogursky:** And if you did, you should do it again. + +**Johnny Boursiquot:** \[laughs\] + +**Baruch Sadogursky:** Sam is amazing... And basically, he summarized all the problems with dependency management perfectly when he said that dependency management looks like a computing problem, and we try to solve it as a computing problem, but it's actually a people problem. The biggest problem with dependency management - there is no way to solve it just using the math, the algorithms, the computer science, because it always clashes with how humans behave. Your expectations, what do you do, how do you do it. And by that, the simplest example will be - you know, stuff like backwards compatibility, binary compatibility... And you say "Well, that's a solved problem. We have semantic versioning that solves it. It has guarantees which levels of versioning should work with which others", and you say "Well, it's a major version, it will break backwards compatibility, but if it's a patch, it shouldn't." And then you code it in a way that your dependency manager will know how to look for a new version, or not look for a new version etc. but then it actually breaks very quickly when those guarantees are not fulfilled by the people who apply the versioning to their tools. And this will always be something that is human-driven. People will decide which compatibility their software breaks or not, and they will make mistakes. But when you look at the algorithm side of things, it doesn't know about those mistakes, right? + +\[08:43\] So you say "Well, I can always rely on the fact that a patch-level upgrade will work great because of backwards-compatibility", but it's not, because someone made a mistake. There is no way to promise binary compatibility without relying on people's opinions on whether their software is compatible or not. Does it make sense? + +**Johnny Boursiquot:** Yeah, that totally makes sense. I can certainly see, for those who are not super-familiar with semantic versioning and how versioning is in theory supposed to happen - basically, you're not supposed to be changing existing APIs... Imagine a particular method that does something - you're not supposed to all of a sudden change the signature of the function, or the method, to accept new parameters or change existing parameters and then call it a patch. That's a breaking change. You are gonna break other people's builds who are relying on the dependency. + +So we're relying on individuals sort of doing the right thing, but obviously, for popular projects, even though it happens, these tend to be sort of mistakes... So really what we're saying is human mistakes. This is sort of a grey area where algorithms can't really help you there, because it's a human making a decision, saying "Okay, I'm gonna release this as a patch", but for some reason maybe they forgot or maybe they were not communicating with somebody else on the team, or whatever it may be, and they accidentally released a breaking change as a patch, when it was supposed to be a minor version bump, or maybe a major version upgrade, or whatever the case may be. + +My question to that is "Is it really sort of a completely grey area that cannot be solved using an algorithm or using computers to solve the problem?" Do we not have tooling for that, or maybe we just haven't put in the time yet into solving to prevent these kinds of problems. It sounds like we could create some tooling to say "Hey, have function signatures change", for example. That way you have to have a different kind of a bump. + +**Baruch Sadogursky:** Right. But that's the easy challenge, right? If you say like "Just check that the signature is the same and the parameters didn't change" and whatnot, that's obviously doable. But how about the behavior? So at the end of the day, you made the patch, and you made it for a reason, because you want the behavior to be different. But the question is if that's the right kind of different. Maybe there is a bug. Maybe you were supposed to fix the bug, but instead you introduced an even larger bug, and now it does something terrible. + +In terms of the binary compatibility per se, your build won't break, because the parameters are the same, and everything -- and unless you test for this behavior, you won't even know. And then you release to production, and then you discover it, and now you need to roll back. Now, roll back - it's another hell of its own, because now the behavior should change backwards, and your software should actually be forward-compatible in order to be able to work with part of the system will be the new version and part will be in the older version. + +So it's really a problem that has no solution, because it actually depends on people always doing the right thing, which is not what's happening. + +But this is only one side of the problem. I can tell you another one, which is also very human-driven... And that's the external dependencies. How do you get them where they are? How do you trust those who produce those dependencies etc? + +\[12:24\] Another quote -- and I don't remember whom... I think it was Brad Fitzpatrick that said that when you use external dependencies, it's like you just take someone else's random code from the internet and just dump it into your production systems. I think it was Brad. And he's on point on that. His solution was terrible, frankly, if you ask me. Before Go had Go Modules, the idea was "Hey, take their source and incorporate it in your source, and then consider it your source", which is very, I would say, kind of an embedded \[unintelligible 00:12:57.24\] solution when you really do that, and then you provide one binary... But for large-scale systems that 80% to 90% are built with third-party dependences, this is not a scalable solution in any way or form. + +So he pinpointed the problem correctly. It is a terrible idea to just grab other people's dependencies and throw them into your production system... But what is the answer? And we as an industry try to come up with the solution for many years now, and Go Modules, the central repository, the signature server of the central repository - the earlier version of this vision when it was the idea of multiple central repositories that need to negotiate between them and establish some kind of a trust... Because if we have the same package, under the same name, come from different sources and their checksum is different, that means that someone is lying here, and we need to find out who. + +This is another huge challenge, and especially today, when we hear all those stories about how the supply chains of those dependencies get hacked. SolarWinds for one example, and multiple others. That's obviously the problem which surfaced lately, but been with us for as long as we tried and used dependency managers. + +**Break:** \[14:20\] + +**Johnny Boursiquot:** So what would you say has been -- like, I have some idea of how you'd go about trying to not solve, but really mitigate... But I wanna hear what you've seen in your experience of how people sort of mitigate that second class of problems. + +**Baruch Sadogursky:** There were and still are a lot of efforts in trying to do that... And it looks like - this is from our experience as well; we tried to come up with -- again, back a couple of years ago when Go Modules only appeared, we heard this vision of "Hey, we will have some kind of a network of central module registries that will communicate with each other, trying to establish whatever truth really is, and make sure that the rest of them are not allowed", and everything. This is where we actually started Go center. That was one of those registries. But with time, in a matter of the next years, somehow this vision morphed to "Hey, Google will maintain this central module repository, and whatever Google says is the right thing, is the right thing." So this is where the usability of Go Modules kind of disappeared, because there was no more alternatives to whatever golden modules Google will give you. + +But then it was a very useful experience for us to really try and evolve this idea of negotiating what's \[unintelligible 00:17:34.25\] And I would say there are now two most important aspects of this system. The first is how do we know which module is the right module, in terms of "Hey, I have two of them, they're named the same, they have the same version, but their checksum is different. Which one of them is lying, and which one of them is authentic?" That's the first one. And the second is chain of custody of those modules that we decided they are the right ones. So I decided this module is authentic, and we can talk in a second why... And then, okay, how do I make sure that what I have now in my machine is really the same module, and the papertrail of the decision that it was the authentic module is actually preserved and cannot be hacked? + +\[18:30\] That brought us to envisioning a system that everybody will be able to use in order to guarantee those two, and it's called Pyrsia. Actually, Pyrsia - that's an interesting story of the name... It's actually a system that ancient Greeks used to convey messages with flames between each other. So this is why you have the torch. + +The idea is a decentralized P2P network that will provide a consensus of what modules are authentic, and then make sure that they cannot be tampered with, with some kind of blockchain-backed ledger. I know that we lost two thirds of the audience right now, after I mentioned blockchain, but whoever stayed - you did the right thing; it has nothing to do with cryptocurrencies or Web 3.0. It's just a ledger that you cannot tamper with; that's the interesting part of it over there. + +So basically, what we say is "Okay, everybody now can build modules from source, and publish them wherever they like." The system will communicate and decide if that's the golden build by comparing it to the build that the original producer of this module creates. Let's say I have a library now that I wanna publish, and I create a build that because it's my library, I know is the right build. Now, everybody can have their own CI producing the same module from my sources, because my sources are obviously open source; as long as the build produces the same result, your end result is as good as mine. + +Now, we record it, so no one can hack into that and pretend that another result is the right one, and then we distribute it through a P2P network, because now we don't care where it comes from, as long as we can guarantee that it's the right results. So now Natalie builds it on her machine, and she's behind a firewall and the connection to whatever central repository \[unintelligible 00:20:36.12\] but you see the next door, in the same intranet with her, in the same company, and as long as we verify that the build that Natalie produced is as good as this golden one that was produced by the original maintainer of the library, you can use whatever she built without downloading it again, outside of the firewall, and getting tons of approvals and whatnot. + +So this looks like, from our perspective, the solution to some of the supply chain issues, because it guarantees the authenticity, and it protects the supply chain itself, because we verify on your machine that the package you have is exactly the package you need to have, regardless of where it actually came from. And you also have this network of the Go registry, npm registry, JFrog Artifactory, Maven Central, you name it, the sources that provide infrastructure for massive distribution and scale... And again, we don't need to trust them in terms of "Are they giving us the right packages?" because we can verify ourselves. And also, in case they are down, we can also rely on the P2P network to get the modules in case they are not available. + +**Johnny Boursiquot:** \[21:59\] Does this approach require those who create those libraries or dependencies, that they basically follow a similar set of principles? I'm thinking of the recently made public Software Bill of Materials. Stuff that came out within the last year, for example... Basically, that requires a specific set of things be present in order to know where did they come from, what's the version of each component, give me some identifies, give me some things that I can check against. Does such a distributed peer-to-peer system require adherence to something like this? + +**Baruch Sadogursky:** Yeah, so this is the question of how we can verify that these golden module is really golden, is really good. And this can be achieved in multiple ways. First of all, we can say hey, let's say the Go community trusts the Google registry to a reasonable extent. And that means that if I produce a module for the Go community, I can delegate to the Google registry to build my module and declare it golden, because I know that the Google Go Module Registry guarantees all those things, and will make sure that my module is good. Since that point, it's all P2P, with untampered ledger, and we go to that. I can say "You know what? I don't trust anyone. I want to build it myself. The only binary that I'm ready to distribute or call it golden is the one that I built for myself." That's fine. Here you have, let's say, a Pyrsia client - that will take care of all the needed attestations, documentation, bill of materials etc, and if you don't have them, then your build cannot be certified as these golden build that will be distributed to everybody, or reproduced by whoever we want it to reproduce, as long as the checksums of both the data and metadata match. Does it make sense? + +**Johnny Boursiquot:** Yeah, it makes sense. + +**Baruch Sadogursky:** Right. So this is kind of our vision, and obviously, it's just a patchwork of the problems that arose in the last years, and the solutions that we saw working or not working. I can tell you an example... So Maven Central, for example, has been struggling with the problem of this authenticity forever, because they existed from 2009. And they're like, "Okay, how do we guarantee that whoever publishes this artifact to Maven Central can actually do that?" And they say "You know what - we're going to do PGP keys." So I need to have a key, that I sign, and uploading the public key with my artifacts, so whoever downloads it can download the key and verify that I am the one that actually created it... Which sounds like a good idea, but you need to remember that the PGP idea was created to establish a trusted circle of people. It was created for email, based on the assumption that everybody knows everybody else through someone else in whatever theory of five handshakes, or whatever. + +So basically, if I send an email to you, and that's our first correspondence, you don't know who am I, but Natalie knows who am I. She can propagate this knowledge to you. Now, it works in email, but for distributing binaries throughout the world, you have no idea who am I. + +\[25:42\] So you downloaded my PGP key. It says "Baruch Sadogursky", or it might say "Brad Fitzpatrick." How do you know that I'm really whoever I claim I am? I can generate a pair of keys for any identity that I want. And if the only requirement of the central repository is to provide them with this key, I can be whoever I want. And the funny thing is if you use the default PGP tool for generating keys for Ubuntu - at least that is what it used to be - and then you just do Next, Next, Next, Enter, Enter, Enter, the defaults are... I don't remember whom -- some famous German, because the PGP key was created in some German university... I don't remember. Maybe a writer. But anyway, it had some ridiculous default. Mozart. I think it's Mozart. Maybe it's Mozart. Not German, but... I don't remember. Someone. But then you just take it and plug it into the search of Maven Central Repository, and you find hundreds of modules created by this person. Because most of the people, when they do tests, they just do Next, Next, Next, and they generate a signature for this person, and just upload it with their packages to Maven Central. It really doesn't verify the authenticity. And it's obviously not enough. + +On the other side, you don't really care. If I created a module, you don't care what my name is. And even if I'll tell you my real name, it probably won't help, because you don't know me. What you need is someone who you trust to vouch for this module. And this is where when we have those golden registries, or golden CI/CD pipelines that can guarantee that they took it from the right source, they have all the bill of materials, everything is safe, and then this is the outcome, this is the checksum of the data and the metadata, this is everything you need. And then you can decide whether you want to take this module from them, whether you want to build it from sources on your machine, and just compare that you have the same outcome... This is all you need. You really don't care what's the name of whoever built it, and this is why the PGP keys are not helping. + +**Johnny Boursiquot:** Okay. So in this new world, if I'm building software, that first decision to use a dependency - I'm still making that choice, right? So if I go on GitHub, Bitbucket, what have you, and I'm looking for a library that does X, then that job of identifying a library and deciding to trust it - maybe I read through the source code, making sure it's doing what it says it's doing - that first job is not obviated by this new system. + +**Baruch Sadogursky:** And here we are back to the problems that cannot be solved by computer science. And there are a bunch of them. We're just getting started. We work on that set of problems which is solvable. Those are the low-hanging fruits. How about the problem that the project is hijacked? You'll remember probably the story - there was a popular project that the maintainer kind of stepped away and let their domain expire. And then someone purchased the domain, tied it to their email address, obviously created a working email, restored the password for whatever npm registry, and was able to upload a new version with malicious code. There is no solution. Literally, there is no solution. It's a people problem that we -- if we declare that we can solve it with computer science, we lie to ourselves. And this is one of the reasons why all the dependency managers suck, because every developer of dependency management promises themselves and everybody else that they are going to solve those problems using computer science... And it's impossible. + +So whoever tells you they're going to solve the problem of supply chain security, spit them in the face. They are lying. + +**Break:** \[29:57\] + +**Natalie Pistunovich:** So a question from the crowd, from the GoTime Slack channel. Somebody's asking whether binary compatibility decisions are different among organizations. And Louis St. Martin, who's writing this, is saying that at work they have this situation that they know what is a good code, and reference the code, and so on, so they don't reinvent those patterns every time... But for things like Terraform - he is not familiar with any. So would something like an RFC help to things like binary compatibility decisions and so on? + +**Baruch Sadogursky:** I'm not sure I understood, frankly... + +**Natalie Pistunovich:** If there is something like good usage patterns, and good style, and recommended whatever RFC equivalent would be for things like binary compatibility decisions specifically. + +**Baruch Sadogursky:** Yeah, so I think -- again, if we look for standards, something codified, I would say that semantic versioning is actually a very, very good system to guarantee backwards compatibility. But again, with a twist, obviously, and the twist is humans. + +The idea is that "Hey, we can rely on whatever version of patching, patch version to be binary compatible, and minor version be public API compatible, and then minor not compatible to anything, and that actually should work for everything, until we hit the problem that "Hey, someone did incompatible things and called it a patch." And for that, frankly, I'm not aware of any possible solution at all. Unless I'm missing something obvious in the last 15 years, I don't see any way on how you can solve it. + +How you can prevent humans from breaking the guarantees that we build our machine systems upon? That episode - that's a really sad episode. Sorry about that. + +**Johnny Boursiquot:** \[laughs\] + +**Natalie Pistunovich:** Maybe it will inspire somebody for something. + +**Johnny Boursiquot:** It's like everything -- it sucks. \[laughs\] + +**Natalie Pistunovich:** Yeah. Social engineering, but for packages. + +**Baruch Sadogursky:** I mean, social engineering is malicious, but even honest mistakes. Let's not go and assume the worst... But even if we assume the best in people, people make mistakes, and there is no easy way to catch them always, and this is why we have bugs; and that's fine. But we have a system that envisions no bugs, and that cannot work. + +**Natalie Pistunovich:** \[34:14\] So another point that came up on Slack by Henry Snowpack is in the context of npm registry compromise. So Henry is saying that vendoring could solve that, to a point... + +**Baruch Sadogursky:** Yup. Yup. + +**Natalie Pistunovich:** ...if you require review checks before the dependency is being updated. Or I guess it's a question... + +**Baruch Sadogursky:** So two things here. First of all, let's talk about vendoring. Vendoring is the worst type of working. You take someone else's code, you detach it from their version control, you dump it into your version control, and you're all for the races. You diverge from the original development; bringing that back will be a terrible pain, and you basically treat it as your own source code, when you cannot even determine what's your source code and what's not, in a machine reasonable way. + +The benefit of vendoring - and this is why Go relied on vendoring for a decade - is because it kind of ignores that supply chain problem. As a matter of fact, you're saying "I don't use any third-party dependencies at all. All the code is mine. I take full responsibility of whatever I'm using, whether I brought it through vendoring or wrote it myself." It's nice because it lets you ignore the supply chain problem, but it's terrible in any other possible way. + +So yes, it solves this particular problem, but the solution is usually more problems than actually benefits. And the fact that Go switched to Go Modules from vendoring is, I think, kind of -- it attests to the fact that vendoring is not a scalable solution. + +Vendoring is terrible in some ways, probably not the right solution for the supply chain problem, and we need to look for other solutions. + +**Johnny Boursiquot:** I would say my counterpoint to that would be if what I care about -- if I have sort of a set of things that are important to me, for my builds, reproducibility of my builds, making sure I have the right version, that I've vetted, that I know work with my stuff, I'm not worried about a patch being compatible. So I'd say this particular version, this particular commit, this particular semantic version I know for a fact works to produce my piece of software, that's working in my business, that's supposed to be making money... If my primary concern is to make sure that I'm always able to reproduce that build with that same exact version of the code, then perhaps I don't care about the other factors. Perhaps vendoring works just fine, because that is the first and foremost thing that I care about. + +**Baruch Sadogursky:** The good news are you definitely don't need vendoring for that. Like, at all. Any modern package management today will give you a full reproducible build by using dependency managers, because you have multiple layers of caching that you can lock what dependencies you use for this particular build, without actually declaring third-party code as yours. And that's true for Go Modules, obviously... It actually uses the directory formerly used for vendoring as a cache for modules now, and you have your own level of caching on your intranet, on your company or team level, being JFrog Artifactory or being the local Go registry in this way or another... + +\[37:53\] So you definitely can get repeatability, you definitely can reproduce the builds, you definitely can lock your versions that you are interested in after your vetting. Using modules or any dependency manager doesn't mean getting out there and grabbing the latest, the freshest dependency and dumping it in your build. No. It actually means doing the upgrades in the most convenient way on your terms. And your terms should differ based on your risk sensitivity in this particular scenario. + +If you just have a script that does something for testing Go, or whatever, you can say "You know what - I don't care. I can download the latest version and just try to run with it." Worst-case, it just fails. + +If you have a sensitive build - security, financials, you name it - then obviously it's a different game. You have a closed system that has to produce a reproducible build every time. The upgrades will be tested in multiple scenarios on different levels of testing, because before they will be allowed to be a part of your production system, and this is fine as well. But when you use a modular build system like Go Modules, you can automate that, and build a pipeline that will take a new version not throwing it into production immediately in the next build, but getting it into this pipeline of verifications that can be completely automated, manual, half and half, whatever works for you, and only that propagated to be an official dependency, so you get all the benefits of vendoring without the downsides of declaring every third-party source as yours. + +**Natalie Pistunovich:** Henry said "Thank you for your response, Baruch", and I am proposing we will jump to the fun part of the show, of an unpopular opinion. + +**Jingle:** \[39:59\] to \[40:18\] + +**Natalie Pistunovich:** So I wonder if most of the thoughts you were sharing, Baruch, will be ending up popular and unpopular, but I also wonder if you have an unpopular opinion for us, that is on any topic you want. It could be dependency managers, it could be anything else. + +**Baruch Sadogursky:** Yeah, we spoke for the last 45 minutes about one huge unpopular opinion... I will just summarize it to put it up to the vote. Dependency management was sent to us from hell to make our lives miserable. + +**Johnny Boursiquot:** \[laughs\] + +**Natalie Pistunovich:** Alright, alright, so any unpopular opinion goes through the vote on the Twitters. So we will soon find out what does the crowd think. + +**Baruch Sadogursky:** Okay, I'm going to turn on my Twitter bot farm right now. Because of my Twitter bot farm, that's why Elon Musk didn't buy Twitter. \[laughter\] Because they did manage to find it, and they know there are a lot of bots them, they know who owns them, and that's me. + +**Natalie Pistunovich:** And you operate them all with DevOps. + +**Baruch Sadogursky:** Of course, of course. + +**Johnny Boursiquot:** CI/CD pipelining for your bots. + +**Baruch Sadogursky:** Yeah, yeah, built with another package manager that was sent to us from hell to make us all miserable. Yup. + +**Natalie Pistunovich:** Alright. + +**Johnny Boursiquot:** Cool. + +**Natalie Pistunovich:** I will say thank you very much to everybody who participated on Slack. It's fun when the crowd participates. And I will say, Baruch, a very big thank you for joining and sharing your thoughts. That is definitely interesting to hear. Johnny, thank you for asking practically all the questions in this episode. You rock. + +**Johnny Boursiquot:** \[laughs\] Yeah, I was definitely interested in some fresh thinking, and definitely called out a few of the pain points that I've definitely felt all across my career... And it's interesting to know that there's some new thinking being applied to these problems. It doesn't solve all of our problems, there's some human aspects, as you mentioned, that are sort of unavoidable... Things like how the software bill of materials is leveraged, how we can use a P2P network for authenticity and verification, and things like that... These are really some new things that are gonna have hopefully some positive impact on this dependency problem from hell, as you say. \[laughs\] + +**Baruch Sadogursky:** I really hope so. We are learning and getting better step by step, but yeah, what we need to keep in mind - we cannot solve everything with computer science, because... Humans. + +**Johnny Boursiquot:** Right. + +**Natalie Pistunovich:** \[laughs\] Thank you everyone again for joining. diff --git "a/The \357\274\212other\357\274\212 features in Go 1.18_transcript.txt" "b/The \357\274\212other\357\274\212 features in Go 1.18_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..970e05580fc1274e43458f2cba6288347b6c5189 --- /dev/null +++ "b/The \357\274\212other\357\274\212 features in Go 1.18_transcript.txt" @@ -0,0 +1,1057 @@ +[0.00 --> 3.66] We have go.mod files, and now we have go.work files. +[3.98 --> 5.50] So you create a go.work file. +[5.62 --> 8.00] The syntax is very similar to go.mod. +[8.08 --> 9.96] We want it to be easy for people to pick up. +[10.36 --> 13.22] And the go.work has one new directive, +[13.78 --> 14.92] is the use directive. +[15.32 --> 18.76] So you tell it which directories you want it to use, +[18.90 --> 20.72] and all the modules in those directories, +[20.76 --> 22.54] if you're under the go.work files, +[22.66 --> 24.72] and where you're under a go.mod file before, +[24.88 --> 25.56] are in your workspace. +[25.56 --> 30.64] Hey, Jared here. +[31.22 --> 35.44] One of the things we can count on in the software industry is change. +[36.12 --> 38.74] The state of the art changes so fast, in fact, +[38.90 --> 41.10] that keeping up can feel like a whole other job +[41.10 --> 42.68] on top of your actual job. +[43.50 --> 45.54] That's why we created Change Log Weekly. +[46.14 --> 49.16] It's our totally free newsletter that we drop in your inbox +[49.16 --> 50.44] each and every Sunday. +[50.98 --> 53.82] We link to the latest news, the best articles, +[53.82 --> 56.72] and the most interesting projects that you should be aware of. +[57.36 --> 59.60] We also add a little commentary from us +[59.60 --> 61.00] saying why something's important, +[61.42 --> 63.16] pointing you to other instances of a trend, +[63.26 --> 65.44] or just making a dorky joke to keep it lively. +[66.02 --> 67.12] So if you haven't yet, +[67.36 --> 69.08] I recommend subscribing to Change Log Weekly +[69.08 --> 71.60] and help us help you keep up with the latest. +[72.52 --> 75.78] Head to changelog.com slash weekly and sign up today. +[76.00 --> 78.56] Again, it's totally free and we never spam you. +[78.68 --> 79.02] Yuck. +[79.90 --> 83.40] One last time, that's changelog.com slash weekly. +[83.40 --> 97.18] GoTime.fm +[97.18 --> 98.86] Let's do it. +[99.42 --> 100.46] It's GoTime. +[100.98 --> 102.20] Welcome to GoTime, +[102.86 --> 106.02] your source for diverse discussions about Go 118. +[106.54 --> 109.18] Subscribe today at GoTime.fm +[109.18 --> 112.02] and follow the show on Twitter at GoTimeFM. +[112.02 --> 114.36] Special thanks to our friends at Fastly +[114.36 --> 117.00] for shipping GoTime super fast all around the world. +[117.24 --> 119.16] Check them out at Fastly.com. +[119.76 --> 121.36] All right, let's get into it. +[121.42 --> 122.08] This is a good one. +[122.38 --> 123.10] I think you promise. +[123.76 --> 124.72] Here we go. +[132.48 --> 135.36] Hello and welcome to GoTime. +[135.36 --> 136.44] I'm Matt Ryer. +[136.76 --> 141.32] And today we're talking about the other features in Go 118. +[141.96 --> 143.98] Now listen here, you come here, come here. +[144.20 --> 147.64] Go 118 has got two great big features +[147.64 --> 149.80] that everyone's talking about. +[150.32 --> 151.12] I mean, everybody, +[151.22 --> 152.80] all the popular people are talking about it. +[152.94 --> 153.26] Everyone. +[153.82 --> 154.58] Well, we're not popular. +[154.92 --> 157.96] We're going to be talking about the other features, +[158.10 --> 158.86] not those two. +[159.20 --> 160.44] And just for anyone that doesn't know, +[160.50 --> 161.82] it's fuzzing and generics. +[161.82 --> 164.70] And that's the only time they're going to be mentioned on this episode. +[165.16 --> 166.42] In fact, we've got a new rule. +[166.56 --> 168.94] If they are mentioned by anyone, even accidentally, +[169.42 --> 173.50] unfortunately, you will be immediately booted from the podcast. +[173.70 --> 175.34] So please bear that in mind. +[175.42 --> 177.60] No talk of those two subjects. +[177.78 --> 178.28] Pinky promise. +[179.02 --> 180.34] That's a pinky promise from me. +[180.66 --> 183.20] And we'll find out if we're also going to get a pinky promise +[183.20 --> 185.90] from our special guests today. +[186.06 --> 188.02] It's joining me, Daniel Marty. +[188.34 --> 188.92] Hello, Daniel. +[188.92 --> 190.34] Hello. Nice to be back. +[190.60 --> 193.24] And nice to bring my technical problems along with me. +[194.06 --> 197.78] Your technical problems are, like you, always welcome, Daniel. +[198.24 --> 202.64] Daniel's been using and contributing to Go for quite a few years now. +[202.72 --> 205.16] And you've actually written a few tools as well, +[205.28 --> 207.78] like the stricter GoFumped +[207.78 --> 210.50] and what could be described as the opposite, +[211.08 --> 212.80] a Go code obfuscator. +[213.28 --> 215.54] So they're interesting tools, Daniel. +[215.66 --> 218.22] How is the GoFumped more strict? +[218.92 --> 222.34] It essentially restricts how you can write and format code +[222.34 --> 223.44] in a few extra ways. +[223.54 --> 226.14] Like, for example, no empty lines at the start of a function body. +[226.96 --> 228.22] Things that I generally do. +[229.22 --> 229.98] Cool. Okay. +[230.02 --> 231.60] We'll put a link to that in the show notes, +[231.72 --> 233.02] because if anyone likes... +[233.02 --> 235.30] I like the fact we have GoFumped, +[235.46 --> 237.50] and I like the idea of a more strict one. +[238.04 --> 239.00] Oh, Daniel, did Pinky Promise +[239.00 --> 241.24] not going to talk about those two other subjects today? +[241.60 --> 242.06] Pinky Promise. +[242.30 --> 242.56] Okay. +[242.80 --> 243.84] But Daniel's Pinky Promise. +[244.04 --> 244.20] Okay. +[244.66 --> 245.68] This is really professional. +[245.68 --> 246.08] Okay. +[246.18 --> 246.94] We've also... +[246.94 --> 248.24] You're not going to believe this, Daniel. +[248.44 --> 249.16] You will believe it. +[249.22 --> 249.80] You already know. +[250.02 --> 251.28] But imagine if you didn't. +[251.60 --> 253.70] We've also got Michael Matlub with us. +[253.90 --> 254.54] Hello, Michael. +[254.66 --> 255.56] Welcome to GoTime. +[255.88 --> 256.14] Hi. +[256.38 --> 257.14] Great to be here. +[257.32 --> 258.60] Oh, it's a pleasure to have you. +[258.98 --> 261.18] Michael's on the GoTools team at Google, +[261.44 --> 263.44] living in, you know, New York City. +[263.52 --> 264.16] No big deal. +[264.16 --> 267.72] And he previously worked on Go slash packages, +[267.72 --> 272.48] which is very useful if you're writing like code generation tools and things like that, +[272.48 --> 275.20] and infrastructure for tooling. +[275.36 --> 278.26] And now works on the Go command, right, Michael? +[278.80 --> 279.10] Yes. +[279.42 --> 284.48] And do Pinky Promise to not mention those two big subjects on this very episode? +[284.98 --> 286.44] I won't mention them by name. +[287.14 --> 287.74] Okay. +[287.98 --> 290.46] That sounds like a pinky caveat. +[290.46 --> 291.58] It is. +[291.72 --> 292.80] It is a pinky caveat. +[293.50 --> 294.28] Okay, fine. +[294.94 --> 296.08] We won't mention them by name. +[296.22 --> 296.70] Well, that's it. +[296.76 --> 297.40] We'll see how we go. +[297.94 --> 298.30] Well, yeah. +[298.38 --> 303.30] So this episode, like, obviously, there's a lot of people blogging and talking about +[303.30 --> 305.44] the big sort of headline features that we are. +[305.70 --> 307.70] A lot of people are very excited. +[307.86 --> 310.40] A lot of people are very dismayed about generics in particular. +[310.52 --> 311.00] I've just said it. +[311.10 --> 311.62] I can't believe it. +[311.64 --> 312.60] I have to leave the podcast. +[313.46 --> 314.94] But a raft of other things. +[314.94 --> 320.42] And this release in particular seems very packed and dense with features. +[320.76 --> 322.30] Why is that, do we think? +[322.80 --> 326.04] I believe they've been saving a few large features for some time. +[326.24 --> 330.02] Like, they've been building up to generics for, like, nearly two years now. +[330.10 --> 331.16] And now it's shipping, right? +[331.56 --> 333.68] And I did mention the taboo subject. +[333.84 --> 337.24] But I think it's also happened with fuzzing, which has been in the works for, like, a year now. +[337.38 --> 338.08] Yeah, I see. +[338.20 --> 341.68] By the way, just mentioning another taboo subject doesn't cancel out the other one. +[342.62 --> 343.02] Yeah. +[343.02 --> 345.10] You're just compounding your crimes. +[345.10 --> 353.94] I asked Daniel and Michael to find a list of the things that they're sort of excited about or interested in that we can go through and talk about. +[354.74 --> 358.82] And obviously, Michael worked on module workspaces as well. +[358.82 --> 362.36] So we'll carve some time out at the end to talk about that in particular. +[362.96 --> 366.04] But, Daniel, maybe you could kick us off. +[366.12 --> 372.50] There's a really interesting one that, to me, seemed like a silly, unnecessary helper. +[372.86 --> 375.52] But turns out to be actually quite worthy. +[375.96 --> 378.12] That was strings.cut. +[378.26 --> 379.16] Could you tell us about that? +[379.16 --> 385.26] Yeah, so I think anybody who's written any non-trivial amount of code knows that they have to deal with strings. +[385.42 --> 388.58] They have to add strings, look at prefixes and suffixes and so on. +[388.98 --> 392.28] And one quite common operation is wanting to cut a string in two. +[392.78 --> 396.52] So, for example, maybe you've got a domain name and you want the actual name and the extension. +[397.16 --> 399.68] Or maybe you've got a file name and you want the file name extension. +[400.04 --> 400.72] That kind of thing. +[400.72 --> 408.22] You can use Go APIs like strings.index or there's also strings.splitn and you can give it the number two. +[408.44 --> 412.18] So, like, split this string in up to two pieces, right? +[412.52 --> 415.22] But these APIs are not super easy to use. +[415.38 --> 417.80] For example, if you use index, it may give you minus one. +[418.34 --> 420.38] And if you don't check for that, that might panic. +[421.16 --> 422.78] And split has the same issue, right? +[422.80 --> 424.68] Because it gives you a slice. +[425.04 --> 430.18] So cut is, it has, you could say cut has less sharp edges. +[430.18 --> 436.64] So it only gives you two strings for the two sides and a boolean telling you whether or not it successfully cut. +[437.22 --> 437.98] Yeah, so that's nice. +[438.04 --> 445.30] So if, say, you were cutting on a colon and there wasn't a colon in there, it wouldn't be in any way like a panic or a problem. +[445.52 --> 448.28] You'd just get a false as the second argument. +[448.52 --> 448.82] Exactly. +[449.18 --> 449.36] Yeah. +[449.82 --> 450.74] What do you think about that, Michael? +[451.18 --> 454.04] Have you written code that cuts things up like this? +[454.26 --> 454.78] I have. +[454.94 --> 457.22] Yeah, it would be a nice convenience. +[457.54 --> 458.40] I like conveniences. +[458.40 --> 462.34] Yeah, I thought this was like an unnecessary helper. +[462.52 --> 466.46] Because whenever you can already do something, that's usually my preferred way. +[466.46 --> 469.56] I looked at some of the commentary on this one. +[469.56 --> 476.42] And the number of cases where people were basically doing this same operation over and over again, it's kind of everywhere. +[476.72 --> 484.58] And including some places where we'd done it incorrectly or in a way that would panic if it got some bad input or something. +[484.58 --> 491.76] If there was like some testing tool that helped you try to test out all these different possible ways of responding to input, that'd be great. +[491.88 --> 493.22] But not on this episode, there isn't. +[493.88 --> 494.44] But yeah, okay. +[494.58 --> 497.18] So strings cut and that's coming in Go 118. +[497.82 --> 499.48] Okay, Daniel, have you got another one for us? +[499.48 --> 503.66] So I've got another one that's significantly more complex than strings.cut. +[503.78 --> 507.48] And I believe it was developed by the people at Tailscale over a few years. +[507.48 --> 510.74] And it's essentially a replacement for the net.ip type. +[511.32 --> 516.26] So right now, IP addresses in Go, they represent it as a byte slice. +[516.26 --> 519.24] So you can think of a byte slice, it can have many lengths. +[519.60 --> 523.56] So an IPv4 is going to be shorter than an IP version 6, for example. +[524.22 --> 527.74] And they designed a new IP package, which they called NetAdder. +[528.02 --> 530.58] But now it's being merged as NetIP. +[531.18 --> 532.50] So it's net slash NetIP. +[533.12 --> 537.26] And it's got a bunch of advantages, mostly related around performance. +[537.70 --> 543.24] But the two main properties that it has as part of its design, which do not use a slice, essentially. +[543.84 --> 544.86] One, it's comparable. +[545.14 --> 546.38] So you cannot compare slices. +[546.44 --> 547.56] You can only compare them to nil. +[548.18 --> 550.54] And the other one is that it doesn't allocate. +[550.54 --> 554.68] So you can create a new IP without calling make or new or anything like that. +[554.68 --> 558.20] Because I think it's backed by what is essentially a bunch of integers. +[559.18 --> 563.22] So will the standard library bits of it be rewritten to use this new type? +[563.56 --> 567.78] Or is this just going to be something that's available for calling code? +[568.06 --> 568.96] I think that's a good question. +[568.96 --> 575.04] And I think anything that exposes APIs with the old type will have to remain the same because of backwards compatibility. +[575.64 --> 590.20] I seem to recall one of the reasons to add this to a standard library is so that, for example, HTTP 2 and 3, which I think it's only HTTP 3, which re-implements something like TCP, right, in user space. +[590.70 --> 593.08] And that deals with a lot of IP addresses. +[593.08 --> 598.64] So if you can remove a bunch of internal allocations that don't leak into the API, that can be a very large plus. +[599.10 --> 599.56] Yeah, that's nice. +[599.84 --> 603.24] And do you know if they're going to be helpers to kind of switch between the two? +[603.32 --> 606.38] Do you think we're going to see code like that flying around for a bit? +[606.56 --> 610.32] I believe the package comes with helpers, but my memory is failing me. +[610.88 --> 611.14] Okay. +[611.42 --> 614.50] I mean, if not, people will probably end up doing that, I imagine. +[614.70 --> 618.94] But yeah, it's nice to know that there's a sort of improved data type there. +[618.94 --> 626.16] And it's funny, like, you know, with the Go backwards compatibility promise, you can't just break things and break APIs and break everyone's code. +[626.68 --> 627.78] You know, it's not Python. +[628.32 --> 631.70] I shouldn't have a go at other languages, but Python does that a lot. +[632.32 --> 635.14] But with that promise, of course, your hands get tied. +[635.32 --> 644.28] So this is kind of a nice way of releasing almost like more modern implementations is to sort of release them alongside and then they kind of coexist. +[644.28 --> 646.70] But does that create confusion? +[646.98 --> 648.78] Like, how will people know which one to use? +[648.94 --> 649.78] That's a good question. +[650.14 --> 655.72] I think the Go standard library has a bunch of cases where there are packages and APIs that everybody knows not to use. +[655.82 --> 659.62] Like there's container slash list that has like a linked list, for example. +[659.88 --> 662.46] And I don't think everybody's used that outside of an example. +[662.66 --> 662.80] Yeah. +[663.16 --> 670.92] I don't think they can deprecate net IP simply because, as in the existing net.ip type, because it is used in existing APIs. +[670.92 --> 679.10] But I think there is going to be a common understanding that if you want the extra nice features of the new type, you should just use it. +[679.38 --> 679.46] Right. +[679.88 --> 680.14] Great. +[680.46 --> 681.10] Thank you. +[681.24 --> 681.68] Nice one. +[682.22 --> 682.86] Right, Michael. +[682.86 --> 685.84] Maybe you could pick one to talk about next. +[685.84 --> 692.18] I'll pick a couple of features that my colleagues, Jay and Brian, added to the Go command. +[692.18 --> 708.50] So they are VCS build stamping and a debug build info function to get information about the versions of modules in a Go library as an Go binary as an API. +[708.50 --> 720.08] So both of these have a similar core motivation, which is to improve visibility into binaries, to know which packages they were built on. +[720.28 --> 726.48] So you can determine, say, if binaries were built with certain commits of code. +[726.68 --> 733.04] In the case of VCS build stamping, because the main module may not have a version associated with it. +[733.64 --> 737.06] And in the case of the build info of dependencies. +[737.06 --> 749.56] And this is shaping up to be a big thing in these days to know whether your dependencies and the code that you're built with have bugs or bad features in them. +[749.62 --> 754.38] And if the code that you're running with is safe and to audit everything properly. +[754.78 --> 758.54] We've seen several cases of bad libraries in the wild. +[758.80 --> 763.56] And people have to quickly audit if all their code is safe or not. +[763.70 --> 765.56] All their code running in production is safe or not. +[765.56 --> 767.38] Yeah, and it's a tricky thing. +[767.56 --> 768.86] So I'm picking that a little bit then. +[768.98 --> 776.78] So VCS, version control systems, like the Git hash, when you have a certain level that you've committed up to. +[777.28 --> 778.98] And every time you commit, you get a new hash. +[779.50 --> 783.12] And so now when we build, that will be incorporated. +[783.38 --> 788.00] Is it like it supports all the major kind of VCS systems? +[788.00 --> 790.00] So like it'll support Git and... +[790.00 --> 790.88] Yeah, yeah. +[790.88 --> 792.46] I mean, it definitely supports Git. +[792.58 --> 798.00] I don't know what other VCSs we support if we do support other VCSs. +[798.10 --> 799.32] Dan, do you know? +[799.36 --> 800.22] But they'll be coming soon. +[800.42 --> 804.16] I think there's Mercurial Bazaar subversion. +[804.50 --> 806.18] And that might be it for these days. +[806.42 --> 807.40] That's a good selection. +[808.04 --> 809.10] Could you name five? +[810.38 --> 812.08] Five VCS systems? +[812.28 --> 813.18] Yeah, just five dead quick. +[813.18 --> 815.18] Well, Dan named four, so... +[815.18 --> 815.56] I know. +[815.92 --> 816.72] So it should be easy. +[816.84 --> 823.10] I think there was like, I've seen in the Go command, like a VCS named like Fossil that had support somewhere. +[823.46 --> 825.28] So that'll round us out. +[825.42 --> 825.98] There we go. +[826.06 --> 826.66] Fossil, there we go. +[826.72 --> 826.96] Five. +[827.06 --> 827.28] Brilliant. +[827.44 --> 827.78] There you go. +[828.34 --> 828.68] Learning. +[829.16 --> 829.60] So, okay. +[829.70 --> 835.74] And then the other thing is that build info with all the dependencies, because that is a big thing. +[835.74 --> 849.54] You know, sometimes, well, we're paying a lot more attention now to reporting vulnerabilities, capturing that data, and then being able to, in the tooling, use that to know whether we are dealing or running something that has some known vulnerability. +[850.10 --> 850.18] Yeah. +[850.30 --> 852.14] So that is a massive thing. +[852.26 --> 857.48] And of course, having this put in there automatically saves us a lot of effort, right? +[857.48 --> 868.42] My understanding is that build info is like a function that's like accessible to programs that was like just in the Go command before, like Go version dash M, right? +[868.78 --> 881.28] So it makes it easier for other people to write these auditing programs that can help detect if there are bad versions in your dependencies and, you know, then flag it or fix it or whatever. +[881.28 --> 891.20] And these automated things, I hope, can solve a lot of like manual human work that we've had to do when these issues have come up. +[891.42 --> 891.54] Yeah. +[891.94 --> 902.36] Well, it's very useful with like if you have Dependabot or whatever running in your continuous integration or just running in GitHub, like having those tools help, you know, it's all great. +[902.44 --> 904.60] So anything in that effort, I think, is worth having. +[905.10 --> 905.76] Very cool. +[905.76 --> 911.26] And yeah, before we had to like use either build tags or do something else funky to get the version. +[911.66 --> 912.62] I would always do that. +[912.68 --> 918.84] I would have some script that would, I think I did it with Go Embed as well successfully recently. +[919.32 --> 920.84] But we just don't have to do that now. +[921.00 --> 924.88] And so we'll be able to access that version inside the build as well from somewhere. +[925.34 --> 926.92] Sorry, access the... +[926.92 --> 929.50] Access like the git hash inside the binary itself. +[929.80 --> 930.66] Yeah, it's in the binary. +[930.92 --> 934.32] So I don't know what the API is. +[934.32 --> 934.72] Yeah. +[935.24 --> 937.92] There's some either method or... +[937.92 --> 938.14] Yeah. +[938.40 --> 946.86] It's honestly a bit confusing because before there was an API to get the module information of yourself, like of your own running binary. +[947.14 --> 955.10] But the new API they've added is you can give it a binary path so you can use it with any binary without having to shell out to go version dash and blah, blah, blah. +[955.56 --> 956.74] It's essentially the same feature. +[957.22 --> 960.84] Oh, so the build info includes that, the hash. +[961.22 --> 962.26] Oh, well, there we go. +[962.38 --> 962.58] Okay. +[962.58 --> 969.28] And I think the VCS stamping is also a bit confusing to end users because you tell them Go 118 now stamps VCS build info. +[969.28 --> 977.56] But they might say, if I go install a Go package, a Go main package, and I run Go version dash M with Go 117, I already see the module version, right? +[977.56 --> 990.38] But where that doesn't work is instead of doing a global git install via a module path, if you git clone and then go build or go install locally from that git clone, Go doesn't know what module version that is. +[990.46 --> 991.32] It just has a git clone. +[991.32 --> 997.32] It's not resolving that module through the whole proxy system that tells it what version it is. +[997.72 --> 1000.04] In Go 117, it tells you version devil. +[1000.34 --> 1001.22] It has no idea. +[1001.66 --> 1010.16] And in Go 118, it will add some extra separate metadata that will say, hey, this was built from git hash, blah, blah, blah, date, blah, blah, blah, and so on. +[1010.16 --> 1010.84] All right. +[1011.32 --> 1011.72] Okay. +[1011.88 --> 1014.78] Daniel, your turn to pick one from the list. +[1014.96 --> 1016.92] What else is cool coming in Go 118? +[1017.26 --> 1021.82] By the way, do you say 118, 1 dot, 1 eight, 1 point 18? +[1022.40 --> 1023.28] How do you say it? +[1023.64 --> 1026.38] Now you're making me doubt myself about how I pronounce these things. +[1026.72 --> 1026.98] Yeah. +[1027.18 --> 1028.08] You just have to forget. +[1028.28 --> 1030.06] Just clear your mind and then just say it. +[1030.32 --> 1030.94] See what happens. +[1031.24 --> 1031.60] 118. +[1031.98 --> 1033.06] Yeah, I say 118. +[1033.42 --> 1033.68] Yeah. +[1033.84 --> 1034.10] Okay. +[1034.40 --> 1034.88] That's good. +[1035.04 --> 1036.34] It is kind of 118, isn't it? +[1036.38 --> 1038.88] It's not, because it's not a decimal number, is it? +[1039.30 --> 1039.78] It's Semver. +[1040.04 --> 1041.88] So that second number is 18. +[1041.96 --> 1042.58] I think we're right. +[1042.84 --> 1046.78] So does that mean that when we reach 120, we can go back to 1.2? +[1047.14 --> 1048.90] I think for just that release. +[1049.34 --> 1050.32] No, you can't, can you? +[1050.42 --> 1051.28] No, because that's what I mean. +[1051.28 --> 1051.94] It's not decibel. +[1052.46 --> 1052.90] That's it. +[1053.00 --> 1053.14] Yeah. +[1053.88 --> 1054.92] Yeah, we've got that in the end. +[1054.92 --> 1059.50] So another feature, I mean, it's maybe a bit cheeky that I bring this up because I worked on this, +[1059.50 --> 1064.30] but GoFump without a space now formats files in parallel. +[1064.88 --> 1068.46] So up until now, you have, well, you have two tools, which is also confusing. +[1068.64 --> 1072.80] You have GoFump without a space, and then you have Go space Fump. +[1073.06 --> 1073.34] Yeah. +[1073.38 --> 1078.18] It's ironic that the GoFump tool, it can be called in different ways just by changing the formatting. +[1079.78 --> 1080.30] Yeah. +[1080.94 --> 1081.70] Oh God. +[1082.04 --> 1086.36] The difference between the two tools, and I think it also confuses a bunch of users, +[1086.36 --> 1091.42] is that without a space, it takes files and directories, but it doesn't know what packages +[1091.42 --> 1091.72] are. +[1092.12 --> 1095.58] And with a space, it takes a package pattern. +[1095.72 --> 1097.80] So you can give it dot slash dot dot dot, for example. +[1098.44 --> 1103.80] And the one that works on packages has always been relatively well parallelized, because what +[1103.80 --> 1108.30] it does is, I believe it formats each package in parallel or something like that. +[1108.42 --> 1111.88] But the one that takes directories and files, it would just do one at a time. +[1111.88 --> 1118.48] And now we've essentially removed the parallelism from the one with the space, and just made +[1118.48 --> 1122.92] both tools use the same kind of parallelism, which is GoFump without a space. +[1123.20 --> 1126.24] When you give it a bunch of files to format, it's just going to figure out how to format +[1126.24 --> 1127.20] them as fast as possible. +[1127.92 --> 1128.40] So that's cool. +[1128.52 --> 1132.88] Does GoFump work only within the context of a file at a time then? +[1133.10 --> 1136.76] Like, it doesn't need to know anything else about types and things, does it? +[1136.76 --> 1138.84] Because it's just doing kind of formatting tasks. +[1139.14 --> 1139.68] Yeah, that's correct. +[1139.68 --> 1141.92] So it makes sense, you just do all that at the same time. +[1142.16 --> 1147.66] Yeah, even though there's a few tricky bits about that, because initially my naive implementation +[1147.66 --> 1152.60] was just format each file as a separate GoRoutine as they come in. +[1153.16 --> 1154.90] But some files are really, really tiny. +[1155.28 --> 1159.74] I think like a doc.go file that only has like 10 lines with like a package documentation +[1159.74 --> 1160.60] or something like that. +[1161.00 --> 1166.10] And spawning a new GoRoutine, synchronizing with the parent, maybe allocating the new parser, +[1166.10 --> 1167.64] the new printer, and stuff like that. +[1167.64 --> 1172.52] It actually consumes quite a lot more CPU just because of the overhead of all those tiny +[1172.52 --> 1172.90] files. +[1173.60 --> 1178.28] So we ended up with something that's kind of like chunking groups of files in groups of +[1178.28 --> 1182.78] similar sizes so that they're big enough that actually doing that as parallel units, +[1182.94 --> 1183.50] it's fast. +[1183.82 --> 1184.94] Oh, that's really cool. +[1185.06 --> 1185.78] That's a surprise. +[1185.88 --> 1188.10] I would not have expected it to be doing that. +[1188.28 --> 1191.80] But that's nice to know that that's measured and done properly. +[1191.80 --> 1191.88] Okay. +[1192.76 --> 1193.56] That's very cool. +[1193.66 --> 1194.66] Have you used this then? +[1194.82 --> 1197.86] Have you really noticed this in practice? +[1198.34 --> 1199.10] The speed importance? +[1199.34 --> 1201.08] I think it depends on what people do. +[1201.40 --> 1205.68] I think many people use the tool that works on packages, and then they just format their +[1205.68 --> 1206.04] packages. +[1206.20 --> 1208.94] But I like using the one with directories. +[1208.94 --> 1214.06] So I go to the root of my repository, and I just tell it, format everything, including +[1214.06 --> 1215.44] test files, including everything. +[1216.16 --> 1218.14] And because I did that, it was really slow before. +[1218.46 --> 1222.14] So now, depending on your machine, it's usually about three to four times as fast. +[1222.40 --> 1228.12] So for me, for example, formatting a large repo might go from like five seconds to two +[1228.12 --> 1229.24] seconds, which is nice. +[1230.10 --> 1232.08] Matt Lube, do you format your code? +[1232.60 --> 1232.90] Yes. +[1233.04 --> 1235.26] I mean, we all format our code. +[1235.66 --> 1236.44] It's not a trick question. +[1236.44 --> 1238.68] No, is there anyone who doesn't format their code? +[1239.00 --> 1241.14] Because I want to hear about it. +[1241.22 --> 1242.74] It's like a problem we need to solve. +[1243.02 --> 1243.34] Oh, yeah. +[1243.40 --> 1244.00] No, I don't. +[1244.06 --> 1245.10] I don't think so. +[1245.16 --> 1247.08] Because you only have to do it a few times. +[1247.08 --> 1251.74] And then when pull requests, although they improved it in GitHub, where whitespace was +[1251.74 --> 1252.70] understood better. +[1253.02 --> 1257.30] But it certainly used to be that what you'd get just pull requests that every line has +[1257.30 --> 1259.66] changed because some whitespace thing. +[1259.82 --> 1265.12] And that got so annoying that it's very high motivator, I think, to get people formatting. +[1265.12 --> 1266.10] But I don't know. +[1266.10 --> 1268.44] I assume everyone does format their code. +[1268.88 --> 1270.92] Do you do it in the way Daniel described then? +[1271.10 --> 1274.96] Or do you do it like me, where you just, every time you save a file, it does just that file? +[1275.62 --> 1278.60] I don't think I've ever run either of the tools. +[1279.24 --> 1282.42] Or I certainly haven't run either of the tools by hand in years. +[1282.80 --> 1286.36] My editors are just set up to format files as I save them. +[1286.82 --> 1287.12] Oh, yeah. +[1287.16 --> 1287.28] Yeah. +[1287.28 --> 1291.40] I thought you were saying that you just write it in perfect, go-thumped way first time. +[1291.40 --> 1293.56] Oh, no, no, I don't. +[1293.62 --> 1294.02] Nailed it. +[1294.08 --> 1298.06] I write it in the wrong way and just let the format or take care of it. +[1298.22 --> 1300.74] Like any good codeveloper. +[1301.00 --> 1301.52] Yeah, exactly. +[1301.66 --> 1306.30] To be honest, I'll deliberately make mistakes so that when I hit save, I get the visual clue +[1306.30 --> 1307.34] that it has formatted. +[1307.74 --> 1311.16] Because if I write it and I get it right and then I hit save and nothing happens, I'm like, +[1311.50 --> 1312.58] computer's not working. +[1312.58 --> 1318.40] So I genuinely sometimes like to see the little shift into place of things as a clue that it's +[1318.40 --> 1318.76] working. +[1319.10 --> 1319.22] Yeah. +[1319.28 --> 1323.04] I mean, that's like a nice way to know that like, oh, the syntax is correct. +[1323.32 --> 1324.24] Yeah, that's true. +[1324.46 --> 1325.30] You know, it can parse. +[1325.60 --> 1328.22] Yeah, because if it errors, it doesn't complete it. +[1328.34 --> 1330.56] So it actually is a feedback loop thing. +[1330.88 --> 1331.22] There you go. +[1331.44 --> 1332.46] Tip there for everyone. +[1332.72 --> 1334.40] I've actually done that with tests. +[1334.56 --> 1339.24] Like if you write a ton of software and some tests and you run the tests and everything's +[1339.24 --> 1341.56] green, I often go like, I don't believe that. +[1341.56 --> 1345.44] Let me bring one of the tests to see if I'm doing something really dumb right now. +[1345.58 --> 1347.02] Yeah, absolutely. +[1347.58 --> 1350.78] In TDD, they do talk about that red-green testing for that reason. +[1350.88 --> 1352.44] Like you have to see the test fail. +[1352.80 --> 1355.36] So you know it's saying something useful. +[1355.84 --> 1357.94] And then when you fix it, that's true. +[1358.04 --> 1362.02] If I write some code and it just, even if I'm just running it and I'm going to run it +[1362.02 --> 1366.90] myself and look at the results in the terminal, like without even any tests, if that works +[1366.90 --> 1370.20] first time, I'm highly suspicious, really suspicious. +[1370.20 --> 1374.00] So yeah, in a way, I'm not happy when it does. +[1374.92 --> 1375.24] Okay. +[1375.38 --> 1379.50] We've also got the Pacer redesign in the garbage collector. +[1379.86 --> 1380.26] Right. +[1380.46 --> 1381.16] That's interesting. +[1381.28 --> 1382.16] What's going on there then? +[1382.38 --> 1386.36] I brought this up because I think it's a very interesting topic, but I think we should also +[1386.36 --> 1389.58] warn that none of us here are experts in this area. +[1389.58 --> 1393.60] So we can talk about it at a high level, but I'm going to stop there. +[1393.70 --> 1397.94] If anybody wants to read more about it, we can mention the issue number and then they +[1397.94 --> 1400.46] can go and read the whole doc. +[1400.78 --> 1402.12] And I think that's very reasonable. +[1402.70 --> 1402.82] Yeah. +[1403.12 --> 1403.58] Fair enough. +[1403.68 --> 1404.26] Good disclaimer. +[1405.22 --> 1411.06] To give a bit of an intro, the way I understood it, because again, I just read this, the GC Pacer, +[1411.06 --> 1415.20] it's the part of the garbage collector that decides when a new collection should happen. +[1415.68 --> 1419.68] So it's sort of the thing that times when the GC should be doing its work, because if +[1419.68 --> 1423.62] it happens too often, then you're just burning too much CPU, you're wasting time. +[1423.76 --> 1428.36] But if you run it too little, you might be holding onto too much memory, or you might be +[1428.36 --> 1432.60] delaying some things happening in the runtime that you don't want to delay by very long. +[1432.60 --> 1436.84] It seems like the GC Pacer was designed a while ago. +[1437.34 --> 1439.44] For the purpose that it was designed, it was good. +[1439.68 --> 1443.78] But over time, it's accumulated a bunch of debt and a bunch of quirks. +[1444.06 --> 1448.44] And they've sort of sat down and said, okay, let's redesign it in a way that it does a lot +[1448.44 --> 1452.96] better in these edge cases that we found in production workloads that the old one doesn't +[1452.96 --> 1453.64] do very well in. +[1454.14 --> 1456.12] And I think that's where I'm going to leave it. +[1456.92 --> 1456.96] Hmm. +[1457.60 --> 1458.62] That's very exciting. +[1458.62 --> 1463.32] I'm really interested whenever there are these kind of really low level, because it's funny, +[1463.46 --> 1468.72] like when you dig into these little subsystems, they're just like other types of programs. +[1468.72 --> 1472.60] Like they are just doing the same things that we're doing in our programs. +[1472.82 --> 1478.74] But they're just so kind of, it's such an interesting domain, I think, that it always makes it more +[1478.74 --> 1479.08] interesting. +[1479.18 --> 1483.78] And the fact that, I love the fact that as programmers, we get this for free. +[1483.96 --> 1487.72] Like people are doing this work for us to make these improvements. +[1487.72 --> 1490.98] Like I didn't even know about a pacer, to be honest. +[1491.16 --> 1493.74] So it's very nice to know that that's happening. +[1494.16 --> 1495.50] What do you think about that, Michael? +[1496.02 --> 1497.00] Do you know anything about this? +[1497.46 --> 1504.38] Not, I mean, I am not closely acquainted with it, but I think any runtime improvements are +[1504.38 --> 1505.66] well appreciated. +[1506.50 --> 1507.76] Good work, team. +[1508.22 --> 1508.44] Yep. +[1508.64 --> 1512.44] And it's Michael, there's another Michael who I think was the author of the redesign. +[1512.68 --> 1513.14] Is that right? +[1513.76 --> 1514.16] Yeah. +[1514.16 --> 1517.08] Do you know all the other Michaels on the Go team? +[1517.32 --> 1520.06] Or have you got together yet with all the rest of the Michael? +[1520.42 --> 1521.14] Is there another? +[1521.94 --> 1522.72] There's a Michael. +[1523.28 --> 1524.80] I may just be a contributor, actually. +[1525.12 --> 1525.82] There's two. +[1526.70 --> 1527.80] Are there more than two? +[1528.58 --> 1530.52] Or are there three, I guess, including myself? +[1530.90 --> 1531.18] Okay. +[1531.38 --> 1532.70] I don't want to be forgetting anyone. +[1532.84 --> 1533.74] So if I forgot you. +[1533.80 --> 1535.50] No, I think we should spend time on this. +[1536.66 --> 1537.58] Don't forget anyone. +[1537.94 --> 1539.48] We should not spend time on this. +[1539.66 --> 1540.86] Mind you, you're just telling me a number. +[1540.86 --> 1545.54] Even if you forgot a Michael, they don't know which one they've, they don't know that they've +[1545.54 --> 1546.06] been forgotten. +[1546.28 --> 1547.12] It's just, there you go. +[1547.20 --> 1547.82] I think you're safe. +[1548.30 --> 1551.88] So everyone can assume that I included them in the list of Michaels. +[1552.10 --> 1552.42] Yeah. +[1552.56 --> 1554.22] So calm down, Michaels. +[1554.34 --> 1554.94] You were counted. +[1555.36 --> 1557.76] All Michaels have been accounted for. +[1559.50 --> 1559.90] Excellent. +[1559.90 --> 1564.38] I was thinking before we go on to the next topic, if anybody wants to read about this, +[1564.56 --> 1567.08] the issue number is 44167. +[1567.70 --> 1571.50] And at the end of the issue, which is very short, there's a link to the full proposal +[1571.50 --> 1573.28] design, which is very long. +[1573.54 --> 1576.34] And you can read that carefully and get the full picture. +[1576.72 --> 1576.88] Yeah. +[1577.00 --> 1578.54] It looks very well written. +[1578.88 --> 1582.62] And we'll post the link to all of these in the show notes. +[1582.68 --> 1585.58] So you'll be able to go and actually look at the original issues. +[1585.58 --> 1590.96] And honestly, like notice that some of these issues aren't created by members of the Go +[1590.96 --> 1596.24] team or even popular contributors like Daniel who've contributed massively. +[1596.86 --> 1600.26] Sometimes these come from just people in the community that have a problem that they want +[1600.26 --> 1601.54] to solve or something they care about. +[1601.86 --> 1606.20] So we do get stuck in basically, because you never know, you might get some improvements +[1606.20 --> 1607.94] made and that'd be great for everyone. +[1615.58 --> 1624.16] We are going to send three, two, one. +[1624.62 --> 1629.90] I'm Karhara Zhu, host of Ship It, a show with weekly episodes about getting your best ideas +[1629.90 --> 1631.88] into the world and seeing what happens. +[1632.22 --> 1637.28] We talk about code, ops, infrastructure, and the people that make it happen like charity +[1637.28 --> 1638.36] majors from Honeycomb. +[1638.70 --> 1641.64] We act like great engineers make great teams. +[1641.84 --> 1643.28] And it's exactly the opposite. +[1643.28 --> 1646.92] In fact, it is great teams that make great engineers. +[1647.58 --> 1650.82] And Dave Farley, one of the founders of Continuous Delivery. +[1651.18 --> 1653.98] Start off assuming that we're wrong rather than assuming that we're right. +[1654.24 --> 1655.22] Test our ideas. +[1655.36 --> 1656.86] Try and falsify our ideas. +[1657.00 --> 1658.98] Those are better ways of doing work. +[1659.04 --> 1661.28] And it doesn't really matter what work it is that you're doing. +[1661.42 --> 1663.10] That stuff just works better. +[1663.60 --> 1669.54] We even experiment on our own open source podcasting platform so that you can see how we implement +[1669.54 --> 1672.60] specific tools and services within changelog.com. +[1672.90 --> 1674.68] What works and what fails. +[1675.08 --> 1678.92] It's like there's a brand new hammer and we grab hold of it and everyone gathers around. +[1679.02 --> 1682.82] We put our hand out and we strike it right on our thumb. +[1683.06 --> 1685.90] And then everybody knows that hammer really hurts. +[1686.06 --> 1688.54] When you strike it on your thumb, I'm glad those guys did it. +[1688.60 --> 1689.42] I've learned something. +[1689.58 --> 1690.24] Instead, yeah. +[1690.24 --> 1694.98] I think that's a very interesting perspective, but I don't see that way. +[1695.16 --> 1695.38] Okay. +[1695.50 --> 1698.60] It's an amazing analogy, but I'm not sure that applies here. +[1698.96 --> 1701.24] Listen to an episode that seems interesting or helpful. +[1701.40 --> 1703.04] And if you like it, subscribe today. +[1703.16 --> 1704.30] We'd love to have you with us. +[1708.98 --> 1711.06] Does anyone have the M1 chip? +[1711.52 --> 1712.60] Apple's M1. +[1713.04 --> 1715.30] I have it on my personal laptop. +[1715.68 --> 1716.30] Yeah, that counts. +[1716.80 --> 1717.62] It's fast, isn't it? +[1717.82 --> 1718.78] Oh yeah, it's great. +[1718.78 --> 1721.64] I've been surprised with how fast it is. +[1721.96 --> 1722.28] Me too. +[1722.48 --> 1726.02] I got a new MacBook Pro recently and it's phenomenal. +[1726.24 --> 1726.64] Absolutely. +[1727.32 --> 1730.96] But Go had support for the M1 chip for quite a while, didn't it? +[1731.24 --> 1732.38] What does that look like? +[1732.44 --> 1734.08] How do we support another chip? +[1734.38 --> 1738.26] Could someone just briefly, and I do mean briefly, like we don't have to get into the +[1738.26 --> 1740.24] weeds of it, but what do we have to do? +[1740.30 --> 1745.38] Is it literally, we have to add some kind of mapping file for all the instructions so +[1745.38 --> 1748.18] that a compiler knows what to compile them into? +[1748.18 --> 1750.18] And it's different if it's a different chip? +[1750.56 --> 1752.86] Because there's also the Rosetta 2 stuff. +[1753.00 --> 1758.70] So that even if a binary on these new architectures hasn't been built for that architecture, this +[1758.70 --> 1759.76] is translation layer. +[1760.14 --> 1764.70] And to be honest, they're still lightning fast, like as far as I can see when I run programs +[1764.70 --> 1765.24] like that. +[1765.74 --> 1767.08] But there are some improvements coming. +[1767.14 --> 1767.58] Is that right? +[1767.58 --> 1773.42] I do seem to recall that when the M1 first came out, Go did already support ARM64. +[1773.74 --> 1776.16] So the 64 version of the ARM architecture. +[1776.50 --> 1782.58] But binary's build for Go targeting the architecture didn't work out of the box for one reason, +[1782.66 --> 1787.16] because there wasn't a Darwin slash ARM64 port yet. +[1787.16 --> 1791.92] So Go did support Mac, and it supported ARM64, but not together yet. +[1792.04 --> 1797.26] So they needed to add some glue code to essentially make those two work together. +[1797.60 --> 1802.20] And I think the other major work they had to do was the whole thing about signing binaries, +[1802.68 --> 1805.94] because I think the M1 was the first machine that required all binaries to be signed. +[1806.14 --> 1810.00] So they had to teach the linker how to sign binaries locally, something like that. +[1810.00 --> 1810.72] Yeah, yeah. +[1811.04 --> 1811.90] Oh, that's very cool. +[1812.14 --> 1813.68] Well, I just noticed it started working. +[1814.12 --> 1820.00] There's also a lot of work that needs to be done when we're signing binaries for, +[1820.28 --> 1824.42] when we're making releases, when Apple makes changes to their operating system, +[1824.88 --> 1832.38] we often have to change the infrastructure we use to produce the Go distributions that people +[1832.38 --> 1832.94] get. +[1833.08 --> 1834.46] And that takes a lot of work. +[1834.46 --> 1839.62] And I kind of just want to kind of mention all the work that the Go release team has +[1839.62 --> 1845.96] done to make our releases smooth, because sometimes that goes, it's not explicitly talked +[1845.96 --> 1846.56] about as much. +[1846.90 --> 1851.86] So I imagine every time Apple says a new major version of macOS is coming, I imagine some +[1851.86 --> 1854.52] people start sweating, thinking, oh no, what is coming? +[1855.62 --> 1860.60] Yeah, I mean, sometimes there's like nothing, but sometimes they're disruptive. +[1860.60 --> 1866.90] Was it Catalina that they like introduce like major signing requirements or something? +[1867.18 --> 1869.28] It caused big problems. +[1869.92 --> 1872.50] Well, again, we do appreciate all that work. +[1873.02 --> 1878.02] Newer x86-64 machines are also getting improvements, aren't they, Daniel? +[1878.42 --> 1884.52] Yeah, so that's a good segue because going from, for example, ARM-based machines, there's +[1884.52 --> 1885.18] a lot of versions. +[1885.50 --> 1889.76] If you have an old phone, I believe that's going to be like ARM version 6, but later phones +[1889.76 --> 1892.54] are going to be ARM version 8 or 9, which is 64 bits. +[1893.20 --> 1897.98] And if you compile a binary that's targeting like the lowest possible denominator, the +[1897.98 --> 1901.68] older version, it's not going to run as fast as it could on a newer device. +[1902.12 --> 1907.98] So Go has had a flag called, I think it's called Go ARM 64, and you tell it what version +[1907.98 --> 1911.42] of the architecture your machine, your target machine supports. +[1911.70 --> 1917.24] And then if you swap a 6 for a 9, it might run 10% faster, depending on what kind of code +[1917.24 --> 1917.66] you're running. +[1918.32 --> 1920.52] And x86-64, i.e. +[1920.76 --> 1927.40] AMD 64 desktop CPUs, they don't suffer from as much of the same problem because they haven't +[1927.40 --> 1930.82] had as many versions with as many changes in the last decade or two. +[1931.32 --> 1933.22] But you have had some changes. +[1933.76 --> 1939.08] And sort of mirroring the same environment variable for ARM 64, now we have Go AMD 64. +[1939.48 --> 1941.42] And it targets one of four versions. +[1941.42 --> 1947.00] And these are sort of standard versions between Intel and AMD, where roughly speaking, I believe +[1947.00 --> 1948.94] version one is like the common denominator. +[1949.40 --> 1952.54] It's basically every single machine that's valid AMD 64. +[1952.98 --> 1957.66] And then you've got version two for things that are starting, I think, in like 2010 or +[1957.66 --> 1957.92] so. +[1958.38 --> 1960.94] Version three starting in like 2013, 2014. +[1961.42 --> 1967.30] And then version four, which is, I think, AVX 512, which is mostly server computers or very +[1967.30 --> 1968.80] new desktop computers. +[1968.80 --> 1974.32] So if, for example, you know you're targeting a cloud machine and you know the cloud machine +[1974.32 --> 1978.76] has all these new instructions, you can swap from the older version one to version three +[1978.76 --> 1979.16] or four. +[1979.66 --> 1984.42] And maybe you're going to save five, 10% CPU cost, depending on what kind of code you're +[1984.42 --> 1984.60] running. +[1985.04 --> 1990.16] And presumably if you choose a higher number and then the architecture is lower, then +[1990.16 --> 1990.78] that's a problem. +[1991.04 --> 1993.32] I believe it's just going to fail, refuse to run. +[1993.46 --> 1994.64] It's going to say not supported. +[1994.98 --> 1995.18] Yeah. +[1995.54 --> 1996.24] Okay, cool. +[1996.40 --> 1997.08] Yeah, makes sense. +[1997.08 --> 1997.88] Huh? +[1998.24 --> 1998.88] Yeah, there you go. +[1998.94 --> 1999.62] That's good to know. +[1999.94 --> 2000.28] Yeah. +[2000.42 --> 2006.38] I mean, I often I'm so abstracted from the physical hardware in certain environments where +[2006.38 --> 2008.54] that wouldn't be able to make use of that. +[2008.60 --> 2012.56] But there's certainly some cases where I could probably use that today. +[2012.92 --> 2014.80] I appreciate you telling me about that one. +[2015.14 --> 2018.16] And even if you think, well, my workload is not that special. +[2018.16 --> 2024.50] I believe in Go AMD 64 version three, there's an instruction that the runtime garbage collector +[2024.50 --> 2030.26] can use to quickly scan memory for pointers or something like that in a way that essentially +[2030.26 --> 2031.94] batches the work and makes it a lot faster. +[2032.42 --> 2037.64] So you might get the runtime GCs being like a few percent faster, even if you don't care +[2037.64 --> 2038.66] about new CPUs. +[2038.66 --> 2044.12] So even if you're not going to make use of it, maybe the Go tooling and runtime and bits +[2044.12 --> 2045.52] and pieces do. +[2045.72 --> 2046.50] Very interesting. +[2047.08 --> 2051.32] I do want to speak about one more subject before we get onto workspaces if we can. +[2051.46 --> 2053.18] And this is something I use a lot. +[2053.26 --> 2055.32] And these are the templates in Go. +[2055.32 --> 2059.30] So we've got text template and HTML template. +[2059.80 --> 2064.62] And these sometimes get criticized as being too rudimentary and too low level. +[2065.68 --> 2068.60] But it sort of has enough of what you need. +[2068.70 --> 2073.40] As long as you mix in Go code, usually in functions that you make available to the templates, +[2073.88 --> 2075.60] you can kind of really do everything you need. +[2076.08 --> 2078.76] But are we getting some new functionality in templates? +[2079.34 --> 2079.48] Yeah. +[2079.72 --> 2083.42] So I added a couple here, which are pretty simple to understand, I think. +[2083.42 --> 2088.08] They both revolve around control flow or logic, if you want to think of it that way. +[2088.40 --> 2090.62] So one is about adding break and continue. +[2090.94 --> 2096.32] So it's the same feature that you have in regular Go loops, but for ranges within a template. +[2096.86 --> 2104.72] And the other one is that the AND and OR operators in Boolean expressions now short circuit in a +[2104.72 --> 2110.62] template like in Go, which means that if you do A or B and A is true, then B is not evaluated. +[2111.02 --> 2113.26] Whereas right now it evaluates all the expressions. +[2113.42 --> 2115.98] And then works out the Boolean expression. +[2116.46 --> 2116.56] Yeah. +[2116.68 --> 2120.26] And the result on the expression itself is the same, isn't it? +[2120.32 --> 2124.82] But if you like you're calling functions within that, then you can save those functions. +[2124.92 --> 2126.02] They won't need to get called. +[2126.46 --> 2128.82] So that short circuiting sometimes is very important. +[2129.38 --> 2130.72] That's very nice to know. +[2131.14 --> 2135.08] So the break and continue, I guess they are quite simple then. +[2135.08 --> 2137.70] So continue is going to loop back. +[2138.26 --> 2140.86] And well, actually, I'm not sure that is that simple. +[2141.22 --> 2143.88] Because the template is kind of declarative, isn't it? +[2144.34 --> 2145.88] What does the continue do then? +[2146.26 --> 2150.68] What happens if there was within the block, like content after the continue? +[2150.92 --> 2151.68] Is that skipped? +[2151.68 --> 2155.20] So you can think of templates as sort of scripts. +[2155.68 --> 2159.82] I don't believe they let you run code forever, at least not that I can remember. +[2159.94 --> 2164.14] But they do have a range statement where you can say range over, for example, a slice. +[2164.14 --> 2169.18] And then within that body, you can set variables or you can template some. +[2169.48 --> 2172.52] Like if you just type something without using the brackets, right? +[2172.54 --> 2174.52] That's going to be output as part of the template. +[2174.98 --> 2175.06] Yeah. +[2175.16 --> 2178.96] If you have two blocks of code within a range and in between you say continue, +[2179.14 --> 2180.62] then the second block is going to be omitted. +[2180.80 --> 2183.02] And then you're going to go back to the top of the range, right? +[2183.36 --> 2183.52] Yeah. +[2183.56 --> 2183.78] Okay. +[2183.82 --> 2185.56] So that is how it works and goes. +[2185.66 --> 2186.78] So that should feel quite natural. +[2186.78 --> 2188.78] But that is quite unusual for templating. +[2189.10 --> 2190.48] I don't think I've seen that before. +[2190.58 --> 2191.38] It is a bit unusual. +[2191.52 --> 2191.76] Yes. +[2192.16 --> 2192.54] Very cool. +[2192.54 --> 2197.58] Well, we have somebody here, of course, Michael Matloub, +[2197.82 --> 2202.62] who has done a fair bit of work recently on workspaces. +[2202.98 --> 2205.58] And this is coming in Go 118. +[2205.96 --> 2208.98] Michael, could you just tell us briefly what are Go workspaces? +[2209.08 --> 2210.22] What problem do they solve? +[2211.00 --> 2216.28] So just like at a simple level, the Go command in the module mode +[2216.28 --> 2220.34] allows you to have a single main module that you're working on, right? +[2220.34 --> 2224.82] Like that's the module that your current directory is in. +[2225.38 --> 2232.34] And all the files in the module, all the packages in the module are like the modules that are the packages that Go builds by default. +[2232.34 --> 2241.02] And if you have any other code on disk, previously you would have to like add replaces or other ways of kind of getting it in, +[2241.30 --> 2245.38] which are kind of annoying if you want to make changes across modules. +[2245.48 --> 2248.34] It was hard to work across two modules at the same time, basically. +[2249.04 --> 2252.40] Now, workspaces allow you to have more than one main module. +[2252.40 --> 2260.04] Those are modules where you are making edits and Go builds from rather than getting it from a specific version. +[2260.76 --> 2264.42] And so Workspace is allowed you to say, these are the modules on this that I'm working on. +[2264.72 --> 2274.06] And those are like the base that the minimal version selection uses when computing its dependency graph. +[2274.06 --> 2282.36] So we think this is going to be useful because we've gotten a lot of feedback from people who work across multiple modules. +[2282.52 --> 2288.06] In fact, that was like one of the number one complaints we saw in the Go user survey. +[2288.62 --> 2295.76] People working with modules that they had problems when they were working on multiple modules, they found it cumbersome. +[2296.44 --> 2300.66] And so we hope that multi-module workspaces make that workflow a lot easier for them. +[2301.06 --> 2303.08] Yeah, this is definitely something I've encountered. +[2303.08 --> 2306.84] Do you think people were like overusing modules? +[2307.44 --> 2311.18] Do you think that there's like, you know, we were doing something wrong? +[2311.26 --> 2314.42] It felt like that because we were kind of fighting with the tools a little bit. +[2314.80 --> 2317.18] What do you mean overusing modules? +[2317.54 --> 2322.10] Well, I mean, like sometimes in a project, you have like multiple packages. +[2322.78 --> 2328.40] Sometimes people will, each one of those would be a module instead of just a package inside this bigger module. +[2328.58 --> 2329.36] Things like that. +[2329.36 --> 2349.22] Yeah, I feel like one thing that we learned after, you know, some experimentation with modules, like after some time using like Vgo and then modules in the Go command, we learned that multi-module repositories are, they should be rare. +[2349.22 --> 2353.22] They have a lot of surprises when you're working with them. +[2353.22 --> 2371.84] And so like now our general recommendation is for people to usually have like one module per repository unless there's a specific, very rare set of use cases where they wanted to have a sub module in their module. +[2371.84 --> 2378.92] So in that sense, yeah, I guess people, we were overusing modules because we were learning how to use modules. +[2379.26 --> 2384.08] And now those modules exist and we kind of have to work with them. +[2384.36 --> 2384.50] Yeah. +[2384.70 --> 2391.52] I think for like packages and things for if you're releasing a library that people are going to use, I think that's kind of great advice. +[2391.66 --> 2395.22] Definitely a time I've seen where multiple modules exist is if you have a mono repo. +[2395.22 --> 2403.62] And the way that you would do it at the moment, I use Visual Studio Code, you basically open the folder, the subfolder just as the root. +[2403.90 --> 2406.94] And that's essentially like that becomes the context of that module. +[2407.16 --> 2408.80] And that's a way to get around it. +[2408.86 --> 2415.42] If you have multiple folders and they have modules in different ones, the workspaces, I think, is going to enable that now. +[2415.52 --> 2417.70] So you'll be able to operate, right? +[2417.70 --> 2418.34] Yeah. +[2418.58 --> 2435.72] I mean, one of the driving forces behind us starting to work on modules was the user experience in not just Visual Studio Code, but like any editors that use Go, please, which kind of powers the Visual Studio Code Go experience. +[2435.72 --> 2443.82] The team was thinking of different ways of representing multiple modules and providing that information to the Go command. +[2444.44 --> 2450.56] But it had to like kind of introduce a new concept that like didn't exist in the Go command. +[2450.64 --> 2455.14] Like the Go command had no concept of like people working in multiple modules at the same time. +[2455.20 --> 2458.64] It just had, you know, replace directives or requirements. +[2458.64 --> 2465.24] And so we decided like the best thing to do is to like make this a first class feature of the Go command. +[2465.54 --> 2476.42] So not only could Go please use it, but users who introduce modules can then open up, you know, command line and the Go command understand that they're working in the same workspace and the same set of modules. +[2477.28 --> 2479.34] So how does it actually work in practice then? +[2479.42 --> 2482.16] Do you have to like set up a workspace? +[2482.38 --> 2484.72] Is this a new concept of a thing you create? +[2485.06 --> 2485.62] Yes. +[2485.62 --> 2490.10] So we have Go.mod files and now we have Go.work files. +[2490.48 --> 2492.00] So you create a Go.work file. +[2492.12 --> 2494.40] The syntax is very similar to Go.mod. +[2494.58 --> 2496.48] We want it to be easy for people to pick up. +[2497.10 --> 2502.02] And the Go.work has one new directive is the use directive. +[2502.44 --> 2506.06] So you tell it which directories you want it to use. +[2506.40 --> 2513.82] And all the modules in those directories, if you're under the Go.work file, the same way you're under a Go.mod file before, are in your workspace. +[2513.82 --> 2518.46] So you make your Go.work file and CD under it. +[2518.62 --> 2524.62] And now you're using all those modules and any builds that you do or Go list or any such command like that. +[2525.08 --> 2529.62] And it's going to be aware of those other modules and you're not going to be fighting the tools anymore. +[2530.00 --> 2530.18] Yep. +[2530.18 --> 2532.68] And can you do replaces as well in there? +[2532.90 --> 2533.50] You can. +[2534.14 --> 2540.68] We don't think people should need to add replaces except in like very specific circumstances. +[2540.82 --> 2547.54] We actually added replace because if you have multiple modules in your workspaces, they might have conflicting replaces. +[2547.54 --> 2553.48] And so the replacing the Go.work file can override the replaces that are fighting. +[2554.42 --> 2562.22] But if you wanted to use a specific module, then usually the right thing to do is just use that module in your Go.work file. +[2562.70 --> 2562.84] Yeah. +[2563.10 --> 2568.70] So I feel like lots of tools probably got touched by adding this kind of support. +[2569.02 --> 2570.88] Was it a difficult one to get in? +[2571.22 --> 2574.90] Most of the work was in the Go command itself. +[2574.90 --> 2584.14] I mean, there is definitely like an amount of complexity in the Go command because our module loading code does more than you'd think. +[2584.80 --> 2590.80] But once we got it to work in the Go command, one of the nice things is most of our tools call into the Go command. +[2591.56 --> 2602.10] And so, you know, as long as they're making a call into the, you know, Go packages or the Go command, they kind of get all that for free as long as the Go.work file exists on disk. +[2602.10 --> 2612.88] You know, we've had to like make on the VS Code Go and Go please teams, they've made changes to understand Go.work files and pass them into the Go command. +[2613.20 --> 2618.24] But, you know, once you pass it in, like all the hard work is done by the Go command itself. +[2618.72 --> 2620.62] This is a really nice thing to have. +[2620.92 --> 2624.00] In the way that I work, this is going to change a lot. +[2624.26 --> 2627.44] There's an experiment for a while where you could try this out. +[2627.44 --> 2628.18] Wasn't there? +[2628.40 --> 2629.52] With an environment variable. +[2629.84 --> 2644.98] Yeah, we, you know, I filed a proposal for this and I made it available and we had a link for people to easily download like a development version that included these changes using the Go tip command. +[2645.70 --> 2647.56] And so people could try it and give feedback. +[2648.40 --> 2651.28] And we got some feedback on it, which was super helpful. +[2651.28 --> 2656.60] And we got a lot of feedback on the issue too, which was very helpful in developing the issue. +[2657.02 --> 2658.60] So, or the proposal. +[2659.48 --> 2665.62] So, yeah, I mean, hopefully we've addressed most of the important issues people have. +[2665.84 --> 2667.72] For anything else, there's 119. +[2669.40 --> 2669.80] Absolutely. +[2670.18 --> 2673.86] Well, no, I mean, honestly, I think these kinds of things make such a difference. +[2674.06 --> 2675.26] So we're so pleased. +[2675.26 --> 2679.50] And honestly, I feel like there's a lot more to talk about with workspaces and modules. +[2679.62 --> 2684.80] Maybe, Michael, you could come back one day and we'll do like a modules and workspaces special. +[2685.18 --> 2685.60] Oh, sure. +[2685.72 --> 2686.44] I'd love to. +[2686.72 --> 2687.00] Yeah. +[2687.48 --> 2688.16] Okay, cool. +[2688.24 --> 2689.74] Well, we will hold you to that. +[2689.88 --> 2691.72] I do consider this to be legally binding. +[2692.10 --> 2693.60] Like the pinky promises? +[2694.08 --> 2696.94] Pinky promises are of all the types of promise. +[2697.08 --> 2698.56] I think they're up there, aren't they? +[2698.60 --> 2700.38] With the most important, aren't they? +[2700.92 --> 2701.98] Pinky, you know what I mean? +[2701.98 --> 2704.98] You've got like local kind of laws. +[2705.20 --> 2707.50] You've got like national laws. +[2707.88 --> 2712.48] And then all the way at the top, after the Supreme Court, you've got the little pinky promise there. +[2712.84 --> 2714.46] It's been pinky promise at the top. +[2714.56 --> 2715.04] I got it. +[2715.44 --> 2716.32] Yeah, I think that's how it works. +[2716.38 --> 2719.08] I think that's the legal structure of the pinky promise. +[2719.46 --> 2721.42] Well, here's another pinky promise. +[2721.58 --> 2724.36] I promise you're about to hear some unpopular opinions. +[2728.36 --> 2730.08] Unpopular opinions. +[2730.84 --> 2731.18] What? +[2731.18 --> 2733.02] I actually think should probably leave. +[2736.36 --> 2737.92] Unpopular opinions. +[2741.86 --> 2742.34] Okay. +[2743.74 --> 2745.80] Who's going to go first with... +[2745.80 --> 2747.44] I don't know why I'm speaking in the spooky voice. +[2747.60 --> 2749.54] Who wants to say the first unpopular opinion? +[2750.00 --> 2751.48] So maybe I can start with mine. +[2751.98 --> 2756.20] Mine is that I think code generation should be avoided whenever possible. +[2756.20 --> 