Datasets:

kerinin

/

hackernews-stories

like 2

Software Bug Delays Ingenuity Helicopter’s 4th Mars Flight

This appears to be the same issue that caused the delay in Ingenuity's first flight timeline. NASA says it's planning to try this one again today, and we should know in a few hours whether or not it was successful. By Ryan Whitwam April 30, 2021 Share on Facebook (opens in a new window) Share on Twitter (opens in a new window) Share on Reddit (opens in a new window) Share on Hacker News (opens in a new window) Share on Flipboard (opens in a new window)

Automatminer: ML Pipelines for Materials Science

Automatminer is a tool for automatically creating strong machine learning pipelines for materials science, including automatic featurization with matminer, feature reduction, and an AutoML backend. Put in a materials dataset, get out a machine that predicts materials properties. How it works Automatminer automatically decorates a dataset using hundreds of descriptor techniques from matminer’s descriptor library, picks the most useful features for learning, and runs a separate AutoML pipeline. Once a pipeline has been fit, it can be summarized in a text file, saved to disk, or used to make predictions on new materials. Automatminer uses pandas dataframes for all of its working objects. Put dataframes in, get dataframes out. Here’s an example of training on known data, and extending the model to out of sample data. from automatminer.pipeline import MatPipe # Fit a pipeline to training data to predict band gap pipe = MatPipe () pipe . fit ( train_df , "band gap" ) # Predict bandgap of some unknown materials predicted_df = pipe . predict ( unknown_df ) Overview Automatminer can work with many kinds of data: both computational and experimental data small (~100 samples) to moderate (~100k samples) sized datasets crystalline datasets composition-only (i.e., unknown phases) datasets datasets containing electronic bandstructures or density of states Many kinds of target properties: electronic mechanical thermodynamic any other kind of property And many featurization (descriptor) techniques: See matminer’s Table of Featurizers for a full (and growing) list. Automatminer is designed to be easy to use and reproducible Save pipelines which are portable across machines Fit a complete pipeline with 1 line of code Predict on new samples with 1 line of code Presets for easy setup Automatminer is automatic and accurate No hand tuning required Comparable in accuracy to hand-tuned models in benchmark tests What’s new? Track changes to automatminer through the changelog. Citing Automatminer or MatBench If you find Automatminer or the MatBench benchmarks helpful in your research, please consider citing our publication in npj Computational Materials: Dunn, A., Wang, Q., Ganose, A., Dopp, D., Jain, A. Benchmarking Materials Property Prediction Methods: The Matbench Test Set and Automatminer Reference Algorithm. npj Computational Materials 6, 138 (2020). https://doi.org/10.1038/s41524-020-00406-3 API documentation Autogenerated API documentation. Beware! Only for the brave. p p p

How to Hide from a Drone

Drones of all sizes are being used by environmental advocates to monitor deforestation, by conservationists to track poachers, and by journalists and activists to document large protests. As a political sociologist who studies social movements and drones, I document a wide range of nonviolent and pro-social drone uses in my new book, “The Good Drone.” I show that these efforts have the potential to democratize surveillance. But when the Department of Homeland Security redirects large, fixed-wing drones from the U.S.-Mexico border to monitor protests, and when towns experiment with using drones to test people for fevers, it’s time to think about how many eyes are in the sky and how to avoid unwanted aerial surveillance. One way that’s within reach of nearly everyone is learning how to simply disappear from view. Over the past decade there’s been an explosion in the public’s use of drones – everyday people with everyday tech doing interesting things. As drones enter already-crowded airspace, the Federal Aviation Administration is struggling to respond. The near future is likely to see even more of these devices in the sky, flown by an ever-growing cast of social, political and economic actors. Public opinion about the use and spread of drones is still up in the air, but burgeoning drone use has sparked numerous efforts to curtail drones. These responses range from public policies exerting community control over local airspace, to the development of sophisticated jamming equipment and tactics for knocking drones out of the sky. From startups to major defense contractors, there is a scramble to deny airspace to drones, to hijack drones digitally, to control drones physically and to shoot drones down. Anti-drone measures range from simple blunt force, 10-gauge shotguns, to the poetic: well-trained hawks. Many of these anti-drone measures are expensive and complicated. Some are illegal. The most affordable – and legal – way to avoid drone technology is hiding. The first thing you can do to hide from a drone is to take advantage of the natural and built environment. It’s possible to wait for bad weather, since smaller devices like those used by local police have a hard time flying in high winds, dense fogs and heavy rains. Trees, walls, alcoves and tunnels are more reliable than the weather, and they offer shelter from the high-flying drones used by the Department of Homeland Security. The second thing you can do is minimize your digital footprints. It’s smart to avoid using wireless devices like mobile phones or GPS systems, since they have digital signatures that can reveal your location. This is useful for evading drones, but is also important for avoiding other privacy-invading technologies. The third thing you can do is confuse a drone. Placing mirrors on the ground, standing over broken glass, and wearing elaborate headgear, machine-readable blankets or sensor-jamming jackets can break up and distort the image a drone sees. Mannequins and other forms of mimicry can confuse both on-board sensors and the analysts charged with monitoring the drone’s video and sensor feeds. Drones equipped with infrared sensors will see right through the mannequin trick, but are confused by tactics that mask the body’s temperature. For example, a space blanket will mask significant amounts of the body’s heat, as will simply hiding in an area that matches the body’s temperature, like a building or sidewalk exhaust vent. The fourth, and most practical, thing you can do to protect yourself from drone surveillance is to get a disguise. The growth of mass surveillance has led to an explosion in creative experiments meant to mask one’s identity. But some of the smartest ideas are decidedly old-school and low-tech. Clothing is the first choice, because hats, glasses, masks and scarves go a long way toward scrambling drone-based facial-recognition software. Your gait is as unique as your fingerprint. As gait-recognition software evolves, it will be important to also mask the key pivot points used in identifying the walker. It may be that the best response is affecting a limp, using a minor leg brace or wearing extremely loose clothing. Artists and scientists have taken these approaches a step further, developing a hoodie wrap that’s intended to shield the owner’s heat signature and to scramble facial recognition software, and glasses intended to foil facial recognition systems. These innovations are alluring, but umbrellas may prove to be the most ubiquitous and robust tactic in this list. They’re affordable, easy to carry, hard to see around and can be disposed of in a hurry. Plus you can build a high-tech one, if you want. It would be nice to live in a world with fewer impositions on privacy, one in which law enforcement did not use small quadcopters and the Department of Homeland Security did not redeploy large Predator drones to surveil protesters. And, for people in some parts of the world, it would be nice not to associate the sound of a drone with impending missile fire. But given that those eyes are in the sky, it’s good to know how to hide. Austin Choi-Fitzpatrick is the author of: The Good Drone: How Social Movements Democratize Surveillance

Malicious NPM package steals Chrome passwords

July 21, 2021 09:00 AM New npm malware has been caught stealing credentials from the Google Chrome web browser by using legitimate password recovery tools on Windows systems. Additionally, this malware listens for incoming connections from the attacker's C2 server and provides advanced capabilities, such as screen and camera access, directory listing, file lookup, file upload, and shell command execution. As seen by BleepingComputer, the identified packages have been sitting on the npm registry since 2018 and scored over 2,000 total downloads at the time of writing. Today, researchers at ReversingLabs have disclosed their findings on two malicious npm packages that secretly steal passwords from your Chrome web browser. These packages were discovered by ReversingLabs' Titanium Platform static analysis engine that employed machine learning algorithms. But, the primary focus of the report is on nodejs_net_server which contains the core malware functionality. The malware targets Windows machines to steal user credentials and also sets up a persistent remote backdoor for the attacker to conduct surveillance activities. To facilitate its credential-stealing activities, the malware—specifically "nodejs_net_server," uses the legitimate ChromePass freeware utility for Windows. ChromePass is a password recovery tool for Windows systems aimed at extracting passwords from the user's Chrome web browser: ChromePass password recovery utility (NirSoft) This utility is packed inside the npm package with cryptic or misleading names, such as a.exe. Regardless, such ChromePass executables have previously been flagged by VirusTotal as malicious. The "nodejs_net_server" has had 12 versions published to date, with the latest one 1.1.2 measuring about 40 MB in size uncompressed. In later versions, though, the malware is seen launching TeamViewer.exe to avoid raising red flags. Most malicious npm packages caught thus far rely on typosquatting or dependency confusion to infect developers. But, that's not the case with these packages, and it isn't yet known how these packages managed to get so many downloads. nodejs_net_server download page on the npm registry (BleepingComputer) "We haven't found any obvious typosquatting target by analyzing the package name." "It is unclear to us how the author intended to trick users into installing the package. We can however see download activity on the package statistics page." "We have contacted NPM to take the package down. We are still waiting on their security team to respond," ReversingLabs' chief software architect and co-founder, Tomislav Pericin told BleepingComputer in an email interview. npm got back to us with: "We removed the package in accordance with npm's acceptable use policy regarding malware, as outlined in its Open-Source Terms," a GitHub spokesperson told BleepingComputer. Interestingly, as soon as the package is installed by the developer, it attempts to gain persistence on the Windows machine by abusing the well-known npm configuration option, "bin". The "bin" option in the package's manifest file, package.json, is aimed at hijacking the popular "jstest" package, should it be pre-installed on a developer's machine. "jstest" is a cross-platform JavaScript testing framework downloaded over 36,000 times to date—meaning, high chances a NodeJS developer would have it. package.json for nodejs_net_server (BleepingComputer) But, having "jstest" pre-installed is by no means a prerequisite for the malicious package to run. Its presence merely helps the malware achieve persistence on the infected machines: "JSTest doesn't need to be installed for this attack to work. Package installation hijacks the command 'jstest' if it was already assigned." "Running that command would ensure that malware gets persistence and that it executes the backdoor functionality," Pericin further told BleepingComputer. The "jstest" file loaded by the malware attempts to overwrite the contents of the existing "jstest" symlink, and further adds another JS file ("test.js") as a Windows service which would now run persistently. Malware attempts to achieve persistence by adding the lib/test.js script as a Windows service (BleepingComputer) This newly added Windows service opens up port 7353 that the attacker to connect to and perform various surveillance activities, including: Malware opens up a socket connection on port 7353 (BleepingComputer) As for temptesttempfile, the package is minimal with just two files, and only implements the remote shell functionality of nodejs_net_server, making it seem like a test package as the name suggests. In an unexpected twist, some versions of nodejs_net_server contain text files with usernames and plaintext passwords of the malware author, extracted from Chrome. ReversingLabs suspects this to be an accident on the author's part: "Fun fact related to versions that contain the password recovery tool is that the package author accidentally published their own, stored login credentials." "It appears that the published versions 1.1.1 and 1.1.2 from the NPM repository include the results of testing the ChromePass tool on the author's personal computer." "These login credentials were stored in the 'a.txt' file located in the same folder as the password recovery tool named 'a.exe,'" said ReversingLabs reverse engineer Karlo Zanki in a blog post. Zanki's observation was also confirmed by BleepingComputer when we noticed two files, a.txt, and b.txt with plaintext credentials, sitting in the aforementioned versions of "nodejs_net_server." NPM malware includes list of passwords believed to be the malware author's (BleepingComputer) Over the last few months, attacks on open source ecosystems including, npm, PyPI and RubyGems have grown steadily. With recent reports of ongoing dependency hijacking attacks flooding open source repos, the problem isn't going away anytime soon. ReversingLabs believes, understanding what's inside your software, or having a Software Bill of Materials (SBOM) is a critical step in defending against these supply chain attacks. "Package repositories offer conveniences for rapid application development, but also come with risks." "Understanding the package dependency tree, or software bill of materials, has become a critical part of defense against software supply chain attacks." "Every component should be looked with scrutiny before installation, or there's a chance malicious code can slip by unnoticed." "We are yet to see a malicious repository package embed itself in the final release image, but that seems like it's only a matter of time with the current state of things," concluded Pericin in his interview with BleepingComputer. Update: July 22, 2021 1:25 AM ET: Added statement from npm's parent company, GitHub, received after publishing. Related Articles: npm packages caught serving TurkoRAT binaries that mimic NodeJS Stealthy SeroXen RAT malware increasingly used to target gamers PyPI announces mandatory use of 2FA for all software publishers PyPI temporarily pauses new users, projects amid high volume of malware Facebook disrupts new NodeStealer information-stealing malware

Crypto exchanges keep getting hacked, and there's little anyone can do

It’s not just lucky investors getting rich from crypto. Hackers have made off with billions of dollars in virtual assets in the past year by compromising some of the cryptocurrency exchanges that have emerged during the bitcoin boom. There have been more than 20 hacks this year where a digital robber stole at least $10 million in digital currencies from a crypto exchange or project. In at least six cases, hackers stole more than $100 million, according to data compiled by NBC News. By comparison, bank robberies netted perpetrators an average of less than $5,000 per heist last year, according to the FBI’s annual crime statistics. Despite the large dollar amounts associated with these thefts, they often lack the drama or attention of traditional bank robberies. But cryptocurrency experts say they offer a warning to would-be crypto investors: Exchanges are now lucrative targets for hackers. “If you hack a Fortune 500 company today, you might steal some usernames and passwords,” said Esteban Castaño, the CEO and co-founder of TRM Labs, a company that builds tools for companies to track digital assets. “If you hack a cryptocurrency exchange, you may have millions of dollars in cryptocurrency.” Once an internet oddity that required a certain level of tech know-how to buy, cryptocurrencies have emerged as a more mainstream investment and speculation tool, spurring more than 300 companies to start up in recent years to offer people an easy way to buy and sell everything from bitcoin to more fringe “altcoins” such as the dog-inspired dogecoin. Crypto exchanges work like traditional money exchanges, setting prices for various currencies and taking a small fee to let users trade one. But while a handful of countries have strict regulations in place, it’s relatively easy for tech entrepreneurs to set up an exchange nearly anywhere in the world and run it however they like. Cryptocurrencies generally offer a certain amount of security — taking their name, in part, from “encryption.” But the exchanges that manage them, especially new ones building their businesses from scratch, often start with a tiny staff, which means few if any full-time cybersecurity professionals. Their developers may work frantically to make the code work, sometimes accidentally leaving flaws that give hackers a foothold. Combined with the fact that a volatile market often leaves them suddenly holding a fortune, exchanges are a particularly ripe target for criminal hackers. Exchanges often keep access to some of their cryptocurrencies in so-called cold wallets, which live safely offline. The rest of it is in “hot wallets,” that are liquid and can be sent to users. That means that if a hacker can gain access to a particular employee account — a common security breach on the internet — they can pull off a major heist, said Dave Jevans, the founder of CipherTrace, a company that tracks theft and fraud in cryptocurrencies. “If you steal the private keys to a hot wallet, it’s not like stealing a database of people’s names and Social Security Numbers,” Jevans said. “You’ve just basically stolen all their money.” If an exchange is wealthy enough and plans ahead to have an emergency fund, it can compensate its customers if its operation is hacked, Jevans said. If not, they often goes out of business. “Not every exchange is so wealthy or has so much foresight. It just goes, pop, ‘We’re out of business. Sorry, you’re all screwed,’” he said. One of the biggest heists happened in early December, when the crypto trading platform Bitmart announced that hackers broke into a company account and stole almost $200 million. The company froze all customer transactions for three days before it allowed them to trade their money again. The problem is exacerbated because many cryptocurrency projects, intent on avoiding government regulations, set up in countries whose law enforcement agencies don’t have much power to go after transnational hackers. Or if they are hacked, they tend to be less likely to call for government help on ideological grounds, said Beth Bisbee, head of U.S. investigations at Chainalysis, a company that tracks cryptocurrency transactions for both private companies and government agencies. Some developers “want to be anti-bank and anti-oversight,” Bisbee said. “So when something like that happens, they’re not necessarily wanting to work with law enforcement, even though they’d be considered to be a victim and it’d be valuable for them to.” While exchange hacks offer some similarities to the bank heists of old, they don’t leave behind the hallmarks that once made them front-page news. Public scrutiny of these hacks can be lacking despite the large dollar amounts. Most exchange hackers are not caught, leaving little closure for consumers. And there is rarely any physical evidence or real-world aftermath: no traumatized bank tellers or perp walks. But some hacks do have happy endings. In one bizarre, public case, a hacker stole $600 million from the cryptocurrency platform Poly Network. Instead of blaming the thief, the company decided to appeal to his better nature, calling him “Mr. White Hat,” which is a cybersecurity term for a researcher working to help make things more secure. Poly Network thanked him for exposing a flaw in its code and asked for the money back. The hacker eventually relented and returned it all. But those instances are rare. Usually, when major law enforcement agencies tackle a major cryptocurrency hack, they try to follow every lead, an exhausting process that moves far slower than the criminals they’re chasing. Claire Georges, the deputy spokesperson for Europol, the European Union’s international law enforcement agency, said the agency is aware of a number of cases against hackers who steal digital assets. But she said building a solid case is a long, slow process that doesn’t keep up with the pace of attacks. “We have a number of investigations going as we speak,” Georges said. “They take a long time, because we also would want to take down the whole criminal network,” she said. “These cases often feed into other cases.” “They could go on forever,” she said. “These investigations usually take time.”

Hedgewars

hedgewars.org technical issues Tue, 2023-03-21 01:28 — sheepluva This website and its related services are currently experiencing various issues due to a sudden loss of server infrastructure. We'll do our best to restore full functionality as soon as we can. Thank you for your patience Update, 2023-03-21 23:25 (UTC): Outgoing mail (i.e. password reset) works again Update, 2023-03-22 01:40 (UTC): Our code repository has returned from the dead Update, 2023-03-22 21:45 (UTC): Even although I'm pretty sure Hedgewars doesn't have any bugs whatsoever, ahem, our bugzilla is back. Update, 2023-03-23 21:50 (UTC): The Knowledge Base's Web-Interface allows KB admins to edit pages again. DLC issues for Hedgewars 1.0.0 on Windows Tue, 2021-10-12 14:27 — nemo [Update 2021-12-02] Thanks to sheepluva for a manual fix to our cert, 1.0.0 direct link should work again in Windows. However, the unit22 link will not, at least for now. One of the interesting things about Windows is that almost all distribution is similar to the "Snap/Flatpak" model - instead of distros managing packaging and updating with a single library used by hundreds of apps, all libraries have to be included by the app, with hundreds of copies of the library instead. This lack of unified package management probably explains 20GB C:\Windows\Installer folders too. It appears the version of OpenSSL in the 1.0.0 Hedgewars is no longer functioning with DLC. This could be due to: https://scotthelme.co.uk/lets-encrypt-old-root-expiration/ As a workaround pending a new release, Windows users can access the DLC page here (it's the exact same page that is shown in client): https://hedgewars.org/content.html You can then download the HWPs off that page and manually copy them to your Data folder which should be My Documents/Hedgewars/Data. The icon in the lower right on the DLC page should also open this folder and should still work. New Antarctic theme Wed, 2020-10-07 18:36 — Wuzzy The forum user “KIRA” recently created a beautiful new theme called “Antarctic”. This theme has been added as optional add-on. You can download this theme in the “Downloadable Content” section of the game. Just click on “Downloadable Content” in the main menu, then on “Antarctic_v4”, restart Hedgewars and you should be ready to go! Enjoy! You can give feedback in the forums. EDIT 2020-10-09: Today, two more cool themes by Kakcoo from the year 2016 have been added as well: Doktor_v1 and Teknologi_v1. Discussion. how do i update it on mint Tue, 2020-04-28 12:47 — nemo So at 1:07 in the morning my time, someone from the other side of the world posted this... 01:07 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint 01:09 < REDACTED> how do i update it on mint Then quit... Seemed like a good opportunity to cover your options when you are on linux. First off, usually your package manager is responsible for this, not the people volunteering to bang code together (although we often end up doing distro tech support too). Mint is an interesting case. It's loosely based off ubuntu, but without the "App Store" so installing packages is a bit more Debian-y. Basically on Ubuntu/Mint in order of preference you should use: 1) Official release package if it is up to date 2) Backports/Updates package if it is up to date https://help.ubuntu.com/community/UbuntuBackports https://backports.debian.org/Instructions/ 3) https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/hedgewars (PPA by the current maintainer of the Debian/Ubuntu packages) Debian/Devuan is about the same except (3) is manual and at your own risk. If all else fails, you can try installing the .debs for the version of Debian or Ubuntu that most closely matches what your distro was derived from, or try our build instructions. https://www.hedgewars.org/kb/BuildingOnLinux I haven't brought up other distros here, but usually we don't have problems with them. Occasionally with Arch. Hedgewars 1.0.0 released! Wed, 2019-10-09 18:15 — Wuzzy R EL EA SE PA RT Y TIME! 1.0.0 is here! That's right, Hedgewars version 1.0.0 has been released! Oh boy, it has finally happened! It was a long, long journey, but we reached our destination! Version 1.0.0 is a major milestone for this game, as its the first version we find confident to share with a wider audience. Compared to version 0.9.25, this release contains mostly a cleanup of existing features and bugfixes. Basically stuff you'd expect for a “worthy” 1.0.0 release. Have fun! Go to the Download page to obtain Hedgewars. Noteworthy changes Campaigns now use your team identity instead of ignoring it Single missions can now be played with your favourite teams and keep track of your progress Hand-drawn maps can be scaled with a slider Quick games are more random Homing bee can be used as secondary ammo Can change hedgehog order in The Specialists Various small HUD improvements Various controls improvements and bugfixes Read the full changelog. Follow this link to read the full post! strong 2 3 4 5 6 7 8 9 p next › last »