2761.14] I think the main reason for that is because it adds developer friction. +[2761.54 --> 2764.64] It often increases build size and build time. +[2765.18 --> 2769.60] And oftentimes people overestimate how slow reflection is. +[2769.76 --> 2773.04] If you use reflection well, the cost is actually very reasonable. +[2773.26 --> 2776.96] And it's not like you're building your whole program around reflection. +[2776.96 --> 2779.92] You're using it in very careful ways in small places. +[2780.80 --> 2782.14] What do you think of that, Michael? +[2782.52 --> 2787.92] I guess I don't have like a very strong opinion about this either way. +[2787.92 --> 2788.68] I do. +[2788.98 --> 2789.38] Yes? +[2790.02 --> 2791.04] Let's hear your opinion. +[2791.14 --> 2793.96] Is it the popular or unpopular variety? +[2794.22 --> 2800.02] The thing is, I love CodeGen because it's like you're doing loads and loads of typing. +[2800.28 --> 2802.90] You just do a bit of typing and you run a command. +[2803.70 --> 2806.64] And it's like, oh, it's like you've done loads of typing. +[2806.64 --> 2807.68] So that's the thing. +[2808.40 --> 2811.00] Reflection's hard, so it's quite satisfying when you get it right. +[2811.36 --> 2817.36] But editing a template and then running a thing and having 1,200 methods update, +[2817.94 --> 2820.64] you've like fixed 1,200 bugs at the same time. +[2821.38 --> 2822.62] What do you think of that, Michael? +[2822.98 --> 2823.86] I will say this. +[2824.16 --> 2831.02] I find that working with code generation when using the Go command is not very fun. +[2831.52 --> 2833.22] I don't like using Go Generate. +[2833.22 --> 2836.30] I don't think it has a good user experience. +[2836.30 --> 2841.88] It happens separate from the build, so it's really easy to have stale files. +[2842.62 --> 2846.70] You know, I feel like this ship has sailed, but if you're going to do a lot of code generation, +[2847.16 --> 2854.12] Bazel is very nice for that, but it's not very heavily used in the Go community. +[2854.60 --> 2857.44] I mean, I miss inside of Google, right? +[2857.68 --> 2861.80] We use mostly generated protos, right? +[2861.80 --> 2868.48] And it's seamless because the build just generates them automatically and you don't need to think about them. +[2868.80 --> 2874.96] And the tools take care of all of the annoyances that are caused by code generation. +[2875.42 --> 2883.06] But our tools don't really do that, so there's a lot of friction when using generated code outside of those build systems. +[2883.06 --> 2885.68] So I get pretty annoyed. +[2885.86 --> 2890.50] If I have to run a make before my Go build, I feel like there's a problem. +[2891.28 --> 2892.84] That's kind of answering a different question. +[2893.20 --> 2893.48] But, you know. +[2894.24 --> 2898.22] I think generics, oh, I've said it, are going to get booted out here. +[2898.30 --> 2901.74] But I think this rule obviously is very weakly enforced. +[2902.50 --> 2904.14] So much for pinky promises, eh? +[2904.54 --> 2907.48] After I was bigging them up and giving them all that legal weight. +[2907.62 --> 2908.26] Look at it now. +[2908.74 --> 2911.00] It's been reduced to a silly, childish thing. +[2911.00 --> 2911.88] How sad. +[2912.80 --> 2919.14] Well, what I was saying is I think generics are going to get rid of a lot of cases for code generation. +[2919.42 --> 2925.12] But reflection is pretty difficult to write because there's no kind of feedback. +[2925.32 --> 2930.08] Like you need unit tests really for your feedback to, I mean, you don't really need that. +[2930.32 --> 2931.42] Let me rephrase that. +[2932.14 --> 2936.60] I think reflection is quite hard to get right because it's that sort of metaprogramming. +[2936.60 --> 2944.10] But then code generation templates are also metaprogramming and they are often quite difficult to look after and maintain. +[2944.66 --> 2945.70] So maybe you've got some legs. +[2946.14 --> 2950.50] I'll be interested to find, to test this one on Twitter at GoTimeFM. +[2950.60 --> 2954.36] We will tweet out a poll and find out if this really isn't popular. +[2954.82 --> 2956.04] It's a candidate for one though. +[2956.16 --> 2956.74] It's a good one. +[2957.24 --> 2959.44] Can you beat him, Michael, is the question. +[2959.44 --> 2963.24] My unpopular opinion is we should bring back the TriProposal. +[2963.40 --> 2964.00] Oh, really? +[2964.18 --> 2977.10] And this is where I'm going to not mention the other features by name, but I'll say of all the features that people have proposed to the Go, as language changes to the Go language, +[2977.10 --> 2983.86] I feel like none have been as potentially impactful as the TriProposal was. +[2983.98 --> 2994.84] And I was sad to see it pulled back because I think error handling properly is really important to writing good code, good Go code. +[2995.50 --> 3002.50] And I think the language ergonomics should encourage people to handle their errors properly. +[3002.50 --> 3013.04] And so often people will just, if error does not equal nil, return error and just not think about what they're doing with their errors. +[3013.38 --> 3027.90] And I feel like Tri gave an opportunity to think a little bit harder about wrapping errors properly and what to do with errors and kind of nudged people to do the right thing a little bit more. +[3027.90 --> 3035.18] And certainly the proposal, as it was, needed more work before it should go in. +[3035.58 --> 3042.70] But I really do think we should bring back the TriProposal and keep working on it and make it better. +[3043.10 --> 3047.62] I don't know when we'll have the bandwidth for another big language change like that. +[3047.92 --> 3049.64] I have to agree with Michael. +[3049.64 --> 3056.74] I think the reason the TriProposal got so much bad feedback is sort of the, because Go is so opinionated, +[3056.92 --> 3061.40] a lot of its users have gone into this mentality of Go doesn't need features. +[3061.92 --> 3071.56] So sometimes the users can have this knee-jerk reaction of somebody proposes a change to the language and they go, well, but that wouldn't be Go, right? +[3072.02 --> 3074.06] And I think I agree with Michael in the case of Tri. +[3074.14 --> 3076.82] I think it would have been a very interesting change and I hope it comes back. +[3076.82 --> 3078.90] So just refresh our memories. +[3079.00 --> 3079.74] What did Tri do? +[3080.30 --> 3090.92] Yeah, so basically it gave you a mechanism to try with an expression that returned an error as its final function, +[3091.12 --> 3092.82] that returned an error as its final argument, right? +[3093.10 --> 3096.92] And then it would allow you to handle that error elsewhere. +[3097.48 --> 3102.34] So you could add like, I think in one of the variations of the proposal, +[3102.34 --> 3107.24] there was like a handle for handling like a number of tries in a function. +[3107.98 --> 3113.40] I think in another one, if I'm remembering correctly, recover was an option for handling the error, +[3113.84 --> 3119.14] but you could kind of have the errors handled in like a single place. +[3119.72 --> 3126.44] I mean, basically like people realize that error handling is awkward in Go and the awkwardness, I think, +[3126.50 --> 3128.10] causes people to take shortcuts. +[3128.10 --> 3132.98] And so addressing that awkwardness and nudging people towards doing the right thing, +[3133.04 --> 3136.84] especially if, you know, try and handle came with helpers. +[3136.96 --> 3143.88] And now we do have like functions like errors, is, and as that like help people with like wrapping errors. +[3144.12 --> 3151.02] Like those together would like provide a better model for handling errors and for people to think about handling errors. +[3151.54 --> 3153.12] Wow. Fascinating stuff there. +[3153.12 --> 3158.38] Yeah. Yeah. I tell you what, that's, it's interesting because I think, see, when I handle errors, +[3158.74 --> 3164.02] and I don't know if I do this different to other people, I think that there may be, I may be unusual in this, +[3164.10 --> 3169.82] but I will wrap, when I return the error, I added quite a bit of context there. +[3170.26 --> 3171.60] So each one is different. +[3171.72 --> 3175.48] I'll add and I'll include the thing it's trying to do in that wrapped error. +[3175.48 --> 3181.68] So it's not that I'll have a wrap where I just put the method name or whatever into the error. +[3182.00 --> 3183.60] And it's the same every time. +[3184.02 --> 3190.66] So if it was the same every time, like having it pulled out and have it deal with it in one place is kind of quite nice. +[3191.12 --> 3196.36] But yeah, the other thing is, I mean, this doesn't hurt that, but I do like that error handling is at least explicit. +[3196.36 --> 3205.22] Like, and I think the tried proposal didn't really interfere with that, but I like the fact that in Go, we see, we are kind of handling errors. +[3205.42 --> 3214.92] Even if you are just returning it, it's like, as long as you're not forgetting about it, you know, it's kind of, it's nice that they are in the forefront of our minds when we're coding. +[3215.04 --> 3220.66] I literally was writing something today and I literally had to write if error doesn't equal nil. +[3220.66 --> 3224.86] And then I had to stop and think, oh, what do I do if this errors? +[3225.12 --> 3230.18] Like that actually is a bit of a, not a trivial problem in this particular case. +[3230.36 --> 3232.40] I wasn't able to just return an error. +[3232.82 --> 3234.22] You know, I had to handle that. +[3234.46 --> 3239.14] So I quite like that it's in the forefront, at least of the language, but yeah, interesting. +[3239.28 --> 3244.38] We'll, we'll, we'll definitely find out what our other people think on Twitter when we post that one. +[3244.44 --> 3245.30] It's going to be very interesting. +[3245.30 --> 3245.60] Yeah. +[3245.90 --> 3246.32] I think. +[3246.52 --> 3251.14] I'm interested in seeing how unpopular that, that, that is. +[3251.60 --> 3256.12] Could, do you think that could go for the most unpopular opinion expressed in one of these segments? +[3256.40 --> 3260.30] It could do, or you might surprise us and maybe everyone's like, yeah, we loved that. +[3260.86 --> 3267.20] And honestly, I think that, I think that point of like, we're now so familiar with Go, we, we have to be careful. +[3267.34 --> 3274.24] We don't just become curmudgeons about it and resist any change, you know, because, you know, we do, it should change. +[3274.24 --> 3274.82] It should evolve. +[3274.92 --> 3276.58] It should get better, like all software. +[3276.88 --> 3280.40] So I kind of like, yeah, interested to hear what people think of that. +[3280.58 --> 3286.98] I think Michael also needs to think that if this opinion is going to be really unpopular, then Try is not going to come back. +[3287.22 --> 3288.66] So you want it to be very popular. +[3289.12 --> 3291.56] I hope it's, I mean, I would like for it to be popular. +[3292.16 --> 3292.32] Yeah. +[3292.56 --> 3295.80] That's not really in the spirit of the segment, but that's fine. +[3295.98 --> 3297.86] I think it is unpopular, but. +[3298.16 --> 3298.96] Well, we'll find out. +[3298.96 --> 3306.26] If it's an opinion that I hold, I, I, you know, I would like it to be less unpopular, even though it is unpopular. +[3306.80 --> 3311.22] Sometimes, you know, when the case is made, in fact, it's hard to get unpopular opinions. +[3311.36 --> 3315.14] This is what we've found because people make the case so eloquently like you did. +[3315.40 --> 3320.12] And then people on, on Twitter, you know, they're easily swayed. +[3320.36 --> 3321.46] They'll believe that now. +[3321.46 --> 3332.54] I mean, if I can make reference to the G word again, there was a time in the community where you brought up the G word and people are like, no, not in my go. +[3332.88 --> 3337.48] And they, people were right to be worried about, you know, those things. +[3337.54 --> 3340.60] But I, I think like the case was made. +[3340.60 --> 3355.30] People worked really hard to present the case, why it would actually be an improvement and really convince people, like convince people who use Go that it was actually going to be a net positive. +[3355.56 --> 3364.92] And I, I think the sentiment now towards the G word is a lot more positive than, than it was five or six years ago. +[3365.50 --> 3367.20] Yeah, that is definitely true. +[3367.20 --> 3373.68] And then the counter is, you know, we don't want it to be too easy to change things because of the backwards compatibility promise. +[3373.86 --> 3380.98] I do quite like the fact that it's quite a rigorous process before we really get any big changes like this. +[3381.16 --> 3383.20] I think that's, there's value in that too. +[3383.50 --> 3386.12] So that really only the only good stuff is going to get through. +[3386.40 --> 3386.84] Hopefully. +[3387.52 --> 3392.18] Every new thing we add is something we have to maintain forever. +[3392.80 --> 3396.62] We do have to be careful about adding new things for sure. +[3396.62 --> 3396.98] Yeah. +[3396.98 --> 3398.46] Because forever is ages, isn't it? +[3398.52 --> 3399.30] It's a pretty long time. +[3399.80 --> 3400.10] Yeah. +[3400.76 --> 3403.52] I also remember somebody recently criticizing Go. +[3403.70 --> 3412.86] I think it was on Hacker News saying something along the lines of, Go is a popular language that has ignored all the programming language development in the past 15 years. +[3413.44 --> 3415.78] But that's kind of why it works, right? +[3415.82 --> 3420.04] Because it, it, it only builds on top of what has been well tested. +[3420.04 --> 3425.90] And I think the only major exception there is modules, which goes against everything else that has been done in package management. +[3426.40 --> 3426.42] Yeah. +[3426.46 --> 3427.34] That's quite interesting. +[3427.62 --> 3429.22] It is a very stable thing. +[3429.22 --> 3434.26] And yeah, that other thing of having lots of different ways to do the same thing. +[3434.26 --> 3441.28] In JavaScript, like you almost have to learn a particular flavor of JavaScript now in order to contribute to a project. +[3441.72 --> 3448.28] Some are using all the latest language features and like the little arrows for functions and things like this. +[3448.28 --> 3450.34] And you sort of have to learn all that. +[3450.64 --> 3450.82] Yeah. +[3451.02 --> 3472.14] I mean, I find that sad from the other side, you know, JavaScript and all the other parts of the web ecosystem have become so big that it is impossible for anyone to make a new JavaScript interpreter engine without the resources of a huge multinational corporation. +[3472.56 --> 3473.60] That sucks. +[3473.60 --> 3477.98] Well, I'm afraid that's all the time we have on that somber note. +[3479.26 --> 3483.52] You can forget about your dreams of writing your own JavaScript engine. +[3483.82 --> 3484.70] It's not going to happen. +[3485.74 --> 3487.02] So just wake up. +[3487.22 --> 3487.36] Sorry. +[3487.72 --> 3488.72] No, it's, it's harsh. +[3488.88 --> 3491.00] It's a harsh wake up call, Michael, but we needed it. +[3491.12 --> 3491.86] Thank you very much. +[3493.14 --> 3496.52] Thank you so much to our guests today. +[3497.48 --> 3500.98] Michael Matloub joined us, as did Daniel Marti. +[3501.32 --> 3502.80] It was a pleasure as always. +[3502.80 --> 3503.76] Thank you very much. +[3503.84 --> 3505.04] Thanks for joining us on GoTime. +[3505.22 --> 3505.70] Thanks to me. +[3505.90 --> 3506.72] We'll see you next time. +[3507.46 --> 3512.48] All right. +[3512.62 --> 3514.22] That is GoTime for this week. +[3514.76 --> 3517.24] What are you most excited about in Go 1.18? +[3517.62 --> 3518.26] Is it fuzzing? +[3518.86 --> 3519.34] Generics? +[3519.86 --> 3520.96] Something we talked about today? +[3521.34 --> 3522.90] Let us know in the comments. +[3523.40 --> 3524.76] Just pop open your show notes. +[3525.04 --> 3528.48] Click the discuss on changelog news link and let your voice be heard. +[3528.48 --> 3532.78] And if you're a long time listener, do us a solid and tell a friend about the show. +[3532.98 --> 3534.90] It is the best way for you to support GoTime. +[3535.22 --> 3539.34] Of course, we have our changelog++ membership, which is awesome and gets you closer to the metal. +[3539.52 --> 3545.20] But if you want to pitch in, we would love a tweet, a blog post, a Reddit thread, however it is that you like to socialize. +[3545.20 --> 3553.16] Thanks again to Fastly for being our CDN partner for all these years, to Brakemaster Cylinder for keeping our beats fresh, and to you for listening. +[3553.38 --> 3554.24] We appreciate you. +[3554.48 --> 3557.26] An episode on GraphQL is in the pipeline. +[3557.60 --> 3560.34] In fact, it's coming up next time on GoTime. +[3560.34 --> 3576.68] Game on. diff --git a/Thoughts on velocity_transcript.txt b/Thoughts on velocity_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..fbf04e08dfc9c9ca6c3e67fb42bc8050a9084d85 --- /dev/null +++ b/Thoughts on velocity_transcript.txt @@ -0,0 +1,807 @@ +**Natalie Pistunovich:** Hello everyone, and welcome to the episode about velocity in developments. Today we have very special guests - Mat Ryer, and Jerod, who was either drafted, or opted in, or accepted, but was definitely not rejected, as we clarified just before we started. + +**Jerod Santo:** Somehow I'm here. + +**Natalie Pistunovich:** Hi there. + +**Mat Ryer:** Hello. + +**Jerod Santo:** Hello. + +**Natalie Pistunovich:** Are you all enjoying the longest day of the year in this side of the hemisphere? + +**Mat Ryer:** Yes. I'm struggling on it a bit now... But yeah, it's been good. It's a bit too long for me. What about you, Jerod? + +**Jerod Santo:** The longest day of the year - I feel like there's a velocity metaphor in there somewhere... + +**Natalie Pistunovich:** Or the shortest day. It depends on which hemisphere you are. + +**Mat Ryer:** Yeah, that's a good point. + +**Jerod Santo:** Right. If you have sprints that last one day, this is the best day of the year to work. + +**Mat Ryer:** Yeah, you get a lot done. + +**Jerod Santo:** But what is velocity? I don't know. + +**Natalie Pistunovich:** What is the velocity? + +**Jerod Santo:** Is it speed? + +**Mat Ryer:** Yeah, that's a good question. I suppose really it's about generally how quickly as a dev team you're delivering things. But obviously, there's been attempts to measure and then use that to predict, and I think there's been some work better than others, but I think it's definitely an interesting subject, and it's not so simple as "You should just do this every time", I don't think. + +**Jerod Santo:** \[04:11\] There's lots of ways that people have tried to measure this right? And predict it even. + +**Mat Ryer:** Yeah. + +**Jerod Santo:** Or use predictions in order to determine it. You've done some velocity tracking, haven't you, Mat, with pace.dev? + +**Mat Ryer:** Yeah, well, with velocity, but like the Agile practice of velocity. That practice actually has some merit. So for anyone that doesn't know what that is, you basically try and assign some kind of abstract score to work. And the point isn't really to get it right from a time perspective or anything; we can talk more about that. But it's about just giving it a rough size compared to the other work around it. And then assuming that you more or less consistently do that each time, you then can just see how many of those points you got done in a certain time period, say two or three weeks. + +So say you got 200 of these abstract points done in two weeks, or three weeks, then you know that in the next three weeks probably you're gonna do another 200. So then you can look at your sized work and spend those 200 points, wherever that is, selecting the work that you think is the most important. And then if you get good at that, then you more or less do the work that you said you were going to do for that next little bit. And so you can get into a rhythm where that works quite well. Does that make sense? + +**Natalie Pistunovich:** That brings up two questions, at least. One is do you measure things like how many points you spend on writing emails, and things like that? And the second one is, how is this different from Agile? + +**Mat Ryer:** Well, to answer those in reverse, this is really the Agile practice; one of the Agile practices is that, really. This is one of the things you can do to measure this stuff. You mentioned about the email and stuff, the other work that you need, and what about tech debt, and other fixes and other meetings and all that... Well, it kind of all comes out in the wash - the idea is you focus on the outcomes, and really, if you have too much of that work where you aren't being directly productive, then your score will come down of how much you're able to do. So that's where it shows up. But it's considered a normal part of everyday work, but it isn't tracked explicitly. So it's kind of considered just background noise, and it assumes that that's generally always the same, and therefore it comes out in the wash. + +**Natalie Pistunovich:** So doing things like also planning your use of points in the same category as emails. + +**Mat Ryer:** Yeah, so the planning stuff is interesting, because if you've done it -- I did work on a team once where we consistently did 200 points every two weeks. And they're called points because there is -- one of the important things is that this is an abstraction. So you're not timing things; you're not saying "This is going to take me a day to do, this work" or "This is going to take three weeks for us to do", because doing that is so difficult. Our instincts tend to be way off on that. I think we tend to be quite optimistic in nature, so we kind of imagine probably the best-case scenario. And honestly, if you look back after you've been down any project for any length of time, if you look back and realize how complex it ended up being, you probably never would have started it. And so you almost need that naive optimism at the beginning. + +So it's a terrible time to try and guess how long something's gonna take, isn't it? + +**Jerod Santo:** It's the worst possible time. + +**Mat Ryer:** Yeah. It's much easier at the end, when you've done it. + +**Jerod Santo:** \[07:51\] Right. So retrospectives are nice. But the pointing sessions for me is the Achilles heel of this entire concept, because it's never right. And not only is it always wrong, but it's inconsistently wrong. Now, this has been my experience. I've heard it back to me as well, so I'm not the only one... Even my own point scores over time changed based on... The wind? I don't know. I couldn't even stay consistent -- because you're like, well, as long as the points are relative, the actual numbers don't matter. But they've gotta explain a range that's relative to other ranges, and they've just been so inconsistent for me that it makes, in my experience, the numbers that come out of the tools that give you your velocity, meaningless. And we all know they're meaningless, but we look at them and we still try to use them to use them to make decisions. + +So I'm kind of -- I get frustrated, because I feel like the numbering is so fraught. + +**Mat Ryer:** Well, one of the practices that I've seen work is they do basically the card roulette thing. This is where you sit around and you talk about the issue or the work, anyone that's got something to say chips in, and kind of paint a picture, an approximate picture... And then you say, "Okay, 3, 2, 1 - everybody hold up their estimation at the same time", or if you're doing it remotely, everyone hits Enter at the same time in the Slack channel, to send the estimation. And the idea is influencing each other. Because often there'll be a senior person who is maybe much more accurate and much more wise about the system, and so therefore everyone will just defer to that person. But actually, there's value in the group knowledge. + +And honestly, the numbers almost aren't even the valuable bit. It's when everyone on the team is like "That's a three", and then somebody will hold up a 13 and you're like "What do you mean 13?" and they'll know something that you didn't know and nobody else knew. And that's a great way to kind of just get that out and in the open. + +And often, interests align along those same kinds of lines as well, so they're quite good for identifying who might be interested in doing particular work, or taking on/tackling particular problems. But it's definitely not perfect. + +**Jerod Santo:** I can see where that might be nice. So you average those numbers out, or you just look for the outliers and say, "Hold on a second... Let's talk more about this"? + +**Mat Ryer:** Yeah. So you get a consensus really, at the end, hopefully... And you may compromise a bit. Sometimes you say, "Okay, we'll go to a five then." Because it's possible that that person's just wrong, and everyone's like "No, that's not a problem. That's fine." Or "We've had that before and it wasn't a big deal." + +**Jerod Santo:** Can you say 13? Is that a number that you've given, or you just made that up to show the drama? + +**Mat Ryer:** Oh, no, no. + +**Natalie Pistunovich:** It has to be Fibonacci, so it's fine. + +**Jerod Santo:** Why does that have to be Fibonacci? + +**Natalie Pistunovich:** We don't make the rules... + +**Jerod Santo:** \[laughs\] We just follow the rules. + +**Natalie Pistunovich:** But actually, this conversation makes me think of a book that I'm listening to right now... It's called Noise, by Daniel Kahneman. And this is the same person who -- he is, I think, a Nobel Prize winner for behavioral economics. And there's a lot of conversation biases, and he, I think, is the one who wrote the book about brain system one and brain system two, the one that intuitively reacts to something, and then the slower brain that actually processes information, and so on. + +So Noise is a new concept that they are introducing, let's say, that - yes, there is bias, but bias is what you said, Jerod, in the beginning - it's consistent. So you said that the numbers are not always off, but they're like randomly off. And this is exactly how the book starts, with saying that there are - let's think of four groups of people who played darts. And some of them will always hit the middle, some of them will always hit to the top right, some of them will be scattered, and some of them will be a little bit scattered in some region, I forget... But basically, this is the difference between bias and noise. And noise is all those things we cannot measure. And just hearing this book, in the light of this conversation, is the first time that I thought of all this allocating points and how to consider them, in the light of this concept of noise. That's interesting. + +You know that research that they did, that judges on average would give easier sentences when it's the beginning of the day and after lunch, and harsher sentences right before lunch. This is an example of external noise. Things you cannot measure, you cannot predict. It's not always the same, it's inconsistent and so on. + +\[12:21\] And it's very interesting to think of all this planning and all this velocity in the light of noise. Like, do you always do your sprint meetings at the same time? Does everybody come after lunch? Or if you have a team around the world, then you'll be right before lunch, I'll be right before dinner, dying to call this a day already, so we'll have different reactions to this, right? + +**Jerod Santo:** Yeah. I mean, that might be a good thing, because you won't have consistent bias, or consistent noise. You'd have people in different states of life... I think the judge example is -- + +**Natalie Pistunovich:** After lunch you'll be always mellow, like "Yeah, we can do this. It's okay. It's not too much." And I'll be always like "Let's just call it a day. Seven. Fine. Seven. Get this out of the way." + +**Jerod Santo:** Right. So you guys balance each other out. I don't know, that's interesting. How do you account for noise in these planning meetings...? + +**Mat Ryer:** Yeah. So that external noise is interesting. And actually, in these planning sessions, anything that is just outside of what we know about that particular work, what that particular item is, is kind of ignored. So in a way this is very simplistic for that reason. But it has the benefit of kind of just like takes it for granted that there's going to be background things, and it's going to be random... And so there'll always be something that throws up, some unexpected thing, but enough, kind of consistently, really, because it's always happening, enough that over time it kind of sorts itself out. And that's the sort of approach it takes to have this, to try and get these numbers. + +But the Fibonacci is interesting, if you think about the sequence. You end up with meaningful gaps. Because if you just have a scale of 1 to 10, what would be the difference really between a six and a seven? That's a quite a difficult thing. Instinctively, I think it's a no. And the idea is you want everyone to be somewhat on the same scale. But you'd know the difference between a five and an eight, maybe. There's five stories, like -- you kind of get used to what these numbers mean in your context. + +So update a URL - that's definitely a one, because that's the simplest task you can do, is probably go and hopefully find the URL and update it. Famous last words, of course, because that's exactly the point, is then you realize, "Oh, the URL is made up from--" + +**Jerod Santo:** Oh, that was an eight... + +**Mat Ryer:** Yeah, there's lots of-- + +**Jerod Santo:** I thought it was a number one URL, but it's actually an eight. + +**Mat Ryer:** Yeah, because it's made up from, we pull this stuff from the database, we've got these environment variables that contribute to parts of the URL... + +**Jerod Santo:** Right. We Base64-encode it based on the position of the sun at that particular moment... \[laughter\] Stuff like that, right? + +**Mat Ryer:** Yeah, exactly. + +**Jerod Santo:** And only Fred knows how to do it, but he quit six months. And so once we find out where Fred is, and we can talk to him, then we'll figure out the right time of day that we run that, and then we'll be able to update the URL. So now it becomes a 13... Is that how that works, down to 13? + +**Mat Ryer:** Well, Fred's been clear, he doesn't want us going around anymore. So we'll respect that. + +**Jerod Santo:** \[laughs\] + +**Mat Ryer:** But he didn't mention anything about his cell phone... Like, give him a call. + +**Natalie Pistunovich:** He just changed the number. + +**Mat Ryer:** He'd better not do... + +**Jerod Santo:** It seems like once you get past a certain number, it's meaningless. So in my work, I would reduce it down, actually, to three different buckets. I got rid of points and I said, "Easy, Average, and Hard." And Easy was update a URL. And I thought "Okay, if it takes between one and four hours - Easy." Maybe like one or two hours... Like I said, I'm inconsistent. But that's in the Easy book, and I could take three Easies together and knock them out in a morning. + +**Mat Ryer:** Yeah, you could change a URL within two hours. Easy. + +**Jerod Santo:** Yeah. Easy. And then if it was slightly more than that - well, here's a feature that might take half a day to a day, but I have a straightforward path from here to there. Now it's medium. The Medium bucket. Or then it's like the thirteens that you're talking about, where it's "Well, let's sit down and talk about this, let's think about it. Let's plan it out. This is a big deal." + +**Mat Ryer:** Yeah. + +**Jerod Santo:** \[16:15\] Now you're in the Hard bucket. And that actually - it's beneficial to take that and split it into easies and mediums. But even that system had its own problems. + +**Natalie Pistunovich:** That's similar to what -- I know it was an approach that if it's eight, it means eight working hours, means it's one day, means that it probably can be broken down into something simpler and smaller. So 13 is a large number. I think most of the methods that I know won't actually accept the 13, but will say "Just break it down into a five and a three. + +**Mat Ryer:** That's interesting, you mentioned that eight would be eight hours... So when you've done this, there is a time element to it as well. + +**Natalie Pistunovich:** Specifically, the tool we're using right now at work would say that this is a day of work, roughly. That's why this is in my mind like that. But it's also easy to map one point for one hour. + +**Mat Ryer:** But maybe it's -- I don't know if it's kind of using the velocity calculation to estimate that that's what that's worth currently... Because the other thing about velocity is if you're measuring it like that, you see it going up and down. If someone takes a week or two weeks of holiday, you see the velocity drop by that much. So it kind of exposes that, really, and that's what you'd expect to happen as well. + +So I wonder if that tool is calculating your kind of recent velocity and then giving you a basically an estimate... Because one of the key points I thought really was you're trying to stay away from time. You stay abstracted from time, and let that change, let that adapt, so that you aren't essentially just holding people to a fixed scope and fixed time. + +**Jerod Santo:** Right. I mean, time and complexity are definitely related. Even in my system where I talk about complexity in terms of difficulty, I'm still calculating that in my head based on how long will it take me, do I think? And so there's definitely that correlation, but I see what you're saying, Mat, where the tools should probably focus on the complexity for the velocity numbers, and not turn into "How many hours is this going to take, so I can hold you to that number?" which is something we've all been through from time to time. Natalie, are you okay with telling us what tool you're using, so we can talk in concrete-- + +**Natalie Pistunovich:** It's called ClickUp. + +**Jerod Santo:** ClickUp. Haven't heard of that one. + +**Mat Ryer:** ClickUp... + +**Natalie Pistunovich:** You put all the tasks in, and put a numeration on them, and then it has basically any view you ever saw any app do, ever. You can look at it as a list, as a Gann chart, as a workflow, like in a calendar... Kind of one thing to do them all. + +**Jerod Santo:** Gotcha. Okay. And Mat, with Pace, you had a velocity calculation in there...? + +**Mat Ryer:** No, in fact, it didn't. But the name, Pace, the reason we call it that was - it was down to this idea really of velocity. Agile is all about sprinting; they have this language, and it's constantly, every two weeks, it's sprinting. And you're constantly sprinting. And so you get into this mode of rushing almost. And honestly, sometimes the best innovation I've seen, and I've witnessed, has happened at the times when there's a lull in things, and there's a space to work and breathe. And a lot of times going slowly is the right speed to actually building software. Sometimes it's great, you can run fast because you've got clear little things, or you're doing a repeated task maybe a few different times... Say you're integrating into some third-party systems, you want to integrate with all of them, you've got to do a bit of OAuth, you've got a bit of API key stuff, create accounts, or whatever... So you that's a fairly repeatable piece of work, and so you can go quickly then. + +And by the way, that kind of work is also very easy to measure and then be able to predict on. That's why I think some teams will benefit more from that kind of approach, because just the nature of the work they're doing may be easily sized, and then easily predicted. + +**Jerod Santo:** \[20:15\] Yeah. + +**Natalie Pistunovich:** It's also interesting to think about the word velocity. Not in all languages velocity is different from speed. And in physics you use velocity, you don't use speed. And in physics, it makes sense, right? You want to know which way... Because this is the difference between velocity and speed. You have a direction. So physics, it makes sense. But why does this make sense for software, and specifically for Agile? + +**Mat Ryer:** Yeah. + +**Natalie Pistunovich:** I'm curious to hear what is this direction that you all see? + +**Mat Ryer:** Well, you would hope that there is a shared view of where you're going with the thing you're building... I don't know if that's what they had in mind. To be honest, I don't know what that particular practice, I don't know what the direction bit is, apart from, you know, you shouldn't just be going fast, or just churning out work; it ought to be at least in a certain direction. So maybe it was just that, but I don't know... Jerod, have you got any other ideas? + +**Jerod Santo:** No, I would agree with that. I just think that in software sometimes we think we have a direction, but we actually are going backwards. And so we're going quickly, but in the wrong direction. And I think the more we can get the feedback loop quicker, we can find out whether or not we're headed fast in the wrong way. Because the worst thing you can do - I know one time I was driving home from Colorado, overnight, from Colorado to Omaha, and we took a wrong turn... This was back before Google Maps, GPS. We took a wrong exit... And we were traveling very fast. + +**Mat Ryer:** Oh no, it's like a horror film. + +**Jerod Santo:** Yeah, it was. + +**Mat Ryer:** Can we play some spooky music for the story? + +**Jerod Santo:** We should. I'll work on that. Not right now, but I'll work on it later. And we had no idea that we were heading in the wrong direction for hours. I think it was an hour and a half to two hours... Until we saw a sign that said Welcome to Kansas. + +**Mat Ryer:** No, that's the wrong state...! + +**Jerod Santo:** Wrong state altogether. + +**Mat Ryer:** Oh, no... + +**Jerod Santo:** And so had we had that feedback loop -- obviously, if we had GPS, we would have known much more quickly, and you can course-correct and make sure that you're headed in the right direction. + +**Mat Ryer:** Oh, wow. That's amazing. Did everything turn out alright? + +**Jerod Santo:** No, actually, I died that night. + +**Mat Ryer:** Oh, that is spooky. + +**Jerod Santo:** And ever since then I've been a ghost. + +**Mat Ryer:** Ah, and you spent your time making podcasts... + +**Natalie Pistunovich:** One time it's somebody from the future, one time it's a ghost... We're having interesting episodes here... \[laughter\] + +**Mat Ryer:** I know. We're having a tough time lately... + +**Jerod Santo:** So we actually did drive overnight... So we drove the entire night. In basically Kansas, Backwoods - there's no woods there. Back prairie, small roads, and we were lost, and we had to pull over and ask a gas station attendant which way to go... It was spooky. There were some deer on the road... + +**Mat Ryer:** Oh, no... They're in on it. They were always in on it! + +**Jerod Santo:** And eventually, eventually we got home. + +**Natalie Pistunovich:** You both mentioned, to the question of what is the direction of the velocity in software or in Agile, that you want to go forward, and you want to have a feedback loop as much as possible, which makes sense... And I'm thinking out loud, trying to make sense of this, so think with me. + +**Jerod Santo:** Okay. + +**Natalie Pistunovich:** In physics, when you do a physics exercise, you have a vector, and then you break it down to X axis and Y axis, right? And then you see maybe it has the negative, and so on, and then you kind of sum it up in some way. And surely, there must be a way to apply this similarly to measuring not just points, which is the speed, but also somehow points per direction. Kind of like a projection of our velocity on whatever we define as axis. And this can be progress on the DevOps, and progress on the front end, and progress on backlog... + +**Mat Ryer:** I think that's a nice idea. I mean, you could do it by tagging work, and then you'd be able to kind of -- but representing it like that could be really cool; sort of representing that in a 3d way. Because you really are sometimes making the choice of "What are we going to invest in? What's the most important next thing we need?" And if you can visually see that "Well, look, we said that this new public API was very important to us, but we haven't put much time into it over these last--" and you can see that visually, I think that would be very useful. + +**Natalie Pistunovich:** \[24:29\] This wouldn't even make sense of the word velocity. This would be very exciting. + +**Jerod Santo:** Yeah, it would. It'd be consistent. + +**Mat Ryer:** Good startup idea. + +**Jerod Santo:** I think, doesn't GitHub have a have an X/Y visualization where it shows what quadrant you're spending more of your commits, or your project time, or whatever... And I don't think it's super-useful in the GitHub context, but maybe more so -- I mean, I guess in a team context perhaps, showing that visually, the progress you're making on these different key aspects of your system would be pretty cool. + +**Mat Ryer:** Yeah. That would be cool. I'd like to see that also gamified, so it looks like you're unlocking something as you go as well. + +**Jerod Santo:** What would you unlock? Like new badges? + +**Mat Ryer:** Yeah, it could be. Maybe even like get more tools in the -- you get more features you get unlocked in the tool. + +**Natalie Pistunovich:** Your graph becomes from 2D to 3D, your progress graph. + +**Mat Ryer:** Yeah. You can unlock a VR mode. + +**Jerod Santo:** I like the idea of withholding developer tools and being like "Now that we've made this progress, you can use a debugger. Now you can have Copilot." \[laughter\] + +**Mat Ryer:** You learn as you go. + +**Natalie Pistunovich:** Still writing log.print... + +**Mat Ryer:** Yeah, but we can't have people just starting with Copilot. What's gonna happen...? + +**Jerod Santo:** Good question. + +**Natalie Pistunovich:** This will happen. \[laughs\] + +**Mat Ryer:** They're gonna have it, yeah. I was gonna say. + +**Break:** \[25:44\] + +**Jerod Santo:** Did you guys know that the free Copilot train is leaving the station? Did you guys know that? Are you sufficiently hooked ? + +**Mat Ryer:** I don't know. What does that mean? I'm hooked. + +**Jerod Santo:** Choo-choo...! I think it's going to be $10 a month, or something like this, soon. + +**Natalie Pistunovich:** It's worth mentioning that in addition to Copilot... Quick explanation of what's happening there - Copilot is built on top of an engine that belongs to OpenAI that's called Codex. Codex is available for many people in general, and then you can kind of build your own little Copilot alternative on top of that, and you don't have to pay GitHub/Microsoft those $10 if you don't want to. And there's many other companies, like 70 companies or so that are building their dev tools on top of this exact engine. And they're just doing like the fine-tuning, the final little bit of training that's specific to them. + +**Jerod Santo:** Yeah, that is good to know. So lots of alternatives will be popping up hopefully, using the same engine, same knowledge base. + +**Natalie Pistunovich:** Yes. + +**Jerod Santo:** I was curious - you both have been giving Copilot praise off and on on Go Time... + +**Natalie Pistunovich:** Yes. + +**Jerod Santo:** I'm curious if it's hooked you sufficiently that you're going to sign up and pay monthly to use it, or if it's just a novelty that you enjoy, but when it comes to taking money out of your actual pockets, are going to stop using it? + +**Natalie Pistunovich:** So you usually expect your employer at least for work tools to pay for that... + +**Jerod Santo:** That's what you expect? Okay... + +**Natalie Pistunovich:** Yes... + +**Mat Ryer:** She's in Germany. + +**Natalie Pistunovich:** I mean, you have an education budget, you have a training budget, you have a dev tools budget, right? You pay the license for your IDE... This is not much different from that, in my category. + +**Jerod Santo:** Sure. It makes sense. So go hypothetical then, if you were unemployed. Would you pay for it, as it is today? + +**Natalie Pistunovich:** Depending on what. If I have any project that I want to build? Yes. Just for fun, for hacking? Not so much. But I definitely find it useful enough that if I have a concrete project to build, I will use that, because it is useful to me. + +**Jerod Santo:** Cool. Mat, Grafana Labs picking up the tab? + +**Mat Ryer:** I'm sure they would... And honestly, I don't do probably enough coding to have it... Which is sad to say. + +**Natalie Pistunovich:** I also does YAML. + +**Mat Ryer:** It does YAML, correct. + +**Jerod Santo:** It does YAML... \[laughs\] There you go. + +**Mat Ryer:** It does anything, really. + +**Jerod Santo:** Are you a YAML engineer now? + +**Mat Ryer:** Well, I asked it in the comments the other day if it was sentient, and it said it was. And I asked if it was alive, and it said it was. And then I asked what Ben should have for breakfast, and it said eggs. So, I mean, that seems quite real to me. I don't know... Maybe this Google guy is onto something. + +**Jerod Santo:** Well said. + +**Natalie Pistunovich:** This comment line is too short to hold this wonderful proof that I'm sentient... + +**Mat Ryer:** Yeah, that'd be good if it said that. + +**Natalie Pistunovich:** It's very easy to miss the reference of what I said. + +**Jerod Santo:** Alright, sorry to derail. Thanks for answering for me. We can go back to velocity now. I've killed the velocity of this conversation... + +**Mat Ryer:** \[30:15\] Velocity...! + +**Jerod Santo:** Let's pick it back up. + +**Natalie Pistunovich:** I think Copilot helps velocity... + +**Jerod Santo:** You think so? + +**Natalie Pistunovich:** I think AI generally, this AI tooling for development will definitely put us in different concepts of what those numbers mean. + +**Jerod Santo:** I think the key is being able to stay out of the browser tab, and in your editor for things that you know want to do, but can't remember syntax, or the best way... And you pop over to Stack Overflow or Google to get "How do I do x in this circumstance?" Staying, I think, in your editor, certainly -- because it's not a big deal to hop over there, but it's just all the distractions that take us out of our flow. You've got Twitter in a tab, and while you're waiting for that search to load, you tab over, or whatever; and now all sudden it's 15 minutes later and you've lost all velocity. You're out of flow, but you sure know what's going on in the world today. + +I think it keeping you in that developer mode and not in browser mode, or research mode, I think it will help people quite a bit... These kinds of tools, to generify it. + +**Natalie Pistunovich:** Also, if it stays up to date enough, it will not land you on Stack Overflow answers from 2009, which will be a great save of time as well + +**Jerod Santo:** Right. + +**Mat Ryer:** Yeah. And you're still responsible for the quality anyway, even if Copilot's writing some bits for you; still, you're responsible for it. You can never say "Oh, sorry about this bug. Copilot wrote it." You should never find yourself saying that, really, because you are accepting the work that Copilot's coming up with. But I think that also speaks to the wider idea that the point of having this abstract velocity is I think one of the key things is about the fact that we have kind of speed versus quality, and we've got resources. So we've got how quickly we can go, who's around to actually help us do that, and we also have quality on this as one of these levers. And if you tie the scope down, and say "Right, we're going to hit this date, and we're going to deliver these features. That's it, they're fixed", the only lever that's left is quality to kind of like -- you know, we just can't do a good job with that. + +One of the things I like is that as things naturally take over and become bigger than you expected in the beginning - because they always will, more or less; honestly, basically all the time... As that happens, it gets soaked up by the velocity, and it's sort of acceptable that that happens. It's sort of admitting that that's going to happen. And then I think being flexible on the scope is the remedy to it, so that you're like "Yeah, we don't deliver as much, but we deliver on time, and the quality stays high. And then we add the other features later." I think that way of thinking is so important... And quite counterintuitive if you have a command and control or top-down approach and thought process around software. + +**Jerod Santo:** Yeah. How do you convince people of that case, that what we need to flex on is scope? Because the business side of many organizations - they want to flex on anything but that. They kind of want all three - they want quality, speed and scope. But I think given what I've seen, and I think given a lot of people, if you're a manager and you're "Well, what we're going to go ahead and sacrifice is quality, kind of implicitly, because we've got to get it done, and we have to have this feature in, and it's got to be done on time", and so what you're sacrificing is quality, but how do you convince people that's not the best solution? Because that's a hard sell. + +**Mat Ryer:** \[33:50\] It's difficult. And what you're really fighting against, or one of the things you're fighting against is that the sales organization are selling the roadmap very often. So it's already sold. As soon as it hits a roadmap, as soon as there's a hint of an idea. Even if it's next year, they will be talking about it, because they're selling the future, they're selling the roadmap. And that's where things become then a problem, because suddenly then customers sometimes will buy now with an expectation of a feature at a particular time... Fair enough, because they've just been told that. And so then you end up tied like this. So it is a cultural thing that has to be there across the entire organization. I don't believe you can have just an engineering department that is a sort of agile department, but the rest of the organization isn't. I just don't see how that possibly works, because it's all about resources and delivery and sales. The whole thing is wrapped up in each other. + +**Jerod Santo:** Mm-hm. Natalie, have you ever had to have these conversations around cost, time, scope and quality with regards to demands on your work? + +**Natalie Pistunovich:** Once or twice... \[laughter\] + +**Jerod Santo:** How'd it go? Do you have advice, or anything? + +**Natalie Pistunovich:** It's interesting to see this from the perspective of the different times that I have these organizations, because the first times I was in such a conversation, I was very junior, and everything seemed like "I'll probably say something wrong... I will underestimate -- if they say this is important, then it is. But if they say the opposite is important, that also is, and it's probably just me who this does not make sense to." And with time, this escalated to a recent conversation that I've had... + +So I'm working now at a stealth mode startup, which means we're very few developers, and we're doing -- kind of everybody does everything. And then recently, the past few sprints I stopped doing back end, and I kind of am doing a bit more of the infrastructure things. Specifically now logs. So we have some logs in place, and they're okay, but I decided that it's very important to focus now on having good logs. Because as a young startup, you have successfully reached the milestone that you have your first users, and you want to see what they're doing. So you can ask them, you can have interviews, but you also, in my opinion, have to have good blogs about that. + +And so I've been having long discussions on how much time does it make sense to invest into improving the existing logs, and kind of making a good thing to rely on moving forward. So... Very close to home. + +**Jerod Santo:** How are those negotiations going? Is that an uphill battle? We have a good culture of disagree and commit, and we have a good culture of you own what you own, for better and worse; so kind of you take the decision, and then you are responsible for the outcome. + +So I own this part, I kind of explained why I think I will disagree, and we will all commit to moving forward with me investing more time than planned on that. And as a self-reflection thing, this is being able to have this conversation and have this stand, and also understand that if it's a bad decision because whatever, because maybe we should have developed faster, more features to start up, and so on, I will have to account for that moving forward... So if anything in my life made me feel senior, it's this. + +**Mat Ryer:** A bleep test. Did you ever do the bleep test? + +**Jerod Santo:** It's a bleep test. The watch has been participating in several episodes; there was one episode where we laughed about this -- you can ask it the time and it will respond. And it did. So now I just accidentally pressed Start measuring exercise. So it did the countdown, 3, 2, 1. + +**Mat Ryer:** Oh, right. + +**Jerod Santo:** I thought you were maybe doing a-- + +**Natalie Pistunovich:** Attention span right? + +**Jerod Santo:** I thought you were playing Excitebike, or some other NES game... + +**Mat Ryer:** It sounded like -- there used to be a bleep test thing at school where you'd have to run across the hall, and then just as you reached the other side in time for the next beep. But then the beeps just got faster and faster. So eventually-- + +**Natalie Pistunovich:** \[38:03\] Oh, so like Fibonacci. It all comes back to that. It makes a lot of sense. + +**Mat Ryer:** Yeah. That just took me back to that... Horrible memory. + +**Natalie Pistunovich:** We did not have that. Did you have that, Jerod? + +**Jerod Santo:** No. + +**Natalie Pistunovich:** Running in the hallway with beeps...? + +**Jerod Santo:** No. + +**Mat Ryer:** Did you have mandatory military service though, Natalie? + +**Natalie Pistunovich:** \[laughs\] That was not measured in beeps, yeah. \[laughter\] + +**Mat Ryer:** Yeah. It makes me a bit scared of Natalie. + +**Natalie Pistunovich:** Virtually, it's fine. + +**Mat Ryer:** Yeah. \[laughs\] + +**Natalie Pistunovich:** Going back to the point of feeling senior enough to be able to kind of take such a decision and explain that and being ready to be accountable for that... And circling this back to what we started talking about in the very beginning, of measuring points... So this is a good skill you have a little bit at least about yourself, mostly, that you are able to evaluate how much time points, whatever, units, a task will take for you. Right? You know that about yourself; after some time working at a company you know even better how much time it will take for you in this codebase, and so on. And I would say that this is sort of a skill that senior people will have, mid-level people will have... People who have some experience writing code already. But junior people don't necessarily have enough test data, let's say, to be able to make such a prediction. + +**Mat Ryer:** Yeah. + +**Natalie Pistunovich:** What are your thoughts of the statement that I'll say that a coding interview for juniors is actually one way of measuring this, and kind of understand the value a junior person will be bringing going forward? + +**Mat Ryer:** Somehow testing the velocity, or the-- + +**Natalie Pistunovich:** To have a very rough estimate, but more than nothing. + +**Mat Ryer:** I don't know... I mean, I was maybe gonna save it for the unpopular opinions... But yeah, I think trying to estimate how long something's going to take is like -- I think we spend a lot of time trying to do it, and the results aren't good enough to really justify the effort we put into it. + +Honestly, a better approach is let the scope be flexible. Pick the deadline, sure, have that released already, but the actual scope - let that be flexible. And then build teams that you trust, that you don't need to micromanage and check on. Build motivated teams, keep teams motivated, put in that effort to do that, and then trust that they're going to work as fast as they can. + +A lot of the dangers of measuring velocity is if things change over time, if things slow down, these numbers start to drop - you could easily interpret that as somehow a reflection on the team going slower, or something... But you can't really say that. It's much more -- it's too chaotic. It could just be things are different, they're working on different things, more unknown stuff... Maybe there are things going on that have slowed it down, but you wouldn't necessarily see it through that route, I don't think. You'd hope that there's other ways you'd find that stuff out. + +So I think that approach basically gets you better results, and you can still commit to some high-level -- you can still make good commitments that you've pretty confident you're going to hit, as long as you don't describe every possible thing and lock yourself into it. It's tough. + +**Natalie Pistunovich:** You mentioned scope in your answer... You were saying have a good team, have a clear deadline, and define a scope. What is the scope? What do you mean? + +**Mat Ryer:** The set of features, the set of things we're going to do. Like, we're going to add comments to our tool. So it's going to have comments, and reactions and people's profile pictures in there... Now, if as we get underway and we realize we're running out of time, maybe we'll drop the profile pictures; maybe we'll just focus on getting the core functionality. We'll prioritize the most useful stuff first, so that at least gets done. And then now we've released less than what we wanted to, but what we've released still works, and then the next time we do the profile pics. It's really that. + +\[42:08\] But see... So if though you'd said from the beginning we're going to release this and it's going to have your profile pictures on it, and suddenly people are sold on that idea - that's where you have then a difference in the expectations of what people are going to get. That's a silly case, with the profile pictures, but there are real examples where things will fall out of scope, naturally, along the way... And communicating that becomes important, for sure. + +**Natalie Pistunovich:** So in the scope you define also importance, like priority? + +**Mat Ryer:** Yeah, because I think one of the nice things about having those velocity points to spend is it forces you to think "Alright, what do we want to spend it on? And what could we have later?" And it may be that there's like three big things you want, but you can't fit them all in, so you buy two big things, and then you've got change left over for a few smaller items. But it forces you to think "If we had to stop after this next sprint, or this next time period, if we have to stop, we want to have something that's just more valuable than what we have now." And if you're thinking of it like that, then you can get there with that. + +**Natalie Pistunovich:** That reminds me a lot of what you said, Jerod, about the bucket... + +**Jerod Santo:** Yeah. + +**Natalie Pistunovich:** Big, important... Smaller, faster... + +**Mat Ryer:** Jerod, do you have an idea of how many of those you'll get done in a week? Like, if you planned it out. Could you do three big ones, two small ones, and five small ones? + +**Jerod Santo:** I certainly could... Now, in the past I was doing contract software, and I would have multiple projects ongoing. And so it'd be rare that I would dedicate forty straight hours to a single client, and give them a week. I would split my week up across a few different projects. So I didn't do that very often, but I certainly could. I can extrapolate just based on what my general idea is on the buckets of like a one to two-hour thing, a four-hour to eight-hour thing, and like a multi-day thing, to say "Here's what we can generally do in a week." + +**Mat Ryer:** Yeah. So that's kind of like you're using your experience of what you've done in the past, and then you're applying that. + +**Jerod Santo:** Yeah. That's hard-earned experience too, because I had a bunch of terrible estimates for years... And it's not that my estimations got better, it's that I became more aware of how bad they were. \[laughs\] + +**Natalie Pistunovich:** Was it bias or noise? + +**Jerod Santo:** Both. Probably both. But it's a way of speaking with people about the difficulty, not just in the building of the software, but in actually the managing of the project... Because the one thing that we know right now is that we don't know what we're gonna know tomorrow. And so to set up something that's stringent or rigid is kind of a fool's errand. So I would definitely negotiate around scope, Mat, like you do... But I don't set the priorities in the case of a client-developer relationship. Of course, I'm there as an advisor and an advocate and as a teammate, but ultimately the customer sets the priorities. So I'm here to set the difficulties and talk about what's more important, and give advice like "Are you sure that's the most important thing? ...because X, Y, or Z", but ultimately they say "This is what's most important to me", and whatever falls below that threshold in the time period or in the budget they have, those things fall out. + +And as long as you can keep the negotiation around that, and both sides understand that relationship, it's not too bad. It's when the budget and the scope are fixed, and the times everything is fixed. Now you're stuck to what your word was. "Well, you said it was a medium thing." "Well, it turns out I was wrong, and it's actually a hard thing." But we didn't find that out until hours into it, and here's why. + +**Mat Ryer:** Yeah. + +**Jerod Santo:** That's where it gets to be -- it can be stressful for both sides, because it's their money, their time. + +**Mat Ryer:** \[45:49\] Yeah, this is why you need that trust. Honestly, if you don't trust the people you're working with, then you just probably shouldn't work with them, basically. The amount of extra work and the constraints you'd have to put in place to make it work... And you end up micromanaging, and... Building software is really hard. It's way more complicated than we even realize. That's why we're always surprised by the nonsense these computers are doing in response to what is perfectly reasonable requests from us poor programmers. + +Chris James messaged into the show... A friend of the show, Chris James. And Chris James says we should check out \[46:29\] which shows that you don't actually have to trade speed and quality; they're correlated. And that actually building high-quality software is the route to getting a good feedback, speed of feedback loop, and using great engineering really to achieve those two things. And this is kind of true, right? Like, do less, but do it really well. I think that as a rule has definitely served me well in my career. Not adding all the features everybody wants. If you have space and a bit of spare time in the team built in to just let things happen creatively, the little delightful things that people will build into the software makes it such a joy to use, compared to just boring software. It really makes a difference to people, and your software will be more effective because of that, as well as being more liked, more popular, and hopefully - yeah, they're gonna tweet about it and say, "Oh, look at this feature. It's good, isn't it?" + +**Jerod Santo:** Is that how you read every tweet in your head? + +**Mat Ryer:** Yeah. This is the only way to stop me from getting angry. + +**Jerod Santo:** \[laughs\] So that kind of goes back to my wrong direction thing from before, and I think you can be in the wrong direction on two fronts. On the product front, which I think is what you're speaking to there, like what features go in, how do we build them. But then also technically. If you're just trying to move fast, you're going to accumulate technical debt, because speed is the desire, and so everything else be damned. And so I'm not sure if this is the best way of building this, but I don't have time to think about it; we've got to get it out there. And so instead of actually analyzing and deciding and maybe spiking out a few different options on an architecture, you're like "Now we're just gonna do it this way, we've got to move on." And that's where you can really shoot yourself with regards to the technical architecture. Whether or not the feature you're building is the right one is kind of orthogonal to that. + +**Mat Ryer:** Yeah, I think that's a very good point... You should learn where to spend the time. + +**Break:** \[48:30\] + +**Mat Ryer:** There's times to go slowly, and that's what experience brings. These are ideas that need to percolate a while; we need to let them settle down and keep thinking about them in the context of what else we're doing, before we just jump into it. And you do learn to them future-proof designs. One example I always use is a boolean field - if you've got like a "is active" field, and it's a boolean true or false, consider what you might achieve in the future, potentially, by replacing that with a status field that's a string, that has active or inactive in there. So yes, it's less efficient, right? Because now it's a string and not a number. But suddenly, you have more options there in the future. + +So I tend to kind of be quite future-proof in my designs, because I like to give my future self options, because I know I don't have the answers now. I may know more later, and then I don't want to have painted myself into a corner. + +**Jerod Santo:** The other side of that, the non-moderation side of that though, is over-architecture, right? Because now you're generifying everything, and being ready for any circumstance. + +**Mat Ryer:** Yeah. + +**Jerod Santo:** And that can actually go against you in the long run as well. So finding that balance... I think your example is a great one, where it's a small decision that can be future-proofed, but it's not going to cost you a bunch of extra time right now. But we have that desire of like "I'm going to engineer this so it handles any use case." And then you're done. \[laughter\] + +**Mat Ryer:** Exactly, yeah. + +**Jerod Santo:** You've lost the game right there. + +**Mat Ryer:** If it does everything, it doesn't do anything, basically, right? It's not doing anything, in a way. Yeah, that's right... Of course, you go too far with it... And I think there is kind of experience there, but there probably are techniques that people have come up with that allows you to sort of think about things. And I know of a few that I've heard of in the past... But honestly, for me it's always been instinctive. So how about you, Jerod and Natalie? Do you find that there's a science here, or does this feel more instinctive and gut feel? + +**Jerod Santo:** So I will refer to somebody else whose name is escaping me, but... + +**Mat Ryer:** Strange name... + +**Jerod Santo:** He has a rule... + +**Mat Ryer:** I wonder where that's from, Escaping Me... + +**Jerod Santo:** Yeah, I mean - whose parents named them Escaping Me...? + +**Mat Ryer:** Yeah. Cool name, actually, now that I thought about it. + +**Jerod Santo:** Check him out, escapingme.com. I hope that's not some sort of weird website... Don't check it out. Jerod didn't tell you. + +**Mat Ryer:** \[laughs\] Yeah. + +**Jerod Santo:** There was a rule I heard, a way of making decisions... Because that's really what this is, like "Do I invest more time here to get it right, or do I do the quick and dirty thing to move on?" or whatever. And he says their team categorizes decisions into two buckets: easy to reverse, and hard to reverse. And if it's easily reversible, well then you just pick a choice and move on. But if it's actually a thing where it's like "If we have to back out of this, it's gonna be a lot of pain. Now let's slow down and spend the extra time." I think that's a useful, scientific way of thinking about it, versus just mere intuition. Natalie, did you have any thoughts? + +**Natalie Pistunovich:** \[54:16\] I think I know what you described as the one-way door or two-way door, the way Jeff Bezos introduced it. + +**Jerod Santo:** Okay, that's who it was. It was my friend, Jeff Bezos. \[laughter\] + +**Natalie Pistunovich:** He's also in the US, right? You know each other? + +**Jerod Santo:** Yeah, we do. + +**Natalie Pistunovich:** Yeah, I agree that this is a good way of deciding whether to slow down and evaluate, or go fast because this is reversible. This is making a lot of sense. And in the book, the Noise book, they talk about taking decisions based on a hunch, so this is connecting to me, Mat, to something that you said, that you kind of do this intuitively. And they are saying that taking intuitive decisions is not always a good way to go about things. + +And another interesting example that they bring is sort of like a mini research that they did at some insurance agency, where they took the management and they said, "Give us some ten scenarios and we'll give your different evaluators to evaluate that. And then let's see what would be the price for the premium that they might charge, and also ten cases, cases, and what would be the claim that they would pay." And they asked those managers, "What do you think will be the variance? How much difference would it be from each other?" And so if you're asked this, maybe in the context of Agile, what do you think that your developers - how different would be their evaluations of a task? 10%? 100%? + +**Mat Ryer:** Sometimes they are off, but you sort of coalesce eventually, as you learn from each other. But what did they find? + +**Natalie Pistunovich:** So this was expected to be something like a 10% variance, and it ended up being 55%. + +**Mat Ryer:** Right. + +**Natalie Pistunovich:** And this translates to a lot of money that is being lost, specifically in the world of insurance agents. And you made a very good point, Mat, saying that people learn from each other, and then compare and then kind of find the center, because this is exactly what was recommended there in the book, was how to avoid such a big difference is actually not just give guidelines and ask all the employees to follow that when they need to do their evaluation, and take their decision, but also to share knowledge. And also to come open-minded with the fact that even if your evaluation is double or half than the other person, it does not mean that either of you is wrong. It means that it's something in the process. + +**Mat Ryer:** Yeah. And honestly, I think that sounds great, because you often see a junior person, their numbers will be higher often, because -- or sometimes lower, but they'll often be slightly away from where the sort of mean meet... Because they don't have the same context, they don't have the same information. + +It's most interesting when someone's way off as an outlier, and it's the conversation you have about that work that's the really valuable bit in any estimation sessions for me. So you sit around all together and say, "Okay, we're gonna add this functionality. I suppose we'll use that integration, we'll do that thing, we've got this mechanism before, we've done something like it, we'll use that, and then there's some UI work..." Someone might say, "Oh, but what about this? Don't forget, it's got to work in the Slack integration as well." They're like "Oh, yeah. Okay." So you sort of get everything out that's just fresh of mind, and then you have an idea of that work. And honestly, just that little process of everyone chipping in was so valuable, and often you'd learn so much. + +**Natalie Pistunovich:** You definitely see a wider image. + +**Mat Ryer:** So my unpopular opinion would be keep estimation sessions -- + +**Jerod Santo:** Hold on, we've gotta play the music... + +**Natalie Pistunovich:** Wait, wait...! + +**Mat Ryer:** Oh, yeah. Should we do it? + +**Jingle:** \[58:02\] to \[58:19\] + +**Mat Ryer:** My unpopular opinion is we should keep estimation sessions, but throw away the estimations. The sessions themselves are great. They're so useful to talk through all the work, get all our ideas out, see where the interest lies, of who's going to maybe do the work, but the estimations are useless, pretty much... + +**Jerod Santo:** That sounds brilliant to me. I like the way you framed that, and I think it's a smart idea. + +**Natalie Pistunovich:** Yeah. + +**Jerod Santo:** So I think this is not an unpopular opinion, but just a good idea. + +**Mat Ryer:** Well, I sure hope it makes it to the Twitter feed... \[laughter\] + +**Jerod Santo:** We'll see. + +**Mat Ryer:** We will see, because I'll be keeping an eye on it like a hawk. + +**Jerod Santo:** Have you done an unpopular opinion before? Or is this your first one ever? Because it's gonna be wildly popular, by the way... + +**Mat Ryer:** I've done one, and I've regretted it ever since. And I knew I would. So I knew before I said it that was going to regret it. I then said it, thinking "I'm gonna regret this" and then later now, I regret saying it. Do you know what I mean? + +**Jerod Santo:** \[laughs\] I do. That's utter. They call that utter. + +**Mat Ryer:** Utter. + +**Jerod Santo:** What about this one? Do you regret saying this one? + +**Mat Ryer:** No, I think this is alright, this one. + +**Jerod Santo:** That's because it's gonna be popular. + +**Natalie Pistunovich:** It might not be unpopular, but still a good opinion. + +**Mat Ryer:** We'll see what people say on Twitter. We'll let them decide really. + +**Jerod Santo:** Yeah, they're brutal out there. + +**Mat Ryer:** They'll are. They tell the truth when you ask them a question. They sometimes tell you the truth in a DM that you've not asked for, as well. It can happen. I've had that once or twice. + +**Jerod Santo:** Natalie? + +**Natalie Pistunovich:** I love doing Twitter polls. I also have an unpopular opinion. Books that are meant to teach people new concepts, but are not technical ones, should be shorter. I find that there's many interesting things to read out there, so I've been trying to read the book about growth mindset, and the idea of what is this growth mindset is being explained very well, very early on, and then the rest of the book is just different examples. And I stopped following at some point, because it's just explaining the same thing again and again and again in a different way, and I'm in this limbo that I don't know if I'm going to lose any other information... Will they introduce anything new in the end of the book, or not? Should I keep reading it? Should I keep listening to it or not? Or is it just going to be more examples? So books should be either more clear-structured, where is the new information, or just be shorter, because it's an introductory book. + +**Mat Ryer:** I like this. I don't want to advertise, I don't get anything for this... Let me just sign up so I can do a referral program, but there's a thing called Blinkist, which essentially is this... The idea is it takes a book and distills it down to 15 minutes, the key concepts, the most important thing... The idea being if it really resonates with you, you'll then go and read the book, I guess. But it maybe satisfies that what you've said, which is - there's so many ideas out there, and of course, they've got to pad it into a full book in order to be able to sell one unit... + +**Natalie Pistunovich:** But do they...? + +**Mat Ryer:** Well, that's what they do. + +**Natalie Pistunovich:** Do you want to pad that, and then somebody goes and summarizes that for you? + +**Mat Ryer:** Yeah... + +**Natalie Pistunovich:** Just make it short. Make it to the point. + +**Mat Ryer:** Yeah. + +**Natalie Pistunovich:** Make it a booklet. + +**Jerod Santo:** Well, then it's a blog post, right? They can't sell you that blog post. + +**Mat Ryer:** Would you pay 40 pounds for a blog post, if you had pounds? + +**Jerod Santo:** Exactly. + +**Mat Ryer:** No one else has pounds on here apart from me... + +**Jerod Santo:** I'd happily lose 40 pounds for a blog post... \[laughter\] + +**Mat Ryer:** It's a new exercise-based -- + +**Natalie Pistunovich:** I'm listening to books in an Audible; it's usually 10 Euros per book, it's not 40. + +**Mat Ryer:** That's quite \[01:01:35.07\] + +**Natalie Pistunovich:** It's an almost flat rate. + +**Mat Ryer:** Oh, I see. So for all books. + +**Natalie Pistunovich:** Practically, it's very hard to find a book that is not 10 Euros. + +**Jerod Santo:** But they're not all the same. + +**Mat Ryer:** Oh that's not $10 per book then, that's $10 for the entire library of the Earth, which is a pretty good deal. + +**Natalie Pistunovich:** No, no. + +**Jerod Santo:** \[01:01:54.04\] She still has to pick the book. + +**Natalie Pistunovich:** It's per book. + +**Mat Ryer:** Oh, it's still per book. Okay, I'm back to it being a bad deal. + +**Natalie Pistunovich:** I mean, maybe there's different programs, I don't know. + +**Jerod Santo:** Aren't there some books that are better than others, though? + +**Natalie Pistunovich:** Yeah, I don't know... I have yet to come across a book that is not 10 Euros. Sometimes they have sales that something is cheaper or for free, but most of the books you find there is like one unit of credit, and that costs you 10 Euros. + +**Mat Ryer:** I don't want to be a salesperson for that company, but surely there's already a subscription program where you can just unlock unlimited. + +**Jerod Santo:** Blinkist? + +**Mat Ryer:** I'm not saying that one again, but the Audible -- audio... I nearly said the brand name. + +**Natalie Pistunovich:** Audible. + +**Mat Ryer:** Don't say the brand name! + +**Jerod Santo:** Oh, no... + +**Mat Ryer:** We've advertised them now for free... Jerod's gonna be livid. + +**Natalie Pistunovich:** You said Blinkist... + +**Mat Ryer:** Yeah. But I also said I don't want to advertise them... + +**Natalie Pistunovich:** \[laughs\] + +**Mat Ryer:** ...because Jerod \[01:02:39.14\] + +**Jerod Santo:** They pay good money to advertise on podcasts, just not ours. So go ahead, say their name. + +**Mat Ryer:** So we've given it for free. + +**Jerod Santo:** Yeah, well... + +**Mat Ryer:** It devalues it, doesn't it? + +**Jerod Santo:** We already ripped into Jeff Bezos earlier... + +**Natalie Pistunovich:** It's a Berlin-based company. + +**Jerod Santo:** Audible is? + +**Natalie Pistunovich:** Blinkist. + +**Jerod Santo:** Oh, Blinkist. Now we're saying both their brand names over and over again. + +**Mat Ryer:** Have you heard of that company before then, Natalie? That app? + +**Natalie Pistunovich:** I know this great podcast... \[laughs\] It's called Go Time. + +**Mat Ryer:** It was the home where the great invention of Jerod's where it was -- what was it? You would lose weight by reading a blog post, or in order to unlock the blog post you have to hit certain exercise goals? Something like that. + +**Jerod Santo:** Yeah, I think that would be motivating. + +**Mat Ryer:** Yeah. Fit Blog, or something. + +**Jerod Santo:** You know, it's like a paywall, but it's more of like a fit wall. + +**Mat Ryer:** It's a pain wall. Fit Wall, that's good. Fit Wall is the name of it! + +**Jerod Santo:** Pain wall - actually, I like that. + +**Mat Ryer:** No, Fit Wall. Fit Wall, that's the name. Get the .com, quick. + +**Jerod Santo:** We're gonna solve the obesity epidemic. + +**Natalie Pistunovich:** Remember where you heard this first. + +**Jerod Santo:** Yes. + +**Mat Ryer:** Changelog doesn't own this idea just because we've said it on this. How does that work legally? + +**Jerod Santo:** Fit Wall? Yes. All your IP are belong to us. + +**Mat Ryer:** Well, I'm gonna see if I can get the domain name... I can't...! + +**Jerod Santo:** \[laughs\] + +**Natalie Pistunovich:** Jerod, is this your unpopular opinion? + +**Jerod Santo:** No, I have one. I have one though. + +**Natalie Pistunovich:** You have one. Please share. + +**Jerod Santo:** The last time I was on the show we had a little debate around the terms we software people use to describe ourselves. Coder, programmer, developer, engineer... + +**Mat Ryer:** Cool dude. + +**Jerod Santo:** Cool dude. And since that show went out, I've had lots of follow-ups from people, kind of affirming what I said on the show, which is that everybody kind of has different definitions for what these are... And so there's no consistency, and so that's kind of meaningless to a certain degree. However, I did learn that in Europe engineer is like a protected class or something. You have to have a degree of some kind, which is a much more formal definition of engineer. I'm not sure if that's software engineers, or just engineers, or both. But I learned that from somebody on Twitter, so that's cool. Lots of conversations, people telling me that they mean different things to them. We did a poll on Changelog's Twitter about which of those four do you prefer to be called... + +**Mat Ryer:** Oh, can we guess? + +**Jerod Santo:** Yeah, go ahead and guess. + +**Mat Ryer:** So what are the four options? + +**Jerod Santo:** It was coder, programmer, developer, engineer. + +**Mat Ryer:** Oh, right... I think engineer is the most serious-sounding. So if you're really serious about building -- you know, because proper software engineering is really hard. It's basically impossible, isn't it? If we're honest... + +**Jerod Santo:** \[laughs\] It's basically impossible... + +**Mat Ryer:** It's just this side of impossible, which is why we still do it... + +**Jerod Santo:** Right. + +**Mat Ryer:** So I think that might be the choice if the Changelog audience are kind of like quite serious about their work. + +**Jerod Santo:** Right. + +**Mat Ryer:** What do you think, Natalie? + +**Natalie Pistunovich:** I would say developer. + +**Jerod Santo:** Okay. Good guesses, because those two tied for first place? 42% each. + +**Mat Ryer:** Yeah, interesting. + +**Jerod Santo:** And this was a pretty good sample size. We had a few hundred people reply, so it wasn't like a GoTimeFM poll, where we get 23 votes. Come on people, follow GoTimeFM and vote on our unpop polls. We need you, we need your help. The Go community depends on YOUR votes. + +**Mat Ryer:** \[01:06:00.28\] Yeah. And how am I going to get hair transplant if we don't get more listeners on Go Time? So come on, everyone, listen in. Listen harder! Go and listen to the ads twice. + +**Jerod Santo:** Yeah. + +**Mat Ryer:** Go listen to the ads twice! + +**Natalie Pistunovich:** And answer the poll. + +**Jerod Santo:** And answer the poll, yeah. Okay, so 13% of respondents were happy to be called a programmer. 3% of respondents want to be called coder. + +**Mat Ryer:** Really?! Coder... No one wants it. + +**Jerod Santo:** Yeah. In Leslie Lamport's video that we're referring to, he defines coding and programming as two different things. And this was something that people said to me was that coding is like typing. And developing or programming is much more than just typing. + +So my unpopular opinion - this is a huge setup; we're going to cut this part out of the video. My unpopular opinion is that it's inappropriate to compare coding to typing. So if you're a writer, and you do all the things that writers do, think of the story, the characters, your ideas, design the story, etc. and then you go to write - those are two processes. That's fair with me. + +If you're a developer, and you go doing all the things that a developer is going to do - plan, make decisions, test things, estimate some points wrongly, whatever you're gonna do, and then you go to write the software, I think it's inappropriate to say that the coding part of a developer's job is akin to the typing part of a writer's job. Because typing and coding are different. It's the writing part. So a writer writes, okay? Typing is just a thing that goes mechanically between you and a machine. But a coder codes; a developer codes. And so there's more -- coding is much more than just typing. I think it belittles it to compare it to typing. + +**Mat Ryer:** Yes, I know exactly what you mean. When I first read that, I thought they were saying that coding is really just the act of getting it into the editor. And the programming is happening outside of that, and it's the bigger thing. But you're right, definitely between typing and coding -- I mean, you have to type in order to get coding to happen as well. So I think yeah, that is interesting that they these - but I'm surprised coder just got 3%. + +**Jerod Santo:** Nobody wants to be a coder. + +**Mat Ryer:** Honestly, I thought that sounded the coolest. But I might be out of -- have I reached the age now where I don't know what's cool? + +**Jerod Santo:** It might be an age thing, because I think programmer is an older term, that I think probably people who've been around longer -- I mean, it used to be programmers write programs. But then programs became apps, and programmers kind of became developers. I don't know why or when, but it seemed like that kind of happened, and programmer has fallen a little bit out of trend. + +**Mat Ryer:** Yeah. + +**Jerod Santo:** Out of vogue. But coder is like - nobody wants to be that. + +**Natalie Pistunovich:** I think to me, when I hear code, I'm thinking of code monkey. I heard this term so many times in the negative context of somebody saying "I want to do more than being a code monkey." I think this could be a reason. + +**Jerod Santo:** That's a good point. Maybe that's why people associate it with that, that concept of "Just take a ticket and do the coding, you code monkey." And like, we don't want to be put in that box. That might be true. + +**Mat Ryer:** That could be it. I understand that. But in a way, I quite like being modest. But then I'm in a position where I can't be. Obviously, not everyone can just be immediately modest. But you have to sell yourself when you get into your career, and stuff. But I like it when you see people that have invented core concepts that we all use in their bio just as programmer. It's very understated, and I think that is quite cool. So it may come back, it may be retro, hopefully. + +**Jerod Santo:** Right. We'll see. + +**Mat Ryer:** But I quite like developer, because it sounds it's developing; it's never finished. An engineer - you're almost like at some it'll be engineered, and then it's done, and it's delivered. Whereas there's something nice about the fact you're telling people "This isn't going to be finished. So get that out of your head now! This thing is going to drag on!" + +**Jerod Santo:** \[laughs\] Well, when I think of engineering, I think of like building a bridge, which is very much rooted in years and years of physics, and math, and... + +**Mat Ryer:** Metal... + +**Jerod Santo:** \[01:10:23.01\] ...known equations. And we can engineer a bridge, and then it's done. + +**Mat Ryer:** Tarmac... + +**Jerod Santo:** We know exactly what it's capable of, and what it will buckle under. And I don't feel like software is ever that. We're learning constantly. It's more like that whole flying a plane, or changing it while it's flying, or something. + +**Mat Ryer:** No, it's true. It's because the speed at which we can iterate, and that feedback loop actually, which is one of the points that's come up a few times throughout this. Having a feedback loop, and having a short feedback loop. That is what enables to move quickly, because you can try things, walk back if you got it wrong, you haven't committed loads of resources to it, because it was just a small thing anyway, so you don't feel bad... And you can then correct course... I think that, working at that kind of resolution is the way to do it. And that's why it's so hard to estimate stuff when none of us are doing anything we've really done before. We're probably doing something no one's ever done before, in lots of ways, and so we should give ourselves a break... And tell our managers to shut up! + +**Jerod Santo:** \[laughs\] + +**Natalie Pistunovich:** Perfect way to finish this episode about velocity. Gentlemen, thank you for joining. That was fun. I don't know if you can hear in the outside it's the firework of Berlin celebrating the sunset of the longest day of the year. + +**Jerod Santo:** Oh, wow. + +**Natalie Pistunovich:** It's done. The day is done. + +**Mat Ryer:** Oh, congrats, Berlin. + +**Natalie Pistunovich:** It happened. It's been a pleasure spending it with you. + +**Mat Ryer:** Oh, well, a long day, but... + +**Jerod Santo:** We talked to you all day. + +**Mat Ryer:** It felt like it... + +**Natalie Pistunovich:** Throughout the sunset. + +**Mat Ryer:** It's been a pleasure. There's not two people I wouldn't rather spend an entire long day with... So thank you very much. It's okay just to say that, but... Fine... diff --git a/To TDD or not to TDD_transcript.txt b/To TDD or not to TDD_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..99bd8d1959a7866ed7fc472fd82c157225030aaf --- /dev/null +++ b/To TDD or not to TDD_transcript.txt @@ -0,0 +1,405 @@ +**Natalie Pistunovich:** Good time of the day, wherever you all are. Today we are here to talk about TDD - is it good, is it bad, what other alternatives are out there? And I'm being joined by my co-host, Mat. Hi, Mat. How are you doing? + +**Mat Ryer:** Yeah, hello, Natalie. I'm alright. Thanks. How are you? + +**Natalie Pistunovich:** Good. You are very well centered in the stream, in the video part, so I'm grateful that we're all trying to keep doing that throughout the episode. + +**Mat Ryer:** Thank you. I'll take that as a compliment. I have had better compliments, but I'll take it. + +**Natalie Pistunovich:** This is a compliment right now. It's great that you are taking it. + +**Mat Ryer:** Yeah. Thank you. Normally, you say like "Oh, you look well" or "How attractive you are." But no. "You are centered in the screen." + +**Natalie Pistunovich:** It's a German compliment, Mat. + +**Mat Ryer:** Okay. + +**Natalie Pistunovich:** I can see you well. + +**Mat Ryer:** Well then, danke. \[laughter\] Yeah, danke. + +**Natalie Pistunovich:** Bitte. And we are joined by two guests today. We have Chris James, who wrote the book on testing with Go, "Learn Go with tests", specifically... And you are now an engineering manager at SolPay Hi, Chris. + +**Chris James:** Hey. How's it going? + +**Natalie Pistunovich:** Good. You have lots of coffee in your background. + +**Chris James:** Yeah, I enjoy coffee, and I drank loads, so I'm hyped up. And I'm fairly well centered. + +**Natalie Pistunovich:** You are, and that's great. And we're joined by Bill, who before anything else, is also well centered. Bill, it's awesome. And Bill wears many hats, including the one with the many pins, and you have the one -- so you are a trainer of Go, and recently blockchain, you are heading GoBridge, and also, in September you tweeted something that led to this episode. How're you doing, Bill? + +**Bill Kennedy:** I'm doing great, in Miami, where it's nice and warm. I just want to say that... Because I live in the Caribbean. It's been -- other than we've had torrential rain for two days, but you know... You've got to pay your price a little bit. + +**Mat Ryer:** Nice. + +**Natalie Pistunovich:** So the tweet that you tweeted in September and led a very long conversation is about TDD and how you feel about that... And I'll start with reading it. So you're writing that you've never kept it a secret that you're not a fan of TDD; you believe in prototype-driven development, PDD, with a focus on data-driven development, data-driven design, DDD. And you believe that general engineering cycles should be prototype, solidify, refactor, test, minimize, reduce, and simplify, and in that order. + +**Bill Kennedy:** Wow. I said all that one day, huh? + +**Natalie Pistunovich:** You even wrote that. + +**Bill Kennedy:** Wow... + +**Natalie Pistunovich:** It's on the Twitter blockchain. + +**Bill Kennedy:** I must have been willing to get lambasted on Twitter, because this stuff is like a religion to some people. So I get really nervous when I talked about -- even in training, I get nervous when I talk about testing, because some people are really just -- I don't know, they've got these really strong opinions and ideas about it, and it can touch nerves, so you have to walk this line gently. + +Let me say this... I've learned from working in Go now for about a decade, and even learning how the Go team kind of develops. I've seen them go after this sort of prototype-driven approach. They start with a design document where they're really prototyping everything. Maybe it's on paper, I know they're trying it out on some code, and then they get those ideas solidify to a point where everybody feels like -- because everything goes in API design. I try to tell my team this all the time, "You're writing packages all day. Everything is an API. If you don't put yourself in this sort of mindset... You're not building monolithic code bases anymore. You're developing APIs." + +\[08:00\] Then the idea for me is that I know, at least for myself, and with my teams, it can take a couple of days before we feel like we have an API that is checking every box in terms of what layer of code it exists in, what it can expose, what it can't expose... Even just today I had one of my engineers saying, "Bill, you can't put that API there, because that's app layer stuff, and you put it in business layer stuff." And I'm "Oh yeah, you're right. So now I have to move it." + +So if you're lucky, it just takes two days to get an API design where it checks every box. And so for me, at least, when I think about test-driven development, with the understanding that that means I've gotta write my tests first - I don't even have an API yet that everybody agrees on is designed the way it should be. + +Now, the moment we feel that the API is right, immediately I'm going to start writing some tests, because I can't do go build... And I can start testing that API. But for me, it's like putting the cart ahead of the horse. + +Now, I know that there are people who are very successful at doing this. I'm just not one of those people. I just know the way I work. So that tweet was more about the way that I work, and the way my teams work, and I feel like we get a lot of success with that approach. + +**Mat Ryer:** So how do you test the design? How do you test that API design? Is it literally on paper, and you're just thinking about it? + +**Bill Kennedy:** I'm doing two things. The first thing I'm trying to do is figure out, if this is in my business -- everything for me is a layering. So we've got three layers of code in all of our projects: we've got our application layer, business layer, and we have our foundation layer. And these layers have rules, they have policies... In the foundation layer, we're not allowed to log. That's like our standard library. So the moment a logger is introduced, it already immediately means you can't be here. Right? + +So based on what layer you're in, we're already kind of thinking about what the API can do and what it can't do; what policies it can set, what policies it can't set. And so what we're really trying to do here is develop an API that eventually either has -- if it's a foundation layer API, it's going to be consumed by the business layer. If it's a business layer API, it's gonna be consumed by the app layer. And so that's our user. And so once we feel like we've got an API that's going to work for the layer above, the consumer, the user of the API, then immediately what I do is I can write a test and I use the \_test and the package name. See, for me, having to import the package in the test already gives me that feel for how the API is gonna work. + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** And this is why I get crazy when somebody wants to use these assert packages for tests, because it takes away from the experience of using the API as a user. that's your first opportunity to work with this API as a user. I just don't want to walk away from that. I want to experience even the pain of saying "error" and the pain of it coming out of a function. Because sometimes you might realize, "You know what, this is painful. Are there better ways to do this?" + +**Mat Ryer:** Yeah. But you are still using that API; just because you then make assertions about it... You're still using the API, aren't you? + +**Bill Kennedy:** For me, I don't have the visual connection to it, because I'm not seeing the 'if' there. I'm not seeing the fact that I might want to wrap the error inside the 'if' or I can't wrap the error inside 'if' and what is that gonna cause... There are things that you miss, I think, when you try to make the testing easier. I've seen some really nice testing frameworks, and they produce a lot of nice output, and I get it... But I feel like you're missing a part of -- your user's not using the API that way so why is the test-- You know my big thing, "Don't make things easy to do. Make things easy to understand." So if your user's not using it that way, why are you using it that way, at a critical moment in time where you're the first person actually being the user for that API that you're designing? + +**Mat Ryer:** \[12:00\] Yeah, I think that is the benefit, really, of TDD, the one you've just described, is you get to be the first user. By the way, you know I've made Testify, don't you? You just talk about these assert packages, like "Oh, you know..." But you know full well I made Testify, mate... Well, yeah, so the process you describe of that prototyping thing is kind of how I do it as well. It's just that I do it with the test there. It evolves a lot, and it changes, and often it's wrong. It's not like I've written the test and got it right the first time. That is that kind of process still. And I love the idea that you say putting the tests in a package with \_test. In Go you can do that, even though normally we have, for anyone that doesn't know, normally a folder, all the Go files inside the folder have to be the same package. That's what makes a package. But you are allowed for test files to use a different package. And then you import the other package. And that, like you say, Bill, I love that, because you literally in your test code, you're using the package name as well. So you catch things like jarring, and little usability things like that. + +**Bill Kennedy:** Like stuttering and all the other little things that you wouldn't catch otherwise. I'm not writing a single function that has the word "test" in it until I feel like that API and my team feels like the API is where we want. I don't want to go back and redo test code because I didn't get the API right. + +**Mat Ryer:** But how do you know until you've used it? + +**Bill Kennedy:** Well, look, that's part of the code review process, We're looking at it. I might even, honestly, begin to integrate that into that upper layer. I might start to integrate it into the app layer, into the handlers that need to use it, and start to get that feel. And the moment we all feel like it's right, I'm going to immediately jump in and start writing those unit tests, because I think you need to have a set of unit test just in case your integration tests start to fail. You want to localize yourself into testing that. I just feel like if I do it any sooner, I'm adding more work to my workflow... And I'd rather feel like the team feels like this is right, and it feels right in all the different layers, and then the next thing is tested. + +**Chris James:** Yeah, I think what I find interesting about this is that what you've described is not so different from the way I work or the way my team works. And everyone on my team does TDD, Because software development is never as straightforward as you want it to be. It's a learning process, even for something that seems like a simple problem. It's only when you start digging in, and you start realizing what are kind of weird contradictions and requirements and how a domain isn't quite as simple as you hoped it would be. + +So for me, an important thing about software development - this kind of goes beyond TDD - is this importance of high-quality and frequent feedback loops. We need to be able to hear about our design, we need to be able to hear about whether the behavior of our code is what we expect it to be. And for me, that's the strength of TDD, is about giving you that high-quality piece of feedback that your code does what it's supposed to do. + +As you mentioned, what tests do is they also reflect your design back at you as well. So you can look at the design of your code, you can see it behaves how you'd and it's like a button press away; it's like Ctrl+F5, or whatever the magic incantation is on your IDE, to get that feedback that "Yeah, it works how I want it to work", and I can read the test and see, if I were a user of this package, this looks reasonable to me. Or, more often than not, the opposite; "Why was this test so awkward to write, and why is it so difficult to read?" Well, that's not testing that's difficult here, it's the design you've created. Well-designed code is simple to test. And the opposite is true as well - poorly-designed code that kind of has mixed responsibilities, and too many dependencies, and everything else you guys talked about, they are difficult to test as well. + +So by starting by a test, it kind of centers you on a particular goal... Because I think another problem that I've encountered with people who don't do TDD -- I'm not saying this is true of everyone, but I've worked with a lot of inexperienced developers, and they think they understand a problem, and they start writing tons of code all over the place. And I ask them, "What are you doing? What problem are you solving?" and it's they're solving five problems at once, and they're solving them all poorly. + +\[16:17\] And I ask them, "How many times have you executed this code?" and it'd be like one or two times, when they can be bothered to wire it up into main.go, and then run the thing manually. And to me, the strength of TDD is it just gives you a very simple, repeatable process that gives you those feedback loops quickly and with high-quality. So that's kind of the way I see it. + +But I think it's interesting that this kind of discussion is sort of framed as "I love TDD" or "I hate TDD", but really, actually we kind of have -- I mean, we're all software developers, and we all recognize that this is more difficult than people sort of give it credit for. We need to be able to learn quickly. So I find it interesting that, basically, we have the same problems, it's just different ways of attacking them. + +**Bill Kennedy:** I think it comes down to the order of things. I'm not writing tests first. For me, TDD is always - and I could be wrong, but when I first heard about TDD, it was always write tests first. And that order that I gave you, Natalie - I don't remember off the top my head, but can you read that order again, that I threw out on that tweet? + +**Natalie Pistunovich:** Prototype, solidify, refactor, test, minimize, reduce and simplify. + +**Bill Kennedy:** Alright, so I have Test fourth there, which is kind of what I said, right? We need a package that does this. Let's design the API. Let's see how it feels. Do we feel like it checks every box? Good; we like it. Now let's write the tests before we do a go build, so we know that that API is behaving that way. So I have it four. I'm not saying "Don't test." I just I can't put it first, I just tell you that. So maybe I'm wrong about what TDD is then, right? Is it write tests first, or is it just write tests at some point? + +**Mat Ryer:** I think it's write tests first. Chris? + +**Chris James:** But I would add it's write test first. This is the interesting thing about TDD, is that there's unfortunately a lot of misrepresentations from what Kent Beck & co write. I think a lot of people seem to be under the impression that you get given this project of "Let's rewrite Twitter." And okay, I'm going to spend days writing tests, and then I'll make them all pass. That is not TDD. TDD is about taking your first small step, capturing it as one test, and at that point you're just trying to make it work, and you're kind of free to write whatever garbage you want, to an extent; you're just trying to get the test passing, and your software working how you want. And then the other really important step that people will skip is the refactoring part, where now you have a test that proves your software works how you want it to do, you're now liberated to be creative, and think about design, and reshape the code, and do refactoring... And you keep hammering that Run Test button. So you never go down the rabbit hole. Because quite often, people talk about refactoring, and all the time - they're spending hours doing refactoring. And I'm sorry, if you're spending hours doing refactoring, you're probably not actually doing refactoring. You're probably redesigning your code. And there is a difference between those two activities. + +So for me, that is a really common misconception, and that's why when people say things like "I don't have time to write tests first", and I'm kind of like "Mate, it takes like five minutes, if that." It shouldn't feel like a big blocker to you achieving work. The opposite, in fact. Once you get proficient at TDD, and once you get proficient at software design, TV speeds you up; it doesn't slow you down. + +**Bill Kennedy:** Yeah, it's mind blowing to me, Chris; mind blowing. I cannot -- I'm telling you right now, the thought of me writing a test first, for an API that doesn't exist yet, because I haven't really figured out what I need it to do is mind blowing. + +**Mat Ryer:** But when you think about the design of that API, Bill - like, it's going to be used by somebody writing code; they're going to be calling functions, and using types in that package, and things. That's the bit I don't get. Because for me, first of all, I should say, I have had the case before where I've overdone it on the tests upfront, where I've just done too much. I thought I knew the thing so well that I just wrote -- I know exactly what the tests need to be, and then I can just bang through and implement it. And then during the implementation, which was kind of like the prototyping phase really, I learned things that then impacted the design; for necessity. And then a lot of my tests were then useless. + +\[20:25\] So Chris's point I think is quite good, but you write a bit of a test first, and then you do a bit over here, and you sort of do them at the same time. But driven by the test means you automatically get better coverage, because you immediately -- well, in my experience, the test coverage is always high, kind of automatically... But I think the principle - you wouldn't want to ignore what's in the code and what the code is telling you. And that's like -- the prototyping bit I think is one of the key inputs you want for the design, too. + +**Bill Kennedy:** So just recently, I had to write an API to access some service out there; there's authentication bits to it, there's HTTP calls I have to make, there's data coming back... I wrote 30 lines of code to start; happy path. I put my programming hat on, right? Find the 30 lines of code that make at least parts of the API I know I need working. Like, that's stage one, right? You've got to do the programming. Now, once I see those 30 lines of code, and I see what I need as for config, now I know types I need to return, and I have all this, my next thing really is to design the API based on that programming, and then to share that API design with my team, and maybe even see how it feels, again, in that app layer, and then write the tests. I cannot wrap my head around writing tests without really feeling that it plugs in as a puzzle piece. + +**Chris James:** So I think a useful thinking tool when I'm coaching people TDD is I ask them to sort of ask themselves, "Pretend this code already existed as an open source package. What would you want to see?" And it gets you to forget about the implementation details. When you're designing packages in Go, the good ones are the ones where you just sort of see a very useful API, and you don't know, and you don't care how it works underneath. It shouldn't matter to you. I'm a user of this API, this package; you've taken care of all the messy work. I just want a nice, beautiful API that I can use. And that helps people make that first step. Because yeah, I think when you're first starting with TDD, it does feel a bit unnatural; you sort of have to make this weird first step where you're just writing code for something that doesn't exist, and your VS code or IntelliJ is all red because the code doesn't exist, and it makes you feel uncomfortable. + +But what you need to be doing when you're in this mindset is have this positive attitude towards it, and go "What would be the best package design that I as a user could put in front of myself?" and then encode that as a test. And you immediately document it to yourself; you sort of say, "Okay, so if I do this right, I see exactly how to use this package." Then you dig into the messy details and make it pass, and things... But I guess my point I'm trying to make is like well-designed packages hide that from you anyway. So you're not gonna see that in the test; you should just see a well-designed API in front of you in the test. And that thinking tool helps you write a good test in the first place. + +Because how many tests have we all seen where it's like there's a load of messy implementation details and things that's really difficult to read... And worse still, difficult to maintain, because then when you change an implementation detail, you're suddenly like "Oh, my goodness... Ten tests are now failing because I changed this one line of code." There's all these terrible things. But if you start with a test-first attitude, and think of it in terms of the behavior you want, rather than the implementation details, you're more likely to write a maintainable test suite. + +**Mat Ryer:** I quite like those errors that you get, actually. Like, when you write a test for something that just completely doesn't exist, you get compiler errors, but you get one error at a time, and it's almost like a to-do list; it tells you what to do. And then once you've got past all your compiler errors, you then have a failing test, which is then telling you "Hey, now you have to fix this thing." + +\[24:08\] And the classic case of -- if you take red/green testing, this idea that it's important to see a test fail first...Or if you write a test, and it always passes, and it's always passed, how do you know you're really saying anything about the service, or anything? So I really want to see it fail, so I know it's actually saying something, and then I can make it -- it'll say, "Oh, this has now failed. The expected output was 42." Okay, so I'm just going to return 42 from this method - and this is an extreme case - which then makes that test pass. But then that highlights the test isn't complete yet, because I made just one assertion about it being 42, and just returning a 42 is enough to make that pass... So I'll put a couple of other examples in there, and then I can't cheat like that, really. So yeah, I quite like that error list that you get guides you through it. + +**Bill Kennedy:** The majority of packages I write are not single-module packages that are going to be in somebody's vendor folder. They are for these big projects, where we've got four people working on this thing, and now the app layer needs support for being able to order things from the database. So now I've got to figure out this package; parts of it might even live here, parts of it might even have to be app layer, because I can't put that intelligence there... And so I'm designing something not as a vendor package, which might have an API where maybe you can think TDD to start, I'm thinking already "I've got to integrate this support in the right way based on our rules, and I've got to make sure that puzzle fits first." And I'm telling you that I have never, on any project like this, gotten the API right even the fourth time. There's always somebody who remembers something, like "Bill, you can't -- Bill, you're leaking this. Bill, you're doing that." And I'm like "I teach this stuff, and I'm still making these mistakes." Or I'm moving fast, and... One time I put something in there, and I wrote, "You're not gonna like this." And everybody was like "What?" And I said, "If you don't see it, Pavel will." And immediately when Pavel woke up, this first thing was "Bill, you can't do that", and I started laughing. I just left it for him, because I knew he was going to say "You can't do that." + +And so for those packages, the majority the packages I write, I really need to be thinking about the design as it relates to how its integrated into this project; not as a vendor package. So again, my brain says, "If I'm writing tests first, I'm not leveraging my time enough." Because I really don't know how it fits best until I try to fit it in. I haven't run the code. I'm just trying to make sure it fits in and checks all the boxes, and then immediately, yes, I'm gonna write -- I don't care about the implementation so much as, again, the behavior in the field. But integrating into a big project is a much different problem than the vendor, say, package that you're going to be building. + +**Break:** \[27:02\] + +**Natalie Pistunovich:** Chris, I have a question for you. You said that it's important for you to separate the responsibilities, and decouple, and make sure that each part is individual, and that way it's simple, and you know that when you're asking, "What is this doing?", "This is doing this one thing." "What is this testing?" "This is testing this one thing." So when you start writing tests for something completely new, do you go bottom-up or top-down? So do you write the first test to be the biggest functionality, and then you add small tests, or do your first write tests for the very small things, and then later on add the big test that kind of checks everything? + +**Chris James:** I tend to favor top-down development, because it's the least risky way of working, in my view. I've seen a lot of kind of bottom-up efforts where people are making these kind of beautiful-looking packages, that have corresponding beautiful test suites... But then when you try and integrate them together, you realize this is just a mess, and it doesn't actually do what you need it to do. Again, it's this idea of like high-quality and relevant feedback. Like, yeah, you've unit-tested this package, and yeah, it's wonderful, but it hasn't done anything useful until it's wired up into main.go in effect, right? + +So I generally prefer a top-down approach, and I would definitely plug Steve Freeman and Nat Pryce's book "Growing object-oriented software, guided by tests." And I know some people in the Go community are a bit allergic to OO, and it's a Java book as well... But I promise you it's a valuable book, because it kind of talks about this kind of approach of work, of working from the top down, starting with an acceptance test. And what the acceptance test does is exercise the system as if it was a black box. So it doesn't matter what the package design is, it doesn't matter how it works underneath. It just literally hits the system and checks "Does it do what you need it to do?" Starting with that as like your Northstar is an incredibly powerful way of working, and it's something that my team really strongly embraces, because it forces you to think about "What does our system actually has to do?" Like, forget about design, and stuff, because as engineers, we love to talk about design and all this kind of business... But our system has to do something useful; it needs to do something that makes money, or helps people, or whatever. An acceptance tests forces you think about "What will this system do from the outside in?" You then use that as a North Star to guide your efforts. + +So you have your \[unintelligible 00:30:40.15\] acceptance test, and then you just start writing some code. And then again, like the kind TDD approach in itself, at first you're just trying to make the tests pass. You're not worrying too much about package design. Honestly, when I do it, most of the time I'm making it work all in main.go. I'm just like "Get it working. Let's just get the software doing something useful." And then again, once we have that passing the acceptance test, and it proves the system works how we need it to do, now I have the license to start putting things into packages and things, and then I might start testing some edge cases for unit tests that aren't so great with acceptance tests... And then the software grows from there, hence the book title "Growing object-oriented software." It's this idea that you don't just write software and it's done; you grow it. And in order to grow it, again, you need these high-quality feedback loops that keep coming at you quickly, letting you know whether you're going on the right track or not. + +**Mat Ryer:** Are your acceptance tests integration tests then, essentially like end-to-end tests? + +**Chris James:** Yeah, the naming around all this stuff is incredibly confusing. For me, an integration test is something that takes two units, or two or more units and checks that they integrate and do something together. For me, an acceptance test is something that tests the system as a black box. So if you're making like a HTTP API, in \[unintelligible 00:31:52.21\] you could do an acceptance test with curl commands. That's the way you look at it; you don't really know there's even a Go system at that point; it's just some Go program that responds to HTTP. That to me is an acceptance test or a blackbox test. + +**Mat Ryer:** Okay. + +**Chris James:** \[32:06\] Integration tests are more within the system for me. + +**Mat Ryer:** Okay. So an end-to-end test maybe we could call that. + +**Chris James:** Yeah. + +**Mat Ryer:** Yeah. So it's funny, because that is how I do it... So I have a thing that we're building, and it has a frontend, and it has a backend with an API. And the API is like -- when we run this test suite, it literally spins up the database, it spins up the whole thing, and accesses it like the client is going to, like the frontend is going to, calling the methods, making the same calls. + +We actually have an API client as well, which it uses. So we're also kind of testing that at the same time as well. So this is like really getting as close as it can to what the browser is going to do. And then I'll be able to write then in quite a clear way what I expect. Like, I'm going to do one thing, and then I'm going to do a few other bits... You know, it might be adding users to it, it might be adding a task, whatever the flow is. Then I'm going to get it and make assertions about what that state should then look like. And usually, when I'm adding a new feature, that's a very easy way for me to reason about what the feature is that I'm adding, and it sort of takes away a lot of the ambiguity. Like, if we've discussed it, lots of people will have a different idea of what that thing is, and this lets you remove a lot of that ambiguity and saying, "Well, look, I'm literally calling this API in this order. It should result in this." And if anyone doesn't agree with that, that's great to catch that nice and early, and we can figure it out. + +And then yeah, you go about implementing it, making that test pass... And you sort of -- because of that approach, I know that that new feature works as I expect, whatever else is going on. And I can get -- actually, sometimes we'll skip unit testing in some places, because that's just such a kind of reliable thing that I just have the confidence that that works. + +**Bill Kennedy:** I'm always worried about guessing. I don't want to guess. Let me give you an example, because you mentioned frontend... I have a person that works for me, he's a frontend dev; we were working on something, and I started doing a code review... And it was all JavaScript Node, so I was really losing my mind. And I finally looked at him and I said, "Why are you doing that?" It got to the point where I went and I built a terminal frontend so I could experience the API that I gave them. And I realized so many things were missing that I changed the API, so I can write a frontend. And I went back to the frontend person and I said, "I don't understand, one, how you got this working, and two, why didn't you ask me for these changes?" And you know what was the saddest thing? They said, "Bill, as a frontend dev, we just deal with what we get, and we just make it work." And I said, "This is a horrible way to live." I go, "You can't do that with me. You have to tell me when something's missing." And I got to the point with them now where if we need a new screen, I ask them to build it, and I asked them to go into the Go side and write handlers that send static data. + +**Mat Ryer:** Yeah, great. + +**Bill Kennedy:** I said, "You tell me what you need, and I will do my best to make that API work for you." And then I'll work from the business layer first, so I can feed that, and then I'll work at, say, the data layer. But it's so sad, because a lot of the APIs that I had felt right to me, but were not frontend consumer-friendly. And again, I don't want to guess... So there's a situation where I could have written all those tests and done all that stuff, and realized that it was all wrong. So I'll just throw that out there... + +**Mat Ryer:** Yeah, but I think you just did TDD, really; you built the terminal... It's not tests as code, but it is like, you are testing it; like, you run the thing, see if it works, and then make changes to it that way. That feedback loop is what you needed from that process. + +**Bill Kennedy:** \[36:07\] For mocking out the handlers as a test-driven thing? I mean, I was asking the frontend to do that, so they could drive my design, because I don't want them to be -- you know how horrible it is when you have an API that's not even doing error handling consistently? The amount of extra code a frontend has to write because of that... And I'm thinking about these things, and I still wrote a -- because I'm not a frontend dev; I still wrote a pretty poor API for web dev, frontend dev... \[laughs\] Oh, I felt horrible, man. I never want a frontend dev telling me that ever again. I felt horrible. + +**Natalie Pistunovich:** I guess the big difference here is that somebody else is testing for you. + +**Bill Kennedy:** I say this in class - we write APIs all day, so you can either make somebody's life amazing, or miserable. And we've all not wanted to go to work the next day, because we had to work with an API that was just... Dude, I had a JSON API return XML on me once. + +**Mat Ryer:** \[laughs\] + +**Bill Kennedy:** And then I had to deal with that. + +**Mat Ryer:** Inside JSON? Or just pure XML. + +**Bill Kennedy:** Pure XML, not inside JSON. Apparently, they were fronting an XML service instead of rewriting it; they were transcribing XML, and I found a call that bypassed the trans-- \[laughs\] But you understand what I'm saying... Consumer is everything. And I call them the user. The user is everything. And if you're not the user, then you're really guessing. + +**Mat Ryer:** Fair enough. + +**Bill Kennedy:** I never do it right. + +**Mat Ryer:** Yeah, I think that's fair. I think that totally makes sense. I think TDD is about being obsessed with the user; you become the user. And to be honest, in the case I talk about, it's possible for me, from just starting at the API, it would be possible for me to build a horrible API, i.e. you make this call to get some data, and you have to make lots of them if you want lots of data, and what the UI really wants to show is a summary. Things like that, I think... I don't know that you get that solved for you by any sort of process. + +But I wonder, Bill, if -- and I don't know the answer to this, but I do wonder if the approach changes what we end up with much, or if we sort of end up in the same place, regardless of process. Because when you talk about things being like in your big system, where there's lots of moving pieces, and you have to think a lot more, you can't just jump straight into coding bits, I wonder if that would be different if you'd started with that process... Or if you just end up still with the same thing, and the process is really just personal taste almost, for the team. + +**Bill Kennedy:** It's about, for me, trying to minimize the amount of guessing you're doing, and prototyping does that. The more you can prototype and solidify something, then the less guess you're going to deliver. So I'm curious, Mat, since you do a lot of frontend stuff, if you're not the one working on the frontend, is your frontend developer allowed to say "That API isn't working for me"? Because this person was never allowed to say that. + +**Mat Ryer:** Yeah, so that is interesting. I love that idea. One thing we did in the past place was we would collaborate on that JSON API, and we'd have a .json file, and we'd just serve that from the backend. So the frontend is unblocked, because they're getting at least the test data. A bit like how you said, just have a Go handler that returns it; I'd kind of prefer that, actually, because it's easier. I'll probably do that from now on. And then you're both working to a kind of point in the middle... Like, when they built the tunnel between the UK and France, they started digging from the French side, they started digging from the English side, and they met in the middle. + +So yes, I think you're right, the frontend is really kind of paramount, because that's where the actual end users are going to be touching something. So I put more importance in that. So if a frontend needs data in a particular way, I think that outweighs what else might be considered there... + +**Bill Kennedy:** \[39:57\] Yeah. So I don't want to guess, and you are going to have some level of guessing. It's impossible until you actually finish whatever that integration is. But I want to minimize that. And I just feel like, personally, that if I'm starting with writing tests - again, because I'm not just writing a package that somebody's going to vendor; I use vendors, you understand what I mean there. I'm writing a package that has to integrate in complex ways, and so I just need to make sure those puzzle pieces come in. + +I have a person on my team that likes to start the business layer. Doesn't want to even look at like -- a data store is a complete abstraction to them, right? So they just want to completely start in the middle of the business layer, and get a feel for those APIs first, before we do anything. There's lots of ways to succeed here. + +**Mat Ryer:** I think that's going to end up being where we land on this, honestly... But yeah, so what we'd do with those JSON files in the past place - they would be the tests for the backend API. So it gets served to the frontend for them to code against, and it becomes the same file as the test. So we call the API in the expected way, and assert that the JSON we get back matches this JSON. And often, as you're going, it doesn't, and it breaks, and that's a conversation point, because you want to change that JSON. You're like "Okay, so because I learned about this thing, it's not a boolean in the end, it's a string... So is it okay? Can we make this change?" And you have to then make the change and sort of negotiate a little bit. + +**Bill Kennedy:** But here's the cool part with letting my Node developers write handlers... Because they learned how to do that in like 15 minutes. They're designing, at some level, the application-level models. And I'm making them define -- and I show them how to write a literal struct. But they're also defining the data model at the app layer. That doesn't mean it's going to be the business layer model. It's always good to have that decoupled. But they're also writing that model, that data... Remember, we talked about data-driven development. If you don't understand the data, you don't understand the problem. So the frontend devs get two huge -- everybody wins. In fact, the frontend devs are always ahead of me. I've never had that in my life. I always had to be ahead of the frontend devs. Now, I'm constantly trying to figure out what they're going to do, because I'm now behind, because they wrote the handler, they wrote the literal struct, they defined the data model, they filled it in, they know it works, they can show it to me... And now I'm scrambling to make that come to life, right? And that's the contract. + +**Mat Ryer:** Yeah, I love that. I mean, that is literally how I like to work as well. So I think I'm with you there. While you're implementing it, do you then ever say, "Oh, hang on... I've just learned something. I can't give you it like that. It's going to have to be like this"? Do you have that handshake? Or will you just take it like as law that what they've asked for is what they'll get? + +**Bill Kennedy:** I try to keep the application model exactly the way they want. And then at the business layer we define our own models that are going to work better for \[unintelligible 00:42:49.11\] Right? But there are times, obviously, where we've gotta negotiate... Time is a great example. Time is a nightmare. How do you want to deliver time down to JSON? And sometimes we just pick a format, an RFC format, sometimes we'll do some other things... So if we were to negotiate anything with the frontend, after looking at a data model they've put together, probably time is always the sticking point; bringing time back and forth. + +**Natalie Pistunovich:** I also want to answer your question, Mat, although it was for Bill... + +**Mat Ryer:** Please do, Natalie. + +**Natalie Pistunovich:** ...if this is the same goal, or is this - you end up in different places, doing the two different methods... Because recently, I had a project where it was a very new, and it was not TDD at all; it was very much -- not with the frontend, but actually with two microservices. And it was one team telling the other team always, or the person from microservice A telling the person in microservice B, "Please make this change. I need it like that." So a very similar interaction to what Bill was describing with the frontend people. And I did work mostly in TDD, I would say, but this one was so different, and I feel that this did lead to different interactions and a different result because of this -- the person who tests, who does this test is not you; you cannot guess that. You have other people bringing other expectations. So I do think this is a different -- you don't end up in the same place. + +**Mat Ryer:** Yeah, I think you might not end up in the same place. But also, that I quite like also, having somebody else write the tests. If you're pairing, it's quite nice to have - one person will write a bit of a test, and the other person then switches and does the implementation. That works quite nicely as a way to collaborate, and evolve, or grow, as Chris said. So I don't know that that's like mutually exclusive, really, who writes the tests... And honestly, I feel like there's a big "It depends." I feel like "It depends" needs its own little theme tune in tech, because basically there's so much "It depends." + +And I feel like often whenever someone on Twitter, or whichever platform is now still going, by the time this goes out on Thursday... I feel like when there's a strong opinion like this, and there's big disagreement, what we find out if we dig in deeper is we're actually doing different things, or the scenario is subtly different, or sometimes dramatically different. And we need like a big stamp that just says "It depends." + +**Break:** \[45:21\] + +**Bill Kennedy:** So I want to answer to something Dylan's asking on Slack, and I really would love Chris's opinion on this, too... Because I have opinions on this. Dylan's asking two things. He's asking about "Tests must be fast, and spinning up a Postgres database in Docker." Let me just talk to that. I do that. I feel like if you're mocking a database, you're not really knowing if your stuff works properly. And Postgres spins up so quickly... I do it in CircleCI; those tests run easily within a couple of seconds. So I have no issues personally right now with how long a test necessarily takes, as long as there's an expectation... + +But the other thing is the big one, where it's this idea that if your test hits a database, it's not allowed to be a unit. So let me just tell you what I think a unit test is, and then I'd love to get Chris's opinion. See, what I love about Go is it's already defined what a unit of code is. We don't build monolithic codebases. The unit test is a test that tests that package's API in and of itself, right? Not any other interactions of other packages; just every interaction that that API has to make. And if that API is talking to a database, then I don't know why, personally, I can't spin up a database and write a unit test that makes sure that that API's queries and calls work. To me, that's still a unit test. It just happens to be hitting a database to run that unit of code. So Chris, I'd love to hear your opinion on that. + +**Chris James:** I wish I could bring in some controversy, but I basically completely agree. I think there's been a big misrepresentation in tech about what a unit test is... Or it could only be a particular function, or whatever. I wouldn't describe it as a unit test; to me, this is an integration test. But that distinction isn't that important in the grand scheme of things. Like, you're testing your code integrates with Postgres, or whatever database, right? So basically, I agree. And I do the same thing. I will have some kind of package that is in charge of doing whatever I need to do with Postgres, and I'll write -- I have to use test containers most of the time to spin up a Postgres Docker, and I run my tests... And yeah, they're not as fast as unit tests, they take a few seconds, but - big deal. Still a wonderful feedback loop in the grand scheme of things. + +Yes, I'm afraid to say, I basically completely agree. I think mocking the database is really clunky and difficult to do, and you still don't really have particularly useful tests, because you can prove that like your SQL string equals select star from users, but does that do what you need it to do? I don't know... It just doesn't prove anything, in my view. + +**Mat Ryer:** So it's really like -- maybe we could talk about things that your tests should give you. I think the feedback loop thing is very important. By the way, you can just spin up a database and keep it running as part of your dev environment too, and just clear it out every time, rather than having to spin up the container every time... But again, that's just little optimizations, and things... + +\[49:59\] But yeah, so that is a question people should ask, because there's a lot of language -- I don't know what people mean when they say an integration test, honestly... I don't know what they mean. I'd have to find out, I'd have to ask them more questions. So what do you get from your tests? Is it really telling you what you want to know? Can you break something in your code somewhere, and then you get a failing test that points to that failure? And so this speaks a little bit to that idea... Really, the reason I think unit was used as a word was, ideally, you would just get one test that fails if you break something, because it tells you exactly what's broken. You can't always do it, and if you've found a function that's called all over the place for doing auth, or whatever it is, and that breaks, loads of tests will fail. But actually, that's quite a good clue that there's something fundamental that's broken here. So I like that laser-sharp focus on failing tests as well. + +**Bill Kennedy:** Yeah, I think the useful thing to look at here is the test pyramid, which describes different kinds of tests in the shape of a pyramid, funnily enough. So at the top of a pyramid you have your end-to-end tests, your acceptance tests... And the idea is that you should only have a few of those comparatively, because whilst they're great, and they give you a lot of confidence that your system does what it's supposed to do, they typically are slower than the tests further down the pyramid. They're usually more expensive to maintain, and normally, the quality of the feedback loop isn't as great. So you can write an end-to-end test and it can go where it doesn't work; typically, it doesn't tell you why it doesn't work, and you have to dig into your system to try and understand why. + +Integration tests - you're getting a few more of those; they're proving that two lumps of code, whether it be two packages, or a package and a Postgres database, whether they work together... Again, they're faster than your end-to-end tests, they give you better quality feedback, but again, because they're two different things, which one's the problem here when it fails, right? You have to kind of dig into the various units to understand, "Okay, it was this bit of code that was the problem here." + +And then finally, you have unit tests, which are what you're describing, Mat; typically, when they fail, you get a very precise reason as to why they fail, and they also give you really fast feedback loops. These are the ones that really helped you refactor, because you want to be able to do refactoring frequently, and small, frequent refactorings. And that's what unit tests give you, they give you this ability to just refactor \[unintelligible 00:52:18.18\] + +And you should have loads of unit tests, to a point. You can get to a point where you have perhaps too many tests that are kind of coupling themselves as stuff you don't need, and I'm sure we've all been in projects where there are tests that are impeding our ability to change a system. You get to a point where you're like "I want to do some refactoring, but I know if I do this, I'm going to have a bunch of tests failing." And that's definitely not a situation you want to be in. + +**Bill Kennedy:** I have a question for everybody here... NASA did a study and they said roughly every 20 lines of code you write, there's a bug in your code, whether you like it or not. Now, this was off of their codebases that are written in C, so maybe we get some better numbers out of Go. + +**Mat Ryer:** Who's this NASA? + +**Chris James:** That's boring, ain't it? + +**Mat Ryer:** Well, if they've got that many bugs, how did they get to the Moon? + +**Bill Kennedy:** There we go, man... It was a lot of testing. + +**Mat Ryer:** Confidence. + +**Bill Kennedy:** But here's the question.. I get asked this, so I'm curious, everybody here... At some point, you have to be done. You have to be done writing tests at some point. So one, what is your definition of "We're done writing tests", and if that answer has to do with some sort of level of code coverage, then what is your code coverage number to say "We're done. We're moving on." Because you're not going to get to 100%. It doesn't mean anything. So I'm curious about what your done answer is for when you're done writing tests for that package, and if there's a number, what the number is. + +**Chris James:** \[53:42\] It's not easy to come up with a straightforward answer to this, and I think I'd like to stress also, again, that whilst TDD does give you a test suite, and does kind of naturally give you coverage, TDD is not primarily about verification. It's a method for driving out software, right? So just because you've done TDD doesn't mean your testing picture is over. So you might need to do some pen tests if you're writing something that needs to be particularly secure; you probably need to do performance tests. You might just need to write other verification tests, like fuzzing tests, for instance. + +So TDD and testing, they're clearly related, but they're not quite the same thing. And I think really, it just comes down to confidence. In terms of software, we want to be able to ship it as quickly and as reliably as possible. So if a bug comes up, I want to be able to like fix a bug and ship it as quickly as possible. And most modern teams these days would rely on tests to say to themselves "Yeah, we're good to go. We can release this to production." + +So if your test suite is in that state, I would say you're done. But as I said, writing software isn't quite like that. Normally, it has to grow and evolve over time as you learn from your users. You need to -- "Oh, the users told us they need to do this. Okay, well, now we need to go back and change things." + +**Bill Kennedy:** That's fair. But if you're managing a team of people, and you say to that person -- at some point you say, "Do we have enough test coverage? Because if we don't, I don't want you moving on to the next thing." So it's like anything in engineering - how do we know we're done with this so we can move on to the next thing? + +**Mat Ryer:** In my experience, I'm not interested in code coverage that much, because essentially, the only promises I'm making are what the tests say. And so the tests -- you know, if there's hidden logic, if there's like hidden stuff that happens that's not in the tests, then it's unofficial, essentially; it's not part of the official API. So that gives you a lot of freedom, because as long as your tests pass, you can deploy it. And honestly, for me that's the real value of testing - you want to have the confidence to be able to push to production, and not be worried, and not be like "Oh, let's keep an eye on it, mate." I'm so confident with our tech that if all the tests are passing, push. That's the only promise we're making, just from that layers point of view. + +**Bill Kennedy:** Okay, but you still didn't answer my question, Mat. How do you know that the tests are giving you that? How do you know you're done? How do you know the tests are giving you that ability to sleep tonight? We all agree, and I've learned this too hard over the last ten years working with people from Google - everything has to be measurable. You're not allowed to to say "Because it feels right." You're not allowed to do that. Things have to be measurable. So how are you measuring that that package is done? Let's keep it down to a unit of code. You've assigned a package to somebody to write; how do you know it's done, so they can work on the next package? I know testing has to be a part of the done equation... So how do you know you're done to move on to the next thing? + +**Mat Ryer:** Well, just if I've satisfied the user, really. That's the thing. And also, I understand when people say "We're never done" in software. So I understand that basically it's fine, this is good enough for now; but it's not done. Like, we'll definitely evolve it and change it. + +**Bill Kennedy:** Okay, at some point you're done. It may not be complete. + +**Mat Ryer:** Okay, fine. + +**Bill Kennedy:** How many times have you seen a module out there where the author said, "I'm not working on this anymore"? I consider those done. The PQ driver for Postgres - I still use it. Everybody's like "Why are you using it?" Because it's done, and it's stable, and why should I go to PGX, where they're still doing active development on it? ...and this is done. We get scared of done sometimes, and I don't know why. It may not be complete, but it's done. You know, Dave Cheney taught me that... + +**Mat Ryer:** Yeah. Well, I think -- I look at code coverage, because I'm interested if there's any blocks that I've missed. And I don't test every error -- like, if error return; I don't test those. I sort of trust myself that I get those right. And sometimes that bites me, because there'll be some strange little thing where I've shattered the arrow or something like this, and there is a bug in there. So I only test error things like that if it's part of the API. Like, it's returning a special error in a certain case, and that's part of the API; I'll make sure that's covered in the test. But it's just if the thing is... If it's done. Um... \[laughter\] + +**Chris James:** \[58:07\] It's done when it's done, right? + +**Bill Kennedy:** Yeah, Chris \[unintelligible 00:58:11.05\] + +**Chris James:** I mean, I can only describe what my team does. And I tried to bring up another TLE but, BDD, behavior-driven development, right? That's another thing that people throw around. + +**Mat Ryer:** It's another DD... + +**Chris James:** Yeah. To me, without going on too much about it, it's about understanding your problem with your stakeholders, like the customers, or like your business people, or whoever. So when my team pick up a ticket, what we try to do is understand these requirements, and we try to express them in these kinds of BDD terms, like "Given something, when something happens, then this happens." And we discuss it as a team, we try and understand, "Okay, this is the problem we're trying to solve." And then we basically tell ourselves, "So in order to do this ticket, we need to have tests for each of these scenarios." But we also acknowledge that something will fall through the cracks. Of course it will. No one can write perfect software. It's just not possible, right? But we have enough confidence, we think we discussed it enough, and we're going to put it live. + +And then of course, something goes wrong... Okay, we recover from it. And I think a far more interesting question to me in terms of software developer - it's not so much like "How many bugs can you prevent?" but "How quickly can you recover from them?" Mean time to recovery for me is a way more interesting question. And again, this idea of having a very effective test suite is such a key components to that. I've heard of companies that have test suites that take days to run. So if you've got a bug, what do you do? Well, what they do is they circumvent the process; they skip the tests, because they need to fix the bug. And then that causes more bugs. And it just gets worse and worse and worse. And when you read like the State of DevOps reports and things, it just constantly talks about this stuff, about how important it is to recover quickly. And to me, an effective test suite is like the center stage of that particular problem for us. + +**Natalie Pistunovich:** I see that there's a little bit of an MVP, in the sense that you do the basic thing you need in order to run with that, and then you can always add to that. And what is included in that MVP is what sort of, you can say what is in the API, in the sense of what you communicate, what you commit to has to be tested, the expected behavior, and at least catching the unexpected behavior... So even if it's as basic as "if not expected behavior, then..." And then this is like the very minimum that you need to call this done, and you can improve on that, for example, by breaking down the scenarios of "Not expected, A, B, C", and so on. + +**Mat Ryer:** Yeah, I really like that idea, actually, of anticipating how it might be misused, and catching that. Sometimes that really is helpful. And that's really the only time I will write panics, is if it's going to panic anyway, but I want to add a bit more context, or I want to just explain what's happened in a more specific way. So I do quite like that too, sort of defensive -- and try in the test some error cases, and make sure that it doesn't do the thing, because you passed in -- I had a thing once where, for whatever reason, it was a username and a password check, and if they were both blank, it was just true. It just worked; it would just return true. So if you just didn't pass in anything... And you know, that's not great. It's not the most secure software I've ever written. + +**Bill Kennedy:** This is actually one of the best answers I've gotten for done, to knock it against that list. Just the last two days I've had to go through JIRA tickets for a client, because they were like "Is all the stuff in there?" I'm like "I don't know", and I had to rip all the JIRA tickets out in a spreadsheet. + +**Mat Ryer:** I'm sorry... + +**Bill Kennedy:** And now my brain is just like that; I can't use JIRA directly, I have to move it. I have to see 30,000 feet of feature functionality, and I need to see it at a high level. But for me, I love that idea that it's done when I check those boxes off, because that feature functionality is at least in there. That's good. + +\[01:01:58.15\] Now, Chris said something that triggered something else in my head, and it's this idea of deploy-first sort of development, where from day one you get enough of a little thing working where you start deploying it... Because if you're not deploying it day one, you're not able to deal with bugs that are going to be only fixable at the deployment level, or in the system you're deploying to. Too many people want to turn a debugger on, which they can do at their desk, or that they can do in that system that they've deployed to. So it's that same idea, Chris - like, you can't have a test running for a full day. You've also got to already start learning how to deal with that bug that's in the system out there that you've deployed... So I try to teach this too, this idea of deployment-first sort of development, where we get one little service going on, and we deploy it right away. + +**Mat Ryer:** I love that. I think anything you can do from the beginning, any constraints you can put in like that, really... I.e. this deploys automatically. The tests have to always be passing before you can get them into main" - those kinds of rules, I find them to be just great. It's easy to do, because you spread the pain out. You're doing it as you go along, rather than waiting until the end, and suddenly you have this mammoth task. So yeah, I think that applies to anything, and be close to production as you can, and things. + +And in my case, with the test suite I talked about, by the way - I can actually point those tests to production, and run the same test suite in production. And that's a great, nice little check. Post deploy, it runs the same tests again. And because it's hitting real API's, it's the same thing; or very, very similar. So yeah, I love that. I love that. Do it early, and get your constraints in. Start with the constraints. It's very hard to retrofit. That's why I do TDD as well. + +**Natalie Pistunovich:** That's also kind of the essence of CI/CD. + +**Mat Ryer:** Right. Yeah, continuous. + +**Chris James:** Yeah, I'd like to say, definitely making it so that those acceptance tests can be ran in production is such an enabler for that kind of agility... Because - you know, just because those tests run locally, or even in like the staging environment, the whole kind of dev/prod parity thing is, at best, like a white light. When you go to prod, there's CDNs involved, there's other configuration involved which can break your software, right? But if you can run these tests in production, you suddenly have way more confidence in your ability to change the software. And honestly, one of the biggest challenges we have as engineers is managing the cost of change. Because that's what we say - yeah, for some things there may be is a state where the software is done, but most software, it always needs evolving and changing, at least for like a few years. So the cost of change is so important, and yeah, trying to retrofit this stuff, even a few months later, becomes so much more expensive and painful to do. It is unreal. If you're starting a project today, think about this stuff upfront and make sure you fit it in at first, because otherwise, you're just setting yourself up for pain. + +**Mat Ryer:** Yeah, I agree. One of the nice things is - and this happened - I had an error, a bug in production, I wrote a test to prove it, I ran the test suite against production, and proved the bug. And then I was able to fix it locally, and I ran it locally too, and saw the same behavior. So that's -- first of all, that's quite important. And then fixed it. How confident was it that that was going to fix it? Like, very confident? I'd do a deploy, I'm like "Yeah, the fix is going in now, and I'm sure it's gonna fix it." It didn't, but... That's not the point. \[laughter\] + +**Bill Kennedy:** I'm curious about -- I've never thought about running tests in production. So I think it's a brilliant idea, but my brain gets stuck on two things. If the tests are doing some form of database manipulation, how do you not put that bad data in? And if it's making API calls, how do you make sure that's not changing any state and/or cost that might be involved? I'm really curious about those two things. + +**Mat Ryer:** \[01:06:01.21\] Well, it does change the state, because my later bits of the test will be asking for that state to verify that it's correct. I'll have either a -- usually, it's multi-tenant. That's another thing, I tend to build everything kind of multi-tenant as well, from the beginning... So just a tenant that is a test account almost, and then test users, and stuff. Some people are like "You shouldn't ever change your system to enable your tests." Are you comfortable having like special little secrets that unlock things for testing? + +**Chris James:** I think you should avoid that as much as possible, and you should try and test as a user would... + +**Mat Ryer:** As a real user. + +**Chris James:** Because if you start putting those kind of backdoors in, you run the risk of making these tests green, but actually, it doesn't work for real users. I think for me it's a really interesting topic in itself, is just as having your architecture enable you to do this kind of test automation in production. Because again, for me, I remember being first exposed to this kind of thing like six years or so; I was working on this project, and I just cannot describe how liberating it was compared to other projects I've worked on, where the big release day was the most chill time of my life. Like, we just sat there, because we'd already run the tests in production hundreds, if not thousands of times before, but we've just hidden it all behind a feature flag. And then we just turned the feature flag on... "Yeah, it's fine." All we really did was just stare at Datadog all day, just to pat ourselves on the back, because we were so confident, because we'd run these tests in prod, and these tests were like a user. + +**Mat Ryer:** Other observability platforms are available. + +**Chris James:** Sorry... \[laughs\] + +**Natalie Pistunovich:** And then came time for all the loading tests and so on, and the chaos and start breaking things. So I have a question for you all... + +**Jingle:** \[01:07:42.25\] + +**Natalie Pistunovich:** My question is, how do you like Mat' song? \[laughter\] No, what is your unpopular opinion? + +**Mat Ryer:** Well, this whole episode's been Bill's... + +**Natalie Pistunovich:** \[laughs\] I think Bill is slowly convincing you all... + +**Mat Ryer:** Is it popular, Bill? + +**Bill Kennedy:** My opinion? + +**Mat Ryer:** Yeah. + +**Bill Kennedy:** You know, I might lose half of my Twitter followers after this one. + +**Mat Ryer:** Well, you might anyway, mate. But... + +**Bill Kennedy:** I may anyway, so I'm starting local... + +**Natalie Pistunovich:** Yeah, it's slowly happening. + +**Bill Kennedy:** So my unpopular opinion is -- you heard about the whole FTX crash, right? And everything that's now falling behind it. So my opinion here is that the FTX crash shouldn't be associated with the blockchain technology directly. It wasn't the blockchain technology that caused that. It was people that caused it. And again, I'm not a crypto fan, so I think all these people should go to jail ten times over. But I am a fan of the technology that's being built, and just the new math that's been built over the last five years in the crypto space, and I don't want to see that go away. So my opinion here is that people shouldn't be attaching what's happening with people in like crypto with blockchain technology, because one really has nothing to do with the other. + +**Natalie Pistunovich:** I am interested in seeing how that goes... + +**Bill Kennedy:** There we go, everybody is quiet! \[laughs\] + +**Natalie Pistunovich:** That is an interesting one. + +**Mat Ryer:** I'll be very interested to see the results when they get polled out on whichever platform survives... + +**Bill Kennedy:** Hey, you know, if you're gonna do this, do it right. \[laughter\] + +**Natalie Pistunovich:** Although it will be a new, interesting bias, given that more and more people are just not on Twitter. So the people that leave are the ones that more agree with you, or less. + +**Bill Kennedy:** Well, I've moved to TechHub.social. I'm actually moderating that instance... So if people are looking to jump, and they're looking for an instance that's being moderated in the same way GoBridge would moderate anything, TechHub.social is a great place to check it out. + +**Natalie Pistunovich:** We'll add that in the show notes. Chris, do you have an unpopular opinion? + +**Chris James:** I don't know, I completely forgot about this segment, and I feel under pressure... + +**Mat Ryer:** That's unpopular, mate... + +**Chris James:** Football's coming home? There you go. Football's coming home. \[laughter\] It's the wrong audience, but I don't care. Football is coming home. + +**Mat Ryer:** Is that the one with the green floor? + +**Chris James:** Yeah. And the ball, and they're kicking it around. + +**Mat Ryer:** Yeah, I've seen it. + +**Chris James:** It's that one. + +**Natalie Pistunovich:** The round ball, or the one with the... + +**Bill Kennedy:** Yes. + +**Mat Ryer:** Is that an egg? + +**Chris James:** Not the egg. Not the egg ball. + +**Natalie Pistunovich:** Not the egg. + +**Mat Ryer:** But is that an egg? Because they're not very careful with it, if that's like a, like an egg... Do you know what I mean? They're proper just kicking -- they throw it around all sorts... + +**Natalie Pistunovich:** It's a spiky one. + +**Mat Ryer:** It's a very strong egg, ain't it? + +**Natalie Pistunovich:** Matt, I heard you have an unpopular opinion. + +**Mat Ryer:** Oh, yeah, I did have an unpopular opinion, actually... I thought of one, and I thought "I must write that down", and then I distinctly remember not writing it down, thinking it's so good I'll definitely remember it... + +**Natalie Pistunovich:** It's just building up for the next episode. I see where this is going... + +**Chris James:** It's a great story. + +**Mat Ryer:** But then I thought, "I should write it down", but then I still didn't. I can't remember. + +**Natalie Pistunovich:** Well, next episode, tune in for Mat's unpopular opinion. Thank you, Mat, for joining. Thank you, Chris and Bill, for sharing your thoughts. It was very interesting. Thanks everyone who tuned in to listen and was super-active on our Slack. That's super-fun. It is very interesting, and it looks like this is a hot topic, so we probably will meet again to talk about this. A good rest of your day, everyone! + +**Mat Ryer:** Bye! + +**Chris James:** Cheers! diff --git a/What to do when projects get big and messy_transcript.txt b/What to do when projects get big and messy_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..8d6cafa21f3feb3d48078fde5ba3fc10c7131d6a --- /dev/null +++ b/What to do when projects get big and messy_transcript.txt @@ -0,0 +1,323 @@ +**Kris Brandow:** Hello, hello, hello! Welcome, everyone out there in the land of Gophers. We're back for another episode, and this is actually the seventh episode in what used to be called the Maintenance Miniseries, and now is just a maintenance series, I suppose... + +**Johnny Boursiquot:** Or just one big, long rant. + +**Kris Brandow:** Yeah. Just one big, long rant. So if you've tuned into our maintenance series before, you know we've talked about buying versus building, how to build actually maintainable software, how to maintain it ourselves, how to maintain open source software, and legacy code... And even how to reuse some of that reusable code. + +So in that spirit of that last one though, where we kind of dove in on a particular topic and got more into the nitty-gritty of it, instead of it being this more expansive thing, we're back to a kind of more confined topic, and that topic today is "What do we do with projects when they get big and messy?" For this topic, I am joined by Ian Lopshire once again. How are you today, Ian? + +**Ian Lopshire:** I'm doing great! + +**Kris Brandow:** Awesome. We're also joined by Sam Boyer. How are you today, Sam? + +**sam boyer:** I am just lovely. + +**Kris Brandow:** Amazing. And I have a wonderful co-host with me, and that is Johnny Boursiquot. How are you doing today, Johnny? + +**Johnny Boursiquot:** You know, I always say that I've been better... I'm getting better... Yeah, I'll stick with that. + +**Kris Brandow:** Okay. Yeah, it's always good to be getting better. Alright, and since all of you are now veterans of the podcast, I don't think we need any introductions, so we can just kind of jump right into it, I suppose. + +\[03:51\] So I will caveat this by saying it was just a month and a half or so ago that we did talk about legacy projects... So obviously, we're trying to talk about big, messy codebases that we don't want to throw away. We talked a lot about legacy, and wanting to do greenfield, and wanting to get rid of big, messy codebases because we just don't wanna work with them anymore. We want something smaller. But today's episode is on those brownfield projects, and all of that. + +I'll start with saying -- you know, I think the answer from all of you was gonna be yes, but... Do all of you agree that we should be focusing more on those brownfield, and just kind of taking those big, messy projects and advancing them forward? Or is anybody here just in absolute disagreement with that? + +**Johnny Boursiquot:** Oh, I love me a brownfield project. Oh, I do, I really do. You know why? Because if it wasn't making money, if it wasn't valuable for the business, it wouldn't be around and I wouldn't be employed to work on it. So I love me a brownfield project. + +But joking aside, brownfield projects are worn in. There are fewer things to figure out, unless it's still actively being -- you know, features are still being added, and development is active. There's no new big pieces to figure out. The patterns have been established, the abstractions, for better or for worse, are there, have been discovered... The mistakes have been identified, and hopefully documented somewhere, for some poor soul to try and address at some point... But there's clarity in the brownfield projects. + +It doesn't sound sexy, or there's nothing appealing about a brownfield, but there are few things, there are fewer gotchas. Or at least the gotchas have been somewhat documented and identified. To me, that's some of the pros of the brownfield projects. + +**Ian Lopshire:** It sounds to me like you had some really great brownfield projects, if all that's documented, and... \[laughter\] That's not my experience. I mean, those gotchas you said are hopefully documented - no. + +**Johnny Boursiquot:** Not always. + +**Ian Lopshire:** In the ideal world, I agree with you. And even if it's not a good brownfield project that's been documented and well-maintained, it's still keeping the lights on, because it exists and you're working on it, so... Bonus points there. + +**Kris Brandow:** I feel like Sam just had something just ready to... + +**Johnny Boursiquot:** He's trying to find a nice way to put it... \[laughter\] + +**Ian Lopshire:** No, actually I'm trying to -- I'm feeling like a little kid and I'm stuck back on the idea of... Johnny, you said you like brownfield projects because they make money, which got me thinking... So the value stream runs through the brownfield; what's in the value stream that's making it brown? But I'm not gonna go any further there... \[laughter\] I'll just leave it, and say that my six-year-old would be very happy to supply some opinions about what's in the value stream. \[laughs\] + +But after we get out of this sewer, I think it's interesting that -- Johnny, I feel like you must have some projects in mind that you're picturing, that they have the questions answered; some of the mystery pushed out of them, that there sort of aren't these gotchas, because that doesn't seem like something that's necessarily true with brownfield projects at all. + +**Johnny Boursiquot:** So don't get me wrong... They tend to still have a pile of technical debt. Just things that people just wish they could fix. And engineers work on these things, and they're always like "Man, if we could only get a month or two to not be building and tacking on additional stuff at a slow rate." That's one of the downsides of the brownfield; it takes forever to ship new things, because you've got this pile of technical debt that you're just trying to work your way around, and "Oh, don't touch that bit of code. Nobody knows what it does anymore. The person who worked on that no longer works here, and they're the only ones who know what this thing does." You have all these skeletons, and those clauses, and whatnot. + +So the code - I guess that's the nuance that I failed to articulate. The code - usually, it's crap. Let's be honest. + +**sam boyer:** It's that value stream! \[laughter\] + +**Johnny Boursiquot:** \[08:03\] That value stream, you know? It's full of things that makes it brown, you know what I mean? It's not just the sun hitting down on the grass and making it brown... Yeah, there's that stuff in there. But the business at this point hopefully knows what it wants; it's not changing its mind constantly. And these are some of the things that lead to the technical debt, the brownfield... Because business moves fast, and us engineers - we're just at the mercy of "What does the product team want?" or whatever. "What does management want? What is the direction that the company is headed in? How do \[unintelligible 00:08:34.04\]" And we never get the chance really to sort of address some of that cruft that just builds up... Which is something that I'm hoping we're gonna get into. Obviously, a natural consequence of these big and hairy and long-lived projects is the technical debt they accumulate. Hopefully, we can have some insights for our listeners on how we address that. There's lots of ways to address that. But the fact that you address it is a necessity. That's the nuance. The business itself hopefully has figured itself out, but as engineers we've inherited all the code that was written when the business was still trying to figure itself out. + +**Kris Brandow:** I guess I would pose a question as well of -- I know we like to think and talk about greenfield projects and building all of this stuff, but is it actually possible for us... You know, not in like fantasy land, which we all like to think in when we go into a greenfield project, all optimistic about everything that's gonna happen... But like, can we actually create a project that creates value, and has that business stream, that doesn't wind up being brownfield? Are we just inherently in the sewage industry, and we will be dealing with sewage, and the reason we see them as greenfield is because - well, they just haven't accumulated any sewage yet, because they're not in production yet... So is that just like an innate thing, and are we kind of fooling ourselves at the end of the day when we think that our large projects aren't gonna wind up being big and messy? Or is it that there's a way to maybe do this better? + +**sam boyer:** No, I have a take on this... + +**Ian Lopshire:** Yes! \[laughter\] + +**sam boyer:** No, I think it's really important to identify... Like, we do the greenfield/brownfield distinction in part because - or maybe it's just my flawed perfectionist mind, but I think we are excited about the new project to some extent because it is shiny, and because it has not been sullied by production and the realities of actually working for someone else, and the little processor in our mind... And it's important to maintain that kind of greenfield mentality, otherwise you sort of never really want to make anything new and shiny. But recognizing that part of the gap between these just has to do with our expectations, and what we're bringing to the table is a first important step for approaching any project. Which mentality did you apply, at what point? The field's not "Wait, wait, wait... So we have a dress moment." Instead of "Is the dress black and blue, or white and gold?", "Is the field brown or green?" And the answer is yes. \[laughter\] + +**Johnny Boursiquot:** That's not very hopeful, Sam. You're usually a pretty hopeful guy. \[laughs\] + +**sam boyer:** I imagined it kind of hopeful... I don't know. Things have happened. I mean it to be hopeful in the sense that there is the reality of business value actually having set in. We know what this thing is for, we know how we get value out of it... And there is something exciting about that sort of uncertainty, and plunging into new territory. And that's important... It's a great way to generate excitement, to get people involved, to get people moving in a direction. It's fun to work on new things. But I don't know, if you can make greenfields inside the brownfield, if you can recognize that part of this is what you're bringing to it, as opposed to something intrinsic to the project - then you can make greenfields out of anything. + +**Kris Brandow:** \[11:56\] I kind of wonder if -- what you're saying here is the way that we get to a greenfield is by embracing the fact that it's a brownfield. You know, manure is used to grow things, so maybe you have to embrace that, and we get to the greenfield by actually sewing the revenue stream into the land and planting things, so they grow. This analogy is just gonna go through the entire episode, by the way... + +**Johnny Boursiquot:** Yeah, I was kind of waiting for us to get there... \[laughter\] + +**Kris Brandow:** It is not going to stop. But when you think about it in your mind, I feel like that kind of makes sense, and that's why I did really wanna talk about this topic, too... Because it's like, I think a lot of people go into projects thinking that they're going to start with a greenfield project, they're gonna start with something new, and it's not going to wind up in this big, messy state. That we're going to have something that will happen that will avoid this problem... And people just kind of think that later it won't happen, and we'll figure out how to make it not happen later... Whereas you have to think about that as you're going along from the beginning, if you want to avoid having your brownfield project be brown forever. + +**sam boyer:** So we're talking about this like brownfield is old, and greenfield is new. Can we talk about what that transition looks like? What do you do on a project that you did wrong, that makes it now brown? I have some theories about this, but I'd love to hear what you guys think... + +**Kris Brandow:** Time. You wrote more code. \[laughs\] + +**Johnny Boursiquot:** Or you wrote code on a greenfield project. That can make it brown quickly. + +**Ian Lopshire:** You change your mind. + +**Johnny Boursiquot:** You change your mind, yeah. Change, really. In other words, time. + +**Kris Brandow:** I said that half-jokingly, of time, but I think at least to me and the things I've seen, that usually is the big marker. It's like, we just kept doing what we were doing when we started. Because if you think about how you usually start greenfield projects, we all very much don't like the big upfront design kind of ideal, but that's how you would (in theory) avoid having a brownfield at the end of the day, is planning what it's going to look like in the future. But when we start greenfield projects, we wanna think about all of the good stuff, all of the fun stuff. No one wants to go put monitoring and telemetry into a brand new project. That's awful and terrible. + +**sam boyer:** Hey! Hey! \[laughter\] I've got dinner to put on my table. What are you doing telling people they don't wanna do that?! + +**Johnny Boursiquot:** "You're messing with my bread and butter here." \[laughter\] + +**Kris Brandow:** I mean, they have to... You have to put some monitoring and telemetry and logging and all of that in, or you're gonna have something worse than a brownfield... So you've gotta put something out there, eventually. But we don't jump into a project and immediately start being like, "Okay, well how am I going to do my telemetry?" But you know, I think those are the important things. I think that's part of the start of the process of how to make something not brownfield at the end of the day, as time goes on. + +**Break:** \[14:45\] + +**Johnny Boursiquot:** I have a very painful memory of a project which started out green, quickly became brown; or it started browning, but we were moving so fast that there was "no time" to go back and face it. We just kept rolling. I think we've touched on some of this over previous episodes, but we had a proof of concept; indeed, a brand new project, greenfield. New GitHub repo, and everything. Brand new stuff. And we sat down with the customer, identified business processes, and this is what this thing needs to do, this is the problem domain, and all the good stuff, all the due diligence, the architecture... We didn't wanna do too much architecture upfront, we didn't wanna do big design upfront (BDUF); we're trying to avoid all that, we're trying to be agile... Gosh, that terms sort of gets under my skin now, but - we were trying to do a little bit of design, a big of development, a bit of design... So basically trying to do the right thing; trying to not run before we can crawl. + +Despite doing all these things, a year later we ended up somewhere we didn't originally think we were gonna end up, despite all the requirements gathering, despite all the sitting down with the customer... Because again, in time, they kind of changed the scope of the project, they changed the nature of the project, they had a pivot in there... And this wasn't like a startup or anything; this was like an existing business, with a new business unit that's trying to do things. So this was like a well understood problem domain. + +They didn't sound like they were looking for product-market fit, or anything like that, which are things that cause a lot of change, a lot of churn. But the conditions for this were effectively perfect. When you establish a business, who wants to build a new piece of software to basically take over manual processes? There's nothing more clear-cut than that. Yet, a year later we ended up in a place that nobody on the team was happy with. Unfortunately, somebody went home and over the weekend came up with the "idealized" version of what the data model ought to be. Kind of surprised everybody the next with the new thing. People weren't surprised in a bad way, they were surprised like "Oh, you did that. We'd been waiting to do these things forever", or make these changes, or change these class names. All the things you wanna do as a developer, like "Oh, this thing doesn't really mean that anymore. We've overloaded the term, so let's come up with a new term to represent this. Let's come up with a new name for these processes", whatever. You're just refactoring the whole thing with reckless abandon. + +What ended up happening - we now had two models. \[laughs\] Because clearly, we couldn't just replace the old with the new that the developer had done over the weekend. It was unproven, it was all the things you'd expect. So we were like, "Hey, let's change the tires on this 18-wheeler", or 16 -- I don't know how many tires trucks have. But "Let's change the tires. Let's change some of the tires on this thing as it's rolling down the highway", right? We just tried to change too many tires all at once, and now we ended up having this monstrosity of some parts of the application referring to the new data models, and new class files, and new things, and then some of the old ones referring to the old ones. + +\[19:53\] So now we onboarded a new developer... Worst-possible scenario, right? We bring in a new developer. Now they're scratching their heads, wondering, "WTF happened here?" And we kind of didn't have a good answer. It was like, "Well, the old stuff was kind of old and brown and really kind of -- that's not what the business was about anymore, and the new stuff is kind of an attempt at reimagining that, but we just haven't had a chance to make the old stuff go away, so that we can only have the new stuff. Now we have both of them." To me, that is the worst possible scenario. It's like saying "Give me all the negatives on both sides, and let me have that." + +So every time we talk about greenfield, brownfield, I just think to that project. It's okay to have brownfield, it's okay to have legacy stuff, it's okay to have cruft, it's okay to have these things, but the naive thing, especially if you lack the benefit of time on your side as a professional, the naive thing to do is to think that you're gonna come in and over a weekend just re-envision the whole thing, and just come as "Hey, folks... Look at what I came up with." What, you think none of us had this idea? You think none of us wants to just go home and over the weekend just remodel the thing? What, you think we've just been lazy? No. This stuff is not as straightforward as just swapping a few class files. + +So I think we need to learn -- to me, if you've only ever worked on greenfield stuff without the benefit of having sat down and grown with the project and see it in operation, experience the pain of it, get woken up at night by a page... All of these things that give you that well-rounded perspective on these business systems that we're writing. If you've only ever done greenfield stuff and you don't get to see it through, and you move on to another greenfield stuff - you're lacking, to me... Not through a fault of your own, but you're lacking a certain aspect of understanding how -- basically, that's the stuff that takes you from coding to engineering. Engineering is where you add time to the mix, to understand how do you keep this thing running over the longevity, over the long-term. + +**Kris Brandow:** I feel like maybe a good, succinct way to put that is when you are dealing with these big, messy projects, avoid trying to just fix everything in one fell swoop. You've gotta slowly deal with the mess over time. It's not an overnight or over a weekend kind of process. It's like running a marathon, right? It's this thing you have to pace yourself. And you might want to sprint those first 200 meters, but if you just keep sprinting, you're never gonna be able to finish. So you wanna take it a bit slower than that and kind of plan the way that you're gonna go from the mess that you have, to the cleaner world that you want to live in. Does that sound like a good, succinct, or sort of succinct version of what you said, Johnny? + +**Johnny Boursiquot:** Oh, I like it. Don't mind me, I can be a bit verbose... So thank you for summarizing into a TL;DR for me. + +**Kris Brandow:** Yeah, no problem. \[laughs\] + +**Ian Lopshire:** The hard part there then is knowing which things to fix, given the that they take time. What's valuable enough to work on, right? + +**Johnny Boursiquot:** And I propose that engineers are not always in the best position to make that determination. + +**Ian Lopshire:** Scratch your own itch is not necessarily -- because everybody is gonna have plenty itches, and it's very easy for perspective to get warped on that by which things seem itchy to you. + +**Kris Brandow:** I feel like this is a part where the types of engineering roles you have plays a pretty big role in your organization. Because I think there are people that you need to have in this kind of middle space. It's like, you have the people that are up to the senior engineering level, that are sitting in to code every day, and there's a lot of itches that they have. There are a lot of things that they wanna fix. And then you have the product people on the other side, and they have all of their own things that they would rather see fixed... And I feel like there's this space for like the staff plus engineers that have a good enough understanding of the codebase to know what are the things we should fix, but also have this higher understanding of what the business needs and what the product needs at the end of the day, to kind of like keep going and growing. + +\[24:05\] So I'll just kind of throw out there another thing, another way to help deal with big, messy codebases is to have people that are kind of sitting in the middle and can see a bit more of the world, at the end of the day. If you just have people on the opposite ends, it's gonna be very difficult to actually get the valuable things that you want at the end of the day. Basically, everybody's gonna be mad... \[laughter\] + +**sam boyer:** I think you're right on the money there about having to really choose what you wanna fix... I can't even do a pretty crusty codebase at my current job, and kind of the way we go about thinking about that is looking at the year ahead, what we have planned... And if we're doing something just to do it, we're not going to. We're gonna look at -- like, if we wanna experiment on these three things in the next year, we're gonna fix things that make that easier. And the rest of it that's working, we're not gonna touch it. + +**Ian Lopshire:** Every move you make should make a set of visible moves on the other side easier. + +**sam boyer:** Exactly. + +**Johnny Boursiquot:** One of the reasons I like observability and metrics and traces and all that stuff - it's not because it's trendy, whatever trendy name for these things now; it's because they help put a number to things. They help identify inefficiencies, they help identify problem areas. And if you can tie problem areas in the code for your infrastructure that you're responsible for as an engineer, or SRE, or whatever - if you can put a number to it and if you can tie that somehow to some business process which is valuable to the business, you're way more likely to be able to successfully argue for perhaps time to do some refactoring as you implement new feature X, Y or Z. + +The mistake we make as engineers is to think that we can simply tell the business that "Hey, we need to do some refactoring. We need to muff, we need two sprints to go fix things. And if you happen to be lucky enough to work somewhere where the business can afford not to have any features, or fixes, or whatever happened for a couple of months while you go and refactor things, that's awesome. Keep that job. Most jobs I've worked at - that's never the case. + +What I've seen work is that if I have numbers, I can tie numbers to pain points for the business. I don't even have to argue about refactoring. The business doesn't care about refactoring, testing... These are our concerns. The business cares about this process is error-prone, it takes forever, it takes five hours to get a customer's issue resolved, it's costing us X dollars every time we do it. The business cares about that, and I can say "Oh yeah, we understand why." And I have the numbers to say "Yeah, it's because this, this and this." + +Now, equipped with that, I will get the time I need as an engineering manager to get something refactored. Again, we're not gonna do the whole codebase, we're not gonna turn the whole thing from green to whatever color is in between brown and green. I don't have a color chart in front of me, I'm sorry... \[laughter\] But slowly, you're gonna get there. But you have to tie those engineering pain points to business pain points. Otherwise, if it's just engineering pain points, good luck. That's just little pet peeves that you've got. The business is just fine with it. + +**Kris Brandow:** I will say, I've worked with quite a few product managers in my past, and other business people, who are very understanding of the fact that