Syncretism

The gods Persephone-Isis and Hades-Serapis, an example of Greco-Egyptian syncretism b () [1] is the practice of combining different beliefs and various schools of thought. Syncretism involves the merging or assimilation of several originally discrete traditions, especially in the theology and mythology of religion, thus asserting an underlying unity and allowing for an inclusive approach to other faiths. Syncretism also occurs commonly in expressions of art and culture, known as eclecticism, as well as in politics, known as syncretic politics. The English word is first attested in the early 17th century, [2] from Modern Latin syncretismus , drawing on the Ancient Greek: συγκρητισμός , romanized:  synkretismos , supposedly meaning "Cretan federation", but this is a spurious etymology from the naive idea in Plutarch's 1st-century AD essay on "Fraternal Love (Peri Philadelphias)" in his collection Moralia . He cites the example of the Cretans, who compromised and reconciled their differences and came together in alliance when faced with external dangers. "And that is their so-called Syncretism [Union of Cretans]". More likely as an etymology is sun- ("with") plus kerannumi ("mix") and its related noun, "krasis," "mixture." Social and political roles The use of elephant-shaped column brackets in buildings of the Lahore Fort reflects Hindu influences on Mughal Architecture during the reign of Akbar. Islam forbids representation of living figures. Overt syncretism in folk belief may show cultural acceptance of an alien or previous tradition, but the "other" cult may survive or infiltrate without authorized i. For example, some conversos developed a sort of cult for martyr-victims of the Spanish Inquisition, thus incorporating elements of Catholicism while resisting it. The Kushite kings who ruled Upper Egypt for approximately a century and the whole of Egypt for approximately 57 years, from 721 to 664 BCE, constituting the Twenty-fifth Dynasty in Manetho's Aegyptiaca , developed a syncretic worship identifying their own god Dedun with the Egyptian Osiris. They maintained that worship even after they had been driven out of Egypt. A temple dedicated to this syncretic god, built by the Kushite ruler Atlanersa, was unearthed at Jebel Barkal. [3] [4] Syncretism was common during the Hellenistic period, with rulers regularly identifying local deities in various parts of their domains with the relevant god or goddess of the Greek Pantheon as a means of increasing the cohesion of their kingdom. This practice was accepted in most locations but vehemently rejected by the Jews, who considered the identification of Yahweh with the Greek Zeus as the worst of blasphemy. All major religious conversions of populations have had elements from prior religious traditions incorporated into legends or doctrine that endure with the newly converted laity. [5] The god Hermanubis, an example of Greco-Egyptian syncretism The god Taranis-Jupiter, an example of Romano-Celtic syncretism Religious syncretism is the blending of two or more religious belief systems into a new system, or the incorporation into a religious tradition of beliefs from unrelated traditions. This can occur for many reasons, and the latter scenario happens quite commonly in areas where multiple religious traditions exist in proximity and function actively in a culture, or when a culture is conquered, and the conquerors bring their religious beliefs with them, but do not succeed in entirely eradicating the old beliefs or (especially) practices. Religions may have syncretic elements to their beliefs or history, but adherents of so-labeled systems often frown on applying the label, especially adherents who belong to "revealed" religious systems, such as the Abrahamic religions, or any system that exhibits an exclusivist approach. Such adherents sometimes see syncretism as a betrayal of their pure truth. By this reasoning, adding an incompatible belief corrupts the original religion, rendering it no longer true. Indeed, critics of a syncretistic trend may use the word or its variants as a disparaging epithet, as a charge implying that those who seek to incorporate a new view, belief, or practice into a religious system pervert the original faith. Non-exclusivist systems of belief, on the other hand, may feel quite free to incorporate other traditions into their own. Keith Ferdinando notes that the term "syncretism" is an elusive one, [6] and can apply to refer to substitution or modification of the central elements of a religion by beliefs or practices introduced from elsewhere. The consequence under such a definition, according to Ferdinando, can lead to a fatal "compromise" of the original religion's "integrity". [7] In modern secular society, religious innovators sometimes construct new faiths or key tenets syncretically, with the added benefit or aim of reducing inter-religious discord. Such chapters often have a side-effect of arousing jealousy and suspicion among authorities and ardent adherents of the pre-existing religion. Such religions tend to inherently appeal to an inclusive, diverse audience. Sometimes the state itself sponsored such new movements, such as the Living Church founded in Soviet Russia and the German Evangelical Church in Nazi Germany, chiefly to stem all outside influences. According to some authors, "Syncretism is often used to describe the product of the large-scale imposition of one alien culture, religion, or body of practices over another that is already present." [8] Others such as Jerry H. Bentley, however, have argued that syncretism has also helped to create cultural compromise. It provides an opportunity to bring beliefs, values, and customs from one cultural tradition into contact with, and to engage different cultural traditions. Such a migration of ideas is generally successful only when there is a resonance between both traditions. While, as Bentley has argued, there are numerous cases where expansive traditions have won popular support in foreign lands, this is not always so. [9] Akbar the Great holding a court discussing Theology In the 16th century, the Mughal emperor Akbar proposed a new religion called the Din-i Ilahi ("Divine Faith"). Sources disagree with respect to whether it was one of many Sufi orders or merged some of the elements of the various religions of his empire. [10] [11] Din-i Ilahi drew elements primarily from Islam and Hinduism but also from Christianity, Jainism, and Zoroastrianism. More resembling a personality cult than a religion, it had no sacred scriptures, no priestly hierarchy, and fewer than 20 disciples, all hand-picked by Akbar himself. It is also accepted that the policy of sulh-i-kul, which formed the essence of the Dīn-i Ilāhī, was adopted by Akbar as a part of general imperial administrative policy. Sulh-i-kul means "universal peace". [12] [13] The syncretic deism of Matthew Tindal undermined Christianity's claim to uniqueness. [14] The modern, rational non-pejorative connotations of syncretism arguably date from Denis Diderot's Encyclopédie articles: Eclecticisme and Syncrétistes, Hénotiques, ou Conciliateurs. Diderot portrayed syncretism as the concordance of eclectic sources. Scientific or legalistic approaches of subjecting all claims to critical thinking prompted at this time much literature in Europe and the Americas studying non-European religions such as Edward Moor's The Hindu Pantheon of 1810, much of which was almost evangelistically appreciative, embracing spirituality and creating the space and tolerance in particular disestablishment of religion (or its stronger form, official secularisation as in France) whereby believers of spiritualism, agnosticism, atheists and in many cases more innovative or pre-Abrahimic based religions could promote and spread their belief system, whether in the family or beyond.[ p ] ^ ^ The Oxford English Dictionary first attests the word syncretism in English in 1618. ^ ^ ^ ^ ^ ^ Peter J. Claus and Margaret A. Mills, : (Garland Publishing, Inc., 2003).South Asian Folklore: An Encyclopedia ^ Jerry Bentley, (New York: Oxford University Press, 1993), viii.Old World Encounters: Cross-Cultural Contacts and Exchanges in Pre-Modern Times ^ ^ ^ [ p ] ^ [ p ] ^

The Search for Life Around Alpha Centauri Just Took a Major Leap Forward

The A.V. Club Deadspin Gizmodo Jalopnik Jezebel Kotaku Quartz The Root The Takeout The Onion The Inventory p The Search for Life Around Alpha Centauri Just Took a Major Leap Forward A telescope dedicated to looking for potentially habitable planets around the nearest stars could be operational by 2023. By George Dvorsky PublishedNovember 18, 2021 Comments (7) Alerts We may earn a commission from links on this page. Our nearest neighbor, Alpha Centauri, is 4.37 light-years from Earth, which is super close from a cosmological perspective but achingly far from a human point of view. A new telescope promises to bring this intriguing star system, and any habitable planets it holds, into closer view. view video p p July 12, 2022 p October 31, 2022 The new mission, called TOLIMAN, was announced today in a press release . TOLIMAN is the ancient Arabic name for Alpha Centauri—the closest star system to Earth—but it’s also an acronym for Telescope for Orbit Locus Interferometric Monitoring of our Astronomical Neighbourhood. Once in space, astronomers will use the orbital observatory to search for potentially habitable exoplanets around Alpha Centauri. The international collaboration includes teams from the University of Sydney, Breakthrough Initiatives, Saber Astronautics, and NASA’s Jet Propulsion Laboratory. Peter Tuthill from the Sydney Institute for Astronomy at the University of Sydney will lead the project. We’re quite fortunate to have such an intriguing next-door neighbor. Alpha Centauri is a triple star system consisting of two Sun-like stars, named Alpha Centauri A and Alpha Centauri B, and a red dwarf known as Proxima Centauri. Two exoplanets are known to orbit Proxima Centauri: an Earth-sized planet parked inside the habitable zone (i.e. that sweet spot within which liquid water is stable at the surface) and a super-Earth located farther out. Alpha Centauri A is suspected to host a Neptune-sized exoplanet, but astronomers aren’t entirely certain. An exoplanet has yet to be discovered in orbit around Alpha Centauri B. Other exoplanets are likely still awaiting detection—and that’s where TOLIMAN comes in. Advertisement “Our nearest stellar neighbours—the Alpha Centauri and Proxima Centauri systems—are turning out to be extraordinarily interesting,” Pete Worden, executive director of Breakthrough Initiatives, said in the press release. “The TOLIMAN mission will be a huge step towards finding out if planets capable of supporting life exist there.” Breakthrough Initiatives, founded by billionaire Yuri Milner, provided seed funding for the project, as did the Australian government through its International Space Investment Expand Capability Grants program. Saber Astronautics, the recipient of AUD$788,00 (USD$573,300) from the Australian government, will provide spaceflight mission operations support, including space traffic management and satellite communications. The firm has facilities in both Australia and the United States. Jason Held, CEO of Saber Astronautics, described TOLIMAN in the press release as “an exciting, bleeding-edge space telescope,” one that will be “supplied by an exceptional international collaboration.” To which he added: “It will be a joy to fly this bird.” TOLIMAN will be custom-tailored for the mission, and its strong suit will be in making extremely fine measurements of the positions of the stars. A key feature of the new telescope is a “diffractive pupil lens.” By dispersing stellar light into flower-like patterns, the lens will make it easier for astronomers to spot wobbles caused by orbiting exoplanets. Once an exoplanet is detected, more specialized telescopes can be recruited to search for potential biosignatures in the atmosphere or surface. The telescope is expected to reach orbit in 2023, as Centauri Dreams reports . In 2019, scientists with Breakthrough Listen, one of several projects supported by Breakthrough Initiatives, identified a candidate signal coming from Proxima Centauri, in what was the first and so far only potential alien technosignature detected by the group. Subsequent research found the signal to be of human origin, ruling out an alien civilization as the source. More: What to know about Kessler Syndrome, the ultimate space disaster .

Top Linux Interview Questions and Answers (2021)

Download Interview guide PDF Before you leave, take this interview guide with you. Linux Interview Questions Linux Interview MCQs Excel at your interview with Masterclasses Know More By Certificate included Feedback Current Employer * Enter company name * Graduation Year * Select an option * 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 Phone Number * OTP will be sent to this number for verification +91 * +91 Phone Number Change Number Phone Number * OTP will be sent to this number for verification +91 * +91 Phone Number Change Number Graduation Year * Graduation Year * 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 Current Employer * Company Name * Please verify your phone number Edit Resend OTP By clicking on Start Test, I agree to be contacted by Scaler in the future. Already have an account? Instructions from Interviewbit Start Test

Why I Broke Wall Street’s Code of Silence

The other day, a colleague and I made $2.5 million. We were paid for doing the right thing, which doesn’t happen often in our neighborhood. Wall Street rarely rewards integrity. A career as a financial analyst isn’t particularly sexy. By and large, we’re not like the high-rolling traders and dealmakers who populate Hollywood scripts, drinking expensive Scotch at swanky downtown bars. We’re the men and women who pore through voluminous amounts of publicly available information and apply complex financial models to assess a company’s strengths and weaknesses. We look at organizations the way a physician examines a patient, who reports symptoms and relies on him or her for a diagnosis. Unlike doctors, for a host of complex reasons, we typically stop short of the cure. We analysts regularly spot irregularities that ignite curiosity. Most often, there are perfectly good explanations for an incongruity. But not always. Sixteen years ago, I discovered what looked like fraud at a financial services company and alerted the Securities and Exchange Commission. Despite providing hundreds of pages of detailed analysis, the regulator didn’t act. I’m not even sure anyone read it. Ultimately, the fraud led to bankruptcy and delisting — shareholders were left with nothing. This unnecessary and avoidable tragedy brought me no gratification. Instead, I grew skeptical of the SEC’s willingness to act on credible and well-researched tips. Going forward, this skepticism led me, and doubtless many of my peers, to turn a blind eye to potential frauds. Other factors in the risk-reward calculus also weighed on my reticence to speak out over the years. Placing anything other than a buy rating — even voicing the slightest criticism — can shut off access to the very management we depend on for critical information. We can be ostracized by our own firms for impeding their ability to conduct investment-banking business with the companies we cover. Many of us operate under strict confidentially agreements that, albeit illegally, may attempt to prevent us from reporting irregularities to regulators. And let’s not forget that being a “rat” violates Wall Street’s implicit code of silence. We have to ask, but we don’t have to tell. When Orthofix crossed my desk seven years ago, a colleague and I spotted irregularities at the medical device manufacturer that reeked of fraud. By that time, the SEC whistleblower program existed, promising anonymity, employment protections, and significant financial incentives. The SEC created a platform for tip intake and formally encouraged whistleblowers — a stark contrast to the indifference I had encountered years earlier. My colleague found a law firm that represented whistleblowers and asked me to join him in filing a claim. We built as close to a bulletproof case as anyone outside the kimono could make. We calculated the scope of revenue overstatements within 2.3 percent of the company’s restatement once it was caught with its pants down. As to whether I outed Orthofix for selfish motives, the answer is no. I’ve never bought or sold short Orthofix stock. It’s just a company that came across my desk in the ordinary course of my workday. The better question is: Where were the official lines of defense? The company’s auditors, executives, and compliance personnel had a fiduciary duty to detect and report misconduct. But they didn’t. The reality is that markets require all hands to clean the street. It’s dirty; we know it. Corruption disrupts our attempts to ensure free and fair markets. Worse, it hurts investors. Analysts are perfect whistleblowers. Using our expertise to identify irregularities and misconduct is not only public service; it also channels what we truly love about our work: the unearthing, the highly complex analysis, and old-fashioned investigative prowess. We can construct blueprints of wrongdoing in ways others can’t, because that’s our everyday. Two and a half million dollars later, are my colleague and I mercenary opportunists? Absolutely not. But make no mistake. If and when we encounter another financial fraud, we will act. By a veteran financial analyst, whose identity Institutional Investor has verified and agreed to keep confidential, owing to fears of retaliation and industry blackballing.

Atmospheric Meditation Nature, Relaxing, Yoga

Your browser isn’t supported anymore. Update it to get the best YouTube experience and our latest features. Learn more

Exxon lobbyists admit they have no intention of fighting climate change

Image: Emily Buchanan Inside Exxon’s playbook: How America’s biggest oil company continues to oppose action on climate change ExxonMobil aims to drastically weaken Biden’s climate plans and used shadow groups to ‘aggressively’ fight climate science, insider tells undercover reporter Inside Exxon’s playbook: How America’s biggest oil company continues to oppose action on climate change ExxonMobil aims to drastically weaken Biden’s climate plans and used shadow groups to ‘aggressively’ fight climate science, insider tells undercover reporter Image: Emily Buchanan ExxonMobil continues to fight efforts to tackle climate change in the United States, despite publicly claiming to support the Paris climate agreement, an undercover investigation by Unearthed has found. A senior lobbyist for Exxon told an undercover reporter that the company had been working to weaken key aspects of President Joe Biden’s flagship initiative on climate change, the American Jobs Plan. Did we aggressively fight against some of the science? Yes. – Keith McCoy, Exxon lobbyist He described Biden’s new plan to slash US greenhouse gas emissions as “insane” and admitted that the company had aggressively fought early climate science through “shadow groups” to protect its investments. Keith McCoy – a senior director in Exxon’s Washington DC government affairs team – told the undercover reporter that he is speaking to the office of influential Democratic senator Joe Manchin every week, with the aim of drastically reducing the scope of Biden’s climate plan so that “negative stuff”, such as rules limiting greenhouse gas emissions and taxes on oil companies, are removed. Last week – after weeks of bipartisan talks – President Biden conditionally endorsed a scaled-back version of his infrastructure plan, which eliminates hundreds of billions of dollars of proposed support for climate initiatives. During the undercover meeting, which took place via Zoom in May, McCoy suggested that Exxon’s public support for a carbon tax as its principal climate policy is an “advocacy tool” and “great talking point” that will never actually happen. “Nobody is going to propose a tax on all Americans and the cynical side of me says, yeah, we kind of know that but it gives us a talking point that we can say, well what is ExxonMobil for? Well, we’re for a carbon tax,” McCoy said. A second Exxon lobbyist, Dan Easley – who left the company in January after working as its chief White House lobbyist throughout the Trump administration – laughed when asked by an undercover reporter if the company had achieved many policy wins under Trump, before outlining victories on fossil fuel permitting and the renegotiation of the NAFTA trade agreement. “The wins are such that it would be difficult to categorise them all,” he said, adding that the biggest victory was Trump’s reduction in the corporate tax rate, which was “probably worth billions to Exxon”. Unearthed reporters posed as recruitment consultants looking to hire a Washington DC lobbyist for a major client and approached McCoy and Easley for meetings over Zoom. During the meetings, the undercover reporter asked about Exxon’s current and historical lobbying on environmental issues. It is important to note that neither McCoy nor Easley were necessarily seeking a new job, but each was willing to talk and provide information to the purported recruiters. Over the coming days, Unearthed, will also reveal: Claims that Exxon covertly fought to prevent a ban on toxic chemicals How Exxon is using its playbook on climate change to head-off regulations on plastic California Congressman, Rep. Ro Khanna, told Unearthed: “For decades, fossil fuel companies have lied to the public, to regulators, and to Congress about the true danger posed by their products. Today’s tape only proves our knowledge that the industry’s disinformation campaign is alive and well. In the coming months, I plan to ask the CEOs of Exxon, Chevron, and other fossil fuel companies to come testify before my Environment subcommittee. We can no longer allow Exxon, or any other companies, to prevent our collective action on the climate crisis.” A spokesman for ExxonMobil said that the allegations put to them: “contained a number of important factual misstatements that are starkly at odds with our positions on a variety of issues, including climate policy and our firm commitment to carbon pricing.” Denial and delay Exxon claims to support global effort to tackle climate change, but it hasn’t always. Throughout the 1990s and early 2000s, the company orchestrated a multimillion-dollar disinformation campaign that manufactured doubt regarding the link between global warming and the burning of fossil fuels. Did we join some of these ‘shadow groups’ to work against some of the early efforts on climate? Yes. – Keith McCoy, Exxon lobbyist It did so through a concerted strategic communications and lobbying push, which provided fringe scientists who denied climate science with funding and a platform, via Exxon-placed op-eds, advertisements, and political briefings. Exxon also helped to found and lead a powerful cross-industry group, the Global Climate Coalition (GCC), which spent tens of millions of dollars campaigning against a binding global climate agreement ahead of the 1997 UN climate summit in Kyoto. The organisation spent $13 million dollars on one advertising campaign alone, aiming to weaken US support for an agreement in Kyoto. The efforts were successful: the US Congress refused to ratify Kyoto and Exxon later lobbied the Bush administration to pull out of the protocol altogether. This left global efforts to rein in greenhouse gas emissions in tatters. Fighting science Exxon continues to deny having misled the public on climate change, and no serving Exxon executive has ever admitted that the company fought climate science to protect the company’s financial interests – until now. McCoy told an undercover Unearthed reporter that although he didn’t believe Exxon had buried its own science, the company had cast doubt on the scientific consensus: “Did we aggressively fight against some of the science? Yes. Did we hide our science, absolutely not. Did we join some of these ‘shadow groups’ to work against some of the early efforts? Yes, that’s true. But there’s nothing illegal about that. You know, we were looking out for our investments, we were looking out for our shareholders.” The reference to “shadow groups” is likely to relate to a powerful network of think tanks and pressure groups through which Exxon fought both the science and political action on climate change. Between 1998 and 2014, the company spent at least $30 million funding climate denial groups, such as the Heartland Institute, Competitive Enterprise Institute, and Heritage Foundation. This network played a critical role in shifting the Republican Party from a position of support for action to cut emissions in the 1980s to its near-total opposition to tackling climate change from the mid-1990s to today. Geoffrey Supran, a researcher at Harvard University who has written a number of scientific papers on Exxon’s efforts to mislead the public on climate change, told Unearthed: “I don’t believe the company has ever publicly acknowledged its role in climate denial… to have active employees of the company acknowledge its past behaviour is significant and certainly relevant to ongoing litigation and investigations against the company.” Supran continued: “The company has thrown up the straw man argument that they never hid or ‘covered up’ the science, but that’s never actually been our point. The point is that they misled the public about climate change by contributing quietly to climate science but loudly to promoting doubt about it, so what this person said seems exactly consistent with that.” A new playbook By the start of the Obama administration in January 2009, it had become untenable for ExxonMobil to continue to publicly cast doubt on climate science. Instead, ahead of Obama’s inauguration, and in an apparent break with Exxon’s long opposition to what it called “near-term policies” on climate change, the company’s then-CEO Rex Tillerson publicly backed a carbon tax. The proposal sought to spread the cost of tackling climate change beyond the fossil fuel sector, taxing businesses in every sector of the economy for each tonne of carbon emitted – either by them directly or embedded in products they sell to consumers. The cost of this would then be passed on to the public. In making the announcement, Tillerson sketched the outlines of a new ExxonMobil playbook on climate change – one that accepts the science but nevertheless seeks to delay rapid emissions cuts. Instead of subsidising renewables, governments should prioritise research and development to discover a “breakthrough” technology. In the meantime, Exxon would remain “fundamentally an oil and gas company because we think that’s what society needs and will have to have for the next 50 years”, Tillerson said. Nobody is going to propose a tax on all Americans… A carbon tax is not going to happen. – Keith McCoy, Exxon lobbyist Crucially, Tillerson explicitly framed the company’s support for a carbon tax in opposition to the incoming Obama administration’s plan to introduce a more stringent cap and trade system. Essentially, Exxon backed a carbon tax as part of its strategy to oppose cap and trade. In 2013 – after the cap and trade bill had been defeated – Tillerson backtracked: “As to our advocacy around a carbon tax—I would not support putting a carbon tax in place today because I think we still have a lot of gains to be made through technology and other less intrusive policies.” Geoffrey Supran explained how this new playbook is a “classic shift from denialism to delayism”. “Sure, over the last decades the company has necessarily shifted its rhetoric, but the end goal remains the same and that’s inaction on climate change. This is just a continuation of their 30-year track record of acting in bad faith on climate change,” Supran continued. Taxing carbon Exxon rejects claims that its support for a carbon tax is not genuine, but Unearthed can reveal that one of its most senior lobbyists believes the company knows the policy has no chance of being enacted but nevertheless uses it as an “advocacy tool.” “Nobody is going to propose a tax on all Americans and the cynical side of me says, yeah, we kind of know that but it gives us a talking point that we can say, well what is ExxonMobil for? Well, we’re for a carbon tax,” McCoy said. When asked by the reporter, “So it’s basically never going to happen right, is the calculation?”, McCoy replied: “Yeah. No it’s not, it’s not. Carbon tax is not going to happen”. He added that other members of the oil industry that have recently announced their support for a carbon tax – such as the American Petroleum Institute (API), an influential lobby group – did so because “they’ve got nothing else, so it’s an easy talking point to say, look I’m for a carbon tax”. McCoy continued: “So that’s the talking point, that is in my mind an effective advocacy tool. Many members of Congress can say, well we don’t believe you, and we’ll say well yeah, we’ve been saying this for over a decade and we’re not new to this. API is new to this, some of these other companies are new to this, but at ExxonMobil we’ve been saying this for a decade. “I think the carbon tax is an effective way of saying to them [members of Congress]: put up or shut up.” McCoy was asked whether Exxon’s support for the “talking point” of a carbon tax makes it easier for the company to oppose more punitive climate regulations, “whilst still having this kind of bold proposal that makes sure you are still for something, not just against it?” He nodded and replied: “Well that’s the danger right, so they realised that they can’t get the sort of large bill put forth so what ends up happening is it’s death by 1,000 cuts right.” McCoy continued: “They just go through the regulatory process and they put a moratorium on federal leasing, they’ll do something on pipelines you know, they’ll look at offshore drilling, they’re looking at the royalty rates now for onshore drilling so there’s going to be each step of the way they’re going to try to cripple the oil and gas sector.” An Exxon spokesman said: “We have been clear in supporting an efficient, economy-wide price on carbon as the best way to achieve the goals of the Paris Agreement. While there is not broad support for a tax, we are actively and publicly discussing other options, including lower-carbon fuels and other sector-based approaches that would place a uniform, predictable cost on carbon.” A spokeswoman for the API said: “We’ve endorsed a host of climate actions and are advocating for a carbon price policy as the most impactful way to spur innovation and reduce emissions across all economic sectors. We’ll continue to advance technology innovation, policy solutions and industry actions to help shape a lower-carbon future.” Infrastructure plan One of the threats facing Exxon, according to McCoy, is President Biden’s proposal to pour billions of dollars into renewable energy and electric vehicles through the $2 trillion American jobs plan, his flagship initiative to tackle climate change. Unearthed can reveal that Exxon has been working hard behind the scenes to eliminate the proposed funding. As alluded to by McCoy, in the absence of major legislation limiting US emissions – such as the cap-and-trade system proposed during the Obama administration – Biden has moved to accelerate the transition to clean energy and transport through government spending. The proposals, the most ambitious clean energy legislation ever put forward by a US President, included more than $100 billion in subsidies for electric vehicles alone and would have been paid for by higher taxes on corporations like Exxon. But last week Biden endorsed an alternative plan, which eliminates the vast majority of spending on climate change after being forced into a compromise by “moderate” Democratic senators, including Joe Manchin, senator for West Virginia. Now, Unearthed can reveal that – according to one of its most senior lobbyists – ExxonMobil targeted a number of these moderate senators, with the aim of scaling back the plan’s ambition by scrapping the tax hikes that would pay for it. Speaking in early May, McCoy said: “We’re playing defence, because President Biden is talking about this big infrastructure package and he’s going to pay for it by increasing corporate taxes. So it’s a delicate balance we’re asking for help with taxes over here [lobbying for subsidies for a carbon capture project] and we’re saying, don’t increase our taxes over here.” He explained that if the plan stuck to “roads and bridges”, the budget could be reduced from $2 trillion to $800 billion, limiting the need for tax rises: “The international tax piece is for, for ExxonMobil is close to a billion dollars.” This would mean “the negative stuff starts to come out, because there’s a germaneness right… that doesn’t make any sense for a highway bill. Why would you put in something on emissions reductions, on climate change to oil refineries in a highway bill?” Targeting senators When asked which senators Exxon is lobbying on these specific points, McCoy said: “Senator [Shelley Moore] Capito [Republican senator for West Virginia]… who’s the ranking member of environment and public works. Joe Manchin, I talk to his office every week, he is the kingmaker on this because he’s a Democrat from West Virginia which is [a] very conservative state, so he is, and he’s not shy about sort of staking his claim early and completely changing the debate.” “On the Democrat side we look for the moderates on these issues”, McCoy continued, highlighting figures including Arizona senator Kyrsten Sinema; John Tester, senator for Montana; and Chris Coons, senator for Delaware, President Biden’s home state. “Senator Coons… has a very close relationship with Senator [President] Biden, so we’ve been working with his office – as a matter of fact our CEO is talking to him next Tuesday and having those conversations and just teeing it up, and then that way I can start working with his staff to let them know where we are on some of these issues,” McCoy said. He added that he targets senators like Mark Kelly, Democrat senator from Arizona, New Hampshire Democrat senator Maggie Hassan, and Florida Republican senator Marco Rubio, who are up for reelection in 2022. “I can’t worry about the 2027 class because they’re not focused on re-election. The 2022 [class] is focused on re-election so I know I have them… you can have those conversations with them because they’re a captive audience, they know they need you and I need them,” McCoy said. With the vast majority of the clean energy and transport spending now stripped from the proposed bipartisan legislation, Biden has signalled that he will attempt to address climate change through a separate budget reconciliation bill, which can be passed with a simple majority vote. But the President faces the same challenge of convincing Joe Manchin, Kyrsten Sinema – and the other senators targeted by Exxon – to support the tax increases necessary to fund the transformation in energy and transport that he originally envisaged. An Exxon spokesman said: Our discussions on the current infrastructure bill are not accurately portrayed. Our lobbying efforts are related to a tax burden that could disadvantage U.S. businesses, and we have made that position known publicly. ExxonMobil stands by our position that increased taxes on American businesses make the U.S. less competitive. “The comments describing interactions with government officials and non-governmental organizations are entirely inconsistent with the way we expect our people to engage. The individuals interviewed were not involved in developing the company’s policy positions on the issues,” he continued. Clean energy Unearthed can further reveal that Exxon’s new strategy for delaying action on climate change goes beyond opposition to specific pieces of legislation. On something like climate change, there’s the forest fires, there’s an increase of, you know, .001 Celsius… that doesn’t affect people’s everyday lives. – Keith McCoy, Exxon lobbyist A core objective for Exxon on Capitol Hill is undermining confidence that a transition to clean energy and transport is possible over the next decade, according to McCoy. “You’re not going to be able to just switch to battery operated vehicles or land for your electricity and it’s having that conversation around why that’s not possible in the next 10 years is critically important to the work that we do… and that’s at every phase: that’s in the Senate, that’s in the House, that’s with the Administration,” he said. McCoy went on to describe President Biden’s new targets to cut US greenhouse gas emissions as “insane”, before suggesting that ambitious action on climate change is unlikely to succeed, because it isn’t an existential threat like the pandemic. “Outside like, of something like Covid where there’s this existential crisis and people rally to support each other. On something like climate change there’s the forest fires, there’s an increase [of] .001 Celsius, that doesn’t affect people’s everyday lives,” he said. ‘A lot of wins’ The investigation also revealed one former Exxon lobbyist’s disbelief at the scale of influence the company had during the Trump administration. Dan Easley, who was a senior director for federal relations at Exxon until February 2021, when he joined a clean technologies firm, laughed when asked by an undercover reporter if the company had achieved many big policy wins under Trump, before outlining victories on fossil fuel permitting and the renegotiation of the NAFTA trade agreement. “You should Google ‘ExxonMobil announcement’ and ‘Donald Trump’. So he live-Facebooked from the West Wing our big drill in the Gulf project, he mentioned us in two States of the Union, we were able to get investor state dispute settlement protection in NAFTA, we were able to rationalise the permit environment and you know, get tonnes of permits out.” “The wins are such that it would be difficult to, to categorise them all. I mean, tax has to be the biggest one right, the reduction of the corporate rate was, you know, it’s probably worth billions to Exxon, so yeah there were a lot of wins,” Easley continued. ExxonMobil’s full statement is available here. Dan Easley did not respond to requests for comment.