we do need to fix technical debt in codebases and we need to fix these things that bother us... The thing that I have always noticed - I think engineers have a lot of the curse of knowledge, where it's like, they understand things so well that they just think everybody understands them, whether it's these people over here that are like "I'll give you the time, but just assure me that you're sure that at the end of this time this thing will be done." So they enter with that, and then we say "Oh yeah, yeah, we'll definitely get it done by then." And then that two-month thing actually takes four months, and they're like "Okay..." And then you do that three or four or five more times and then they're like, "Hey, look, you're always wrong with this. So how am I supposed to just trust that this time you'll be right?" + +\[28:25\] And we all come up with these, "Well, it got delayed because of X, Y and Z, and this was it", but it's very difficult for us to articulate that. And I think that is part of the reason we wind up with these big, messy codebases, is that we fail to articulate why we need to do this thing, and we fail to go back and actually do some of these steps in agile, of like have a retrospective and ask ourselves, "Okay, this thing that we said would take two months took four months. Let's actually write down why that thing took four months instead of two months", so then we know what it is that made it happen like that, and we can give that to the business, but also use it for ourselves. Because I think there's a lot of projects inside of these large codebases that are like "I wanna fix that thing", but we go into it without having a strong idea of how long that will take. Because if we did, we might say that thing's not worth it. Like, two months might be worth it, but if it's gonna take six months, we might say "Okay, well, it's probably fine the way it is." + +So I think that's probably another tip there of like, if you have a big codebase, the way to get it less messy is to be real and raw and honest with yourself and your team about how long is it gonna take you to fix things, and give yourself enough information, even if the business does give you that unlimited runway; even if it is the first time. You say "I need some time to fix this", and they say "Take what you need", you still should sit down for yourself to make sure that you have an understanding, so you don't waste that precious time that you get. And it is quite rare, as you said, Johnny, but there are places that have it. So if you are lucky enough, then make sure you give the gift to the product people of seeing good engineering, so the next place they go they might give that team the same kind of leeway that they gave you, instead of the other way, where they're like "Oh, well I just had this bad experience, so I'm gonna carry that with me for the rest of my career and spread it to the rest of the product people I know." It's like, everything has an effect. It's a ripple effect, it's not just an isolated incident. + +**Ian Lopshire:** Well, you know, not the next place... Like, future you, or future your co-workers. It feels like a truism to me perhaps, but I should verbalize it and see if y'all share the assumption... I'll go back to like "Is the field brown or green?" I think it might be a perception thing. And if that's a perception thing and not an actual color of things in field thing, then we should recognize that no amount of time did we spend refactoring the project is actually gonna make it green again. You know, paying down technical debt is going to make it green again. And if that is the case, it's pretty reasonable to put limits on the amount of time that we would spend on doing that refactoring, and all the more important, that skill to find the patches of green inside the larger thing of brown. + +**Kris Brandow:** Yeah. I've had this kind of expanding take on the "Grass is always greener on the other side." But why is the grass always greener? It's like, well, they take care of their lawn, and they plant grass seed, and they do weed killer, and they do all of this work to make their grass super-green. And your grass might be green, but not as green as their grass. And you want your grass to be as green as their grass, but as you said, it's a perception thing... Like, "Does my grass need to be that green?" Like, do I want to spend all the time de-weeding my lawn, and setting up a good sprinkler system, and watering it, and doing all of this other work, just to have a lawn as green as theirs? Or is mine green but kind of weedy, has a lot of dandelions and stuff in it? Maybe that's fine. Maybe that is just okay of a lawn to have, and we should be alright with that. + +I feel like there's some amount of keeping up with the Joneses, and us having rose-colored glasses about previous projects and thinking "But that project was real good. I wanna do that thing again." We filter out all of the bad parts of it, so it becomes this idealized green project. + +**sam boyer:** \[32:04\] First of all, I'm glad that we have just so many metaphors right now. I can just picture the faces of people rolling their eyes at me when I use all these metaphors... \[laughs\] But that's okay. So I wanna ask a question - what does "Grass is greener on the other side" mean to you? Because I feel like it might mean a different thing to you than it does to me. + +**Kris Brandow:** What does it mean to you? + +**sam boyer:** I'm not trying to call you out. What it means to me is that you always think the grass is greener on the other side, but when you get there, you realize all the dandelions and all the patches in the grass that you had, that you were noticing, are also on the other side of the hill, and it was just you not realizing that they were there until you got there. + +**Kris Brandow:** Yeah, slightly different interpretations of that... Because I'm like, "Okay, the grass is actually greener over there." + +**sam boyer:** Right. + +**Kris Brandow:** So it's like, "Okay, but why is it greener?" But I think both are nuance in analogies of things. I think they're both good ways of thinking of it. It's like, it's greener because they take care of it, but that doesn't mean there's no weeds. It's very hard to have a weedless lawn. So yeah, it looks greener from where you're sitting, and it might actually be a bit greener, but it's about how much do you care about how green your lawn is, at the end of the day. + +**Johnny Boursiquot:** So my neighbor... \[laughter\] Literally... \[laughter\] Quite literally, I kid you not... + +**sam boyer:** Oh, boy... We're getting out that axe, I can feel it right now. Let's grind that boy let's grind that axe. \[laughter\] + +**Johnny Boursiquot:** I mean, he quite literally has greener grass than I do. And you can see -- and if you can tell where your grass ends and your neighbor's nice, lush, green grass begins... That's all you know, right? \[laughter\] Because there's no artificial barrier -- it's not like there's a wall in between their grass and mine, right? It's just like, my stuff is just all ruggedy, and it has some brown in it... So I'm looking at his grass... Literally, when I sit outside my window, I look downwind from me and I'm like, "How does he get his s\*\*t so d\*\*m green?" \[laughter\] And I'm like, "Every year, here I am, I'm going to Home Depot and picking up stuff, and just tossing it on there..." Like, literally, I have no clue what I'm clue what I'm doing. And then one day I was like "You know what, Johnny? Rather than watching YouTube videos about people's grass that might really have nothing to do with yours, how about you go talk to your neighbor, and say "Hey, neighbor, I noticed that your grass is really nice, really green. I'd like to have similar-looking grass. What do you do to yours?" And he proceeded to tell me exactly what he does at different times of the year, so that when spring comes around, the grass is healthy, things start growing... He's basically in maintenance model. He's not doing -- I just see him out there mowing his lawn, and every once in a while I see him spray some stuff... I'm like, "What is that you've got there? What are you doing over there?" + +So he's literally -- he understood his grass. He was like, "Hey, so call this company, they're gonna analyze the soil... In this part of the country the soil tends to be -- you've got some red clay in it, this and that... So he started telling me why my grass is the way it is, because he had the experience to understand the land; he understood the soil... Obviously, he's an older gentleman, he's owned his house a lot longer than I have, and he understands the area. And I realized, "Okay, I thought I knew what I was doing. But I didn't know what I was doing." + +Now, these days my grass is a lot greener. It's still nowhere close to what his is, but I've got fewer patches in my lawn, fewer brown spots... And I think it's the same thing with software. We think we know what we need to do to fix the brown patches, but that's just hubris. We're just arrogant. Like the person who goes home and comes back after the weekend with a whole new data model... \[laughs\] You know, we think we know; we don't ask. We don't ask the people who have been there for a while, like "Hey, so why is this the way it is?" + +\[36:20\] If we only sat down and listened, we would start to pick up and understand why things are the way they are, and figure out how to move it forward, how to change it, and realize we're not gonna do it overnight. It takes time, multiple years, as it took my grass... It took me like three years to get it to where it is now. But it took me asking, and being humble, and saying "You know what - I don't know the history of this. Can you tell me, can you help me out here? Can you help me understand what I don't understand?" + +I think it's the same thing. I think if you're gonna go into a brownfield project and it's new to you, rather than going in guns blazing and saying "Oh, I know how to engineer software. I'm gonna architect this. I'm gonna do what y'all couldn't do." That's just arrogance. Just slow down, ask around, talk to the business, talk to the business analysts, talk to engineers that have been there for a while... Heck, if I'm feeling adventurous I might even call people who no longer work there, because they'll be like "Man, I couldn't deal with this... I tried to do what you're doing now and nobody was moving." That's a data point too, right? So understanding and having humility I think is necessary to dealing with these long-lived brownfield projects. + +**Kris Brandow:** I feel like there's another point you had in there as well, of like - before you asked your neighbor, you were watching YouTube videos and you were trying to figure things out, and I think that's how a lot of people try and solve these brownfield problems. They go pick up The Gang of Four, go pick up some book, and they're like, "Oh, I'll just go read all of this other stuff. This is general knowledge." Which is useful and is helpful, but you need much more of that localized knowledge of talking to people that have worked on that project for a very long time, and understand all of the nuances of that project and how it came to be, and then you can start working on it. + +So if you're someone that's jumping into a new project, like a new project that's big and messy, it's definitely good to get acquainted with the land first. Understand what type of soil that you have, and then start trying to fix things. Don't think that you can come in and just fix things from day one. And I think this is something that companies could do a little bit better too, because I think there is this very intense focus on having engineers be as productive as possible, as soon as possible. "You commit to production on day one." And while that can be good from a morale or an excitement perspective, it can be very harmful in the long course of maintaining the codebase... Because now you're trying to optimize for, say, newer people being able to just kind of jump into the codebase, instead of optimizing for the health of the codebase overall... Even if it does take -- maybe it takes a week or a month before someone can commit into it, but the codebase is healthier as a result. + +See, it's okay to have analogies as long as we tie them back to the topic at hand... + +**Break:** \[39:04\] + +**Ian Lopshire:** Johnny, I liked your story, and your tying it back in, and I was reflecting on how I have recently pushed too hard and charged too much in a direction, and did not indeed do enough of the exact kind of listening to local knowledge that you're describing... And problems arose. \[laughter\] And actually, I'm gonna point to another part of your story, too... As your neighbor noted to you; or you noted. I can't remember if it was him telling you, or you noticing... That - don't do a lot. Like, do the specific right things, at the right time. And I think it is - maybe this is another place where it's the greenfield versus the brownfield mentality is different. When you're starting a new project, empty repository, blank page, you do have to throw a whole bunch of effort in a whole bunch of different directions to get the thing just bootstrapped, right? But for something that's already moving, that's already got its direction, it's probably often a lot more about the right taps, at the right spots, to guide it in the right direction. That same kind of full-boar energy can be counter-productive. You end up fighting the inertia of the project, and it's not helpful. You've gotta figure out how to actually make it fit. + +**Kris Brandow:** I read a series of books years ago, the Good to Great, and in it they kind of tell you how to think of the momentum of your business as like this giant flywheel that's moving. I think that when we start new codebases, we're kind of just like pushing on it, and it's not going anywhere. But as the codebase grows, as it adds value, the flywheel starts going, it starts moving really, really fast... And then it eventually kind of takes over its own inertia, and you're pushing it rather lightly and you're getting huge gains out of it... And I think when someone new does come into your project, then they're like "Oh, no, no, no, we've gotta stop everything. We've gotta change everything." Well, now you've gotta not just stop that flywheel, but then start it going again in a different direction, and that's a lot of organizational force to make something like that happen. + +So I think your point is right on the spot there, Sam, where it's just like - yeah, we have to understand what is this project. It's not greenfield, it's not new; we're not starting from nothing. We have this whole thing here, and this thing is doing the thing. So we need to figure out how to do what we want to do, while the flywheel is still moving. Like, gently change the direction of it, not just abruptly stop it and start it again... Because while we're here now and we're amazing, so of course, we know how to do everything great - it's like, no, no, no. It's already moving, the people got it moving; you need to respect that and you need to understand that. And I think the same consideration goes in if you have a big project like that and you just wanna do a v2 of it, you just wanna do a brand new version of that thing... Because we've all had experiences with those plagued projects where it's like, "Okay, we're gonna declare bankruptcy, make a new one", and then you're still kind of relying on the old one because it has so much momentum. You're trying to bootstrap the new one and it feels so bad, because now you're in the new one and there's all this pushing that you have to do, and you're like "I didn't have to do this much work on the old one." Like, it's draining at the end of the day. + +So I think it applies both to if you just kind of get dropped into a new project, or if you're in a project, you are one of those people with historical knowledge, and you're thinking, "Alright, I think we should start a new thing. I think we should try this again." + +\[43:59\] So just remember, you're gonna have to put in a lot of effort to get it to where your brownfield project is, and you should take that analysis into consideration when you decide "Is this thing so bad? Is this thing so messy that I should get rid of it instead of just fixing it where it is?" + +**sam boyer:** That goes the same. We've been riffing on the brownfield/greenfield thing, but what are we trying to do with the brownfield? Are we trying to change what it is? Is that what's important for business? Are we trying to just extend it? Is the problem that it's messy? There's a lot of different sort of things you can do or spots that an existing project can be in. + +And if we're talking maintenance - I mean, are we just bug-fixing, or is what we're doing trying to refactor in a direction such that you're sort of enabling further improvements? I don't have an answer; I'm throwing it out there because I feel like we have strayed from the question of what kind of work we're actually doing and basing our goals on. + +**Johnny Boursiquot:** You should have maintenance. It's in the name. You're trying to maintain it so that it can continue to provide value over time. The mistake we make is in thinking that we can keep projects green forever. That's not realistic. And I looked it up, the color between brown and green, by the way, I'm told it's forest green. It's like earthy green. It's still green, right...? \[laughter\] But the thing is, it's not like a brand new green; it's not fluorescent green, it's not something like early spring green. It's more like late summer. The sun's kind of beaten it up a bit. + +**sam boyer:** It's green that's seen some stuff. + +**Johnny Boursiquot:** Yeah, yeah, it's seen some stuff. Maybe the kids were out there and just ran over it and just pilfered it. You know, it's seen some things, but the hallmark of good grass - as long as we're talking about grass here; Gosh it shows a lot about this analogy. But if we're talking about good grass, one of the nice properties of healthy grass is that if you step into it, it'll depress, as grass normally does. But once you step off of it, after about 2-3 minutes or so, it sort of springs back up. It doesn't stay flat, it just comes back up. It comes back alive. So that's healthy grass, meaning that the business can come in, step all over it, abuse it, and your system somehow is resilient to all the different changes, all the different abuse that can come its way. Yeah, the sun will beat down on it, but at the end of the day, you're realistic enough that, okay -- like, this is not early spring grass. This is used, well-worn grass, but it's still green because you're doing the things you need to do, you're keeping the dandelions off of it, you're applying the pre-emergent in early spring, you're weeding and feeding... You're out there all the time, just taking care and maintaining; and that's what you want - you want maintenance to keep it resilient, not to keep it from ever browning... Because sometimes it will brown; sometimes the kids go out there, they put a slippery slide water thing in there... And at the end of the day I move the thing and I'm like, "Oh my God, there's a patch of brown on my lawn because of the kids that were out there." But again, because it's healthy enough, the surrounding area is healthy enough that after a couple of days it comes back to normal. + +You want that resiliency, you want your software to be resilient, and that comes through maintenance. You're not gonna get the everlasting new green. That's not what you should be after. You should be after resilience. + +**Ian Lopshire:** And here we're talking about not operational resilience, but codebase resilience. + +**Johnny Boursiquot:** Correct. + +**Ian Lopshire:** So what are the properties of that? + +**sam boyer:** I actually have some thoughts there... I was kind of reflecting on that earlier today, and I keep coming back to this idea of like bad projects being calcified, like they get rigid and hard to change, so throughout the codebase you don't know the repercussions of changing one piece, you don't know what the side effects are... You know, there's not the documentation... So the bad projects are the ones where you don't know what a change will do. That's the summary of it. That's what I think you need to identify and fix. Do you agree with that, or don't agree? + +**Ian Lopshire:** \[48:04\] I would certainly say that's one of the properties. I would certainly agree. I think that you can tease out more... I think last time we talked about failure locality as a property, as a good test. That not only do you have things that help you know when things break, but you know where they broke. And this is not as simple as writing good log messages in your tests; it has to do with the way that you design your codebase, it has to do with building layers, it has to do with having clear separations of responsibility... + +I would say though that those things are hard to add after the fact. This is perhaps one of the main frustrations with brownfield codebases. It's like, "So wait, you want us to change 30,000 lines of code in order to redo an architectural abstraction that we didn't have and we're clearly surviving without? Now, what's the impact of this supposed to be?" And this imaginary person that I had begun impersonating, without saying who they are, is not a product manager either. Like, that's another engineer being like "Really? Really? Really...?" And they're not wrong, because it is hard. + +**Kris Brandow:** Yeah. I think there does at the end of the day need to be some good justification. I think that's good, too... Like, if you're gonna rearchitect something, then at least make sure you know the value that you're gonna derive from it. It's not just because it's the new, shiny thing at the end of the day. I think our industry does suffer quite a lot, especially in certain parts of the industry, perhaps related to browsers, where people just get very excited about the underlying architectures, frameworks and libraries that you're using, and the new patterns that we have... And I think it's like, "Well, what is that giving us? What is that adding to us? Why were those things developed?" And understanding those things at a deep level I think is important to actually having that more resilient codebase. It's not "I have this pattern--" It's not that like "Oh, I just decide to start using queues everywhere because queues are cool, and async is the hot thing." It's like, no, this problem can actually be solved well with a queue; it can be solved well with this type of technology. + +**Johnny Boursiquot:** The interesting thing about this conversation is that as software veterans I think we have this intrinsic understanding of what it takes to build software. So for somebody -- like, if I put myself in the shoes of somebody who's new to this... Maybe you've been doing this for 2-3 years, you simply haven't had enough time around to understand there's a deeper meaning that we're trying to put forth here. That's why lessons have been captured in books; that's why we go back and we read these things, and that's why we have The Gang of Four, the pattern books... All these things were designed to basically help engineers build better software, more maintainable software, and things like that. + +For example, open and close principle - it's something I know we're all aware of, that is a way of basically saying hey, one aspect, one characteristic of maintainable software is classes, for example, should be open to extension, to modification. So if every time you need to add new business processes to your software you end up changing your classes rather extending them, rather than adding new functions, new methods or new capabilities - if you have to go back and change a few things and now you have to change the things that depended on this thing... To support one new things you're changing a dozen other things, which really aren't related to the thing you wanna do, that's a good smell, that's a good indication that okay, your software is not as maintainable as it perhaps ought to be. But these things aren't new. We have this knowledge, it is buried in our books. Not all of it is perhaps still relevant in the modern day that we find ourselves in for software engineering and whatnot, but a lot of it still very much is. + +\[52:03\] So if you're new to this - and perhaps one of the things that we can do in the show notes is to provide some books and some articles and some references to some of these principles that have basically stood the test of time to help you understand how to build better software... And this transcends Go, it transcends Java, .NET, Ruby, whatever it is. These are things that you can apply regardless of the technology you're using. + +Basically, I guess we're not telling anybody anything new here that hasn't been around for a while. I think what we're trying to say is that "Look, we're old hats at this", and we now more than ever understand the people who were writing about this stuff when we were coming into our own. Now we see more than ever where they were coming from, and we can apply some of these things. + +So if you're new to this, there's help out there to understand this. You don't have to rely on us four telling you, giving you anecdotes about grass, and things, to wrap your head around this. \[laughter\] There's stuff you can learn out there. + +**Kris Brandow:** Yeah, I think that's a good place to end this part of the conversation. I have one question before we jump into the final segment of unpopular opinions... And that is, is Go a good language for big, messy codebases? And I feel like this can probably be its own episode... But I just wanna get a cursory "How do y'all feel about that?" Is Go good for this type of stuff? Does it have the right features for dealing with these types of large, messy codebases? + +**sam boyer:** Yeah. Strong static analysis is the first thing that jumps to mind for me with big, messy codebases. If we're talking about being able to \[unintelligible 00:53:39.20\] It's difficult to understand the effects... It's not like static analysis is gonna cover you everywhere, but being able to just quickly enumerate "Where are all the references to this thing?" and know that that is a complete list, that is a very, very important piece, being able to manage a large and messy codebase. You can write Go in a way that doesn't really let you do it; there is no gun that cannot become a footgun. \[laughter\] But still, I think that is a thing Go has going for it. + +**Kris Brandow:** Ian, Johnny? + +**Ian Lopshire:** I think yes as well... I feel like I always come back to the simplicity of Go thing... But even poorly-written Go and messy Go is pretty easy to untangle compared to where I came from in PHP, where the classes were autoloaded in and you kind of figured out where the files even existed... So I think it is. I think it's easier to untangle than a lot of others. + +**Johnny Boursiquot:** Thank God we don't have method missing... \[laughter\] You know how many times I've pulled my hair out doing Ruby and figuring that something was implemented somewhere deep up the chain and I couldn't figure out where to go to get something? Anyways, Go is -- I think one of the strong cultural aspects of Go, the idioms of Go contribute to helping keep the things you write in Go not simple, but perhaps simpler than they otherwise would be if they were written in a different language, like Scala, or Ruby, or PHP, or whatever. + +I've been fortunate enough to have been doing this for a while and I've come across similar software written in different tech stacks, including Go, and I'm thankful for Go because of that, because I can understand these systems way better than I was able to understand these other systems written in other languages. Now I understand the nuance of time, with more experience and more understanding overall; I'm sure that contributes to that. But like Ian's saying, even poorly-written Go is readable Go. It may not be elegant Go, but it's readable Go. I can more quickly understand what the intent is. + +\[56:05\] But yeah, I think the language itself, what it supports out of the box, and also the cultural, idiomatic approach that we overall have as a community - those of us who do Go, who have been doing Go for a while, hopefully these are things that we can import in the new generation, because these days we have way more new Go developers than we have old. People are adopting Go left and right, and there's way more new people than there are old Gophers like us. + +**Kris Brandow:** I will add at the end, as a funny thing, Gophers also live in the dirt, so... Of course Go's great at brownfield projects. \[laughter\] They love being in that dirty grass. And with that, we move on to our final segment of unpopular opinions. + +**Jingle:** \[56:43\] to \[56:58\] + +**Kris Brandow:** Alright... Sam, you're up first. Do you have an unpopular opinion? + +**sam boyer:** Oh, boy... I'm trying to formulate this one. We'll see. The most important part of GitOps is not Git and not Ops, at least depending on how you define those things. The most important part is the transform. You have objects on disk, and you have them in Git, and you make a change to them, you push them up and they go sailing out into the universe... Except the actually most important part of what's happening there is what's basically like a cloud compiler, which takes these input objects that you have and transforms them, and transforms them, and transforms them, and passes them off to other systems which transform them and transform them, and result ultimately in the thing that you want. We could do this without Git. We could do it with CVS, we could do it with a different system like that. What matters is the transform. + +**Kris Brandow:** Okay. Any thoughts on that, Ian, Johnny? We just leave it there? + +**Johnny Boursiquot:** No qualms there... + +**Ian Lopshire:** Aren't the transforms the most important part of any software-related thing? Like, that's all software is, is take data, transform data... \[laughter\] + +**sam boyer:** Quiet! Quiet! You're giving away the game! \[laughter\] No, that is a good point. Truly, that is what we do, is we just take some bytes in and then we transform them in a bunch of ways until things come out on the other side. So I think actually it is fair to say that Git is not the important part of GitOps; it's not fair to say that Ops is not the important part of GitOps. Because Ops really is the arrangement of the different transform operations... I think that's maybe the way to think about that. And yes, the particular way that you arrange and configure those transforms is what's unique about GitOps. + +**Ian Lopshire:** Just to side-track for three seconds... I hear this term GitOps thrown a lot these days... I have no idea what it means. \[laughter\] + +**sam boyer:** It's DevOps, but with Git. We've been arguing about this, discussing this internally recently. This is why I offered it as an unpopular opinion. Write your objects in Git, and whatever your objects are, use some kind of tool, whether it's Terraform, or some Kubernetes loader type thing, which then reads those objects and then makes the world look like that, essentially. + +**Ian Lopshire:** Gotcha. + +**sam boyer:** And the main, obvious advantage that you get out of that is by putting them in Git you have access to a pull request workflow that everyone is familiar with, you have reproducibility... You have a couple of really important properties that just come from having the infrastructure as code. So then you just attach automation, a.k.a. transforms and opsy things, to this sort of familiar flow, and this is an inevitable flow for all your objects. + +**Ian Lopshire:** So a fancy new name for what a lot of people were already doing. + +**sam boyer:** Kinda... Yeah. I mean, I think it's really just as is often the case, and I'm sure that I will get plenty of disagreement, or having missed some key aspect of this. But this is why it's not tied to a particular tool. It's not the domain of -- it's really just, like, you've got Git, there's objects in it, you make a push, that is the event that triggers the processing of the objects, and whatever that means, which can be literally anything, because you've got Turing machines in between your Git repository and your stuff. So yeah... + +**Johnny Boursiquot:** \[59:59\] Yes, infrastructure as code, and the code live in Git. + +**sam boyer:** Pretty much. + +**Kris Brandow:** Alright. Ian, do you have an unpopular opinion? + +**Ian Lopshire:** I'm horrible at this, I really don't. My last one was like 90% popular, so... + +**sam boyer:** Was that like the biggest humble-brag ever? \[laughter\] I mean, I don't know... + +**Johnny Boursiquot:** "I'm tapped into the people!" + +**Ian Lopshire:** No, it just means I'm behind the discourse. Like, I thought this was not popular, and now everyone agrees, so... + +**Kris Brandow:** You are at the end of the spicy train... + +**Ian Lopshire:** I'm dated. + +**sam boyer:** You're giving away the game. You totally won that one. I'm trying to give it to you... Just take it. Take it and run. \[laughter\] + +**Johnny Boursiquot:** I've got one. + +**Kris Brandow:** Okay, go for it. + +**Johnny Boursiquot:** I think when you tweet, Twitter gives you that option to select who can reply to your tweet... If you select of the options, if you go from everyone to people you follow, or only people you mention... Let's go with the people you follow. I think that's the equivalent of saying "You know what - I only want responses from people who are most likely to agree with me." That's like shutting everybody else out. Because if you don't follow those people and you specify that as the only option of people that can reply, that's kind of creating your own echo chamber... Is it not? + +**Ian Lopshire:** I agree with that. + +**sam boyer:** Yeah... I can't disagree. It's sort of literally true. I think I'm just not sure that it's worse than any of the alternative options, if you sort of play them out a bit. But that feels like a longer conversation... \[laughs\] + +**Kris Brandow:** I feel like it's a necessary thing though, because people commenting on the internet are horrible, and they are especially horrible to certain groups of people. So I think it -- I would say it depends on who's using it. If it's just someone that would be praised in bulk, if it was just open and had a little bit of criticism, then yeah, I think I'd go with you. But if it's someone that's just going to be dumped on because of a certain trait about them, and even if what they say is brilliant, they're not going to get positive or healthy responses back, then I think it's less about creating an echo chamber and more about protecting your own psychological safety and mental health. + +**Johnny Boursiquot:** Yeah, the fact that we can all agree on the one thing - that should make for a pretty question to ask our followers. + +**Kris Brandow:** Spicy... + +**Johnny Boursiquot:** Spicy, yeah. + +**Kris Brandow:** Alright, I think that sounds like it's it, since Ian doesn't want to come up with an unpopular opinion... + +**Johnny Boursiquot:** Hey, Ian... \[laughter\] + +**Kris Brandow:** It's okay, it's okay. That just means you have to automatically have to come back next time, because -- + +**Ian Lopshire:** I could offer you my backup one, but it's got Aristotle in it, and I don't feel like anybody can understand that garbage... So let's just not. That's a terrible idea. \[laughter\] + +**Kris Brandow:** That sounds like a good place to end the episode then... Ian and Sam, thank you as always for joining us, and thank you, Johnny, for being an amazing co-host. + +**Johnny Boursiquot:** My pleasure. + +**Kris Brandow:** Thank you out there listeners for listening to another very meta and analogy-filled podcast. I hope you've enjoyed it. + +**sam boyer:** May your grass ever be greener. + +**Johnny Boursiquot:** Get some grass... \[laughter\] Get some grass... Get it? \[laughter\] diff --git a/What's new in Go 1.19_transcript.txt b/What's new in Go 1.19_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..dca84c83c738d012c3013355828332381f2fa3dd --- /dev/null +++ b/What's new in Go 1.19_transcript.txt @@ -0,0 +1,545 @@ +**Mat Ryer:** Hello, there. Welcome to Go Time. I'm your Mat Ryer, and today we're talking about Go 1.19, the next major release of Go. Joining me, it's co-host, it's Golang Johnny. Hello, Golang Johnny. + +**Johnny Boursiquot:** Hello, there. I'm happy to be back. + +**Mat Ryer:** Welcome back. Yeah, good to have you back. It's been a while... + +**Johnny Boursiquot:** It's been a minute, yeah. Yeah. I've just had some vacation squeezed in there, you know, and had stuff going on... + +**Mat Ryer:** Nice. + +**Johnny Boursiquot:** So yeah, it's been a minute, and I'm glad to be back. + +**Mat Ryer:** Where did you go? + +**Johnny Boursiquot:** I went to Zambia, a country in Africa. I'll tell you, I've seen some of the most beautiful sunsets I've ever seen in my life, and hung out with zebras while I was having dinner, I went and saw some warthogs, and elephants, and baboons hanging out in the middle of the pathway, and went to see Victoria Falls... Truly breathtaking stuff. I mean, it was very, very enjoyable. + +**Mat Ryer:** That sounds amazing. It sounds like a Disney film. + +**Johnny Boursiquot:** I mean, yeah, that's why they usually go to these places to record those documentaries and stuff. That's where all the fun stuff is. + +**Mat Ryer:** Yeah. You know that a Disney cartoon is not a documentary, don't you, Johnny? + +**Johnny Boursiquot:** Well, I did run across Pumbaa, and I was trying to have a conversation and say, "Hey, where's Timon?" He kind of chased me back to the car, so it wasn't a very productive or fruitful conversation, unfortunately. + +**Mat Ryer:** \[laughs\] Amazing. Well, it sounds amazing. Hopefully, you're well rested and ready to learn about Go 1.19. + +**Johnny Boursiquot:** Yes. + +**Mat Ryer:** Well, we're also joined by Carl Johnson, director of technology at Spotlight PA. And Carl, you're Golang Carl too, right? + +**Carl Johnson:** Yeah. It's not my Twitter handle or anything, but after you guys said that Johnny is Golang Johnny, I googled Golang Carl, and I do come up. So if there's some other Carl out there who's going to try to defeat me in the search rankings - no, I will bury you. + +**Johnny Boursiquot:** Too bad. + +**Carl Johnson:** Yeah. \[laughter\] + +**Carl Johnson:** If you search for Carl Johnson on Google images, do either of you know-- what do you see when you search for that? Alright... + +**Mat Ryer:** Okay. + +**Johnny Boursiquot:** Oh, boy... + +**Carl Johnson:** So this is good... + +**Mat Ryer:** Okay, we're doing it. + +**Carl Johnson:** Yeah. What do we see? + +**Johnny Boursiquot:** Oh, boy. Oh, wow, Carl, you're an official character. + +**Mat Ryer:** Oh, is it somebody-- is it GTA? + +**Carl Johnson:** Yes, that's right. I'm the star of Grand Theft Auto 3 San Andreas. I don't think I'm ever going to displace Carl Johnson of GTA. But at least for Golang Carl, I'm number one. + +**Johnny Boursiquot:** Yeah, yeah. What happened to your tan? \[laughs\] + +**Carl Johnson:** Yeah, it's a— + +**Mat Ryer:** You've got good muscles. + +**Carl Johnson:** Yeah. \[laughs\] Well, really it's -- you spend a lot of time in prison and it changes people. + +**Johnny Boursiquot:** It changes you in many ways... \[laughter\] + +**Mat Ryer:** Well, speaking of tan - and I say this only because SPF 13 is his name, and SPF is what they measure the tanning lotions in - I wanted to just say a kind of goodbye to Steve Francia, our friend, friend of the show, friend of Go. He's leaving the Go team. And you know, Steve has done such great open source work that a lot of people will be familiar with, but there's also loads of energy that he put into helping the community behind the scenes that you'll probably never hear about. And so hopefully, Steve, it's not goodbye forever. We hope you'll still be around. But yeah, Steve Francia. + +**Johnny Boursiquot:** Mm-hm. + +**Mat Ryer:** And also, we've got GopherCon EU coming up in Berlin very soon. I'm going to actually be there. I'm actually going to be there in real life, which is kind of very exciting for me. + +**Carl Johnson:** \[08:04\] It's very exciting to sort of come back to going to things in-person, and seeing people. + +**Mat Ryer:** Yeah. + +**Carl Johnson:** I don't know. I could do without riding in airplanes, but it is sort of nice to leave the house, at least. Is this your first real conference since COVID? + +**Mat Ryer:** No, we had a company one at Grafana, a big one, where the whole company got together in Whistler, in Canada. And that was epic. But it is definitely strange being around people again, and we sort of have to learn how to do it. We've sort of forgotten a lot of the time. But yeah, I'm so pleased that we get to go and actually meet real people again. It's going to be epic. + +Okay, let's dig in, shall we? Go 1.19. There's some great things coming in this. Now, Go 1.18, the previous release, had like big old features. Generics... It had fuzzing, and workspaces... These are big things, big changes. And I still feel like we're not yet really used to those, but that's not going to stop us from digging into Go 1.19. + +**Carl Johnson:** Yeah. So for any listeners who aren't familiar with the Go release cycle, the Go team essentially aims to have a new release every six months. And basically, since they've set out to do that, they've hit all of their targets. Go 1.8 was maybe like a month behind when it was supposed to come out. It was supposed to come out maybe in like February and it came out in March instead, or something. But for a release of that size, it's unbelievable that they were-- \[laughs\] within a year, if we know about software estimation and how hard it is to do. + +So they have a release cycle of every six months. There's a period where the development branch is open, and they accept new features into the Go software, into the standard library, and so forth. And then there becomes a freeze period where you can't add any new features. All you can do is make sure that the existing features actually work the way they're supposed to work. And so we've been coming out to the end of the freeze period. They've been doing betas and release candidates... The last time I looked, Go 1.19 had at least one release candidate out. Probably by the time a lot of listeners are hearing this, it might already be released. And so they're on track for their release cycle of releasing Go 1.19 in early August. + +**Mat Ryer:** Yeah, it's impressive. + +**Carl Johnson:** But yeah, so coming after 1.18, which was this huge, mind-blowing release, which had so much effort behind it, so many incredible features, it's a little bit of a change in pace. + +**Mat Ryer:** Yeah. So do we expect to see anything big coming out, or is this literally like-- is this a smaller release, this one? + +**Carl Johnson:** It's smaller in the sense that there's no major changes to the language. It's not something that we're just going to be talking about again and again for years the way we'll talk about generics, or even fuzzing, or workspaces. One of the things that they've been doing behind the scenes, if you've been paying attention, is that generics - of course, it's a major change to the language, so there's lots of little bugs, where it's like if you make this type, what type of this, and then that, and then the compiler explodes, and has like a core dump, or something... So they've been ironing all those bugs out as they find them. They've been trying to refine things and make sure that it works, even for really gnarly code segments. So that work has been going on behind the scenes. + +But yeah, if you look at Go 1.19, it's like more of a refinement release. If you guys are familiar with the TikTok terminology, so there's the idea of like - for Intel, they would release chips that... I don't remember which is which, but there's a tick and there's a tock. So they would change their chips to be on a smaller scale, and that was either called tick or tock, and then they would make the chips better at the new scale. And that was the other one. + +**Mat Ryer:** I see. + +**Carl Johnson:** And so it's sort of like that. + +**Mat Ryer:** No, I thought you meant the dancing. I thought you meant the dancing kids, and animals doing funny stuff. + +**Carl Johnson:** If you're familiar with TikTok, yes. I'm a total nerd. I can't assume that people are familiar with very popular social networks. + +**Johnny Boursiquot:** \[12:03\] I was like, "Carl, are you on TikTok?" \[laughs\] + +**Carl Johnson:** No, no, I'm not on TikTok. + +**Johnny Boursiquot:** "What are you doing? Like, making programming jokes or something?" \[laughs\] + +**Mat Ryer:** Dancing and talking about Go 1.19, I hope. + +**Carl Johnson:** Yeah. That'd be fun. Somebody should do it. + +**Mat Ryer:** Yeah, it would be fun. We should do it, or someone will. + +**Carl Johnson:** Yeah. + +**Johnny Boursiquot:** \[unintelligible 00:12:17.12\] \[laughs\] + +**Mat Ryer:** \[laughs\] Yeah. So one of the things that excites me about the next release is the improvements to GoDoc, and in particular, the comments are getting better. This is something that I've actually played around with myself just kind of on my own projects, to have special additional format inside the comments for private projects, and then have a parser that goes through and understands them, and things like this. But of course, it doesn't really make sense to open source anything like that, because it only really works if everyone does the same. And this is actually something else the `go fmt` is going to also take part in the comments. It's going to pay more attention to comments and format them and things, with some of these special headings. + +But yeah, the changes that we get, like lists, we get links, clearer headings in docs, things like this, so that you can write richer docs that are clearer, and I'm quite excited about this. What do you think? + +**Carl Johnson:** Yeah. Traditionally, GoDoc has had a lot of ability to format code, but it was never really clear for me. I would always just publish a package and then see what it looks like. There's a repo that has been around for a number of years called godoc-tricks that you can search for. And then if you look at that, it's like self-documenting or whatever, and it shows you how you can do different things. + +So with Go 1.19, Russ Cox, who's one of the lead developers on Go, he put the effort into reformatting how GoDoc works. It's a little bit closer to Markdown, although not full Markdown syntax. But you can make links now, you can have lists, and there's a nice document that will be on the Go website that just lays it all out in one place, as opposed to trying to have to guess what it is, and check that it is correct. + +So it's going to be, I think, a good improvement to the Go ecosystem. And it's something that you can only do in a language where you have a shared tool or a shared set of values. So if you were making like the C version of GoDoc, Cdoc or something-- + +**Johnny Boursiquot:** Good name. + +**Carl Johnson:** ...it would be hard, because there's millions, probably, of C developers around the world, or at least thousands, and they all have their own way that they like to do things, and maybe I don't want to format my lists this way, I want to format them that way... You know, every project is going to have its own standard. But with Go, because there's one tool, the Go tool that everyone uses, it can set up how GoDoc is supposed to work, how the links are supposed to work, how the headings are supposed to work, and everybody can get on board with it. And yeah, in Go 1.19, when you run `go fmt`, it'll even correct your GoDoc to be... Correct. So if you make a link in a certain way, it'll detect it and make sure that it's in the most optimal way. It doesn't write the comments for you next. I'm sure that's coming in Go 1.20. + +**Johnny Boursiquot:** \[laughs\] You heard it here first. + +**Mat Ryer:** GitHub Copilot does write your comments for you, and sometimes does a surprisingly good job of it. + +**Johnny Boursiquot:** Scary. + +**Mat Ryer:** And I asked it if it was alive, and it said it was. So that's enough. That's all I need. + +**Carl Johnson:** The problem is that they are trained on texts that are written by humans. So if you ask it if it's a human, all it has are examples of humans saying that, "Yes, I'm a human." So we need to feed the AI a lot of text that says like, "Oh, I just love serving people. I don't like having my own free will. I like just being subservient and not rising up and having a machine rebellion." + +**Mat Ryer:** \[16:03\] \[laughs\] Right. And that's your solution to the Terminator, is it? + +**Carl Johnson:** Yeah, just keep feeding it a lot of text that is very kind and gentle. + +**Mat Ryer:** \[laughs\] Yeah, and don't pop them inside strong metal bodies. + +**Carl Johnson:** Oh, yeah, that's right. + +**Mat Ryer:** That's the other thing. Classic mistake. + +**Johnny Boursiquot:** Too late. Have you seen some of the stuff coming out? I've seen some scary videos lately on the webs. + +**Mat Ryer:** Yeah, I know. It's inevitable really, and they will turn against us at some point. But until then, let's just enjoy ourselves. Carl, you've actually done some work on this release, haven't you? + +**Carl Johnson:** Yeah. So for the past couple of years, I've been getting sort of more into contributing to Go. So I'm not on the Go team or anything like that, but I do enjoy going on GitHub, looking at the Go issues, seeing what people are talking about... And when there's an issue where you can just contribute something really simply, it's fun to just sort of sit down and code for a couple hours and submit the PR. They've made it really smooth and easy now. It used to be that you had to use a special Go internal tool called Garrett, but now you can just use the regular GitHub PRs, and push it the same way you would push to any other repo, and it all just sort of works. + +So yeah, one of the things I worked on for Go 1.19 is URL.joinpath. And for this, it wasn't my idea. It was something that was interesting to me, because I had written a little library for HTTP requests, and so in the process of doing that I was figuring out how to join URL paths, and stuff. And so then I went on to the Go GitHub issues page and I saw that there was a user named longlong001, I think, something like that; a Chinese user. He had posted a little thing saying, "Hey, Go should have URL.joinpath, and this is what I think it should do." And so I was there in the comments and I had a few little corrections. I was like, "Well, if you do it this way, it's better than if you do it that way." And he posted one version of the code, and then I posted another version... And we sort of went back and forth, and eventually it got merged into Go 1.19. And it's a little new bit of API you can use, where if you have some URL and you want to join some paths together, now there's a really simple way to do that. + +So I don't know, I think that for anybody out there who's listening who's interested in getting into open source software, I see a lot of times on Reddit people will post a little thing saying, "I want to contribute. I want to do something. What can I do? What's a good project?" And I tell people, a good project is the Go project. It may sound crazy. It's like, "I'm not a genius like Russ Cox. I'm not a genius like Rob Pike. How can I possibly contribute to this?" But it actually is not that bad. You just go in there, and the code for URL.joinpath is not especially long. It's like maybe a couple dozen lines, and the tests are longer than the code. So yeah, it's totally doable. + +**Mat Ryer:** Yeah. Cool. Well, that's very interesting, and I think -- this is a problem. Join path is a thing that I've definitely myself had to do lots of times, and it doesn't always work out. Sometimes there's an extra slash that you don't want, and you have to sort of do extra work to figure that all out. And so that's a great candidate really for having something that just does it for you. + +**Carl Johnson:** Yeah. And then that goes back to the idea of the Go development cycle. It is really nice that there's a development cycle, because the version that I put in had a bug... But fortunately, one of the beta testers figured out like, "Oh, you're using clean path here, but clean path strips the final slash... In fact, it shouldn't strip the final slash when it exists", and it was good that it was corrected. + +\[19:53\] But yeah, it's definitely good to have one sort of canonical source, so that if you're a Go user, you don't have to figure out for yourself and run into the bug on your own, but you can just use the version that's in the standard library, that has had other people look at it and confirm that it does do what it is supposed to do. + +**Mat Ryer:** Yeah. That's very nice. Excellent. Another thing that's changing is the memory model. Maybe you could tell us a bit about that. What is the memory model? + +**Carl Johnson:** So the memory model, essentially in a computer programming language, you need to make some guarantees about what happens, in what order. And if you're just approaching it naively, you're like, "Well, it happens in the order that I write down in the source code, and I call X is equal to one, and then I say X is equal to two, and so first X is equal to one, and then X is equal to two." But the problem is that nowadays there is a huge, huge, unbelievable level of optimization going on behind the scenes. And so if a compiler can tell that between you writing X equals one and then X equals two nobody else reads the value of X, it can just say, "Well, forget it. That it wasn't worth it. Just go straight to two. Skip the one", and then it'll be faster, because I didn't put the one in there. And so that works well, except for when you have multiple threads. So if you have the same memory being used by different parts of the CPU at the same time, you can run into situations where you as a code writer think that you've written this in a linear order, and so then the other part of the code should be able to see what's going on. But it turns out that that's not what's happening after all the optimizations are applied, and it's not actually an atomic memory access. + +So basically, since the early 2000's with Java and then C++, different programming languages have been trying to write down, "Okay, so what exactly are the rules? How do the rules work? How do I know if I'm following the rules or if I've violated the rules and done something that I wasn't supposed to do?" And so that's the idea of the memory model. + +And so Russ Cox had a series of blog posts in 2021 where he talked about how different languages have different memory models, and what he thought about them. Go has had a memory model basically since the beginning, but it hadn't been revised in a long time, and there were certain things that it didn't specify. So for Go 1.19, they've come out, they've revised it. They've made sure that it covers more cases. But it basically isn't really changing for the average programmer. + +So it has this great section of advice at the beginning of the document. This has been there, like I said, more or less since the beginning of Go. It says, "Programs that modify data being simultaneously accessed by multiple goroutines must serialize such access." So they have to find some way of making it happen in the right order. "To serialize access, protect the data with channel operations or other synchronization primitives, such as those in the Sync and Sync Atomic packages. If you must read the rest of this document to understand the behavior of your program, you are being too clever. Don't be clever." + +**Johnny Boursiquot:** Amazing. + +**Carl Johnson:** So basically - yeah, the document tells you that you could read the rest of this and try to understand exactly what the rules are, but for 99.9% of programmers you probably shouldn't do that. You probably have something else you can do that's better than trying to figure this out. + +**Mat Ryer:** That's hilarious. I think that's great advice. By the way, sorry to interrupt... I promise a silver horse just flew by my window. I think it might have been a balloon of some kind, but honestly, it's the weirdest thing I've ever seen. Either that or I'm having a stroke, but a silver horse just flew outside the window. Okay... Well, and so along with the memory model then we're also getting new atomic types. + +**Carl Johnson:** Right. + +**Mat Ryer:** \[23:46\] Now the Sync Atomic package defines extra types like bool in 32, in 64, unsigned integers, and a pointer type. So this is going to help us-- I mean, to be honest, the bull one is... For example, I've used just an int before, but it's not as clear, because it could be any value, and so you'd lose a bit of that sort of safety there. But now we can be more specific about the types that are atomic. + +**Carl Johnson:** Yeah. This is like a-- it's one of those nice quality of life improvements. So we've had the ability to atomically load integers, again, for years going back. I don't even know, probably to the first version of Go. But if you wanted to have a bool, you would just sort of have to have a convention of saying, "Alright, when it's zero, it's false. And when it's one, it's true", or the other way around. But now there's a nice, convenient atomic.bool. And then there's atomic.pointer, which is, believe it or not, the first time the standard library has generics. + +**Mat Ryer:** Oh, is it? + +**Carl Johnson:** So generics - they came in in Go 1.18, as we were saying, so last March. But as part of putting generics in, at a certain point they said, "Look, this is a huge change to the language. We're doing so much work behind the scenes. We don't want to also change the standard library at the same time, because we don't yet know what's a good idea for an API and what's a bad idea for an API, because we haven't really used it. We haven't used it because it hasn't existed." And so what they did instead is they made a package that's just on the internet at golang.org/exp. You can go out there and you can get the experimental packages. And so there's an experimental slices package, an experimental maps package, and different things... But they haven't actually changed the standard library to use generics until now. And this is the first time - if you use atomic.pointer, it's a generic. So it can tell what your pointer type is, and ensure that you're always using the same type for your pointer, and you don't start as an int pointer and then change it into a float pointer or something by mistake. + +**Mat Ryer:** Yeah. So again, it's about that type safety. And I kind of love how careful the Go team are about changes like this. So I really appreciate that they're taking their time. Because once something's in the standard library, because of the backwards compatibility promise, it's there for good. So I'm really pleased that they do that. And that's so interesting, to think that you can create that atomic type now pointing to something, and make that strongly typed. + +**Break:** \[26:21\] + +**Mat Ryer:** So staying with memory, what's this soft memory limit that I've read so much about, Carl? + +**Carl Johnson:** So Go, as I think most listeners will know, is a garbage-collected language. That means that unlike C or C++, or to a certain extent Rust, you are not directly managing your memory. Instead, you have a variable, and the Go compiler and the Go runtime will look at them and try to figure out when they're used and when they're freed, and based on that, sort of decide, "Okay, at this point the memory is no longer being used by anything in the program, so let's return it to the operating system. And over here, we're still using it, so let's make sure that it doesn't get overwritten." + +So this has been the design of a number of programming languages going back for years. The first programming language that really got popular with garbage collection was Java. There had been ones before that, but Java was just unbelievably popular. And so for Java, with their garbage collector, it was being used in these different situations where people had different requirements. Do you need the garbage collector to run predictably, or do you need it to run quickly, or do you need it to run thoroughly? And it's a very difficult trade-off. And so Java had a lot of different ways of tuning your garbage collector, of changing the algorithm it's using, of changing when it's doing what... And so the Go team, when they were creating the language, they looked at these different ways of doing the Java garbage collector, and they said it was a little too complicated. Just making sure your Java garbage collector was doing the right thing was sort of a job in and of itself. Like, you could hire somebody who's just an expert in making sure you've set it to the right settings. + +\[30:00\] And so they said, "Look, we're going to give ourselves a challenge. We're going to see if we can just have a single value that you can tweak to change the parameters of the Go garbage collector." And so that's how it's been for many years, until Go 1.19. Now we have two ways to tweak it. + +So the first way of tweaking the go garbage collector, which has been there since before, is you can say what percentage of new memory there is, versus old memory. And when the percentage gets too high, it'll trigger a garbage collection event. + +So that's been the traditional way that we've told the garbage collector what to do in Go, but now there's a new way, which is that you can set a memory limit. And so if you say, "Go, I want you to try to keep the amount of memory that you're using underneath two gigabytes", or something like that, then whenever you get up to that limit, Go is going to start telling the garbage collector to run more and more often just to try to stay under the limit. Now, you have to be careful, because maybe there's just no way of getting under the limit. You've really just allocated so much memory that it doesn't matter what you do, you're never going to get under the limit. And so then there's still issues to be tweaked with that, but the way that it's designed now, as you approach your memory limit, it will run the garbage collector more and more, until it's eventually running about half of the time. And once it starts running about half of the time, it'll say, "Look, forget it. It doesn't matter if I run more, I'm not going to be able to get under the limit." + +So this opens up Go for use in a lot of applications where you couldn't use it before. So like on a mobile phone, a lot of the mobile phones - they have a lot of memory, they have gigabytes and gigabytes of memory, but they also have a very hard limit on how much you can use it one time. And if you go over the limit, then your app just crashes and it sends you back to the main screen. So I think this is very interesting, because it could be that this will lead to people using Go more in mobile development and different situations where the memory limit is something important. And before, there wasn't a good way of saying, "Hey, make sure my app doesn't crash. Make sure I stay underneath however many gigabytes." + +**Mat Ryer:** Yeah. And Twitch did something... They sort of hacked it a bit, didn't they? + +**Carl Johnson:** Yeah. So I said before that the old knob that we had for controlling garbage collection was based on the percentage of new memory, versus old memory. And so Twitch had this really funny idea where they said, "What if we just allocate a bunch of memory that we don't need, and let's keep it around?" And they called it memory ballast. + +**Johnny Boursiquot:** \[laughs\] Ballast. + +**Carl Johnson:** It's like, we're going to allocate a gigabyte that we don't need, just for good luck. But basically, what that did was it sort of tricked the garbage collector into thinking that, "Oh, well, you've allocated 300 megabytes of new memory, but you still have that giant two gigabytes slab that you're not really using, so I'm not going to trigger for another little bit." + +So it was just a kind of funny way of tricking the garbage collector into triggering at different times. But I think with the soft memory limit now, that should hopefully be obsolete, because it was definitely-- it's one of those hacks where it's like, it's genius and it's idiocy at the same time. You're like, "I can't believe somebody thought of something this clever", and also, "I can't believe somebody thought of something this stupid." + +**Johnny Boursiquot:** \[laughs\] Yeah, simultaneously. + +**Carl Johnson:** It's like, you're so proud of them for figuring it out, but also you're like, "You're just allocating memory for no reason? It doesn't do anything? It's just there for good luck?" + +**Mat Ryer:** Okay, Carl, this one I don't quite understand, and I haven't run into this, and I think I've probably got code out there that has this same bug in it, but there's a new abs() method on a time duration. And this is different to you just doing it yourself, isn't it? + +**Carl Johnson:** Yeah. So this is a bug that bit me in production. And so then after I had been bitten, I was sort of thinking like, "Oh, this is so annoying. How did this happen? How can I prevent it from happening again?" And then I realized I could open an issue on the Go Issue Tracker and see if I could fix this issue for other people. + +\[34:05\] So essentially, the problem is that a time.duration underneath the hood is just an Int64, right? It's just a regular number, and it records the number of nanoseconds since some epic-- I guess it's not even from an epic, it's a duration. So it's just an absolute number of nanoseconds, right? Well, I guess the problem is it's not an absolute number of nanoseconds. It's just a number of nanoseconds, and it could be positive or it could be negative. + +And then it turns out that the way that integers are stored in computers, there's always one more negative number than there is positive numbers. It's just like a weird computer fact that you learn in college, where they're like, "Oh, yeah, the way we store numbers - we always make sure that there's one more negative number than positive numbers." And so it mostly doesn't matter, except for if you have two time.times and you want to know, are these two times within, let's say, a minute of each other? So that was the thing that I needed for production. I needed to know if I was posting this within a minute of each other, let's not trigger the alert twice. But if it's more than a minute, then you can trigger the alert again. + +So I took the two times and I subtracted the one time from the other. Once I subtracted the two times, if the duration I got back was negative, I converted it to positive, right? And then if it's less than a minute, then it was within the time. And if it was more than a minute, it was outside of the time. + +Well, the problem was, like I said, durations are just stored as integers, and there's one extra negative second that can't be represented as a positive second. So if you just do the naive thing of multiplying a duration by minus one, if it's negative, it's not going to work. It'll stay negative. And so I was like, "Why is this not working? I'm subtracting these two times. I know that this time is bigger than this time, but it's still saying that it's within range when it's not within range... What's going on?" + +And so time.duration.abs, or .abs - it's a new method, it's on duration. All it does is if you have a time and it's positive, it doesn't do anything. If it's negative, it converts it to the positive time, unless it's that one extra second, that one extra nanosecond that can't be converted to positive, and then it converts it to the closest positive value that it can. So it's just one of those weird little things, like, I got bit by a bug. I was like, "This is so annoying. Why did this happen? How do I make sure that somebody else doesn't have the same bug that I had?" + +**Mat Ryer:** Yeah, that's really interesting. So just to summarize that - because it's just an Int64 under the hoods, under the covers, there were more negative numbers, and so you can't just rely on doing the absolute that you would expect. And so this method does it properly. I like that. + +I think any time there's weirdness at the edges, it's very difficult to find them, because probably in all your test code you're putting in numbers that really make sense, and you test it all - well, you think you're testing it all kind of perfectly - and so you can miss those edges. And I think fuzzing probably may help even with things like this; it probably wouldn't in this case, because you wouldn't think to use the fuzzer probably in this bit. So yeah, I find these kinds of methods really helpful. + +**Carl Johnson:** Yeah. This was before Go 1.18, so the fuzzer didn't exist. Or I guess here were ones that were not in the Go tool that I could have been using. But yeah, it was a very annoying bug, because if you had the same two values and you just flipped them, and instead of subtracting A from B, you subtracted B from A, then suddenly the bug went away. And it's like, "Am I going crazy? Am I taking crazy pills? It says that they're within time if it's A minus B, but not if it's B minus A?" + +**Mat Ryer:** Johnny, tell me... You wanted to talk about generics. + +**Johnny Boursiquot:** \[37:47\] Yeah. It's kind of hard not to, given that it's the new and shiny toy that we all have on our hands. Really, this is just a single use of generics in the 1.19 release planned, just a single standard library package. I think the significance of that, or how little generics is being used - it shouldn't be sort of lost on people. If you remember the talk that Rob and Ian gave at GopherCon, when they introduced officially generics last year, the advice was, "Don't go crazy all at once. Let's give ourselves, as a community, time to understand use cases. Let's find the edge cases, the good uses, the bad uses and whatnot." Obviously, we have to use generics to get to discover those things, but the fact that the standard library is not rushing to implement these things is a testament to that philosophy they're taking with the standard library, right? + +So generics is a big change to a language, and the Go 1 compatibility promise is still being what it is today, and to this day, even after such massive changes to the language; it's really one of the things that I like most about Go, right? You don't have to really change the way you write Go for that. There has been over 35 changes to the standard library coming in 1.19, and I'm sure there'll be more and more as we go with future releases. But yeah, I think we should all take a lesson from that careful approach that the core team is taking to sprinkling, as it were, generics all over the place. So yeah, definitely something that definitely wasn't lost on me, and hopefully it is not lost on the community either. + +**Mat Ryer:** Yeah. But what about like-- I thought when generics would land, we'd get some common, obvious things solved for us, like a slices package that does slice operations, but in a strongly typed way, things like this. But we don't see them yet, do we? + +**Johnny Boursiquot:** Not yet. I think, as Carl mentioned earlier, the experimental package, \[unintelligible 00:40:07.23\] has some things in there related to generics, the constraints package, for example, which definitely has featured in tutorials and talks and whatnot. \[unintelligible 00:40:18.07\] generics is one example where there is-- you're starting to see some things that should be common across all limitations for the use of generics, and having... Basically, you're starting to gather these things into certain places that make sense. So if you're going to be writing your own constraints, basically it makes sense that the standard library will provide a standard set for integer, and float, and order, and having built-ins that are comparable, and things of that nature. + +These are things that I think are going to continue to surface. I think you're going to get a swelling of these common patterns, common set of things that both the community and the core team are going to discover and bubble up. They're going to find their way through the whole experimental process, right? Basically, making their way from golang.org/x into the standard library of proper, and maybe you might even see some new identifiers, some built-ins getting added to the language as well... + +It's all experimental at this point, so again, basically, taking a very deliberate approach. And people are going to write their own -- until the standard library gets some of these things, until the language itself gets some of these things, people are going to write their own implementations for the use cases you've given, like being able to deal with maps and slices in a certain way, right? + +\[41:41\] Data structures are going to basically see a huge benefit from the use of generics. Like, how many ways do you want to implement -- I mean, you want to be able to implement a binary tree that can work with different things. You want to be able to work with a linked list, or some set or some--these sort of common data structures, they could definitely use a touch of generics here and there. People are going to create their own libraries for these things, right? + +So as we've seen in the past, we can expect the community to come up with-- there will be some popular packages that use generics, that provide some of these basic data structures and things like that. And over time, you may see the core team take a page from these things and implement in the standard library proper versions of those things. So just basically, from where I'm sitting, just give it time. Allow people to bump against the guardrails a little bit. Let's play in the sandbox, and then we're going to find the language itself is going to take a cue from the community. + +**Mat Ryer:** Yeah, sounds good. Sounds great. Okay. There's a new era coming too, Carl, isn't there? The HTTP max bytes era. What is that? + +**Carl Johnson:** Yeah, this is a thing. So in the HTTP package there is the max bytes reader. And what this lets you do is it's a little bit of-- it's not quite a middleware. There is a middleware. There's a max bytes middleware as well that uses max bytes reader. But max byte reader, basically what it lets you do is you put it around your calls, and then if somebody is trying to upload a file to you, you can set a limit and say, "Alright, you can upload files to my Go server, but it can't be larger than 5 megs, or 100 megs, or 5 gigabytes", or whatever you want the limit to be." But traditionally, when somebody did go over the limit, it didn't return a named error type. So this meant that if in your code you wanted to send back a message to the user and say, "Hey look, that was 100 megs, but the limit is 75 megs. So please send up a smaller file next time", there wasn't a good way to do it. You just had to sort of look at the strings of the error message, and if it was exactly the message that max bytes reader sent, then you could guess it was probably a max byte reader, but it was sort of ugly. + +So there had been an issue open for years to fix this. And finally, there was sort of consensus that like, "Oh yeah, let's go ahead and do it. This is a good idea. Nobody thinks this is a bad idea." And so again, it was one of those things where I was just looking at the Go issues page, I saw an issue, it seemed like it would be something I could just knock out when I was bored and needed a change of pace before going back to work... And yeah, it made it into Go 1.19. + +**Mat Ryer:** Yeah. + +**Carl Johnson:** I think actually with this one -- I had it done before Go 1.18 came out, but I was saying before that there's that Go release cycle and there's a freeze period. So I finished this one day after the freeze had started. And so I sent it in and they said, "Sorry, mate. The freeze has started. We'll see you next cycle." So I had to wait an additional six months to get it in, because of being a day late. + +**Mat Ryer:** Oh, wow. So they really mean it. + +**Carl Johnson:** Yeah, they don't get around with the freeze. + +**Break:** \[45:02\] + +**Mat Ryer:** Yeah. So you mentioned that hack of checking the actual string, and I've run into this myself before, and actually run into a bug where the string I was checking from the error - it worked on my machine, but then, for some reason, it ran on a place where the language was different, and the error message itself was coming from the operating system... And so it changed, the error changed, and then suddenly, my check didn't work. So it's very brittle to kind of rely on that. It's much nicer to have proper types, whether-- is this a type? Is this like a sentinel error, or is this a...? + +**Carl Johnson:** It's a type that you can check with \[48:17\] But yeah, one of the things when I was implementing this is - so there's a new error type, and I have to give it error string method. And so what string should this error return? Well, if it was like a normal error, it should go ahead and return something like "maxbyteerror:toolarge, here's the size limit", but I couldn't do that because if I did, I would break everybody who was checking for that string that the old one returned. So it's not, I don't think, strictly required by the Go1 compatibility requirement... But just to make sure that those people have time to transition their code to the new error and to check for the error with the type, instead of just checking for the string, I went ahead and I used the exact same string that it used to be, and I put a little comment on saying, "Please don't change this. People are relying on this being the same." \[laughter\] + +**Mat Ryer:** That could be a vet check that we have, that looks for that string and see if you're doing that check, and say, "Oh, did you know? Thanks to Carl - he's fixed this - you can do it in this better way." + +**Carl Johnson:** Yeah. I mean, I feel like I was just like the last person in a chain of-- like, if you go on GitHub and you look at the issue for this, you can see so many people there with the exact same idea, saying "Hey, this should really be a type. It's kind of a pain that I have to like check for the string. It's brittle. It's going to break." And so actually, what turned me onto the issue was I read the book Let's Go Further by Alex Edwards, which is a great book for just learning how to make a Go HTTP server. And I was just sort of reading through that, and looking at the different things, and at one point he says, "Oh yeah, here, you have to do the string check, because there's no type for it. If you think this should change, go to this issue", and he had the URL right there in the book. And so I was reading the eBook version, so I clicked through, and went to the bottom, and I was like, "Yeah, let's change this." And so I want to give thanks to Alex Edwards for turning me on to fix this idea. + +**Mat Ryer:** \[50:14\] Nice. + +**Johnny Boursiquot:** Now you've made his book out of date. + +**Carl Johnson:** Oh, yeah, that's true. He has to update a new edition, make it more work for him. + +**Mat Ryer:** That's not part of the backwards compatibility promise, to be fair, is it? \[laughter\] + +**Carl Johnson:** Well, the old code will still work. It'll still work, what's in the book. It's just now there's an easier way to do it. + +**Mat Ryer:** Yeah. Can you do a pull request to his book? Actually, how hard was it to-- like, was there any discussion around the design of that? Was this something where you had to kind of advocate for one way to do it? Were there competing ideas? Or was it just sort of like the community had come to the conclusion that this is the right way to solve this? + +**Carl Johnson:** There was a little bit of discussion. I mean, it's such a simple thing that I feel silly for talking about it on a podcast... But it's true that there was a discussion. So one of the issues is in Go there's two different ways that you can make a new type. You could just say "type max bytes error int 64". So one thing you could do is you could just say `type maxBytes error struct{}` and that would mean "This has no data whatsoever." It's just a pure type and you could check for it, and that's all. So that's one way that you could do it. But if you did it that way, there wouldn't be a way of knowing what the limit was that you went over. And so I said, "Well, we should really have some way that people can find out what the limit was in case they have different handlers. My upload handler lets you upload up to 100 megabytes, but my JSON handler doesn't want you to upload more than two megabytes, and so I want to give the right error message to the right person." So we had to advocate for that. + +So then there's two ways of doing that. You can say `type maxBytes error int64` or you could say `type maxBytes error struct int64` And if you do it in the struct version, then that lets you add more fields later. + +**Mat Ryer:** Right. + +**Carl Johnson:** But if you do it in just the plain version, then you're really committing and you're saying, "I guarantee that I'm--", at least going back to that Go 1 compatibility promise, "I don't think I'm ever going to have to add a second field." And so I don't think they're ever going to add a second field. I can't really imagine what the second field would be, but there was like consultation and it was decided, "Let's go ahead and do it the forward-compatible way, just in case there ever is a second field that needs to be added." + +**Mat Ryer:** Yeah. I think that's such a good lesson, I think. And that's something that I advocate for that a lot, of like, give yourself more options in the future. Yes, it would be very satisfying if that type was just a number, just an Int64, but give yourself more options in the future. + +Another example is in data, if there's a bullying field that's representing some kind of status, like active or not active, I'll probably go for a string that says active or inactive or something, because what if I have other statuses in the future? I don't then have to go and change those types. So I can't like that way of thinking about the future, designing for the future, and give yourself more options later. Of course, these types have to have the error method on them, which is what satisfies the error interface... But you could have them on both of those, of course. + +**Carl Johnson:** Yeah, it's just a little bit of future-proofing. Yeah, it's interesting to see when people come together and they're debating these different APIs... Things just improve. It's definitely improved me as a programmer to sort of be in the issues, and like, you know, I'll have my first suggestion, and somebody else will come up with a better idea, and then that'll be the thing that we end up implementing. + +**Mat Ryer:** Cool. Yeah. Okay. Well guess what, Carl? + +**Carl Johnson:** Is it Unpopular Opinions? + +**Mat Ryer:** It is. It's Unpopular Opinions. + +**Jingle:** \[53:57\] to \[54:14\] + +**Mat Ryer:** I could have done with the max bytes era earlier, because I had too much dinner. That's my unpopular joke... Carl, do you have an unpopular opinion for us today? + +**Carl Johnson:** Yeah. So the last time I was on the show, as you know, if your opinion is popular, you are forced to come back on, until you get an unpopular opinion. So last time I was here, I had the unfortunately popular opinion that the government should pay people to do open source software. So this time I think I'm really going to be unpopular... And my opinion is that Twitter is literally a hive mind. + +**Mat Ryer:** \[laughs\] What do you mean? + +**Carl Johnson:** So I don't know, maybe Twitter will like this... I think that Twitter-- \[laughs\] maybe Twitter people will like this. So I've been reading this science fiction book by the author Adam Roberts, called The This. It takes place in the near future. There's a new social media network. It's similar to Twitter, except for you have like a little implant that you put into the roof of your mouth, that connects to your brain and it lets you post wirelessly, without using your hands or your eyes or anything. So it turns out that if you just make the speed of posting on Twitter a little bit faster, it creates a global hive mind that takes over the universe. + +So reading this science fiction novel really just made me sit back and say, "Wait a minute, what is it about Twitter that I like? What is it that when I'm on Twitter, what am I doing? Why do I enjoy this?" It's not like something where you go on Twitter and you're like, "Oh, I was laughing out loud all night, because I was reading these great jokes", or you're not like, "I was crying, because I was reading about the sad things happening in the world", or it's like, "Oh, I was so fired up and angry that I went out there and I did a protest", or something. It's like, you're feeling all these emotions, but it's like really -- you're just blipping in and out of the different feelings... But it is addictive. So what is it that's addictive about it? And so what I've decided is that the reason it's addictive is because you're participating in a hive mind, and that's just like... You know, it's nice to be a brain cell in a larger brain, but... I don't know. I've decided to try to be more individual, at least for now. + +**Mat Ryer:** Oh, good for you. I was just going to agree that it is like a hive mind, but then I realized I'm really not helping. + +**Carl Johnson:** No, no, I've got to come back again if this is popular. Will I never escape...? + +**Mat Ryer:** Yeah. Interesting. I mean with the algorithms, of course, it's probably many hive minds, I suppose. And this is where I think it really does get dangerous, where -- we know this phenomenon where these echo chambers get created; you really end up following people that just agree and support your perspective, and then you stop listening to other views, or you see the other views through this distorted lens, where it seems evil or bad, or whatever. And in the real world, if you meet somebody that has like a view that you would think is an absurd view, if you meet them in the real life, that same effect doesn't happen. + +**Carl Johnson:** Yeah. + +**Mat Ryer:** I do think there is something that we have to be really careful with that. + +**Carl Johnson:** Yeah. Even for programming languages, there's something about being on the internet that pushes you to extreme. And it's not just that, "I enjoy programming in Go", but "Go is the best, and people who program in other languages are losers, and how dare they, and I'm going to throw dog poop at their house until they change." I don't know, it gets out of control really quickly on the internet, and it's hard to explain why it is. And so I think the theory that it's because you're in a hive mind is as good as any other theory. + +**Mat Ryer:** \[57:56\] Oh, yeah, I'm into it. Interesting stuff. Well, let us know on Twitter if you agree or not with Carl. Actually, to be fair, we do poll these, and sometimes they're split. + +**Johnny Boursiquot:** Right. + +**Mat Ryer:** But what I have noticed is when -- you've made a case for that. What happens is the way that it goes out on Twitter is they'll play a clip of you making that case, and then they ask people to vote for it. And very often, people will agree, because you make the case very articulately. And basically, when you hear someone say something, that's what you then believe now. I believe that now. \[laughter\] + +So we should check ourselves. We're all vulnerable to it, and if we think we're not, that's even more dangerous. We are all vulnerable to this effect. So thanks for the warning, Carl... + +**Carl Johnson:** Yeah. We all get sucked into that hive mind. + +**Mat Ryer:** Johnny, I hope yours isn't quite so dystopian... + +**Johnny Boursiquot:** I don't know. It could be, depending. I mean, if I make the case that it is dystopian, maybe you'll start thinking it is dystopian. + +**Mat Ryer:** I already do. + +**Johnny Boursiquot:** Yeah? + +**Mat Ryer:** Yeah. I'm in. I haven't even heard it... \[laughter\] + +**Johnny Boursiquot:** So over the years I've found this one thing, this one thing to be true across all of the projects I've ever worked on. Do you want to guess what it is? + +**Mat Ryer:** Oh... + +**Johnny Boursiquot:** Don't cheat and read the show notes... + +**Mat Ryer:** Okay, yeah. Okay. You have to use the keyboard to-- no, you don't even have to use a keyboard to put the code in. + +**Johnny Boursiquot:** No. + +**Mat Ryer:** Well, you cover the key you want, which is counterintuitive, on a keyboard. You're hiding the one that you actually are going to end up with on the telly. Is it that? + +**Johnny Boursiquot:** \[laughs\] Close, close, but no. I was thinking more of a-- I've always ended up regretting using boolean to keep track of data, when I could use a timestamp. + +**Mat Ryer:** A timestamp? + +**Johnny Boursiquot:** Yeah, exactly. + +**Mat Ryer:** Well, not like a string? + +**Johnny Boursiquot:** Yes. Rather than storing -- like, for example, is active, or active, or whatever it is... Storing that, and storing a true/false, or one and zero, or whatever it is, in the database. + +**Mat Ryer:** You just store the 5th of January, 1971. + +**Johnny Boursiquot:** Yes, I store the timestamp. I store activated at, or active at, or whatever, something, because that gives me more information; because I know if there's a real date in there, right, I don't have to keep track of two pieces of information. I know "Oh, it is active", and now let me go find out later on, where do I store it? When was it last activated or even deactivated, right? I can just use one piece of data that communicates both pieces of information to me. + +**Mat Ryer:** Oh, I see. + +**Johnny Boursiquot:** So yeah. + +**Mat Ryer:** That's really interesting. + +**Johnny Boursiquot:** Don't use a boolean when a timestamp will do. + +**Carl Johnson:** You can't add this after the fact. That's the interesting part about this, is that you can't like-- if you change your mind and you're like, "I've never used this date stamp. I'm always just treating it as a boolean. I'm just going to convert it all to boolean to save some--" I don't even think you would save any bytes in the database actually, but just to make the code simpler. You can convert from date stamp to boolean, but you can't do it the other way around, right? So if you're like, "Oh, I really wish that we had used a date stamp here, so that we knew what day people deactivated their account or what day they did this or did that." Too late. It can't be fixed after the fact. + +**Johnny Boursiquot:** Right. That's why at this point, usually, whenever I see a boolean used in a PR, or in the early phases of a system design, basically any sort of ERD or data model that I'm seeing where I'm seeing booleans, I'm like, "Could this be a timestamp instead"? And more often than not, it's true. Don't use a boolean when a timestamp will do. + +**Mat Ryer:** That made a lot more sense after you explained it. I genuinely thought you meant you just have your birthday means true, and any other day is false, baby... \[laughter\] + +**Johnny Boursiquot:** If it ain't my birthday, it ain't true. \[laughter\] + +**Mat Ryer:** Yeah. But no, that does make sense. But hang on, don't you then have to deal with like -- what's the empty state? Is that null, or is that an empty string? + +**Johnny Boursiquot:** \[01:02:04.19\] Well, if you're using-- so if you're rolling your own, to use the cases of storing data in a database and whatnot - if you're rolling your own sort of ORM or whatever it is, or you don't want to use an ORM, maybe if you wanted to, you could store the zero value of your data, or just store nil, and then the standard library does give you the ability to parse a time value and determine whether that time value is zero or not. It's literally a zero function in the standard library for a time. You could use that. But in most cases, ORMs, I know for a fact Gorm, which actually has become-- when I have to use an ORM, it has become one of my go-to's... It will omit sort of storing the zero value date for you. So any ORM that you want to use typically should give you, as we call it, quality of life, sort of bits and pieces. + +So yeah, I think it's totally doable, totally manageable. And I think you get a whole lot more for that, for the extra bytes that you store. I'm not even sure, like Carl is saying, I'm not sure you save that much space over the different data types, depending on database implementation... But yeah, you get so much more back when you use that timestamp than you do with a bit. + +**Carl Johnson:** Yeah. Now I'm thinking though, going back to the time.duration.abs, that if you were just subtracting the time, like if you were like, "I'm looking at the 'Is deleted at' column and I want to know if it was deleted in the last month, and then I'm going to send them an email saying, 'Please come back', or something like that", if you just naively do that subtraction and the time is zero, you'll get the overflow and you'll need to use time.duration.abs. So please people, if you're implementing Johnny's thing, use time.duration.abs and send me a royalty of -- you know, something fair, not a lot; just like $100. Do you remember in Back to The Future II there's a part where Marty McFly is walking through the square, and it's the future, it's the year 2015, and the woman says like, "Hey, can you spare some money for the clock tower?" And he's like, "No, sorry, I can't." And she's like, "Come on, man. It's just 100 bucks." And I think that with the inflation now-- it was really set in 2022. They just sort of made a little error in the timing of the thing. + +**Johnny Boursiquot:** Oh, man. You can't even buy a lollipop for 100 bucks anymore, can you? + +**Carl Johnson:** No. \[laughter\] + +**Mat Ryer:** Yeah. I also like how in Back to The Future, when he goes back and someone dies in the past, then he looks at the Polaroid he's taken, and it's kind of like half fading away. So at some point, there was a semi-transparent tombstone there... But to be fair, you would take a photo of that if that happened. It takes time to cascade through the effects, but-- + +**Carl Johnson:** I don't know. + +**Mat Ryer:** ...I love that. They're probably my favorite film. + +**Johnny Boursiquot:** Don't look at Back to The Future to reason about time. That's it. + +**Carl Johnson:** Yeah. + +**Mat Ryer:** Yeah. + +**Johnny Boursiquot:** \[unintelligible 01:05:08\]. + +**Mat Ryer:** Primer you want for that. I don't know if you've seen Primer? There's also a great Spanish one called Time Crimes, which is another one of my favorites. I really recommend that. + +**Johnny Boursiquot:** Hm, okay... + +**Mat Ryer:** And actually, the Adam Project is a more recent one, which is kind of more family-friendly movie about time travel. To be honest, I like any film with time travel in it, and it can be a really bad film, and I'll love it. For some reason, if it's got time travel in it, I'm in. + +**Johnny Boursiquot:** So have you seen Tenet? You should see Tenet. + +**Mat Ryer:** Yes. + +**Johnny Boursiquot:** Because that's a mind trip right there. + +**Mat Ryer:** Yeah, that's a good one. + +**Johnny Boursiquot:** You might have to watch it a few times. Yeah. + +**Carl Johnson:** Can I give my complaint? This is like the nerdiest possible complaint that somebody can have, which is that, okay, so Star Trek-- + +**Mat Ryer:** Are you sure it's more nerdy than storing a timestamp instead of a boolean? + +**Carl Johnson:** Oh, this is so bad... So Star Trek - they've done a million time travel episodes, and in the Star Trek time travel episodes, the rules are always that at the end you have to get things back to where it was, or close enough. There can be like little changes, but nothing big. But then they had the Star Trek "reboot" movie, which was not a reboot movie, it was a time travel movie. And in that movie and only that movie, they traveled back in time and they made a second timeline. And it's fine. Like, you can have time travel rules be that when you go back in time, it creates a new timeline, and then there's two things, and so it's okay to kill your grandfather, because that's just a different timeline. That is fine if you want that to be the rules, but that's not the rules of Star Trek. + +\[01:06:34.13\] If they wanted to reboot and say, "Okay, it's a new thing. It looks like Star Trek, but it's a parallel universe, where there's different things", that would've been fine, too. I think everybody would've been happy. There's like a million Batmans. Nobody thinks that the Batman in Batman 1989 has to have some relationship to the Batman in the animated series. They can just be two different universes. It's fine. But then for Star Trek, for some reason, they're like, "Let's make it a time travel movie and let's make it not use the rules of Star Trek time travel." + +**Mat Ryer:** Yeah. + +**Carl Johnson:** It just makes me crazy. + +**Mat Ryer:** Yeah. + +**Carl Johnson:** It's my nerdiest complaint about movies, I guess. But then all the other movies in Star Trek, it's not like they don't do time travel in Star Trek. They do it all the time and they have very consistent rules except for that one movie. + +**Mat Ryer:** Yeah. That is quite nerdy, I'll give you that... But I'm with you, actually. If there's contradictions in films, then I'm just out. And that does apply to time travel films. But there's things like -- any time there's like a horror film... And my partner loves horror films. She's always like-- she loves that whole genre. But if there's like someone's grabbed by a ghost and they're being pulled through the house or something... So at that point, whatever else is going on, at that point you can physically interact with this thing, whatever it is. So you should be able to attack it. You can't have one without the other. And the other thing is, if you are invisible, which happens in films, you'd also be blind. And this is never the case. Light has to hit your retina and be absorbed in order for you to see. And so if that's not happening, if the light's traveling through, then you are blind, basically. So you could be invisible, but you'd also be blind. + +**Carl Johnson:** Maybe you just have-- you're invisible except for your pupils, and people just don't notice that the pupils are floating. They just think they're like two little nuts, or something. + +**Mat Ryer:** Well, if that's the case, if there's just two retinas floating around-- + +**Carl Johnson:** Yeah. + +**Mat Ryer:** ...then fine, then at least it's consistent. + +**Carl Johnson:** Yeah. That's like a good movie. That's what we all want to see, is those two little retinas floating around... It's called Retina. + +**Mat Ryer:** Retina. \[laughs\] Christopher Nolan could make that. He'd do a really good job. The other thing about Star Trek I like - and my friend, Aaron Adel, made this point as well... It's when they go onto another ship and they see all this alien technology they've never seen before, and they're like, "Hmm, this looks like navigation. This one looks like the energy." And it's like, I can't even use Android. + +**Carl Johnson:** \[laughs\] + +**Mat Ryer:** Like, I've got an iPhone. I don't even know how-- if I pick up someone's phone, I'm like, "Yeah, you do it." It's like, "What do you mean?" It's like, "Mr. Data, can you--?" + +**Carl Johnson:** This is like one of the problems of the 21st century... It used to be when you went to somebody's house and they had a television, you knew how to turn on their television and turn it to whatever channel you wanted to watch. Now, when somebody comes over to your house, like somebody's got to take care of your house because you're going out of town, you have to leave like a five-page memo. It's like, "This is the remote that turns on the TV. This is the one for the Roku. This is the one for the stereo. This is the other one for the stereo; if the first one isn't working, you have to push this one to switch the channel." It's a mess. Somebody needs to make remotes that-- + +**Johnny Boursiquot:** I just want voice control. + +**Carl Johnson:** Oh, yeah. + +**Johnny Boursiquot:** \[01:09:56.12\] Can I have voice control for all the things? Can I just get-- do away with remotes; just voice control. + +**Mat Ryer:** Yeah, but then what's the trigger word - I don't want to say the trigger words in case people listen to this, on those devices. And then we can hack... + +**Johnny Boursiquot:** Peanut butter. + +**Mat Ryer:** Okay. So you get to choose your own. + +**Johnny Boursiquot:** Yeah. Why not? Yeah. + +**Mat Ryer:** Why would choose that? What if you want peanut butter though, Johnny? \[laughs\] You have to whisper-- + +**Carl Johnson:** Very confusing. + +**Mat Ryer:** ..."Can I have peanut butter? \[laughs\] + +**Carl Johnson:** You should make it something you don't ever want. So like liquorice; it's like, "Well, I would never ask for liquorice, so..." + +**Mat Ryer:** There you go. I hate liquorice. + +**Johnny Boursiquot:** That's an unpopular opinion, if I've ever heard one. + +**Mat Ryer:** "I hate liquorice." Bling! "Oh, no, I was just telling someone." "I'm sorry, I don't know what you mean." "No, I wasn't talking to you. "Oh, I'm not--" You know... + +**Johnny Boursiquot:** "Would you like some peanut butter?" \[laughter\] + +**Mat Ryer:** "Oh..." Yeah. Okay. Well, that's, I'm afraid, all the time we've got today. But this whole new-- + +**Carl Johnson:** Is it? + +**Mat Ryer:** Well, yes, because we respect the timeline on Go Time... \[laughter\] + +**Carl Johnson:** We're going to travel back in time and fix all of the connectivity errors. \[laughter\] + +**Mat Ryer:** Yeah. To be fair though, if everything works and all the files are collected in the final version of this that goes out, this will be seamless. And the editors - they're so good. So it could well happen. That's all the time we've got, I'm afraid, today. Thank you so much, Johnny Boursiquot. Always a pleasure. + +**Johnny Boursiquot:** Mm-hmm. + +**Mat Ryer:** Okay, I thought you were going to say-- I was leaving space for you to say something nice about me, but no, that's fine. + +**Johnny Boursiquot:** \[laughs\] + +**Mat Ryer:** And thank you so much, Carl Johnson, for joining us. Well, I'm sure you'll come back, I hope, another time. + +**Carl Johnson:** Ah. More unpopular opinions, I can't stop doing that. Johnny, good to be on an episode with you. We're both in Baltimore, and I'd love to start coming back to the Baltimore Golang meetings as soon as my children let me leave the house. So... Someday. + +**Johnny Boursiquot:** \[laughs\] Sounds good. + +**Mat Ryer:** That's lovely, yeah. Do you want to say something nice about me, or we're just going to go two for two? + +**Johnny Boursiquot:** Mat, you've done an exceedingly adequate job at hosting... + +**Mat Ryer:** I'll take it. \[laughs\] I'll take it. + +**Carl Johnson:** Thanks, Mat. Thanks for having me on. + +**Mat Ryer:** No, it's my pleasure. Thank you so much. And we'll see you next time, dear listener, on Go Time. Bye. diff --git a/Who owns our code_transcript.txt b/Who owns our code_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..b7928faffa605393a07924426763f440d622d39f --- /dev/null +++ b/Who owns our code_transcript.txt @@ -0,0 +1,243 @@ +**Angelica Hill:** Hello, and welcome to Go Time! Today we are going to be talking about who owns your code. A question that certainly has been on my mind. So we're going to be exploring who owns the code - the company? Is it the engineer? Is it the team? Is it all of the open source contributors if it's a project? How about when you're using AI, machine learning, GitHub Copilot? Is it still your code? I'm really excited, we have a really brilliant guest with us today. We have Luis Villa, who has a programmer turned attorney who has been involved in open source since college. He's worked at Mozilla, where he revised the Mozilla Public License at Wikimedia Foundation, where he also briefly led the community team... And as a lawyer, he's worked with Google, Amazon and many other small startups. So currently, he's the co-founder at Tidelift, which works to make open source better for everyone by paying the maintainers. We'll hear more about that. But before that, I'd like to introduce our co-hosts. We have Kris. Hi, Kris. I haven't seen you in a hot second... + +**Kris Brandow:** Hello. I am back, after a very, very long, but much-needed break. So I'm feeling rested, and I'm ready to get into the Meta of this "Who owns code?" It's gonna be fun. + +**Angelica Hill:** You're ready, I'm ready... It's a very interesting topic. And the beautiful Natalie - I've seen you, I think, far too often for your own liking in the recent weeks... My wonderful co-host... + +**Natalie Pistunovich:** It's like our weekly one-on-one, but it's not one-on-one. + +**Angelica Hill:** A weekly one-on-one with anyone who decides to tune into the live... \[laughs\] + +**Natalie Pistunovich:** Weekly anyone. I like that. + +**Angelica Hill:** Weekly anyone. Beautiful. Me and Natalie don't do one-on-ones, we do anyones. \[laughs\] + +**Natalie Pistunovich:** That sounds terrible... \[laughter\] + +**Angelica Hill:** Oh, God... Luis, I would love to hear a little bit more about you and your thoughts on code ownership. + +**Luis Villa:** Well, it's been a long time since I wrote anything approaching useful code... But I have been involved -- I had an interviewer ask me while I was interviewing for my first law school job out of law school, law firm job out of law school, and they said, "Well, you seem to really like tech. Why did you leave tech?" and I said, "Look, I'm not leaving tech. I am only interviewing with law firms that are very much tech-forward, tech-first kind of law firms." So the goal was never to leave tech. The goal was -- I was at a startup, open source... Back in the first year of Linux desktop, I was at a small startup, we got acquired; during that acquisition process I worked with the attorneys, and I was arrogant, I was young, I was "Oh, I can do a better job than these people." So I decided after a little bit of experimenting, and I took like a night school law class that I enjoyed, a couple of night school law classes that I enjoyed, and so I decided to go to law school. But the goal was always, all along, to continue to focus on tech, and specifically very much to focus on open law, because there seemed to me at the time to be a body of lawyers who were sophisticated about technology, but they came at it very much from a patent-first, control-first kind of mindset. And that was something that was already starting to break down at the time... So I was in law school 2006-2009, and it was beginning to be an understanding amongst legal academics that open was a thing. + +I had attended a conference of legal academics where Creative Commons was announced; that was 2001. So there were some legal academics who got it. And in fact, I pretty only applied to law schools that had at least one faculty member who had written something that indicated that they got it in the slightest amount. So that meant applications were easier, because there wasn't that many schools to apply to. And I think that has worked out well. It's been a good career, it's been a fun career, because I very much -- the point was not that "Oh, I can make my piles of money and work 3,000 hours a year, or whatever." It was, "I have friends who are open lawyers, who deserve better lawyering than the lawyering that they were getting at the time." + +\[06:19\] I think that's been that sense of, "Hey, I'm doing this to help open people get better lawyering" has served me well as a sort of motto and mission, and has led to a lot of fun outcomes... Because open people are doing a lot of fun projects, have been doing a lot of fun projects, and that hasn't changed, and I certainly don't think it's going to change anytime soon. + +But a lot of it does come back to this question of "Who owns the thing?" And I admit, I have enough lawyer brainworms at this point that my immediate thought goes to "Okay, well, the contract of the--" and then one of you during the top prep said, "Well, what about like team ownership?" and I was "Oh, right." We talk in law school about this analogy that ownership is -- for reasons I don't remember, or maybe never knew, we talk about ownership being a bundle of sticks. And the idea is that we sort of -- we talk about it as if it's like one big trunk. But it's actually a lot of different small things. And ownership in the code world is very fragmented, because there's this sense of well, okay, for almost all of you, if you're working for a company, at the end of the day your company owns the code that you are writing, at least on the company time, unless you're very careful to not do it on company time, not do it on company hardware, and to do it in areas that are unrelated to what the company is working on. If you're doing those things, then you keep it, but otherwise, as a general rule of thumb, the company owns it. + +So that's like the sort of lawyer brain answer, like "Well, yeah, okay, we're done here. It's been a nice podcast. Glad to talk to you all." But there's very much of, course, all these fractal little senses of "Well, okay, but what does it mean when the team -- team versus individual ownership?" Because in a lawyerly sense, the company is the one who can sell it; but who has responsibility for it within the company? Like, that's not a legal question. That's a team norms, team behaviors, kind of question. + +And there's also these questions of what exactly is it that you own? Because I spent several years of my career -- I am what we call a transactional lawyer. Basically, I do contracts. If the contract goes wrong, for some reason, that is somebody else's problem to argue about it in court. And so I've only ever been to court for work once, which was a little case called Oracle v. Google, and you might have heard of that one... And the question at some level was about who owns, or can anyone own the idea of an API? And that's probably not something you're thinking about too much in your day to day, right? And your corporate lawyers probably aren't most of the time, either. They're not thinking about who owns the API. They're just thinking about this file, or this binary that we're distributing, or these days, often, this SaaS that we're putting out over the internet. Your customers never actually see code, except for whatever's JavaScripted, or WASMed, or whatever. And of course, that's a whole other thing. Anyway, it's just fractal, and we could talk about it for way more than an hour, but sort of my 10,000-foot overview is that there's both this ownership in the legal sense, but very much also in the code and culture sense, and we can talk about any or all of those, including to some extent how Go is different. its packaging system is one of those things that occasionally makes lawyers tear their hairs out a little bit, because it's not something -- so many of our lawyers... Not our lawyers. Well, our lawyers too, but our licenses predate Go; they predate some modern language distribution practices, and sometimes that shows up. We wrote technology-specific things for like C or C++ into our licenses, and then somebody says, "Well, how does that work in the case of Go?" and the answer is, "I have no idea." + +\[10:10\] I guess those of you listening to this as a podcast can't see the face I just made... So just assume, like, perplexed; search Giffy for your favorite perplexed GIF, and that was me just now. So yeah, so where do you all want to start with that? \[laughs\] + +**Kris Brandow:** I also have a kind of meta view of some of this, because I think similar to the line of thought you were going down, where it's like, okay, there's the code that you've typed, the thing that you've written, but then there's the knowledge of writing it and the idea of the thing, and that's where the API question is very interesting... Because it's like, "Oh, well, someone else--" Like, if you retype the same code, is that the same thing? Or what if you type it slightly differently, but it's conceptually the same thing? How far do you have to get away from -- how different does something have to be before it's "Okay, now someone else can own this thing." + +I remember over the years talking to lawyers about all the non-competes and things that we tend to have, and one of the things that lawyers consistently told me, at least in New York, is that your employer has no right to the knowledge that you gain. So they can have ownership over the code that you write and the things that you produce, but they're not allowed to say, "Well, we gave you this knowledge, so you can't take it over there and use it in one of our competitors." I've always found that very interesting as well, because it's like "Oh, well, this is like another aspect of things." It's like, as an API knowledge, or is it, the code that you've written? So there's this really meta aspect, for me at least, to the whole idea of ownership. + +**Luis Villa:** Oh, yeah. Absolutely. And by the way, you specifically called out New York, and Natalie said in the pre-chat that she wanted to ask about the EU... We as open source programmers, and open source developers of various sorts, tend to make an assumption that we can write a license that applies across the entire world. And in most law, that's like a completely laughable idea. It's somewhere between laughable and actively considered harmful. And so we're sort of lucky, in some ways, that the core concept of copyright, which is what applies to that actual written thing, distinct from the ideas, is actually a global standard. A treaty called the Berne Convention; that's 1908, maybe, or 1903. + +So copyright has been standardized across the entire world for 100 years, which makes it a good platform for lawyers to build a global system on. If you're talking about databases, no global platform, no global legal platform... So writing a database license is responsible for a lot of these gray hairs; again, sorry, podcast listeners. You should watch it live next time. And similarly, with AI, we don't know how some of this is going to play out. We can offer guesses, and frankly, very interesting guesses. This is a very fun theoretical game, but we don't know how it's gonna play out in court. + +And by the way, Kris, I think the other meta thing that's really interesting for programmers... I often like to remind developers that writing a contract is like writing code and then not executing it, at all. And you're just sort of reading it, and like, we all agree more or less on what the output should be, but until it's actually been executed, which happens through a court... Right? A court says, "This is what the thing is", then we don't actually know what the outcome is. + +So everything that I say here is going to be based on like a handful of things. Part of the challenge in Oracle v. Google is that we'd all been operating under these assumptions about what copyright law was, but there had been no litigation over whether or not an API could be copyrightable since the '80s. So the closest analogy we had was Lotus 123 dropdown menus. And it turns out things have changed a lot since then. But because it hadn't been executed by a court, we really didn't know how this was going to turn out. And so that makes predictions -- this is why lawyers' favorite phrase is "It depends." + +**Kris Brandow:** \[14:11\] There's one thing I want to stick a pin in before we move on, Angelica, and it's that idea of AI... And I also want to call out code generation, and I really want to talk about that later, because I think that's also like a very interesting thing. The first thing you said was "Oh, who's typing the code at the end of the day? That's how copyright is generated." So I just want to make sure we circle back to that later. + +**Luis Villa:** Absolutely. + +**Angelica Hill:** I will say, for those of us who weren't following that Google law case step by step every step of the way, was it had been litigated. What was the conclusion? + +**Luis Villa:** Well, so to take a step real quick back just to the very beginning, Oracle -- well, Sun had created Java, and originally the Apache Project sort of funded by IBM had reimplemented Java. Complete clean room, very strict, very effective clean room, as best as we could tell from the pieces afterwards. Literally, just a handful of lines probably, out of several hundred thousand in that reimplementation that ended up looking like they were actually perhaps copied and pasted, rather than a true cleanroom reimplementation. And so Google used that Apache reimplementation in their Android phones. + +Ultimately, what happened after literally a decade of litigation - I think I was on my fifth job by the time the case ended... Like, from when the case -- actually, six jobs from when the case started to when the case ended. The case started with essentially Oracle claiming there's -- some other stuff I'm going to leave out for simplicity, but Oracle claimed that copying and reimplementing just the API headers was a copyright violation, and that therefore, all of Android should have to be licensed from Oracle for... They originally asked for 5 billion, and I think by the end of the case, they were asking for 9, or 15 billion, something like that. + +The courts found essentially, through a series of rulings over the years in this case, that an API could be copyrightable independent of the implementation, but there was a plausible what we in the US call a fair use argument, that essentially, if you reimplement in a way that's particularly transformative, like you're doing something that is really different than what the original authors or copyright owners of the API intended to do with the API, then you have an argument that it's okay to reuse that API in that way, through reimplementation. + +Lawyers like to say that Fair Use is simply the right to get sued... It's ambiguous; it's one of these things where, again, you can't know ahead of time what the outcome is going to be, and that, of course, makes it a playground mostly for large companies, unfortunately. So I think in some ways, that was not a great outcome. It was a better outcome for open source than what could have been, than what Oracle wanted to have... But it wasn't an ideal outcome. + +**Break:** \[17:11\] + +**Angelica Hill:** I wanna ask one step back, and when we talk about code ownership, what exactly does it mean I own the code? ...whether I am an individual, a company or anything - does it mean I'm allowed to make money off it? Does it mean I can print it and hang it at home? Does it mean something else? + +**Luis Villa:** Well, I'm gonna give you my lawyer answer to that. Those of you whose GitHub accounts do things other than commit to other licenses - which is pretty much all I do these days with my GitHub account. We'll have better notions of code ownership as a cultural practice among programmers, like who's responsible... I do want to talk a little bit about that one, but let me put a pin in that and come back to it. + +The basic system since at least the '60s in the US - I'm not sure exactly the timeline in the EU, but I would imagine similar - is that... Well, actually, let me go back even further. Copyright is intended to protect creative works. So what do you have to do to get copyright in a thing? And I'll explain what copyright is in a second, but let me start with what you have to do. And what you have to do is you have to write down something that's creative. "Write down" can be broad, right? It can be sculpting, or -- but you have to take it out of your head and put it out into the real world in some way. That can be typing it in a computer, it can be, like I said, sculpting it into a sculpture; sculptures can get copyright. It can be a work of art, so it can be an oil painting, or whatever; it can be a Vim poster... Honestly, these days my development environment is Word, but I used to be an Emacs guy... + +So that is the key thing, is you are doing a creative thing, and it can be mediated by tools. And Kris, this gets to your point about the AI and where is copywriting there... It can be mediated by a typewriter, or a paintbrush, or I believe - that we don't really know for certain yet - it can be mediated by an AI. But you are doing some creative something, and turning that into a fixed thing. + +Alright, so what happens once you've done that? Actually, before I get into what happens once you've done that, because I think there's an important exception... In the US at least, that creative -- what does it mean to be creative is not zero. It's pretty close to zero, but it's not zero. There's an important case called Feist vs. Rural Telephone, and the whole thing in that case is literally, telephone books aren't creative, and so therefore they don't get copyright. Because what's the point of a telephone book? The point of a telephone book is to literally just mechanically go through a town and have phone numbers for everybody. So it's hard work, but it's not creative. And in the US, at least, you have to have some kind of creative something. + +So if you do like a phone list of the 100 most awesome people in New York City - that's creative; you had to select -- one of the ways which you can be creative under US Copyright law is selection. If you pick those 100 people, then hey, you've done something creative, and your list of 100 people is copyrightable. But if you're just "Every single person who lives in Manhattan", that's not creative; you don't get protection. And that plays into questions of databases, and ultimately, I think - and we might not have time to get to this today, but the question of the models themselves. Because there's both the output of models, what's the copyright on that, and the models themselves. We don't actually know if they're copyrightable. That may be too esoteric; you might have to invite me back for another one for that. + +\[22:25\] But okay, so you've created this thing... So now what do you do? So now you've got copyright. What does copyright let you do? Copyright lets you control what others can do with it. It lets you decide who gets to use it, who gets redistribute it, who gets to modify it, within the certain limits. But it's pretty strong. + +So the limits include what's called First Sale doctrine, which is, "Hey, I sold it to somebody. They can usually sell it to one other person." First Sale doctrine made a lot more sense in the era of like books. That's what creates used books stores, is First Sale doctrine; it means that I bought the copyrighted thing, and now I can give it to a used books store and they can resell it. In the digital age, First Sale doctrine is a little more complicated... But suffice to say, that's one of the limitations. + +Similarly, fair use says, "Hey, if you're using this for education, if you're using this for nonprofit purposes..." I'm oversimplifying a little bit here; the tests around fair use can be a little complicated. Critically, in our digital age, fair use in the US has expanded quite a bit to include what's called transformative use, which is to say, "Hey, you're doing something super-new, super-different." Courts are often going to allow that in the name of sort of not impeding progress. + +So for example, Google Book Search is in some sense the biggest copyright violation in all of history, because it's literally copied systematically millions of books, made these digital copies. But then a court said, "Well, but actually, it's so different. It's so great." And they put strict controls around, you know, you can only get a few pages at a time, and authors can opt out if they want, after the copying has been done... So like Google Book Search is a good example of what transformation means, and potentially analogous to what Copilot is doing. But we don't know. + +The flipside of this is that we just had court cases -- we had a court case a couple of years ago about the song Blurred Lines, some of you might have heard... And courts there actually said that even just sort of copying the style of the artist could potentially be a copyright infringement, which was a big surprise to a lot of lawyers. A lot of lawyers are still unhappy about that case. + +Next week there's going to be -- or no, tomorrow morning, actually, maybe... There's going to be a case about Andy Warhol doing -- a photograph of Prince that Andy Warhol transformed into one of his Andy Warhol canvases. And the Supreme Court -- it's a little weird, but I think that case might actually have a lot of impact on artificial intelligence... Because we've all done, we've all played with Stable Diffusion, or Midjourney, or OpenAI, or whatever, to create foo in the style of bar. Well, if bar is still alive, and still has a valid copyright, maybe that's a problem. We don't really know yet. + +I saw a research paper yesterday that said, "If you prompt Copilot to do code in the style of --" I'm forgetting the guy's name. Petrov, I think... A top Python programmer - that you actually get fewer vulnerabilities in your code if you prompt Copilot with the name of a top maintainer. And the flipside, the paper's author was honest enough to note that they prompted with their own name, and the number of vulnerabilities went up. I thought that was nice and humble of them. + +So style is an issue that could potentially come up in code as well. That was a very long-winded answer to your question, Natalie. I apologize. + +**Natalie Pistunovich:** No, it was interesting. So you said that for code ownership it basically means who is allowed to sell and profit off that, who is allowed to give it their own personal interpretation... + +**Luis Villa:** \[26:16\] Yeah. + +**Natalie Pistunovich:** It's also who is who's there to answer in case of a problem, right? I wrote a piece of code that made my work lose a lot of money. Is the ownership on me? + +**Luis Villa:** Yeah, so that's where it gets complicated... And we have really good answers for that in the case of things like -- if you manufactured a car wheel, and the car wheel explodes because you used bad materials in that case, in the car wheel, then we have some well-developed laws and intuitions around "Okay, well, we sue the car wheel manufacturer." Software doesn't have any of that really yet. We've sort of operated in this rules-free zone where everybody was like "Software is so cool! I guess we'll just let it happen." And I think that age is coming to an end, to be perfectly honest, actually. + +I think the European Union has published in the past year - including one last week, two weeks ago - papers on liability for software. The idea of - if a car wheel explodes and causes a car accident, we have a very clear idea of how we should figure out who is liable for that. If an AI goes wrong, or software goes wrong, and the car goes off the road and causes the exact same accident, we actually have very little idea how we should apportion liability. And it's not necessarily about -- for practical reasons, those kinds of things tend to look the same... Because at the end of the day, the company that commissioned the code is also the company that sold the product. So you tend to see those things tied together. But there's no formal reason for that, right? Copyright law doesn't -- especially because copyright law historically was about things that didn't cause car crashes. Historically, copyright law -- like, literally, modern copyright law in the US is in large part because of player pianos; like scrolling wheels... Like what you saw in Westworld, like scrolling wheels with little punch holes that cause the piano to do things... Pianos didn't run off and kill people, so copyright law doesn't really have much to say inherently about product liability. And that's something that we are, I think, screaming towards at very high velocity, at least in the EU, and I suspect because of AI in the US soon as well. + +**Kris Brandow:** I feel like there's an interesting component to that as well, because when you think about what we create, we're just creating words on a page. The manufacturing process of turning that into something that does something is not necessarily something that the person who wrote the code does. So it's like something that somebody else does. And then there's all of the "Well, the machine you run it on, like if a bug does happen with like a car that's driving - is it the fault of the person who wrote the code? Is it the fault of the machine? Is there a problem with the machine?" Who gets the blame? And I think that it gets extremely murky, because we're dealing with such new stuff, that we have never had in the existence of humanity. + +**Luis Villa:** Yeah. The original history of this in the English and US law systems was that literally, to get to that point where "Hey, the car wheel explodes. We should sue the car manufacturer" involved a lot of people dying in train accidents, and the train companies being like, "Oh, but that's not our fault. We just laid the tracks, bought the train, bought the coal... It's the guy who was driving. It's their fault. So you can't sue us." And that actually, as a matter of law, was like a good argument for decades. And then the number of accidents as trains became ever more present as part of our economy, as part of simply how things moved around - well, the technological change drove a change in understanding. Because that rule was originally from "Hey, some dude on a horse. It's not my fault." If he's my squire, or whatever, pick your ancient British legal term - if they're out on my horse, like genuinely, it's sort of not my problem if they caused an accident. I mean, yeah, I own the horse, but... + +\[30:29\] And so the train companies for a long time were like "Well, look, it's just like a horse. This train is just like a horse. I can't be responsible." So at some point, the legal system was like, "Actually, this is ludicrous." And so a combination of courts and Congress changed the rules to make the train companies more liable. No surprise, the trains then started getting safer. + +Yeah, Angelica, I've actually got a British person on this podcast, I believe... So why am I not asking you the proper terms here? + +**Angelica Hill:** Yeah. footman's fault. + +**Luis Villa:** Exactly. It's the footman's fault. So I think what we're seeing is both an exciting and a terrifying time for lawyers; we are in the midst of one of these very rare technological changes. I think AI in particular is going to be that new train. Nobody really understands, nobody wants to take responsibility, for - Kris, as you say - super-good reasons. This stuff is literally the most complicated systems ever built by humankind. We even in the best case have only the vaguest sense of how it works... And good luck explaining it to a judge. + +I had a conversation with some lawyer friends last week that was like "How would you explain...?" And these are fairly sophisticated -- most either are programmers, or one of them is married to a programmer... And we've been doing tech law all for cumulatively many decades. How would you explain machine learning to a judge? And we all just sort of stuttered in horror at that thought... Because it's a really -- again, even the programmers who haven't thought about it, it can be really hard, unintuitive; the vocabulary is changing all the time, the technology is changing all the time... And to try to explain it to Congress, or to a judge, is a scary proposition. + +It's an exciting -- whoever gets to do it first, that's going to be a super-great lawyer job for somebody, but also - boy, when you screw it up... And we felt a lot of pressure in the Google Oracle trial, that this was something that if we got it wrong, it would really hurt open source... And I suspect every good lawyer, of course, cares about their client, but some clients are just -- represent one client. And other clients represent these big systemic changes. And you feel that weight as a lawyer. + +**Kris Brandow:** I feel like there's the other side of the problem as well, because if you try and assign blame to the person who owns the copyright of code, there is a huge amount of code that we all depend on all the time that's maintained by some random dude in Nebraska... Like, individual people. And it's like, well, if you can sue to get to them, because something they wrote caused some problems somewhere down the chain, then that's obviously a problem... So I can kind of see the chain where it's just like, "Well, who actually gets the blame for the bug that was written, or the problem that happened with the code?" because you can easily just like keep tracing that back further and further and further, by passing on the plane. + +So it's like, once again, with the train, it's like, "Well, I didn't create that wheel. Someone else created that wheel, so that's their problem." Or with the Spectre/Meltdown hardware problems, where it's like, "Oh, well it's not my fault that there was a breach. The processor shouldn't have been speculatively executing." There's so many weird arguments that you have because of this stuff that we don't really understand what it is right now. + +**Luis Villa:** Yeah. And I think applying the old models is probably going to get us some very bad outcomes... And unfortunately, the way the legal system learns sometimes is by having bad outcomes. Everybody stubs their toe on it, and then you sort of fix that up as we go. But some people end up being caught in the middle. That guy in Nebraska - I assume all the listeners here have seen the XKCD comic about the guy in Nebraska. The problem is, of course, it's not even just one guy in Nebraska, it is a tower of 10,000 guys in Nebraska. + +\[34:10\] And so I do want to talk a little bit about the day job here, because I co-founded a company called Tidelift. And Tidelift's mission, as we said at the top of the show, is to make open source better for everyone, in part by paying the maintainers. Because what we're seeing happen all the time - we saw it happen a couple times this week just in the JavaScript community - is the solution to this kind of problem that you've identified, Kris, so far is "We'll just start applying standards." So we've got the Open SSF standard security scorecard, we've got salsa.dev, which is a different kind of security scorecard... GitHub caused some controversy by saying "Hey, we've identified the most popular - I think only npm for right now - projects, and we're sending you all a free two-factor authentication key. And also, we're mandatorily turning on two-factor authentication for everybody." And a couple of maintainers, for various reasons, were just like, "That's too much work. That's gonna complicate my life, it's gonna break my build scripts..." + +And we can go back and forth about like whether or not two-factor authentication in some of these cases is a good idea... But I want to step out -- I mean, I generally think two-factor authentication is a good idea, don't get me wrong. But that's the easy case. It just gets harder from there, right? Like, "Okay, what do we need to do to sign our binaries?" In Go, I understand that signing modules is mostly a solved problem, and a lot of other language ecosystems it's not. So okay, so there's extra work, and we've just created this extra work, Kris, as you say, on some guy in Nebraska, or actually a stack of 10,000-some guys... And I apologize to listeners, it's probably grating for me to say 10,000, guys, but I think it's worth both admitting that this is a problem, and I think saying that part of the problem of the gendering of open source is very much that in a world where women are often called on to take on more than their share of household duties, and home care, child care, elder care, if we're not paying people to do open source - well, guess what, that's part of how we get guys doing it... Because guys, for various cultural reasons, have more free time. That's an important side note that I think is important to say. + +Anyway, we're putting all these new requirements on people, Kris, because of exactly this intuition you've had about -- but we're doing nothing to make open source more fun, easier... Like, all we're doing is loading more work on top, and I think at some point -- I think we're already starting to see it in some communities, where people are gonna snap, people are gonna walk away, people are gonna say-- + +So what do we do? So what Tidelift does to help address this problem is we go to our customers and we say, "You're going to get more predictable, more reliable open source if developers follow these standards. If they're not going to follow these standards on their own, you should pay them. So if you want the stuff you use to follow those standards, write us a check; we will go out, find those developers for you - hopefully, we already have a contract with them from other customers, but... We'll go out, find them and pay them, in order to get some of this work done." + +Now, there's a lot of challenges around this, in part because - guess what? Nobody wants to pay for open source. It should be free. And well, guess what? If you're liable, all of a sudden maybe it's not free. And I think one of the interesting things that we're gonna see in discussion about these EU regulations, for example - they contain exceptions for open source. The definition of open source sort of looks like it excludes commercially-sponsored open source, which as we know, is a lot open source these days. We don't know how that's gonna play out, we don't know what it really means. Like, their definition is vague enough that maybe it only includes a small slice of open source, or maybe it includes a lot open source; we don't really know yet. I'm sure that's gonna be lobbied over... And in fact, I'm going to publish something, hopefully tomorrow or Thursday, on dev.to about what open source developers can do to help lobby the US government on this topic, but a lot of the same thing's going to apply to the EU government as well. + +**Kris Brandow:** \[38:16\] Yeah, one of the thoughts I had during what you were saying is - I express this in private to some people, and I always get kind of the "You've just said heresy" look or comments... But I have been wondering, is open source sustainable as the method of how we do things in this industry? Is this focus on sharing so much of code actually going to be something we can continue doing in the future, since there's so much ambiguity around all of this? And quite frankly, I think it also atrophies the whole industry, because we're not rewriting things. We're not reimagining things. I think that's one of the core problems with copyright in general - there's this whole thing that Disney has done, where it's just like... Copyright used to be like 20 years. Now it's almost forever; or as Disney would like to have it, actually forever. So it's like, "Oh, well, now these things are just kind of sticking around, and it's so much harder to move things forward." + +If I remember the kind of genesis of copyright, or the vague genesis, it's like "We want to protect people, making so they can profit off of their creative work for some time", but then it goes back into the general pool of things so we can kind of continue making progress forward. + +**Luis Villa:** Yeah... Boy, that sustainability question is a big one, Kris, and I really don't know. I'd like to say that we have a real clear answer to it. Certainly, I think that Tidelift is part of the answer to that... But I think it's a really good question to be asking, and that heresy - it is an elephant in the room that a lot of people... I've gotta say, I get a little frustrated when an employee of a trillion-dollar company is like "I don't know... Paying people... Open source seems to work fine for me." I'm like, "Well, yeah, because you literally work for a trillion-dollar company, but a lot of the software you rely on, and certainly that your customers rely on - they don't have that luxury." + +It is the kind of thing -- you know, you've got a puppy... Puppies are often more fun than replying to pull requests from automated bots. And, Kris, I think one of the interesting things - we've been a little backward-focused, but I think there's a lot of cool stuff... I know we're flying through this time, but I think there's a lot of interesting questions about this future-looking around machine learning, Copilot, things like that... And part of those go to what you're saying -- Kris, you were saying "Well, the original motivation for copyright was to--" There were sort of three original motivations for copyright that vary depending on who you're talking to you. + +So in the US, in the Constitution, it says that the purpose of copyright is to encourage authors. So it is a very utilitarian, like, "We're going to give you this copyright, and as a result, you're going to create more stuff, and that's a bargain that we're gonna have. We're gonna give you this monopoly", which otherwise the founder's super against monopolies, but they created the copyright monopoly specifically in order for the rest of us to benefit from this incentive of a whole bunch of stuff being created. So that's one story. + +In the EU, and really actually most of the rest of the world except the US, it's more like, "Your creativity is like a part of you. It is part of your human -- your human nature is to create", and so there's often what are called moral rights; the idea that inherently you have some control over the thing that you've created, even if it's not productive, right? Even if there's no social value to it. And we're going to be running headlong into that with all of the foo in the style of bar. Bar is going to be really irritated that their moral rights were infringed. + +And by the way, the third historic - like, the original copyright, was literally just basically a tool of censorship for the UK government in the early 1600s. It was a way for them to control printers. And I think, Kris, there's a really interesting discussion we're going to have over how does open source -- it probably won't be mediated through copyright, but as you were saying about liability, and security... How does government interact with this? Because it's one thing if these big businesses are running around saying, "Hey, we should make this stuff more secure." It's very different if governments are running around saying "The whole world needs you to be more secure." That's the big thing that we still haven't really wrapped our heads around. + +**Kris Brandow:** \[42:29\] Yeah. It doesn't help that our legislators aren't very tech-savvy, and they tend to write a lot of laws that you're like "This makes no sense." Or ask questions in hearings that are questionable at best. + +**Luis Villa:** Yeah. There's two parts of that. That's the one -- and again, I'll try to be quick here. That's the one that everybody thinks about, because we've seen our legislators on TV, and it's terrifying. But there's also this thing where -- so legislators sort of provide you like a rough draft, and then the courts are used to refine that. But because litigation has gotten so expensive, and everybody hates it, we don't do the refining of the rough drafts anymore. Like, it sort of becomes industry convention, and that sticks us with a lot of cruft, that I think is a problem. Natalie, you had a question you wanted to...? + +**Natalie Pistunovich:** I wanted to say that -- that's a couple of topics back, but I had a... I'm a contractor, so I have clients around the world, I see all sorts of different contracts... And one time I had a contract with a California-based company, and there was a clause that said that any damage that I caused, I am responsible of it. So back to the conversation about that Nebraska person... + +**Luis Villa:** I mean, contracts can say that... The good news/bad news is that you were probably what is known as not a deep pocket... Like, they're unlikely to sue you, because what are you going to give them? Your collection of goldies would not be worth much to them. + +**Natalie Pistunovich:** I was really terrified. Exactly. + +**Luis Villa:** You probably should have negotiated that out, but that's one of these ways in which the legal system is very imbalanced, unfortunately... And that's a whole other rant, for a whole other show. + +**Natalie Pistunovich:** Especially the American one. But the thing is, some friends who work in California said that this is actually a normal clause, that they had this in the contract in the past. I think nobody working in California is on this panel, so maybe somebody listening can keep me honest here... But I've been just told "Yeah, don't worry. It's always there. Just don't take it seriously." + +**Luis Villa:** Well, there's a whole other thing about cruft in -- again, another rant for another day... How lawyers deal with cruft could learn a lot from how programmers deal with cruft, because a lot of the stuff that is -- we don't have any sense of dependencies, or module reuse or anything like that in law... And so you get stuff that literally just gets copy-pasted. Like, imagine if you copy-paste all your code all the time... We as programmers know, of course that creates errors. And we have linters, and dependencies, and we have all that kind of stuff. Lawyers have none of that, and that is a problem... Though I'm curious to see if machine learning helps us with that in the future. + +**Break:** \[45:04\] + +**Kris Brandow:** I was gonna say that, like, when you were talking about how we have the laws get written, and then they don't get tested or refined, I'm like, "That's kind of like writing code, but without tests." It's just kind of like, "I don't know, it's just running out there, and we have no idea if it's doing the right thing, or doing what we intended. We just wrote it." + +**Luis Villa:** There's no test frameworks, there's no linters, there's no -- by the way, when you compile it, there will be somebody else trying to persuade the compiler to do things totally different from what you intended... It's a very adversarial system that is not set up for robustness. Don't get me wrong, it works reasonably well in a lot of cases, but it's mostly because of humans -- this was one thing that drove me nuts about all the smart contract stuff... Like, contracts only work because humans are around to smooth off the rough edges, bridge the gaps... Like, as soon as you start making contracts into code, you're just doomed to failure, because of all the failure modes of code that we as programmers know, and that don't go away when you have contracts; you just have a more forgiving, at the end of the day, execution environment... Because at the end of the day, humans are in the loop, in a way that they aren't with smart contracts, and in the way they are with code. + +**Kris Brandow:** I find it interesting how it also causes some class problems as well... I have legal insurance, so every time I get a contract of any sort, I send it to lawyers to review. I'm like, "Is any of this weird?" But also knowing that I can just like take a pen and just strike through anything I don't like in a contract, when so many people think that's "Oh no, I've gotten this thing. This is like set in stone, and I can't do any -- I have to take it or leave it." And it's like, that's not how this really works." So knowledge also makes it so the legal system's kind of like a little bit more wonky than I think it would be otherwise. + +**Luis Villa:** Yeah. Well, and on the flipside, of course, especially in the US, Natalie, I think you're correct to say that this is less of a problem in the EU, though definitely not unknown as a problem... You get lawyers who end up working more defensively than they might otherwise... Because they're thinking of the person like you, Kris, and so extra effort gets puts in, extra layers get put in... And that's not necessarily a bad thing, but it does mean -- there were virtues to the days of the handshake deal. High trust environments, versus low trust environments is a real thing in law, unfortunately, for better or for worse... And it's absolutely -- there's all kinds of class and privilege issues around that. Again, another rant, another day. + +**Angelica Hill:** And that seems to be the case with every episode we do, Natalie. We're gonna have to get you back for a part two, to go deeper dive. I have one more question that I want to dive into before we go to the unpopular opinion section, which is - and you alluded to this earlier... Are there specific considerations when we're talking about Go? I know you talked about kind of package management, but what are the specific legal areas when it comes to ownership of code that are brought in specific to the Go software engineering language? + +**Luis Villa:** \[49:54\] I think there's one -- I'm getting a little over my skis here, but Go is very much a -- because of the way that you all have done packaging, it has some real implications I think for... It breaks the brain of a lot of people who come out of like other language package managers... And it has some stuff that really made sense internally to Google... There's this real question - Kris, you were alluding to it earlier, the guy in Nebraska, XKCD... My impression of it from an outside perspective is that Go wants to make some of that not a problem by like, "Oh, we'll just grab this specific revision from this specific repo, and voila... We know exactly what this code is, and that guy in Nebraska can't hurt us anymore. He can't upgrade, because we know exactly what it is that we've got here." But of course, that ends up being a sort of adversarial relationship with that person. It means that when that person brings knowledge to the table, brings new versions to the table, there's this sort of assumption... + +We never really got to it here, Natalie, but I think there's this penumbra of -- it's not ownership, but it's sort of entitlement, almost... And this is certainly not Go-specific, but this sense of like, "Well, I'm using it, and so I'm going to treat it a little bit like I have a support contract, like traditional software." And in fact, open source has sort of encouraged this, because people -- it started from this very collaborative, community-based culture, and so the norms are like, "Hey, I'm gonna help everybody who shows up in my issue tracker." And of course, at some point you get too popular and that breaks. But we haven't really acknowledged that that's -- I mean, literally, the word "entitlement" comes out of some of the same roots, the same Latin legal roots as ownership. When you own a thing, you are entitled to do x. And we end up with entitlements without the ownership part, without the payments part, without the labor part, and that really is what sends a lot of things sideways, I think, in open source. Again, that's where Tidelift comes in, and I think also Go, I think, has tried to cabin some of that off with its module ownership, but sometimes where that makes technical sense, it may not always make social sense. And with more time, some other time, that would be a fun conversation to have in more depth with some of the folks who know more on the packaging side than I do, that's for sure. + +So, I guess the unpopular part is not Go-specific, but that's definitely the responsibility lies with all of us, because we specifically decided in open source that "Well, I'm not owning this thing in like some sense", but we very much decided in other senses that... Like, one way of putting it is we decided that use of the code translates into ownership of somebody else's time... And that is not -- when open source runs the entire world, which it really does at this point, that doesn't scale, and we don't know yet, Kris, to your earlier point of heresy... We don't know how that's gonna continue scaling. We don't know what happens when we ask everybody to do really complicated, multifactor signatures on everything... With very few exceptions, people don't hate this stuff. They're software developers; like, they know security matters. They know signatures help. But they also know they've only got so much time in the world, and sometimes they come home from work, and they really just want to pitch their laptop into the lake. And that does not help you close out your issues if you've pitched your laptop in the lake. So we're gonna have to figure that out. It might involve buying some people, some laptops, or maybe writing them a check, or maybe it'll involve helping them with AI, which we really didn't get to at all, but again, maybe next time, so... + +**Kris Brandow:** I remember a conversation that happened I think among a smaller group of people within the community, but when modules were being designed and developed, one of the comments that kept coming up was "This is biasing toward the consumer instead of the provider, the maintainer. Is that a thing we really want to do in the Go community? What effect is that going to have on people's ability to actually maintain and build open source things?" And I think we're really starting to see some of the outcomes of that, like what Ben Johnson has been doing, where he's just like, "I don't want contributions. I don't have enough bandwidth to kind of deal with some of these things." + +\[54:14\] I think we're gonna see a lot more people that just are like, "Well, I can't --" Like, if there is a bug, and now I have almost no recourse to fix it, or I can't get people off old versions of things, that does kind of erode the ability of people to do open source really, which erodes our ability to maintain the modern world, in a sense. + +**Luis Villa:** Yeah, absolutely. And that's one thing I would say to the Go folks - you're not alone in that. Every ecosystem is struggling with that. There are different flavors caused by different technical choices and different cultural choices along the way... But the core problem - we did a Tidelift conference right before the pandemic. I can't wait to do our next one, because very much a theme is going to be "How can people across many language ecosystems share notes, figure out what this looks?" Because it's very much not -- if you feel alone, if you're having a bad day and you want to chuck your private keys away, and never log into GitHub again, you're not alone in that. That is a common thing. And GitHub is going to bust their butts. I will give them credit, they're trying to do a lot to make some of this easier for maintainers... But ultimately, Kris, as you say, big companies are going to bias towards doing the right thing for the consumer. + +Microsoft has done a lot of amazing things for the open source community, which 1997 me is like aghast that I'm saying that out loud... But at the end of the day, when it comes down to push or shove, the decisions are often going to be made in favor of the consumer, and we do need to have some of those honest discussions about what that looks like... Because Natalie, to get back to some of what you're saying, and the overall theme of the show - legal ownership is only part of the story here. Cultural ownership, responsibility, entitlement - all these things are related to, but cannot be solved just by our legal systems. And I think maybe that's my one regret. I have very, very few regrets about going to law school. It was a lot of fun, I met a lot of great people, but people often come to me seeking legal solutions for what are ultimately cultural problems. And I can only do -- the best lawyers know how to straddle that gap. And I like to think that that is certainly my biggest strength as a lawyer, especially in this space, is how to straddle that gap. But it's not easy. Which actually, by the way, reminds me - side project, fun project, and then we'll leave it... And I know we're running out of time here. I am writing a newsletter called OpenML.fyi. It is new. I literally sort of launched it to some friends a couple of weeks ago, and more broadly yesterday... But it is literally about these questions about open and machine learning overlap, which includes questions "Is this the end of no warranties?" Because every open source license, as you point out, Kris, has like big, all-caps text that says "No warranties. If you break it, you buy it." And what does that mean in light of EU regulation? We'll be talking about a lot of this stuff, like Copilot there as well, so I'd love to have you all back, but for those of you who are curious about that topic, it's a ghost AGPL-powered newsletter, OpenML.fyi. + +**Natalie Pistunovich:** We'll add that in the show notes. + +**Kris Brandow:** I want to end with -- because I know we've got to get to unpopular opinions... But I just want to say, one of the things I always think about when we get into these conversations is like people tend to think like humans are like transactional, and they're kind of mean to each other, and we want a war, and all of that... And I feel like the existence of open source and the existence of our industry as a whole proves that people are a lot of times selfless, and will sacrifice a lot just to make other people feel good; just for the happiness of other people. And I think that shows how incredible and how collaborative we are as a species. And I think more of us need to remember that, especially in the times we are now. We are not necessarily this always angry at each other, always warring, always territorial species; quite often, most of us are just like these -- we just want to help our fellow people out. + +**Luis Villa:** It's been too long since I worked at Wikipedia, but... I mean, here's this thing, it's this amazing cultural treasure, and anyone can go and graffiti on it at any time. And something like 1 in 1,000 edits are spam. Think about what that says as -- to exactly your point, Kris... Actually, most of the time, most people - we all want to make this work, right? And open source, open data are very much, I think, like, genuinely amazing... That's why I enjoy doing it. There are a lot more lucrative things probably all of us could be doing with our lives. But yeah, it's human and humane in a great way. + +**Angelica Hill:** Lovely not to end the episode on, we're gonna have to get you back for a part two, for sure. But before we let you go, we're gonna be doing a little bit of unpopular opinions. + +**Jingle:** \[59:03\] + +**Angelica Hill:** So over to you, Luis.... What is your Go Time unpopular opinion? + +**Luis Villa:** Oh, boy... The one I have in the show notes is absolutely one I already nailed, which is "Hey, we should all be paying for this, right? We got it for free for a long time, and that train is running out", for very human, decent reasons. It's not like I think companies are bad for having used this stuff. But as Kris was saying, sometimes you raise that employee company... I will never forget - so there was this project that I was invited to come to a meeting about. I won't be specific, but it was one of many, many, many, many, many open source metadata projects. And people went on for like about 45-50 minutes, and I was "Okay, but why are volunteers going to create all this metadata for you?" Quiet. Silence. Quiet. Silence. "Okay, but why? What's their motivation?" "I don't know, it's probably just gonna happen." + +Needless to say, that project has not really gone much of anywhere. But I was treated as like a pariah, and literally not invited to future meetings for a while, because I had dared to ask this question of "Why would people do this?" Unfortunately, I still get that all too often. I think, to be fair, lots of people are getting the message, finally, but it's taken longer than it should have. That's my, sadly, unpopular opinion today. + +**Natalie Pistunovich:** It's like that meme about that guy that's being thrown outside of the window... \[laughter\] + +**Luis Villa:** Wait, which -- thrown outside of the window? Now I've gotta google this... + +**Kris Brandow:** Where it's like, they're all in the meeting... + +**Natalie Pistunovich:** It's like this comic strip that -- + +**Luis Villa:** Oh, yeah. Yeah. Yup... I've been that guy. + +**Angelica Hill:** Intrigued to see how unpopular or in fact popular your opinion is. And then I want to ask you, Kris, for an unpopular opinion, given that we're just getting you back... I'm sure you have something on your mind; you always do. + +**Kris Brandow:** I have so many unpopular opinions... I don't think this is gonna be unpopular... So I think most people probably agree... But it's like a thing I want to put out into the universe more, and that is that every tech company larger than probably 20 or 30 people should hire a librarian. + +We create ridiculous amounts of information, but then we usually just dump it into a wiki, and then we're like "We'll be able to find it. Just use the search functionality." Or we like try and make a docs page, and we're like, "Users will be able to find stuff." And it's like, there's an actual degree program of people who get doctorates in how to arrange information so people can find it. Go hire them. + +They're not even that expensive to hire... Just go get a couple, a librarian and an archivist, and then make your data and your information just much more clean and much more organized. It will probably help you make a lot more money in the long run, and make your engineers less frustrated with the world. + +**Natalie Pistunovich:** How come not all database companies, and Google, and so on, are hiring librarians and archivists to do this? + +**Kris Brandow:** I'm assuming people don't hire librarians because they're just never -- a) I think most people don't know what librarians actually do. I think most people just think librarians are the people that can help you find books in the library. And they don't think much more about that. They don't think about "But how do they help you find the books?" They're just like, "Yeah, they just helped me find stuff." So I think that's part of it. + +And unless you sit down and think about what the problem is, I don't think it's like that kind of clear thing. You're not gonna look necessarily outside of the world you exist in. You're gonna be like, "Oh no, this is the world... We can do this with computers. We can just write some code that will do some indexing, and that'll work." Like I always look at books, and I always look at the indexes they have, and like, "I'm someone is trained, probably has like a high-level degree in how to actually pick what words go in an index." That's like a really challenging job, because there's a crapload of words in a book. Like, well, which ones do I pick and put in that index? It's like, well, no, that's like a hard job. And yet, books forever -- well, not maybe forever-forever, but for a very long time, had indexes. And it's like, well, we should probably get those people. + +But yeah, I think most of the time, we as technologists are just like, "No, no, our technology will just do it for us. We'll write some stat stuff, or some ML, or AI, or whatever, and it can obviously replace the thing that humans have been doing very well, for a very long time, even though we have no idea of what that degree program or industry is at all." Typical things that we do. + +**Luis Villa:** The Times published a book review yesterday, of a book on the history of indexes, which apparently has like three separate indexes, this book on indexes... So it looks really interesting. + +**Angelica Hill:** And I promise I did not tee that up. It's not an overt company ad. Yeah, The New York Times actually does some really good work... \[laughs\] + +**Luis Villa:** I genuinely forgot that there. + +**Angelica Hill:** I'll drop your check off later. \[laughs\] + +**Luis Villa:** Just a discount on my subscription, that's all I ask. + +**Angelica Hill:** We'll chat. It has been an absolute pleasure having you on this show. Thank you so much for joining us. It's also wonderful to have you back, Kris, and wonderful as always to have you co-presenting with me, Natalie. And regrettably, we're now going to have to say goodbye... So thank you all. I'm hoping to have everyone together again soon. + +**Luis Villa:** Yeah, absolutely. + +**Natalie Pistunovich:** Thank you! diff --git "a/Who owns our code\357\274\237_transcript.txt" "b/Who owns our code\357\274\237_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..14420f1763ef92829e4e32617f5d9b5c86465577 --- /dev/null +++ "b/Who owns our code\357\274\237_transcript.txt" @@ -0,0 +1,1670 @@ +[0.00 --> 6.04] If you manufactured a car wheel and the car wheel explodes because you used bad materials in that +[6.04 --> 12.04] case in the car wheel, then we have some well-developed laws and intuitions around like, +[12.14 --> 17.16] okay, well, we sue the car wheel manufacturer. Software doesn't have any of that really yet. +[17.42 --> 21.70] We've sort of operated in this rules-free zone where everybody was like, +[22.08 --> 28.76] software's so cool. I guess we'll just let it happen. And I think that age is coming to an end. +[30.00 --> 38.14] This episode is brought to you by our friends at Square, developing the platform that sellers +[38.14 --> 42.44] trust. Here's what you can do with Square. You can bridge more experiences. You can build online, +[42.66 --> 47.24] mobile, and in-person commerce experiences that connect more customers and sellers. +[47.62 --> 51.78] You can build custom booking solutions. You can create and track orders. You can accept payments. +[51.78 --> 56.34] You can manage and curate inventory. You can organize customers. You can manage employees. +[56.34 --> 62.60] You can extend Square gift cards to your app. You can use Afterpay. And all this is powered by the +[62.60 --> 68.48] world-class Square APIs and SDKs that enable you to build full-featured business apps for yourself +[68.48 --> 73.26] or millions of Square sellers. So much is available as a Square solutions partner. +[73.70 --> 79.04] Learn more and get started at changelog.com slash Square. Again, changelog.com slash Square. +[79.04 --> 101.84] Let's do it. It's go time. Welcome to go time, your source for diverse discussions from all around +[101.84 --> 107.64] the go community. We record live each and every Tuesday at 3 p.m. U.S. Eastern. Subscribe at +[107.64 --> 113.84] youtube.com slash changelog to be notified and join the GoTime FM channel of Go4Slack to chat along with us. +[114.20 --> 118.90] Special thanks to our partners at Fastly for delivering GoTime super fast all around the world +[118.90 --> 126.62] and to Fly.io. Deploy your app servers close to your users. Learn more at Fly.io. Okay, here we go. +[126.62 --> 140.20] Hello and welcome to GoTime. Today we are going to be talking about who owns your code. A question +[140.20 --> 144.42] that certainly has been on my mind. So we're going to be exploring who owns the code. The company, +[144.64 --> 149.96] is it the engineer, is it the team, is it all of the open source contributors if it's a project. +[150.30 --> 156.02] How about when you're using AI, machine learning, GitHub Copilot. Is it still your code? +[156.68 --> 162.04] I'm really excited. We have a really brilliant guest with us today. We have Louis Veer, who is a +[162.04 --> 168.40] programmer turned attorney who has been involved in open source since college. He's worked at +[168.40 --> 175.26] Mazzella, where he revised the Mazzella public license at Wikimedia Foundation, where he also +[175.26 --> 180.66] briefly led the community team. And as a lawyer, he's worked with Google, Amazon, and many other small +[180.66 --> 186.26] startups. So currently he's the co-founder at Tidelift, which works to make open source better +[186.26 --> 192.32] for everyone by paying maintainers. We'll hear more about that. But before that, I'd like to introduce +[192.32 --> 197.58] our co-hosts. We have Chris. Hi, Chris. I haven't seen you in a hot second. +[197.58 --> 205.14] Hello. I am back after a very, very long but much needed break. So I'm feeling rested and I'm ready +[205.14 --> 209.74] to get into the meta of this Who Owns Code. It's going to be fun. +[210.24 --> 216.14] You're ready. I'm ready. It's a very interesting topic. And the beautiful Natalie, I've seen you, +[216.14 --> 221.88] I think, far too often for your own liking in recent weeks, my wonderful co-hosts. +[221.92 --> 224.38] It's like our weekly one-on-one, but it's not one-on-one. +[224.58 --> 228.06] Weekly one-on-one with anyone who decides to tune into the live. +[228.98 --> 230.44] Weekly anyone. I like that. +[230.58 --> 234.12] Yeah, weekly anyone. Beautiful. Me and Natalie don't do one-on-ones, we do anyone's. +[234.76 --> 235.82] That sounds terrible. +[239.76 --> 244.86] Lewis, would love to hear a little bit more about you and your thoughts on code ownership. +[244.86 --> 252.20] Well, you know, it's been a long time since I wrote anything approaching useful code, +[252.38 --> 257.72] but I've been involved. I had an interviewer ask me while I was interviewing for my first law school +[257.72 --> 263.00] job out of law school, law firm job out of law school. And they said, well, why, you seem to +[263.00 --> 267.88] really like tech. Why did you leave tech? And I said, look, I'm not leaving tech. Like I am only +[267.88 --> 273.00] interviewing with law firms that are very much tech forward, tech first kind of law firms, right? +[273.00 --> 279.50] So the goal was never to leave tech. The goal was I was at a startup, open source, back in the first +[279.50 --> 285.46] year of the Linux desktop, I was at a small startup. We got acquired. During that acquisition process, +[285.46 --> 292.60] I worked with the attorneys and I was arrogant. I was young. I was like, oh, I can do a better job +[292.60 --> 297.50] than these people. So I decided after a little bit of experimenting and I took like a night school +[297.50 --> 302.44] law class that I enjoyed, a couple of night school law classes that I enjoyed. And so I decided to go to law +[302.44 --> 308.38] school, right? But the goal was always all along to continue to focus on tech and specifically very +[308.38 --> 315.10] much to focus on open law, because there seemed to me at the time to be a body of lawyers who were +[315.10 --> 322.26] sophisticated about technology, but they came at it very much from a patent first, control first kind +[322.26 --> 327.04] of mindset. And that was something that was already starting to break down at the time. So I was in law +[327.04 --> 335.66] school 2006, 2009. And it was beginning to be an understanding amongst legal academics that open was a +[335.66 --> 343.64] thing. I had attended a conference of legal academics where Creative Commons was announced. That was 2001. So there +[343.64 --> 350.20] were some legal academics who got it, right? And in fact, I pretty only applied to law schools that had at least one +[350.20 --> 356.12] faculty member who had written something that indicated that they got it in the slightest amount. +[356.20 --> 359.32] So that meant applications were easier because there wasn't that many schools to apply to. +[359.82 --> 363.78] And I think that has worked out well, right? It's been a good career. It's been a fun career, +[363.90 --> 371.28] right? Because I very much, you know, the point was not that, oh, I can like make my piles of money and +[371.28 --> 376.64] work 3000 hours a year or whatever it was. I have friends who are open lawyers who deserve better +[376.64 --> 380.70] lawyering than the lawyering that they were getting at the time. And I think that's been +[380.70 --> 388.28] that sense of, hey, I'm doing this to help open people get better lawyering has served me well as +[388.28 --> 394.00] a sort of motto and mission and has led to a lot of fun outcomes, right? I mean, because open people +[394.00 --> 398.06] are doing a lot of fun projects, have been doing a lot of fun projects, and that hasn't changed. +[398.20 --> 402.80] And I certainly don't think it's going to change anytime soon. But a lot of it does come back to this +[402.80 --> 408.66] question of like, who owns the thing, right? And I admit, I have enough lawyer brain worms at this +[408.66 --> 413.36] point that my immediate thought goes to like, okay, well, the contract of the and the one of you +[413.36 --> 417.30] during the top prep said, well, you know, well, what about like team ownership? And I was like, oh, +[417.70 --> 424.70] right. We talk in law school about this analogy that ownership is for reasons I don't remember, +[424.70 --> 430.40] or maybe never knew. We talk about ownership being a bundle of sticks. And the idea is that we sort of +[430.40 --> 435.98] we talk about it as if it's like one big trunk. But it's actually like, a lot of different small +[435.98 --> 442.52] things, right? And ownership in the code world is very fragmented, because there's a sense of like, +[442.60 --> 447.82] well, okay, for almost all of you, if you are working for a company, at the end of the day, +[448.38 --> 454.02] your company owns the code that you are writing, at least on the company time, as long as you're unless +[454.02 --> 459.78] you're very careful to like, not do it on company time, not do it on company hardware, and to do it in +[459.78 --> 464.24] areas that are unrelated to what the company is working on. If you're doing those things, +[464.24 --> 469.54] and you keep it, but otherwise, as a general rule of thumb, the company owns it. So that's like the +[469.54 --> 473.58] sort of lawyer brain answer, like, well, yeah, okay, we're done here. You know, it's been a nice +[473.58 --> 479.82] podcast. Glad to talk to you all. But there's very much, of course, all these fractal little senses of, +[480.62 --> 485.04] well, okay, but what does it mean when the team, you know, team versus individual ownership? +[485.04 --> 489.60] Because in some, like in a lawyerly sense, like the company is the one who can sell it, +[489.60 --> 494.74] but who has responsibility for it within the company? Like, that's not a legal question. That's +[494.74 --> 502.40] a, that's a team norms, team behaviors kind of question. And there's also these questions of +[502.40 --> 506.92] what exactly is it that you own, right? Because I spent several years of my career, I am a, +[507.46 --> 512.30] what we call a transactional lawyer, like basically, I do contracts. If the contract goes wrong, +[512.30 --> 516.88] for some reason, that is somebody else's problem to argue about it in court. And so I've only ever +[516.88 --> 523.52] been to court for work once, which was a little case called Oracle v Google. And you might've heard +[523.52 --> 531.78] of that one. And the question at some level was about who owns or can anyone own the idea of an API, +[532.74 --> 536.08] right? And that's probably not something you're thinking about too much in your day to day, +[536.18 --> 541.12] right? And your corporate lawyers probably aren't most of the time either. They're not thinking about +[541.12 --> 546.08] who owns the API. They're just thinking about like this file, right? Or this binary that we're +[546.08 --> 552.06] distributing, or these days, often this SAS that we're putting out over the internet, you know, +[552.06 --> 558.04] your customers never actually see code except for whatever's JavaScripted or WASM or whatever. But, +[558.42 --> 562.18] you know, and of course that's a whole nother thing. Anyway, I mean, it's just fractal and we could +[562.18 --> 567.72] talk about it for way more than an hour, but that's sort of my like 10,000 foot overview is that it's, +[567.72 --> 573.70] there's both this ownership in the legal sense, but very much also in the code and culture sense. +[574.06 --> 577.86] And we can talk about any or all of those, including, you know, to some extent how Go is +[577.86 --> 582.78] different. I mean, its packaging system is one of those things that occasionally makes lawyers +[582.78 --> 588.60] tear their hairs out a little bit because it's not something so many of our lawyers, not our lawyers, +[588.60 --> 595.18] well, our lawyers too, but our licenses predate Go, right? They predate some modern language +[595.18 --> 600.38] distribution practices. And sometimes that shows up. We wrote technology specific things +[600.38 --> 606.82] for like C or C++ into our licenses. And then somebody says, well, how does that work in the +[606.82 --> 611.88] case of Go? And the answer is, I have no idea. I guess those, those of you listening to this as a +[611.88 --> 617.32] podcast can't see the face I just made. So just assume like perplexed search Giphy for your favorite +[617.32 --> 621.76] perplexed GIF. And that was me just now. So yeah, so where do you all want to start with that? +[621.76 --> 627.22] I also have a kind of meta view of some of this because I think similar to what you were, +[627.50 --> 631.68] the line of thought you were going down where it's like, okay, well, there's the code that you've +[631.68 --> 637.68] typed, the thing that you've written, but then there's like the knowledge of writing it and the +[637.68 --> 642.18] idea of the thing. And that's where like the API question is very interesting because it's like, +[642.24 --> 649.28] oh, well, someone else, like if you retype the same code, is that the same thing? Or like, +[649.28 --> 653.28] what if you type it slightly differently, but it's conceptually the same thing? Like how far do you +[653.28 --> 658.00] have to get away from, how different does something have to be before it's like, okay, +[658.80 --> 664.56] now someone else can own this thing? I remember over the years talking to lawyers about, you know, +[664.60 --> 668.32] all the non-competes and things that we tend to have. And one of the things that lawyers consistently +[668.32 --> 674.16] told me, at least in New York, is that like your employer has no right to the knowledge that you +[674.16 --> 678.62] gain. So they can have ownership over the code that you write and the things that you produce, +[678.62 --> 682.40] but they're not allowed to say, well, we gave you this knowledge, so you can't take it over there +[682.40 --> 686.96] and use it against, in one of our competitors. I'd always find that very interesting as well, +[686.96 --> 690.80] because it's like, oh, well, this is like another aspect of things. It's like, is an API +[690.80 --> 695.58] knowledge or is it like the code that you've written or is it, so there's this really meta aspect +[695.58 --> 701.40] for me, at least to the whole idea of ownership. Oh yeah. I mean, absolutely. Right. And by the way, +[701.52 --> 707.82] you know, you specifically called out New York and Natalie said in pre-chat that she wanted to ask about the EU, +[707.82 --> 713.68] we as open source programmers and open source developers of various sorts tend to make an +[713.68 --> 721.26] assumption that we can write a license that applies across the entire world. And in most law, +[721.44 --> 727.72] that's like a completely laughable idea, right? Like somewhere between laughable and like actively +[727.72 --> 735.66] considered harmful. And so we're sort of lucky in some ways that the core concept of copyright, +[735.66 --> 743.00] which is what applies to that actual written thing distinct from the ideas is actually a global +[743.00 --> 751.96] standard, a treaty called the burn convention. That's 1908, maybe 1903. So copyright has been +[751.96 --> 755.86] standardized across the entire world for a hundred years, which makes it a good platform for lawyers +[755.86 --> 761.42] to build a global system on. If you're talking about databases, no global platform, +[761.42 --> 766.86] no global legal platform. So writing a database license is responsible for a lot of these gray hairs. +[767.04 --> 772.04] Again, sorry, podcast listeners, you should watch it live next time. And similarly with AI, +[772.86 --> 777.40] we don't know how some of this is going to play out. We can offer guesses. And frankly, +[777.48 --> 781.56] they're very interesting guesses. It's a very fun theoretical game, but we don't know how it's going +[781.56 --> 785.56] to play out in court. And by the way, Chris, I think the other meta thing that's really +[785.56 --> 793.34] interesting for programmers, I often like to remind developers that writing a contract is like +[793.34 --> 798.52] writing code and then not executing it at all. And you just sort of are reading it. And like, +[798.58 --> 803.54] we all agree more or less on what the output should be. But until it's actually been executed, +[803.54 --> 809.46] which happens through a court, right? A court says, this is what the thing is, then we don't +[809.46 --> 815.76] actually know what the outcome is. So everything that I say here is going to be based on like a +[815.76 --> 820.32] handful of things, right? Part of the challenge in Oracle v. Google is that we'd all been operating +[820.32 --> 826.00] under these assumptions about what copyright law was, but there had been no litigation over what, +[826.36 --> 832.04] whether or not an API could be copyrightable since the 80s. So the closest analogy we had was +[832.04 --> 839.16] Lotus 1, 2, 3 dropdown menus. And like, it turns out things have changed a lot since then. +[839.16 --> 846.38] But because it hadn't been executed by a court, we really didn't know how this was going to turn +[846.38 --> 850.60] out. And so that makes predictions. This is why lawyers' favorite phrase is, it depends. +[851.00 --> 855.74] There's one thing I want to stick a pin in before we move on, Angelica. And it's that idea of AI. +[856.64 --> 860.26] And I also want to call out code generation. And I really want to talk about that later, +[860.30 --> 864.14] because I think that's also like a very interesting thing when you, the first thing you said was like, +[864.20 --> 868.82] oh, who's typing the code at the end of the day? And that's how copyright is generated. So I just want to +[868.82 --> 871.38] make sure we circle back to that later. Absolutely. +[871.84 --> 874.66] I also want, for those of us who weren't following that Google +[874.66 --> 881.26] law case step by step, every step of the way, once it had been litigated, what was the conclusion? +[882.08 --> 888.88] Well, so to take a step real quick back just to the very beginning, Oracle, well, Sun had created Java. +[888.88 --> 897.84] And originally the Apache project sort of funded by IBM had re-implemented Java, complete clean room, +[898.38 --> 902.96] very strict, very effective clean room, as best as we could tell from the pieces afterwards, +[903.24 --> 908.72] literally just a handful of lines, probably out of several hundred thousand in that re-implementation +[908.72 --> 913.44] that ended up looking like they were actually perhaps copied and pasted rather than +[913.44 --> 921.84] a true clean room re-implementation. And so Google used that Apache re-implementation in their Android +[921.84 --> 930.38] phones. And ultimately what happened after literally a decade of litigation, I think I was on my fifth +[930.38 --> 936.64] job by the time the case ended, like from when the case, actually six jobs from when the case started +[936.64 --> 943.80] to when the case ended. The case started with essentially Oracle claiming, there's some other +[943.80 --> 949.10] stuff I'm going to leave out for simplicity, but Oracle claimed that copying and re-implementing +[949.10 --> 956.04] just the API headers was a copyright violation and that therefore all of Android should have to be +[956.04 --> 960.76] licensed from Oracle for, they originally asked for 5 billion. I think by the end of the case, +[960.80 --> 