How to Deploy ML Models with Google Cloud Run

This blog post is part of our series on Serverless Inference for Machine Learning models accompanying our KubeCon 2020 talk: Serverless for ML Inference on Kubernetes: Panacea or Folly? - Manasi Vartak, Verta Inc. We'll be hosting a live replay on December 2nd at 2 PM ET for anyone who missed it. As builders of an MLOps platform, we often get asked whether serverless is the right paradigm to deploy models. The cost savings touted by serverless seem extremely appealing for ML workloads as for other traditional workloads. However, the special requirements of ML models as related to hardware and resources can cause impediments to using serverless architectures. To provide the best solution to our customers, we ran extensive benchmarking to compare serverless to traditional computing for inference workloads. In particular, we evaluated inference workloads on different systems including AWS Lambda, Google Cloud Run, and Verta. In this series of posts, we cover how to deploy ML models on each of the above platforms and summarize our results in our benchmarking blog post. This post talks about how to get started with deploying models on Google Cloud Run, along with the pros and cons of using this system for inference. Cloud Run is a serverless application running in a container. Cloud Run allows containerized applications to run either on a managed container service or in a customer’s (k8s) kubernetes cluster via Anthos. Cloud Run is powered by Knative, which bills itself as a “Kubernetes-based platform to deploy and manage modern serverless workloads.” Knative supports various service mesh frameworks but in the case of Cloud Run is using the Istio k8s service mesh. Cloud Run allows containerized applications to support modern serverless workloads including scale-to-zero and request concurrency scaling. Cloud Run is the paid / managed version of knative / istio / k8s allowing teams to leverage modern autoscaling container technology without any of the hassle of running the infrastructure. Creating an endpoint with Google Cloud Run is very similar to creating an always-on container based deployment. But before we can deploy anything we need to perform some Google Cloud Platform setup. You will need to create a GCP account, attach billing information and create a project. If you are a new GCP user then you are in luck-- you will receive $300 in credits to use for your first year! First, let’s login and create a new gcloud project (to make cleaning things up easier) A browser should open prompting you to log into your google account. If authentication is successful you should see the google cloud console landing page in your browser. Return to the terminal and set the current project where PROJECT_ID is the project you created earlier. Set default project for further commands: Using a separate project is a good idea so that we can find everything we started in order to clean up our account and avoid perpetual charges. At this time we can also set a default region for Cloud Run. Feel free to use a different region closer to you. If you have a POSIX terminal (linux / mac / windows wsl)  you can save off the project ID for later use: Finally, let’s create a directory for some files we’ll add soon. For this blog post we will set up DistilBERT and perform question and answer predictions. There are two python source files, the bootstrap to download and save the model file and the app to serve predictions. Having a separate bootstrap script that we run before creating the endpoint step will greatly improve the cold start time of the endpoint by skipping the need to download the entire model before servicing a request. Create these files in your cloudrun_example folder: p There should be three files: This Dockerfile references two python scripts from earlier-- bootstrap.py and app.py. Because we have gcloud installed we can offload the docker build and publish to Google Cloud Build: If the command completes with a SUCCESS then the docker container has been built and published. If there is an error, then a google search of the error should lead you quickly towards the answer. We are now ready to deploy a cloud run endpoint: If successful, the terminal will return the URL of the new endpoint. e.g. https://distilbert-3bz2zvlxbq-uc.a.run.app We are now ready to make a prediction! Try it out: Now that we’ve seen how easy it is to deploy a serverless ML application to Cloud Run, let’s have some real talk on some limitations. Cloud Run, being a platform that targets a wide variety of applications comes with a set of defaults and constraints that preclude some ML workloads. For example, the max CPUs that can be requested is 4 and the max memory is 4G. When deploying a ML workload, one will probably want to dive into settings to lower container concurrency which is the number of open requests per container. The maximum response time for Google Cloud Run is 60 minutes. While there is no container size limit, containers must be able to load and start within 5 minutes. Cloud Run containers are given 10s to terminate before being forcefully terminated. When building this example, issues with long cold starts necessitated the bootstrap.py file to preload the model into the container. While running benchmarks, it was noted that Cloud Run seemed to suffer from “continuous cold starts” such that even with a very steady throughput, new instances would continuously spin up and the unlucky requestor to that instance would have to wait ~10s for a response. Because of this, cloud run is never able to truly achieve a steady state. It would appear that pods are continuously recycled. In Cloud Run there exists no configuration to customize this behavior compared to a self hosted Knative installation. Since Cloud Run is built almost entirely on open source technologies, one could choose to install and configure their own k8s cluster, ingress controllers, Istio and Knative to run their own “serverless” workloads directly. Of course to call such an installation “serverless” actually demonstrates how ridiculous the term “serverless” actually is. Of course there are servers… but when using a managed serverless platform like Cloud Run what you, as the customer, release is the responsibility of directly managing server infrastructure. Because Cloud Run is mostly built on open source technologies you always have the choice to host your own infrastructure to run similar workloads. However, this author will confirm that teams looking to go this route are in for quite the ride. Configuring and running k8s, Istio and Knative are all non-trivial tasks. When you go with a managed platform, like Cloud Run, the intricacies of these platforms are mostly hidden from the user. Self hosting will open up a world of knobs and switches that can be set. While this can be a huge advantage to teams needing to highly tune specialized workloads, for most use cases is a big deep rabbit hole of testing, tuning and more testing. Not yet afraid? Well we haven’t even covered logs, metrics, revision control and dashboards. With Cloud Run you get web and command line tools for viewing and managing your workloads. These tools are the part of the k8s / Istio / Knative stack that is not open source. Now you are on the hook to add all of these to be able to effectively manage and provide visibility to your workloads. If one insists on setting up their own infrastructure, then consider taking a look at Anthos. Anthos is yet another installation on a k8s cluster that provides a rich set of integration with Google Cloud Console. An Anthos enabled cluster can be deployed via the Google Cloud Console Cloud Run dashboard as well as connect with all of Google Cloud’s existing monitoring tools. With Anthos customers can control and customize their Knative installation while also having access to all of Google's proprietary console tools. Google Cloud Run is great for the use cases that fit within the container contract. Because Google Cloud Run is built on top of open source technologies, there is also a path towards self-hosting. In addition, since Cloud Run can scale to zero, you pay as you go. However, you will end up paying more for that compute compared to simply buying Google Cloud Compute instance outright and hosting the service directly. In this case you would also create a floor on expense because you would not be scaling to zero. Google Cloud Run, with its default settings, might be best suited for smaller HTTP workloads than machine learning. There are hard limits to CPU and memory that prevent running larger ML workloads. The default request concurrency (80) makes sense for parallelizable workloads but will create serious issues for ML workloads. Care should be taken when creating the deployable container to ensure that the container is able to start, run and stop within the constraints of Cloud Run. Stay tuned for our next blog in this series! John spent over ten years in the enterprise software/embedded/mobile development tools space before transitioning to fullstack web development for the past ten years. John currently leads the frontend team at Verta. Verta provides AI/ML model management and operations software that helps enterprise data science teams to manage inherently complex model-based products. Verta’s production-ready systems help data science and IT operations teams to focus on their strengths and rapidly bring AI/ML advances to market. Based in Palo Alto, Verta is backed by Intel Capital and General Catalyst. For more information, go to  www.verta.ai  or follow @VertaAI.

Facebook has 'tentatively friended' us again, Australia says

CANBERRA (Reuters) - Facebook Inc is back at the negotiating table, Australian Prime Minister Scott Morrison said on Saturday after the tech giant this week blocked news on its site in the country. Facebook’s abrupt decision to stop Australians from sharing news on the site and strip the pages of domestic and foreign news outlets also erased several state government and emergency department accounts, causing widespread anger. The company has “tentatively friended us again,” Morrison told a news conference in Sydney. “What I’m pleased about it that Facebook is back at the table again.” Facebook has publicly indicated no change in its opposition to a proposed law requiring social media platforms to pay for links to news content. Morrison was not asked about that. Australia’s Treasurer Josh Frydenberg said on Friday he had spoken with Facebook CEO Mark Zuckerberg and further talks were expected over the weekend. It was not clear whether those talks have happened. A Facebook spokeswoman and representatives for Frydenberg did not immediately respond to requests for comment. The stand-off comes as Australia’s vows to press ahead with the landmark legislation, which could set a global precedent as countries like Canada express interest in taking similar action. The Australian law, which would force Facebook and Alphabet Inc’s Google to reach commercial deals with Australian publishers or face compulsory arbitration, has cleared the lower house of parliament and is expected to be passed by the Senate within the next week. Simon Milner, Facebook’s Asia-Pacific policy director of policy for the Asia-Pacific region, was quoted on Saturday as telling the Sydney Morning Herald the company had three main objections to the legislation. Facebook objects to being barred from discriminating between different news outlets that ask for money, to arbitration models that allow an independent body to select one payment over another, and to the obligation to enter commercial negotiations with Australian media companies, Milner said. Facebook declined to make Milner available to speak with Reuters. Australia’s legislation is being widely watched overseas. Canadian Heritage Minister Steven Guilbeault said on Thursday his country would adopt the Australian approach as it crafts its own legislation in coming months. Google, which has initially threatened to close its search engine in Australia, has announced host of preemptive licensing deals over the past week, including a global agreement with News Corp. Facebook’s move had an immediate impact on traffic to Australian new sites, according to early data from New York-based analytics firm Chartbeat. Total traffic to the Australian news sites from various platforms fell from the day before the ban by around 13% within the country. Reporting by Colin Packham; Editing by William Mallard Our Standards: The Thomson Reuters Trust Principles.

Stages of Self-Development

What if we had a map that could show us where we are in the self-development journey? This could help us find out what we need to focus on right now to improve ourselves. We could identify the bottleneck of our improvement easily. Instead of focusing on thousand small things and failing like most people do, we could find one or two most painful issues, fix these, and then move on to the next stage. In bodybuilding and strength training, we understand the differences between beginner, intermediate, and advanced athletes. Each level requires a different training regimen. However, in self-development, no one distinguishes between whether people are beginners, intermediates, or advanced practitioners. This only leads to frustration because people focus on the wrong things at the wrong time. For example, you read Getting Things Done. You think it’s great. You try it for a couple of days but then you fail. Why? Because the book is meant for intermediates and you’re a beginner. As a beginner, you probably struggle with being consistent and not avoiding work in the first place. It’s the same as trying to work on little nuances at the bottom of the squat when the trainee has the main issue of not going to the gym regularly. His squat technique can be perfect but if he avoids the gym, it doesn’t matter. We’re all somewhere on this spectrum. No one is statically positioned. We can move on the spectrum every day by our actions. We can count the first 3 stages as beginners. Intermediates are Motivated and Achievers. Advanced are High Achievers. As a beginner what you do is way more important than how you do it. In the first 2 stages, it’s necessary to get out of a rut and start doing something. If you start doing something and get excited about it, you can quite quickly catapult yourself to the Motivated stage, skipping the Unmotivated stage altogether. This is the worst starting position in self-development. We don’t usually begin our journey here but some of us slip here. It can be a very difficult starting position because people often don’t care and don’t have any motivation to do anything. When you’re depressed, you might even assume that only now you think clearly. All the hopes and dreams you had were pathetic. Only now you can see the world as it truly is. This is a frequent perception distortion. Not all episodes of depression are equal and they differ in severity. Please get professional help. Some people might be genetically predisposed and that’s why you need all the tools available to get out of this state. Get professional help from a psychiatrist who won’t just give you pills and send you home. If you know someone who might be clinically depressed, please talk to them and try to get them to seek professional help. Other episodes of depression aren’t that severe and don’t last long. Usually, if you feel down for less than two weeks and you see glimpses of hope here and there, it’s not that serious. Serious depression is when you can’t see any hope in sight. You might show weak symptoms of being depressed. These are called moods and the good thing is they’re fleeting and can go away fast. Especially, if you help the process by achieving something. If we were to use the gym metaphor again, the equivalent of the Depressed stage is being obese and spending all one’s time in bed or on the couch. To move forward, the person has to start walking. The goal here is to spark any motivation at all. Any motivation to take a step forward, and then another step, and another. That can range from the smallest things like taking care of your hygiene regularly or cleaning your room to bigger things like starting to exercise or work. The good news is, depending on the strength of your chronic procrastination habit, if you start moving forward, and you start feeling good about it, you can skip the other 2 beginner stages entirely and jump straight to intermediates. I just don't care about myself Why should you even bother improving yourself? (Not a fan, but it’s a smart message.) This is another beginner stage. Most people also don’t start here. However, many people slip here during high-school, college, or a particularly boring or stressful period at work. We explained chronic procrastination here. Basically, being a chronic procrastinator means that the habit of procrastination is strongly ingrained. Most chronic procrastinators live lives haunted by regret and shame and can't enjoy leisure. Chronic procrastination for a longer period of time can lead to depression. People in this stage often show symptoms of being depressed because they’re unsatisfied with their lives. The main difference between chronic procrastinators and depressed people is that chronic procrastinators want to change themselves. They might even have glimmers of temporary motivation to do so. However, most of the time they try, they fail, and dig themselves into an even deeper hole than before. They usually fail because they take on many things at once, focus on wrong areas, or self-sabotage themselves because they get scared (of responsibility, failure…). Following the gym analogy, a chronic procrastinator is like a chubby, out-of-shape person that wants to start working out but doesn’t know how. To move forward, the person has to start exercising regularly, no matter how small the feat is (10 pushups per day). To get out of this stage, there are the two main goals: 1. Reduce distractions You have to reduce distractions to have room to work on yourself. Distractions like games can be so captivating that people can always escape and stop thinking about what’s happening in their real life. If you spend 8 hours playing games, it’s hard to make any progress in any area of your life. 2. Start showing up consistently (no matter how small the steps are) That can mean writing 50 words every day if you’re a writer, studying for 10 minutes every single day if you’re a student, not avoiding the most important responsibility if you’re an employee. It can be whatever you select it to be. If this works consistently, you can start exploring the reasons why you procrastinate in the first place. What is chronic procrastination and how to overcome it Reading our guide Watch this short talk on how small steps can add up Unmotivated people are still beginners, but unlike chronic procrastinators they appear to function normally on the outside. By unmotivated, we mean lacking internal motivation. People here function by getting motivated externally (getting a diploma, getting a salary, etc). They do things because they have to. This is the stage where most people spend the majority of their lives. Unmotivated people sometimes stumble upon glimpses of internal motivation. For example, when they start a new hobby or a new project they get excited about it. However, their habits usually aren’t strong enough to sustain the initial motivation and without a positive feedback loop, they fall back into their old unproductive ways. Due to these reasons, their motivation is temporary and dissipates after the initial excitement fades. Unmotivated people often struggle with procrastination in all forms. They resent work most of the time and feel like they HAVE TO do this and that. In gym terms, this is a person who “has” to go to the gym because someone told them it’s good for them, but they don’t have any positive feelings towards exercising. It’s a chore for them. If the person wants to move forward, the goal is to go to the gym regularly and slowly start liking it. That might entail finding physical activities that they enjoy. In the Unmotivated stage, there are 3 main goals: 1. Reduce distractions Similar to chronic procrastinators, an unmotivated person needs to cut distractions. Distractions can serve as an easy form of escapism that let people forget about their own lives. This is often the reason why people spend their lives in the unmotivated stage. They kill all their waking time by being distracted, which doesn’t let them think about how their life is going. Don’t believe us? Go sit in your room staring at a wall in silence for an hour not doing anything. You will get bored quickly and will want to do something like grab your phone or stand up and go somewhere. Anything to escape looking inward and grappling with the issues on your mind. People simply hate getting bored. When you reduce distractions, you will feel uncomfortable and want to do something. The key is to change the meaning of something from getting distracted to improving yourself (in any way you can manage). If you reduce and block distractions, this will become easier because now the choices are to be bored or do something productive, instead of to get distracted or do something productive. If you want to know more about this, check out this blog post. 2. Find your internal motivation If you want to get motivated for the long term, you want to find ways to rely on internal motivation. That means finding personal reasons why you want to improve. It also means that you want to discover or create a positive feedback loop. What’s that? It’s when you feel satisfied after working hard and you can clearly see the connection between your effort and how it impacts your life. For more on finding motivation, read this short article. 3. Establish regular working habits Showing up consistently is something that will make your progress almost guaranteed. Once you do this, the work usually becomes more exciting because you start making visible progress. In this stage, you need to learn how to start and how to start often. That also means resolving your fears and insecurities. You can still live with them but resolving these will make starting and working easier. You also need to learn to enjoy deserved leisure so you don’t have to fear infinite dreary work. Another part of establishing regular working habits is using a calendar on a regular basis and planning your days. These behaviors aren’t viable in the previous stages, because people usually stop looking at their calendar to avoid responsibilities or don’t follow the plans for the day. But once you’re in this stage, you know that to be a responsible person, you need some form of reminders to actually accomplish things on a consistent basis. Fear of infinite work Read our guide Intermediates start to stumble upon a new problem. For beginners, the main problem is doing things they planned or set out to do. Intermediates don’t struggle with doing things most of the time and they’re better at starting, resolving fears, and not giving in to instant gratification. Instead, they struggle with the issue of how to do more or work more effectively in general. People at this stage have already discovered a positive feedback loop that fuels their internal motivation. They don’t dread work like people at the stages before. They understand that difficult tasks have to be done to succeed. They also like how good it feels to be disciplined. This is not to say they’re successful all the time, but people here are much better at saying no to immediate gratification. In gym terms, these are talented athletes. They’re showing a lot of potential but plenty of work awaits them to reach full potential. To get more out of their training, they need to start focusing on the little details. 1. Create or maintain a positive feedback loop The main goal here is to get addicted to the positive feedback loop work brings you. You already know how satisfying it is when you start achieving things. You can more easily reject instant gratification because you know making progress feels better. Setting goals and tracking them helps maintain the positive feedback loop. 2. Start working on how you do things You know you can achieve plenty of things, but to be able to do them you need external help. You take on too much and want to build your ability to rely on yourself. That’s where things like the GTD system start making sense. Motivated people can start playing with productivity systems. They have enough self-confidence to use them consistently enough to be able to compare and see the differences. They start to see why all the weekly reviewing, goal planning, and note taking aren’t a waste of time. Read GTD or any methodology that will suit you to have a mind like water. Read How to take smart notes Be sure to check out Forte Labs At this stage, people don’t struggle with what they do. They know how to begin, how to divide a large task into smaller subtasks they can complete, and they can successfully overcome fear most of the time. If they have fears, stress, and anxiety they acknowledge them but they don’t act on them. They very rarely choose something that isn’t in their best interests. They know that the long term choice will bring them more happiness through delayed gratification. This stage becomes fun because people get excited about what they do and about the future possibilities. For achievers, the bottleneck of productivity is how they do things. They have reached the upper limits of what they can do with the current methods they’ve adopted or come up with. To expand their horizons further, they need to start looking for outside methodologies to boost their productivity. When they start optimizing things, it can appear to onlookers that these people are wackos. The reality is, they’re showing up every day and want to get more out of every second of their time and energy. However, these people can still struggle with self-confidence. This is called imposter syndrome. They’re good at what they do, but don’t they know they're good. Many achievers underestimate how effective they are compared to other people and can feel insecure about their abilities and efforts. In gym terms, this person is an amateur competitor who always places in the top 3. If he keeps doing what he is doing, he might become a professional. 1. Find new methodologies and study the nuances Now you can properly test GTD, personal knowledge management, and note-taking. Not that you can’t function without these, but you should see the reasons why these methodologies allow you to do more. Try out these external productivity systems or various tips you find to see which methods work best for you and allow you to do more than you otherwise would. Read GTD or any methodology that will suit you to have a mind like water. Read How to take smart notes Be sure to check out Forte Labs If you’re here, congrats! You’re the 1%. You probably live a satisfying life. This stage is what internal motivation can lead to. High achievers appear to other people like gods of discipline. But they aren’t. They just established really good working habits and love the satisfaction their work brings them. They really enjoy what they do and might say they’re following their passion. When high achievers focus on something, they get great at it. The main difference between high achievers and achievers is that they’ve been doing it for longer. High achievers have true self-confidence. They know what they stand for and can achieve anything they set out to do because they’ve proved it many times to themselves and to others. Often this self-confidence is demonstrated by their humility and quiet certitude about their abilities. In gym terms, this person is a pro competitor. strong You can usually double down on knowledge management and note-taking. You benefit the most from these because it fuels your curiosity even more. If you feel you have any gaps there, be sure to check out Forte Labs. Or maybe read this. 2. Experiment Based on your knowledge of what you’re good at and how you operate, try out new techniques or variations of someone else’s methods. Experiment and find what fits into your productivity system and what doesn’t. 3. Prioritize, focus on high impact activities, delegate If you get to the high achiever level, it becomes crucial to focus on the most important activities first. You can focus often and intensely, but if you focus on the wrong things, you’ll limit yourself. Start here. At this stage, it also makes sense to hire others to do less important busywork. That’s it. While reading, you probably found the stage you’re currently at. You can now identify more easily what to focus on to improve yourself. Use this framework to improve your life today.

.NET Open Source: What Happens When the Free Lunch Ends?