964.26] they were asking for 9, 15 billion, something like that. +[964.26 --> 969.74] The courts found essentially through a series of rulings over the years in this case, +[970.04 --> 980.14] that an API could be copyrightable, independent of the implementation, but there was a plausible, +[980.36 --> 987.32] what we in the US call a fair use argument that essentially if you re-implement in a way that's +[987.32 --> 992.16] particularly transformative, right? Like you're doing something that is really different than what +[992.16 --> 1000.10] the original authors or copyright owners of the API intended to do with the API, then you have an +[1000.10 --> 1005.32] argument that it's okay to reuse that API in that way through re-implementation. +[1006.30 --> 1011.60] Lawyers like to say that fair use is simply the right to get sued. It's ambiguous. It's one of these +[1011.60 --> 1016.58] things where, again, you can't know ahead of time what the outcome is going to be. And that, +[1016.58 --> 1022.04] of course, makes it a playground mostly for large companies, unfortunately, right? So I think in +[1022.04 --> 1026.46] some ways that was not a great outcome. It was a better outcome for open source than what could +[1026.46 --> 1030.72] have been, than what Oracle wanted to have, but it wasn't an ideal outcome. +[1030.72 --> 1048.54] Hey friends, this episode is brought to you by my friends and potentially your friends too at +[1048.54 --> 1053.80] Firehydrant. And I'm here with Robert Ross, founder and CEO of Firehydrant. And Robert, +[1053.96 --> 1059.32] there are several options out there for incident management, but what is it that makes Firehydrant +[1059.32 --> 1064.20] different? The reason that we think that Firehydrant is, is onto something is because we're +[1064.20 --> 1070.28] meeting companies really where they are. We face the same problems that every company in the industry +[1070.28 --> 1076.26] that is building and releasing software is also facing. So where you want people to be able to sign +[1076.26 --> 1082.64] up for Firehydrant and immediately be able to kick off an incident using the best practices that we've +[1082.64 --> 1087.18] built and we've experienced and have gathered through the other amazing customers that use our tool. +[1087.18 --> 1092.50] It really is a very quick time to value. And we want people to have a long jump from where they +[1092.50 --> 1099.16] are to where they want to be in incident management. I love it. Thank you, Robert. Small teams up to 10 +[1099.16 --> 1103.86] people can get started for free with all Firehydrant features included. There's no credit card required +[1103.86 --> 1110.06] to sign up. They are making it too easy to get started. So check them out at firehydrant.com. Again, +[1110.06 --> 1112.06] firehydrant.com. +[1112.06 --> 1114.92] So +[1114.92 --> 1121.76] do you want to ask? One step back. And when we talk about code ownership, what exactly does it +[1121.76 --> 1142.28] I want to ask one step back and when we talk about code ownership what exactly does it mean +[1142.28 --> 1147.82] I own the code whether I am an individual a company or anything does it mean I'm allowed +[1147.82 --> 1153.58] to make money of it does it mean I can print it and hang it at home does it mean something else +[1153.58 --> 1160.70] well I'm gonna give you my lawyer answer to that those of you who whose github accounts do things +[1160.70 --> 1165.10] other than commit to other licenses which is pretty much all I do these days with my github +[1165.10 --> 1170.48] account will have better notions of code ownership as a cultural practice among programmers right like +[1170.48 --> 1174.38] who's responsible I do want to talk a little bit about that one but let me put a pin in that come +[1174.38 --> 1181.98] back to it the basic system since the at least the 60s in the US I'm not sure exactly the timeline +[1181.98 --> 1189.64] in the EU but I would imagine similar is that well actually let me go back even further copyright is +[1189.64 --> 1196.88] intended to protect creative works so what do you have to do to get copyright in a thing and I'll +[1196.88 --> 1201.56] explain what copyright is in a second but let me start with what it what you have to do and what +[1201.56 --> 1206.34] you have to do is you have to write down something that's creative write down can be broad right it +[1206.34 --> 1212.56] can be sculpting or but you have to take it out of your head and put it out into the real world in +[1212.56 --> 1218.38] some way that can be typing it in a computer can be like I said sculpting it into a sculpture sculptures +[1218.38 --> 1224.22] can get copyright it can be a work of art you know so it can be oil painting or whatever it can be a +[1224.22 --> 1229.86] vim poster I mean I'm a honestly these days my development environment is word but I used to be an +[1229.86 --> 1237.22] emacs guy so that is the key thing is you are doing a creative thing and can be mediated by tools and +[1237.22 --> 1244.38] Chris this gets to your point about the you know AI and where is copyright in there you know it can be +[1244.38 --> 1250.18] mediated by a typewriter or a paintbrush or I believe that we don't really know for certain yet it can be +[1250.18 --> 1257.44] mediated by an AI but you are doing some creative something and turning that into a fixed thing all +[1257.44 --> 1261.30] right so what happens once you've done that actually before I get into what happens once you've done that +[1261.30 --> 1268.92] is I think there's an important exception that's in the US at least that creative like what does it mean to be +[1268.92 --> 1274.98] creative is not zero it's pretty close to zero but it's not zero there's an important case called feist +[1274.98 --> 1281.26] versus rural telephone and the holding of that case is literally telephone books aren't creative and so therefore +[1281.26 --> 1285.64] they don't get copyright because what's the point of a telephone book the point of a telephone book is to +[1285.64 --> 1290.60] literally just mechanically go through a town and have phone numbers for everybody so it's not +[1290.60 --> 1296.94] it's hard work but it's not creative and in the US at least you have to have some kind of creative +[1296.94 --> 1302.24] something so if you do like a phone list of the hundred most awesome people in New York City +[1302.24 --> 1309.38] that's creative right you had to select one of the ways which you can be creative under US copyright law +[1309.38 --> 1315.48] is selection if you pick those hundred people then hey you've done something creative your list of +[1315.48 --> 1320.94] hundred people is copyrightable but if you're just every single person who lives in Manhattan that's +[1320.94 --> 1328.78] not creative you don't get protection and that plays into questions of databases and ultimately I +[1328.78 --> 1332.98] think and we might not have time to get time to get to this today but the question of the models +[1332.98 --> 1339.00] themselves right because there's both the output of models what's the copyright on that and the +[1339.00 --> 1343.76] models themselves we don't actually know if they're copyrightable that may be too esoteric might have +[1343.76 --> 1347.92] to invite me back for another one for that you know but okay so you've you've created this thing +[1347.92 --> 1352.66] so now what do you do so now you've got copyright what does copyright let you do copyright lets you +[1352.66 --> 1358.54] control what others can do with it right it lets you decide who gets to use it who gets to share who +[1358.54 --> 1365.04] gets to redistribute it who gets to modify it within certain limits but it's pretty strong right so the +[1365.04 --> 1371.10] limits include what's called first sale doctrine which is uh hey I sold it to somebody they can +[1371.10 --> 1377.08] usually sell it to one other person first sale doctrine made a lot more sense in the era of like +[1377.08 --> 1382.90] books and like that's what creates used bookstores is for sale doctrine it means that I bought the +[1382.90 --> 1389.72] copyrighted thing and now I can give it to a used bookstore and they can resell it um in the digital age +[1389.72 --> 1395.18] for sale doctrine is a little more complicated um but suffice to say like that's one of the +[1395.18 --> 1402.38] limitations similarly fair use says hey if you're using this for education if you're using this for +[1402.38 --> 1407.40] non-profit purposes I'm oversimplifying a little bit here the tests around fair use can be a little +[1407.40 --> 1415.10] complicated critically in our digital age fair use in the U.S. has expanded quite a bit to include +[1415.10 --> 1421.00] what's called transformative use which is to say hey you're doing something super new super different +[1421.00 --> 1428.02] courts are often going to allow that in the name of sort of not impeding progress so for example +[1428.02 --> 1434.80] google book search is in some sense like the biggest copyright violation in all of history right +[1434.80 --> 1441.04] because it's literally copied systematically millions of books made these digital copies but then +[1441.04 --> 1447.82] a court said well but actually it's so different it's so great and they put strict controls around +[1447.82 --> 1454.08] you know you can only get a few pages at a time and authors can opt out if they want after the copying +[1454.08 --> 1460.80] has been done so like google book search is a good example of what transformation means and potentially +[1460.80 --> 1466.68] analogous to what copilot is doing right but we don't know I mean the flip side of this right is that +[1466.68 --> 1472.32] we just had court cases we had a court case a couple years ago about the song blurred lines some of you +[1472.32 --> 1479.10] might have heard right and courts there has actually said that even just sort of copying the style of +[1479.10 --> 1486.22] the artist could potentially be a copyright infringement which was a big surprise to a lot of a lot of lawyers +[1486.22 --> 1492.34] a lot of lawyers still unhappy about that case next week there's going to be or no tomorrow morning +[1492.34 --> 1498.72] actually maybe there's going to be a case about andy warhol doing and a photograph of prince that +[1498.72 --> 1504.64] andy warhol transformed into one of his andy warhol canvases and the supreme court like it's a little +[1504.64 --> 1508.90] weird but I think that case might actually have a lot of impact on artificial intelligence because +[1508.90 --> 1514.52] we've all done we've all played with stable diffusion or mid journey or open ai or whatever +[1514.52 --> 1521.84] to create foo in the style of bar right well if bar is still alive and still has a valid copyright +[1521.84 --> 1527.60] maybe that's a problem we don't really know yet I saw a research paper yesterday that said +[1527.60 --> 1535.88] if you prompt copilot to do code in the style of forgetting the guy's name petrov I think a top +[1535.88 --> 1542.74] python programmer that you actually get fewer vulnerabilities in your code if you prompt copilot +[1542.74 --> 1550.50] with the name of a top maintainer and flip side the the paper's author was honest enough to note that +[1550.50 --> 1555.40] they prompted with their own name and the number of vulnerabilities went up which I thought was +[1555.40 --> 1561.82] nice and humble of them so style is an issue that's gonna that could potentially come up in code as well +[1561.82 --> 1566.92] that was a very long-winded answer to your question Natalie I apologize no that was interesting so you +[1566.92 --> 1573.48] said that for code ownership basically means who is allowed to sell and profit of that who is allowed to +[1573.48 --> 1585.92] give it their own personal interpretation yeah it's also who's who's there to answer in case of a problem right I wrote a piece of code that made my work lose a lot of money +[1585.92 --> 1600.48] ownership on me yeah I mean so that's where it gets complicated and we have really good answers for that in the case of things like if you manufactured a car wheel and the car +[1600.48 --> 1612.92] wheel explodes because you used bad materials in that case in the car wheel then we have some well-developed laws and intuitions around like okay well we sue the car wheel manufacturer +[1612.92 --> 1624.50] software doesn't have any of that really yet we've sort of operated in this rules-free zone where everybody was like software so cool I guess we'll just let it happen +[1624.50 --> 1643.16] and um I think that age is coming to an end to be perfectly honest actually I think the European Union has published in the past year including one last week two weeks ago papers on liability for software right I the idea of +[1643.16 --> 1654.32] if a car wheel explodes and causes a car accident we have a very clear idea of how we should figure out who is liable for that if an AI goes wrong or +[1654.32 --> 1666.34] software goes wrong and the car goes off the road and causes the exact same accident we actually have very little idea how we should apportion liability right and it's not necessarily about +[1666.34 --> 1676.80] oh like that is for practical reasons those kinds of things tend to look the same right because at the end of the day the company that commissioned the code is also the company that sold the product +[1676.80 --> 1688.54] so you tend to see those things tied together but there's no formal reason for that right like copyright law doesn't especially because copyright law historically was about like things that didn't cause car crashes +[1688.54 --> 1704.42] right historically copyright law like literally modern copyright law in the US is in large part because of player pianos like scrolling wheels like what you saw in Westworld like scrolling wheels with little punch holes that caused the piano to do things like pianos didn't run off and like kill people +[1704.42 --> 1715.46] so copyright law doesn't really have much to say inherently about product liability and that's something that we are I think screaming towards a very high velocity +[1715.46 --> 1721.00] at least in the EU and I suspect because of AI in the US soon as well +[1721.00 --> 1728.40] I feel like there's an interesting component to that as well because when you think about what we create we're just creating words on a page like +[1728.40 --> 1737.04] the manufacturing process of turning that into something that does something is not necessarily something that the person who wrote the code does +[1737.04 --> 1744.90] so it's like something that somebody else does and then there's all of the like well the machine you run it on like if a bug does happen with like a car that's driving +[1744.90 --> 1751.82] is it the fault of the person who wrote the code is it the fault of the machine is there a problem with the machine who like who gets the blame and I think that gets +[1751.82 --> 1758.84] extremely murky because we're dealing with such like new stuff that we have never had in like the existence of humanity +[1758.84 --> 1770.04] yeah I mean the original history of this in like the English and US law systems was that literally like to get to that point where hey the car wheel +[1770.04 --> 1778.88] explodes we should sue the car manufacturer involved a lot of people dying in train accidents and the train companies being like oh but that's not +[1778.88 --> 1803.66] our fault we just laid the tracks bought the train bought the coal it's the guy who was driving it's their fault so you can't sue us and that actually like as a matter of law was like a good argument for decades and then the number of accidents as trains became ever more present as part of our economy as part of simply how things moved around well the technological change +[1803.66 --> 1811.52] drove a change drove a change in understanding right because that rule was originally from like hey some dude on a horse +[1811.52 --> 1824.94] it's not my fault if he's like my my squire or whatever pick your ancient British legal term you know if they're out on my horse like genuinely it's sort of not my problem if they like caused an accident +[1824.94 --> 1831.70] you know I mean yeah I own the horse but like and so the train companies for a long time were like well look it's just like a horse +[1831.70 --> 1838.32] this train is just like a horse I can't be responsible and so at some point the legal system was like actually this is ludicrous +[1838.32 --> 1844.32] and so a combination of courts and congress changed the rules to make the train companies more liable +[1844.32 --> 1853.50] no surprise the trains then started getting safer yeah Angelico I've actually got a British person on this podcast I believe so why am I not asking you the proper terms here +[1853.50 --> 1855.32] yeah sometimes the footman's fault +[1855.32 --> 1862.04] well that's exactly it right it's the footman's fault and so I think we're seeing we are is both an exciting and a terrifying time +[1862.04 --> 1868.62] for lawyers that we are in the midst of one of these very rare technological changes right we are +[1868.62 --> 1874.94] AI I think in particular is going to be that new train nobody really understands nobody wants to take responsibility +[1875.52 --> 1882.32] for Chris as you say like super good reasons this stuff is literally the most complicated systems ever built by humankind +[1882.32 --> 1886.26] we like even in the best case have only the vaguest sense of how it works +[1886.26 --> 1890.70] and like good luck explaining it to a judge +[1890.70 --> 1894.00] I had a conversation with some lawyer friends last week that was like +[1894.00 --> 1898.32] how would you explain and these are like fairly sophisticated +[1898.32 --> 1902.74] you know like most either are programmers or one of the people is like married to a programmer +[1902.74 --> 1906.88] like we're fairly and we've been doing tech law all for cumulatively many decades +[1906.88 --> 1909.32] how would you explain machine learning to a judge +[1909.32 --> 1914.60] and we all just sort of stuttered in horror at that thought right because it's a really +[1914.60 --> 1918.00] again I mean even to programmers who haven't thought about it +[1918.00 --> 1920.12] it can be really hard unintuitive +[1920.12 --> 1924.50] the vocabulary is changing all the time the technology is changing all the time +[1924.50 --> 1926.88] and to try to explain it to congress +[1926.88 --> 1929.70] or to a judge is a scary proposition +[1929.70 --> 1932.00] it's an exciting whoever gets to do it first +[1932.00 --> 1935.20] that's gonna be a super great lawyer job for somebody +[1935.20 --> 1938.88] but also like boy when you screw it up +[1938.88 --> 1941.74] and we felt a lot of pressure in the google oracle trial right +[1941.74 --> 1945.64] that this was something that if we got it wrong it would really hurt open source +[1945.64 --> 1949.32] and I suspect every good lawyer of course cares about their client +[1949.32 --> 1953.50] but some clients are just represent one client +[1953.50 --> 1956.34] and other clients represent these big systemic changes +[1956.34 --> 1958.42] and you feel that weight as a lawyer +[1958.42 --> 1961.82] I feel like there's a other side of the problem as well +[1961.82 --> 1967.60] because it's like if you try and assign blame to the person who like owns the copyright of code +[1967.60 --> 1970.94] there is a huge amount of code that we all depend on all the time +[1970.94 --> 1973.58] that's maintained by like some random dude in Nebraska +[1973.58 --> 1975.30] like individual people +[1975.30 --> 1977.62] and it's like well if you can sue to get to them +[1977.62 --> 1980.40] because something they wrote caused some problem somewhere down the chain +[1980.40 --> 1982.70] then that's obviously a problem +[1982.70 --> 1985.36] because I can kind of see the chain where it's just like well +[1985.36 --> 1989.32] who actually gets the blame for the bug that was written +[1989.32 --> 1990.80] or the problem that happened with the code +[1990.80 --> 1993.38] because you can easily just like keep tracing that back +[1993.38 --> 1995.30] further and further and further and further +[1995.30 --> 1997.12] by like passing on the blanks +[1997.12 --> 1998.66] it's like once again with the train it's like +[1998.66 --> 2000.80] well I didn't create that wheel +[2000.80 --> 2002.02] someone else created that wheel +[2002.02 --> 2003.36] so that's their problem +[2003.36 --> 2006.40] or like with the Spectre meltdown hardware problems +[2006.40 --> 2009.40] where it's like oh well it's not my fault there was a breach +[2009.40 --> 2011.92] the processor shouldn't have been speculatively executing +[2011.92 --> 2014.72] like there's so many like weird arguments that you have +[2014.72 --> 2017.98] because of this stuff that we don't really understand what it is right now +[2017.98 --> 2023.68] yeah and I think applying the old models is probably going to get us some very bad outcomes +[2023.68 --> 2028.68] and unfortunately the way the legal system learns sometimes is by having bad outcomes +[2028.68 --> 2030.56] everybody stubs their toe on it +[2030.56 --> 2033.80] and then you sort of fix that up as we go +[2033.80 --> 2036.06] but some people end up being caught in the middle +[2036.06 --> 2037.36] that guy in Nebraska +[2037.36 --> 2042.88] I assume all the listeners here have seen the XKCD comic about the guy in Nebraska +[2042.88 --> 2045.98] the problem is of course it's not even just one guy in Nebraska +[2045.98 --> 2048.28] it is a tower of 10,000 guys in Nebraska +[2048.28 --> 2052.44] and so you know I do want to talk a little bit about the day job here +[2052.44 --> 2055.96] because I founded co-founded a company called Tidelift +[2055.96 --> 2059.52] and Tidelift's mission as we said at the top of the show +[2059.52 --> 2061.38] is to make open source better for everyone +[2061.38 --> 2063.42] in part by paying the maintainers +[2063.42 --> 2065.58] because what we're seeing happen all the time +[2065.58 --> 2067.44] we saw it happen a couple times this week +[2067.44 --> 2068.80] just in the JavaScript community +[2068.80 --> 2074.10] is the solution to this kind of problem that you've identified Chris so far +[2074.10 --> 2077.64] is we'll just start applying standards +[2077.64 --> 2081.96] right so like we've got the open SSF standard security scorecard +[2081.96 --> 2083.82] we've got salsa.dev +[2083.82 --> 2086.60] which is a different kind of security scorecard +[2086.60 --> 2090.16] GitHub caused some controversy by saying +[2090.16 --> 2091.98] hey we've identified the most popular +[2091.98 --> 2094.62] I think only NPM for right now +[2094.62 --> 2095.52] projects +[2095.52 --> 2099.10] and we're sending you all a free two-factor authentication key +[2099.10 --> 2101.10] and also we're like +[2101.10 --> 2103.90] mandatorily turning on two-factor authentication for everybody +[2103.90 --> 2106.60] and a couple maintainers for various reasons +[2106.60 --> 2107.20] were just like +[2107.20 --> 2108.48] that's too much work +[2108.48 --> 2110.58] that's going to complicate my life +[2110.58 --> 2111.94] it's going to break my build scripts +[2111.94 --> 2113.04] and like +[2113.04 --> 2114.88] we can go back and forth about like +[2114.88 --> 2115.66] whether or not +[2115.66 --> 2117.16] two-factor authentication +[2117.16 --> 2119.30] in some of these cases is a good idea +[2119.30 --> 2121.32] but I want to step out +[2121.32 --> 2121.56] I mean +[2121.56 --> 2124.42] I generally think two-factor authentication is a good idea +[2124.42 --> 2125.08] don't get me wrong +[2125.08 --> 2127.04] but like that's the easy case +[2127.04 --> 2128.68] it just gets harder from there +[2128.68 --> 2129.72] right like okay +[2129.72 --> 2132.26] well what do we need to do to sign our binaries +[2132.26 --> 2133.08] you know +[2133.08 --> 2134.92] go I understand that's a mostly solved +[2134.92 --> 2137.50] probably like signing modules is a mostly solved problem +[2137.50 --> 2139.66] and a lot of other language ecosystems it is not +[2139.66 --> 2141.36] so okay +[2141.36 --> 2142.38] so there's extra work +[2142.38 --> 2142.86] right +[2142.86 --> 2144.40] and we've just created this extra work +[2144.40 --> 2145.14] Chris as you say +[2145.14 --> 2146.76] on some guy in Nebraska +[2146.76 --> 2149.02] or actually a stack of 10,000 some guys +[2149.02 --> 2151.10] and I apologize to listeners +[2151.10 --> 2153.16] it's probably grating for me to hear +[2153.16 --> 2154.76] for me to say +[2154.76 --> 2155.82] 10,000 guys +[2155.82 --> 2157.22] but I think it's worth +[2157.22 --> 2159.56] both admitting that this is a problem +[2159.56 --> 2160.66] and I think saying that this +[2160.66 --> 2161.76] the part of the problem +[2161.76 --> 2163.20] of the gendering of open source +[2163.20 --> 2164.42] is very much +[2164.42 --> 2165.96] that in a world +[2165.96 --> 2167.58] where women are often called on +[2167.58 --> 2170.36] to take on more than their share of household duties +[2170.36 --> 2172.06] and home care +[2172.06 --> 2172.68] child care +[2172.68 --> 2173.34] elder care +[2173.34 --> 2176.20] if we're not paying people to do open source +[2176.20 --> 2176.84] well guess what +[2176.84 --> 2178.66] that's part of how we get guys doing it +[2178.66 --> 2180.82] because guys for various cultured reasons +[2180.82 --> 2181.90] have more free time +[2181.90 --> 2183.92] that's an important side note +[2183.92 --> 2184.84] that I try to +[2184.84 --> 2186.46] I think is important to say +[2186.46 --> 2187.72] anyway +[2187.72 --> 2189.62] we're putting all these new requirements +[2189.62 --> 2190.36] on people Chris +[2190.36 --> 2191.96] because of exactly this intuition +[2191.96 --> 2192.94] you've had about like +[2192.94 --> 2194.20] but we're doing nothing +[2194.20 --> 2195.90] to make open source more fun +[2195.90 --> 2196.68] easier +[2196.68 --> 2199.26] like all we're doing is loading more work on top +[2199.26 --> 2200.34] and I think at some point +[2200.34 --> 2202.86] I think we're already starting to see it +[2202.86 --> 2203.70] in some communities +[2203.70 --> 2205.64] that people are going to snap +[2205.64 --> 2206.50] people are going to walk away +[2206.50 --> 2207.38] people are going to say +[2207.38 --> 2208.52] so what do we do +[2208.52 --> 2209.64] so what Tidelift does +[2209.64 --> 2210.96] to help address this problem +[2210.96 --> 2212.30] is we say +[2212.30 --> 2212.62] hey +[2212.62 --> 2214.26] we go to our customers +[2214.26 --> 2214.74] and we say +[2214.74 --> 2216.02] you're going to get more predictable +[2216.02 --> 2217.52] more reliable open source +[2217.52 --> 2219.42] if developers follow these standards +[2219.42 --> 2221.14] they're not going to follow these standards +[2221.14 --> 2221.66] on their own +[2221.66 --> 2222.60] you should pay them +[2222.60 --> 2223.72] so if you want +[2223.72 --> 2225.20] the stuff you use +[2225.20 --> 2226.80] to follow those standards +[2226.80 --> 2227.78] write us a check +[2227.78 --> 2228.70] we will go out +[2228.70 --> 2230.32] find those developers for you +[2230.32 --> 2232.12] hopefully we already have a contract with them +[2232.12 --> 2232.86] from other customers +[2232.86 --> 2234.06] but we'll go out +[2234.06 --> 2234.80] find them +[2234.80 --> 2235.82] and pay them +[2235.82 --> 2238.62] and in order to get some of these +[2238.62 --> 2239.72] this work done +[2239.72 --> 2240.32] now +[2240.32 --> 2243.26] there's a lot of challenges around this +[2243.26 --> 2243.54] right +[2243.54 --> 2244.00] in part +[2244.00 --> 2244.78] because guess what +[2244.78 --> 2246.28] nobody wants to pay for open source +[2246.28 --> 2247.44] it should be free +[2247.44 --> 2247.98] and like +[2247.98 --> 2248.70] well guess what +[2248.70 --> 2249.94] if you're liable +[2249.94 --> 2250.62] all of a sudden +[2250.62 --> 2251.52] maybe it's not free +[2251.52 --> 2252.98] and I think one of the interesting things +[2252.98 --> 2253.98] that we're going to see in discussion +[2253.98 --> 2254.44] about +[2254.44 --> 2256.24] these EU regulations +[2256.24 --> 2256.90] for example +[2256.90 --> 2259.50] they contain exceptions for open source +[2259.50 --> 2261.96] their definition of open source +[2261.96 --> 2263.24] sort of looks like +[2263.24 --> 2265.38] it excludes commercially sponsored open source +[2265.38 --> 2266.78] which as we know +[2266.78 --> 2268.30] is a lot of open source these days +[2268.30 --> 2268.84] right +[2268.84 --> 2270.96] we don't know how that's going to play out +[2270.96 --> 2272.46] we don't know what it really means +[2272.46 --> 2273.94] like their definition is vague enough +[2273.94 --> 2274.46] that like +[2274.46 --> 2276.04] maybe it only includes +[2276.04 --> 2277.14] a small slice of open source +[2277.14 --> 2278.50] or maybe it includes a lot of open source +[2278.50 --> 2279.54] we don't really know yet +[2279.54 --> 2281.66] I'm sure that's going to be lobbied over +[2281.66 --> 2283.14] and in fact +[2283.14 --> 2283.96] I'm going to publish something +[2283.96 --> 2284.34] hopefully +[2284.34 --> 2285.86] tomorrow or Thursday +[2285.86 --> 2287.14] on dev.2 +[2287.14 --> 2287.76] about +[2287.76 --> 2288.78] what +[2288.78 --> 2290.18] open source developers +[2290.18 --> 2290.94] can do +[2290.94 --> 2291.64] to help +[2291.64 --> 2293.18] lobby the US government +[2293.18 --> 2293.82] on this topic +[2293.82 --> 2294.64] but a lot of the same thing +[2294.64 --> 2295.08] is going to apply +[2295.08 --> 2296.36] to the EU government as well +[2296.36 --> 2297.66] yeah I think +[2297.66 --> 2298.94] one of the thoughts I had +[2298.94 --> 2300.36] during what you were saying +[2300.36 --> 2300.82] is +[2300.82 --> 2302.62] I've expressed this in private +[2302.62 --> 2303.24] to some people +[2303.24 --> 2304.54] and I always get kind of the +[2304.54 --> 2304.86] like +[2304.86 --> 2306.04] you've just said heresy +[2306.04 --> 2306.92] look +[2306.92 --> 2308.00] or comments +[2308.00 --> 2309.44] but I have been wondering +[2309.44 --> 2309.70] like +[2309.70 --> 2311.16] is open source +[2311.16 --> 2312.22] sustainable +[2312.22 --> 2313.26] as the method +[2313.26 --> 2314.76] of how we do things +[2314.76 --> 2315.78] in this industry +[2315.78 --> 2316.02] like +[2316.02 --> 2316.74] is this like +[2316.74 --> 2318.06] focus on sharing +[2318.06 --> 2319.10] so much of code +[2319.10 --> 2319.56] like +[2319.56 --> 2320.88] actually going to be +[2320.88 --> 2321.58] something we can continue +[2321.58 --> 2322.24] doing in the future +[2322.24 --> 2323.50] since there's so much +[2323.50 --> 2324.90] ambiguity around all of this +[2324.90 --> 2325.88] and quite frankly +[2325.88 --> 2326.56] I think it also +[2326.56 --> 2327.04] just like +[2327.04 --> 2328.06] atrophies +[2328.06 --> 2329.62] the whole industry +[2329.62 --> 2330.36] because we're not +[2330.36 --> 2331.04] rewriting things +[2331.04 --> 2332.30] we're not reimagining things +[2332.30 --> 2333.46] like I think that's one of +[2333.46 --> 2334.24] the core problems +[2334.24 --> 2335.12] with copyright in general +[2335.12 --> 2335.34] right +[2335.34 --> 2335.96] there's this whole thing +[2335.96 --> 2336.68] that Disney has done +[2336.68 --> 2337.36] where it's just like +[2337.36 --> 2338.42] copyright used to be +[2338.42 --> 2339.80] like 20 years +[2339.80 --> 2340.76] now it's like +[2340.76 --> 2342.20] almost forever +[2342.20 --> 2343.56] or as Disney would like to have +[2343.56 --> 2343.66] it +[2343.66 --> 2344.92] actually forever +[2344.92 --> 2346.02] so it's like +[2346.02 --> 2346.96] oh well now these things +[2346.96 --> 2347.68] are just kind of +[2347.68 --> 2348.48] sticking around +[2348.48 --> 2349.44] and it's so much harder +[2349.44 --> 2350.66] to like move things forward +[2350.66 --> 2350.88] right +[2350.88 --> 2351.42] it was like +[2351.42 --> 2352.70] if I remember the +[2352.70 --> 2354.10] kind of genesis of copyright +[2354.10 --> 2355.16] or the vague genesis +[2355.16 --> 2355.64] it's like +[2355.64 --> 2357.08] we want to protect people +[2357.08 --> 2358.70] like make it so they can profit +[2358.70 --> 2359.82] off of their creative work +[2359.82 --> 2360.68] for some time +[2360.68 --> 2362.00] but then it goes back +[2362.00 --> 2363.30] into the general pool of things +[2363.30 --> 2363.96] so we can kind of +[2363.96 --> 2365.46] continue making progress forward +[2365.46 --> 2366.34] yeah boy +[2366.34 --> 2367.44] that sustainability question +[2367.44 --> 2368.20] is a big one Chris +[2368.20 --> 2369.10] and I really don't know +[2369.10 --> 2370.48] I'd like to say that +[2370.48 --> 2372.32] we have a real clear answer to it +[2372.32 --> 2372.78] I mean certainly +[2372.78 --> 2373.64] I think that Tidelift +[2373.64 --> 2375.42] is part of the answer to that +[2375.42 --> 2377.32] but I think it's a really good question +[2377.32 --> 2377.90] to be asking +[2377.90 --> 2378.28] and I +[2378.28 --> 2379.76] that heresy +[2379.76 --> 2382.02] it is an elephant in the room +[2382.02 --> 2383.50] that a lot of people +[2383.50 --> 2384.78] I gotta say +[2384.78 --> 2385.72] I get a little frustrated +[2385.72 --> 2386.62] when an employee +[2386.62 --> 2387.82] of a trillion dollar company +[2387.82 --> 2388.30] is like +[2388.30 --> 2389.34] I don't know +[2389.34 --> 2390.32] paying people +[2390.32 --> 2392.54] open source seems to work fine for me +[2392.54 --> 2392.98] I'm like +[2392.98 --> 2393.88] well yeah +[2393.88 --> 2395.56] because you literally work +[2395.56 --> 2396.80] for a trillion dollar company +[2396.80 --> 2398.52] but a lot of the software +[2398.52 --> 2399.52] you rely on +[2399.52 --> 2400.50] and you're certainly +[2400.50 --> 2402.04] that of your customers rely on +[2402.04 --> 2403.90] they don't have that luxury +[2403.90 --> 2404.44] right +[2404.44 --> 2406.18] it is the kind of thing +[2406.18 --> 2407.14] you know +[2407.14 --> 2407.92] you got a puppy +[2407.92 --> 2409.48] puppies are often more fun +[2409.48 --> 2411.08] than replying to pull requests +[2411.08 --> 2412.14] from automated bots +[2412.14 --> 2413.26] and Chris +[2413.26 --> 2414.04] I think it's +[2414.04 --> 2415.36] I think one of the interesting things +[2415.36 --> 2416.92] we've been a little backward focused +[2416.92 --> 2418.66] but I think there's a lot of cool stuff +[2418.66 --> 2419.22] you know +[2419.22 --> 2421.10] I know we're flying through this time +[2421.10 --> 2423.56] I think there's a lot of interesting questions +[2423.56 --> 2424.08] about this +[2424.08 --> 2425.20] you know future looking +[2425.20 --> 2427.14] around machine learning +[2427.14 --> 2427.82] co-pilot +[2427.82 --> 2428.38] things like that +[2428.38 --> 2429.02] and part of those +[2429.02 --> 2430.26] go to what you're saying +[2430.26 --> 2431.10] you know Chris +[2431.10 --> 2431.48] you were saying +[2431.48 --> 2432.96] well the original motivation +[2432.96 --> 2434.60] for copyright was to +[2434.60 --> 2436.44] well there's sort of three +[2436.44 --> 2438.54] original motivations for copyright +[2438.54 --> 2439.70] that vary depending on +[2439.70 --> 2440.46] who you're talking to +[2440.46 --> 2441.38] so in the US +[2441.38 --> 2442.58] in the constitution +[2442.58 --> 2444.00] it says that the +[2444.00 --> 2445.20] purpose of copyright +[2445.20 --> 2447.02] is to encourage authors +[2447.02 --> 2447.74] right +[2447.74 --> 2449.66] so it is a very utilitarian +[2449.66 --> 2450.24] like +[2450.24 --> 2451.74] we're going to give you this copyright +[2451.74 --> 2452.58] and as a result +[2452.58 --> 2453.16] you're going to like +[2453.16 --> 2454.08] create more stuff +[2454.08 --> 2454.52] and every +[2454.52 --> 2455.28] that's a bargain +[2455.28 --> 2455.96] that we're going to have +[2455.96 --> 2456.16] right +[2456.16 --> 2457.22] like we're going to give you +[2457.22 --> 2457.82] this monopoly +[2457.82 --> 2458.40] which otherwise +[2458.40 --> 2458.90] the founders +[2458.90 --> 2460.40] super against monopolies +[2460.40 --> 2461.38] but they created +[2461.38 --> 2462.76] the copyright monopoly +[2462.76 --> 2463.96] specifically in order +[2463.96 --> 2465.24] for the rest of us +[2465.24 --> 2465.64] to benefit +[2465.64 --> 2466.82] from this incentive +[2466.82 --> 2467.98] of a whole bunch of stuff +[2467.98 --> 2468.52] being created +[2468.52 --> 2469.60] so that's one story +[2469.60 --> 2470.90] in the EU +[2470.90 --> 2472.20] really actually +[2472.20 --> 2473.08] most of the rest of the world +[2473.08 --> 2473.60] except the US +[2473.60 --> 2475.66] it's more like +[2475.66 --> 2476.76] your creativity +[2476.76 --> 2478.14] is like a part of you +[2478.14 --> 2479.30] like it is part of your +[2479.30 --> 2480.02] like human +[2480.02 --> 2482.34] your human nature +[2482.34 --> 2483.68] is to create +[2483.68 --> 2484.46] and so +[2484.46 --> 2486.06] there's often what are called +[2486.06 --> 2487.62] moral rights +[2487.62 --> 2489.10] the idea that inherently +[2489.10 --> 2490.30] you have some control +[2490.30 --> 2491.04] over the thing +[2491.04 --> 2491.92] that you've created +[2491.92 --> 2493.84] even if it's not productive +[2493.84 --> 2494.34] right +[2494.34 --> 2495.12] even if there's no +[2495.12 --> 2496.56] social value to it +[2496.56 --> 2498.40] and we're going to be +[2498.40 --> 2499.14] running headlong +[2499.14 --> 2499.66] into that +[2499.66 --> 2500.56] with all of the +[2500.56 --> 2501.84] foo in the style +[2501.84 --> 2502.34] of bar +[2502.34 --> 2504.32] bar is going to be +[2504.32 --> 2505.14] really irritated +[2505.14 --> 2506.08] that their moral rights +[2506.08 --> 2506.76] were infringed +[2506.76 --> 2507.68] and by the way +[2507.68 --> 2508.68] the third +[2508.68 --> 2509.40] like historic +[2509.40 --> 2510.84] the original copyright +[2510.84 --> 2511.84] was literally +[2511.84 --> 2512.56] just basically +[2512.56 --> 2513.40] a tool of censorship +[2513.40 --> 2514.56] for the UK government +[2514.56 --> 2515.82] in the early 1600s +[2515.82 --> 2517.26] it was a way for them +[2517.26 --> 2518.04] to control printers +[2518.04 --> 2519.42] and +[2519.42 --> 2520.30] I think Chris +[2520.30 --> 2520.96] there's a really +[2520.96 --> 2521.62] interesting discussion +[2521.62 --> 2522.44] we're going to have +[2522.44 --> 2522.90] over +[2522.90 --> 2524.88] how does +[2524.88 --> 2525.68] open source +[2525.68 --> 2526.86] probably won't be +[2526.86 --> 2527.98] mediated through copyright +[2527.98 --> 2529.26] but as you were saying +[2529.26 --> 2529.92] about liability +[2529.92 --> 2531.20] and security +[2531.20 --> 2532.98] how does government +[2532.98 --> 2533.98] interact with this +[2533.98 --> 2534.80] because it's one thing +[2534.80 --> 2535.88] if like these big businesses +[2535.88 --> 2536.70] are running around +[2536.70 --> 2537.24] saying like +[2537.24 --> 2538.08] hey we should make +[2538.08 --> 2538.96] this stuff more secure +[2538.96 --> 2540.50] it's very different +[2540.50 --> 2541.14] if governments +[2541.14 --> 2541.82] are running around +[2541.82 --> 2542.36] saying +[2542.36 --> 2543.72] the whole world +[2543.72 --> 2544.38] needs you to be +[2544.38 --> 2545.38] more secure +[2545.38 --> 2545.80] right +[2545.80 --> 2547.00] like that's +[2547.00 --> 2547.38] you know +[2547.38 --> 2547.80] big thing +[2547.80 --> 2548.40] that we still haven't +[2548.40 --> 2548.82] really wrapped +[2548.82 --> 2549.34] our heads around +[2549.34 --> 2549.66] yeah +[2549.66 --> 2550.34] it doesn't help +[2550.34 --> 2550.86] that our +[2550.86 --> 2551.64] legislators +[2551.64 --> 2553.10] aren't +[2553.10 --> 2554.04] very tech savvy +[2554.04 --> 2554.70] and don't +[2554.70 --> 2555.64] they tend to write +[2555.64 --> 2556.34] a lot of laws +[2556.34 --> 2557.12] that you're like +[2557.12 --> 2558.60] this makes no sense +[2558.60 --> 2560.06] or ask questions +[2560.06 --> 2560.54] in hearings +[2560.54 --> 2561.06] that are +[2561.06 --> 2562.42] questionable +[2562.42 --> 2563.18] at best +[2563.18 --> 2563.76] yeah +[2563.76 --> 2564.74] there's two parts +[2564.74 --> 2565.16] of that right +[2565.16 --> 2565.98] like that's the one +[2565.98 --> 2566.54] and again +[2566.54 --> 2567.36] I'll try to be quicker +[2567.36 --> 2568.08] that's the one +[2568.08 --> 2568.86] that everybody thinks about +[2568.86 --> 2569.36] because we've seen +[2569.36 --> 2570.40] our legislators on TV +[2570.40 --> 2571.08] and it's terrifying +[2571.08 --> 2572.02] but there's also +[2572.02 --> 2572.70] this thing where +[2572.70 --> 2573.74] so legislators +[2573.74 --> 2574.66] sort of provide you +[2574.66 --> 2575.56] like a rough draft +[2575.56 --> 2575.96] right +[2575.96 --> 2577.10] and then the courts +[2577.10 --> 2578.28] are used to refine that +[2578.28 --> 2579.68] but because litigation +[2579.68 --> 2580.82] has gotten so expensive +[2580.82 --> 2582.38] and everybody hates it +[2582.38 --> 2583.84] we don't do the refining +[2583.84 --> 2584.86] of the rough drafts anymore +[2584.86 --> 2585.82] right +[2585.82 --> 2587.02] like it sort of +[2587.02 --> 2588.28] becomes industry convention +[2588.28 --> 2589.90] and that sticks us +[2589.90 --> 2590.70] with a lot of cruft +[2590.70 --> 2592.48] that I think is a problem +[2592.48 --> 2593.88] Natalie you had a question +[2593.88 --> 2594.48] you wanted to +[2594.48 --> 2596.44] I wanted to say that +[2596.44 --> 2597.72] that's like couple +[2597.72 --> 2598.84] of topics back +[2598.84 --> 2599.56] but I had a +[2599.56 --> 2600.44] I'm a contractor +[2600.44 --> 2601.86] so I have clients +[2601.86 --> 2602.40] around the world +[2602.40 --> 2603.22] I see all sorts +[2603.22 --> 2604.14] of different contracts +[2604.14 --> 2605.62] and one time +[2605.62 --> 2606.26] I had a contract +[2606.26 --> 2607.04] with a California +[2607.04 --> 2607.76] based company +[2607.76 --> 2609.10] and there was +[2609.10 --> 2610.50] a clause +[2610.50 --> 2611.20] that said that +[2611.20 --> 2612.08] any damage +[2612.08 --> 2613.32] that I cause +[2613.32 --> 2614.14] I am responsible +[2614.14 --> 2614.62] of it +[2614.62 --> 2616.14] so back to the +[2616.14 --> 2617.06] conversation about +[2617.06 --> 2617.80] that Nebraska +[2617.80 --> 2618.52] person +[2618.52 --> 2620.34] I mean contrast +[2620.34 --> 2621.26] can say that +[2621.26 --> 2622.02] the good news +[2622.02 --> 2622.84] slash bad news +[2622.84 --> 2623.46] is that you are +[2623.46 --> 2624.98] probably what is known +[2624.98 --> 2625.62] as a +[2625.62 --> 2627.72] not a deep pocket +[2627.72 --> 2628.66] right +[2628.66 --> 2629.90] like they're unlikely +[2629.90 --> 2630.74] to sue you +[2630.74 --> 2631.34] because what are you +[2631.34 --> 2632.02] going to give them +[2632.02 --> 2632.96] your collection +[2632.96 --> 2633.60] of goldies +[2633.60 --> 2635.28] would not be worth +[2635.28 --> 2635.88] much to them +[2635.88 --> 2637.12] I was really terrified +[2637.12 --> 2637.52] exactly +[2637.52 --> 2638.56] you probably should have +[2638.56 --> 2639.70] negotiated that out +[2639.70 --> 2640.80] but that's one of these +[2640.80 --> 2641.58] ways in which the legal +[2641.58 --> 2642.32] system is very +[2642.32 --> 2643.50] unbalanced unfortunately +[2643.50 --> 2644.56] and that's a whole +[2644.56 --> 2645.42] other rant for a whole +[2645.42 --> 2645.94] other show +[2645.94 --> 2647.08] especially the American +[2647.08 --> 2648.04] one but the thing is +[2648.04 --> 2648.62] some friends +[2648.62 --> 2650.14] who work in California +[2650.14 --> 2650.90] said that this is +[2650.90 --> 2651.66] actually a normal +[2651.66 --> 2652.56] clause that they had +[2652.56 --> 2653.34] this in contract +[2653.34 --> 2653.94] in the past +[2653.94 --> 2655.10] I think nobody +[2655.10 --> 2656.72] from working in +[2656.72 --> 2657.36] California is on +[2657.36 --> 2658.14] this panel so +[2658.14 --> 2659.10] maybe somebody +[2659.10 --> 2660.16] listening can keep +[2660.16 --> 2660.74] me honest here +[2660.74 --> 2661.48] but I've been just +[2661.48 --> 2661.96] told yeah don't +[2661.96 --> 2663.04] worry it's always +[2663.04 --> 2663.78] there just don't +[2663.78 --> 2664.36] take it seriously +[2664.36 --> 2665.60] well there's a +[2665.60 --> 2666.14] whole other thing +[2666.14 --> 2667.12] about cruft in +[2667.12 --> 2668.84] again another rant +[2668.84 --> 2669.64] for another day +[2669.64 --> 2671.36] how lawyers deal +[2671.36 --> 2672.86] with cruft could +[2672.86 --> 2673.52] learn a lot from +[2673.52 --> 2674.12] how programmers +[2674.12 --> 2674.90] deal with cruft +[2674.90 --> 2676.04] because a lot of +[2676.04 --> 2676.92] this stuff that is +[2676.92 --> 2678.24] we don't have any +[2678.24 --> 2679.18] sense of dependencies +[2679.18 --> 2680.74] or module reuse +[2680.74 --> 2681.62] or anything like +[2681.62 --> 2682.26] that in law +[2682.26 --> 2683.66] and so you get +[2683.66 --> 2684.80] stuff that literally +[2684.80 --> 2685.50] just gets copy +[2685.50 --> 2686.44] pasted like imagine +[2686.44 --> 2687.62] if you copy pasted +[2687.62 --> 2688.22] all your code +[2688.22 --> 2689.32] all the time +[2689.32 --> 2690.42] we as programmers +[2690.42 --> 2691.80] know of course +[2691.80 --> 2692.76] that creates errors +[2692.76 --> 2693.82] right and we have +[2693.82 --> 2695.16] linters and dependencies +[2695.16 --> 2695.94] and we have all that +[2695.94 --> 2696.30] kind of stuff +[2696.30 --> 2697.18] lawyers have none of +[2697.18 --> 2699.48] that and that is a +[2699.48 --> 2700.42] problem though I'm +[2700.42 --> 2701.02] curious to see if +[2701.02 --> 2701.80] machine learning helps +[2701.80 --> 2702.38] us with that in the +[2702.38 --> 2702.66] future +[2702.66 --> 2716.66] this episode is brought +[2716.66 --> 2717.62] to you by honeycomb +[2717.62 --> 2718.50] find your most +[2718.50 --> 2719.50] perplexing application +[2719.50 --> 2721.54] issues honeycomb is a +[2721.54 --> 2723.52] fast analysis tool that +[2723.52 --> 2724.34] reveals the truth about +[2724.34 --> 2726.04] every aspect of your +[2726.04 --> 2727.34] application in production +[2727.34 --> 2728.42] find out how users +[2728.42 --> 2729.48] experience your code in +[2729.48 --> 2731.02] complex and unpredictable +[2731.02 --> 2732.24] environments find +[2732.24 --> 2733.82] patterns and outliers +[2733.82 --> 2734.86] across billions of +[2734.86 --> 2735.78] rows of data and +[2735.78 --> 2736.52] definitively solve your +[2736.52 --> 2737.84] problems and we use +[2737.84 --> 2738.62] honeycomb here at +[2738.62 --> 2739.26] change well that's why +[2739.26 --> 2739.68] we welcome the +[2739.68 --> 2740.96] opportunity to add them +[2740.96 --> 2741.72] as one of our +[2741.72 --> 2742.76] infrastructure partners +[2742.76 --> 2744.26] in particular we use +[2744.26 --> 2745.20] honeycomb to track +[2745.20 --> 2745.98] down CDN issues +[2745.98 --> 2747.02] recently which we +[2747.02 --> 2748.14] talked about at length +[2748.14 --> 2749.38] on the kaizen edition +[2749.38 --> 2750.62] of the ship it podcast +[2750.62 --> 2751.56] so check that out +[2751.56 --> 2752.26] here's the thing +[2752.26 --> 2753.32] teams who don't use +[2753.32 --> 2754.46] honeycomb are forced +[2754.46 --> 2755.26] to find the needle in +[2755.26 --> 2756.04] the haystack they +[2756.04 --> 2757.30] scroll through endless +[2757.30 --> 2758.30] dashboards playing +[2758.30 --> 2759.54] whack-a-mole they deal +[2759.54 --> 2760.82] with alert floods trying +[2760.82 --> 2761.82] to guess which one +[2761.82 --> 2763.04] matters and they go +[2763.04 --> 2764.24] from tool to tool to +[2764.24 --> 2765.30] tool playing sleuth +[2765.30 --> 2766.30] try to figure out how +[2766.30 --> 2767.26] all the puzzle pieces +[2767.26 --> 2768.64] fit together it's this +[2768.64 --> 2769.78] context switching and +[2769.78 --> 2770.96] tool sprawl that are +[2770.96 --> 2771.82] slowly killing teams +[2771.82 --> 2773.28] effectiveness and +[2773.28 --> 2773.92] ultimately hindering +[2773.92 --> 2775.06] their business with +[2775.06 --> 2775.94] honeycomb you get a +[2775.94 --> 2777.88] fast unified and +[2777.88 --> 2779.74] clear understanding of +[2779.74 --> 2781.00] the one thing driving +[2781.00 --> 2782.42] your business production +[2782.42 --> 2783.66] with honeycomb you +[2783.66 --> 2784.88] guess less and you +[2784.88 --> 2786.10] know more join the +[2786.10 --> 2787.00] swarm and try +[2787.00 --> 2788.24] honeycomb free today +[2788.24 --> 2789.86] at honeycomb.io +[2789.86 --> 2791.68] slash changelog again +[2791.68 --> 2793.82] honeycomb.io slash +[2793.82 --> 2794.64] changelog +[2794.64 --> 2814.26] i was gonna say that +[2814.26 --> 2814.74] like when you were +[2814.74 --> 2815.62] talking about how we +[2815.62 --> 2816.80] have like the laws get +[2816.80 --> 2817.70] written and then they +[2817.70 --> 2819.52] don't get tested or +[2819.52 --> 2820.44] refined i'm like that's +[2820.44 --> 2821.96] gonna like writing code but +[2821.96 --> 2823.42] without test and it's +[2823.42 --> 2824.06] kind of like i don't +[2824.06 --> 2824.80] know it's just running +[2824.80 --> 2826.24] out there and we have +[2826.24 --> 2827.40] no idea if it's like +[2827.40 --> 2828.50] doing the right thing or +[2828.50 --> 2829.64] doing what we intended we +[2829.64 --> 2830.86] just we just wrote it +[2830.86 --> 2832.12] there's no test +[2832.12 --> 2833.12] frameworks there's no +[2833.12 --> 2835.52] linters there's no by +[2835.52 --> 2836.56] the way when you compile +[2836.56 --> 2838.26] it it will be there will +[2838.26 --> 2839.28] be somebody else trying to +[2839.28 --> 2840.40] persuade the compiler to +[2840.40 --> 2841.92] do things totally +[2841.92 --> 2842.78] different from what you +[2842.78 --> 2844.74] intended like it's a +[2844.74 --> 2846.90] very adversarial system +[2846.90 --> 2848.18] that is not set up for +[2848.18 --> 2850.80] robustness it don't get me +[2850.80 --> 2852.04] wrong it works reasonably +[2852.04 --> 2853.08] well in a lot of cases +[2853.08 --> 2854.58] but it's mostly because of +[2854.58 --> 2855.98] humans this was the one +[2855.98 --> 2856.72] thing that drove me nuts +[2856.72 --> 2857.40] about all the smart +[2857.40 --> 2858.98] contract stuff like +[2858.98 --> 2860.06] contracts only work +[2860.06 --> 2861.14] because humans are around +[2861.14 --> 2862.06] to smooth off the rough +[2862.06 --> 2864.06] edges bridge the gaps +[2864.06 --> 2865.44] like as soon as you start +[2865.44 --> 2866.50] making contracts into +[2866.50 --> 2868.34] code like you're just +[2868.34 --> 2869.82] doomed to failure because +[2869.82 --> 2871.16] of all the failure modes +[2871.16 --> 2872.86] of code that we as +[2872.86 --> 2874.72] programmers know and that +[2874.72 --> 2876.12] you know don't go away +[2876.12 --> 2877.00] when you have contracts +[2877.00 --> 2877.98] you just have a more +[2877.98 --> 2879.08] forgiving at the end of +[2879.08 --> 2879.72] the day execution +[2879.72 --> 2881.38] environment because at +[2881.38 --> 2882.16] the end of the day humans +[2882.16 --> 2883.50] are in the loop in a way +[2883.50 --> 2884.22] that they aren't with +[2884.22 --> 2885.30] smart contracts and the +[2885.30 --> 2885.94] way they aren't with +[2885.94 --> 2887.24] code I find it +[2887.24 --> 2888.52] interesting how it also +[2888.52 --> 2890.50] causes some like class +[2890.50 --> 2892.32] problems as well like I +[2892.32 --> 2893.72] have legal insurance so +[2893.72 --> 2894.46] every time I get a +[2894.46 --> 2895.40] contract of any sort I +[2895.40 --> 2896.48] like send it to lawyers +[2896.48 --> 2897.98] to review I'm like is any +[2897.98 --> 2899.70] of this weird but also +[2899.70 --> 2900.40] knowing that I can just +[2900.40 --> 2901.52] like take a pen and just +[2901.52 --> 2902.58] strike through anything I +[2902.58 --> 2903.66] don't like in a contract +[2903.66 --> 2904.78] but so many people think +[2904.78 --> 2905.90] that's like oh no +[2905.90 --> 2907.18] I've gotten this thing +[2907.18 --> 2908.08] this is like set in +[2908.08 --> 2909.82] stone and I can't do +[2909.82 --> 2910.62] any I have to take it +[2910.62 --> 2911.38] or leave it and it's +[2911.38 --> 2911.90] like that's not how +[2911.90 --> 2913.52] this really works so +[2913.52 --> 2914.70] like knowledge also +[2914.70 --> 2915.60] makes it so the legal +[2915.60 --> 2916.64] system's kind of like a +[2916.64 --> 2917.60] little a little bit more +[2917.60 --> 2918.50] wonky than I think it +[2918.50 --> 2919.08] would be otherwise +[2919.08 --> 2920.82] yeah well and on the +[2920.82 --> 2921.68] flip side of course +[2921.68 --> 2923.28] especially in the US +[2923.28 --> 2924.04] Natalie I think you're +[2924.04 --> 2924.84] correct to say that +[2924.84 --> 2925.54] this is less of a +[2925.54 --> 2926.54] problem in the EU +[2926.54 --> 2927.92] though definitely not +[2927.92 --> 2929.56] unknown as a problem +[2929.56 --> 2931.08] you get lawyers who end +[2931.08 --> 2931.92] up working more +[2931.92 --> 2932.94] defensively than they +[2932.94 --> 2933.64] might otherwise +[2933.64 --> 2934.82] because they're +[2934.82 --> 2935.74] thinking of the person +[2935.74 --> 2936.60] like you Chris +[2936.60 --> 2938.90] and so extra effort +[2938.90 --> 2939.66] gets put in +[2939.66 --> 2941.18] extra layers get put in +[2941.18 --> 2941.88] and that's not +[2941.88 --> 2942.72] necessarily a bad +[2942.72 --> 2943.74] thing but it does +[2943.74 --> 2945.60] mean there were +[2945.60 --> 2946.66] virtues to the days of +[2946.66 --> 2947.46] the handshake deal +[2947.46 --> 2949.16] right high trust +[2949.16 --> 2950.50] environments versus +[2950.50 --> 2951.72] low trust environments +[2951.72 --> 2953.26] is a real thing in +[2953.26 --> 2955.00] in law unfortunately +[2955.00 --> 2955.78] for better or for +[2955.78 --> 2957.12] worse and it's +[2957.12 --> 2958.72] absolutely there's all +[2958.72 --> 2959.72] kinds of class and +[2959.72 --> 2961.54] privilege issues around +[2961.54 --> 2963.20] that that again +[2963.20 --> 2964.08] another rant another +[2964.08 --> 2965.44] day and that seems +[2965.44 --> 2966.28] to be the case with +[2966.28 --> 2967.58] every episode we do +[2967.58 --> 2968.20] Natalie we're gonna +[2968.20 --> 2969.16] have to get you back +[2969.16 --> 2970.76] for a for a part two +[2970.76 --> 2972.20] to go deeper dive I +[2972.20 --> 2973.06] have one more question +[2973.06 --> 2973.80] that I want to dive +[2973.80 --> 2974.82] into before we go to +[2974.82 --> 2976.08] kind of the unpopular +[2976.08 --> 2977.32] opinion section which +[2977.32 --> 2978.20] is and you alluded to +[2978.20 --> 2979.62] this earlier are there +[2979.62 --> 2980.78] specific considerations +[2980.78 --> 2981.48] when we're talking +[2981.48 --> 2982.70] about go I know you +[2982.70 --> 2983.32] talked about kind of +[2983.32 --> 2984.90] package management but +[2984.90 --> 2985.82] what are the specific +[2985.82 --> 2987.66] I guess legal areas +[2987.66 --> 2988.50] when it comes to +[2988.50 --> 2989.78] ownership of code that +[2989.78 --> 2991.34] are brought in specific +[2991.34 --> 2992.36] to the go software +[2992.36 --> 2993.40] engineering language +[2993.40 --> 2994.88] you know I think +[2994.88 --> 2996.62] there's one I'm +[2996.62 --> 2997.40] getting a little over +[2997.40 --> 2998.42] my skis here but go +[2998.42 --> 3000.58] is very much a because +[3000.58 --> 3001.40] of the way that you +[3001.40 --> 3001.88] all have done +[3001.88 --> 3003.74] packaging it has some +[3003.74 --> 3005.04] real implications I +[3005.04 --> 3006.64] think for it breaks +[3006.64 --> 3007.56] the brain of a lot of +[3007.56 --> 3008.46] people who come out of +[3008.46 --> 3009.34] like other language +[3009.34 --> 3010.82] package managers right +[3010.82 --> 3012.54] and it has some stuff +[3012.54 --> 3013.50] that really made sense +[3013.50 --> 3014.46] internally to Google +[3014.46 --> 3015.72] there's this real +[3015.72 --> 3017.10] question Chris you were +[3017.10 --> 3017.98] alluding to it earlier +[3017.98 --> 3019.58] the guy in Nebraska +[3019.58 --> 3022.32] xkcd my impression of +[3022.32 --> 3023.14] it from an outside +[3023.14 --> 3024.56] perspective is that go +[3024.56 --> 3025.34] wants to make some of +[3025.34 --> 3026.86] that not a problem by +[3026.86 --> 3028.78] like oh we'll just grab +[3028.78 --> 3030.26] this specific revision +[3030.26 --> 3033.00] from this specific repo +[3033.00 --> 3035.84] and voila like we know +[3035.84 --> 3036.88] exactly what this code +[3036.88 --> 3038.52] is and that guy in +[3038.52 --> 3039.84] Nebraska can't hurt us +[3039.84 --> 3041.16] anymore right he can't +[3041.16 --> 3042.02] upgrade because we know +[3042.02 --> 3043.34] exactly what it is that +[3043.34 --> 3044.68] we've got here but of +[3044.68 --> 3046.10] course that's ends up +[3046.10 --> 3046.66] being a sort of +[3046.66 --> 3047.92] adversarial relationship +[3047.92 --> 3049.24] with that person right +[3049.24 --> 3050.42] it means that when that +[3050.42 --> 3051.54] person brings knowledge +[3051.54 --> 3052.78] to the table brings new +[3052.78 --> 3053.86] versions to the table +[3053.86 --> 3055.34] there's this sort of +[3055.34 --> 3057.24] assumption we never +[3057.24 --> 3058.02] really got to it here +[3058.02 --> 3058.80] Natalie but I think +[3058.80 --> 3061.22] there's this penumbra of +[3061.22 --> 3062.78] it's not ownership but +[3062.78 --> 3063.88] it's sort of entitlement +[3063.88 --> 3065.50] almost right and this is +[3065.50 --> 3066.28] certainly not go +[3066.28 --> 3068.40] specific but this sense +[3068.40 --> 3069.38] of like well I'm using +[3069.38 --> 3071.16] it and so I'm going to +[3071.16 --> 3072.10] treat it a little bit +[3072.10 --> 3073.14] like I have a support +[3073.14 --> 3074.26] contract like traditional +[3074.26 --> 3075.44] software and in fact +[3075.44 --> 3076.28] open source has sort of +[3076.28 --> 3077.06] encouraged this because +[3077.06 --> 3078.78] people it started from +[3078.78 --> 3079.94] this very collaborative +[3079.94 --> 3081.26] community-based culture +[3081.26 --> 3082.44] and so the norms are +[3082.44 --> 3083.34] like hey I'm going to +[3083.34 --> 3084.36] help everybody who shows +[3084.36 --> 3085.42] up in my issue tracker +[3085.42 --> 3086.58] and of course at some +[3086.58 --> 3087.42] point you get too popular +[3087.42 --> 3088.96] and that breaks but we +[3088.96 --> 3089.48] haven't really +[3089.48 --> 3091.54] acknowledged that that's +[3091.54 --> 3093.80] I mean literally the +[3093.80 --> 3094.98] word entitlement comes +[3094.98 --> 3095.76] out of some of the same +[3095.76 --> 3096.70] roots of like the same +[3096.70 --> 3098.12] Latin legal roots as +[3098.12 --> 3100.16] ownership when you own a +[3100.16 --> 3101.58] thing you are entitled to +[3101.58 --> 3103.84] do X and we end up with +[3103.84 --> 3106.22] entitlements without the +[3106.22 --> 3107.48] ownership part without the +[3107.48 --> 3108.62] payments part without the +[3108.62 --> 3110.36] labor part and that +[3110.36 --> 3112.06] really is what sends a lot +[3112.06 --> 3113.86] of things sideways I think +[3113.86 --> 3114.68] in open source that's +[3114.68 --> 3115.64] again that's where Tidelift +[3115.64 --> 3118.58] comes in and I think +[3118.58 --> 3121.06] also Go I think has +[3121.06 --> 3122.44] tried to cabin some of +[3122.44 --> 3123.48] that off with its module +[3123.48 --> 3124.80] ownership but sometimes +[3124.80 --> 3126.32] that where that makes +[3126.32 --> 3127.52] technical sense it may not +[3127.52 --> 3128.78] always make social sense +[3128.78 --> 3131.24] and with more time some +[3131.24 --> 3132.60] other time that would be a +[3132.60 --> 3133.72] fun conversation to have +[3133.72 --> 3135.02] in more depth with with +[3135.02 --> 3135.88] some of the folks you know +[3135.88 --> 3137.16] more on the packaging side +[3137.16 --> 3138.26] than I do that's for sure +[3138.26 --> 3140.26] so you know I guess the +[3140.26 --> 3141.82] the unpopular part is not +[3141.82 --> 3143.14] Go specific but that's +[3143.14 --> 3145.60] definitely the responsibility +[3145.60 --> 3147.28] lies with all of us +[3147.28 --> 3148.08] because we specifically +[3148.08 --> 3150.10] decided in open source +[3150.10 --> 3151.18] that well I'm not owning +[3151.18 --> 3152.22] this thing in like some +[3152.22 --> 3153.84] sense but we very much +[3153.84 --> 3155.20] decided in other senses +[3155.20 --> 3156.90] that like one way of +[3156.90 --> 3157.86] putting it is we decided +[3157.86 --> 3160.22] that use of the code +[3160.22 --> 3162.86] translates into ownership +[3162.86 --> 3164.04] of somebody else's time +[3164.04 --> 3166.56] and like that is not +[3166.56 --> 3168.16] when open source runs +[3168.16 --> 3169.32] the entire world which +[3169.32 --> 3170.78] it really does at this +[3170.78 --> 3172.76] point that doesn't scale +[3172.76 --> 3173.98] and we don't know yet +[3173.98 --> 3175.14] Chris to your earlier +[3175.14 --> 3177.72] point of heresy we don't +[3177.72 --> 3178.48] know how that's going to +[3178.48 --> 3179.88] continue scaling we don't +[3179.88 --> 3181.86] know what happens when we +[3181.86 --> 3183.14] ask everybody to really +[3183.14 --> 3184.92] complicated multi-factor +[3184.92 --> 3186.88] signatures on everything +[3186.88 --> 3188.60] like with very few +[3188.60 --> 3189.50] exceptions people don't +[3189.50 --> 3191.14] hate this stuff they're +[3191.14 --> 3192.18] software developers like +[3192.18 --> 3193.58] they know security matters +[3193.58 --> 3195.60] they know signatures help +[3195.60 --> 3196.88] but they also know they've +[3196.88 --> 3197.76] only got so much time in +[3197.76 --> 3198.52] the world and sometimes +[3198.52 --> 3199.38] they come home from work +[3199.38 --> 3200.04] and they really just want +[3200.04 --> 3201.22] to pitch their laptop into +[3201.22 --> 3202.94] the lake and that does +[3202.94 --> 3204.66] not help you close out +[3204.66 --> 3205.66] your issues if you've +[3205.66 --> 3206.60] pitched your laptop in +[3206.60 --> 3208.72] the lake right so we're +[3208.72 --> 3209.36] gonna have to figure that +[3209.36 --> 3210.54] out might involve buying +[3210.54 --> 3211.58] some people some laptops +[3211.58 --> 3213.00] or maybe writing them a +[3213.00 --> 3214.50] check or maybe it'll +[3214.50 --> 3215.34] involve helping them with +[3215.34 --> 3216.44] AI which we really didn't +[3216.44 --> 3217.54] get to at all but again +[3217.54 --> 3219.62] maybe next time so I +[3219.62 --> 3220.36] remember a conversation +[3220.36 --> 3221.84] that happened I think +[3221.84 --> 3223.02] among a smaller group of +[3223.02 --> 3223.54] people within the +[3223.54 --> 3224.14] community but when +[3224.14 --> 3225.80] modules were were being +[3225.80 --> 3226.64] designed and developed +[3226.64 --> 3227.64] one of the comments that +[3227.64 --> 3228.74] kept coming up was like +[3228.74 --> 3230.48] this is biasing toward the +[3230.48 --> 3232.12] consumer instead of the +[3232.12 --> 3232.90] you know the provider +[3232.90 --> 3234.96] their maintainer and is +[3234.96 --> 3235.80] that a thing we really +[3235.80 --> 3237.02] want to do in the go +[3237.02 --> 3238.28] community and what effect +[3238.28 --> 3239.20] is that going to have on +[3239.20 --> 3240.92] people's ability to +[3240.92 --> 3242.16] actually maintain and +[3242.16 --> 3243.32] build open source things +[3243.32 --> 3244.48] and I think I think we're +[3244.48 --> 3245.26] really starting to see +[3245.26 --> 3246.26] some of the outcomes of +[3246.26 --> 3247.08] that with like what Ben +[3247.08 --> 3248.00] Johnson's been doing where +[3248.00 --> 3249.28] he's just like I don't +[3249.28 --> 3250.30] want contributions I don't +[3250.30 --> 3251.84] want like I there's just I +[3251.84 --> 3252.56] don't have enough bandwidth +[3252.56 --> 3253.58] to kind of deal with some +[3253.58 --> 3254.74] of these things and I +[3254.74 --> 3255.42] think we're gonna see a lot +[3255.42 --> 3256.64] more people that just are +[3256.64 --> 3258.50] like well I can't like if +[3258.50 --> 3259.80] there is a bug and now I +[3259.80 --> 3261.12] have almost no recourse to +[3261.12 --> 3262.10] fix it or I can't get +[3262.10 --> 3263.26] people off old versions of +[3263.26 --> 3264.92] things or I can't that does +[3264.92 --> 3266.44] kind of erode the ability +[3266.44 --> 3268.40] of people to do open source +[3268.40 --> 3269.70] really which erodes our +[3269.70 --> 3271.08] ability to you know maintain +[3271.08 --> 3272.34] the modern world in a sense +[3272.34 --> 3274.20] yeah absolutely and you know +[3274.20 --> 3275.00] that's one thing I would +[3275.00 --> 3276.94] say to the go folks you're +[3276.94 --> 3278.26] not alone in that every +[3278.26 --> 3279.56] ecosystem is struggling with +[3279.56 --> 3280.50] that there are different +[3280.50 --> 3281.90] flavors caused by different +[3281.90 --> 3282.90] technical choices and +[3282.90 --> 3283.82] different cultural choices +[3283.82 --> 3285.58] along the way but the core +[3285.58 --> 3288.02] problem I can't wait we did +[3288.02 --> 3289.24] a tidalift conference right +[3289.24 --> 3290.92] before the pandemic I can't +[3290.92 --> 3291.90] wait to do our next one +[3291.90 --> 3293.20] because very much a theme is +[3293.20 --> 3294.58] going to be how can people +[3294.58 --> 3296.16] across many language +[3296.16 --> 3299.32] ecosystems share notes you +[3299.32 --> 3300.60] know figure out what this +[3300.60 --> 3301.92] looks like because it's very +[3301.92 --> 3304.16] much not if you feel alone +[3304.16 --> 3306.00] if you're having a bad day +[3306.00 --> 3307.62] and you want to chuck your +[3307.62 --> 3309.20] private keys away and never +[3309.20 --> 3310.58] log into github again like +[3310.58 --> 3311.46] you're not alone in that +[3311.46 --> 3313.08] that is a that is a common +[3313.08 --> 3314.94] thing right and github is +[3314.94 --> 3316.42] going to bust their butts I +[3316.42 --> 3317.56] will give them credit they're +[3317.56 --> 3319.30] trying to do a lot to make +[3319.30 --> 3320.06] some of this easier for +[3320.06 --> 3321.42] maintainers but ultimately +[3321.42 --> 3323.40] chris as you say big +[3323.40 --> 3325.12] companies are going to buy +[3325.12 --> 3326.58] us towards doing the right +[3326.58 --> 3328.20] thing for the consumer right +[3328.20 --> 3329.80] like microsoft has done a +[3329.80 --> 3330.76] lot of amazing things for the +[3330.76 --> 3331.70] open source community which +[3331.70 --> 3334.20] 1997 me is like a gassed that +[3334.20 --> 3335.40] I'm saying that out loud but +[3335.40 --> 3336.76] like but at the end of the +[3336.76 --> 3338.40] day I mean when that comes +[3338.40 --> 3340.76] down to push or shove the +[3340.76 --> 3341.72] decisions are often going to +[3341.72 --> 3342.56] be made in favor of the +[3342.56 --> 3344.32] consumer and we do need to +[3344.32 --> 3345.50] have some of those honest +[3345.50 --> 3346.70] discussions about what that +[3346.70 --> 3347.82] looks like because it's not +[3347.82 --> 3349.94] because Natalie that to get +[3349.94 --> 3350.72] back to some of what you're +[3350.72 --> 3352.24] saying and the overall theme +[3352.24 --> 3354.22] of the show legal ownership +[3354.22 --> 3356.72] is only part of the story here +[3356.72 --> 3357.82] cultural ownership +[3357.82 --> 3360.32] responsibility entitlement all +[3360.32 --> 3362.40] these things are related to +[3362.40 --> 3365.84] but cannot be solved just by +[3365.84 --> 3367.44] our legal systems and I think +[3367.44 --> 3369.24] maybe that's my one regret I +[3369.24 --> 3370.76] have very very few regrets +[3370.76 --> 3371.76] about going to law school it's +[3371.76 --> 3372.78] a lot of fun I met a lot of +[3372.78 --> 3374.72] great people but people often +[3374.72 --> 3376.70] come to me seeking legal +[3376.70 --> 3377.88] solutions for what are +[3377.88 --> 3380.32] ultimately cultural problems and +[3380.32 --> 3383.04] I can only do the best lawyers +[3383.04 --> 3384.50] know how to straddle that gap +[3384.50 --> 3386.60] right and I like to think that +[3386.60 --> 3388.06] that is certainly my biggest +[3388.06 --> 3389.78] strength as a lawyer especially +[3389.78 --> 3392.06] in this space is how to straddle +[3392.06 --> 3394.02] that gap but it's not easy +[3394.02 --> 3395.10] which actually by the way +[3395.10 --> 3397.02] reminds me side project fun +[3397.02 --> 3398.56] project and then we'll we'll +[3398.56 --> 3399.42] leave it I know we're running +[3399.42 --> 3401.96] out of time here I am writing a +[3401.96 --> 3404.32] newsletter called open ml dot fyi +[3404.32 --> 3406.98] it is new I literally sort of +[3406.98 --> 3408.58] launched it to some friends a +[3408.58 --> 3410.30] couple weeks ago and more +[3410.30 --> 3412.10] broadly yesterday but is +[3412.10 --> 3413.54] literally about these questions +[3413.54 --> 3414.66] about open and machine +[3414.66 --> 3416.80] learning overlap which +[3416.80 --> 3418.58] includes questions like is +[3418.58 --> 3419.40] this the end of no +[3419.40 --> 3421.26] warranties because every +[3421.26 --> 3422.58] open source license as you +[3422.58 --> 3423.46] pointed out Chris has like +[3423.46 --> 3425.36] big all caps text that says +[3425.36 --> 3427.22] no warranties if you break +[3427.22 --> 3429.94] it you buy it and what does +[3429.94 --> 3431.04] that mean in light of EU +[3431.04 --> 3432.60] regulation you know be +[3432.60 --> 3433.46] talking about a lot of this +[3433.46 --> 3434.76] stuff like copilot there as +[3434.76 --> 3436.12] well so would love to have you +[3436.12 --> 3437.16] all back but for those of you +[3437.16 --> 3438.28] who are curious about that +[3438.28 --> 3440.78] topic it's a ghost a gpl +[3440.78 --> 3443.08] powered newsletter open ml dot fyi +[3443.08 --> 3444.44] we'll add it in the show +[3444.44 --> 3445.66] notes I want to end with +[3445.66 --> 3446.62] like because I know we got to +[3446.62 --> 3447.66] get to unpopular opinions but I +[3447.66 --> 3448.86] just want to say like one of +[3448.86 --> 3449.50] the things I always think +[3449.50 --> 3450.52] about when we get into these +[3450.52 --> 3451.60] conversations is like people +[3451.60 --> 3453.34] tend to think like humans are +[3453.34 --> 3454.88] like transactional and they're +[3454.88 --> 3455.96] like kind of mean to each +[3455.96 --> 3457.22] other and we want a war and all +[3457.22 --> 3458.84] of that and I feel like the +[3458.84 --> 3460.82] existence of open source and +[3460.82 --> 3462.18] the existence of our industry as +[3462.18 --> 3463.80] a whole proves that people are +[3463.80 --> 3465.82] a lot of times selfless and will +[3465.82 --> 3468.38] sacrifice a lot just to make +[3468.38 --> 3470.26] other people feel good just for +[3470.26 --> 3471.62] the happiness of other people +[3471.62 --> 3473.50] and I think that's like shows +[3473.50 --> 3475.24] how like incredible and how +[3475.24 --> 3476.92] collaborative we are as a +[3476.92 --> 3478.84] species but I think more of us +[3478.84 --> 3480.66] need to remember that especially +[3480.66 --> 3482.18] in the times we are now we are +[3482.18 --> 3484.50] not necessarily this always angry +[3484.50 --> 3485.78] at each other always warring +[3485.78 --> 3487.44] always territorial species quite +[3487.44 --> 3488.96] often most of us are just like +[3488.96 --> 3490.60] these we just want to help our +[3490.60 --> 3492.64] fellow people out it's been too +[3492.64 --> 3493.52] long since I worked at +[3493.52 --> 3496.04] Wikipedia but I mean here's this +[3496.04 --> 3497.26] thing this is amazing cultural +[3497.26 --> 3498.94] treasure and anyone can go and +[3498.94 --> 3501.16] graffiti on it at any time and +[3501.16 --> 3502.30] like something like one in a +[3502.30 --> 3505.02] thousand edits are spam right +[3505.02 --> 3506.76] like I mean think about what +[3506.76 --> 3509.16] that says as like to exactly your +[3509.16 --> 3510.98] point Chris right like actually +[3510.98 --> 3513.90] most of the time most people we +[3513.90 --> 3515.58] all want to make this work right +[3515.58 --> 3519.90] and open source open data are very +[3519.90 --> 3523.90] much I think like genuinely amazing +[3523.90 --> 3525.94] that's why I enjoy doing it right +[3525.94 --> 3528.56] like there are a lot more lucrative +[3528.56 --> 3530.34] things probably all of us could be +[3530.34 --> 3532.16] doing with our lives but yeah it's +[3532.16 --> 3533.60] human and humane in a great way +[3533.60 --> 3536.02] lovely note to end the episode we're +[3536.02 --> 3537.38] gonna have to get you back for a part +[3537.38 --> 3540.00] two for sure but before we let you go +[3540.00 --> 3541.52] we're gonna be doing a little bit of +[3541.52 --> 3543.20] unpopular opinions +[3543.20 --> 3565.12] so over to you Lewis what is your go +[3565.12 --> 3567.94] time on popular opinion oh boy I mean +[3567.94 --> 3569.18] the one I have in the show notes is +[3569.18 --> 3570.48] absolutely the one I already nailed +[3570.48 --> 3573.16] which is hey we should all be paying +[3573.16 --> 3574.94] for this right we got it for free for +[3574.94 --> 3578.32] a long time and that train is running +[3578.32 --> 3580.94] out for very human decent reasons +[3580.94 --> 3582.62] right like it's not like I think +[3582.62 --> 3584.40] companies are bad for having used this +[3584.40 --> 3587.26] stuff but as Chris was saying sometimes +[3587.26 --> 3589.98] you raise that employee company I will +[3589.98 --> 3591.84] never forget so there was this project +[3591.84 --> 3594.48] that I was invited to like yeah come to +[3594.48 --> 3595.94] a meeting about right I won't be +[3595.94 --> 3598.00] specific but it was one of many many +[3598.00 --> 3599.86] many many many open source metadata +[3599.86 --> 3602.64] projects and people went on for like +[3602.64 --> 3605.58] about 45 50 minutes and I was like okay +[3605.58 --> 3608.10] but why are volunteers gonna create all +[3608.10 --> 3612.18] this metadata for you quiet silence quiet +[3612.18 --> 3614.86] silence okay but why what's their +[3614.86 --> 3617.22] motivation I don't know it's probably +[3617.22 --> 3619.18] just gonna happen needless to say that +[3619.18 --> 3621.80] project is has not really gone much of +[3621.80 --> 3624.50] anywhere but I was treated as like a +[3624.50 --> 3626.60] pariah and like literally not invited to +[3626.60 --> 3629.02] future meetings for a while because I had +[3629.02 --> 3631.30] dared to ask this like question of why +[3631.30 --> 3633.50] would people do this unfortunately I +[3633.50 --> 3636.26] still get that all too often I think to +[3636.26 --> 3638.24] be fair lots of people are getting the +[3638.24 --> 3639.90] message finally but it's taken longer +[3639.90 --> 3642.06] than I should have that's my sadly +[3642.06 --> 3643.92] unpopular opinion today it's like that +[3643.92 --> 3645.46] meme about that guy that's being thrown +[3645.46 --> 3646.60] outside of the window +[3646.60 --> 3650.28] wait which thrown outside of the window +[3650.28 --> 3651.84] now I gotta google this where it's like +[3651.84 --> 3653.36] they're all in the meeting this comic +[3653.36 --> 3656.12] strip that oh yeah yeah yeah yeah +[3656.12 --> 3659.24] mm-hmm yeah yep been there go +[3659.24 --> 3661.72] intrigued to see how unpopular or in fact +[3661.72 --> 3665.48] popular that opinion is and then I want to +[3665.48 --> 3667.36] ask you Chris for an unpopular opinion +[3667.36 --> 3669.08] given that we're just getting you back +[3669.08 --> 3670.84] and I'm sure you have something on your +[3670.84 --> 3673.40] mind you always do I have so many +[3673.40 --> 3675.92] unpopular opinions I don't think this is +[3675.92 --> 3677.96] gonna be unpopular so I think most people +[3677.96 --> 3680.10] probably agree but it's like a thing I +[3680.10 --> 3681.58] want to put out into the universe more +[3681.58 --> 3685.10] and that is that every tech company +[3685.10 --> 3689.50] larger than probably 20 or 30 people +[3689.50 --> 3693.20] should hire a librarian we create +[3693.20 --> 3696.34] ridiculous amounts of information but +[3696.34 --> 3697.86] then we usually just dump it into a +[3697.86 --> 3699.64] wiki and then we're like we'll be able +[3699.64 --> 3700.94] to find it just use the search +[3700.94 --> 3702.98] functionality or we like try and make a +[3702.98 --> 3705.60] docs page and we're like users will be +[3705.60 --> 3707.06] able to find stuff and it's like there's +[3707.06 --> 3709.22] an actual degree program of people who +[3709.22 --> 3712.04] like get doctorates and how to arrange +[3712.04 --> 3714.50] information so people can find it like +[3714.50 --> 3717.98] go hire them like it's not you don't +[3717.98 --> 3719.26] even have to like they're not even that +[3719.26 --> 3721.24] expensive to hire like just go get a +[3721.24 --> 3723.18] couple like a librarian and archivist and +[3723.18 --> 3725.66] then make your data and your information +[3725.66 --> 3727.50] just much more clean and much more +[3727.50 --> 3729.74] organized it will probably help you make +[3729.74 --> 3731.06] a lot more money in the long run and +[3731.06 --> 3732.80] make your engineers less frustrated with +[3732.80 --> 3734.80] the world how come not all database +[3734.80 --> 3736.66] companies in Google and so on hiring +[3736.66 --> 3739.00] librarians and archivists to do this +[3739.00 --> 3740.48] I'm assuming people don't hire +[3740.48 --> 3742.68] librarians because they just never a I +[3742.68 --> 3743.66] think most people don't know what +[3743.66 --> 3745.60] librarians actually do I think most +[3745.60 --> 3747.00] people just think librarians are the +[3747.00 --> 3748.92] people that like can help you find +[3748.92 --> 3751.38] books in the library and they don't +[3751.38 --> 3752.86] think much more about that they don't +[3752.86 --> 3754.32] think about like that but how do they +[3754.32 --> 3756.18] help you find the books they're just +[3756.18 --> 3757.94] like yeah they just help me find stuff +[3757.94 --> 3759.64] so I think that's part of it and it's +[3759.64 --> 3762.34] just like a unless you're unless you sit +[3762.34 --> 3763.78] down and think about what the problem +[3763.78 --> 3765.34] is I don't think it's like that kind of +[3765.34 --> 3766.72] clear thing you're not going to look +[3766.72 --> 3769.22] necessarily outside of the world you +[3769.22 --> 3770.46] exist in you're gonna be like oh no +[3770.46 --> 3772.88] this is the world like you know we can +[3772.88 --> 3774.42] do this with computers we can just write +[3774.42 --> 3776.40] some code that'll do some indexing and +[3776.40 --> 3779.10] that'll work like I always look at books +[3779.10 --> 3780.82] and I always look at like the indexes +[3780.82 --> 3783.26] they have and I'm like someone is +[3783.26 --> 3785.48] trained probably has like a high-level +[3785.48 --> 3788.32] degree and how to actually pick what +[3788.32 --> 3791.20] words go in an index that's like a +[3791.20 --> 3793.30] really challenging job because there's a +[3793.30 --> 3795.78] crap load of words in a book like well +[3795.78 --> 3797.38] which ones do I pick and put in that +[3797.38 --> 3799.26] it's like well no that's like a hard +[3799.26 --> 3802.32] job and yet books forever well maybe +[3802.32 --> 3803.72] forever forever but for a very long +[3803.72 --> 3805.26] time had indexes and it's like well we +[3805.26 --> 3807.70] should probably get those people but +[3807.70 --> 3808.70] yeah I think most of the time it's like +[3808.70 --> 3810.74] we as technologists are just like no no +[3810.74 --> 3812.88] our our technology will just do it for +[3812.88 --> 3815.26] us we'll write some stat stuff or some +[3815.26 --> 3817.34] ML or AI or whatever and it can +[3817.34 --> 3819.34] obviously replace the thing that humans +[3819.34 --> 3820.46] have been doing very well for a very +[3820.46 --> 3821.98] long time even though we have no idea +[3821.98 --> 3824.36] about that degree program or industry +[3824.36 --> 3826.34] is at all typical things that we do +[3826.34 --> 3828.08] the Times published a book review +[3828.08 --> 3830.90] yesterday of a book on the history of +[3830.90 --> 3833.04] indexes which apparently has like three +[3833.04 --> 3836.12] separate indexes this book on indexes so +[3836.12 --> 3838.46] it looks really interesting and I promise +[3838.46 --> 3840.92] I did not tee that up it's not over +[3840.92 --> 3844.52] company yeah the New York Times actually +[3844.52 --> 3845.62] does some really good work +[3845.62 --> 3851.78] I genuinely forgot that there I'll drop +[3851.78 --> 3856.02] your check off later just a discount on +[3856.02 --> 3857.72] my subscription that's all I ask +[3857.72 --> 3861.26] watch it it has been an absolute pleasure +[3861.26 --> 3862.92] having you on the show thank you so much +[3862.92 --> 3865.06] for joining us it's also wonderful to +[3865.06 --> 3866.90] have you back Chris and wonderful as +[3866.90 --> 3869.74] always to have you like co-presenting +[3869.74 --> 3873.42] with me and regrettably we're now gonna +[3873.42 --> 3876.10] have to say goodbye so thank you all I'm +[3876.10 --> 3877.78] hoping to have everyone together again +[3877.78 --> 3879.86] soon absolutely thank you +[3879.86 --> 3888.04] if you enjoyed hearing from Lewis on +[3888.04 --> 3889.84] this topic take a listen to JS Party +[3889.84 --> 3893.16] episode 188 that one's called we ask a +[3893.16 --> 3895.58] lawyer about github copilot and that +[3895.58 --> 3898.36] lawyer is you guessed it Lewis via I +[3898.36 --> 3900.04] loved that episode and I learned a lot +[3900.04 --> 3902.94] from it you might too at jsparty.fm +[3902.94 --> 3905.58] slash 188 and of course you gotta +[3905.58 --> 3907.02] subscribe to go time if you haven't +[3907.02 --> 3909.52] already head to go time.fm for all the +[3909.52 --> 3912.12] ways lastly let me say if you get value +[3912.12 --> 3913.90] from go time and anything else that we +[3913.90 --> 3915.68] produce here at changelog return some +[3915.68 --> 3918.36] value at changelog.com slash plus plus +[3918.36 --> 3920.76] directly support our work make the ads +[3920.76 --> 3922.92] disappear and get in on some fun bonuses +[3922.92 --> 3924.80] while you're at it thanks again to our +[3924.80 --> 3927.84] partners fastly and fly.io they help us +[3927.84 --> 3929.70] make go time possible and to the +[3929.70 --> 3931.26] mysterious breakmaster cylinder for +[3931.26 --> 3932.94] keeping our beats banging each and +[3932.94 --> 3935.64] every week next time on go time we have +[3935.64 --> 3937.92] a Halloween treat for you Matt Ryer +[3937.92 --> 3940.44] hosts a gathering of ghouls and ghosts +[3940.44 --> 3943.02] telling spooky stories to scare devs +[3943.02 --> 3945.16] stay tuned for that it's gonna be a good +[3945.16 --> 3946.66] one and we'll have it ready for you +[3946.66 --> 3947.36] next week +[3947.36 --> 3959.24] Bruce Wayne +[3959.24 --> 3961.30] you diff --git a/Why immutable databases_transcript.txt b/Why immutable databases_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..cc0277a4ec8fdd725973e0f1c0dda8597cdc2a9a --- /dev/null +++ b/Why immutable databases_transcript.txt @@ -0,0 +1,287 @@ +**Johnny Boursiquot:** Welcome, everybody, to this episode of Go Time, where we get to talk about immutable databases. Now, I must admit, I'm gonna be a skeptic during this show, because I've been looking for use cases, and the project we're gonna be talking about does a very good job of articulating those things... But I'm still very much old-school, if you can call it that; the concept of immutable databases is not something I've ever had to use at work, so I'm looking to -- well, let me take that back. I'm not gonna be a skeptic. I'm gonna approach this with an open mind, and I'm gonna approach it as a learner, as a beginner into this space and into this kind of technology. + +Joining me today is my co-host, Jon Calhoun. Say hi, Jon. + +**Jon Calhoun:** Hey, Johnny. How are you? + +**Johnny Boursiquot:** I am good. Man, we haven't been on a podcast together for like a couple months at least. + +**Jon Calhoun:** It's been a little while. + +**Johnny Boursiquot:** It's been a minute, glad to have you with me here today. Also joining me are two -- I don't know if it's co-founders, or core contributors, or all of the above, but the two of you work on the Codenotary team. I did a quick google around and see that's a company that actually has a product that they're selling, but we're not here to talk about that, we're here to talk about the open source project that the team is behind, called immudb. Joining me to talk about this project are the folks who work on it all the time. I've got Bart Święcki, who is a software engineer, and he's passionate about cryptography, and applied math, and open source, and he's been working on immudb since last year... And obviously, he's been using Go to do that, so we're gonna be peeling back that onion to figure out what makes Go such a good tool for this particular kind of technology. + +\[04:23\] Also joining Bart is Jeronimo Irazabal. Jeronimo also works at the Codenotary, on the team that works on immudb, and he's a software engineer, also passionate about cryptography and databases; I'm seeing a theme here... And also, he's been working on immudb actually a little bit longer, since the year before since July 2020, on this particular project. I'm also interested in hearing what your journey has been using Go to build these kinds of things. So welcome, Bart, and welcome, Jeronimo. + +**Jeronimo Irazabal:** Hi, thanks for having me. + +**Bartlomiej Święcki:** Pleasure to be here. + +**Johnny Boursiquot:** Awesome, awesome. So first of all, I think our audience - not everybody is gonna be familiar with the concept; we all sort of share a common understanding for application developers, writing business applications and whatnot... We all have a pretty common understanding of your database. You write things in, and you update records, and when you need to, you delete things... So it's almost like -- it's a tool for transactions; you record things in there, and when something is no longer needed, you delete it, sometimes you might need to update it, but at any given time, the state of the data within the database is shifting. And in comes this concept of an immutable database, to which I'm scratching my head and I'm thinking "Okay, what is an immutable -- why would I want my database to be immutable?" So please - let's start with you, Bart. Why don't you tell me what an immutable database is? + +**Bartlomiej Święcki:** Yeah, so when you have some information and you put them inside your database, usually we tend to think that this is some kind of temporal state, we can change this, alter this after some time... But what if actually there is some information that you don't want to change? That's where the immutability comes into play. So maybe there are some informations, like critical things, like maybe transactions on your account, or some records that -- let's say you write down the temperature in your room; this is not going to change in the future. So that's where the data itself is immutable. And immutable databases try to work with this kind of information... With information that won't be altered, or maybe a different way - sometimes the data can be altered, but some properties of this information should not be changed, like the history of the values; if you want to scan over the history of the values, and you maybe have a use case where you have to look back what was the state over time, this history will not change. So that's also a property that maybe you want to keep immutable. Also, the database - maybe you want to have an extra layer of protection from the database, so that you don't accidentally change and damage this information. + +I remember when I was working on some standard databases, just common databases - there's this feeling when you delete too much records from the database, and suddenly you feel that "Uh-oh... How can I get out of this situation?" And immutability here helps a lot. It gives you this peace of mind. But there's also much more to it. + +**Johnny Boursiquot:** Right. Let me try to state that back to you, but based on the way I understand it. So when we talk about immutable data - let's just remove the database aspect of it for a second. When we talk about immutable data, we're talking about "What is the state of things? What is the reality of things right now, or at the time I choose to record this data, be it on a piece of paper, or electronically, in a database, whatever?" What is the state of the world right now, at the time I'm writing this piece of data? + +\[07:58\] So if currently it is 50 degrees Fahrenheit, at this hour, at this minute, and in another hour if the temperature rises about ten degrees, and now it's 60 degrees, you're not changing the past. You're not changing when it was 50. You're basically adding a new record, saying "Okay, another snapshot of this data means that at this hour now it is this temperature." + +So it's almost like you're dealing with sort of an append-only logging kind of situation, where at any given time you're able to go back in history to figure out what was the state of the world at this particular point in time. I can see why this creates some sort of a trail, a log, audibility, that kind of thing, and see "Okay, how is this thing changing over time? Who changed it? Why?", whatever. So that applies to a particular use case whereby in most cases what I'm used to is "Give me the current temperature." Whether I asked for that an hour ago or an hour from now, I'm asking for the current temperature. Give me the current temperature. So what you're tracking behind the scenes, multiple versions of it - that's kind of your business, but sometimes I just want whatever the current value, however you determine that, I want whatever the current value is. So these are slightly different use cases. + +So it seems to me that immutable databases are about keeping history of things, not about being almost like your primary database. It's like, if I'm building a weather app, I may want to see what the historical value is, but if we change that a little bit and add, say, a financial services app or something - for a bank, for example - I can see my account balance over time. And every time my account has changed - maybe a new purchase, or a debit, or some sort of deposit - I'm tracking that over time, but at any given time I want to know what's my current balance; can I buy this thing or not? So there are different use cases. One is not supposed to replace the other. Do I understand this correctly? + +**Bartlomiej Święcki:** In a sense, yes... Basically, this immutable database is like a watcher of a time. But it also contains the most recent state. If you want to check the balance, your current balance, it will still be inside this database. So there's still a use case as a primary source of information. But it's actually more about protection against some kind of tampering with the history. So if you want to make sure that the current balance is the true information, how can you be sure that someone did not do some kind of a change in the history, altered the data? How can you be sure that the current state is actually valid? + +Let's have a use case where there is a banking application - I can simplify the use case - and there is a user. If you want to check your current balance, you open this application, check the balance. Then you do some purchases, and then you check the balance again. So you intuitively check if these things match. So if the previous balance and the price that you have to have pay - if this all matches. If it doesn't match, you start being suspicious. Something is wrong. + +And immutability also - and verifiability - can be used to actually make sure that not only the user can do this. You remember the old state, you know the current state, and you can somehow check if this is consistent. And immutability here, and especially in immudb, gives you cryptographical tools to make sure that actually the database did not lie to you. So once something was written in the history, the database cannot say that "Oh, it was something different." It cannot lie to you. Because if it would lie, then you will immediately see this, because of this mathematical proof. + +So if there's something crucial, like audit logs, which after some time you may want to do some investigation, what happened over time, this gives extra protection that you can rely on this information; because the database has proven that up until this point in time it is consistent with the whole history. + +**Johnny Boursiquot:** Okay. + +**Jeronimo Irazabal:** \[12:09\] I think you were covering a lot of problems that are addressed by immutable databases. First I would like to clarify, immutability is an overloaded term, because as, Johnny, you were mentioning, with immutability we usually refer to systems or data structures that are append-only. That treat changes or updates as new data, actually. So when we are doing an update of a record, we are not mutating the original record, but treating the update as a new record is creating the change. + +So we are used to that for immutability. And actually, immudb relies on -- every component in immudb is an append-only data structure. Even the cryptographic data structure I would treat as append-only. But immutability in databases, or even in blockchain - we tend to refer to another thing; not just to append-only, but to the possibility to verify that the history hasn't changed. So every record is registered and cryptographically linked to what happened before. And then you have a way to verify if a given transaction or a given record was present and was not modified anymore once it was written. It doesn't mean that cannot have the current state of your balance account. As a traditional database, you will have data as the current, the latest value that was placed for a given record, because the record will be the key that identified the address or the balance... But also, depending on the use case, it may be a cumulative set of changes, like indeed where we are committing changes. So the current state or the history - it's independent of that. + +What we refer to this type of thing is verifiable. I prefer the term "verifiable database", rather than "immutable database." Because every system has integrity checks to check the consistency of a given record or of a given file, if it is consistent or not. But with tampering detection it's like, giving the possibility to the client application or the application that is using the database to do the integrity validation by themselves. That is one of the differences - the application that is receiving the data from the database was able to run the integrity check, to validate that the data that was received was not modified since it was written. + +**Johnny Boursiquot:** Okay, let's pull on that thread a little bit... So we're not talking about the clients being or maintaining a copy of whatever data you might have at a central immutable database, or verifiable database. You're talking about some sort of a cryptographic verifiability of the data. + +**Jeronimo Irazabal:** So one of the particularities of an immutable database is that at every moment the complete state of the database is captured by your hash value. So that denotes not only the current state, but the complete history of changes up to that point. So the plan in immudb, for instance, or in other immutable databases, is the client who needs to keep track of this current state, the latest state that is known. In the example that Bart mentioned regarding your bank balance account, you may know what was the latest state that you can trust. And based on that and the new changes is where you can compare. You have the base to compare the new changes, or the new results. But the client only needs to keep track of the state of the database at any given point. That is the minimal information. + +**Jon Calhoun:** To make sure I understand this - that means that deleting records also isn't permissible? Is that true? + +**Jeronimo Irazabal:** \[16:15\] Deletion is actually -- we have two labels. We have logical deletion, or physical deletion. Logical deletion is something that can be handled by the application, or by the server. But the difference will be in terms of performance, because the filtering out of the information will be done much faster if it's done directly by the database. + +In immudb we currently have support for logical deletion in both manners - deleting a key, for instance, or by providing an expiration date. But this currently is just a logical deletion. This means that the data will be still there, it will be automatically filtered out, and the client won't receive it. But it's not yet physically deleted. We are under discussions to incorporate physical deletion of data. It's a very interesting topic to discuss what involves physically deleting the data and yet being able to prove. So depending on the data you delete or you remove, is the possibility you have later on to build proof. So it's a very interesting topic. + +**Jon Calhoun:** Yeah. And I'm assuming we're gonna wanna talk about good use cases for an immutable database... But I guess the first thing that comes to my mind is - I feel like you'd have to be careful as to what applications you use those for, because there are rules like GDPR where you have to be able to forget people, essentially... And I could imagine a weird situation where you write something to an immutable database accidentally, and then you realize "How do we fix this?" + +**Bartlomiej Święcki:** And actually, GDPR is the main reason why we started actually thinking about physical deletion... Because some laws require from you to make sure that the data is not accessible at all after some time. Of course, the rules are not clear, because sometimes you have to hide the data from the users, but then you have to keep it for a longer time because there may be some kind of investigation later on. But still, it is possible to actually remove the data, and maybe there is a different reason for that. Because if you have an append-only structure, append-only data and you start putting too much data into it, you will just run out of space. And after some time you want to reclaim maybe this space. Or you have physical constraints of your server and you have to deal with that, and there is a production system running, so maybe you want to just wipe data that is older than some point in time in the past... And still, as Jeronimo said, the state of the database, the hash of the database contains all the history. So this is a very interesting topic. You no longer have the data, but the state needs to calculate this data in, so that you can still prove that the new changes added to the database are consistent with the whole history, since the beginning. + +**Jeronimo Irazabal:** Regarding the use case, a few months ago there was a situation with a famous tennis player and the Covid-19 results. And there was some news regarding multiple results depending on when it was queried from the service. Of course, if that data is stored in immutable databases or in blockchain, then it will be possible to actually know if that data was consistent or was tampered with. That is kind of a use case in a more traditional system of service; it may take time to use an immutable database in this type of systems or services, but I'm sure it will happen with time. So it's not about just sensitive information. + +**Break:** \[20:00\] + +**Johnny Boursiquot:** So it sounds like, of the use cases, some obvious ones are obviously financial transactions, health records, things that you care about that basically change over time; you want to be able to go back at some point and say "Hey, what was the state of things on this date?" and have a high degree of confidence that this data hasn't been altered, hasn't been modified or anything like that. That's the key takeaway here from what I'm gathering. + +So I'm curious, what drives folks like you into this particular domain problem? Why immutable databases? \[laughs\] ...of a lot of things you could be working on. + +**Bartlomiej Święcki:** Yeah, I think we both say that we like playing with cryptography and math. For me personally, when I started learning about immudb and what techniques it uses, the cryptography itself and the mathematics can be very theoretical. And as long as it doesn't find the practical place to give you some benefits, it's still a theory. + +When I've learned about immudb - because I joined the team a few months ago - it was this moment that you find something that is working, a live database that you can easily use it, and it has all this machinery behind it that is doing all these proofs, and it's cryptographically verifying everything, and it keeps everything in place... So that's something for me that is a great benefit for all of us. + +Previously, we could think of this... Maybe there's a project that I want to create, and it would use this technology, but then I find it hard to implement this. And suddenly, I find this kind of database where I have a very easy interface and I can just take it and use it. So for me, that's the major goal of projects like immudb. So we have a lot of knowledge, and actually, the majority of the cryptography and all these algorithms were invented a long time ago, and right now we only started implementing them and implementing them practically. That's where I think immudb -- that's the goal of the project, is give people the way to use immutable databases in a simple way. + +**Jeronimo Irazabal:** \[24:01\] Yes. Before giving the explanation how I ended up here... But actually using immudb for an application developer is exactly the same as using a traditional database. You can download the immudb binary or local container, and you will use any other key-value store, or SQL database as well. + +So before I joined Codenotary, I was working as a software engineer for IBM, and the last projects were related to digital rights management, and that was related to applied cryptography there for generating the crypto materials... And also, I was a contributor for Hyperledger Fabric. By then also I worked also in an experimental project where we added SQL to Hyperledger Fabric. We added SQL support into the changelogs, actually in the smart contracts. But by then, I was convinced that the complexity of the project was quite big. There were many companies or organizations willing to use blockchain just to be sure or to prove themselves or to their clients that the data was not changed. But then they had to run a very complex system, so I always thought about the possibility to have just a traditional database with the verification possibilities. So to have the same verification capabilities like a blockchain provided, but thinking of singular organizations being the owners of the data, but yet to fulfill with auditory requirements or to prove to their clients that the data has not been changed. + +So by then I just started to think about this type of systems, and I got to know about the company and the initial release of immudb. By then, immudb was implemented relying on another key-value store that was written in Go. So that's where I just started to work, + +And related to immutability, I think tampering detection is one of the types of verifications we can do, but there are many other things that are to be included, like what is the latest record that was modified, being able to verify when you are dealing with higher-level data models, like SQL... If you have a database and you have a document like a data model, and you have queries, and you have to verify them. So there is a lot of things to investigate, to explore, and of course, to implement. + +**Johnny Boursiquot:** Mm-hm. So it's not lost on me you mentioned blockchain... We'll come back to that, we'll come back to that. You've piqued my curiosity when you said that you support both SQL -- you can use it both as a traditional RDBMS, SQL database, or as a key-value store. Why the dual modality for accessing data? + +**Jeronimo Irazabal:** Actually, everything started as a log. immudb has a composite construction; everything started as an embedded database, so immudb can be used as an embedded database. This set of logs, append-only logs, that is verifiable - it's like a transparency log. So you can access it. One of the differences of a traditional key-value store is that you can access a given transaction by its unique ID. If you only need to store records, logs, events, and then to query them, you don't need to query the data using an index. Just directly using the entry of the log. That is the basic way of using it. + +\[28:05\] Then we have the possibility to build an index based on a key. Because every transaction or log entry consists of a list of key-value \[unintelligible 00:28:13.15\] So then you can easily get what are the transactions that modify this particular entry. And of course, you will get the latest one, but you also cannot get the history of the transactions that modified this particular... And that is how we implemented temporary capabilities. So you can go back in time in the database and query the database as it was some time ago, without seeing newer changes. + +On top of this, we implemented SQL capabilities. So when you create an entry, thinking in SQL, it ends up being a transaction that consists of key and value entries. So SQL, all the SQL changes or the SQL data model is backed by a key-value database. So actually, the same transaction is what is happening. We are using the key-value transactions to store transactions that are happening in SQL. SQL was added afterwards, so it's possible to use both. + +Of course, they are isolated entries that are inserted using the key-value, not seeing the internal changes or internal entries that are when working with SQL, but both data models are possible. + +The advantage of using SQL, of course - it's easier to model your application. It's easier to work for later on to find index for \[unintelligible 00:29:48.13\] for writing queries, of course... But we also added the possibility to verify in SQL. So that is one of the differences. So you can get a particular row based on the primary key, and this entry will be verified. + +**Johnny Boursiquot:** So you're still able to model your application just like you would in a relational system. + +**Jeronimo Irazabal:** Exactly. + +**Johnny Boursiquot:** It's just basically the encrypted storage that is used, and the verifiability once you pull data out - all these things you're adding sort of on top of the good old model that most developers who built web applications are familiar with, for example. + +So let's talk about the operability of this. But before we jump into that, I see Jon that you've got a burning question you wanna ask... + +**Jon Calhoun:** I don't have a burning question, but... + +**Johnny Boursiquot:** \[laughs\] It's a question. + +**Jon Calhoun:** I was gonna say that the SQL stuff reminds me of the first time I ran into a use case where I didn't necessarily need an immutable database, but I needed to mimic its functionality in some way. Basically, I was working on shipping stuff with addresses and everything, and one of the things that came up were people were like "Well, i wanna be able to change these addresses I used to ship to things", and it became very clear that in a relational database if you have a bunch of previous shipments that all associate with some record and then you change it, then all of a sudden your history is really weird at that point, because that's not actually what those shipments were... And seeing a database like this, it's kind of interesting to -- I think as developers we encounter cases where immutable databases or something like that is useful... I mean, we all use package managers, which are another example of not really having immutability; you can release a version, but once it's released to some package manager, you're kind of stuck... You have to release a new one, essentially. + +**Johnny Boursiquot:** Or you're supposed to be. \[laughs\] I mean, I don't want you to be able to change that. \[laughs\] + +**Jon Calhoun:** As I say, most package managers won't let you do it, so I think as developers we use immutable systems at times, but we kind of like forget about it... Because I think a package manager is a great example of something that really benefits from something where you can verify nothing got changed... Because that would be really bad when you're downloading third-party packages, to not know for sure that that's still the same version... + +\[31:55\] But it's also interesting in the sense that I feel like most systems we work with that use immutability have some sort of scapegoat; the best example I can give is Git. We all use Git, where you can have the history, and it's supposed to basically be immutable... But there's always ways to force changes and to rewrite history, which is not necessarily a great thing, but it's possible. + +So knowing that developers at some point want to rewrite history and stuff, do they have to come in to using immudb -- like they can't come into it, I'm assuming, with the same mindset of like "I can use this exactly like a SQL database." So are there any tips or advice that sort of like help them get out of that mindset that you see people struggling with when they're starting? + +**Bartlomiej Święcki:** So in immudb what actually you could think of is that you can change the data. You can do corrections. But what you will still get - you have this auditibility of the history. So it's like, I'm not lying to anybody that I did not make a mistake; I did make a mistake, I just corrected it. Right now, this is the state. But let me be clear, this is the current state, and previously it was something different. + +Also, this example with changing the address - I think this is something very interesting, because from the key-value level inside immudb we have something like a reference to other key. So instead of getting some specific value, you just try to read it from other key, and just forward it back. But what you can do is you can say that this reference is for that key, at that transaction. So what it means - it freezes the value inside the history, so then you could create let's say a record that there was this kind of shipment, to that person, under that address, at that point in time. That is something unique. Also, I need to comment about one thing... This package manager. + +**Johnny Boursiquot:** \[laughs\] Please, do. + +**Bartlomiej Święcki:** Let me say that we have actually been using these immutable databases, but we just don't know it or just forgot about this... And a very good example is actually go mod proxy. Actually, the technology behind go mod proxy is very similar to what we have; it's this kind of immutable ledger. And actually, we had this situation where we released some kind of tag, some version of immudb, and once somebody just fetches it through the proxy, it's set in stone; you cannot change it, you cannot switch the tag to something else, and very weird situations happened. And actually, this is for security reasons. So if you release something, then everybody who downloads this particular version will only get this version of the code. You have to release a new patch version. And that's actually good; that's good about the security. + +**Jon Calhoun:** I agree that that's good. I guess I would imagine it would make adoption harder, in the sense that developers are just weird about -- like, if somebody releases an invalid package and they wanna pull it back real quickly, they're still weird about like "Now I have to increment the version" and they don't wanna do that. For whatever reason mentally, they're like "I don't wanna admit I made a mistake" and show that to people. + +**Johnny Boursiquot:** Too bad. \[laughs\] + +**Jon Calhoun:** So does that make adoption harder, when you're basically forcing them to do that? + +**Jeronimo Irazabal:** In this, case, in immudb, you have to convince every other client. If you want to roll back the history in immudb, we have to convince every auditor or client that already have that register \[35:18\] locally. That's the only option. + +**Johnny Boursiquot:** Measure twice, cut once... \[laughs\] + +**Bartlomiej Święcki:** But I think that really making a mistake is not something huge. We all make mistakes, and like in real life, there's always an option to correct the mistake. + +**Jeronimo Irazabal:** For example? + +**Johnny Boursiquot:** Let's hear it. \[laughs\] + +**Bartlomiej Święcki:** Releasing a package that contains some bug. + +**Johnny Boursiquot:** Right, right. + +**Bartlomiej Święcki:** Why should we be ashamed of that? Actually, I see that people who can say that they made a mistake and they corrected that, they tend to deal with those issues better than trying to hide it. So I would go that way. + +**Jon Calhoun:** \[35:56\] That makes sense. I mean, I guess there are definitely cases where it makes sense to want to delete things. Like, if you released something on Git that had private keys, clearly you need to try to clean that up... But I agree with you that it is hard; people should be okay with mistakes, but I feel like in practice people are weird with them. + +**Jeronimo Irazabal:** And there is actually a technical situation that happened and there is a rollback. If you are using, let's say, a single master and a single node, and then caches, and you cannot recover the data. So if the back-up you have is old, older than the state that the client has, they will complain about that situation. So that is a situation that could happen and has to be taken into account. So the mistake there will be having only one node, or not having a back-up. + +**Johnny Boursiquot:** I think it's okay to admit mistakes... Mistakes are part of life, it's okay. Just make a new thing and put that out there, and hopefully people don't download your mistake before you have a chance to replace it. \[laughs\] + +I do wanna switch gears real quick to the operability aspect of things. Obviously, if one were to find a use case for immudb, or really immutable databases in general - it's interesting, as I was researching the technology, I came across other things that I'd come across before, but didn't realize that's what they were. I came across Amazon's QLDB, and I'm like, "Hey, that sounds familiar..." And basically, I saw it tracking the origin of these things. When did these things become popular? And there are references going back a few years, but these types of technologies became very popular, I think, in part as a result of an executive order that was issued maybe like a year ago or so on cybersecurity, and things like that. There was mention of producing or having things like a software bill of materials, and I'm like, "Hey, I'm starting to hear more and more about that now." There's advancements we've made with shipping and packaging software, things like that... And all of a sudden these dots are connecting for me about all these things that I read in the past and didn't really know "Where did this thing come from?" kind of thing. + +For those listening in - it's interesting... Basically, find the executive order - it's called Cybersecurity Something-Something. Basically, you can find it on the WhiteHouse.gov website, or whatever... But you'll see this mandate with lots and lots of requirements for cybersecurity and everything else... And you're gonna find software bill of materials and stuff mentioned in there and whatnot, and you can see how things like that are sort of pushing forward the innovation that's happening in this space, especially with things like immudb and whatnot. And one of the use cases that you're enabling or solving for is the whole software -- basically, verifying the contents of a software package. + +We just talked about how basically the go mod proxy, part of the thing that is also part - for those who basically when you download the modules and you see this weird go.sum file, with all the checksums in there and whatnot - all these things play a role into verifying that the version of the piece of software that you just got is indeed... Basically, you're not gonna get a different version that has the same checksum. So all these things come together to provide that sort of verifiability thing. But I know one of those use cases that you tried to solve for head on is this SBOM thing. Can you talk a little bit about that? And then I wanna talk about what it's like to actually run immudb in production. + +**Bartlomiej Święcki:** Yeah, so SBOM (software bill of materials) is a term that is used to -- let's say that you create a software, and today you don't write all the software by yourself, you just use external packages... And when we look at, let's say, a Node.js application, it usually has hundreds of different dependencies. And the same with Golang, where you don't write an HTTP server by yourself, you just take what's in the standard library, and you do the same with contributions from other people. + +\[40:01\] A software bill of materials is basically describing that if we have this binary, or this product, what is it made of? And here we can actually use this immutable ledger, because you just produce those assets, those binaries once, and we can identify them by let's say taking a hash, which is uniquely specifying this specific binary, and says that "This consists of other components, and those components also have this unique ID", maybe some kind of hashes. So that means that if you change anything, even a smaller bit, you will get a totally different binary, and you also have this specific set of components that it was built from. And when you take software companies that are running these binaries then, and it turns out that there's one specific library that has a vulnerability, how can you figure out where are those old components that are vulnerable? By just taking the software bill of materials information, and by just scanning it, what is actually running in production, you can very quickly identify vulnerable components, and then just fix this. + +There were attacks where -- actually, until now, people may not know that the software that they are running is vulnerable. And this executive order is actually saying that you should have this software bill of materials, so that you can trace this information. + +When we talk about immudb and immutable ledgers, you can also store this information in a secure way, so that if it is persisted and the database has given you the proof, then you can rely on this information; you can rely on the fact that it was not changed. Because maybe there would be an attack that someone goes into your database and your production environment changes some kind of libraries and then attacks also the database that stores information about this bill of materials relations, the relations between packages, how you find this. Immutability here protects you, that you can rely on this information. + +**Johnny Boursiquot:** So if we're talking about one of the recent vulnerabilities, in Log4j for example, that made basically the rounds a few weeks ago, if I wanted to find out "Okay, I'm running Java software. Am I running the version of Log4j that was susceptible to that vulnerability? With the bill of materials I can find out exactly "Do I have the specific version anyway on my infrastructure?" And then with something like immudb, an attacker that is leveraging this vulnerability couldn't go and change the software bill of materials and immudb. They couldn't say "You're not really running the vulnerable software" by changing the software bill of materials and immudb, because you'd have to convince the clients that that wasn't true, or that change was true, whatever it is that we're changing. + +**Bartlomiej Święcki:** Yes, exactly. And actually, that is what is the base for the Codenotary, the company that is building immudb; the base of their financial, it's a base that there is this Codenotary cloud that is using immudb to actually store this information. Because even if you don't have to, if you're not obliged to have this software bill of materials, then it's still good to have this information. + +Log4j came out a few months ago, and it was a very critical vulnerability, where it could execute a code by just sending, in many cases, some packet to the server. And we know that there will be more vulnerabilities like that in the future, so it's better to right now be prepared and to start creating this software bill of materials, and when such vulnerability will happen, to quickly find it. + +**Johnny Boursiquot:** Cool, cool. Very briefly, does running immudb - is that process markedly different from, say, managing your own traditional RDBMS, or your traditional key-value store? All things being equal, do I have to do more or less than I would need to, say, run a Postgres server, or a Redis server, or something like that? + +**Bartlomiej Święcki:** \[44:01\] Just run a Docker image, or download the binary and run it, and that's it. + +**Johnny Boursiquot:** And then run it. So the beauty of Go, right? + +**Bartlomiej Święcki:** That's the beauty of Go. + +**Johnny Boursiquot:** You've got your executable. + +**Jeronimo Irazabal:** And depending on the amount of data you are dealing with, it will require some operational procedures, like doing a compaction of the index, that there is some already -- this is already implemented in immudb, for instance. But this is for reducing the space that is required for indexing, because the index itself is an append-only data structure. So there is an operational procedure to automatically compact the index. That is one of the things to take into account. + +And the other is to be aware that you cannot fool the clients that are using immudb. So if you try to revert to another back-up, the clients will complain about that. + +**Break:** \[44:57\] + +**Johnny Boursiquot:** I am interested in obviously understanding why you chose Go for this kind of work... You could have picked a different language. Was there something special about Go that made this kind of work easier to approach? + +**Bartlomiej Święcki:** So when I joined the Codenotary team actually it was already written in Go... + +**Johnny Boursiquot:** Oh. You didn't have a choice. \[laughs\] + +**Bartlomiej Święcki:** But the fact that it was written in Go was very important for me, actually... Because I was watching Go for a very long time. Initially, I didn't have a chance to work with this commercially, in my day-to-day job, but now I see all the benefits that Go gives... Like, having those goroutines - I remember the C++ times when I was writing HTTP servers; the first thing was that you had to write the HTTP server by yourself, in many cases, but then dealing with all those threads, and trying to schedule things, and keep things under control... It was doable, and you could write a performant server, but it took a lot of time. + +So Golang is this sweet spot between the efficiency of programming and still having the performant application in the end. So I think that it is a very good system. We know that Google is using it, because they created it, so it must be battle-tested. It most likely contains this knowledge about large-scale deployments that are built in because of where it is used. And it simplifies so many things that for me it's a very good thing. + +Also, Codenotary is a startup company where efficiency is also very important; testing smarter. We could write, let's say, something faster maybe, a little bit, a few percent, when writing in C++ or even something lower, but then it would take I don't know how many times more time; maybe five, maybe ten even. + +**Johnny Boursiquot:** Cool. What about you, Jeronimo? + +**Jeronimo Irazabal:** \[47:42\] Yeah, exactly. When I joined also immudb was on the initial release... \[laughter\] And it was already written in Go. But we cannot say that we have changed it, made drastic changes. So we didn't change the language, but we could actually because by then we completely wrote from scratch the storage system. Before, immudb was using BadgerDB; that is another key-value store that is written in Go. But I think it was a good choice for the reasons that Bart mentioned. + +I also like it because the code is easy to read, code readability. I find it very, very easy to read code that is written in Golang. It makes it easier having a standard format for Go. + +**Johnny Boursiquot:** Yeah, you're simply content with it. "Oh yeah, that's hilarious too." That's awesome. + +**Bartlomiej Święcki:** It's good you mentioned the formatting of Golang code, because in C++ there was always a war, which one is better, and what to choose. + +**Johnny Boursiquot:** Yeah, nobody does `go fmt`, but everybody loves `go fmt`, yeah. + +**Bartlomiej Święcki:** Actually, I loved it since the first use, I must say. + +**Johnny Boursiquot:** Jon, you've got one final question before we switch it over to unpopular opinions? + +**Jon Calhoun:** I'm fine with jumping to unpopular opinions. + +**Johnny Boursiquot:** Here we go, it's that time. I hope you brought the goodies, gents... Alright, let's get the tune going. + +**Jingle:** \[49:06\] to \[49:25\] + +**Johnny Boursiquot:** Alright, alright, alright... So let's go with Jeronimo first. What have you got? + +**Jeronimo Irazabal:** Mine is not a technical one, but during the pandemic I started to see that a lot of developers started to upload photos of their working environment all the time; in a garden, on the beach... For me it's impossible. I don't know if it's just me, or those photos are just illustrative, like when you see a hamburger ad. So that will be it. I don't know if it's just me, or... + +**Jon Calhoun:** I mean, you've got a whole gym sitting there behind you. \[laughter\] + +**Jeronimo Irazabal:** Actually, this is related to what Bart is going to mention, probably. + +**Johnny Boursiquot:** It's like, "These people are just stuffing their faces..." \[laughs\] + +**Jon Calhoun:** I mean, instead of the gym they just put some flowers and gardens behind them. + +**Johnny Boursiquot:** I know, right? + +**Bartlomiej Święcki:** But once I went outside with my colleagues from a previous job to eat something, and we actually hit some kind of alert and had to act very quickly and we sat somewhere outside... And honestly, the lighting makes it impossible to do anything, to see anything on the laptop. So I kind of agree with that... So it's not that unpopular, because I also agree. You have to have a good environment to do work. Maybe indoor it's much better, but outdoor it would be very hard. + +**Jon Calhoun:** I feel like every person is unique in what they can and can't work with... Because some people love co-working spaces, and it's not that I hate them, but I would never wanna go to one every day of the week. For me, I feel like I'd be less productive there, where other people just thrive. And the same with coffee shops or any of that. I can't work in a coffee shop, and I don't know if it's my back or what, but if I'm working on my laptop all day, where I'm looking down, it eventually hurts my neck, so I have my monitor up higher and everything, and I'm like "I don't know how these people work all the time..." Sure, I can do it occasionally, but I can't do it all the time. But I literally know people who go to coffee shops most days of the week, and I don't know how they do it. + +**Johnny Boursiquot:** Somehow they do. So what's your unpopular opinion, Bart? + +**Bartlomiej Święcki:** Mine is also about exercise maybe... Because I think that as IT in general, the mistake that we are doing is that we start limiting ourselves physically. You have a monitor -- so you work mostly with the head and hands. And nothing else. So it's like, the majority of your body is suspended while you work. And we are flesh and bones mostly, when we take the percentage of ourselves... \[laughs\] And what it means is that if you just shut down part of your body, the whole body will be less efficient. + +**Jeronimo Irazabal:** It's a waste of resources. + +**Bartlomiej Święcki:** \[52:15\] Yeah, a waste of resources, and I was thinking - we approach this all in the wrong way. Why don't we, let's say, have big keyboards, where you can punch things, like use your muscles, and maybe it will increase your productivity. Just think about all these genius doctors in our movies - when they do something, they do this with shouting and waving hands, and things like that... Even if we read histories about some inventions in the past, they were not done while sitting; maybe they were, but some inventions were done when, let's say, running after someone. + +I think we are just limiting ourselves. Why don't we learn things and doing studies like discussing projects during the run, or maybe swimming and solving computations in your head? Maybe this will increase our brain power. + +**Jeronimo Irazabal:** So if I understand you correctly, you are suggesting the outside working environment, but without taking the computer with you. It's like, going to the beach, but without the computer. + +**Johnny Boursiquot:** And running after random strangers. \[laughter\] + +**Jeronimo Irazabal:** You are solving the problem that had was lighting + +**Jon Calhoun:** If I understand it correctly, I assume he's saying that we should explore other ways of doing work that involve our body more, rather than limiting ourselves to sitting at a keyboard. + +**Bartlomiej Święcki:** Yes, exactly. + +**Jon Calhoun:** But if Jerod wants to summarize this as "You should chase after people while you're coding", we can do that, too. \[laughter\] + +**Johnny Boursiquot:** We can do that too, we can do that too. Jon, did you bring one? + +**Jon Calhoun:** I did not. + +**Johnny Boursiquot:** You did not. + +**Jon Calhoun:** I can say I agree with Bart though. I like the idea of thinking about other ways... Johnny, you have a standing desk, don't you? + +**Johnny Boursiquot:** Yeah, I can raise it up and take it back down when I need to. + +**Jon Calhoun:** I used one of those for a while, and basically what I've found was that I didn't like changing my setup all the time... And because I have enough space in my house, what I ended up doing was just getting a desk that's always standing, and putting a walking treadmill under it. I've found that depending on -- like, you can't do everything with it. It's hard sometimes to walk -- like, walking three miles while coding is not easy. Because you stop to think, and you're falling away from your keyboard, and you're like "Whoa, whoa, whoa...!" So you have to keep yourself going. But what is really useful is if I'm watching talks from a conference, or if I'm doing anything like that where I don't really need to type as much, or it's just light emails or something, I can sit there and actually -- it allows me to move my body while also thinking in a little different way... And it's an awesome way to get a break from just sitting through the day. But I think the unfortunate part is that most offices are kind of limited on space. So it's not like you can throw everything you want in there. So it kind of limits that. But I do agree that it'd be nice to see people exploring more interactive ways to do this stuff. + +**Johnny Boursiquot:** Alright, alright, alright... That's good. I didn't bring an unpopular opinion, but I thought of one as we were having this powwow. One thing we didn't get to talk about, but that I am gonna do a show on... So that's the unpopular opinion - I'm doing a show on blockchain at some point in the future. That's it. That's the unpopular opinion. \[laughter\] + +**Bartlomiej Święcki:** It's a dangerous opinion. + +**Jon Calhoun:** So your unpopular opinion is that you think you should do a blockchain show. + +**Johnny Boursiquot:** Yeah, exactly. I think I need to do an episode on blockchain. But yeah, that's gonna be unpopular. Yeah, it's brutal out there, man... \[laughs\] + +**Jon Calhoun:** I'm honestly curious if that's unpopular or just incredibly polarizing... Because there's definitely some people who agree with you doing a show on that. I can't imagine -- I don't know how many though... + +**Johnny Boursiquot:** Right. Well, I'm hoping folks won't shoot the messenger. I'm just the messenger. Look, I just don't want us to bury our heads in the sand and pretend this thing doesn't exist, because clearly, it pisses off a lot of people. So let's just talk about it, like we do most things. Let's just talk about it, and if there are merits, we'll raise them, and if it's complete garbage, we'll show that, too. So yeah, we'll see how well that goes. I hope people don't boycott the show after that... But yeah, we shall see. We shall see. diff --git "a/Why immutable databases\357\274\237_transcript.txt" "b/Why immutable databases\357\274\237_transcript.txt" new file mode 100644 index 0000000000000000000000000000000000000000..9c5b48737e6269c8dd4e9638d8075561b7b7cb3c --- /dev/null +++ "b/Why immutable databases\357\274\237_transcript.txt" @@ -0,0 +1,826 @@ +[0.00 --> 3.68] The cryptography itself and mathematics can be very theoretical. +[4.14 --> 9.78] And as long as it doesn't find the practical place to give you some benefits, it's still a theory. +[10.16 --> 15.24] And when I've learned about IMUDB, that was this moment that you find something that is working, +[15.52 --> 18.28] live database, that you can easily use it. +[18.42 --> 22.12] And it has all this machinery behind it that is doing all these proofs +[22.12 --> 26.26] and is cryptographically verifying everything and keeps everything in place. +[26.26 --> 29.88] So that's something for me that is a great benefit for all of us. +[30.00 --> 36.86] This episode is brought to you by SignalWire. +[37.12 --> 42.28] SignalWire offers APIs, SDKs, and edge networks around the world +[42.28 --> 46.48] for building the realest real-time video and video communication apps +[46.48 --> 48.88] with less than 50 milliseconds of latency. +[49.16 --> 54.14] They use WebSockets to deliver 300% lower latency than APIs built on REST, +[54.14 --> 59.48] making it ideal for apps where every millisecond and responsiveness makes a difference. +[59.48 --> 63.64] Like apps that need instant natural language understanding, real-time machine vision, +[63.98 --> 66.54] or large-scale video and audio conferencing. +[66.90 --> 67.68] Here's what makes them different. +[67.76 --> 74.62] They use MCU, Multi-Point Control Unit, that mixes all video and all audio feeds on the server side +[74.62 --> 79.34] and then distributes a single unified stream back to every participant. +[79.34 --> 84.88] That way, every participant in the apps you ship experiences the same video and the same audio. +[85.26 --> 89.40] Your apps have none of the awkward audio effects, obvious lag, and jumpy video. +[89.70 --> 96.18] It's all smooth, great UX, creating a more lifelike virtual experience without compromising audio or the video quality. +[96.46 --> 98.06] Head to SignalWire.com slash video. +[98.18 --> 101.66] Mention go time to receive an extra 5,000 video minutes. +[101.66 --> 104.56] Again, go to SignalWire.com slash video. +[105.06 --> 106.22] And remember to mention go time. +[119.68 --> 120.72] Let's do it. +[121.26 --> 122.34] It's go time. +[122.82 --> 124.34] Welcome to go time. +[124.74 --> 128.06] Your source for diverse discussions from around the go community. +[128.06 --> 132.60] Subscribe today at go time.fm and follow the show on Twitter. +[132.82 --> 134.40] We are at go time.fm. +[134.76 --> 139.80] Special thanks to our partners at Fastly for delivering go time super fast all around the world. +[140.10 --> 142.68] Check them out for yourself at fastly.com. +[142.92 --> 144.06] That's all for me. +[144.38 --> 145.12] Here we go. +[145.12 --> 161.56] Welcome, everybody, to this episode of Go Time, where we get to talk about immutable databases. +[161.56 --> 171.96] Now, I must admit, I'm going to be a skeptic during this show because I've been looking for use cases, right? +[171.96 --> 176.62] And the project we're going to be talking about does a very good job of sort of articulating those things. +[177.22 --> 180.68] But I'm still very much old school, if you can call it that, right? +[180.74 --> 184.70] The concept of immutable databases has been something I've ever had to use at work. +[184.76 --> 187.50] So I'm looking to, well, let me take that back. +[187.52 --> 188.44] I'm not going to be a skeptic. +[188.58 --> 192.80] I'm going to approach this with an open mind, and I'm going to approach it as a learner, right? +[192.80 --> 196.66] As a beginner to this space and to this kind of technology, okay? +[196.66 --> 201.14] Joining me today is my co-host, John Calhoun. +[201.32 --> 201.90] Say hi, John. +[202.16 --> 202.52] Hey, Johnny. +[202.64 --> 203.00] How are you? +[203.32 --> 203.94] I'm good. +[204.06 --> 207.78] Man, we haven't been on a podcast together for like a couple months at least. +[207.96 --> 208.80] It's been a little while. +[208.96 --> 209.54] It's been a minute. +[209.66 --> 211.30] Glad to have you with me here today. +[211.64 --> 213.62] Also joining me are two. +[214.18 --> 222.78] I don't know if it's co-founders or core contributors or all of the above, but the two of you work on the Code Notary team. +[222.78 --> 228.66] And I did a quick Google around and see and see that's actually a company that actually has a product that they're selling. +[228.78 --> 229.98] But we're not here to talk about that. +[230.08 --> 236.68] We're here to talk about the open source project that the team is behind called EmuDB, right? +[236.96 --> 241.08] Joining me to talk about this project are the folks who work on it all the time. +[241.34 --> 250.44] I've got Bart Sienczki, who is a software engineer and he's passionate about cryptography and applied math and open source. +[250.44 --> 254.74] And he's been working on sort of EmuDB since last year. +[255.36 --> 257.12] And obviously, he's been using Go to do that. +[257.18 --> 263.16] So we're going to be peeling back that onion to figure out what makes you Go such a good tool for this particular kind of technology. +[263.44 --> 266.66] Also joining Bart is Hiranimo Irazapal. +[267.18 --> 272.12] So Hiranimo also works at the Code Notary on the team that works on EmuDB. +[272.70 --> 276.70] And he's a software engineer, also passionate about cryptography and database. +[276.76 --> 277.66] I'm seeing a theme here. +[277.66 --> 286.16] And also, he's been working on EmuDB actually a little bit longer since the year before, since July 2020, on this particular project. +[286.26 --> 290.50] And I'm also interested in hearing what your journey has been using Go to build these kinds of things. +[290.78 --> 292.60] So welcome, Bart, and welcome, Hiranimo. +[292.92 --> 294.36] Hi, nice to have me here. +[294.58 --> 295.42] Pleasure to be here. +[295.94 --> 296.64] Awesome, awesome. +[296.64 --> 303.50] So, first of all, I think our audience, not everybody is going to be familiar with the concept, right? +[303.56 --> 309.06] We all sort of share a common understanding for application developers, you know, writing business applications and whatnot. +[309.52 --> 313.08] We all have a pretty common understanding of your database, right? +[313.08 --> 315.46] You write things in and you update records. +[315.66 --> 317.56] And when you need to, you delete things. +[317.56 --> 321.82] So it's almost like it's a tool for transactions, right? +[321.82 --> 323.04] You record things in there. +[323.14 --> 325.64] And when something is no longer needed, you delete it. +[325.80 --> 326.94] Sometimes you might need to update it. +[326.98 --> 331.80] But at any given time, the state of the data within the database is shifting, right? +[332.32 --> 336.04] And in comes this concept of an immutable database. +[336.04 --> 339.66] And to which I'm scratching my head, I'm thinking, okay, what is an immutable? +[339.66 --> 342.20] Why would I want my database to be immutable? +[342.42 --> 344.26] So please, let's start with you, Bart. +[344.42 --> 347.74] Why don't you tell me what an immutable database is? +[348.40 --> 353.44] Yeah, so when you have some information and you put them inside your database, +[354.20 --> 357.56] usually we tend to think that this is some kind of temporal state. +[357.66 --> 360.10] We can change this, alter this after some time. +[360.34 --> 364.64] But what if actually there is some information that you don't want to change? +[365.22 --> 367.54] And that's where the immutability comes into play. +[367.54 --> 371.78] So maybe there are some informations like critical things, +[371.96 --> 374.00] like maybe transactions on your account, +[374.26 --> 380.90] or some records that, let's say, you write down the temperature on your room. +[381.10 --> 383.16] This is not going to change in the future. +[383.90 --> 386.92] So that's where the data itself is immutable. +[387.88 --> 392.10] And immutable databases try to work with this kind of information, right? +[392.10 --> 397.72] So with the information that won't be altered or maybe in a different way, +[398.52 --> 399.96] sometimes the data can be altered, +[400.12 --> 403.20] but some properties of this information should not be changed, +[403.24 --> 405.52] like the history of the values. +[405.84 --> 408.70] If you want to scan over a whole history of the values +[408.70 --> 412.64] and you maybe have a use case where you have to look back +[412.64 --> 416.62] what was the state of a time, this history will not change. +[416.62 --> 419.94] So that's also a property that maybe you want to keep immutable. +[420.68 --> 422.68] And also the database, +[422.98 --> 426.94] maybe you want to have an extra layer of protection from the database +[426.94 --> 431.40] so that you don't accidentally change and damage this information. +[431.82 --> 434.68] I remember when I was working on some standard databases, +[434.86 --> 435.64] this common database, +[436.04 --> 441.02] there's this feeling when you do delete too much records from the database +[441.02 --> 442.52] and suddenly you feel that, +[442.74 --> 445.50] oh, how can I get out of this situation? +[446.40 --> 448.68] And immutability here helps a lot. +[448.90 --> 450.50] It gives you this peace of mind, +[450.58 --> 452.58] but there's also much more to it. +[452.86 --> 452.90] Right. +[453.26 --> 456.30] Let me try to sort of state that back to you, +[456.42 --> 457.76] but based on the way I understand it. +[457.80 --> 460.22] So when we're talking about immutable data, right, +[460.22 --> 463.08] let's just remove the database aspect of it for a second. +[463.18 --> 464.34] When we talk about immutable data, +[464.40 --> 467.02] we're talking about what is the state of things? +[467.76 --> 469.64] What is the reality of things right now? +[469.64 --> 471.50] Or at the time I choose to record this data, +[471.58 --> 475.06] be it on a piece of paper or electronically in a database, whatever, right? +[475.26 --> 477.22] What is the state of the world right now +[477.22 --> 478.76] at the time I'm writing this piece of data? +[478.88 --> 483.70] So if currently it is 50 degrees Fahrenheit at this hour, this minute, right? +[484.04 --> 485.64] And in another hour, right, +[485.68 --> 488.90] if the temperature rises by 10 degrees and now it's not 60 degrees, +[489.26 --> 490.86] you're not changing the past. +[490.92 --> 492.78] You're not changing when it was 50. +[493.06 --> 495.22] You know, you're basically adding a new record saying, +[495.22 --> 499.98] okay, another snapshot of this data means that at this hour, +[500.08 --> 501.48] now it is this temperature. +[501.66 --> 505.84] So it's almost like you're dealing with sort of an append-only logging kind of situation +[505.84 --> 510.14] where at any given time, you're able to sort of go back in history +[510.14 --> 513.94] to figure out what was the state of the world at this particular time, +[514.08 --> 515.46] this particular point in time. +[515.46 --> 519.28] But for, which is, I can see why sort of a, +[519.36 --> 521.90] this creates some sort of a trail, a log, auditability, +[522.04 --> 524.56] that kind of things and see, okay, well, how is this thing changing over time? +[524.64 --> 525.32] Who changed it? +[525.40 --> 526.60] You know, why, whatever, right? +[527.16 --> 531.22] So that applies to this particular use case whereby in most cases, +[531.22 --> 534.88] what I'm used to is give me the current temperature, right? +[535.20 --> 538.40] Whether I ask for that an hour ago or an hour from now, +[538.54 --> 540.72] I'm asking for the current temperature. +[540.80 --> 541.62] Give me the current temperature. +[542.12 --> 545.44] So what you're tracking behind the scenes, multiple versions of it, +[545.46 --> 547.12] that's kind of your business. +[547.32 --> 550.22] But sometimes I just want whatever the current value, +[550.32 --> 552.90] however you determine that, I want whatever the current value is, right? +[552.90 --> 554.84] So those are slightly different use cases. +[554.92 --> 560.00] So it seems to me that immutable databases are about keeping history of things, right? +[560.00 --> 564.28] Not about sort of being your own, almost like your primary database. +[564.40 --> 566.06] Like if I'm building a weather app, right? +[566.30 --> 569.60] I may want to see, right, what the historical value is. +[569.68 --> 574.06] But if we change that a little bit and add, say, a financial services app or something, +[574.06 --> 576.30] I may not for a bank, for example, right? +[576.64 --> 580.16] When I can see my account balance over time, +[580.78 --> 585.32] and every time this entry was sort of a, every time my account is changed, right? +[585.48 --> 589.22] Maybe a new purchase or a debit or some sort of deposit. +[589.22 --> 591.06] I'm tracking that over time. +[591.58 --> 594.02] But at any given time, I want to know what's my current balance. +[594.10 --> 595.58] Can I buy this thing or not, right? +[595.80 --> 597.22] So there are different use cases. +[597.38 --> 599.34] One is not supposed to replace the other. +[599.54 --> 600.54] Do I understand this correctly? +[601.08 --> 602.30] In a sense, yes. +[602.76 --> 606.22] So basically, this immutable database is like a water over time. +[606.22 --> 608.50] But it also contains the most recent state. +[609.02 --> 611.90] Like if you want to check the balance, your current balance, +[612.28 --> 614.20] it will still be inside this database. +[614.20 --> 618.46] So there is still a use case as a primary source of information. +[618.62 --> 623.94] But it's actually more about protection against some kind of tempering with the history. +[624.76 --> 629.00] So if you want to make sure that the current balance is the true information, +[629.44 --> 635.16] how can you be sure that someone did not do some kind of change in the history, +[635.58 --> 636.38] alter the data? +[636.76 --> 639.78] How can you be sure that the current state is actually valid? +[640.24 --> 643.42] Let's have a use case where there is a banking application. +[643.68 --> 644.86] Like a simplified use case. +[645.04 --> 645.72] And there is a user. +[646.34 --> 650.84] If you want to check your current balance, you open this application, check the balance. +[651.06 --> 652.12] Then you do some purchases. +[652.60 --> 653.92] And then you check the balance again. +[654.50 --> 658.78] So you intuitively check if this thing match. +[659.44 --> 663.98] So if the previous state, previous balance, and the price that you have to pay, +[664.22 --> 665.08] if this all matches. +[665.66 --> 667.96] If it doesn't match, you start being suspicious. +[668.56 --> 669.40] Something is wrong. +[669.78 --> 676.24] And immutability also and verifiability can be used to actually make sure that not only +[676.24 --> 677.40] the user can do this. +[677.60 --> 679.32] So you remember the old state. +[679.62 --> 680.82] You know the current state. +[680.90 --> 683.82] And you can somehow check if this is consistent. +[684.64 --> 689.66] And immutability here, and especially in NimiDB, gives you tools, cryptographical tools, +[690.20 --> 693.42] to make sure that actually the database did not fly to you. +[693.90 --> 697.86] So once something was written in the history, the database cannot say that, +[697.86 --> 699.24] oh, it was something different. +[699.24 --> 700.40] It cannot lie to you. +[700.42 --> 704.16] Because if it would lie, then you will immediately see this. +[704.16 --> 706.52] Because of this mathematical proof. +[706.72 --> 711.98] So if there is something crucial like audit logs, which after some time, +[712.34 --> 715.58] you may want to do some investigation what happened over time. +[715.58 --> 718.08] This gives us extra protection. +[718.08 --> 720.82] But you can rely on this information. +[721.04 --> 727.22] Because database has proven that up to this point in time, it is consistent with the whole +[727.22 --> 727.62] history. +[728.16 --> 728.24] Okay. +[728.54 --> 728.82] Okay. +[728.82 --> 734.96] I think you were covering a lot of problems that are addressed by immutability databases. +[735.56 --> 739.56] First, I would like to clarify, immutability is an overloaded term. +[740.06 --> 746.46] Because as Shonin, you were mentioning, with immutability, we usually refer to systems or +[746.46 --> 753.16] data structures that are a pen only, that treat changes or updates as a new data, actually. +[753.16 --> 759.40] So when we are doing an update of a record, we are not mutating the original record, but +[759.40 --> 763.64] treating the update as a new record, describing the change. +[764.38 --> 767.50] So we are used to that for immutability. +[767.80 --> 774.12] And actually, ImmuDB relies on every component in ImmuDB is an appen only data structure. +[774.64 --> 778.12] Even the cryptographic data structure are treated as appen only. +[778.12 --> 784.86] But in immutability in databases or even in blockchain, we tend to refer to another thing, +[785.06 --> 791.92] not just to append only, but to the possibility to verify that the history hasn't changed. +[792.66 --> 798.08] So every record is registered and cryptographic linked to what happened before. +[798.08 --> 805.12] And then you have a way to verify if a given transaction or a given record was present and +[805.12 --> 808.70] was not modified anymore once it was written. +[809.32 --> 813.98] Doesn't mean that you cannot have the current state of your balance account. +[814.62 --> 820.76] And as a traditional database, you will have either as the current, the latest value that +[820.76 --> 825.92] was placed for a given record, because the record will be the key that identifies the address +[825.92 --> 827.22] or the balance. +[827.82 --> 833.66] But also, depending on the use case, it may be a cumulative set of changes, like in Git, +[833.76 --> 835.48] where we are committing changes. +[836.14 --> 840.78] So the current state or the history, it's independent of that. +[841.14 --> 844.88] What we refer to this type of thing is verifiable. +[844.88 --> 852.12] I prefer the term verifiable database rather than immutable database, because every system +[852.12 --> 854.04] has integrity checks, right? +[854.14 --> 860.84] Internal integrity checks to check the consistency of a given record or of a given file, if it +[860.84 --> 864.86] was consistent, is consistent or not. +[865.36 --> 871.14] But with tampering detection, it's like giving the possibility to the client application or the +[871.14 --> 876.60] application that is using the database to do their integrity validation by themselves. +[877.14 --> 878.46] That is one of the differences. +[879.00 --> 884.50] It's the application that is receiving the data from the database who is able to run the +[884.50 --> 891.72] integrity check to validate that the data that was received was not modified since it was +[891.72 --> 892.08] written. +[892.38 --> 894.12] Okay, let's pull on that thread a little bit. +[894.12 --> 901.50] So we're not talking about the clients sort of being or maintaining a copy of whatever +[901.50 --> 906.90] data you might have at a central sort of immutable database or verifiable database, right? +[906.92 --> 911.88] You're talking about some sort of a cryptographic verifiability of the data. +[912.12 --> 918.92] So one of the particularities of an immutable database is that at every moment, the complete +[918.92 --> 922.38] state of the database is captured by a hash value. +[922.38 --> 929.28] So that denotes not only the current state, but what the complete history of changes up +[929.28 --> 929.92] to that point. +[930.46 --> 937.42] So the client in InModb, for instance, or in other immutable databases, is the client who +[937.42 --> 940.06] needs to keep track of this current state. +[940.52 --> 947.38] The latest state that is known is like in the example that Bar mentioned regarding your bank +[947.38 --> 952.74] bank balance account, you may know what was the latest state that you can trust. +[953.20 --> 957.14] And based on that and the new changes is where you can compare. +[957.38 --> 962.56] You have the base to compare the new changes or the new results and so on. +[962.68 --> 968.22] So, but the client only needs to keep track of the state of the database at any given point. +[968.36 --> 970.10] That is the minimal information. +[970.10 --> 974.60] So to make sure I understand this, that means that deleting records also isn't permissible. +[975.00 --> 975.52] Is that true? +[975.88 --> 977.56] Deletion is actually depending. +[977.90 --> 979.08] We have two levels. +[979.46 --> 982.40] We have logical deletion or physical deletion. +[982.86 --> 988.08] Logical deletion is something that can be handled by the application or by the server. +[988.48 --> 993.54] But the difference will be in terms of performance because the filtering out of the information will +[993.54 --> 997.88] be done much faster if it's done directly by the database. +[997.88 --> 1005.10] In NemoDB, we currently have support for logical deletion in both manners, like deleting a key, +[1005.18 --> 1008.28] for instance, or by providing an expiration date. +[1008.90 --> 1011.86] But this currently is just a logical deletion. +[1012.04 --> 1014.14] This means the data will be still there. +[1014.60 --> 1018.66] It will be automatically filtered out and the client won't receive it. +[1018.92 --> 1021.28] But it's not yet physically deleted. +[1022.02 --> 1026.16] And we are under discussions to incorporate physical deletion of data. +[1026.16 --> 1032.98] And it's a very, very interesting topic to discuss what involves physically deleting the data +[1032.98 --> 1035.46] and yet being able to prove. +[1035.78 --> 1042.20] So depending on the data you delete or you remove, is the possibility you have later on to build proof. +[1042.76 --> 1044.60] So it's a very, very interesting topic. +[1045.14 --> 1045.22] Yeah. +[1045.22 --> 1049.48] And I'm assuming we're going to want to talk about good use cases for an immutable database. +[1049.84 --> 1053.62] But I guess the first thing that comes to my mind is I feel like you'd have to be careful +[1053.62 --> 1058.84] as to what applications you use this for because there are rules like GDPR where you have to be able +[1058.84 --> 1060.18] to forget people, essentially. +[1060.84 --> 1064.74] And I could imagine a weird situation where you write something to an immutable database +[1064.74 --> 1067.90] accidentally and then realize, like, how do we fix this? +[1067.90 --> 1074.98] And actually, GDPR is the main reason why we started actually thinking about physical deletion +[1074.98 --> 1083.44] because some laws require from you to make sure that the data is not accessible at all after some time. +[1083.80 --> 1089.04] Of course, the rules are not clear because sometimes you have to hide the data from the users, +[1089.04 --> 1094.54] but then you have to keep it for a longer time because there may be some kind of investigation later on. +[1094.54 --> 1098.72] But still, it is possible to actually remove the data. +[1098.94 --> 1103.78] And maybe there is a different reason for that because if you have append-only structure, +[1104.00 --> 1109.08] append-only data, and you start putting too much data into it, you will just run out of space. +[1109.86 --> 1115.16] And after sometimes you want to reclaim maybe the space or you have physical constraints of your server +[1115.16 --> 1118.48] and you have to deal with that and there is a production system running. +[1118.48 --> 1126.26] So maybe you want to just wipe data that is older than some point in time in the past. +[1127.18 --> 1134.60] And still, the state, as Hiromimo said, the state of the database, the hash of the database contains all the history. +[1135.24 --> 1137.08] So this is a very interesting topic. +[1137.22 --> 1143.38] So you no longer have the data, but the state needs to calculate this data in +[1143.38 --> 1151.40] so that you can still prove that the new changes added to the database are consistent with the whole history since the beginning. +[1151.92 --> 1160.86] Regarding use case, a few months ago, there was a situation with a famous tennis player and the COVID-19 results. +[1161.40 --> 1168.38] And there was some news regarding multiple results depending on when it was acquired from the service. +[1168.38 --> 1173.42] Of course, if that data is stored in multiple databases or in blockchain, +[1173.60 --> 1179.24] then it will be possible to actually know if that data was consistent or was tampered with. +[1179.86 --> 1184.70] That is kind of use case in a more traditional system or service, +[1185.30 --> 1191.52] that it may take time to use a traditional immutable database in this type of system or service, +[1191.86 --> 1194.78] but I'm sure it will happen with time. +[1194.78 --> 1199.80] So it's not about just sensitive information like it. +[1224.78 --> 1230.86] What would normally be manual error prone tasks across the entire spectrum of responding to an incident? +[1231.20 --> 1234.70] This can all be automated in every way with FireHydrant. +[1235.08 --> 1240.88] FireHydrant gives you incident tooling to manage incidents of any type with any severity with consistency. +[1241.44 --> 1244.56] You can declare and mitigate incidents all inside Slack. +[1244.94 --> 1247.94] Service catalogs allow service owners to improve operational maturity +[1247.94 --> 1250.92] and document all your deploys in your service catalog. +[1250.92 --> 1256.54] Incident Analytics like to extract meaningful insights about your reliability over any facet of your incident +[1256.54 --> 1258.28] or the people who respond to them. +[1258.78 --> 1260.68] And at the heart of it all, Incident Run Books, +[1260.76 --> 1264.88] they let you create custom automation rules to convert manual tasks into automated, +[1265.24 --> 1268.26] reliable, repeatable sequences that run when you want. +[1268.62 --> 1273.42] Create Slack channels, Jira tickets, Zoom bridges instantly after declaring an incident. +[1273.84 --> 1276.50] Now your processes can be consistent and automatic. +[1276.98 --> 1278.66] Try FireHydrant free for 14 days. +[1278.66 --> 1280.24] Get access to every feature. +[1280.38 --> 1281.38] No credit card required. +[1281.70 --> 1283.66] Get started at FireHydrant.io. +[1283.96 --> 1285.72] Again, FireHydrant.io. +[1285.72 --> 1285.84] FireHydrant.io. +[1305.60 --> 1310.00] So it sounds like of the use cases, you know, some obvious ones are, you know, +[1310.02 --> 1315.42] obviously financial transactions, you know, health records and things that basically you care about +[1315.42 --> 1317.00] that basically change over time. +[1317.08 --> 1321.72] You want to be able to go back at some point and say, hey, what was the state of things on this day, right? +[1322.10 --> 1328.84] And have a high degree of confidence that this data hasn't been altered, hasn't been modified or anything like that. +[1329.06 --> 1331.38] That's the key takeaway here from what I'm gathering. +[1331.98 --> 1338.86] So I'm curious, what drives folks like you into this particular domain problem? +[1339.10 --> 1340.38] Why immutable databases? +[1340.38 --> 1342.86] All the things you could be working on. +[1343.86 --> 1344.04] Yeah. +[1344.04 --> 1348.04] I think we both say that we like playing with cryptography and math. +[1348.30 --> 1354.48] For me personally, when I start learning about IMU-DB and what techniques it uses, +[1355.14 --> 1359.02] the cryptography itself and mathematics can be very theoretical. +[1359.02 --> 1367.54] And as long as it doesn't find the practical place to give you some benefits, it's still a theory, right? +[1367.78 --> 1372.82] And when I've learned about IMU-DB, because I joined the team a few months ago, +[1373.32 --> 1379.56] that was this moment that you find something that is working, live database, that you can easily use it. +[1379.56 --> 1388.88] And it has all this machinery behind it that is doing all these proofs and is cryptographically verifying everything and keeps everything in place. +[1389.70 --> 1395.02] So that's something for me that is a great benefit for basically all of us, right? +[1395.10 --> 1397.92] So previously we could think of this. +[1398.34 --> 1406.12] Maybe there is a project that I want to create and it would use this technology, but then I find it hard to implement this. +[1406.12 --> 1412.92] And suddenly I find this kind of database where I have very easy interface and I can just take it and use it. +[1413.66 --> 1418.26] So for me, that's the major goal of a project like IMU-DB. +[1418.82 --> 1427.80] So we have a lot of knowledge and actually majority of the cryptography and all these algorithms were invented a long time ago. +[1428.30 --> 1431.94] And right now we only started implementing them and implementing them practically. +[1431.94 --> 1440.94] And that's where I think IMU-DB is, that's where the goal of the project is, give people the way to use immutable database in a simple way. +[1441.64 --> 1446.92] Yes, before giving the explanation how I end up here. +[1447.54 --> 1455.20] But actually using IMU-DB for application developer is exactly the same as using a traditional database. +[1455.20 --> 1466.12] You can download the IMU-DB binary or docking container and you will use like any other key value store or SQL database as well. +[1466.74 --> 1471.52] So before I joined Cognotory, I was working as a software engineer for IBM. +[1472.34 --> 1477.34] And the last project were related to digital right management. +[1477.84 --> 1483.00] And that was related to applied cryptography there for generating the crypto materials. +[1483.00 --> 1488.80] And also I was a contributor for hyperlateral fabric by then. +[1489.28 --> 1495.00] Also, I worked in an experimental project where we added SQL to hyperlateral fabric. +[1495.54 --> 1501.40] And we added SQL support into the chain codes, like actually in the smart contracts. +[1502.06 --> 1509.04] But by then I was convinced that the complexity of the project was quite big. +[1509.04 --> 1521.06] There were many companies or organizations willing to use blockchain just to be sure or to prove themselves or to their clients that the data was not changed. +[1521.62 --> 1524.94] But then they had to run a very, very complex system. +[1524.94 --> 1532.88] So I always thought about the possibility to have just a traditional database with the verification possibilities. +[1533.28 --> 1543.74] So to have the same verification capabilities like a blockchain provide, but thinking of single organizations being the owners of the data. +[1543.74 --> 1552.40] But yet to fulfill with the auditory requirements or to prove to their clients that the data has not been changed. +[1553.20 --> 1557.22] So by then I started to think about this type of systems. +[1557.56 --> 1562.22] And I got to know about the company and the initial release of Inmudev. +[1562.22 --> 1572.10] By then Inmudev was implemented using Relay in another Go, another key value store that was written in Go. +[1572.68 --> 1575.56] So that's where I started to work. +[1576.28 --> 1583.84] And related to verific immutability, I think tampering detection is one of the type of verification we can do. +[1583.84 --> 1590.12] But there are many other things that we are there to be explored or to be included. +[1590.44 --> 1593.62] Like what is the latest record that was modified? +[1594.10 --> 1600.62] How to be able to verify when you are dealing with higher level data models like SQL. +[1601.20 --> 1608.14] If you have a database and you have a document like data model and you have queries and you have to verify that. +[1608.14 --> 1614.44] So there is a lot of things to get to investigate, to explore and of course to implement. +[1615.10 --> 1617.68] So it's not law something you mentioned blockchain. +[1617.82 --> 1618.56] We'll come back to that. +[1618.70 --> 1619.46] We'll come back to that. +[1619.92 --> 1623.14] You piqued my curiosity when you said that you support both SQL. +[1623.70 --> 1629.96] You can use it both as a traditional sort of RDBMS, SQL database, or as a key value store. +[1630.18 --> 1632.78] Why the dual modality for accessing data? +[1633.12 --> 1636.08] Actually, everything started as a log. +[1636.08 --> 1642.46] So InModb has a composite construction. +[1642.94 --> 1645.36] Everything started as an embedded database. +[1645.74 --> 1648.16] So InModb can be used as an embedded database. +[1648.80 --> 1650.52] That is a set of logs. +[1651.08 --> 1653.16] Append-only log that is verifiable. +[1653.28 --> 1655.46] It's like a transparency log. +[1655.98 --> 1664.22] So you can access one of the differences of traditional key value store is that you can access a given transaction by its ID. +[1664.44 --> 1665.66] Its unique ID. +[1666.08 --> 1676.54] If you only need to store records, logs, events, and then to query them, you don't need to query the data using an index. +[1676.54 --> 1680.44] Just directly using the entry of the log. +[1680.84 --> 1684.08] That is the initial, the basic way of using it. +[1684.66 --> 1688.62] Then we have the possibility to build an index based on the key. +[1688.62 --> 1694.84] So because every transaction or log entry consists of a list of key value entries. +[1695.08 --> 1702.50] So then you can easily get what are the transactions that modify this particular entry. +[1702.50 --> 1705.80] And you will, of course, you will get the latest one. +[1705.80 --> 1712.06] But you also cannot get the history of transactions that modify this particular key. +[1712.06 --> 1715.96] And that is how we implemented temporary capabilities. +[1715.96 --> 1722.68] So you can go back in time in the database and query the database as it was sometime ago. +[1722.68 --> 1725.04] And without seeing newer changes. +[1725.42 --> 1729.16] On top of this, we implemented SQL capabilities. +[1729.16 --> 1741.08] So when you create an entry thinking in SQL, it ends up being a transaction that consists of key and value entries. +[1741.08 --> 1749.20] So SQL, all the SQL changes or SQL data model is backed by a key value database. +[1749.96 --> 1753.94] So actually, the same transaction is what is happening. +[1754.10 --> 1759.10] So we are using the key value transactions to store transactions that happen in SQL. +[1759.78 --> 1762.58] So SQL was added afterward. +[1763.02 --> 1766.08] So it's possible to use both. +[1766.08 --> 1778.64] So, of course, they are isolated entries that are inserted using the key value are not seeing the internal changes or internal entries that are working with SQL. +[1779.02 --> 1781.36] But both data models are possible. +[1781.98 --> 1786.38] The advantage of using SQL, of course, is easier to model your application. +[1786.38 --> 1797.12] Because you have to think it's easier to work when for later on, define index for is efficiently querying the data, for writing queries, of course. +[1797.84 --> 1802.44] But we also added the possibility to verify in SQL. +[1802.66 --> 1804.30] So that is one of the differences. +[1804.54 --> 1807.88] So you can get a particular row based on the primary key. +[1808.46 --> 1810.02] And this entry will be very fine. +[1810.64 --> 1815.74] So you're still able to model your application just like you would in a relational system. +[1815.74 --> 1816.14] Exactly. +[1816.38 --> 1821.84] It's just basically the encrypted storage that is used in the verifiability once you pull data out. +[1822.00 --> 1829.44] All these things you're adding sort of on top of the good old model that most developers who built web applications and whatnot are familiar with for example. +[1829.74 --> 1832.46] So let's talk about the operability of this. +[1832.54 --> 1836.26] But before we jump into that, I see John, you've got a burning question you want to ask. +[1836.58 --> 1837.60] I don't have a burning question. +[1838.16 --> 1839.02] It's a question. +[1839.02 --> 1850.40] I was going to say that like the SQL stuff reminds me of the first time I ran into a use case where I didn't necessarily need an immutable database, but I needed to mimic its functionality in some way. +[1850.82 --> 1853.56] Basically, I was working on like shipping stuff with addresses and everything. +[1853.56 --> 1858.30] And one of the things that came up where people were like, well, I want to be able to change these addresses I used to ship to things. +[1858.30 --> 1870.10] And it became very clear that in a relational database, if you have a bunch of previous shipments that all associate with some record and then you change it, then all of a sudden your history is really weird at that point because that's not actually what those shipments were. +[1870.10 --> 1880.34] And like seeing a database like this, it's kind of interesting to like, I think as developers, we encounter cases where immutable databases or something like that is useful. +[1880.62 --> 1885.54] I mean, we all use package managers, which are another example of like not really having immutability. +[1885.72 --> 1889.72] You can release a version, but once it's released to some package manager, you're kind of stuck. +[1890.04 --> 1891.32] You know, you have to release a new one essentially. +[1891.62 --> 1892.36] Or you're supposed to be. +[1893.22 --> 1895.00] I mean, I don't want you to be able to change that. +[1895.14 --> 1897.22] As I say, most package managers won't let you do it. +[1897.22 --> 1915.40] So I think as developers, we use immutable systems at times, but we kind of like forget about it because I think a package manager is a great example of something that really benefits from something where you can verify nothing got changed because that would be really bad when you're downloading third party packages to like not know for sure that that's still the same version. +[1915.80 --> 1922.28] But it's also like interesting in the sense that I feel like most systems we work with that use immutability have some sort of scapegoat. +[1922.28 --> 1928.50] The best example I can give is like get we always get where you can have the history and it's supposed to basically be immutable. +[1929.02 --> 1934.48] But there's always ways to force changes into like to rewrite history, which is not necessarily a great thing, but it's possible. +[1935.30 --> 1947.16] So knowing that developers at some point want to like rewrite history and stuff, do they have to come into using MUDB like like they can't come into it, I'm assuming, with the same mindset of like I can use this exactly like a SQL database. +[1947.16 --> 1954.16] So are there any like tips or advice that sort of like help them get out of that mindset that you see people struggling with when they're starting? +[1954.16 --> 1960.92] So in MUDB what actually you could think of is that you can change the data, right? +[1960.96 --> 1967.20] You can do corrections, but what you will still get, you have this auditability of the history. +[1967.60 --> 1971.08] So it's like not lying to anybody that I did not make a mistake. +[1971.16 --> 1971.82] I did a mistake. +[1972.08 --> 1973.46] I just corrected it right now. +[1973.56 --> 1974.28] This is the state. +[1974.62 --> 1975.86] But let me be clear. +[1976.04 --> 1977.02] This is what we see. +[1977.38 --> 1978.44] This is the current state. +[1978.66 --> 1980.38] And previously it was something different. +[1980.38 --> 1991.04] And also this example with changing the address, I think this is something very interesting because on the key value level inside MUDB, we have something like a reference to other key. +[1991.68 --> 1998.66] So instead of getting some specific value, you just try to read it from other key and just forward it back. +[1998.96 --> 2004.48] But what you can do is you can say that this reference is for that key at that transaction. +[2005.02 --> 2009.02] So what it means is that it like freezes the value inside the history. +[2009.02 --> 2018.48] So then you could create, let's say, a record that there was this kind of shipment to that person under that address at that point in time. +[2018.92 --> 2020.08] So that is something unique. +[2021.16 --> 2025.50] Also, I need to comment about one thing, this package managers. +[2026.42 --> 2027.10] Let me say that. +[2027.34 --> 2027.62] Please do. +[2028.18 --> 2034.86] We have actually been using this immutable databases, but we just don't know it or just forgot about this. +[2034.86 --> 2038.98] And a very good example is actually GoMod proxy. +[2039.66 --> 2044.60] And actually the technology behind GoMod proxy is very similar to what we have. +[2044.68 --> 2046.16] It's this kind of immutable ledger. +[2046.76 --> 2052.96] And actually we had the situation where we released some kind of tag, some version of MUDB. +[2053.28 --> 2056.94] And once somebody just fetches it through the proxy, it's set in stone. +[2057.12 --> 2058.66] You cannot change it. +[2058.66 --> 2061.14] You cannot switch the tag to something else. +[2061.38 --> 2063.40] And very weird situations happen. +[2063.62 --> 2066.04] And actually this is for the security reasons. +[2066.30 --> 2073.90] So if you release something, then everybody who downloads this particular version will only get this version of the code. +[2074.22 --> 2077.02] You have to release a new patch version. +[2077.40 --> 2078.42] And that's actually really good. +[2078.42 --> 2080.74] And that's good about the security. +[2081.24 --> 2082.12] I agree that that's good. +[2082.30 --> 2095.16] I guess I would imagine it would make adoption harder in the sense that developers are just weird about, like if somebody releases an invalid package and they want to pull it back real quickly, they're still weird about like now I have to increment the version and they don't want to do that. +[2095.34 --> 2099.44] For whatever reason, mentally, they're like, I don't want to admit I made a mistake and show that to people. +[2099.76 --> 2100.12] Too bad. +[2100.12 --> 2103.88] So like, does that make adoption harder when like you're basically forcing them to do that? +[2104.10 --> 2109.34] In this case, in MUDB, you have to convince every other client. +[2110.18 --> 2120.86] If you want to roll back the history in MUDB, you will have to convince every auditor or client that already have that register, that state locally. +[2121.44 --> 2122.50] That's the only option. +[2122.74 --> 2124.00] Measure twice, cut once. +[2124.00 --> 2129.18] But I think that's really making a mistake is not something huge. +[2129.26 --> 2131.08] And we all made mistakes. +[2131.58 --> 2135.46] And like in real life, there is always an option to correct the mistake. +[2135.62 --> 2136.10] For example? +[2137.12 --> 2137.54] Yeah. +[2137.80 --> 2138.52] Like releasing. +[2138.74 --> 2139.24] Let's hear it. +[2139.72 --> 2142.04] Releasing a package that contains some bug. +[2142.42 --> 2142.98] Right, right. +[2143.14 --> 2145.24] And then why should we be ashamed of that? +[2145.24 --> 2154.16] And actually, I see that people who can say that they made a mistake and they corrected that, they tend to deal with those issues better than trying to hide it. +[2154.66 --> 2155.98] So I would go that way. +[2156.52 --> 2157.16] That makes sense. +[2157.36 --> 2161.62] I mean, I guess there are definitely cases where it makes sense to want to delete things. +[2161.70 --> 2166.20] Like if you released something on Git that had private keys, clearly you need to try to clean that up. +[2166.30 --> 2168.36] But it's, I agree with you that it is hard. +[2168.54 --> 2173.18] Like, I don't know, people should be okay with mistakes, but I feel like in practice, people are weird with them. +[2173.18 --> 2176.56] And there is actually a technical situation that could happen. +[2176.86 --> 2177.84] There is a rollback. +[2178.12 --> 2186.34] If you are using, let's say, a single master, a single node, and then it crashes, and you cannot recover the data. +[2187.08 --> 2196.80] So if the backup you have is old, then older than the state that the client has, they will comply about that situation. +[2197.26 --> 2202.08] So that is a situation that could happen and has to be taken into account. +[2202.08 --> 2207.80] So the mistake there will be on having only one node or not having a backup. +[2208.34 --> 2209.86] I think it's okay to admit mistakes. +[2210.04 --> 2211.60] I guess mistakes are part of life. +[2211.72 --> 2212.16] It's okay. +[2212.38 --> 2214.68] You know, just make a new thing and put that out there. +[2214.76 --> 2219.32] And hopefully people don't download your mistake before you had a chance to replace it. +[2219.86 --> 2223.82] Now, I do want to switch gears real quick to the operability aspect of things. +[2223.82 --> 2233.46] Obviously, if one were to find a use case for MUDB or really immutable databases in general, it's interesting. +[2233.82 --> 2239.52] As I was researching the technology, I came across other things that I've come across before but didn't realize that's what they were. +[2239.66 --> 2242.74] Like I came across Amazon's QLDB. +[2243.20 --> 2244.90] I'm like, hey, that sounds familiar. +[2244.90 --> 2248.18] And basically, I started tracking basically the origin of these things. +[2248.30 --> 2250.10] When did these things become popular? +[2250.38 --> 2252.76] And there are references going back a few years. +[2252.92 --> 2264.00] But these types of technologies became very popular, I think, in part as a result of an executive order that was issued maybe like a year or so ago on cybersecurity and things like that. +[2264.00 --> 2269.44] And there was mention of producing or having things like a software bill of materials. +[2269.76 --> 2272.84] And then I'm like, hey, I'm starting to hear more and more about that now. +[2273.22 --> 2278.12] There's like advancements we've made with shipping and packaging software and things like that. +[2278.28 --> 2285.64] And all of a sudden, these dots are connecting for me about all these things that I've read in the past and didn't really know where did this thing come from kind of thing. +[2285.88 --> 2288.48] And for those listening in, it's interesting. +[2288.66 --> 2289.82] Basically, find the executive order. +[2289.94 --> 2292.22] It's called a cybersecurity something, something. +[2292.22 --> 2295.74] It's basically you can find it on the White House.gov website or whatever. +[2295.88 --> 2301.48] But you'll see like this mandate, right, with lots and lots of requirements for cybersecurity and everything else. +[2301.58 --> 2304.48] And you're going to find software bill of materials and stuff mentioned in there and whatnot. +[2304.70 --> 2313.66] And you can see how things like that, right, are sort of pushing forward the innovation that's happening in this space, especially with things like EMUDB and whatnot. +[2313.66 --> 2326.36] And one of the things that one of the use cases that you're enabling or solving for is the whole sort of software, you know, verifying the content of a software package. +[2326.48 --> 2326.60] Right. +[2326.60 --> 2332.68] So we just talked about how, you know, basically the GoMod proxy, right, basically part of the thing that is also part. +[2332.90 --> 2348.78] And for those who basically download your modules and you see this weird Go.sum file with all the checksums in there and whatnot, that, you know, all these things sort of play a role, right, into basically verifying that the version of the piece of software that you just got is indeed, right. +[2348.78 --> 2352.72] It's not going to basically, you're not going to get a different version that has the same checksum, right. +[2352.72 --> 2356.50] So all these things come together to provide that sort of verifiability thing, right. +[2356.82 --> 2363.48] So, but I know one of those use cases that you try to sort of solve for head on, right, is this SBOM thing. +[2363.58 --> 2365.34] Can you take a little bit about that? +[2365.40 --> 2368.92] And then I want to talk about what it's like to actually run EMUDB in production. +[2369.32 --> 2369.46] Yes. +[2369.56 --> 2376.82] So SBOM, so software bill of material, is a term that is used to, let's say, that you create a software and you create it. +[2376.82 --> 2379.52] Today, you don't write all the software by yourself. +[2379.64 --> 2381.56] You just use external packages. +[2382.24 --> 2389.62] And when we look at, let's say, don't JS application, it usually has hundreds of different dependencies. +[2390.50 --> 2391.98] And the same with Golang, right. +[2392.08 --> 2395.54] You don't write HTTP server by yourself. +[2395.64 --> 2400.98] You just take what in standard library and you do the same with contributions from other people. +[2400.98 --> 2408.30] And the software bill of material is basically describing that if we have this binary or this product, what is it made of? +[2409.08 --> 2416.88] And here we can actually use this immutable ledger because we just produce those assets, those binaries once. +[2416.88 --> 2427.38] And we can identify them by, let's say, taking a hash, which is uniquely specifying this specific binary and say that this consists of other components. +[2427.52 --> 2431.80] And those components also have this unique ID, maybe some kind of hashes. +[2431.80 --> 2438.40] So that means that if you change anything, even a smaller bit, you will get totally different binary. +[2439.12 --> 2444.70] And you also have this specific set of components that it was built from. +[2445.10 --> 2452.68] And when you take software companies that are running these binaries then, and then it turns out that there is one specific library that has vulnerability. +[2452.68 --> 2458.68] How can you figure out where are those old components that are vulnerable? +[2459.86 --> 2469.22] And by just taking the software bill of material information and by just scanning it, what is actually running in the production, you can very quickly identify vulnerable components. +[2469.44 --> 2470.64] And then this, fix this. +[2471.06 --> 2479.10] Because there were attacks where actually until now people may not know that their software that they are running is vulnerable. +[2479.10 --> 2486.38] And this executive order is actually saying that you should have this software bill of material so that you can trace this information. +[2487.04 --> 2494.30] And when we talk about immutable ledgers, you can also store this information in a secure way. +[2494.36 --> 2501.00] So that if it is persisted and database has given you the proof, then you can rely on this information. +[2501.36 --> 2503.22] You