It’s a Thursday, which means: .NET open source drama. Last October, Dominick Baier and Brock Allen, the two creators and maintainers of IdentityServer, announced that IS’s current business model was inherently unsustainable and they’d be moving to a paid licensing model using the Reciprocal Public License (RPL) and under a new company, Duende Software, beginning with IdentityServer5. Last month Microsoft announced that they were going to continue to include Duende’s IdentityServer dependency in their templates for ASP.NET 6 - IdentityServer had, historically, been a free open source product licensed under the permissive Apache 2.0 license and has been a popular choice among ASP.NET developers for handling OpenID and OAuth 2.0 tokens, hence why they’ve been included in so many of Microsoft’s default ASP.NET templates for years and years. Now the version of IdentityServer being included in Microsoft’s popular templates requires that users earning more than $1m per year pay license fees as low as $1,500 per year. The .NET community responded to this announcement graciously; they took a moment to recognize their role in making the .NET open source ecosystem more innovative by supporting independent software vendors; and agreed that having IdentityServer maintain itself through recurring license fees was a highly preferable alternative to having the maintainers abandon the project. Just kidding. The .NET community collectively freaked out and demanded to speak with the manager of .NET OSS - over what’s included in some templates. strong. The worst part is that apparently the .NET OSS ecosystem does still, in fact, have a manager: Microsoft. I went into some detail on this on my Twitch stream, and I’ve included a highlight from my YouTube channel (both of those are new!) where we discuss the details: However, since this thread and discussion around Microsoft, IdentityServer, and the role of free vs. paid “open source” software refuses to cease - I think it’s worth exploring the “end of the free lunch” for .NET OSS users. In the context of people exploiting a free / under-priced / unauthorized resources, there is an expression I love: “little piggies get to live another day; hogs get slaughtered.” When it comes to theft: someone who shoplifts a candy-bar from a convenience store probably won’t be prosecuted with the full resources of the plaintiff and law enforcement but someone who robs a bank might. When it comes to open source software: it’s inexpensive for maintainers to support a small number of users with relatively similar demands - but once a project achieves critical mass and the demand on the maintainers exceeds their desire to supply, something will have to give. IdentityServer’s users are hogs and it’s off to the chopping block. A less grotesque analogy: most IdentityServer users been dining greedily on Dominick and Brock’s tab for the better part of 10 years and now the bill has come due. Inevitably, the thread with Microsoft is full of hogs squealing: The hysterics over what Microsoft chooses to include in some templates, not a core library that will render the RPL terms viral for end users, is as absurd as it is inevitable. Suddenly when asked to pay $1,500, $4,000, or whatever per year for a service that is “essential to our business,” per the words of some of these commenters, these developers suddenly plead poverty. You can rarely buy developer expertise with a credit card - paying for an excellent, battle-tested, well-documented, and highly re-usable solution like IdentityServer built by domain experts is not only significantly cheaper than paying your own developers to do it but it’s also inherently lower risk. The failure costs of getting something as critical as authentication and authorization wrong in your application can be catastrophic. If you’re in charge of this area of your company’s software and you’re agonizing over the dollar-cost of a Duende license, please do your company a favor and fire yourself from that position: you’re not qualified for it. I suspect the reason the chumps on the thread are squealing about licensing costs and playing poor has nothing to do with the cost and everything to do with dealing with their procurement department. One of the greatest reasons why open source technology spreads so quickly and accrues so much value: it is permissionless - anyone can adopt, use, modify, and redistribute a vetted piece of open source software without having to encounter the department budget. But once maintainers affix a dollar amount as the entry fee to benefit from all of their institutionalized knowledge and expertise developers now have no choice other than violating the license terms (legal won’t stand for that) or dealing with the procurement bureaucracy to allocate company money for the purchase. The raison d’être of the procurement bureaucracy is to thoroughly vet every vendor in the supply chain for “risk” - a largely performative song and dance that involves Dun & Bradstreet numbers, certificates of general liability insurance, and insisting on getting the governing law of the statement of work changed from wherever the vendor is to wherever the buyer is. The procurement bureaucracy typically does not produce any meaningful outcomes other than making it more expensive and difficult for both parties to transact with each other, hence the disincentive for software developers to engage it. It is a slow-going experience that requires developers to engage in frightening acts like “cost / benefits justification” and “writing emails.” Thus the white-hot rage in the Microsoft “I want to speak to the manager!” thread - Duende has now put these .NET developers in a position where they must justify a frankly trivial dollar-cost to the procurement bureaucracy and Microsoft doesn’t care. “HOW CAN YOU NOT CARE?!?!” I’m not particularly sad at the plight of these .NET developers - this is mostly their fault, after all. You can only be a free rider for so long before the tab-payer takes notice - and when they do, you’re at their mercy. In the case of IdentityServer, it’s being asked to pay for new releases of the product under (my opinion) very generous terms - with OSS support for IdentityServer ongoing through November 2022 still! In the case of other projects, total abandonment and leaving you holding the bag. I’m personally dealing with the latter scenario with DotNetty, a critical part of our networking stack now abandoned by Microsoft once their technology choices for Azure IoT (the product it was built for) changed. It’s unpleasant but we have the resources and expertise to migrate to / create something else to fill that gap. Most companies do not. Yours very likely does not. OSS is becoming increasingly popular in the .NET ecosystem and that trend will only accelerate over time - and so you should expect the sustainability problem to become more common in .NET, not less. Begging Microsoft to answer every possible question any user with any amount of money might ask with a free library is what turned the .NET ecosystem into a flaming pile of dogshit years ago. We’re never going back. Your free lunch is already over - this is your wake-up call. When you select packages and technologies to maintain and build your .NET applications, start pricing in the expectation to pay money for it - because that’s the only way to avoid surprises and supply chain shocks in the future: by pricing them in today. Get in the habit of sending value back upstream to your dependencies. That value can be in form of contributing to the projects you use, monthly donations, or even better: buying value-added products and services from the maintainers. Other projects might want to help promote the project with blog posts, videos, and PluralSight courses. Maybe a testimonial from your company might help! There are lots of ways to give value back to the people who build the components you use to help run your business software - and most users choose none. Creating virtuous cycles where you continuously exchange value with OSS producers is the inevitable conclusion to the “Open Source Sustainability Crisis” - and everyone will be better off for it. So you should start the conversation with your team and find some projects to support - because it’s in your own self-interest to see them sustained.

Solipsism

b (; from Latin i 'alone', and i 'self') [1] is the philosophical idea that only one's mind is sure to exist. As an epistemological position, solipsism holds that knowledge of anything outside one's own mind is unsure; the external world and other minds cannot be known and might not exist outside the mind. There are varying degrees of solipsism that parallel the varying degrees of skepticism: Metaphysical solipsism is a variety of solipsism. Based on a philosophy of subjective idealism, metaphysical solipsists maintain that the self is the only existing reality and that all other realities, including the external world and other persons, are representations of that self, and have no independent existence.[ p ] There are several versions of metaphysical solipsism, such as Caspar Hare's egocentric presentism (or perspectival realism), in which other people are conscious, but their experiences are simply not present. Epistemological solipsism is the variety of idealism according to which only the directly accessible mental contents of the solipsistic philosopher can be known. The existence of an external world is regarded as an unresolvable question rather than actually false. [2] Further, one cannot also be certain as to what extent the external world exists independently of one's mind. For instance, it may be that a God-like being controls the sensations received by mind, making it appear as if there is an external world when most of it (excluding the God-like being and oneself) is false. However, the point remains that epistemological solipsists consider this an "unresolvable" question. [2] Methodological solipsism is an agnostic variant of solipsism. It exists in opposition to the strict epistemological requirements for "knowledge" (e.g. the requirement that knowledge must be certain). It still entertains the points that any induction is fallible. Methodological solipsism sometimes goes even further to say that even what we perceive as the brain is actually part of the external world, for it is only through our senses that we can see or feel the mind. Only the existence of thoughts is known for certain. Methodological solipsists do not intend to conclude that the stronger forms of solipsism are actually true. They simply emphasize that justifications of an external world must be founded on indisputable facts about their own consciousness. The methodological solipsist believes that subjective impressions (empiricism) or innate knowledge (rationalism) are the sole possible or proper starting point for philosophical construction. [3] Often methodological solipsism is not held as a belief system, but rather used as a thought experiment to assist skepticism (e.g. Descartes' Cartesian skepticism).[ p ] Denial of material existence, in itself, does not constitute solipsism. A feature of the metaphysical solipsistic worldview is the denial of the existence of other minds. Since personal experiences are private and often considered ineffable, another being's experience can be known only by analogy. Philosophers try to build knowledge on more than an inference or analogy. The failure of Descartes' epistemological enterprise brought to popularity the idea that all certain knowledge may go no further than "I think; therefore I exist" [4] without providing any real details about the nature of the "I" that has been proven to exist. The theory of solipsism also merits close examination because it relates to three widely held philosophical presuppositions, each itself fundamental and wide-ranging in importance: [4] To expand on the second point, the conceptual problem here is that the previous assumes mind or consciousness (which are attributes) can exist independent of some entity having this attribute (a capability in this case), i.e., that an attribute of an existent can exist apart from the existent itself. If one admits to the existence of an independent entity (e.g., the brain) having that attribute, the door is open to an independent reality. (See Brain in a vat) Some people hold that, while it cannot be proven that anything independent of one's mind exists, the point that solipsism makes is irrelevant. This is because, whether the world as we perceive it exists independently or not, we cannot escape this perception, hence it is best to act assuming that the world is independent of our minds. (See Falsifiability and testability below) [5] However, being aware simply acknowledges its existence; it does not identify the actual creations until they are observed by the user. Origins of solipsist thought are found in Greece and later Enlightenment thinkers such as Thomas Hobbes [6] [7] and Descartes. Solipsism was first recorded by the Greek presocratic sophist, Gorgias (c. 483–375 BC) who is quoted by the Roman sceptic Sextus Empiricus as having stated: [8] Much of the point of the sophists was to show that objective knowledge was a literal impossibility. The foundations of solipsism are in turn the foundations of the view that the individual's understanding of any and all psychological concepts (thinking, willing, perceiving, etc.) is accomplished by making an analogy with their own mental states; i.e., by abstraction from inner experience. And this view, or some variant of it, has been influential in philosophy since Descartes elevated the search for incontrovertible certainty to the status of the primary goal of epistemology, whilst also elevating epistemology to "first philosophy".[ p ] Portrait of George Berkeley by John Smybert, 1727 George Berkeley's arguments against materialism in favour of idealism provide the solipsist with a number of arguments not found in Descartes. While Descartes defends ontological dualism, thus accepting the existence of a material world ( res extensa ) as well as immaterial minds ( res cogitans ) and God, Berkeley denies the existence of matter but not minds, of which God is one. [9] One of the most fundamental debates in philosophy concerns the "true" nature of the world—whether it is some ethereal plane of ideas or a reality of atomic particles and energy. Materialism [10] posits a real "world out there", as well as in and through us, that can be sensed—seen, heard, tasted, touched and felt, sometimes with prosthetic technologies corresponding to human sensing organs. (Materialists do not claim that human senses or even their prosthetics can, even when collected, sense the totality of the universe; simply that they collectively cannot sense what cannot in any way be known to us.) Materialists do not find this a useful way of thinking about the ontology and ontogeny of ideas, but we might say that from a materialist perspective pushed to a logical extreme communicable to an idealist, ideas are ultimately reducible to a physically communicated, organically, socially and environmentally embedded 'brain state'. While reflexive existence is not considered by materialists to be experienced on the atomic level, the individual's physical and mental experiences are ultimately reducible to the unique tripartite combination of environmentally determined, genetically determined, and randomly determined interactions of firing neurons and atomic collisions. For materialists, ideas have no primary reality as essences separate from our physical existence. From a materialist perspective, ideas are social (rather than purely biological), and formed and transmitted and modified through the interactions between social organisms and their social and physical environments. This materialist perspective informs scientific methodology, insofar as that methodology assumes that humans have no access to omniscience and that therefore human knowledge is an ongoing, collective enterprise that is best produced via scientific and logical conventions adjusted specifically for material human capacities and limitations.[ p ] Modern idealists believe that the mind and its thoughts are the only true things that exist. This is the reverse of what is sometimes called "classical idealism" or, somewhat confusingly, "Platonic idealism" due to the influence of Plato's theory of forms (εἶδος eidos or ἰδέα idea) which were not products of our thinking. [11] The material world is ephemeral, but a perfect triangle or "beauty" is eternal. Religious thinking tends to be some form of idealism, as God usually becomes the highest ideal (such as neoplatonism). [10] [12] [13] On this scale, solipsism can be classed as idealism. Thoughts and concepts are all that exist, and furthermore, only the solipsist's own thoughts and consciousness exist. The so-called "reality" is nothing more than an idea that the solipsist has (perhaps unconsciously) created. There is another option: the belief that both ideals and "reality" exist. Dualists commonly argue that the distinction between the mind (or 'ideas') and matter can be proven by employing Leibniz's principle of the identity of indiscernibles , which states that if two things share exactly the same qualities, then they must be identical, as in indistinguishable from each other and therefore one and the same thing. Dualists then attempt to identify attributes of mind that are lacked by matter (such as privacy or intentionality) or vice versa (such as having a certain temperature or electrical charge). [14] [15] One notable application of the identity of indiscernibles was by René Descartes in his Meditations on First Philosophy . Descartes concluded that he could not doubt the existence of himself (the famous cogito ergo sum argument), but that he could doubt the (separate) existence of his body. From this, he inferred that the person Descartes must not be identical to the Descartes body since one possessed a characteristic that the other did not: namely, it could be known to exist. Solipsism agrees with Descartes in this aspect, and goes further: only things that can be known to exist for sure should be considered to exist. The Descartes body could only exist as an idea in the mind of the person Descartes. [16] [17] Descartes and dualism aim to prove the actual existence of reality as opposed to a phantom existence (as well as the existence of God in Descartes' case), using the realm of ideas merely as a starting point, but solipsism usually finds those further arguments unconvincing. The solipsist instead proposes that their own unconscious is the author of all seemingly "external" events from "reality". Philosophy of Schopenhauer The World as Will and Representation is the central work of Arthur Schopenhauer. Schopenhauer saw the human will as our one window to the world behind the representation, the Kantian thing-in-itself. He believed, therefore, that we could gain knowledge about the thing-in-itself, something Kant said was impossible, since the rest of the relationship between representation and thing-in-itself could be understood by analogy as the relationship between human will and human body. The idealist philosopher George Berkeley argued that physical objects do not exist independently of the mind that perceives them. An item truly exists only as long as it is observed; otherwise, it is not only meaningless but simply nonexistent. Berkeley does attempt to show things can and do exist apart from the human mind and our perception, but only because there is an all-encompassing Mind in which all "ideas" are perceived – in other words, God, who observes all. Solipsism agrees that nothing exists outside of perception, but would argue that Berkeley falls prey to the egocentric predicament – he can only make his own observations, and thus cannot be truly sure that this God or other people exist to observe "reality". The solipsist would say it is better to disregard the unreliable observations of alleged other people and rely upon the immediate certainty of one's own perceptions. [18] Rationalism is the philosophical position that truth is best discovered by the use of reasoning and logic rather than by the use of the senses (see Plato's theory of forms). Solipsism is also skeptical of sense-data. The theory of solipsism crosses over with the theory of the philosophical zombie in that other seemingly conscious beings may actually lack true consciousness, instead they only display traits of consciousness to the observer, who may be the only conscious being there is. Falsifiability and testability Solipsism is not a falsifiable hypothesis as described by Karl Popper: there does not seem to be an imaginable disproof. [19] According to Popper: a hypothesis that cannot be falsified is not scientific, and a solipsist can observe "the success of sciences" (see also no miracles argument). One critical test is nevertheless to consider the induction from experience that the externally observable world does not seem, at first approach, to be directly manipulable purely by mental energies alone. One can indirectly manipulate the world through the medium of the physical body, but it seems impossible to do so through pure thought (psychokinesis). It might be argued that if the external world were merely a construct of a single consciousness, i.e. the self, it could then follow that the external world should be somehow directly manipulable by that consciousness, and if it is not, then solipsism is false. An argument against this states that this argument is circular and incoherent. It assumes at the beginning a "construct of a single consciousness" meaning something false, and then tries to manipulate the external world that it just assumed was false. Of course this is an impossible task, but it does not disprove solipsism. It is simply poor reasoning when considering pure idealized logic and that's why David Deutsch states that when also other scientific methods are used (not only logic) solipsism is "indefensible", also when using the simplest explanations: "If, according to the simplest explanation, an entity is complex and autonomous, then that entity is real." [20] The method of the typical scientist is naturalist: they first assume that the external world exists and can be known. But the scientific method, in the sense of a predict-observe-modify loop, does not require the assumption of an external world. A solipsist may perform a psychological test on themselves, to discern the nature of the reality in their mind – however David Deutsch uses this fact to counter-argue: "outer parts" of solipsist, behave independently so they are independent for "narrowly" defined (i) self. [20] A solipsist's investigations may not be proper science, however, since it would not include the co-operative and communitarian aspects of scientific inquiry that normally serve to diminish bias. Solipsism is a form of logical minimalism. Many people are intuitively unconvinced of the nonexistence of the external world from the basic arguments of solipsism, but a solid proof of its existence is not available at present. The central assertion of solipsism rests on the nonexistence of such a proof, and strong solipsism (as opposed to weak solipsism) asserts that no such proof can be made. In this sense, solipsism is logically related to agnosticism in religion: the distinction between believing you do not know, and believing you could not have known. However, minimality (or parsimony) is not the only logical virtue. A common misapprehension of Occam's razor has it that the simpler theory is always the best. In fact, the principle is that the simpler of two theories of equal explanatory power is to be preferred. In other words: additional "entities" can pay their way with enhanced explanatory power. So the naturalist can claim that, while their world view is more complex, it is more satisfying as an explanation. Some developmental psychologists believe that infants are solipsistic, and that eventually children infer that others have experiences much like theirs and reject solipsism. [21] The earliest reference to Solipsism is found in the ideas in Hindu philosophy in the Brihadaranyaka Upanishad, dated to early 1st millennium BC. [22] The Upanishad holds the mind to be the only god and all actions in the universe are thought to be a result of the mind assuming infinite forms. [23] After the development of distinct schools of Indian philosophy, Advaita Vedanta and Samkhya schools are thought to have originated concepts similar to solipsism.[ p ] Advaita is one of the six most known Hindu philosophical systems and literally means "non-duality". Its first great consolidator was Adi Shankaracharya, who continued the work of some of the Upanishadic teachers, and that of his teacher's teacher Gaudapada. By using various arguments, such as the analysis of the three states of experience—wakefulness, dream, and deep sleep, he established the singular reality of Brahman, in which Brahman, the universe and the Atman or the Self, were one and the same. One who sees everything as nothing but the Self, and the Self in everything one sees, such a seer withdraws from nothing. For the enlightened, all that exists is nothing but the Self, so how could any suffering or delusion continue for those who know this oneness? The concept of the Self in the philosophy of Advaita could be interpreted as solipsism. However, the theological definition of the Self in Advaita protect it from true solipsism as found in the west. Similarly, the Vedantic text Yogavasistha, escapes charge of solipsism because the real "I" is thought to be nothing but the absolute whole looked at through a particular unique point of interest. [24] It is mentioned in Yoga Vasistha that “…..according to them (we can safely assume that them are present Solipsists) this world is mental in nature. There is no reality other than the ideas of one’s own mind. This view is incorrect, because the world cannot be the content of an individual’s mind. If it were so, an individual would have created and destroyed the world according to his whims. This theory is called atma khyati – the pervasion of the little self (intellect). [25] Yoga Vasistha - Nirvana Prakarana - Uttarardha (Volume - 6) Page 107 by Swami Jyotirmayananda Samkhya philosophy, which is sometimes seen as the basis of Yogic thought, [26] adopts a view that matter exists independently of individual minds. Representation of an object in an individual mind is held to be a mental approximation of the object in the external world. [27] Therefore, Samkhya chooses representational realism over epistemological solipsism. Having established this distinction between the external world and the mind, Samkhya posits the existence of two metaphysical realities Prakriti (matter) and Purusha (consciousness). Some interpretations of Buddhism assert that external reality is an illusion, and sometimes this position is [mis]understood as metaphysical solipsism. Buddhist philosophy, though, generally holds that the mind and external phenomena are both equally transient, and that they arise from each other. The mind cannot exist without external phenomena, nor can external phenomena exist without the mind. This relation is known as "dependent arising" (pratityasamutpada). The Buddha stated, "Within this fathom long body is the world, the origin of the world, the cessation of the world and the path leading to the cessation of the world". [28] Whilst not rejecting the occurrence of external phenomena, the Buddha focused on the illusion created within the mind of the perceiver by the process of ascribing permanence to impermanent phenomena, satisfaction to unsatisfying experiences, and a sense of reality to things that were effectively insubstantial. Mahayana Buddhism also challenges the illusion of the idea that one can experience an 'objective' reality independent of individual perceiving minds. From the standpoint of Prasangika (a branch of Madhyamaka thought), external objects do exist, but are devoid of any type of inherent identity: "Just as objects of mind do not exist [inherently], mind also does not exist [inherently]". [29] In other words, even though a chair may physically exist, individuals can only experience it through the medium of their own mind, each with their own literal point of view. Therefore, an independent, purely 'objective' reality could never be experienced. The Yogacara (sometimes translated as "Mind only") school of Buddhist philosophy contends that all human experience is constructed by mind. Some later representatives of one Yogacara subschool (Prajnakaragupta, Ratnakīrti) propounded a form of idealism that has been interpreted as solipsism. A view of this sort is contained in the 11th-century treatise of Ratnakirti, "Refutation of the existence of other minds" (Santanantara dusana), which provides a philosophical refutation of external mind-streams from the Buddhist standpoint of ultimate truth (as distinct from the perspective of everyday reality). [30] In addition to this, the Bardo Thodol, Tibet's famous book of the dead, repeatedly states that all of reality is a figment of one's perception, although this occurs within the "Bardo" realm (post-mortem). For instance, within the sixth part of the section titled "The Root Verses of the Six Bardos", there appears the following line: "May I recognize whatever appeareth as being mine own thought-forms"; [31] there are many lines in similar ideal. Anathema Antiscience Aseity Alfred Binet – The mind and the brain Cartesian doubt Centered world Cognitive closure (philosophy) Consensus reality Cotard delusion - the opposite Dream argument Eliminative materialism - the idea that even aspects of one’s mind may not be sure to exist Ethical solipsism Existential nihilism Externism Heinlein's World as Myth Henry Rollins's Solipsist Immaterialism LaVeyan Satanism Metaphysical nihilism Mind over matter Model-dependent realism Object permanence Objective idealism Open individualism Panpsychism Personal horizon Phaneron Phenomenalism Philosophical realism Postmodernism Post-structuralism Primary/secondary quality distinction – John Locke's response to solipsism Problem of other minds Protagoras of Abdera Solipsism syndrome Stream of consciousness Subjectivity The Egg The Truman Show delusion Vertiginous question ^ ^ a b ^ ^ a b ^ ^ ^ ^ ^ ^ a b ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ a b Deutsch, David. (1997) Fabric of Reality ^ ^ ^ Krishnananda, (Swami). The Brihadaranyaka Upanishad. Divine Life Society, Rishikesh. P. 248. ^ O'Flaherty, Wendy Doniger. Dreams, Illusion, and Other Realities. University of Chicago, 1984. pp. 120–1. ISBN 0-226-61855-2. ^ ^ Radhakrishnan, Indian Philosophy, London, George Allen & Unwin Ltd., 1971 edition, Volume II, p. 342. ^ Isaac, J. R.; Dangwal, Ritu; Chakraborty, C. Proceedings. International conference on cognitive systems (1997). Allied Publishers Ltd. pp. 341–2. ISBN 81-7023-746-7. ^ ^ Chandrakirti, Guide to the Middle Way 6:71cd, translation in Ocean of Nectar: Wisdom and Compassion in Mahayana Buddhism, London: Tharpa Publications, p. 253. ^ ^

DuckDuckGo reveals it picked up $100M in secondary investment last year

Privacy tech continues cooking on gas. To wit: Non-tracking search engine DuckDuckGo has just revealed that it beefed up its balance sheet at the back end of last year with $100 million+ in “mainly secondary investment” — from a mix of existing and new investors. Its blog post name-checks Omers Ventures, Thrive, GP Bullhound, Impact America Fund and WhatsApp founder Brian Acton; inventor of the world wide web Tim Berners-Lee; VC and diversity activist Freada Kapor Klein; and entrepreneur Mitch Kapor as being among the participating investors. So quite the line up. DuckDuckGo said the secondary investment allowed some of its early employees and investors to cash out a chunk of their equity while bolstering its financial position. Although it also says its business — which has been profitable since 2014 — is “thriving”, reporting that revenues are now running at more than $100 million a year. Hence it not needing to keep dipping into an external investor pot. Its last VC raise was in 2018 when it took in $10 million after being actively pursued by Omers Ventures — who convinced it to take the money to help support growth objectives (especially internationally). DDG has a few other metrics to throw around now: Over the last 12 months it said its apps were downloaded over 50 million times — more than in all prior years combined. It’s also revealed that its monthly search traffic increased 55% and says market share trackers indicate that it grabbed the No. 2 spot for search engine on mobile in a number of countries, including the U.S., Canada, Australia and the Netherlands (StatCounter/Wikipedia). “We don’t track our users so we can’t say for sure how many we have, but based on market share estimates, download numbers, and national surveys, we believe there are between 70-100 million DuckDuckGo users,” it added. A looming shift to Google’s Android choice screen in Europe, where regulators have forced the company to present users of mobile devices that run its OS with rival options when they’re setting a default search engine, looks likely to further boost DuckDuckGo’s regional fortunes. Google ditches pay-to-play Android search choice auction for free version after EU pressure Google will be ditching the current paid auction model — so rivals which have a valuable alternative proposition for users (like privacy) combined with strong brand awareness (and, well, everyone likes ducks…) have the best chance yet to take slices out of Google’s market share. DuckDuckGo’s blog post confirms it’ll be dialling up its marketing in Europe and other regions. “Our thriving business also gives us the resources to tell more people there is a simple solution for online privacy they can use right now. Over the last month, we’ve rolled out billboard, radio, and TV ads in 175 metro areas across the U.S., with additional efforts planned for Europe and other countries around the world,” it notes. So it look like a good chunk of DDG’s secondary funding will be spent on growth marketing — as it seeks to capitalize on rising public attention to online privacy, tracking and creepy ads, itself fuelled by years of data scandals. Awareness is also now being actively driven by Apple’s recent switch to inform iOS users of third-party app tracking and to give people a simple way to say no — which includes slick, Cupertino-funded ad campaigns (such as the one below) which are clearly intended to turn and engage mainstream heads… It’s fair to say it’s probably never been easier to craft a simple and compelling marketing message around privacy — and that’s also a testament to how far privacy tech has come in terms of usability and accessibility. So, yes, DuckDuckGo’s business sure looks like it’s sitting pretty at this juncture of the web’s evolution. And its blog post talks about “becoming a household name for simple privacy protection”. So the scale of its ambition is clear. “Privacy skeptics have dominated the discussion about online privacy for too long. “Sure people care about privacy, but they’ll never do anything about it.” It’s time to lay this bad take to rest,” it adds. More products are also on the slate from the 13-year veteran privacy player. It already bolted on tracker-blocking back in 2018 but is looking to go further — saying that it will be rolling out additional privacy features to what it bills as its “all-in-one privacy bundle”, including an email protection tool that will be launched in beta “in a few weeks” and which it says will “give users more privacy without having to get a new inbox”. “Later this summer, app tracker blocking will be available in beta for Android devices, allowing users to block app trackers and providing more transparency on what’s happening behind the scenes on their device. And Before the end of the year, we also plan to release a brand-new desktop version of our existing mobile app which people can use as a primary browser,” it goes on, adding: “By continuing to expand our simple and seamless privacy bundle, we continue to make our product vision, ‘Privacy, simplified.’ a reality.” That’s another trend we’re seeing in privacy tech: Innovators who have carefully and credibly built up a solid reputation around one type of tech tool (such as search or email) find themselves — as usage grows — perfectly positioned to branch out into offering a whole bundle or suite of apps they can wrap in the same protective promise. Another player, ProtonMail, for example, has morphed into Proton, a privacy-centric company which offers freemium tools for not just end-to-end encrypted email but also cloud storage, calendar and a VPN — all neatly positioned under its pro-privacy umbrella. Expect more development momentum as privacy tech continues to accelerate into the mainstream. ProtonMail gets a slick new look, as privacy tech eyes the mainstream DuckDuckGo gets $10M from Omers for global privacy push DuckDuckGo presses the case for true ‘one-click’ search competition on Android

Audi’s Latest Concept Is an Electric Car That Expands and Contracts

To continue, please click the box below to let us know you're not a robot.

The Deep Learning Tool We Wish We Had in Grad School

Last Updated on January 6, 2021 by Machine learning PhD students are in a unique position: they often need to run large-scale experiments to conduct state-of-the-art research but they don’t have the support of the platform teams that industrial ML engineers can rely on. As a result, PhD students waste countless hours writing boilerplate code, ad-hoc scripts, and hacking together infrastructure — rather than doing research. As former PhD students ourselves, we recount our hands-on experience with these challenges and explain how open-source tools like Determined would have made grad school a lot less painful. When we started graduate school as PhD students at Carnegie Mellon University (CMU), we thought the challenge laid in having novel ideas, testing hypotheses, and presenting research. Instead, the most difficult part was building out the tooling and infrastructure needed to run deep learning experiments. While industry labs like Google Brain and FAIR have teams of engineers to provide this kind of support, independent researchers and graduate students are left to manage on their own. This meant that during our PhDs, the majority of our attention was spent wrangling hundreds of models, dozens of experiments and hyperparameter searches, and a fleet of machines. Our ad-hoc workflows prevented us from doing better research, as tasks like starting new experiments and distributing training would cause increasingly more strain on the existing workflows and infrastructure. Every project started the same, with a lone graduate student tasked with implementing a research prototype and performing a virtually endless number of experiments to test its promise. There was little infrastructure, scarce resources, and no process. So we would start writing one-off scripts: scripts to create the prototype, scripts to kick off dozens of experiments, and even more scripts to interpret the logs from these experiments. These scripts were run on whatever machines we could find: the labs’ machines, friends’ lab’s machines, AWS spot instances, or even our professors’ personal machines. As a result, we’d have gigabytes of logs in various drummed up formats, model checkpoints, and PDFs of graphs showcasing our results, scattered about the file systems of the machines we used. We quickly learned that to survive as ML graduate students, becoming well-versed in engineering, system administration, and infrastructure management was table-stakes. The first time each of us realized that it didn’t have to be this way was when we did industry internships. As interns in a place like Google Brain, we had access to Google’s internal training infrastructure that allowed us to focus on research as opposed to operations. It was daunting to leave a place like Google, knowing that as independent researchers, we would be back to managing our own infrastructure and that this would come at the cost of doing our research. Fortunately, you don’t have to do grad school the way we did. Open-source tools for deep learning training have matured and can empower individual researchers to spend less time wrangling machines, managing files, and writing boilerplate code, and spend more of their time forming hypotheses, designing experiments, interpreting results, and sharing their findings with the community. But in the throes of conducting research and surviving grad school, it is difficult to invest time to learn a new tool without the guarantee that it will increase your productivity. To help future graduate students get over that hurdle, we share the ML research pain points that Determined AI would have alleviated for us. Throughout the life cycle of a deep learning research project, you’re bound to run into several common pain points. Today, many of these can be alleviated with foresight and the right tooling. In this section, we share the pain points we commonly encountered and how tooling like Determined can help. A single deep learning experiment can run for days or weeks and requires constant monitoring. In grad school, we would typically monitor experiments by tailing a log file or SSH’ing into the cluster and using tmux to monitor the job’s console output. This required remembering to start the experiment in a tmux session, to log key metrics, and to manage output log file naming and organization. When running several experiments at the same time, this also required tracking which experiment was running on which machine. Tools like Determined shed this overhead by automatically tracking and persisting key metrics and by logging them to TensorBoard. These results and more are available in a web UI for users to monitor their experiments in real-time and can be shared with peers, advisors, and community members with a single link. An experiment can also crash due to transient errors that are out of our control. This issue is exacerbated when running on preemptible cloud instances as a cost-saving measure. In these situations, we could easily lose hours or days worth of work and would then need to relaunch an experiment manually by passing a command via SSH. With Determined, the system automatically retries failed jobs for you, so no time is wasted when an error occurs. Automated experiment logging helps you diagnose and track where failures are happening across machines. Checkpoint saving also ensures that little progress is lost when a failure occurs. Determined manages checkpoints automatically: users can specify policies to control how often checkpoints are taken and which checkpoints should be preserved for future use. The result of days and hours of experimentation are artifacts like log files, model checkpoints, and results from subsequent analyses. It’s necessary to persist results in all stages of the project to retroactively report them to the community. Initially, managing this data is straightforward to do on the file system with careful naming and folder organization. But as a project progresses, it becomes an unwieldy way to track the gigs and gigs of emerging data. For us, it was common to have to redo a long-running and resource-intensive experiment because we lost track of a particular experiment graph or the script and model checkpoint to reproduce an earlier result. Instead, it’s better to start with an experiment tracking platform early into a project’s lifecycle, so that all experiment data is managed for you. Using Determined, model source code, library dependencies, hyperparameters, and configuration settings are automatically persisted to allow you to easily reproduce an earlier experiment. The built-in model registry can be used to track your trained models and identify model versions that are promising or significant. Deep learning training requires a huge amount of resources, with state-of-the-art results sometimes requiring tens of thousands of GPU hours. Almost inevitably, independent researchers need to scale their own training experiments to more than a single GPU. By relying on native PyTorch or TF distribution strategies, we were still left with tasks like setting up networking between machines and alleviating errors from machine failures and stragglers. At Determined, distributed training is rolled out for you by infrastructure experts. By writing your model code in Determined’s Trial API format, you can distribute your code with a single config file change. Determined takes care of things like provisioning machines, setting up networking, communicating between machines, efficient distributed data loading, and fault tolerance. Many deep learning practitioners rely on cloud platforms like AWS and GCP to run resource-heavy experiments. However, when operating within tight academic budgets, cloud platforms were often prohibitively expensive. Instead, we would run on cheaper spot instances without guaranteed uptime. Consequently, we had to manually restart stopped instances, checkpoint experiments constantly, or in absence of this, suffer lost results. To make the most use of available resources, Determined manages cloud compute resources automatically for the user depending on what jobs are queued. When using AWS spot instances or GCP preemptible instances to reduce cost, Determined maintains reproducibility with fault-tolerant checkpointing. Hyperparameter tuning is a necessary step to achieve state-of-the-art model performance. However, these were one of the most difficult experiments to run as they scale up all of the pain points previously discussed. Running a grid search is simple in theory, but ends up being orders of magnitude more costly and longer to run than traditional training. Algorithms that employ early-stopping like SHA and ASHA can be dramatically more efficient but are difficult to implement. (Well, not for Liam who invented these algorithms, but it’s difficult for the rest of us!) Hyperparameter searches also generate a lot more experimental metadata to manage and are harder to rerun when things go wrong. With Determined AI, you can run hyperparameter searches with state-of-the-art algorithms by changing a config file. And just like in regular training, you get experiment tracking, distributed training, and resource management out of the box. You can also pause, resume, or restart hyperparameter tuning jobs on-the-fly. Building upon empirical results, either by ourselves or the broader community, requires being able to reliably reproduce said results. Reproducibility is becoming a first-class goal of the ML research community, with initiatives like the Reproducibility Challenges or Artifact Evaluations. During our PhD, we found it difficult to anticipate all the data needed to ensure full reproducibility of our results. For instance, we may start by saving the experimentation script and the code (via git SHA) that led to a particular result, only to come to find that we could not reliably reproduce the result without knowing the machine the experiment ran on. With Determined AI, the data you need for reproducibility is automatically persisted for you, including the code used to create the model, the model weights, the full environment used to train the model, and the data preprocessing code. If you are like us, you may find yourself spending most of your time on operations, not research. Fortunately, with the emergence of ML infrastructure tools, you don’t have to do ML research the way we did. Tools like Determined provide researchers the foundation to build state-of-the-art and even production grade models. If you feel like you could benefit from the backing of a training platform, we encourage you to give Determined a spin. To get started, check out our quick start guide. If you have any questions along the way, hop on our community Slack or visit our GitHub repository — we’d love to help! Originally published at https://determined.ai on November 5, 2020. The Deep Learning Tool We Wish We Had In Grad School was originally published in Towards AI on Medium, where people are continuing the conversation by highlighting and responding to this story. Published via Towards AI

Katzenpost/catchat: Traffic analysis resistant messaging app

{{ message }} This repository has been archived by the owner on Jan 6, 2023. It is now read-only. katzenpost/catchat You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.

DeepMind: Generally capable agents emerge from open-ended play

Crossposted from the AI Alignment Forum. May contain more technical jargon than usual. This is a linkpost for https://deepmind.com/blog/article/generally-capable-agents-emerge-from-open-ended-play p p p p p p p

Skweak: Weak Supervision for NLP

{{ message }} NorskRegnesentral/skweak You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.

Crime Boss or Tech CEO? Encrypted Phone Co Sues the Government to Save Itself

Hacking. Disinformation. Surveillance. CYBER is Motherboard's podcast and reporting on the dark underbelly of the internet. See More → In an illustration on his company’s website, Canadian entrepreneur and CEO Jean-François Eap is smiling with his head slightly tilted to the side alongside the staff of his technology company. Next to Eap are images of the Chief Operating Officer, the Chief Technology Officer, the company’s Controller, as well as the Vice President of Revenue Operations. They all work for Eap at Sky Global, a firm that develops privacy-focused mobile phones with custom software for sending encrypted messages. Advertisement But Eap is not an ordinary tech CEO. In March, the U.S. Department of Justice indicted him for allegedly helping distribute at least 5 kilograms of cocaine by providing his customized phones to criminals. The indictment also charged Eap and an alleged co-conspirator under the Racketeer Influenced and Corrupt Organizations Act (RICO), a law originally designed to prosecute mob bosses. As part of the operation against Sky, U.S. authorities also seized over 100 Sky web domains. In parallel, European law enforcement officials said they managed to intercept and decrypt some half a billion messages sent by Sky devices. Ordinarily this would require somehow bypassing Sky’s end-to-end encryption, suggesting a sophisticated law enforcement operation. Sky told Motherboard at the time that this was done via a rogue version of Sky’s app. In the encrypted phone world, these charges and operations are not that unusual. Law enforcement agencies have targeted Phantom Secure, Encrochat, and other companies. In the wake of some of these law enforcement actions, the owners and employees of the companies were arrested or have become fugitives. Eap, however, has not disappeared. In fact, his lawyers say he didn't do the crimes alleged in the indictment, that Sky was and is a legitimate, above-board company, and they are pushing back and trying to regain control of the seized web domains. On Thursday, Eap’s counsel submitted a detailed filing in the Southern District of California laying out why they think the government took over the domains illegally, and saying they have offered to cooperate with the Department of Justice. The filing, reviewed by Motherboard, also included detailed internal Sky documents showing how parts of the company operated. Advertisement “The government should be ordered to return the seized internet domains because the government’s seizure and retention of the domains is improper and contrary to law,” the filing, signed by Ashwin J. Ram and Steven R. Welk, attorneys for Sky from law firm Steptoe & Johnson LLP, reads. It adds that the indictment is “falsely alleging that Sky ECC was created by Mr. Eap and used by Mr. Herdman [his alleged co-conspirator] to facilitate drug trafficking and other illegal activity.” Without its domains, Sky has essentially been rendered a dead company. The filing notes that the firm had to let go of 27 staff and 14 contractors after the law enforcement action. But if a court agrees with Sky, perhaps the company could try to resurrect itself. Beyond reviewing the documents included with the court filing, Motherboard has spoken to multiple people who sold phones on behalf of Sky, a source familiar with the investigation into the company, and obtained another cache of documents that Sky provides to its resellers. The documents and interviews show that Sky did try to enforce policies to keep crooks off its platform. But those efforts cannot erase the past—at least at one point, Sky was a preferred phone of choice for some serious organized criminals. “Of course they know. Of course they know,” one person who sold Sky phones to criminals told Motherboard when asked if they thought Sky itself knew it was providing devices to criminals. Motherboard granted the source anonymity to speak more candidly about potentially criminal behaviour. Advertisement * Eap started Sky in 2010 in Vancouver, Canada. The main product was Sky ECC, a phone that came preloaded with an app for sending encrypted messages to other Sky users. The phones also included an unlimited data roaming package, according to the filing. Sky offered the ability to remotely wipe a device; a customer could contact their reseller who would either wipe the device themselves or send a request to Sky itself to do so. The company worked on a distributor model, where Sky entered agreements with distributors who would then hire their own agents who would sell the devices to end-users. The filing collectively describes these sellers as “partners” and says Sky did this because, being a startup, it did not have its own sales or marketing channels at first. By March of this year, the company had around 120,000 active users, according to the filing. Many of Sky’s customers, however, were criminals, the Department of Justice alleges, and former Sky sellers told Motherboard. In Australia, Sky sellers competed with other encrypted phone companies such as Ciphr and Phantom Secure. Criminal users of these sorts of companies have used their respective remote wipe features to, sometimes successfully, remove incriminating evidence from phones after the police seized the devices. “Ciphr and Sky started moving into the Australian market with superior products,” an encrypted phone industry source told Motherboard. Motherboard provided them anonymity to protect them from retaliation. Advertisement Did you previously use phones sold by Sky, Anom, Ciphr, Phantom, or any other similar company? Did you used to work for any of those companies? We'd love to hear from you. Using a non-work phone or computer, you can contact Joseph Cox securely on Signal on +44 20 8133 5190, Wickr on josephcox, or email joseph.cox@vice.com. A Department of Justice press release announcing the indictment against Eap alleged that Sky’s purpose was to create, maintain, and control a secure communication method to facilitate the trade of heroin, cocaine, and methamphetamine across Australia, Asia, Europe, and North America, including in the United States and Canada. The announcement alleged that Sky has facilitated the criminal activity of transnational criminal organizations for more than a decade. “The indictment alleges that Sky Global generated hundreds of millions of dollars providing a service that allowed criminal networks around the world to hide their international drug trafficking activity from law enforcement,” Acting U.S. Attorney Randy Grossman said in the announcement. In the wake of the mass interception of Sky messages by European authorities, police arrested over 80 people and confiscated phones and weapons. Advertisement In March 2018, the U.S. charged Vincent Ramos, the CEO of Phantom Secure, with similar charges to what Eap now faces. Ramos later pled guilty and received a nine-year sentence. That prosecution appears to have been something of a wake-up call to the encrypted phone industry—U.S. prosecutors and the FBI now potentially saw them as targets, not neutral technology companies like Apple or Google. Soon after, other companies started to change their positioning and ramp up their enforcement efforts. A screenshot of a Sky employee saying they won't wipe devices that are being investigated. Image: Motherboard “Sky now enforces contracts for distributors and sub-distributors,” the industry source told Motherboard in the wake of the Phantom shutdown. “No one ever speaks of criminals using the products.” The idea was to create a buffer between the company and the criminal users, the source suggested. “Sky is 2 layers removed from the sale at minimum,” they added. The exhibits included with Sky’s filing show in new detail the effort Sky took to remove criminals from its platform. “You may not knowingly sell or otherwise provide the Products and Services to any Customer for illicit, illegal or criminal use,” a Sky terms of use document reads. “This ECC ID has been flagged for breaching our terms of service. It will be deactivated immediately,” one email sent by a Sky support worker to a user reads. The message from Sky points to a section of the company’s terms of use that does mention non-permitted uses such as promoting criminal activity. Advertisement Another email sent by Sky’s chief operating officer to what appears to be a Sky distributor says that one of the distributor’s agents has violated the company’s terms of service because of their willingness to sell the Sky ECC product to someone wanting to use it for illicit activity, as well as other violations. One document shows a submitted application by someone called Agramovic Srdjan to become a Sky reseller. When asked in the application form who they intended to sell Sky phones to, Srdjan wrote “political party, and criminals,” according to the document. Sky’s Head of Channel Sales then responded that selling to criminals is a breach of the Sky terms and conditions and denied the application. Srdjan confirmed to Motherboard in an email that he did apply to become a Sky reseller, but said his “application was a joke.” It was “just to test what is happening. Everybody [was] talking about it, so I was just curious.” Another email shows a Sky seller reporting another applicant for their intention to sell Sky phones to criminals. A screenshot of the denied application from Srdjan. Additional redactions by Motherboard. Image: Motherboard. One of the more detailed documents is a spreadsheet listing various websites, classified ads, and Facebook, Twitter, and Instagram accounts which were advertising or discussing Sky phones and that Sky itself has tried to police. Where applicable the spreadsheet includes notes such as the social media post having an “illegal undertone,” and that Sky has contacted the account owner. “Referencing authorities. Have sent a request to have the ad corrected/removed,” another reads. This and another document mention “secret shopping,” suggesting that Sky engaged with the alleged sellers to try and ascertain if they were official resellers or not. Advertisement The most crucial documents may be those that discuss Sky’s remote wipe feature. An email chain between an apparent Sky seller and Sky itself discusses how to phrase marketing material around the feature. Originally, a website was going to read “We can remotely delete all data if the Unit is lost or detained by Third Parties.” A Sky employee suggested the website remove the phrase about the phone being “detained.” Other documents show Sky employees declining to wipe devices when asked to do so by resellers. “Hello. Please delete this ECC ID, the police have it,” one message sent to Sky reads. “PLEASE HELP!!! Two customers have problems with the police. Their devices were confiscated,” another adds. In both cases, Sky declined to wipe the phones, according to the emails. In practice, some resellers found a workaround to Sky’s policies. “What is a solution? Don’t tell me. Just ask me to wipe the phone; I’m not going to ask the reason. I’m going to do what you ask, and that’s it.” one former seller of Sky phones to criminal users told Motherboard. And even if they did know if the wipe was to remove content from a phone in police custody: “To be honest, I would probably wipe it.” The Department of Justice announcement alleges that Sky instituted an “ask nothing/do nothing” approach after the takedown of Phantom so it could claim plausible deniability. Ram, one of the attorneys who signed the filing, told Motherboard in an emailed statement that Sky had no control over its third-party sellers, “beyond contractual obligations prohibiting the unlawful use and marketing of its technology.” Advertisement “Sky and Phantom Secure were fundamentally different companies. While Phantom operated in the dark and wiped devices held by law enforcement, Sky had a public website and platform, and created policies to prevent that from happening and had an actual track record of refusing to wipe devices that were known to be under investigation. More fundamentally, Sky took steps to root out bad actors, and when Sky became aware of illegal activity, it terminated the source of that activity. Phantom did the opposite,” Ram added. A screenshot of a Sky employee suggesting changes to particular language. Image: Motherboard But all of the documents which Sky provided and Motherboard reviewed which show enforcement of Sky’s policies were created in 2019 or later, after the U.S. prosecution of Phantom Secure’s Ramos. In a statement, Ram said “There were compliance measures that explicitly prohibited criminal activity long before the Phantom indictment in 2018.  As with any legitimate company, these compliance measures were continually improved over time, including as a part of the well-documented ‘Sky 2.0’ launch, which pre-dated the Phantom indictment announced in March 2018.” A source familiar with the investigation into Sky said they believe that the company is similar to Phantom Secure. That is, they believe the company is aware it was providing phones to criminal end-users, and Eap specifically knew that the customer base he had was mostly, if not almost exclusively, criminal, they said. Motherboard granted the source anonymity as they weren’t authorized to speak publicly about the case. The Department of Justice announcement says that Sky “knowingly and intentionally participated in a criminal enterprise.” Eap declined to comment through his attorney on whether he knew Sky’s users included criminals. Advertisement In its filing, Sky maintains that it advertised its phones to individuals with heightened data privacy concerns, “such as doctors, lawyers, government contractors, celebrities, and even law enforcement agencies.” One email chain suggests the company did have interest from at least one agency: the Ontario Provincial Police (OPP) in Canada. In an email thread starting in 2018, Timothy Brown, an official at the OPP, asked to purchase at least one Sky device. “We are impressed with the service provided by the SkyECC phones. As previously noted the level of security is a bit too restrictive for our usual business purposes. However, we may still have a need for your product,” Brown wrote. It is not clear what exactly the OPP wanted the devices for. The OPP did not respond to a request for comment. * Despite the high profile nature of the indictment, in the filing Eap’s counsel claims that the U.S. government’s efforts to pursue the prosecution “appear to be minimal.” “The government has not, to my knowledge, initiated extradition proceedings against Mr. Eap, a Canadian citizen and resident,” the filing adds. The FBI did not respond to a request for comment. Ian McLeod, public affairs and issues management at the Canadian Department of Justice declined to comment on whether extradition proceedings have started, saying that “extradition requests are confidential state-to-state communications.” Advertisement Sgt. Kris Clark, community services and media coordinator at the Royal Canadian Mounted Police (RCMP) told Motherboard in an email that “The RCMP continues to work closely with international policing partners in order to collect and preserve evidence and investigate transnational crime. To date no charges have been laid in Canada. Given the investigation is ongoing, no additional information is available for release.” A screenshot of the Ontario Provincial Police trying to purchase Sky phones. Redaction by Motherboard. Image: Motherboard Eap’s lawyers also say they contacted the U.S. government in July to offer to cooperate with the investigation, and express their concerns over the seizures of Sky’s web domains. The filing says that the government declined to engage in substantive discussions. Kelly Thornton, director of media relations at the U.S. Attorney’s Office in the Southern District of California, declined to comment. Eap has responded in an unusually public manner to the charges against him. Shortly after the indictment against him, Eap told Motherboard that he would clear his name. Then several months ago, an employee at high profile public relations firm Berlin Rosen contacted Motherboard to arrange an interview with Eap. For months, the public relations firm planned an on-the-record interview with Eap. At the last moment, Eap’s counsel asked instead for the interview to be conducted on background. That is, attributing the information anonymously to someone familiar with the case. Motherboard declined: it is not typical to allow company CEOs to speak on background, because they are the ones ultimately accountable for a company’s actions. With the sections of the filing that specifically say the government should return Sky’s seized domains, “Since its inception, Sky Global has taken all reasonable and necessary steps to ensure that its domains were used for their intended, legitimate purposes. Nevertheless, the government seized Sky Global’s property, directly resulting in the sudden, involuntary suspension of an ongoing legitimate business with global operations and customers.” Ram added in the emailed statement that “The motion lays out exactly how law enforcement significantly overstepped its authority and violated the free speech and due process rights of Sky Global. Anyone concerned about privacy should be deeply troubled by how the government almost shut down a legitimate, law-abiding company that was attempting to address critical issues around data protection and privacy.” Update: This piece has been updated to clarify that Eap's counsel asked for the interview to be conducted on background, and not Eap himself. The piece has also been updated to include more information from Ram. Subscribe to our cybersecurity podcast, CYBER. Subscribe to our new Twitch channel.