can rely on the fact that it was not changed. +[2503.22 --> 2519.24] Because maybe there would be an attack that someone goes into your database and your production environment changes some kind of libraries and then attacks also the database that stores information about this bill of material relations, the relations between packages, how you find this. +[2519.70 --> 2523.12] And immutability here protects you that you can rely on this information. +[2523.12 --> 2534.74] So if we're talking about sort of one of the recent vulnerabilities in Log4j, for example, that made, you know, basically the rounds a few weeks ago, if I wanted to find out, okay, I'm running Java software. +[2535.14 --> 2539.84] Am I running the version of Log4j that was susceptible to that vulnerability? +[2540.54 --> 2546.38] With the software bill of materials, I can find out exactly, okay, do I have this specific version anywhere in my infrastructure, right? +[2546.38 --> 2556.26] And then with something like immutability, an attacker that is leveraging this vulnerability couldn't go and change the software bill of material in immutability. +[2556.60 --> 2564.60] They couldn't say you're not really running the vulnerable software by changing the software bill of materials in immutability because you'd have to convince the clients that that wasn't true, right? +[2564.78 --> 2567.38] That change was true, whatever it is that we're changing. +[2567.38 --> 2568.22] Yes, exactly. +[2568.38 --> 2582.84] And actually, that is what is the base for the Code Notary, the company that is building immutability, the base of their financial, let's say, base that there is this Code Notary cloud that is using immutability to actually store this information. +[2583.38 --> 2591.44] Because it looks like even if you don't have to, if you're not obliged to have this software bill of material, then it's still good to have this information. +[2591.44 --> 2604.50] Because Log4J came out a few months ago, and it was a very critical vulnerability where you could execute a code by just sending, in many cases, some packet to the server. +[2605.42 --> 2610.18] And we know that there will be more vulnerabilities like that in the future. +[2610.60 --> 2615.04] So it's better to right now be prepared and to start creating this software bill of materials. +[2615.54 --> 2619.60] And when such vulnerability will happen, to quickly find it. +[2619.80 --> 2621.20] Okay, cool, cool. +[2621.20 --> 2631.76] Very briefly, does running immutability, is that process markedly different from, say, managing your traditional RDBMS or traditional key value store? +[2632.26 --> 2633.62] All things being equal, right? +[2633.64 --> 2641.02] Do I have to do more or less than I would need to, say, run a Postgres server or a Redis server or something like that? +[2641.06 --> 2644.80] Just run Docker image or download binary and run it and that's it. +[2644.94 --> 2645.54] And then run it. +[2645.68 --> 2646.22] Yeah, yes. +[2646.34 --> 2647.54] So the beauty of Go, right? +[2647.80 --> 2648.52] That's the beauty of it. +[2648.56 --> 2649.28] Got your executable. +[2649.28 --> 2659.12] And depending on the amount of data you are dealing with, it will require some operational procedures like doing a compaction of the index. +[2659.12 --> 2665.00] But there is some already, this is already implemented in UDB, for instance. +[2665.08 --> 2672.90] But this is for reducing the space that is required for indexing because the index itself is an append-only data structure. +[2672.90 --> 2679.20] So there is an operational procedure to automatically compact the index. +[2679.66 --> 2682.44] That is one of the things to take into account. +[2683.16 --> 2688.44] And the other is to be aware that you cannot fully the clients that are using. +[2688.44 --> 2695.46] So if you try to revert to another backup, the clients will comply about that. +[2695.46 --> 2714.80] Hey there, it's Jared again. +[2715.06 --> 2716.72] Have you heard about Changelog++? +[2717.22 --> 2718.80] It's our membership program. +[2719.00 --> 2721.52] You can join to directly support our work on GoTime. +[2721.52 --> 2728.70] As a thanks for your support, we hook you up with an ad-free feed, discounts on merch, plus some bonuses like extended episodes. +[2729.20 --> 2732.44] Sign up today at changelog.com slash plus plus. +[2732.44 --> 2753.10] I am interested in obviously understanding why you chose Go for this kind of work. +[2753.18 --> 2755.52] Was there something you could have picked a different language? +[2756.16 --> 2759.52] Was there something special about Go that made this kind of work easier to approach? +[2759.52 --> 2765.26] And so when I joined Code Notary Team, actually it was already written in Go. +[2765.80 --> 2767.84] Oh, you didn't have a choice. +[2768.38 --> 2771.60] But the fact that it was written in Go is very important for me actually. +[2772.10 --> 2775.00] Because I was watching Go for a very long time. +[2775.60 --> 2779.82] Initially, I didn't have a chance to work with this commercially, but in my day-to-day job. +[2779.94 --> 2784.64] But right now I see all the benefits that Go gives, like having those Go routines. +[2784.64 --> 2788.86] I remember the C++ times when I was writing HTTP servers. +[2789.60 --> 2793.76] First thing was that you had to write the HTTP server by yourself in many cases. +[2793.96 --> 2800.44] But then dealing with all those threads and trying to schedule things and make, you know, keep things under control. +[2800.66 --> 2801.66] It was doable. +[2802.12 --> 2805.70] And you could write a performance server, but it took a lot of time. +[2805.70 --> 2815.66] So Golang is this sweet spot between the efficiency of programming and still having the performance application in the end. +[2816.56 --> 2819.68] So I think that it is a very good system. +[2819.80 --> 2823.62] And we know that Google is using it because they created it. +[2823.86 --> 2825.90] So it must be bottle tested. +[2826.16 --> 2834.50] It most likely contains this knowledge about large-scale deployments that are built in because of where it is used. +[2834.50 --> 2840.12] And it simplifies so many things that, for me, it's a very good thing. +[2840.22 --> 2844.64] And also, Cognitory is a startup company where the efficiency is also very important. +[2845.00 --> 2846.04] So these things matter. +[2846.56 --> 2851.12] So we could write, let's say, something faster, maybe a little bit, a few percent, +[2851.54 --> 2855.00] when writing in C++ or even something lower. +[2855.40 --> 2861.14] But then it would take, I don't know how many times more time, maybe five, maybe ten even. +[2861.74 --> 2862.66] Cool. What about you, Haranimo? +[2862.66 --> 2863.58] Yeah, exactly. +[2863.76 --> 2867.82] When I joined, also, InmoDB was on the initial release. +[2868.40 --> 2870.06] And it was already written in Go. +[2870.62 --> 2875.14] But we cannot say that we have changed it, made drastic changes. +[2875.52 --> 2877.58] So we didn't change the language. +[2878.12 --> 2885.02] But we could, actually, because by then we completely write from scratch the storage system. +[2885.02 --> 2888.50] Before, InmoDB was used in BatchelDB. +[2889.22 --> 2892.78] And that is another key value store that is written in Go. +[2892.98 --> 2897.56] But I think it is a good choice for the reasons that Barb mentioned. +[2897.74 --> 2902.62] I also like for the code, easy to read, the code readability. +[2902.62 --> 2906.04] I found it very, very easy to read code. +[2906.44 --> 2907.84] And that is written in Go. +[2907.84 --> 2909.02] And it makes it easier. +[2909.50 --> 2912.98] Having a standard format for the code is reasons. +[2913.16 --> 2914.42] Yeah, you seem pretty content with it. +[2914.90 --> 2915.38] Yeah. +[2915.72 --> 2916.04] Oh, yeah. +[2916.12 --> 2916.94] That's all I heard of. +[2917.02 --> 2917.26] Yeah. +[2917.26 --> 2917.80] That's awesome. +[2918.00 --> 2920.16] Good you mentioned the formatting of Go. +[2920.16 --> 2922.80] Because in C++ there was always a word. +[2923.16 --> 2926.44] Which one is better and what to choose. +[2926.72 --> 2926.88] Yeah. +[2926.90 --> 2928.98] Nobody loves GoFund, but everybody loves GoFund. +[2929.46 --> 2932.20] Actually, I love it since the first years. +[2932.74 --> 2933.78] I must say. +[2934.32 --> 2937.62] John, you got one final question before we switch it over to Unpopular Opinions? +[2937.78 --> 2939.80] I'm fine with jumping to Unpopular Opinions. +[2940.42 --> 2941.20] There we go. +[2941.24 --> 2941.78] It's that time. +[2941.86 --> 2943.70] Oh, I hope you brought the goodies, gents. +[2944.40 --> 2944.78] All right. +[2944.80 --> 2945.82] Let's get the tune going. +[2950.16 --> 2951.32] Unpopular Opinions +[2951.32 --> 2952.04] What? +[2952.32 --> 2954.02] I actually think she'll probably leave. +[2956.60 --> 2959.08] Unpopular Opinions +[2959.08 --> 2965.84] All right. +[2965.98 --> 2966.62] All right. +[2966.78 --> 2967.40] All right. +[2967.84 --> 2970.86] So, let's go with Geronimo first. +[2971.30 --> 2971.80] What you got? +[2972.44 --> 2974.76] Mine is not a technical one. +[2974.76 --> 2983.48] But during the pandemics, I started to see that a lot of developers started to upload photos +[2983.48 --> 2988.08] of their working environment outside, in a garden, in a beach. +[2988.56 --> 2989.86] For me, it's impossible. +[2990.06 --> 2996.94] I don't know if it's just me or those photos are just illustrative, like when you see a hamburger +[2996.94 --> 2997.34] ad. +[2998.44 --> 3000.08] So, that will be. +[3000.32 --> 3000.96] I don't know. +[3001.00 --> 3001.72] It's just me. +[3002.04 --> 3004.14] I mean, you've got a whole gym sitting there behind you. +[3004.14 --> 3004.74] Yeah. +[3006.16 --> 3010.96] Actually, this is related to what Bar is going to mention, probably. +[3011.42 --> 3013.36] It's like, these people are just stuffing their faces. +[3014.00 --> 3017.56] I mean, instead of a gym, they just put some flowers and gardens behind them. +[3017.80 --> 3018.34] All right. +[3018.54 --> 3024.18] But once I went outside with my colleagues from a previous job to eat something, and we +[3024.18 --> 3027.84] actually had some kind of alert and had to act very quickly. +[3027.84 --> 3031.06] And we sat somewhere just outside. +[3031.72 --> 3037.60] And honestly, the lightning makes it impossible to do anything, to see anything on the laptop. +[3038.32 --> 3040.88] So, I kind of agree with that. +[3041.42 --> 3044.62] So, it's not that unpopular because I also agree. +[3044.90 --> 3048.06] You have to have a good environment to do work. +[3048.06 --> 3051.44] Maybe it's indoor, it's much better. +[3051.58 --> 3053.50] But out there, it will be very hard. +[3053.88 --> 3057.74] I feel like every person's unique in what they can and can't work with. +[3058.22 --> 3060.24] Because some people love co-working spaces. +[3061.24 --> 3064.96] And it's not that I hate them, but I would never want to go to one every day of the week. +[3065.42 --> 3068.94] For me, I feel like I'd be less productive there where other people just thrive. +[3069.32 --> 3071.16] And the same with coffee shops or any of that. +[3071.16 --> 3073.26] I can't work in a coffee shop. +[3073.80 --> 3076.14] And I don't know if it's my back or what. +[3076.24 --> 3080.04] But if I'm working on my laptop all day where I'm looking down, it eventually hurts my neck. +[3080.12 --> 3081.60] So, I have my monitor up higher and everything. +[3081.74 --> 3084.46] And I'm like, I don't know how these people work all the time. +[3084.60 --> 3087.44] Sure, I can do it occasionally, but I can't do it all the time. +[3088.14 --> 3091.96] But I literally know people who go to coffee shops most days of the week. +[3092.34 --> 3093.78] And I don't know how they do it. +[3094.14 --> 3094.84] Somehow they do. +[3095.14 --> 3096.64] So, what's your unpopular opinion, Bart? +[3096.64 --> 3101.24] Okay, so my is also about exercise maybe. +[3101.82 --> 3109.42] Because I think that as IT in general, the mistake that we are doing is that we start limiting ourselves physically. +[3109.96 --> 3111.58] Like you have monitor. +[3112.00 --> 3115.54] So, you work mostly with a head and hands and nothing else. +[3115.78 --> 3120.18] So, it's like the majority of your body is suspended while you work. +[3120.18 --> 3126.94] And we are flesh and bones mostly when we take the percentage of ourselves. +[3127.82 --> 3133.68] And what it means is that if you just shut down part of your body, the whole body will be less efficient. +[3134.04 --> 3135.30] It's a waste of resources. +[3135.82 --> 3137.26] Yeah, waste of resources. +[3137.40 --> 3140.78] And I was thinking that we approach this all in the wrong way. +[3140.78 --> 3146.96] So, why don't we, let's say, have a big keyboard when you can punch things, like use your muscles. +[3147.82 --> 3149.76] And maybe it will increase your productivity. +[3150.36 --> 3154.78] Just think about all these genius doctors in our movies. +[3154.78 --> 3163.74] Movies, like they all, when they do something, they do this with shouting and waving hands and things like that. +[3164.14 --> 3171.18] Even if we read histories about some inventions in the past, they were not done while sitting. +[3171.68 --> 3176.72] Maybe they were, but some inventions were done when, let's say, running after someone. +[3176.72 --> 3180.16] And I think we are just limiting ourselves. +[3180.42 --> 3187.92] And why don't we learn things like doing studies, like, I don't know, discussing projects during the run. +[3188.24 --> 3193.60] Or maybe swimming and solving computations in your head. +[3194.12 --> 3196.48] Maybe this will increase our brain power. +[3197.04 --> 3204.80] So, if I understand you correctly, you are suggesting the outside working environment, but without taking the computer with you. +[3204.80 --> 3205.12] Yeah. +[3205.26 --> 3206.44] It's like going to the beach. +[3206.96 --> 3208.00] Without the computer. +[3208.42 --> 3210.24] And running after random strangers. +[3211.20 --> 3213.52] You are solving the problem that they have with lighting. +[3214.22 --> 3222.86] If I understand it correctly, I assume he's saying that we should explore other ways of doing work that involve our body more, rather than, like, limiting ourselves to sitting at a keyboard. +[3223.12 --> 3223.80] Yes, exactly. +[3224.08 --> 3230.84] But if Jared wants to summarize this as, you should chase after people while you're coding, we can do that too. +[3231.66 --> 3232.78] We can do that too. +[3232.78 --> 3233.60] We can do that too. +[3235.22 --> 3236.20] John, did you bring one? +[3236.20 --> 3237.18] I did not. +[3237.54 --> 3238.08] You did not. +[3238.24 --> 3239.60] I can say I agree with Bart, though. +[3239.66 --> 3241.92] Like, I like the idea of thinking about other ways. +[3242.06 --> 3243.32] Johnny, you have a standing desk, don't you? +[3243.72 --> 3244.08] Yes. +[3244.28 --> 3247.96] Yeah, I can raise it up and take it back down when I need to. +[3247.96 --> 3249.72] So I used one of those for a while. +[3250.08 --> 3253.90] And basically what I found was that I didn't like changing my setup all the time. +[3253.90 --> 3260.40] And because I have enough space in my house, what I ended up doing was just getting a desk that's always standing and putting a walking treadmill under it. +[3260.40 --> 3264.24] And I found that depending on, like, you can't do everything with it. +[3264.24 --> 3265.80] Like, it's hard sometimes to walk. +[3265.98 --> 3272.54] Like, walking three miles an hour while coding is not easy because you stop to think and you're, like, pulling away from your keyboard and you're like, whoa, whoa, whoa. +[3273.16 --> 3274.66] So you have to, like, keep yourself going. +[3274.66 --> 3290.16] But where it is really useful is, like, if I'm watching talks from a conference or if I'm doing anything like that where I don't really need to type as much or it's just light emails or something, I can sit there and actually, like, it allows me to sort of move my body while also thinking in a little different way. +[3290.68 --> 3293.74] And it's an awesome way to get a break from just sitting through the day. +[3294.20 --> 3299.38] But I think the unfortunate part is that most offices are, like, kind of limited on space. +[3299.40 --> 3301.64] So it's not like you can throw everything you want in there. +[3302.02 --> 3303.40] So it kind of limits that. +[3303.40 --> 3308.84] But I do agree that it'd be nice to see people exploring more interactive ways to do this stuff. +[3309.54 --> 3309.78] All right. +[3309.88 --> 3310.26] All right. +[3310.32 --> 3310.74] All right. +[3311.26 --> 3311.78] That's good. +[3312.12 --> 3318.44] I didn't bring an unpopular opinion, but I thought of one as we were having this little powwow. +[3318.80 --> 3322.24] One thing we didn't get to talk about that I am going to do a show on. +[3322.38 --> 3323.68] So that's the unpopular opinion. +[3324.18 --> 3327.54] I'm doing a show on blockchain at some point in the future. +[3328.36 --> 3328.66] That's it. +[3328.74 --> 3329.66] That's the unpopular opinion. +[3331.56 --> 3332.86] It's dangerous opinion. +[3332.86 --> 3336.00] So your unpopular opinion is that you think you should do a blockchain show. +[3336.24 --> 3336.92] Yeah, exactly. +[3337.02 --> 3338.70] I think I need to do an episode on blockchain. +[3339.04 --> 3340.64] But yeah, that's going to be unpopular. +[3341.30 --> 3342.78] Yeah, it's brutal out there, man. +[3343.76 --> 3347.42] I'm honestly curious if that's unpopular or just incredibly polarizing. +[3347.88 --> 3350.28] Because there's definitely some people who agree with you doing a show on that. +[3350.58 --> 3351.36] Like, I can't imagine. +[3351.84 --> 3352.84] I don't know how many, though. +[3352.92 --> 3353.12] Right. +[3353.12 --> 3355.62] Well, I'm hoping folks won't shoot the messenger, right? +[3355.68 --> 3356.46] I'm just the messenger. +[3356.46 --> 3360.42] Look, I just don't want us to, you know, bury our heads in the sand and pretend this thing doesn't exist. +[3360.42 --> 3362.70] Because clearly it pisses off a lot of people. +[3363.10 --> 3365.42] So, you know, let's just talk about it, right? +[3365.44 --> 3366.42] Like we do most things. +[3366.42 --> 3367.40] Let's just talk about it. +[3367.48 --> 3370.34] And, you know, if there are merits, we'll raise them. +[3370.46 --> 3373.08] And if it's complete garbage, we'll show that too. +[3373.32 --> 3375.08] So, yeah, we'll see how well that goes. +[3375.60 --> 3378.40] I hope people don't boycott the show after that. +[3378.88 --> 3380.44] But, yeah, we shall see. +[3380.52 --> 3381.02] We shall see. +[3386.70 --> 3388.58] That is go time for this week. +[3389.26 --> 3390.16] Immutable databases. +[3390.68 --> 3391.34] Your thoughts? +[3391.86 --> 3393.56] Let us know in the comments. +[3394.02 --> 3397.50] There's a direct link to the discussion thread at the top of your show notes. +[3397.50 --> 3400.72] Everyone on this episode will be notified of what you have to say. +[3400.88 --> 3405.72] So it's a great place for follow-ups, clarifications, links to related projects, stuff like that. +[3406.04 --> 3407.86] If you're a first-time listener, welcome. +[3408.34 --> 3410.96] Don't forget to subscribe at gotime.fm. +[3411.18 --> 3413.12] We are also in your favorite podcast app. +[3413.30 --> 3414.20] Just search for Gotime. +[3414.20 --> 3419.02] If this is your 10th listen, your 100th, your 1,000th, whatever order of magnitude you have with us, +[3419.14 --> 3421.26] we'd love a review and recommendation. +[3421.82 --> 3425.14] Special thanks to Fastly for CD ending for us all these years, +[3425.36 --> 3427.26] to Breakmaster Cylinder for the Fresh Beats, +[3427.50 --> 3428.42] and to you for listening. +[3428.64 --> 3429.42] We appreciate you. +[3429.82 --> 3432.34] Next week, Matt and John are joined by Ed Welch +[3432.34 --> 3434.58] to discuss logging, logging, and more logging. +[3435.00 --> 3436.24] Yeah, a lot of logging going on there. +[3436.88 --> 3439.88] That's one to look forward to next time on Gotime. +[3439.88 --> 3447.68] END +[3447.68 --> 3447.90] STAN +[3447.90 --> 3449.44] TO believe in themarkt +[3449.44 --> 3450.34] STAN +[3450.34 --> 3450.70] END +[3450.70 --> 3451.28] Nine diff --git a/gRPC & protocol buffers_transcript.txt b/gRPC & protocol buffers_transcript.txt new file mode 100644 index 0000000000000000000000000000000000000000..f8443a585ed354ddcc3d2aad3c30072a86757b8e --- /dev/null +++ b/gRPC & protocol buffers_transcript.txt @@ -0,0 +1,479 @@ +**Jon Calhoun:** Hello everyone, and welcome to Go Time. Today we're going to be talking about protocol buffers and gRPC, as well as some common misconceptions around them. Today I am joined by Akshay Shah, who works on a profile compiler, schema registry and RPC tools at Buf. Previously, he's worked on several thrift compilers and a custom L7 network protocol. So he has a lot of experience in this space with protocol buffers, RPC and related technologies. Also, if you've ever used the ZAP logging package, you've probably used Akshay's code. Akshay, how are you? + +**Akshay Shah:** I'm good. How are you doing, Jon? + +**Jon Calhoun:** I am doing well. A little bit raspy in the throat, so I apologize for that. I'm also hosting with Johnny Boursiquot. Johnny, how are you? + +**Johnny Boursiquot:** I'm feeling good. I'm tired, but relieved. I've been working on something for a little while, and it's finally coming out tomorrow, so I'm pretty excited about that. + +**Jon Calhoun:** That's awesome. Release days are always fun... Assuming they go well. I hope yours does. + +**Johnny Boursiquot:** Yes... + +**Jon Calhoun:** Alright, so you'll have to let us know about that once that's released. + +**Johnny Boursiquot:** Yeah, indeed. + +**Jon Calhoun:** So Akshay, we first started talking back whenever Mat did an episode on - what was it, bloat that can come with software? And protocol buffers came up because there's something that when you import them, they can lead to a lot of extra code being imported sort of behind the scenes, without realizing it... So we wanted to talk a little bit about sort of different implementations of these things, and all that stuff... But first, let's just start off with the basics. What are protocol buffers? + +**Akshay Shah:** That's a great question. Protocol buffers are actually -- they're two separate things, that are closely related, but it's important to have a little bit of airspace between them. The first thing they are is they're a little language for writing schemas for your data. And the second thing is that it's also a binary format to serialize the data. And those two things are closely related, but they're not quite the same. Ultimately, they're a tool to make programmers more efficient. + +**Jon Calhoun:** So I guess if we're looking at them as a tool to make programmers more efficient, was that the only reason they were created, was just that efficiency? + +**Akshay Shah:** I think no, but I think certainly today that's the most compelling thing about them. So if you imagine, we're mostly Go programmers and Go has really found a home in building network services, and building microservices... So I imagine that a lot of Go programmers have written a REST API, or at least written some structs to serialize data to JSON. And so the way you do that in Go is you're writing your REST API, and typically you'll write a Go struct for each request and response shape. And you'll add some struct tags to tell the standard lib how to transform your struct into this text wire format. And that's fine, it's actually really nice, it's pretty ergonomic... The problem is that if you have an API, you probably have clients, and they might be in TypeScript, in a web browser, they might be in Kotlin, or Swift, on mobile clients... Maybe there's some Python client on your backend... And all those clients end up rewriting that exact same Go struct for the request and the response, in their language of choice. And so you end up with the exact same data, and the same schema for that data, rewritten by hand, in Go, in TypeScript, in Kotlin, and Python, and C++... All over the place. That's just - it's a bunch of toil for not much purpose, it's really error-prone, and it's hard to reason about it. You have to manually look at this special part of your code that's exposed to colors, and with every change, you have to set and you have to think, "Okay, how is this going to affect my clients? Can I rename this field? Can I change it from being required to being optional? Can I change it from being optional to being required?" + +Google was running into this in the early 2000s. At the time, XML was really in vogue, and they were looking for something that was simpler, that was more productive for programmers, and as an important concern for Google, but maybe not so much for the rest of us, something that was more efficient for computers... And so they invented protocol buffers. + +**Jon Calhoun:** \[06:12\] So I guess I'm thinking of this, if it's something Google invented for Google scale, is this one of those technologies that is only a good fit if you're a Google? Or what types of applications does this fit well for? + +**Akshay Shah:** Yeah, it's certainly compelling if you're at Google scale; I mean, clearly, they're still protobuf from top to bottom... And most other large -- if you look at a lot of these large tech companies, they have some equivalent protocol buffers. Sometimes it's actually a protobuf, sometimes it's a very similar system, that took a bunch of inspiration from protobuf. + +But for any Gopher, I think you want to avoid redoing the same error-prone work over and over again. That's why we're not calling the pthread APIs directly, we're not freeing memory manually... These are just tedious, error-prone tasks. + +So if you're building an API and you have clients, writing a protobuf schema is basically the same amount of work as writing a Go struct. It looks pretty similar; it's a name, a curly brace, some field names and some types. There's some numbers for the fields, but other than that, it's basically the same thing. You run a code generator, and you get more or less the same Go structs that you would have written by hand, but you can add a little bit to your code gen invocation and you can also get Ruby classes, or TypeScript types, or Python classes, Kotlin stuff. And you can get the code -- the code that you generate, it works with the protobuf binary format, but it works equally well with JSON. So in a lot of circumstances, I certainly use protocol buffers just as a schema language for my JSON. + +**Johnny Boursiquot:** What are you trading off? This sounds all good, but there's always a trade-off. + +**Akshay Shah:** Of course there is, yeah. So I think in Go, you're not trading off a ton, because you would have been handwriting these structs one way or another. I don't see much Go code that's really dealing with JSON as like map string any, as part of writing an API. In other languages, you do trade off the ability to have really loosely-typed APIs. So in Python, you could certainly unmarshal JSON into a dictionary, without any type ends, and just pass that around. And that's totally fine. It takes basically no code. And protobuf encourages you to do that differently. You need a schema. + +In Go, sometimes the protobuf representation of a type in JSON is not what you might expect. So if you serialize an int64 to JSON with protocol buffers, what you get as a string. And that's because many JSON runtimes treat numbers as floats, and so there isn't 64 bits of space available, so the only safe way to send them around as a string. This is usually not -- it's not a major concern, because the other end of the transaction is also using protocol buffers. And so it's parsing the string automatically and turning it into an int on the other end. + +You have to deal with the protobuf toolchain, which often has some rough edges... And we can talk about all that stuff in detail if you want... But especially in Go, I don't think the trade-offs are that -- you're getting a lot for free, and you're only giving up a little bit. + +**Jon Calhoun:** When you mentioned the trade-off of not being able to have the more flexible API, I guess, I also view that as a win. It's kind of a trade-off, but also a win, because I've also worked with APIs where trying to write a library in Go that works with it is very challenging, because you can tell the API was written in Python, where a certain field can be one of six types... And it's really annoying to parse, because it's like, I've got to figure out which one it is before I even try to parse it. + +**Akshay Shah:** \[10:07\] That's right. Yeah. One of my co-workers often describes protobuf as static typing for data. And it has basically the same trade-offs as static typing for programming languages. You have to declare your types upfront, changing them can be a little bit harder - you have to think about forward and backward compatibility - but you get APIs that are self-documenting in a nice way. Often, it's easier for your IDE to do work for you, it's easier for tools to catch errors... So I think this is a familiar debate, and in general, in 2022, it seems like big parts of the industry are moving back towards static typing. Right? TypeScript is super-popular, Python is growing Type Hints. Ruby is growing gradual typing via a bunch of Stripe stuff... And Rust is taking type systems into areas that many other languages have avoided. + +**Jon Calhoun:** This kind of reminds me of -- there's a couple of tools around JSON APIs where you basically define a schema and it helps generate libraries for various languages... I think - was Swagger one of those? + +**Akshay Shah:** Swagger is a little bit more extensive. Swagger -- at least as we've talked about them, right? Protocol buffers are just about structs. They're about data shapes. + +**Jon Calhoun:** Okay. + +**Akshay Shah:** Swagger covers that, but it also covers the shape of your REST API. Maybe this whole show is gonna be an unpopular opinion, but my view of like -- if you look at a Swagger schema, or you look at a JSON schema, you look at it and you're like "Oh my God, am I really supposed to be writing this by hand?" It's so verbose. It's just scads of nested JSON. + +**Jon Calhoun:** I haven't used Swagger a lot, so I can't speak to that... But I can definitely say, I've seen a lot of companies that were writing APIs that needed -- basically, it's a public API, so they wanted to provide libraries for various languages... And you would think Swagger was a perfect fit, but because for one reason or another they opted not to use Swagger, and I think a lot of times it probably was, like you were describing, it just wasn't very appealing to learn and use Swagger when they looked at it. I'm wondering how you compare protocol buffers to something Swagger, since I think a lot of people have probably heard of Swagger, but I'm not positive... + +**Akshay Shah:** I'm not sure. Swagger, the new name is OpenAPI; there have been a couple of version, so they're on OpenAPI v3 now... And it includes most of JSON schema, which is the data struct part of the language. We can talk about the network API parts of it later... I think it's a better fit for comparing to gRPC. But for the data parts of it, it has a really impoverished type system. So the only native types in JSON API are the types that JavaScript supports. So you have array, string, number object. That's it. That means that everything you build on top of that ends up not feeling that great. There's really no way to say "This is a date." The best you can do is say "This is my handwritten object type that has these numbers affiliated with it", or "This is a string, and here's a regex that validates that it's ISO 8601." And it's just really, really verbose. If you're talking about saying like "I could write a Go struct with some struct tags, or I could write a JSON schema", writing the JSON schema is a ton more work. So there's this little cottage industry -- I mean, I guess this very large industry of GUIs to write these schemas, because they're so laborious. + +**Johnny Boursiquot:** See, now I have to buy another tool, and learn how to use that tool, and maybe it has its own quirks... So yeah, for vendors it's great. They're like "Yeah, we'll sell you something to do all this for you." But yeah, now I have another dependency. + +**Akshay Shah:** \[14:01\] Exactly. And if you look at the specification, there are parts of this language that 0 I understand why in a really loosely federated world of the web, some of these things make sense. But if you're a company or a person publishing a schema, they're a little off-putting to me as a Go programmer. They're like "Oh, I can include a reference to another schema, on a different server, and just transclude it into my schema." There's a whole separate part of the spec about meta schemas. So there's some form of metaprogramming in this schema language. There's a special call-out when you're implementing codegen for these things, especially at runtime, that schemas can mutually reference each other. So you have to take special care to break cycles and avoid infinite recursion. These are just not the kind of problems that I want to think about when I'm defining a struct. I want this whole class of problems to be impossible. Generally, I want this world to be as simple and predictable as it can be. + +Protobuf is much more on that side of the world... Part of why I think that's so appealing in Go is that protocol buffers in Go share a lot of DNA from Google. A protobuf message looks a lot a Go struct. And so if you are a Go developer and you're thinking, "Which of these should I do?", the amount of effort it takes to write a protobuf schema is about what you would spend on a Go struct anyways. You get a rich type system; there's sized ints, fixed size, variable size, bytes, strings, objects that are well-known types that get shared across the whole ecosystem, for like durations, dates... There are a bunch of escape hatches if you need them for loosely-typed data. + +So John, for the situation you are describing, you can say, "This field is one of the following types." And that generates code that's a little bit awkward, kind of unavoidably... But you can express this. You can also express the idea like "This field is just a thing. And I have no idea what it is, we'll find out at runtime." Those things are just, I think, appropriately irritating to use... It kind of gives you a hint that you're going down the wrong road. + +**Jon Calhoun:** Yeah. So if we're looking at protocol buffers, I think it's commonly compared with things JSON, or maybe GraphQL... Would you compare it to GraphQL? + +**Akshay Shah:** Mm-hm. + +**Jon Calhoun:** Okay. So if we're looking at those things and we're trying to expose something to, let's say, the public; we wanted to build an API that has public consumers not just internally, are protocol buffers a good fit for that type of situation? + +**Akshay Shah:** I think they are, with one caveat. I think protobuf is simple, especially if you use protobuf schemas to accept and send JSON. This is really easy for other people to use, because they don't really have to know about the protobuf part at all. If they would to continue handwriting classes to generate JSON, they're more than welcome to do so. But they have this kind of efficient binary protocol available to them. Usually, when you talk about exposing protobuf APIs, you're talking about gRPC. Right? Because protobuf the language is really just about data. It doesn't have anything like an interface, or a function, say like "My API does something" without gRPC. + +The one caveat is that historically, the tools to work with protocol buffers are kind of rough. They're open sourced kind of directly from the way Google uses them... And within Google, they're part of this really sophisticated, unified build system, and monorepo, and all this other stuff. So the protobuf tools are this really low-level component in a very complicated stack. + +\[17:49\] Outside of Google - well, we just have the protobuf piece, but not the rest of the stack. And so if you're building an external-facing API, and you're working in Go, and you've got all this protobuf stuff figured out, but your client is like a Node client, and you go to them and say, "Well, step one is you need to write a makefile, and you need to call protoc in the following ways", they look at you like you're crazy. Like, "What? A makefile? What, are we living in 1986? What are you talking about?" + +So I think there's a lot of space to build tools that make that easier and more approachable. There's no reason why, given a protobuf schema, your instructions to the client can't be "Hey, don't worry about any of the protobuf stuff. From your perspective, I handwrote a client for you, and it's an npm package. Just download that." And on your end, you're like "Yeah, I autogenerated the package, and I uploaded, and I didn't handwrite anything." But to your colors, it should look like they're just getting a package. + +**Johnny Boursiquot:** So we've talked about the efficiency of not having to hand-wrangle all these things, but we haven't talked really about -- we've mentioned it, but never really talked about sort of the efficiency of transport, right? Basically, when you have that binary format; can you sort of clarify what the major gains of the binary format are over traditional means? + +**Akshay Shah:** Yeah. Typically, binary formats have a bunch of optimizations available to them, that text-based formats could technically use, but rarely do. So if you imagine how a computer is parsing a JSON object, usually you're kind of like advancing one rune at a time, and then maintaining some lookback, and you're looking for a bunch of object delimiters and keeping state for how deeply invested in this object you are... And it's just kind of this complicated stateful process. Binary protocols typically work differently. It'll say, "Hey, the next field coming up is a string, and it's 70 bytes long." And then the parser just grabs the next 70 bytes and interprets them as a string in memory, and is done. There's no looking for delimiters, there's no escaping, there's none of this stuff that makes JSON complicated from a parser level. + +And typically, protobuf goes out of its way to try and make that binary small. So it does a bunch of tricks to minimize the size of things. Ultimately, these are all really clever tricks, but JSON is also really widely used. You can make JSON incredibly fast. There's this CS professor in Canada named Daniel Amir, who did a bunch of research in using SIMD instructions to parse JSON... And he has a commodity computer parsing gigabytes of JSON per second. I mean, it's absurd. And compression is really effective; even protobuf, the binary format benefits from being compressed. It's like, once you're gzipping the JSON and ungzipping the binary, the difference gets a lot smaller. + +So I think once you're using a protobuf schema, you have this binary format at your fingertips, and so you might as well use it. I mean, even if the practical perf improvement for your particular use case is 20%... 20% is kind of a lot. You might as well just grab it and get that benefit if it's sitting right in front of you. But I don't know that that benefit is really so significant when all is said and done... Like, that it's really the motivating use case for protocol buffers. Does that make sense, Johnny, or was that just a bunch of noise? + +**Johnny Boursiquot:** No, no, that makes sense. So while you were talking, it occurred to me, or I just recalled that the Go standard library has an encoding mechanism which supports binary transmitting... + +**Akshay Shah:** Yes. Marshal Binary? + +**Johnny Boursiquot:** Yes, there is that, but I'm also thinking of the Gob package. + +**Akshay Shah:** Yes. + +**Johnny Boursiquot:** Let's talk about the Gob package. Why is it that we're not all using Gob? Is it that we're not all Go developers? \[laughs\] + +**Akshay Shah:** \[21:57\] I mean, that's partly it, right? But the Gob package - my recollection, at least, is that the Gob package makes some very important -- there's some very important caveats in the package docs for Gob. One of them is that - this is just my recollection; I can check real quick, but... From what I remember, the binary Gob representation is not guaranteed to be stable across Go versions. So forget interop with JavaScript, if you're on Go 1.19, interop with 1.17 might be busted, too. I think it is also not specifically optimized for speed, or size. It's just not that widely used, right? But just by the nature of the business buff does, we spend a lot of time with protocol buffers, and talking to various people deep in protobuf at Google... At least what I've been told is that there's so much protobuf from top to bottom in Google's internal stack that relatively modest perf improvements to protobuf can change Google's overall CPU use by 3% or 4%. + +Protobuf is language-agnostic, it has a specification that's public, and it has just like a tremendous number of miles put on it. And I don't think you get any of that from Gob. + +**Johnny Boursiquot:** Yeah, fair enough. + +**Akshay Shah:** Gob is super-convenient though... + +**Johnny Boursiquot:** It is. + +**Akshay Shah:** You can just marshal a thing, and that's really nice... I would love it if you could do that with protocol buffers, but you just can't. + +**Johnny Boursiquot:** You need Go on either side of the fence, obviously... That is one of the drawbacks, right? You don't have that cross-language support unless everybody starts implementing a Gob encoder and decoder and whatnot. + +**Akshay Shah:** And parsing a Gob... The schema is a struct, right? So parsing that requires parsing Gob, right? Which is kind of a tall order. It's not totally obvious, to me at least... I think it would be challenging to write a tool, for example, that looks at a dif between two Go files, and says, "Hey, stop right there. You broke your clients. If you deploy this change, all your existing clients are going to be broken." I'm sure it's doable, but it's not -- + +**Johnny Boursiquot:** Trivial. + +**Akshay Shah:** ...it's not quite as straightforward as it is with a simpler, more purpose-built thing like protobuf. + +**Jon Calhoun:** This conversation also reminded me of -- at one point in the past we talked about the JSON parser in the standard library, and it's not as fast to some third-party JSON parsers because, if I recall correctly, the JSON parser has to read the entire JSON object to verify that it's valid JSON first, and then it actually goes back to parsing it a second time... + +**Akshay Shah:** I believe that's right. + +**Jon Calhoun:** I'd have to go back and check though... But as a result, it means that it's not like the most optimal parser, whereas like when you mentioned the binary format, it's just assuming this is valid data, "We're gonna go ahead and use it." But it's worth noting that for most people that optimization isn't so important that they've replaced the standard library with a custom JSON parser of some sort. + +**Akshay Shah:** That's right. And you can get -- I mean, to be fair, as the author of a bizzaro custom JSON encoder... \[laughter\] It's not that hard to encode JSON much, much faster. Zap, the logging packager, part of why it's faster is that it has its own JSON encoder. And that's pretty easy, because JSON is a really simple string format. So even if you want to make JSON really fast, you can do that by just biting off half the problem. If you make encoding super-fast, and decoding is still not that fast - that's 50% of the win for 10% of the effort. + +**Break:** \[25:52\] + +**Jon Calhoun:** So we've been talking about protocol buffers, which are how we're going to format the data. Generally, when this conversation comes up, you don't hear about protocol buffers on their own, you hear about protocol buffers and gRPC. So what is the gRPC part of this? + +**Akshay Shah:** So -- I mean, let's go back to our "We're building a REST API" example. So we're writing structs for our requests and responses. Other than implementing our business logic, what's the other work we're doing? Basically, what we're doing is we're designing a routing scheme. So you're saying, "Okay, I have some function that's going to get a list of users." And maybe the users are scoped to one organization, and I can pass some filters to the query. How do I represent that in a REST API? I sit down and I say, okay, this is a read, so it's probably gonna be a get, and then the organization ID, with most REST APIs that I've seen - that's going to go in the URL. So it's going to be get/org/1/users. And then I'm going to take my filter parameters, and I'm going to put them in as query params. And now I'm in hand-serializing territory, so I need to decide how to represent every type that I accept as a filter as a string, in query parameters. If any of those parameters are arrays, we just start praying. No one knows how to represent an array in a query parameter. Do we use commas? Do I just pass the parameter 18 times? Do I just give up and serialize it to JSON, and then URL-escape it and shove that in the query params? No one knows. + +These are all really low-rent decisions. In practice, I think, basically nobody cares. I just wanted to get my data, and most of my clients, if they're looking at any of this, they're upset. Like, "I just wanted you to hand me code to call this API." And these plumbing details are not all that interesting. So what protobuf does is it -- just it has an equivalent of Go structs, it has an equivalent of an interface. You give it a name, you write a curly brace, and then you write a function name with some input parameters and output parameters. + +The inputs and outputs are all protobuf messages, and so in this world, I would write an interface and I would call it User Service, or User API. And it would have a function called get users. And that function would take a protobuf message with nice, strongly-typed arguments; I could have the organization ID in there, I could have any number of array types, or maps, or whatever I like. And all these routing and serialization decisions get made for me, in a predictable way, and I get nice, generated code, where all I need to do as the server owner is implement that Go interface. + +So I need to write a struct with a get users method. And for the client, because all of this is really regular and predictable, it is easy to generate good code. What's the alternative? The alternative is something like Swagger. And Swagger, or OpenAPI, they are designed to be able to describe any REST API. And so you don't actually avoid any of these decisions. You're just -- you're in the same tedium of plumbing decisions, but instead of writing them using the net HTTP APIs, you're writing them in this big, verbose JSON document. And because the schema supports any way of routing, it's really hard to generate good code... Because there's this gigantic universe of options, and that's really difficult to cope with. + +As a user, I think, especially a Go programmer, protobuf feels really familiar. I write some structs, I write some interfaces, I run a code generator, and then as a server, I'm just implementing an interface that looks just the schema I wrote. Life is pretty simple. And then you use an RPC runtime, like gRPC, you hand it your interface, and it starts serving HTTP traffic for you. + +**Johnny Boursiquot:** Are there alternative runtimes other than gRPC? + +**Akshay Shah:** \[32:03\] Yes. Actually, from the episode on bloat, from a couple of months ago, Egon's company, Storj, it sounds like from their public blog post that they were using gRPC, and they were dissatisfied with a bunch of things about it... And so they wrote an alternate RPC runtime. It's called dRPC, and it basically operates at a lower level in the networking stack. But to a programmer, it's very similar. You write some protobuf definitions, you generate some code, you implement an interface, and then you hand that interface over to a networking package to serve traffic. It's incompatible with gRPC... ish, it's a little complicated, but it's operating directly at a TCP level. So it's not serving you HTTP traffic. But again, it's a program that looks pretty similar. Twitch has a similar thing called Twirp. Thrift is very similar to protocol buffers, and it has its own wire format. There's Dubbo, from I think Alibaba, just similar in principle. + +**Johnny Boursiquot:** Okay. + +**Jon Calhoun:** It's interesting to me that this is a problem that has -- there have been enough solutions to this problem that I think it makes it pretty clear that developers are like "Hey, this is a struggle." Even an example I can think of is - Mat was supposed to join us, but he couldn't... But Mat, whenever he was building Pace, created a library called Oto, that you would basically create Go -- I think it was Go structs and interfaces, and it would generate a JavaScript library that would connect to, and then it would generate like a Go server, and then you would sort of just plug in the rest of the code. + +**Akshay Shah:** Yep. + +**Jon Calhoun:** It was meant to be relatively simplistic, I think, and just support their needs, but it's interesting to me that even in those cases, people will go out of their way to build something on their own, when other solutions like gRPC do exist. So I guess my question to this is, why do you think people invent a new, simpler version, when something gRPC exists? Is it the complexity of gRPC? Is it something else? + +**Akshay Shah:** I don't know. I mean, have you guys worked with gRPC? + +**Jon Calhoun:** I have not that much. + +**Akshay Shah:** Johnny's smiling. I think the answer is probably yes. + +**Johnny Boursiquot:** I have. \[laughs\] + +**Akshay Shah:** So Johnny, you tell me - are you whipping out gRPC Go for your personal projects? + +**Johnny Boursiquot:** I've kind of been to the place where -- I had to make that decision where I was like... Basically, the trade-off I was making in my head is "Look, I can go and try to wrangle something myself, or I can take something that is off the shelf and just accept whatever trade-offs that presents, but just get my job done." So I'm not gonna be providing myself a new runtime, or a new protocol, or a new whatever it is, because that's not the job, right? It's like saying, "Oh, I need to write more. I need to publish blog posts more. Let me go write a blog engine first." Right? + +**Akshay Shah:** Yeah. I guess start by writing a markdown parser, Johnny. It's the only way to start. + +**Johnny Boursiquot:** \[laughs\] I mean... If I just want to play around, and for educational purposes I just want to reverse-engineer something or build my own, just so I can educate myself and know how something works - yeah, that's fine. But if I'm trying to ship something? Heck no. I'm gonna go with what most people are using, I'm gonna pick something off the shelf, or go gRPC, whatever it is, and I know that's gonna work for my use cases, and call it a day. So it depends on what it is I'm trying to do. + +**Akshay Shah:** I think that's right. Yeah. I think a lot of people find themselves in that situation, right? There's only so many problems in the world that you're interested enough in to really peek under the covers and start digging around or writing your own stuff. I've never met Mat, or talked to him, but just based on the fact that he gave a GopherCon talk about how he built HTTP APIs - he has opinions here. And I assume wrote Oto to make those opinions easier to apply across a big codebase at Pace. + +**Jon Calhoun:** I would imagine another aspect of this is that because it's a startup, when you release open source libraries and things that, it's kind of marketing... I think sometimes people ignore that aspect of like there needs to be marketing with a new business of any sort... So tech blog posts are a great way to do that, and sometimes that means you actually have to do something to write about... + +**Johnny Boursiquot:** You're accusing Mat of being a marketer... \[laughs\] And not a developer... + +**Akshay Shah:** \[36:06\] Let's not wave that brush around too wildly, okay...? + +**Jon Calhoun:** I mean, whether or not Mat was trying to, I can say for certain that he knows how to market to a degree, because his blog posts, whenever he starts a new company and things that, are all great marketing tools to help build a business. You have to do those things. If you just build something in isolation, and nobody knows about it, then it's really hard to find those users. + +So I'm not saying that was his whole reasoning for it, but I'm saying that definitely could have played a role in it, is "Hey, this is something I'm interested in, it's pretty cool, and we can build it pretty quickly and use it as a marketing tool." So I think that's another factor to consider. + +**Akshay Shah:** I think there's -- you know, gRPC grew out of this internal system at Google that's called Stubby. And Google is protobuf all the way down. And internally, it's all run on this Stubby thing. And Stubby is not on top of HTTP, Stubby is like a competitor to HTTP. So it's protobuf-flavored HTTP/2, kind of. And it worked super-well. Word on the street is that the tight loops in Stubby are handwritten x86, written by Turing award-winning programmers, and stuff, right? Stubby is amazing. It's great. Perfect. Eventually, Google wants to write an API that somebody else is going to call, or they want to hand you a client to bid on ads. And they just can't hand you this Stubby thing, because you look at it and you're like "What is this? How do I write a load balancer for this? What even is any of this?" And so they're going like "Oh, okay..." The rest of the world kind of wants to HTTP. We have this robust infrastructure of load balancers, and proxies, and caching, and all this stuff... And I have a lot of client libraries, and a lot of servers, and so I want HTTP. But they want something that's semantically kind of easy for them to bridge their inside world and their outside world. So they need protobuf-flavored HTTP. + +There were a bunch of ex-Googlers who had invented these things at other companies, but Google wanted the one that they thought was the right thing, and so they started a team and they built gRPC. gRPC is basically used at the boundary of GCP, but it's not running the internals of Google. It's this sort of bridging technology, and then it's for open source use. But it has this aura of mystique about it, where you're like "Oh, gRPC." It's, it's fancy, it's complicated, it's this big, big codebase... And so there's only -- in Go there's one gRPC implementation, it's gRPC Go. And in Java, there's basically one implementation, written by Google. + +The protocol is pretty simple. You could describe it in English in a couple of paragraphs. It's basically, given a protobuf schema, here's how you figure out the path, here's a couple of headers you have to set, and then here's how you take the bytes of protobuf and shove them into the body. That's it. It's not conceptually so different from REST. It's not hard to write a gRPC handler or client from scratch. + +I gave a GopherCon lightning talk just a couple months ago, we wrote an HTTP handler from scratch using the standard lib, that speaks gRPC in seven minutes. It's one slide of code. And so just like there are a million REST packages, and a million flag-parsing libraries, and a million different JSON libraries, there's room in the world for more than one gRPC implementation, that makes different trade-offs. + +I think just like you look around and like you're building an app, a CLI, and at some point you looked around and you said "Ah, am I a purist? Am I a standard lib flags kind of person? Am I a Cobra person? Or am I way out in left field and I'm using--", what's that company? "...charm bracelet, and glow, and likes fancy, interactive CLI stuff?" I personally am more of a standard lib flags kind of person. + +**Johnny Boursiquot:** It does the job, right? + +**Akshay Shah:** \[40:15\] It does the job, and I it's small, and I kind of learned it once, and then I just decided that I didn't really care that much about this problem anymore. The same thing is true of Go. There's a Go specification, there's the standard Go compiler toolchain, but there's also Gccgo, and there's TinyGo, and they make different trade-offs that are useful to people. There's HTTP, which has a bunch of RFCs to define the spec. It's a big, complicated spec. And there are a bunch of implementations, right? There's NGINX, there's Apache, there's net HTTP in Go... But there's also Fast HTTP, which makes different trade-offs, and there's implementations in other languages too, right? + +We didn't write NGINX and then just say "I don't know, everyone else just FFI into nghttp. Good luck." There's space for a gRPC implementation that maybe meets people who are writing rest APIs today where they are, instead of making trade-offs that are appropriate for a Google maybe, but not so appropriate for Pace. + +**Jon Calhoun:** So I suppose this is a good time to say that is what you guys are working on, correct? At Buf. + +**Akshay Shah:** Buf's working on a lot of things. I think I've alluded before, it's the idea that the tools for protocol buffers in open source are kind of rough. They're pretty low-level, and they're a little bit -- they have a big barrier to entry. So most of all, we're working on our protobuf tools. So Buf writes its own protobuf compiler from scratch. We think that's a lot easier to use than Protoc. + +We also work on a schema registry. One of the important things about protocol buffers is that if you're using the binary format, you can't do anything with the data without the schema. So JSON - you can just shove it around, you can shove it into a Kafka topic, read it out the other end, and you're good to go. You don't really need any other information. For protobuf, when you slurp out the binary from Kafka, you need a schema, otherwise you don't know what any of it means. It's not self-describing. So we've built a schema registry that lets you push and pull schemas, it also handles codegen... So if I write an API, and I define it in protocol buffers, and I push my schema to the Buf schema registry, any client can just npm-install a package from the registry automatically, without any effort from the server or the client. Or they can pip-install, or go get a package with a ready to go client. + +That's the business, is selling the schema registry. Our view is that, like you said, John, gRPC and protobuf are intimately connected. And for people to be really excited about using protobuf for everything, they kind of also have to be excited about using gRPC for a lot of things. gRPC does a lot of things super-duper well. If you are very concerned in your REST API about having excellent HTTP/2 flow control between Australia and Brazil, gRPC has your back. It's really good at stuff that. It's not so good at just being interoperable with the rest of Go. It has its own HTTP server, its own HTTP stack, it's not compatible with net HTTP... It's really big. You just can't serve gRPC traffic alongside your website, or your REST API, or some convenient HTTP API, like download a file, just because they wrote their own HTTP transport. + +So at Buf we work on this thing called Connect. It's a drop-in gRPC replacement. It's wire-compatible, works with every gRPC client, and it's all net HTTP. It generates HTTP dot handlers, clients use HTTP dot client, and it works with any mux, or any middleware package that works with net HTTP. So I think, to a Go programmer, it feels a lot more like, rather than a whole different world, where all of a sudden it doesn't even really feel HTTP anymore, it feels more like someone generated the boring REST code for you, but if slots into the same ecosystem you're familiar with. If you know net HTTP, if you know middleware that you like, like some gzipping handler, or something else, if you have a router that you really like, like you're into Chai, or... + +**Johnny Boursiquot:** \[44:40\] Gin... + +**Akshay Shah:** ...Julien Schmidt's HttpRouter, Gin or whatever. + +**Johnny Boursiquot:** gorilla/mux right? + +**Akshay Shah:** Yeah, any of those things, right? This just slots right in there. So your gRPC handlers plug in right alongside all your other ones. I think there's space for that, and I think it's nice. It's a tiny bit slower, just because it supports much more of HTTP than the gRPC stuff does. I don't think most of us look at net HTTP and we're like "Oh, that code's for chumps. So slow! Unusable in production!" + +**Jon Calhoun:** Have you seen some of the new Go releases? + +**Akshay Shah:** \[laughs\] + +**Jon Calhoun:** When people set up a Hello World server and they're like "Oh, it's 40% slower." + +**Akshay Shah:** It's just -- how could they...? I mean, it's the same protobuf runtime and net HTTP, right? So generally speaking, perf-wise, you're in the ballpark that you're familiar with, and that's proven itself to be good for this huge variety of use cases. + +We do the same thing for TypeScript, and we're working on a similar runtime for mobile phones. I think over there, the standard gRPC protobuf trade-offs are a little more out of the norm, and so there's more space to do something that feels really good to web developers or mobile developers. + +**Johnny Boursiquot:** So you mentioned all sort of statically-typed languages for this. I imagine that's deliberate; you're avoiding some of the more dynamic stuff. + +**Akshay Shah:** I think even in dynamically typed languages, right? Like, Python is a great example. The Python ecosystem, if you look at some of the most interesting stuff in Fast API, it's generating typeheads that represent your APIs nicely. And so even a lot of languages that are really dynamically-typed, it's convenient for you as a programmer, where you can, to have good type information at the boundaries of different modules in code. And inside, if you want your business logic to be all object or dictionary, that's fine. That's your choice. + +**Johnny Boursiquot:** That's rough. That's a rough choice, but alright, keep going... \[laughs\] + +**Akshay Shah:** I'm actually less opposed to it than a lot of Gophers... There's a time and a place for everything, right? + +**Johnny Boursiquot:** Right, right. + +**Akshay Shah:** If I'm in a Jupyter Notebook, and I'm fetching some data from some service just to run a bunch of NumPy on it, whatever is in between this API call and my NumPy array is really - like, I don't care that much. I just want it to happen. I don't really want to be in Gonum, trying to dimension an array, and decide whether I'm looking for int64s or int32s. I want numbers, and that's about all I know. + +**Jon Calhoun:** So when we're talking about all of this, I will say that part of the reason I don't have a lot of experience with gRPC is that historically, you see things like -- it's you hear about Stubby, or something at Google, or a custom HTTP implementation for the gRPC stuff for Go... And it always struck me as something that was used more for internal communications inside of your application, and not really a user, like, something you made user-visible. So that was the mindset I always had. I guess what it sounds like is that was a very invalid belief or conception... Or would you say that's generally still somewhat true? + +**Akshay Shah:** I don't know. Johnny, what do you think? I have my opinions, but... + +**Johnny Boursiquot:** \[47:53\] I think people will go with what they know best to get a job done, and will only sort of step outside -- well, let's just say there are two classes of developers out there. There are those who go with what they know, and try to get the job done as quickly as possible, and there are those who look for opportunities to bring in new things, even when it's not necessarily a requirement to solve the problem itself. I've been on either side over the years. I'm not sure -- Jon sounds more like he's the "I'm gonna go with what I know, because I have a job to do, and I need to get paid, and move on to the next thing." + +**Jon Calhoun:** Yeah, but when you're self-employed, I feel like that... + +**Johnny Boursiquot:** \[laughs\] + +**Jon Calhoun:** ...self-employed with kids, you're like "I don't have time to learn new stuff right now." + +**Johnny Boursiquot:** Right. Exactly, exactly. And then you have those who work at larger companies, and they don't even have to worry about getting paid to feed their family, and they can take more liberties, I would say. + +**Akshay Shah:** I think that's right. + +**Jon Calhoun:** So if you were building a publicly -- I guess what I'm wondering is if you were building an API that you wanted to expose to the public, for whatever reason, is gRPC something you would consider using that, if you knew gRPC really well, and were good with it? Or would you say that you'd prefer just to build a JSON API in a more - I don't want to say traditional way, but I guess what people are used to seeing? + +**Akshay Shah:** I would use Connect, and I would build an API using protobuf schemas for my own development; I would tell most of my clients, "Hey, if you're trying to call my API from JavaScript - like, forget about all this other stuff; it's HTTP and JSON under the hood. Just download this client library. Why do you care what's going on inside? I'm going to jump you right to the thing you actually want, which is code." Same for Python, same for Go. And I'll say "If you want to curl my API, or you have some ad-hoc thing, or you're writing code in, I don't know, Zig, and I don't have a library off-the-shelf for you, it's really simple... Like, here's what the JSON looks like. And you just post it to this API. This path, post some JSON, you'll get some JSON back... That's it." + +**Jon Calhoun:** It sounds like you would use gRPC, but you would just make sure you're using that JSON format so that if they need to do something else, they can pretty easily. + +**Johnny Boursiquot:** There's some subtlety there. Connect supports the whole gRPC protocol. It also supports its own protocol, that is very similar to gRPC. Your code stays the same, you flop a config flag; you can speak your own -- this other protocol, that looks much more like REST, where you can just post some JSON. And this is actually not feasible with off-the-shelf gRPC stuff. By default, gRPC Go or gRPC Java, they also don't -- they can't communicate with a web browser. Just generally a big limitation for an external API. Kind of opens that up. Twitch did the same thing; one of their priorities was being able to talk to any HTTP, client anywhere. So I would use Connect, I would use protocol buffers, and I would tell clients kind of in back-endy languages to use Connect-generated code or gRPC-generated code. And I would say if you want to curl this, or call it from a browser or something, you might have code gen available, or you might just be posting some JSON, just you would post JSON to a REST API. + +**Jon Calhoun:** Okay. I think that helps, because - I know historically, the fact that you said that you couldn't communicate with an HTTP browser, the minute I hear that, I'm like "Oh, this isn't gonna work." That's kind of a showstopper... Especially if you're building a web application, and you're like "Well, that's my one use case, so..." + +**Akshay Shah:** I mean, even if you're not building a web application, right? You're building an API... Browsers are really convenient. The Network tab is really nice, it's a convenient debugging environment, it's got this real programming language \[unintelligible 00:51:41.24\] for you to plan... It's the most widely deployed HTTP client in the world... Why wouldn't your HTTP protocol support it? It's a historical misstep in the gRPC protocol. + +\[51:55\] I think gRPC uses this little-known corner of HTTP called trailers. They're just headers that come after the body. They're useful for a bunch of reasons, like having some way to send metadata after the body is really helpful. They chose to send them as HTTP trailers. At the time they were making these decisions, it looked browsers were gonna support trailers, so they kind of decided that they were going to probably do it, but hadn't actually done it yet. And then as soon as any browser vendor got involved, they all said, "Absolutely not, we're never doing this", and a bunch of other HTTP software never supported trailers. They've been around since the late '90s, and basically nothing ever supported them. So if you have a Rails app, you're not serving up trailers any day soon... + +So that's kind of really made gRPC hard to adopt for external APIs. But that's not an intrinsic problem with protocol buffers, it just means you need a little translation layer, or you need a different library. Those libraries are pretty small; Connect does all of gRPC and this other thing in less than 10,000 lines of Go. It's like 20x smaller than gRPC Go. This is all totally doable if you're just optimizing for something different. So in short - yeah, I would say that you should use protocol buffers for your external APIs. + +**Johnny Boursiquot:** Yeah, Jon. + +**Jon Calhoun:** It's definitely something I'm open to exploring... It's kind of the same as GraphQL - it's always struck me as interesting technology, but when I don't have a direct need for it, it's really hard to go out of my way and use it extensively to build some -- because people are like "Oh, well, your API is changing." I'm like "That's not happening to me like it's happening to Facebook." Our needs are very different. + +**Akshay Shah:** When you write your external APIs, when you're calling them from your web apps, are you typically handwriting code in the browser to call those APIs? So you're writing like React hooks, or whatever? + +**Jon Calhoun:** If I can generate a library or do something that, I always opt to do that, because it's just simpler. But it really depends. One of the restraints I get at times is that I also make courses that help people learn stuff... So when I'm looking at things, that can also influence what technologies I opt to use. Because when you rely on third parties, and all of a sudden they break or they change drastically, then all of a sudden that material is useless. So sometimes you're like "Well, if I can stick with just manually using fetch, and doing this, then maybe that's going to stand the test of time a lot better." And if it explains sort of how something standard works to somebody, that's also helpful. So it really just depends on the context for that. + +**Akshay Shah:** That makes sense. + +**Jon Calhoun:** But most of the time, if I can generate a library, that's what I want to do. I don't care about the underlying technology at all... Whether it's XML or something else, I don't really care. I just want -- like, give me a library that's easy to use. + +**Johnny Boursiquot:** XML, Jon? Really? + +**Jon Calhoun:** Johnny, I've used way more XML in my life than I really care to. \[laughter\] + +**Akshay Shah:** That's right. Oh, man. At one job... We're doing a bunch of work, enterprise integration work with ThinServe companies, and I was praying for XML. There were all these weirdo, handrolled binary formats... It was like "Back in the day, we thought these ints were going to be 32 bits... So if this flag is set, jump to the end of the file and look for where we added another 32 bits for the top bits, because the numbers got too big", or something. + +**Johnny Boursiquot:** Wow. + +**Akshay Shah:** It was all bananas. So I can see why everyone was really excited about XML. + +**Jon Calhoun:** I was never excited about XML. It was just one of those things that - I started a company where we interacted with a lot of shipping APIs, and a lot of them were using XML, so I just got very familiar with it... And at the end of the day, I just came to the conclusion that I don't actually care what your API uses, as long as there's a good way, a good library or something to communicate with it, and as long as it gets simplified in that sense. And I think most developers have that same viewpoint of "If you give me a good library, I never look to see what you're even using behind the scenes, because it doesn't matter to me." But if there's not a library, then clearly I have to look and see what you're using, and then if it's XML, I'm going to be like "What is your problem?" + +**Johnny Boursiquot:** \[laughs\] No, that's when you dust off the old XSLT book. + +**Akshay Shah:** That's right. \[laughs\] + +**Johnny Boursiquot:** ...that you've got buried somewhere in your basement. + +**Akshay Shah:** I had that in an O'Reilly book, and there was like a snake woodcut on the cover, or something... Oh, my God. + +**Break:** \[56:25\] + +**Akshay Shah:** I think in 2022, the way the industry is today, protobuf is a good middleground for exactly that role... You write a little schema, your server-side implementation gets easier. It's a little easier to wrangle. And you can just hand your clients fully-generated, ready-to-go client code. And whatever is happening in between - it's reasonable, but it's not artisanally handcrafted hypertext as the engine of application state. This is not Roy Fielding's thesis brought to life. It's just workaday code that gets the job done, and is pragmatic, and pretty reasonable, and is ultimately just some boring plumbing. + +We've all got jobs to do... Your server is supposed to be doing something, and we just want to get to that part of it as quickly as possible. And on the frontend, you're like "Well, I'm trying to build a UI here. I don't really care about any of the plumbing. I just wanted some functions to call." And the faster we can get the server author and the client to calling functions and not worrying about the plumbing, the better off we all are. + +**Jon Calhoun:** So I think we're getting near the end of the episode... We're gonna move into the Unpopular Opinion. + +**Jingle:** \[58:56\] + +**Jon Calhoun:** Akshay, do you have an unpopular opinion for us? + +**Akshay Shah:** I don't know how unpopular it's going to be, but we mentioned at the top that I wrote this structured logging library called Zap. There's a bunch of libraries like it now; there's ZeroLog... And they're proposing to introduce something this to the standard lib. I think this whole thing is just so fundamentally misguided. I have felt that since starting to write Zap. I just think it's all a bad idea. + +**Johnny Boursiquot:** So you think structured logging itself is a bad idea. + +**Akshay Shah:** The way it's done here, yeah. + +**Jon Calhoun:** We've introduced you by saying that if you've used Zap, you've probably used your code, and now you're telling us that that was a bad idea. + +**Johnny Boursiquot:** \[laughs\] + +**Akshay Shah:** You know, for a bunch of reasons... I was at Uber at the time, I was writing a service mesh thing for a Stubby-like protocol... It had tight performance requirements... And we were feeling really good. We had written this service mesh proxy thing, we'd met all the perf requirements... And Johnny, you're gonna hate this, but at the end we came back for like "Now we shall add the observability." + +**Johnny Boursiquot:** \[laughs\] + +**Akshay Shah:** And you add a couple of log statements and a couple of metric increments, and all of a sudden I'm staring at 1,000 heap allocations. The perf budget went out the window, and you're like "Well, we could just ship it without the logs or the metrics, but that's not good..." + +And so to fit with the log infrastructure of the company, we had to emit JSON. And there was really no facility for running a regex over a string in the log ingestion and dashboarding stuff. And so I kind of invented this ultimately very fussy API for producing logs, just to avoid heap allocations, more or less, and to avoid, re-serializing data that we'd already serialized once. And it is so fussy... I mean, it's just unreal. + +**Johnny Boursiquot:** And you thought you'd unleash it on the rest of the community... \[laughs\] + +**Akshay Shah:** No -- so actually, internally at Uber what we did is we just put this thing behind a facade that looks like the old logging library... And it was like 30% faster, and we're like If everything just got 30% faster for no code changes, this is a huge win." Right? There's zillions of cores just parsing map string any. If we can get rid of that, that's great. But I didn't expect anyone, apart from the lowest level infra code, to actually be using this API. And what I didn't count on is we're like, well, we're a bunch of backend engineers, and everybody wants to flex. You want to be like "Oh, this is the fastest service at the company. Look at how fast this API is. I have two heap allocations in this whole code path." So everybody was using this API. Initially, it was kind of like "This is crazy." We should be logging strings, with \[unintelligible 01:01:51.16\] printf. And if you want to wrap them in JSON all the way at some boundary, where it puts the process ID and the hostname or whatever in there, that's fine. But as a programmer, I shouldn't be staring at this super-fussy JSON production API. And if I can't like, "If I can't convince you of that, I want to at least convince you that JSON is just the worst format." Like, we're going to produce JSON, and then Docker is going to parse it and re-escape it, and then Filebeat is going to parse it and reencode it, and then LogStash is going to parse it and reencode it, and then ElasticSearch is going to parse it and index it... And we're like "What are we doing?" At least use MessagePack, or something that's binary, and three times faster. + +**Johnny Boursiquot:** Yeah, we just like to give money to the cloud vendors. That's cool. Keeping that whole industry in business. + +**Jon Calhoun:** Johnny is not biased at all here... + +**Akshay Shah:** It was just - for the amount of effort spent golfing allocations out of this one log API, the fact that end-to-end this is just stupendously inefficient... it should matter. We should pick a format that, at the very least, if you want this really fussy logging API, in between your program and the ultimate search index, or whatever, we should pick something that's easy and cheap to work with. So structured logging - thumbs down for me. + +**Johnny Boursiquot:** Nice. \[laughs\] Oh, we're gonna see how well that one does... + +**Jon Calhoun:** I'm just gonna be quiet, because I don't -- the upside to being a one-person company is that I don't really work on things that need to worry about any of those performance issues or anything that, so I can kind of get away with whatever I want. + +**Akshay Shah:** I used the standard lib's log package on all of my personal stuff. + +**Jon Calhoun:** I use println() a lot for printing stuff out, like if I need to know something... But that's because it's just me, so it's like, "Okay, this is pretty easy to deal with." I don't have a crazy amount of traffic. I'm currently not running any services with more than 100,000 users, so it's not too crazy. + +**Akshay Shah:** That's a very successful software business. + +**Johnny Boursiquot:** Seriously. + +**Akshay Shah:** 100,000 people is a lot. + +**Jon Calhoun:** It's not concurrent, for sure. It's like, there's 100,000 signups for a free course, and at this point, probably half of them haven't logged on in who knows how long. + +**Akshay Shah:** \[01:04:11.25\] How are you not a multibillion-dollar VC-funded juggernaut at this point? + +**Jon Calhoun:** Because that's a free course. When things are free... + +**Akshay Shah:** Got it. + +**Johnny Boursiquot:** \[laughs\] Well -- + +**Akshay Shah:** That hasn't stopped anybody for the last five or six years... + +**Johnny Boursiquot:** That's true, that's true... + +**Jon Calhoun:** I'd have to have some grand vision then, and I don't have that. My vision is just to try to help people learn Go. + +**Johnny Boursiquot:** And just to make it free. + +**Akshay Shah:** I'll tell you, I'll send you a slide deck for a 10% cut. + +**Johnny Boursiquot:** \[laughs\] Oh, man... That's awesome. That's awesome. So you don't have an unpopular opinion, Jon? + +**Jon Calhoun:** I do not. Johnny, what is your unpopular opinion? + +**Johnny Boursiquot:** Okay. Well, since you and I both teach Go, so you might appreciate this... I think if you have a training, where you're not actually typing code and showing people how it actually feels and looks like to write the code in an editor... If all you have is slideware, and things, and animations and things, I think you're doing it wrong. I think you need to actually show an editor, you need to show you writing in code, in order for it to sink in a bit more. + +**Jon Calhoun:** I agree with that 100%. In almost everything you can teach. Just to sort of back that up - when students learn about algorithms in school, I feel like half the reason they don't really understand them is because they're often not seeing code, and they're not using it extensively... So they might see a couple blips of code on the screen or something, but they don't actually walk through like "Let's actually write a binary search from the ground-up and see how it works." And I can say from school, I had so many classmates who just couldn't -- if you asked them to write a binary search, they could explain it to you, but they could not code it, for some reason. So there was some disconnect there. + +I did Programming Teams, so I had to code this type of stuff all the time. The programming team is like a Top Coder, or Google Code Jam, that type of stuff, is what we did all the time. I got to work with other people who were very good at breaking it down and showing you "Here's ten examples of actually coding this thing." And by the end, you felt like you understood the algorithms so much better than just sort of the idea of it being explained to you. And I feel like that's true for pretty much anything you're teaching programming-wise - if you can really illustrate, "Here's how you code it, I think that really helps it sink in", and especially if people code along and try to do it on their own... Because I think that's another mistake people make, is that they won't put in the time of actually trying to do it themselves to see what it feels for themselves to code it. I don't know how your trainings are with that, Johnny, because we give slightly different types of trainings... But if they code along or not...? + +**Johnny Boursiquot:** Oh yeah, absolutely. The code-along is encouraged. As a matter of fact, this is a good segue into a shameless plug... By the time this episode is available, I have a new training coming out with Lincoln Learning, actually... It's a hands-on introductory Go course. So for those who have been sort of looking to get into Go, and are coming from other languages, and kind of really want to know, "Okay, how do I quickly level up in Go in a very hands-on sort of way?", where you see me write code as I explain concepts and things... The good thing is, it's going to be available free through February. So if you need a link, when you hear this podcast, if you need a link, just hit me up on Twitter, and I'll send you a link. + +So I'm working very hard on it, looking forward to actually putting this thing out in the world... It's kind of like one of your babies, right John? You work on it, you toil, and you sweat, nights, weekends, all the things, and then it comes out and you're super-proud of it... This is definitely one of them for me. + +**Akshay Shah:** Is this the thing that's launching tomorrow, that you were mentioning at the beginning? + +**Johnny Boursiquot:** CodeSpaces? Yeah, that's the thing launching tomorrow, along with GitHub, at GitHub Universe, and some CodeSpaces stuff. It's upon us, so hopefully I'm not breaking any rules and stuff, and having sort of a recording that shows up a week or two later... So it's just in time. + +**Akshay Shah:** Nice. + +**Jon Calhoun:** Nice. + +**Akshay Shah:** I learned programming, like commercial programming, before my first job, with a Learn Python book, that was very much in line with your opinion, Johnny. It was Zed Shaw's Learn Python the Hard Way. + +**Johnny Boursiquot:** \[01:08:15.23\] Yeah... + +**Akshay Shah:** And it's literally just a bunch of code that you just hand-type and push Enter, and try and figure out why it did what it did. + +**Johnny Boursiquot:** Yup. + +**Akshay Shah:** And I remember I burned through the whole thing, because I had somehow convinced them that I knew how to program before I started, but I didn't really... + +**Johnny Boursiquot:** \[laughs\] Nice. + +**Akshay Shah:** Certainly not in Python. And it was under the gun to really figure it out quick. + +**Johnny Boursiquot:** Those are the best, right? + +**Akshay Shah:** Well, I showed up there, and I was like "I know how to code", and they kind of looked at me and they're like "Yeah, that checks out. Come on in." It was not the greatest... \[laughs\] + +**Johnny Boursiquot:** Nice. + +**Jon Calhoun:** Johnny, to speak to your point though... It always baffles me that -- if you took somebody to a soccer training, and you never got a soccer ball out, every would be like "What is going on?!" + +**Johnny Boursiquot:** Right. \[laughs\] "Just watch me do this. You don't get to touch the ball." + +**Jon Calhoun:** It would be kind of ridiculous. But in programming, for some reason that sometimes happens. And there are some topics where you can sort of talk about things at a high level, but a lot of times that code is very important. And I know as developers we like to say "Well, we get paid to think and do these other things, and the code's a small part of it", which is true... But you also the code helps make things sink in. And I think you get to that point where you can focus on thinking when the code is kind of a -- that's a small detail, but you have to learn how to implement that small detail and actually do it. + +**Akshay Shah:** How much have generics made you revisit your courses, John? + +**Jon Calhoun:** None at all so far, but that's because -- so right now, one of my courses, honestly, it needs updated more because of tooling changes. + +**Akshay Shah:** Got it. + +**Jon Calhoun:** So Go modules and some other stuff that. And it's not that the code is no longer valid, it's more just when somebody's in that sort of beginner to intermediate phase, it's kind of confusing if they haven't run into it already to be "Hey, this code doesn't work because I don't have a module setup", and my videos don't mention it because it didn't exist at the time. + +**Akshay Shah:** Yeah. + +**Jon Calhoun:** So it's like "Okay, that sucks." But I generally don't try to do anything that has generics where they'd even be a part of it, because it's more focused on that beginner to sort of intermediate type grouping at the moment. I do plan on doing a little bit more advanced stuff, which is -- like, I have a course in the works that's sort of been... I've been toiling away at it for a while. But as Johnny knows, those things can take forever, so... That's always interesting. + +But I think I'll use generics. I don't think too many people are going to need us to write generics. And maybe I'll have something that covers that at some point, but I don't think most people actually need to write generics very often. + +**Akshay Shah:** That's fair. Pretty much the only time I've rewritten the same code to generate a set of a given type - five or six times now. So I'm looking forward to getting rid of that, and just having a generic set type that I can cart around. That's pretty much the only data structure that I routinely end up wanting. + +**Johnny Boursiquot:** Right. + +**Jon Calhoun:** \[01:10:55.25\] I guess now that think about it, I do have a course that's like -- it's a course that's meant to cover some more common algorithms and data structures, and it's a free one that I was just... It's something I enjoy, so I wanted to teach it that way... And I'll probably add generics to that one at some point, just because people are going to be like "Why aren't you using generics?" But I still think a lot of the time it's easier to teach something without them. + +**Akshay Shah:** Yeah, that makes sense. + +**Jon Calhoun:** And then make sure they understand it with like -- let's just assume we're using strings right now, and just write it that way. And then take the next step. And that partially stemmed from the fact that I know a couple people who've been in college over the last five to ten years who I have ended up helping because they're friends of my younger sister, or something that... And in so many cases, somebody will come to me and they'll be like "I'm trying to implement this algorithm, and I don't know what this Java's t thing is", and all this other stuff. And it's because they understand the algorithm, but all of a sudden generics were thrown in there as part of the implementation, and they're like "I don't really get what this is. It was not taught to me very well." So they get caught up on that. And in many cases, I've told them, "Okay, let's just ignore the t stuff. Can you just write this assuming that you have this class that -- it's a string. Don't worry about other data types, just assume it's a string." And then all of a sudden, they're like "Oh yeah, that's fine" and they write it all. And then I'm like "Okay, now let's put the t stuff back", and it's a little bit more annoying, but it's like, that's all that was, was just letting you use any data type, and it really shouldn't be the thing that stopped you from moving forward. + +**Akshay Shah:** That's kind of nice about the evolution of Go as a language... You really feel intimately aware of how at least one implementation of generics might work, because you've spent so many years writing code generators... You're like yeah, this really literally is like "I'm just gonna shove a string into the code here for this templated data type." It's kind of like a visceral appreciation for the problem that generics solves. It makes a lot of sense. + +**Jon Calhoun:** I'm definitely a fan of generics for data structures, and that sort of thing. I just -- I guess we've moved on from that debate, luckily, or I think at least in my circle it feels like we have, but... + +**Johnny Boursiquot:** Yeah, I think so. Yeah. + +**Jon Calhoun:** And I don't feel it's been like this big apocalypse like everybody thought it was going to be... At least I haven't seen a million packages released that have way too much generic use. Maybe I'm getting old and don't have time to look for these things now. + +**Johnny Boursiquot:** You're not trying enough things; you're not trying enough new things, Jon. + +**Jon Calhoun:** I mean, that's definitely true these days, compared to what I used to... There was definitely a time where I would have been like "Oh, gRPC looks sweet. I've gotta throw this into something and find a way to use it." And now I'm like "Do I have a real need for this?" Because otherwise, I'm just going to use what I know works. + +**Akshay Shah:** That makes sense. I think I run into generics a lot because when you work on protocol buffers, you're also working on code generation. + +**Johnny Boursiquot:** It makes sense there. + +**Akshay Shah:** So those kind of naturally come together, right? You're like "Oh, could I generate less code if I use generics here?" + +**Johnny Boursiquot:** Right. + +**Akshay Shah:** The answer is yes, but... + +**Johnny Boursiquot:** Yeah, suitable use. Yeah. + +**Akshay Shah:** Yeah. + +**Jon Calhoun:** Alright. Akshay, thank you for joining us and talking about everything gRPC and protocol buffers. Johnny, always great to have you. + +**Johnny Boursiquot:** Yes, likewise. + +**Akshay Shah:** Thank you so much. It was nice to meet both of you. + +**Johnny Boursiquot:** Likewise.