OwnCloud now uses Go for rewriting its eponymous file platform

Ready, steady, Go PHP has long been the tool of choice for ownCloud. But the scripting language no longer meets ownCloud’s requirements for developing a modern content collaboration application. Time for a change and a switch to the programming language Go. Yes, PHP has become increasingly powerful over time thanks to numerous adaptations, such as object orientation, standardized access to databases, an integrated web server and improved memory management. PHP interpreters and compilers have further increased the speed of script execution. But all these improvements cannot hide the fact that in the years since ownCloud was first developed, the requirements regarding performance and scalability have shifted. Meanwhile, Go emerged as a new language that offers significant advantages for modern application development in general and to suit the new ownCloud vision in particular. ownCloud now uses Go for rewriting its eponymous file platform. Based on our experience with Go in developing ownCloud Infinite Scale, we can confidently share a list of its most important advantages: Cross-platform Go runs on all operating systems natively. Go’s tool set provides excellent utilities for cross-compiling, enabling the fast and immediate deployment of the application for many different platforms. Speed Go does not require a script interpreter or a virtual runtime environment. This makes Go as fast as the programming language C, which is normally used for particularly high-performance requirements. Concurrency Go offers the possibility to perform several computations in parallel. “This makes it easier for developers to parallelize resource-intensive operations and thus develop a high-performance application,” adds Felix Böhm, CTO at ownCloud. Syntax Go is very easy to learn and has many parallels with the widely used programming language C. Developers can easily get up to speed with Go or switch from their previous tool stack. “It is a lot of fun to program in Go,” emphasizes Alex Unger, Senior Software Developer at ownCloud. Ecosystem The language was developed by Google teams and first presented in 2009. It is an open-source project and benefits from community involvement. Many tools for cloud-native projects are written in Go, such as the container solutions Docker and Kubernetes. “Go makes ownCloud appealing to a worldwide developer community,” Felix Böhm concludes. Go has many other benefits, such as automatic memory management and a direct and flexible implementation of object orientation. Go applications are more performant and use significantly less server resources, thus reducing both hardware costs and energy consumption. The new ownCloud generation, ownCloud Infinite Scale, will go one step further and stores the metadata of the files directly with the data on the storage and therefore no longer requires a database. Thus, ownCloud Infinite Scale is optimized for scale – as in the number of files, users and instances, but also as in file size. “The developers at ownCloud are enthusiastic about the switch and are thrilled about the ongoing collaboration with the very professional Go community,” summarizes Felix Böhm. p p Ready to see what’s next? * Mandatory field By submitting this form I agree that I want to to receive notifications and services via email, phone or personalized ads. Therefore, I agree, that ownCloud stores and uses my contact data for further information and in order to optimize and adapt the offer to my individual interests. I can revoke my consent for the future at any time, either directly via the link in emails or by email to . For further information please also see the Privacy Statement. We care about protecting your data. Here’s our Privacy Policy. Read now: 5 years of GDPR: We can do without holey shields The European Union’s General Data Protection Regulation (GDPR) might be exhausting, but it is a success story. The European Commission should use the fifth anniversary of the GDPR to reconsider “Privacy Shield 2.0”, opines Holger Dyroff, Co-Founder and COO of ownCloud. read more Sprint Review: Hybrid User Attributes In recent weeks, ownCloud developers presented a bunch of helpful new features in user/group and identity management. It is now possible to override ID settings by the local ownCloud server, and a logout will end all of the user’s sessions in the backend. Apart from that, there’s a new side panel for user settings, and much more. read more Powerful File Management Platforms for Digital Collaboration If you are developing a new software application and need functionality for managing unstructured data, you don’t have to reinvent the wheel. ownCloud explains why it is worth integrating an existing file management platform instead. read more

Game Stack Live – two days of all things game dev from Microsoft (April 20-21)

Game Stack Live, Now On-Demand Watch or re-watch all the talks from Game Stack Live Stay Connected Join the conversation on Discord to stay current on news, get help from the community, and engage industry experts from Microsoft. Announcing ID@Azure! Empowering game developers to take full advantage of the cloud Graphics Achieve amazing visuals and performance on next-gen Windows and Xbox graphics hardware with the latest DirectX Tools and features. New consoles and graphics cards usher in a new wave of more immersive video games. In this session, we will cover the latest innovations to our graphics platform, making it easier for game developers to push the limits of graphics on both PC and Xbox with DirectX, HLSL, PIX, HDR, and more. The DirectX team has been working on a way to change the game for developers who want to be on the bleeding edge of graphics. We're finally ready to show the world how we're pushing the envelope with an exciting new update. Introducing HLSL Shader Model 6.6: granting shader developers increased flexibility to enhance and expand existing rendering approaches and devise all new ones! ​ High Dynamic Range (HDR) is part of the gold standard for the newest generation of graphically rich games. In this talk, you'll learn about our ongoing work to improve DirectX's HDR support on PCs. We are helping to solve the problem of HDR display ecosystem variability, providing best practices for optimizing your native HDR implementations for displays, and working on exciting new technology to extend the reach of HDR PC gaming.​ This talk will discuss the usage of Variable Rate Shading in Gears 5 and Gears Tactics. Our implementation targets zero perceptual drop in visual quality, works seamlessly with dynamic resolution, and runs across Xbox Series X|S and PC. Performance improvements will be discussed as well as how VRS was tuned to work across different rendering passes. We will also discuss some challenges and integration tips that can be applied to any game engine. With raytraced visuals bumping rendering quality even higher than ever before, a significant amount of fine tuning is required to maintain real-time performance. A typical way to achieve this is to trace fewer rays, and to make sense of the noisier output that method delivers. This presentation will explain how the AMD FidelityFX Denoiser allows for high-quality raytracing results without increasing rays per pixel, and deep dives into specific RDNA2-based optimizations that benefit both Xbox Series X|S and PC. This talk is aimed at graphics engineers that have little or no experience with ray tracing. It serves as a gentle introduction to many topics, including "What is ray tracing?", "How many rays do you need to make an image?", "The importance of [importance] sampling. (And more importantly, what is importance sampling?)", "Denoising", "The problem with small bright things". Along the way, you will learn about specific implementation details from Minecraft. How the addition of a path traced rendering mode in Minecraft inspired our creative team. We will do a brief review of "original" Minecraft, and it's pixel art roots. We reflect on the game's unique look, and how players and creators can tell new and different stories in path traced Minecraft. Then we will discuss how this new rendering option is made available to our creators community via our texture pack system and then show the results, and speak to the aesthetic benefits and challenges of the new tech. RTXDXI offers realistic lighting of dynamic scenes that require computing shadows from millions of area lights. Until now, this has not been possible in video games. Traditionally, game developers have baked most lighting and supported a small number of "hero" lights that are computed at runtime. This talk gives an overview of RTXDI and offers a deep dive into previously undisclosed details that enable high performance. System & Tools Learn how to take advantage of features on Windows and Xbox Series Consoles for improving development, performance, load times, engine integration and more. The new Xbox Velocity Architecture in the Xbox Series X|S consoles enables developers to re-imagine how to build their games. By combining a super-fast SSD, hardware decompression, and the new DirectStorage functionality, game developers have tools to develop immersive experiences with a minimal load time for users. This session will review the benefits of the Velocity Architecture and show a real-world example. Microsoft is excited to bring DirectStorage, an API in the DirectX family originally designed for the Velocity Architecture to Windows PCs! DirectStorage will bring best-in-class IO tech to both PC and console just as DirectX 12 Ultimate does with rendering tech. With a DirectStorage capable PC and a DirectStorage enabled game, you can look forward to vastly reduced load times and virtual worlds that are more expansive and detailed than ever. In this session, we will be discussing the details of this technology will help you build your next-generation PC games. PIX is the essential performance tuning and debugging tool for DirectX 12 applications on Xbox and Windows. As the game industry pushes the boundaries of what is possible on the latest generation hardware, we are building the tools you need to get detailed information about how those games are running. In this talk you will learn how AAA studios use PIX in their development process to make their games better. You'll hear about the latest features in both Timing Captures and GPU Captures, as well as what's coming up in 2021 and beyond.​ The Unity Physics DOTS package, developed in partnership with Havok, needed to be cache free in order to better support scenarios such as simulation rollback for networked simulations. Stacking rigid bodies tends to require a cache, spanning multiple frames, to produce stable results. Here we will talk through some heuristics needed to resolve this paradox and to achieve a more stable physics simulation without sacrificing performance or adding additional state to the simulation.​ Join AMD on an adventure thru "Zen 2" and "Zen 3" processors which power today's game consoles and PCs. Dive into instruction sets, cache hierarchies, resource sharing, and simultaneous multi-threading. Journey across the sands of silicon to master microarchitecture and uncover best practices! In the pursuit of enabling developers to create the most amazing gaming experiences yet, the Xbox hardware team developed many innovations to deliver the most powerful Xbox ever, the Xbox Series X. We will discuss the silicon and memory architectures that enable a generational leap in GPU, CPU and storage performance. The team will also dive into the features that enabled the console to be cool, quiet and power efficient, such as the innovative tower form factor, split motherboard design, optimized power delivery and the thermal architecture. Cloud services are critical components of making games today and we expect the reliance on the cloud by game creators to continue to grow in the future. The Game Development Experiences team at Microsoft is investing deeply in our Azure infrastructure to ensure that it meets the demands of game developers around the world. In this talk we'll discuss some of the challenges of working in a decentralized and diverse world and share present and future ideas for how Microsoft can help accelerate the sharing of ideas, code, content, and creativity for game creators of all sizes. Production & Publishing Ship your games on Xbox faster using tips and tricks from our experts on release management, publishing, certification, QA, and production. The future of gaming is in the cloud. Xbox Game Streaming enables gamers to play the content they want, on the device they want, when they want. But why as a developer would you want to make the investment in optimizing your game for cloud streaming? In this session we will highlight what cloud optimizations are possible, how these optimizations are influencing consumer behavior in the current Xbox Game Streaming offerings, and where we see the impact to consumer behavior in the future. Microsoft Partner Center gives you the ability to publish, or "flight" your titles in the RETAIL environment in a private, secure manner to a controlled audience.  This is a great way to validate your builds for a small group.  The Xbox Insider Program supercharges this feature and lets you scale your flight to a much larger audience either publicly or privately, with built-in tools to get actionable feedback. This presentation walks you through this flighting process step by step to illustrate just how easy it is. Game telemetry is a powerful, but underutilized tool in the test/quality space. We aim to flip the status quo of games testing by empowering everyone from developers to active users to contribute to the passive verification of game features and behaviors. We will highlight a method employing the scalability of the Azure ecosystem and flexibility of Python to turn our game telemetry into actionable insights and bugs. Let everyone who touches the game, test the game. In this presentation, you will learn about building cloud native CI/CD pipelines for Unity mobile game apps with managed services like Azure Pipelines and Visual Studio App Center, as well as setting up cloud ready licensing infrastructure with Unity Build Server. Also, you will discover how to achieve build performance improvement and cost savings at the same time by deploying Azure VMSS agent pools and preparing VM images with Azure Image Builder. Disclaimer: This talk is about our own experiences and workflows. The workflows and techniques described within is not created by, supported or endorsed by Unity. In this talk, we will share the evolution of our bot testing tools from complex, fragile code to a much more powerful and flexible toolset. Our talk focuses on the advantages (and challenges) of utilizing game state information and re-usable actions to contextually control the bots. We'll demonstrate some of the system's newest capabilities to control bots in coordinated tests and how these have already been successfully utilized on titles such as Forza, Gears of War, and Minecraft Dungeons. Accessibility & Inclusion Reach more gamers and develop for inclusivity with accessibility tools and partnerships shared through the Gaming for Everyone initiative. Each studio is a little different and each dev team has their own needs. That's why having champions for accessibility embedded in the studios is a key to success. In this talk, representatives from Undead Labs, Turn 10 and the Coalition describe their journey to success in creating an inclusive atmosphere to impact their product. Xbox made history when it became the first platform to publish its own set of accessibility guidelines for developers to reference. After their release, we received feedback on how they can be even more powerful. We've embarked on a massive refresh, with new background information, clarifications and example contact to drastically enhance their usefulness. In this sessions, viewers will learn about all the new updates that have been made and why we made them. This talk is a retrospective of the inclusive production and community practices used during the creation and release of DONTNOD's latest game, Tell Me Why. Our panel of game developers from across Xbox Game Studios talk about their specific experiences being hired, on-boarding and working in remote scenarios for Xbox Game Studios teams. Attendees will hear details on how game professionals are both joining, working and being successful as part of top-tier development teams at Xbox during a global health crisis. Audio accessibility can be a daunting task for developer to consider but it doesn't have to be. Microsoft offers many technologies, platform features and techniques for bringing audio accessibility for both Gamers without Sight and Deaf and Hard of hearing gamers. We will explore WHY audio accessibility is so important, WHAT platform features, tools and technologies Microsoft offers, and HOW developers can use those to more easily bring great audio experiences which not only help make games more accessible but give everyone better, more immersive gaming experiences. Audio Develop more immersive, dynamic, and optimized game audio using the Microsoft Spatial Sound platform, Project Acoustics, and Xbox Series console hardware. This presentation is an audio content focused look at a variety of ways that Spatial Sound can be used to enhance the audio experience in a game. We will explore using Spatial to enhance Immersion, expand Story Telling, bring better Gameplay, push Spectacle, and give Clarity and mix space to the audio environment. This video will have pre-rendered HRTF playback and will be best listened to over headphones. All content will be played in A/B fashion to show difference between Stereo and Spatial rendering. This presentation will focus on integrating the Microsoft Spatial Sound platform into your title. It is intended to be a primer on getting started with using Spatial Sound on our platforms. We will show how to integrate using middleware, and discuss best practices when integrating spatial sound into a game. We will cover the different types of usage scenarios possible for playback within a game, and how the Spatial platform's user choices might affect those choices. This presentation will explore how we can take immersive audio to the next level in our games. By taking advantage of platform features, technologies and next gen hardware offerings, developers have the ability to bring deeply immersive audio experiences to their games, without sacrificing resources to accomplish it. We will look at Spatial Audio with Dolby Atmos and DTS:X, Advanced acoustic technologies such as Project acoustics, the possibilities of ray traced reflection techniques, and tie it all together with the promises of Series X/S audio HW processors. Multiplayer Build next-gen cross-network multiplayer games with server hosting, player communication, matchmaking, and more powered by Azure and PlayFab. This session is an introductory level overview of Azure PlayFab Multiplayer Servers, with guided steps to help developers successfully build & deploy a sample game server on PlayFab. We'll dig into compute options optimized for gaming and share best practices to quickly & easily integrate existing game server code with the game server developer kit (GSDK).​​ This intermediate level session will help developers learn advanced scaling techniques that adjust game sever hosting capacity to meet dynamic shifts in player demand. This is a great session for learning how to keep server hosting costs low while maintaining capacity to grow dynamically with your player base.​ A Conceptual overview of Party Concepts and Features followed by examples of scenarios studios are using Party for. Game Streaming presents opportunities to bring your game to new players and devices. This talk will cover ways to make your game stand out by mitigating streaming latency. Along the way, we'll cover time machines, coyotes, and saving the world. We're going to show you how we took the console and PC version of Minecraft Dungeons and updated it to play great on mobile in just seven weeks.  We will walk through what we learned in this process, and show examples of how you might be able to apply what we learned to your own game.​ Migrating technology stacks can be scary. But it doesn't have to be. In this talk, we'll show you how Mojang ported Minecraft Realms from a custom infrastructure-as-a-service implementation to a managed solution running on Azure, while limiting impact to players​. To create the most realistic next-gen Flight Simulator we had to design a virtual world at Earth-Scale. It means to create the most realistic Earth digital Twin ever seen from the air, with its geography, its sky, air model and weather, and its real-time activities. But also, to distribute this data to players all around the globe and gather them in a one unified world connected by our game services. This presentation is about how these challenges have been achieved using Azure scaleable solutions for computation, distribution, and services as a backbone. Being able to pull actionable insights from your game analytics can help teams grind out a win, turnaround a difficult launch, or accelerate growth. In this talk, we'll cover how Azure Data Services provide powerful data analytics for game developers all around the world, helping studios acquire better insights and make smarter decisions. We'll also dive into best practices for migrating your data analytics platform from an on-premise solution to the cloud.​​ Game developers today want to be able to start small and then grow to meet the load when their game goes viral. Azure Cosmos DB is Microsoft's fully-managed NoSQL database service in Azure that provides extreme low latency, unmatched availability and unlimited scale. You'll also learn how Azure Cosmos DB can allow you to build near real-time game analytics not possible with any other database in Azure. You'll see what other game studios have to say about using Cosmos DB. Community Connections *The Americas Only Don't miss these exciting events hosted by the Xbox Employee Communities in the Americas time zone. They include a mix of content including Panels, Group Mentoring, and Community Connection social hours for you to participate in, learn more, and make new connections! Asian Game Developers Panel Discussion Hosted by Xbox Community Leaders​ Join us during Game Stack Live for a panel discussion and an opportunity to hear from many from the game industry including Team Xbox.  The two major topics will be: what it is like being Asian in our respective companies, and what culture work we each do to help support the work and careers of our Asian teammates. Black Game Developers Panel Discussion Hosted by Xbox Community Leaders​ Join us during Game Stack Live for an opportunity to hear from other Black developers in the gaming industry. We will talk about the wide range of Black experiences in the industry and how we can work to support the work and careers of our colleagues. We look forward to seeing you there! Gaming & Disability in Gaming Happy Hour at Game Stack Live​ Join us during Game Stack Live for an opportunity to meet Gaming & Disability industry professionals for connection and community. After a welcome from members of Gaming & Disability, join the chat about what you're playing, find other community members near you, and network about job or mentoring opportunities using the event profile system. We look forward to the chance to meet you! This event is hosted and moderated by members of Gaming & Disability. Team Xbox Latinx Career Panel with Leaders in Gaming You're invited to join the Xbox Latinx community to listen to game industry professionals reflect on their career paths and retrace their journeys through the industry in their respective fields. This hour-long talk aims to shed light on some of the various career paths and disciplines that bring the gaming industry to life followed by Q&A at the end. We hope to see you there! Team Xbox Latinx Group Mentorships with Gaming Professionals Join us for an opportunity to participate in some group mentorships with various gaming industry professionals. After a welcome from members of Team Xbox Latinx, you can join a video call on the topic of your choice and these experts will answer questions they frequently come across in their mentoring as well as questions from those of you who attend. We look forward to meeting you and bringing these protips your way! LGBTQIA in Gaming Happy Hour Join us live to connect and build community. This session kicks off with a welcome from the Team Xbox LGBTQIA leadership, then everyone can chat about your gig, connect with other LGBTQIA people in the industry, or just talk about what you're playing. Expand your network using the event profile system to find people to learn from or partner with! This event is hosted and moderated by members of Team Xbox LGBTQIA who are looking forward to the chance to meet you!​ Women in Gaming Happy Hour Join us during Game Stack Live for an opportunity to meet Women in Gaming industry professionals for connection and community. After a welcome from members of Women in Gaming, join the chat about what you're playing, find other community members near you, and network about job or mentoring opportunities using the event profile system. We look forward to the chance to meet you! This event is hosted and moderated by members of Women in Gaming. DirectX 12 Ultimate: Ask the Experts Panel We've already seen the first wave of titles that have unlocked next-gen graphics with DirectX 12 Ultimate on the Xbox Series X|S and a range of PC hardware. But who are the engineers who built DirectX 12 Ultimate? And can you ask them questions about how to bring the state-of-the-art to your game? YES! Come join video chat to ask your questions or to meet the developers who are building DirectX on PC and Xbox. Game Stack Live is open to game developers around the world. The 24 hour event kicks off in The Americas at 8:00am PDT (UTC-7) on April 20, moves to Asia Pacific at 8:00am JST (UTC+9) on April 21 and finishes up in Europe, Middle-East, Africa at 9:00am CEST (UTC+2) on April 21. Microsoft strives to create an inclusive and accessible environment for all attendees. At Game Stack Live you will find sessions offered with ASL, closed captions in 15+ languages and additional session resources in Japanese and Korean. If you have any questions about the accommodations to be provided during this event, please email .

Netflix suspends three employees who crashed an execute meeting

Netflix Suspends Three Employees, Including Trans Person Who Spoke Out Against Dave Chappelle, for Crashing Leadership Meeting has suspended three employees for crashing a meeting of its top executives, including an out trans person who criticized a new comedy special from , sources tell Variety. Terra Field, a senior software engineer based in San Francisco, was among those suspended late last week for attending the “QBR” — Netflix’s quarterly business review, a two-day affair that convenes the top 500 employees at the company. Field, who identifies as queer and trans, and the other employees were not invited to the virtual gathering, according to insiders. Netflix did not suspend Field over recent tweets decrying what she called anti-trans jokes in the Chappelle special, individuals familiar with Netflix said, but for attending uninvited. An investigation has been launched into the three employees. Field did not immediately respond to requests for comment. “It is absolutely untrue to say that we have suspended any employees for tweeting about this show. Our employees are encouraged to disagree openly and we support their right to do so,” a Netflix spokesperson told Variety. Last Wednesday, Field wrote a lengthy Twitter thread about “,” the new Chappelle stand-up special in which the comic accuses the trans community of having “thin skin.” In her tweets, Field writes, “ Field went on to say of Chappelle, “our existence is ‘funny’ to him – and when we object to his harm, we’re ‘offended.'” She then listed numerous names of trans people, specifically highlighting trans women of color, killed in hate crimes. At the meeting, Netflix co-CEO Ted Sarandos fielded questions about how leadership should handle employees and talent upset over Chappelle’s remarks. He addressed the entire group in a memo after the event, which Variety obtained on Monday. “Chapelle is one of the most popular stand-up comedians today, and we have a long standing deal with him. His last special ‘Sticks & Stones,’ also controversial, is our most watched, stickiest and most award-winning stand-up special to date,” Sarandos wrote. “As with our other talent, we work hard to support their creative freedom — even though this means there will always be content on Netflix some people believe is harmful,” he said. Netflix talent speaking out against the Chappelle special includes Jaclyn Moore, the showrunner of its original series “Dear White People.” Moore told Variety last week that she “never loved Dave’s trans material before but this time it felt different. This is the first time I felt like, ‘Oh, people are laughing at this joke and they’re agreeing that it’s absurd to call me a woman.'” Read Field’s full Twitter thread: I work at @netflix. Yesterday we launched another Chappelle special where he attacks the trans community, and the very validity of transness – all while trying to pit us against other marginalized groups. You're going to hear a lot of talk about "offense". We are not offended 🧵 — 🎃 Terra Fied 👻 (@RainofTerra) October 7, 2021

Austrian website’s use of Google Analytics found to breach GDPR

A decision by Austria’s data protection watchdog upholding a complaint against a website related to its use of Google Analytics does not bode well for use of US cloud services in Europe. The decision raises a big red flag over routine use of tools that require transferring Europeans’ personal data to the US for processing — with the watchdog finding that IP address and identifiers in cookie data are the personal data of site visitors, meaning these transfers fall under the purview of EU data protection law. In this specific case, an IP address “anonymization” function had not been properly implemented on the website. But, regardless of that technical wrinkle, the regulator found IP address data to be personal data given the potential for it to be combined — like a “puzzle piece” — with other digital data to identify a visitor. Consequently the Austrian DPA found that the website in question — a health focused site called netdoktor.at, which had been exporting visitors’ data to the US as a result of implementing Google Analytics — had violated Chapter V of the EU’s General Data Protection Regulation (GDPR), which deals with data transfers out of the bloc. “US intelligence services use certain online identifiers (such as the IP address or unique identification numbers) as a starting point for the surveillance of individuals,” the regulator notes in the decision [via a machine translation of the German language text], adding: “In particular, it cannot be excluded that these intelligence services have already collected information with the help of which the data transmitted here can be traced back to the person of the complainant.” In reaching its conclusion, the regulator assessed various measures Google said it had implemented to protect the data in the US — such as encryption at rest in its data centers; or its claim that the data “must be considered as pseudonymous” — but did not find sufficient safeguards had been put in place to effectively block US intelligence services from accessing the data, as required to meet the GDPR’s standard. “As long as the second respondent himself [i.e. Google] has the possibility to access data in plain text, the technical measures invoked cannot be considered effective in the sense of the above considerations,” it notes at one point, dismissing the type of encryption used as inadequate protection. Austria’s regulator also quotes earlier guidance from German DPAs to back up its dismissal of Google’s “pseudonymous” claim — noting that this states: ” …the use of IP addresses, cookie IDs, advertising IDs, unique user IDs or other identifiers to (re)identify users do not constitute appropriate safeguards to comply with data protection principles or to safeguard the rights of data subjects. This is because, unlike in cases where data is pseudonymised in order to disguise or delete the identifying data so that the data subjects can no longer be addressed, IDs or identifiers are used to make the individuals distinguishable and addressable. Consequently, there is no protective effect. They are therefore not pseudonymisations within the meaning of Recital 28, which reduce the risks for the data subjects and assist data controllers and processors in complying with their data protection obligations.” The DPA’s wholesale dismissal of any legally relevant impact of the bundle of aforementioned “Technical and Organizational Measures” (such as standard encryption) — which were cited by Google to try to fend off the complaint — is significant because such claims are the prevailing tactic used by US-based cloud giants to try to massage compliance and ensure EU-to-US data transfers continue so they can continue business as usual. So if this tactic is getting called out here, as a result of a single website’s use of Google Analytics, it can and will be sanctioned by EU regulators elsewhere. After all, Google Analytics is everywhere online. (See also the extensive list of extremely standard measures cited by Facebook in an internal assessment of its EU-to-US data transfers’ — in which it too tries to claim ‘compliance’ with EU law, per an earlier document reveal.) The complaint back story here is that back in August 2020 European privacy campaign group noyb filed a full 101 complaints with DPAs across the bloc targeting websites with regional operators that it had identified as sending data to the US via Google Analytics and/or Facebook Connect integrations. Use of such analytics tools may seem intensely normal but — legally speaking, in the EU — it’s anything but because EU-to-US transfers of personal data have been clouded in legal uncertainty for years. The underlying conflict boils down to a clash between European privacy rights and US surveillance law — as the latter affords foreigners zero rights over how their data is scooped up and snooped on, nor any route to legal redress for whatever happens to their information when it’s in the US, making it extremely difficult for exported EU data to get the necessary standard of “essentially equivalent” protection that it gets at home when it’s abroad. To radically simplify: EU law says European levels of protection must travel with data. While US law says ‘we’re taking your data; we’re not telling you what we’re doing; and you can’t do anything about it anyway, sucker!’. US cloud providers that are subject to Section 702 of the Foreign Intelligence Surveillance Act (FISA) are all in the frame — which takes in a broad sweep of tech giants, including Google and Facebook, since this law applies broadly to “electronic communications services”. While Executive Order 12,333, a Reagan era mandate that’s also relevant as it also expanded intelligence agency powers to acquire data, is thought to target vulnerabilities in telecoms infrastructure. The EU-US legal clash between privacy and surveillance dates back almost a decade at this point. It was catalyized by the 2013 Snowden disclosures which revealed the extent of US government mass surveillance programs — and led, back in 2015, to the EU’s Court of Justice to invalidate the Safe Harbor arrangement between the bloc and the US on the grounds that EU data could no longer be considered safe when it went over the pond. And whereas Safe Harbor had stood for around 15 years, its hastily agreed replacement — the EU-US Privacy Shield — lasted just four. So the lifespan of commercially minded European Commission decisions seeking to grease transatlantic data flows in spite of the massive privacy risks has been shrinking radically. Some complaints about risky EU-to-US data transfers also date back almost a decade at this point. But there’s fresh enforcement energy in the air since a landmark ruling by the CJEU in July 2020 — which struck down the Commission’s reupped data transfer arrangement (Privacy Shield), which — since 2016 — had been relied upon by thousands of companies to rubberstamp their US transfers. The court did not outlaw personal data transfers to so-called third countries entirely. Which is why these data flows didn’t cease overnight smack bang in the middle of 2020. However it clarified that such data flows must be assessed on a case by case basis for risks. And it made it clear that DPAs could not just turn a blind eye to compliance — hi Ireland! — rather they must proactively step in and suspend transfers in cases where they believe data is flowing to a risky location like the US. In a much watched for follow-on interpretation of the court ruling, the European Data Protection Board’s (EDPB) guidance confirmed that personal data transfers out of the EU may still be possible — if a set of narrow circumstances and/or conditions apply. Such as the data can be genuinely anonymized so that it is truly no longer personal data. Or if you can apply a suite of supplementary measures (such as technical stuff like applying robust end-to-end encryption — meaning there’s zero access to decrypted data possible by a US entity) — in order to raise the level of legal protection. The problem for adtech firms like Google and Facebook is that their business models are all about accessing people’s data. So it’s not clear how such data-mining giants could apply supplementary measures that radically limit their own access to this core business data without a radical change of model. Or, well, federating their services — and localizing European data and processing in the EU. The Austrian DPA decision makes it clear that Google’s current package of measures, related to how it operates Google Analytics, is not adequate because it does not remove the risk of surveillance agencies accessing people’s data. The decision puts heavy underscoring on the need for any such supplementary measures to actually e nhance standard provisions if they’re to do anything at all for your chances of compliance. Europe puts out advice on fixing international data transfers that’s cold comfort for Facebook Supplementary of course means extra. tl;dr you can’t pass off totally standard security processes, procedures, policies, protocols and measures as some kind of special Schrems II-busting legal magic, no matter how much you might want to. (A quick comparable scenario that might hammer home the point: One can’t — legally speaking — hold a party during a pandemic if lockdown rules ban social gatherings simply by branding a ‘bring your own bottle’ garden soirée as a work event. Not even if you’re the prime minister of the UK. At least not if you want to remain in post for long, anyway… ) It’s fair to say that the the tech industry response to the Schrems II ruling has been a massive, collective putting of heads into sand. Or, as the eponymous Max Schrems himself, honorary chair of noyb, puts it in a statement: “Instead of adapting services to be GDPR compliant, US companies have tried to simply add some text to their privacy policies and ignore the Court of Justice. Many EU companies have followed the lead instead of switching to legal options.” This charade has been possible because — to date — there hasn’t been much regulatory renforcement following the July 2020 ruling. Despite the European Data Protection Board warning immediately that there would be no grace period for coming into compliance. To the untrained eye that might suggest the industry’s collective strategy — of ignoring the legal nightmare wrapping EU-to-US transfers in the hopes the problem would just go away — has been working. But, as the Austria decision indicates, regulatory gears are grinding towards a bunch of rude awakenings. The European Commission — which remains eager for a replacement to the EU-US Privacy Shield — has also warned there will be no quick fix this time around, suggesting major reforms of US surveillance law are required to bridge the legal divide. (Although negotiations between the Commission and the US on a replacement data transfer agreement are continuing.) In the meanwhile Schrems II enforcements are starting to flow — and orders to cease US data flows may soon follow. In another sign of enforcement ramping up, the European Data Protection Supervisor (EDPS) — just this week — upheld a complaint against the European Parliament over US data transfers involving use of Google Analytics and Stripe. The EDPS’ decision reprimands the parliament and also orders it to fix outstanding issues within one month. The other 101 complaints noyb filed back in 2020 are also still awaiting decisions. And as Schrems notes EU DPAs have been coordinating their response to the data transfer issue. So there’s likely to be a pipeline of enforcements striking at usage of US cloud services in the coming months. And, well, a lot of sand falling out of eyes. Here’s Schrems on the Austria DPA’s reasoning again: “This is a very detailed and sound decision. The bottom line is: Companies can’t use US cloud services in Europe anymore. It has now been 1.5 years since the Court of Justice confirmed this a second time, so it is more than time that the law is also enforced.” “We expect similar decisions to now drop gradually in most EU member states,” he adds, further noting that Member State authorities have been coordinating their response to the flotilla of complaints (the EDPB announced a taskforce on the issue last fall). “In the long run we either need proper protections in the US, or we will end up with separate products for the US and the EU,” Schrems also said, adding: “I would personally prefer better protections in the US, but this is up to the US legislator — not to anyone in Europe.” While netdoktor has been found to have violated the GDPR, it’s not clear whether it will face a penalty as yet. It may also seek to appeal the Austrian DPA’s decision. The company has since moved its HQ to Germany, which complicates the regulatory jurisdiction component of this process — and means it may face additional enforcement, such as an order banning transfers, in a follow on action by a German regulator. There is another notable element of the decision that has gone Google’s way — for now. While the regulator upheld the complaint against netdoktor it did not find against Google’s US business for receiving/processing the data — deciding that the rules on data transfers only apply to EU entities and not to the US recipients. That bit of the decision is a disappointment to noyb which is considering whether to appeal — with Schrems arguing: “It is crucial that the US providers cannot just shift the problem to EU customers.” noyb further flags that Google may still face some pending sanction, however, as the Austria DPA has said it will investigate further in relation to potential violations of Article 5, 28 and 29 GDPR (related to whether Google is allowed to provide personal data to the US government without an explicit order by the EU data exporter). The DPA has said it will issue a separate decision on that. So Google may yet be on the hook for a GDPR breach in Austria. Penalties under the regulation can scale as high as 4% of a company’s annual global turnover. Although orders to ban data transfers may ultimately prove a lot more costly to certain types of data-mining business models. To wit: Long time EU privacy watchers will be aware that Facebook’s European business is on penalty time in Ireland over this same EU-US transfers issue. A preliminary order that Facebook suspend transfers was issued by Ireland in fall 2020 — triggering legal action from the social media giant to try to block the order. Facebook’s court challenge failed but a final decision remains pending from the Irish regulator — which promised noyb a swift resolution of the vintage complaint a full year ago. So the clock really is ticking on that data transfer complaint. And someone should phone Meta’s chief spin doctor, Nick Clegg, to ask if he’s ready to pull the plug on Facebook’s European service yet? Legal clouds gather over US cloud services, after CJEU ruling Max Schrems on the EU court ruling that could cut Facebook in two EU websites’ use of Google Analytics and Facebook Connect targeted by post-Schrems II privacy complaints European parliament found to have broken EU rules on data transfers and cookie consents

AppKit is done – macOS and SwiftUI

Well, not like Carbon. Don’t be so dramatic! More like Core Foundation. It’s still there behind the scenes, but programmers use high-level Objective-C and Swift wrappers from Foundation. If something is missing, you can call an underlying C API. The relation between SwiftUI and AppKit is similar, for now . This is a native macOS app written entirely in SwiftUI, from @main to bottom. Not a prototype, not a toy. A full-featured app. The intention is to deliver the best macOS experience possible. The two main reasons this app is possible are SwiftUI and Big Sur. I have extensive experience with iOS, so I started with it. When the time came to adopt it for macOS, I was anxious. I had no idea what to expect from SwiftUI on a Mac. But thanks to the new SwiftUI APIs and the Big Sur design changes, everything just made sense! Total Lines: 7124 iOS: 593 (8%) macOS: 513 (7%) Shared: 6018 (85%) App Size: 2.5 MB Want to learn how to build macOS apps or see some videos with rad UI interactions? This post is for you. The first step was to get the navigation right. I knew I wanted to use triple-column navigation . SwiftUI makes this setup ease. struct ContentView : View { var body : some View { NavigationView { SidebarView () Text ( "No Sidebar Selection" ) // You won't see this in practice (default selection) Text ( "No Message Selection" ) // You will see this } } } The key is the type system. What you used to configure in code, e.g. by passing a number of columns to UISplitViewController, you now configure using generic parameters. In my case, the type of the view is NavigationView<TupleView<(Sidebar, Text, Text)>>. Use navigation links as usual. To make the first tab active by default, pass isActive in the NavigationLink initializer with a value of true. struct SidebarView : View { @State private var isDefaultItemActive = true var body : some View { List { NavigationLink ( destination : ConsoleView (), isActive : $ isDefaultItemActive ) { Label ( "Console" , systemImage : "message" ) } // ... } . listStyle ( SidebarListStyle ()) // Gives you this sweet sidebar look } } By default, the toolbar displays a title in the main panel. To disable it, I used UnifiedWindowToolbarStyle(showsTitle: false) . To learn more about triple-column navigation, see Triple Trouble . Don’t use the List selection API for navigation. It only exists to support single or multiple item selection of items in edit mode. The “Toggle Sidebar” button does not appear by default. To add it, use .toolbar modifier. SidebarView ( model : model ) . p { Button ( action : toggleSidebar ) { Image ( systemName : "sidebar.left" ) . help ( "Toggle Sidebar" ) } } } } Unfortunately, there doesn’t seem to be a way to hide the sidebar programmatically using NavigationView. I had to use NSApp APIs. One of the few AppKit usages in the app. private func toggleSidebar () { NSApp . keyWindow ? . firstResponder ? . tryToPerform ( #selector( NSSplitViewController.toggleSidebar(_:) ) , with : nil ) } The relationship between SwiftUI and AppKit are not documented and not guaranteed to be supported. This workaround is useful for now, but might stop working in the future. To add a “Toggle Sidebar” shortcut, use SidebarCommands . If you have a group of items, put them in ToolbarItemGroup . This will ensure proper positioning and spacing. Customize the positioning on buttons with ToolbarItemPlacement . The positioning that you can see on the video is the default option, no customization is needed. To create a collapsing section, use DisclosureGroup . You can even nest sections. To open a section programmatically, use isExpanded binding. DisclosureGroup ( content : { Toggle ( "Current Session" , isOn : $ model . isCurrentSession ) // ... }, label : { Label ( "Log Level" , systemImage : "flag" ) }) To learn more about outlines, see WWDC 2020: Stacks, Grids, and Outlines in SwiftUI With the new WindowGroup API introduces in Big Sur, SwiftUI takes care of certain platform behaviors. For example, users can open more than one window from the group simultaneously. When the user double-clicks on a list item in the main panel, it opens the details in a separate window. This happens automatically. It also comes with a standard set of shortcuts: Command-Tilde to switch between open windows, Command+W to close, Command+M to minimize, etc. You can also open a new window group using Command+N. In my case, the app remembers the database you selected. This way you can have multiple views on the same set of logs with different filters. Very useful when investigating complex issues! @main struct PulseApp : App { @StateObject var model = AppViewModel () var body : some Scene { WindowGroup { contents } . windowToolbarStyle ( UnifiedWindowToolbarStyle ( showsTitle : false )) . commands { // Adds `Command+Option+S` shortcut to toggle the sidebar SidebarCommands () // ... } } @ViewBuilder private var contents : some View { if let store = model . selectedStore { MainView ( messageStore : store ) } else { WelcomeView () } } } The way WindowGroup works is clever: it’s just struct copying. Every window created from the group maintains independent state. For example, for each new window created from the group the system allocates new storage for any State or StateObject variables instantiated by the scene’s view hierarchy. To learn more about data flow in SwiftUI, see SwiftUI Data Flow. Search is probably the most complex UI feature I built. It felt like I was pushing it. The search field, the search toolbar, and how the searches are performed is not particularly interesting – same as in AppKit. What is, is how do you scroll to the next item and activate the navigation link each time the user hits the Return (↩) key? For programmatic scrolling, I used ScrollViewReader . struct ConsoleContentView : View { @ObservedObject private(set) var model : ConsoleViewModel var body : some View { VStack ( spacing : 0 ) { if ! model . isSearchBarHidden { searchToolbar } ScrollViewReader { proxy in ConsoleMessageListView ( model : model ) . onReceive ( model . $ selectedObjectId ) { objectId in guard let objectId = objectId else { return } proxy . scrollTo ( objectId ) } } filterToolbar } } } When the user hits Return (↩), ConsoleViewModel updates the published selectedObjectId property value observed by the view. Every declarative framework is imperative if you try hard enough. Now, what about activating the navigation links? We already know how to do that using isActive . But what do you do in a dynamic list? The solution I came up with is to create an array of bindings and grow it every time the number of items increases. final class ConsoleViewModel : ObservableObject { @Published private(set) var messages : [ MessageEntity ] @Published var isLinkActive : [ Bool ] private func refresh () { // .... messages = /* new array */ if isLinkActive . count < messages . count { isLinkActive += Array ( repeating : false , count : messages . count - isLinkActive . count ) } } private func scrollToMatch ( _ match : Match ) { selectedObjectId = match . objectID isLinkActive [ match . index ] = true } } struct ConsoleMessageListView : View { var body : some View { List ( messages , rowContent : makeListItem ) } private func makeListItem ( message : MessageEntity ) -> some View { NavigationLink ( destination : DetailsView ( message ), isActive : $ model . isLinkActive [ row . index ]) { ConsoleMessageView ( message ) } } } By the way, this is not the only way to navigate a list with a keyboard. By default, on macOS, List can also be navigated with “Up” and “Down” arrows! List is not great. The performance with a large number of items (10000+) is unacceptable. The reason is the diff-based mechanism used for computing animations that you can’t disable. If you are considering displaying a large number of items in a List, it most likely won’t work for you. A similar search is also available on a details screen. This is the only place (except for NSSearchField) where I had to use AppKit components directly. SwiftUI doesn’t yet have a wrapper for NSTextView (except for TextEditor , but it doesn’t support attributed strings that I needed for syntax highlighting). So I had to create one. Now here is the question. Does it mean that SwiftUI is somewhat incomplete? Maybe. After all, text view is a common component. But SwiftUI is built on top of AppKit, it is not going anywhere. Just because I had to wrap two components doesn’t mean that SwiftUI was in the way. Quite the opposite! The AppKit integration works great. Despite the rumors of SwiftUI’s lack of programmatic navigation, it exists. One of the cool Pulse features is the ability to pin important messages. Imagine you are investigating a bug report, you found some related messages, but you are not sure. You can pin them and get back to them later in the “Pins” tab. You can also pin a selected item using Command+P shortcut. But this isn’t enough, is it? What you want to do is the ability to open pinned messages in the list. Pulse allows you to just that. To implement it, PinsViewModel sends a message to MainViewModel: “hey, open a message with this ID”. MainViewModel activates the console navigation link and asks ConsoleViewModel to scroll to the given link. The console uses the same implementation from search. Simple. To add a context menu to the view, use .contextMenu modifier. It’s relatively easy to use and there are plenty of code samples online, including the official documentation. The context menus look different on different platforms. On iOS, you invoke a context menu using a long-press. On macOS, it’s a right-click. You can add context menu to pretty much any views. Menu is a related component. Unlike .contextMenu, a menu has a label (can be a button) to be displayed on the screen. To invoke a menu, you press the button. On macOS the menu also automatically adds an arrow to the control. Menus can have pickers, this is how you create hierarchy. Menu ( content : { Picker ( searchOptions . isCaseSensitive ? "Case Sensitive" : "Case Insensitive" , selection : $ searchOptions . isCaseSensitive ) { Text ( "Case Sensitive" ) . tag ( true ) Text ( "Case Insensitive" ) . tag ( false ) } // ... } The menu from the video can be improved even further. When there are only a couple of items, there is no need for a hierarchal menu. To flatten the hierarchy, use InlinePickerStyle. Unlike iOS, there are multiple ways to implement sharing on macOS. A common approach is to fetch the list of sharing services available for the current item and put them in a context menu. My initial implementation looked like this: Menu { ForEach ( NSSharingService . sharingServices ( forItems : preview ), id : \ . title ) { service in Button ( action : { service . perform ( withItems : self . items ()) }) { Image ( nsImage : service . image ) Text ( service . title ) } } } Now, here is a problem, the first time you call NSSharingService.sharingServices it is slow. And when you create a menu, its content is evaluated eagerly. When I added the menu to the details screen, the first time I open it, it would be unacceptably slow. My first attempt to fix it was to load services asynchronously when the menu appears while showing a placeholder “Loading…” button. This didn’t work. The menu just wouldn’t reload when the request is done. What I ended up doing – and I’m not sure I’m happy about – is pre-fetching the list of services when the app starts and just hoping that the user doesn’t open the share menu before that. If that wasn’t enough, here is a dozen more tips. Use .help to configure the view’s accessibility hint and its tooltip (“help tag”) Use .keyboardShortcut to add shortcuts for common actions that are not in the app menu. For example, .keyboardShortcut("f", modifiers: [.shift, .option]). Use .keyboardShortcut(.defaultAction) on a button to make it a default button (will use accent color) and register it for Return (↩) keyboard shortcut. Use .cancelAction for Escape. To change your app’s accent color, add a new color to your assets catalog and set it as a “Global Accent Color Name” in the build settings. On Big Sur, the system uses your app’s accent color by default, but the user can change the accent color to one of the predefined colors. Make sure your app looks great with all of them. Use GeometryReader for layouts where SwiftUI layout system doesn’t cut it. I used it for metrics chart. When adding commands, use CommandGroupPlacement to place items in one of the standard locations. For example, for “Open” command, use .newItem . SwiftUI has several built-in command groups, e.g. SidebarCommands , ToolbarCommands , TextEditingCommands , TextFormattingCommands . SwiftUI layout system and data flow are the same on macOS and iOS, see the linked posts to learn more. Gotta learn these first. Use Settings to add a settings screen, don’t add it anywhere in the app itself Instead of navigationBarTitle, use navigationTitle . macOS also supports subtitles. Use Alert to show alerts, same API as iOS Prioritize watching WWDC and reading official documentation, there is now a lot of outdated information about SwiftUI online SwiftUI isn’t perfect. I had to compromise in a few places. But I don’t have a lot of bugs to report . Maybe I’m just getting better at avoiding things that don’t quite work as expected. There are some limitations. But the AppKit integration is always there for me. One of the criticisms I hear a lot is: “I started using SwiftUI and it took me T amount of time to build X, it would’ve taken me a fraction of time to do that in AppKit”. I voiced this criticism. SwiftUI is complex and is not magic. It is almost nothing like AppKit which I think is a good thing. But it means that you need to learn a lot before you can use it efficiently, even if you are already familiar with the core principles behind it. There are still a lot of things that can’t be done using only SwiftUI. I can sacrifice certain features for my app, but I can’t sacrifice performance. So far, List was the main impediment (as long as you have a decent number of items to display). If you want to learn more about List limitations, see the next post. Is SwiftUI a game-changer for macOS? It’s economics. It might seem like web technologies are dominating the desktop. On the one hand, there is M1 which is finally powerful and energy-efficient enough to run Slack, rejoice! On the other hand, the calculation is changing thanks to SwiftUI. AppKit is arguably one of the main impediments for developing apps for a Mac. But SwiftUI opens this market for millions of iOS engineers. The path to delivering great native experience on Apple platforms is becoming clearer than ever . I heard you can even put Swift on a server, but was not able to reach Tim Cook for comment .

Great Stirrup Controversy

Avar stirrups from 7-8th century Hungary. The b is the academic debate about the b, the theory that feudalism in Europe developed largely as a result of the introduction of the stirrup to cavalry [1] [2] in the 8th century AD. It relates to the hypothesis suggested by Lynn Townsend White Jr. in his 1962 book, Medieval Technology and Social Change. [3] White believed that the stirrup enabled heavy cavalry and shock combat, which in turn prompted the Carolingian dynasty of the 8th and 9th centuries to organize its territory into a vassalage system, rewarding mounted warriors with land grants for their service. White's book has proved very influential, but others have accused him of speculation, oversimplification, and ignoring contradictory evidence on the subject. Scholars have debated whether the stirrup actually provided the impetus for this social change, or whether the rise of heavy cavalry resulted from political changes in Medieval Europe. White begins by tracing the research of the 19th century German historian Heinrich Brunner, who claimed that the switch to mounted warfare occurred after the Battle of Tours with a Saracen army in 732. Brunner pointed out that Pepin the Short began demanding horses as tribute from the Saxons in 758, citing this as evidence of an increasingly cavalry-dependent army. [4] : 3  Brunner also claimed that the Muslim incursion into Europe prompted Charles Martel to confiscate church lands to support cavalry. [4] : 5 White used linguistic changes and evidence of a drastic change in weapons to support his claim that this change to mounted shock combat occurred in the early 8th century. [4] : 27  He claimed that the francisca (Frankish throwing ax) was replaced by longswords and lances — weapons designed to be used from horseback. The lance, White says, is the strongest evidence that the Franks had adopted the stirrup by this time. [4] : 28  He further claimed: "The feudal class of the European Middle Ages existed to be armed horsemen, cavaliers fighting in a particular manner which was made possible by the stirrup." [4] : 28  He believed that the stirrup had enabled the knight to exist. This hypothesis was also supported by Canadian media theorist Marshall McLuhan. McLuhan believed transformative new technologies, like the stirrup or printing press, extend a man's abilities to the point where the current social structure must change to accommodate it. Just as the car created the Interstate Highway System, the suburb, and the oil industry, so the stirrup helped create a specialized weapon system (knights) that required land and pasture to support it and provide for training and material. Criticisms of White's ideas Despite the great influence of White's book, his ideas of technological determinism were met with criticism in the following decades. It is agreed that cavalry replaced infantry in Carolingian France as the preferred mode of combat around the same time that feudalism emerged in that area, but whether this shift to cavalry was caused by the introduction of the stirrup is a contentious issue among historians. It has been asserted that armored cavalry were used successfully without stirrups before their introduction, and that the transition to cavalry was not a result of new technologies. The first fully armoured cataphracts appeared in the third century BC, almost 1000 years before the Carolingian dynasty. White argued that they were "essentially armoured bowmen." [4] : 9 Sawyer and Hilton's critique In an April 1963 review of White's book, the scholars Peter Sawyer, of the University of Birmingham, and R. H. Hilton, were quick to point out that "the most serious weakness in this argument is that the introduction of the stirrup is not in itself an adequate explanation for any changes that may have occurred. The stirrup made new methods possible, not inevitable ... the stirrup cannot alone explain the changes that it made possible." [5] Sawyer and Hilton further point out that the scant archaeological evidence makes it difficult to determine when the stirrup reached the Franks, as they were already Christian by the 7th century and had largely abandoned elaborate burials and grave goods. [5] : 93  They also stated White's footnotes often contradict his thesis and evidence. Stephen Morillo's argument Military historian Stephen Morillo, of Wabash College, offered a different explanation for the rise of cavalry in Medieval warfare: that of a lack of centralized government. Morillo contends that cavalry-dependent militaries are common in societies that do not have strong central governments, and cites Medieval Japan and China as analogous examples to 8th century Europe. A central government, he explains, is crucial to the development of a highly trained infantry, but cavalry can be maintained, however loosely, by an already horse-owning noble class. He writes: "Rural warrior elites were in fact a common feature of many traditional civilizations. Sons of such classes were raised to the military lifestyle, trained in small groups built from the social connections among the class, and exercised military force in the interest of maintaining their own position in the hierarchy of power." [6] : 52  Furthermore, Morillo examines cases of Frankish warriors fighting on foot—and defeating mounted knights in the process. Even White quoted Brunner as admitting that a good infantry could break a cavalry charge if its soldiers held their ranks. [4] : 5  Morillo used the example of the familia regis , an elite Anglo-Norman infantry unit,[ p ] as further evidence that a strong central government was the key to developing a strong infantry. Therefore, Morillo considers feudalism a political construct rather than a military one. Objections from archaeology and experiment It has also been asserted by some, including Richard Alvarez, [7] that modern reenactment and experimental archaeology have shown that the stirrup provides very little benefit for a mounted lancer, and a cantled saddle and spurs have a greater effect. White noted the importance of the prior emergence of the saddle, but argued, "The stirrup made possible—although it did not demand—a vastly more effective mode of attack" (than a blow "delivered with the strength of shoulder and biceps"): "now the rider could lay his lance at rest, held between the upper arm and the body, and make at his foe, delivering the blow not with his muscles but with the combined weight of himself and his charging stallion." [4] : 1–2, 7  Stirrups provide stability for striking in melee after the initial cavalry charge. ^ ^ Stix, Gary. "The Stirrup". Scientific American 301 (3) p.78 ^ ^ a b c d e f g h ^ a b ^ ^

Ontology Is Overrated: Categories, Links, and Tags [pdf]

euN@&i)WсVU<o%E}1,![Fs]'y_  _OFѭ:tk߻85 1[ĐAGj*jg?Qab7o?\}EԿ4ɭxw.J?g/5?*EY[+{L7n\
O`{%+|GſcyLwzqQ/ &o {O$B+zmuGd|v+?0Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@
|;>a_K (<'Y{d,kؼM`%a?/GCl/&b~k^&]MOĿ rVq?#oǙe;~p[ ^:灾s>#-Kx%q yo'pἿ4&GgPx¿/))Džc塼+a9xE|G4n_ʁ*'3\8ya?|}EԿ}-_x{5Gat|KǕ6>TNgBqO[[>#iZLzu\ix}.TQ9Jdgis?g?z_ |~"ޡ|Fxۦd]<Υr239' %Ϯj:'k-7FKk2Oʞ[EJdgasDzako$>?dVR}l4I]y==(~"u_qI[^y#K EM[9.:@9 ܘJポk~~v=~)ibM>8UzgqM&~4hCO Be~i|+aD}%sx.'(H=z}o Aѿ+j 6_o"c?Dɔ_LG 2Ah?/71e_o"c?Dɔ Aѿ(&Q/EF&?L_LG 2Ah?/71e_o"c?Dɔ Aѿ(&Q/EF&?L_LG 2Ah?/71e_o"c?Dɔ Aѿ(&Q/EF&?L_LG 2Ah?/71e_o"c?Dɔ Aѿ(&Q/EF&?L_LG 2Ah?/71e_o"c?Dɔ Aѿ(&Q/EF&?L_LG 2Ah?/71e_o"c?Dɔ Aѿ(&Q/EF&?L_LG 2:Veֵ8BcTKŲ(s6O;wzh (|q%ǿJ`FهM rV"8Rf'&o(K`%g=lf _OM3 rR|C?\Gx_ɏ?3cwoD0._aOu72ZiD0;?&_a\|G~~>1-g[tT T׵>|E䘚w^q[FM˗;BXRqyZ+S ( ( ( ( ( ( ( ( ( ( ( ( (w }GڿO]MP
ExO?LY
0ɱx.J^>"_L?ּM`%wGakr =ɦx7.JOh~'=#K1GxLnm8|7l!)rB&\K_ (|'lfpm?4xĿ?ekxJY0dILk&C 7)ɸ50ЧU^2M?G{ßzVyImɃն]-tg$
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP-@.(g?Qab7o?\}EԿ4ɭxw.J?g/5?*EY[+{L7n\
O`{%+|GſcyLwzqQ/ &o C2SM;L~Q? O iK{LO h El:gFf<]ȭBʃ#WWώ 5gFCKg6ѽbrqZ:<JG*U(7.[io3|- o~^RBm#I֭ϟAEQ@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@
|;>a_K (<'Y{d,kؼM`%a?/GCl/&b~k^&]MOĿ rVq?#o?4xm.#.\dsYNO½?^\Et0kWSgO*.~Giթ8]o͛,WzO[-Z<f n9vͼC#Rhj*P-G0Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@
|;>a_K (<'Y{d,kؼM`%a?/GCl/&b~k^&]MOĿ rVq?#o\r0pjj ( ( ( ( ( ( ( ( ( ( (w }GڿO]MTյ4˭NK[0@ ;cyv=כCcᯈg˙|azn2CLҁ!`ػ<@Wӯt" }WPtĖY?͚I*nt{s3!xCJ ZON%䇎Eᕇ8*y4
8%0f\&o+ z?e_3hZ7\~_j~%U3|W'&o)>ɡ_.J#?+əҗŷ?// '_3g/4-;Bo+['&͗(&ݏ_e
_|a7Cګ<>"e_$߉o>CK(4/-A:Z<@@:{տ>j^ H~DJ̯ xrʫ\VO; B6jIrs[^,i:ŭ go3('+^%Npg%e-hS ((((((((((QS@5W#n_DPXEu3A ɏ<ʹB#~) <*)ilox摭HsL2!$O'×Z7W~3KQMtA`ys>BdcԾO OߥΟԢKq,4̀HA g|7_xǟq3miJo/Vʄ1
Gm!QZo9 z?%_37Ý?wŶ}ԉNՈG8{h'H`)d ]-ӮH,/)lpY8Nmq:T_=wŲ]g8,~|E_0XO /ğ:=b|[-{js`˛mqڏt}?_dZecjgFqe.rN`x8=( ILGf;OC[{E6>(l,,=NJ>-8٢p-CouNå؄(¿PwD>hW٧x wų]I5.QY[ }L|gs?5u-Ťf jwv`1m GbT~ `?O'-ළC(
9\NGF×|]tu;ͲhS<~;4Wm{~^վͦɨxSh }kK㦁uZ|;Ԣ:ԞrN蘺P976 B&Uu?ͯ<[τ1cC/!hlwپx6dĕi_ I6cjQJTMvM8PÏ/M8PÏ/M9C8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M8PÏ/M τz\'g-cU9/.弝og<JެX$i@Q@,=Sj5~?l^&mȶ_1?f5.EG&_]+8c7Z܂~i&M?u=?CoLQ8?8.[*%a ar|{&JܶУ'ɗ/CJ'a a2?[O )~/ɜ[i__|e<mBӿ뽟(¿Qe"_hW|Bi7}:*}EW0aKM:O*m/U\7G~wOQWQ_*ҩWm{]!/'Y!τo~^UZo'%YN0pW</[rGㆩЗK#켙yyZ>9i|Y/ \rW^իnk_ySh坝HNn99㟎:g!xo׺LgPtc'5|RӴoW&muH8 ֫jպmB*rÏWZ:E結y?mիM=6)ҷ:ך{0Q@Q@Q@Q@Q@Q@Q@Q@
|;>a_K (<'Y{d,kؼM`%a?/GCl/&b~k^&]MOĿ rVq?#oh_`܂\׹b߈p YЅWY~4yS9_M,?&) i+m85x_ӿl6ҭuKe6H=HL|RNHFwW?Q?Oғ|5Bk8J7Gi:n~^ivfCs)lbT:UAߖOYiF? tqhCs)6N1]ݯįxu8Dۘ@CІ<Tj_~8Y\uJjJ/k,(O (((((((QS@Q@,=Sj5~?l^&mȶ_1?f5.EG&_]+8c7Z܂~i&M?u=?CoLQ8?8.[*%a ar|{&JܶУ'ɗ/CJ'a a2?[O )~/ɜ[i__|e<mBӿ뽟(¿Qe"_hW|Bi7}:*}EW0aKM:O*m/U\7G~wOQWQ_*ҩWm{]!/'Y!τo~^UZOڧG%?A V܁SJ?j?( oO0_^kAh. 1]o_8O,BĿC<- #яGL/]kODׄ4jzRהTO =+RzE]?*cF~՚$'E6ǫ(,‹ W^f9#$mˏ\j5Vrsa7*}g]i k.ʺ7y#n>\}kg|Mo) ȁ_~8OWӫ_qʼn_˧˘O |J5HBkI(2ɂ~Z|? x#k61>*A[ӌԒ}1Yul%Jj[.d͊*΢ ( ( ( ( (9o=t
EPK?py'ECp?-~O٣Mk rQ9ɩW.J?g?z_ gw&(O[?--*gC}a7\ɒ-(oe҉`wL&gCJ_gkDW,`&`,뭟дg *}Yn>_6_n_v?Ίo~kRyw4jS_Ka qߵS_A~ʴ}Ue
^WK7~IVs7}h_`6׺x~_3&7v?iOxU ~=|'ڏJ'-~ ך_;K_?W[N? ?//p?*g%= CGc|5?-`Z? 5A?`OJT?zE]?*cOORh'Yi.W+5?
< v >!Jt͙Gqg%[B42Iį$mU%rqǜT33п(x[4J{dծg2!x´f.?ixXb'p Tcascʯeo+w7OH>qwrEQZ|@f}+P-H-+`!s_̳,2tm riH9ӮtɭBO0:!}'ҵ>g}?RѠm.[iۼ8.դjrv3'XY]RqOϚ_|L KWq J$"7yǽu JJXj:ҡUZQv`# 斨 ( ( ( (w }GڿO]MP
ExO?LY
0ɱx.J^>"_L?ּM`%wGakr =ɦx7.JOh~'=#K1GxLnm8|7l!)rB&\K_ (|'lfpm?4&qmO~ g LkKM N E1~9ɣ_e >=ɦc_톿/7cM?FWpIW߅??G]GJ__0wʴg>VUl{U#?o?i?j>cߪ/[rM#(|'}?@y5su<` i`Ҧr^*зt4xOF?S2_Au=^ѩK^QS]? ./C_פ_ʅB/??h?/RxGO/ş'S gN~4OWٗ~_UaI#G%7*ct0 $D.?:?dO|{a;ҧ>a~?7W?d.:)O]fؿWC4S~yKn.$i3 ˿[a}q GvRX㎔R)ŵMw_ڗhl}~ y=*^-]^ij==.2aNl^Hn:aFR|H k
/%Mamap>CHqS|<x]OJ:K"%"' NHRݽ.rW)C(5_] ~>|\Bca%ܶC:A\U qzK8,Z&GTVlU|\ʦỌ=UoenxW⿄Omx4%pn6I=5+zMGya{xVD=ԭv#W.~*@ (9o=t
Eyo4_c%TMwE p8 wL*rr6~ RA1jp +#<f 'x/>|?q82|6M
v7r~a5 h?爼7Qņ6t#$q0cX7;?pu/9I){jnN]]>r5~!εGmȎS+,lOM
~9|?gcXbm.c$q6Qbt$%Ij/-Vj'ߎ_o٧>(bYJX8cb P"-8|oIG]S⇆-5OS;,"ێN@ uQkRZc{ޚv%/?W O^\46=N M 83Ŀ ~6=N &*b 9Qz0x~j[ZG]eᛍn1nEB*nqԿ~8dEt>c6|:2\+BA6@҉Em>B Q.\DZGhG/z좞>(bX58dR m9|տeH|1QjppYKmON,.>4kOU}_e>*|q7~~(bWkQ% U1OQkg]֑G7K}NgB aO˧
\'>K7h/ZçxoGV Q- nX9|?fm/Ú/ _]yVڜ2nX|([ӢC.P.yZmOZ=|O~ j<-+wgX5Ŵ2KGˠlVW׉ß|1 >=ٳ@Kcdϥ:mJ˱[Ni[^kK;k|5ß1&x$k=N F;FAI|s5wg7>)xcTO.SupF;GJqoNsM.XUOU|O?i/Myß1EﻒSuv.UQ~? xUeuߊ^!/w% ^\ɢqoN B 4iOU|OwVp9.ÿ|1Stp+ɤ9|/x{◆5;m>BsZjpΐ.eb{'@C|KTZ7d7un3xᖡ>t/w= ^]QSG׈oR jVy/uqi4p.XZ'A1t!NiFzOwV]n/ǯϋ⏆/ s}NaXk2o|Akw2vnSV 6OCX.ήN/]swZϊ>9ھ/ ,bF0oC.h/MX/ b-Hu8~ݷ?.W2г]kX/ YD'S۽<sѪ|s4+o>]'ۺcݷ? /t7ݔw_ﯖy~W^+◅E6 [S[wmD EY^_^ME[S[A.hQzi. /wr|r5k⩾(a4i PmNl%ݴlxoڷľ({}ưjpn0>i[AªSZj$u s3aVx/oݺ,poD?C]_ÿ fwVZKh#yk=3NjuϏQwI-zKUguro?>KRƙoj3z0'1`c)?f|3w"2 Fy-$g]enT%s|B׹f2mVCdk> <׈)xcKQ58`l27 w~9Ǎ3|QƖ2ZN<ߜ#D#%uy*OM>gKkmF xVo>ZKe̻9QGׅG"7 i2^Il04˻ v#(.K B&4٭RCNk¶*7 iR\o 2azZ?d_+|1Oy={L`Xr9֊q2ЩIR.k.߅_ i
|WC⯅[+);m'5GzO |1iMNppŕm'4(4$Pqiq1|>|oJAmK⏆-utɕ48R+"ێH q(Oޛ0º [k2*ipV EH8QM>Utiq*KU~7I;߉[KNRnYX(IԾnL׼#uGkk `vm88e쎦.Qsn}O= [~.>(b=qtE|ݿ'sW< E|/?e]SE1h>)MZ$C#ݼ8FKH# Q8qi⹞=\?m7t0,˩C%3nf#=nckoIM|D4itan81I
ڭ9? T5S}mw6-'|ek Ij-gBHytCrE ^@Q3}ɻ '|f *WEZ~N4oCg48s"k]:+ݺpYw53ftտpxsRV!0iNz6%B~z+t9s⹛Z+Zھ]^`G㟃dB ;w *G*ǁT]8#:ͯ!|W xgXtsX{{9xe_2q]
ᖁ2RմCFIl5-&* p<)}^|%1ii6kkbQFXz<I<usdҺ`ѝ7ȊifRۧ|LQ=s@<xN^Ӥwڽךvۧ𷃴ͮOpj^kQ4xS_Z}2K:.s>M"Fq}ht \ir\?֩6q>M*_EkWzW խnuNUqMx~QG</Px7Ji|͑l?#cf[x/Lmait][r x3J<pvmJ(ʼnchǃ4 x6taejli_tG-K~e.1CҼs@<cv> Ӣi_t [z_x?Lׂ| c%~dM+ıc-. 3þ -iq4C F#nuBxIOӴȴG!".[m_DB?_N|+;Rz׀Sܗ#L2Iv@P͎\@|e-7_k,ꂏKMφ/:Z^̆$ txrēFc,ꢓƾcuu*(sX-f)X1񯃴MbKycnï>Q^Ӽ_CjrNCyozK
>QQx_5i.;BR ܀z<_/ƶ}%&a}%A^yk E n|9q-ͨ4x#]nh#n;,ZoMwQ7K׼GK=QȪ~b g^]7r\-߇UG7w(^'мYu%x}nQ';\~4c:dotT|#-IwR柨xWNԼSy'-ݭ.L&MùGLzs@ <#j>/Ѽk<UGt.si=9𖛨xHN/[>U],z|޴^OMiaX_iꭈWOllZmtLNPt)eٴ[ր  4ԼzTzw3:ߍ7ʳ5/bMԙҀ 3^ wƶ\[YZ+cT󼭋؟>}xJnx^񝴗QYCtVdX.8ȕGt~ 񅬗khj[#${n֍^K)kwkxz4_zE9.KxEh XS`F}MnKkf}؞r _`P
~ UJXծ݉|}=jO ZXj~-Ӭ/'`Hmt7 e2n;o_ h"DTDr\Z
 >׵&ևm;Y\\6%۹Yܞ<N&k7z]uy{ 9s$vC (;Y^(PY h͎0V*3)ǡ 6+U_ 4縷{[KqIJ.R#R@8A׎煥|IjڭܖHŬ"Y&F A@=Nk}:}2 ㉚:h§h >;;kZ-V рJ$$:v']M5i"Bg7 ;YeYܨ* ƀ;F_ \B-!{tY7Fas+>7`<i iwM]đm?+z@~ X떖=ղOgm{SZʒ(a{}nĞ |@߇&^ԋ^̋ rS|*9f ],F+ܨ?x 2|__V,:5-؎46 SGg!o3^u]K<Es" +d4
2A"~"vmj͆:yy5q q&̟6Dusfέ<-*h&x_-GxfJ>iͥ^%屒f|Q#\M{-xVῇ~KNHu90{n*QDٸQτUiWf]\_ztEM$oپ㱂4f A/{h>4{˽K9n?zҘmP0 vWc@S04]~¦/h:5L-Ȏn@[Xex>5
k χuyt>Z`H.rh޷5߈]gX}m],.-d@[Dj`#$&:z|_oo A^m%ƥDlQA:o⧊>%]W᫝R{×-$Xwu3J F$ >nw@|Ft" 4 CX_дwwoзHW<|l~/Myuj\:|{c]i,C潴RI(x_!X7@A_:T$AgK7 xYsYLz.c,j B6l|
|>#źVjj=<ۧdlEF~e&OJufÚe&o_.4.šߴ<5Fk"i֗v\\ŢMD$;~m]^%Z.]h>!}6 +KI/0b<߰mPT@=F(-5? >n&ƙY\Wc: ~ūI K^߷ꚦ!Yne qĊUL5+I;{6n!X$crґ(μ+Z?/|OygPҥ1ȫXK3}t>h <9fe}mčIbOEUUP
j~Ěu8O]@SA:` fmcZ0Fr\V IbDbx#7!е_k+[+V X;8!"<Y>'ů f~$9fK󢜘d[enyP| ^f},Sh,D >HՇWS> Zd^R<I bfff8d(•3\kw'jYʷ[~suyW]W<e.XO>\Kx[|7ZH\ o˾ݏrI'^8-τؽƛ*PJހ8[xo<|?Ѽ{=H]A-Ry.7n-Q$n궱KPMu%̈ZiH@P7?í_ +O^ 5Yȗ15żc#tl|O2|ViwxŹub($XP*W3^(6/ğݸԴL7,ۤ匰;I<M:rͣ_:moil" X=my߀ ,Bf$g/0 TG b$ݖw}i مݲ*A+Df\*GIIE$P^w#ߊen"x} Jb+KF">jۋ<gۊg|UBxVWu)5X巶>pn%_x AjBߊz/Z^yyh Y7A#,Bi>Ye HD^o~ h?b𭍖^ X KHXA4
6oK"M7Zׯuac
qmgoS2 xc29pN} <ߍSo!'P}WLf+K[dP,*m08usP!v-V[mRI&g<<GpA
@S1~ ^[W/~(o]iZEv62Hk$$<|E_vijQZ@aⵆ(#1RĞ쌂2F{
5ևį&jm};Cgas -"9U(˼)7WE> endobj 69 0 obj<> endobj 70 0 obj<>stream ExifMM*
z







(

1
2
<

File created with CoreGraphicsH
H
QuickTime 6.5.22005:04:19 13:26:07Mac OS X 10.3.9 AppleMark ۄ       
  



















 

}
!1AQa"q2#BR$3br %&'()*456789:CDEFGHIJSTUVWXYZcdefghijstuvwxyz

w
!1AQaq"2B #3Rbr $4%&'()*56789:CDEFGHIJSTUVWXYZcdefghijstuvwxyz
7B
!


?SKg-:=U3I}bY4`HdbM Q+nǹ@@~7^'U&{ [%4i3$ b𗃓˶2yj˝&7FW1L"83,S*np
˩G_F>=|o}cQ]wUigmm#\ݏ㔰*Sv@]ig[i_h}[%6UR+|R2L.yV[e mj ?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[ ?w[
?w[]S4nAm,lFBFG(]G9fI>TQ@WSmbwPz( ( ( ( ( ( ( ( ( ( ( ( ( ( ( (  S)[PѶ_r_*z(+?ҴnL8һ2Ja' -IEؿ%ŅĤ%v c9 ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( (:?E}ep%BLtiOkڅtˠNb%U[sc%>-xZhiYxytBwZXv˽VT#̄ 3M>8jkWxN-BxNŴpƮ,H@Ti#}^\趺g&xS(@x䏭|ZƵf^#бN?$_v/ =ܽO85ۖ8/JMλk$a@$Fuq՟O5#47,kO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,h_?OƀO5#47,jvm"MVKY$F,up:p8@]YgvTW~/|S ,CIĺ۬Coa7EcݤEu+Gj$"{!WX#pG[*sWٻG|Wu{JZ}J[Y[(5k}?Uoٞ97 rCd(+?J*'YOFhi [#((((((((((((((((
my\ ( {ƻr~eSњG!V ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( (:?E}ep%BNi+?y IoMPdXP w1<5wWǏ֝=KZpZ^<7۬ܙ7ˍqՏ⶘y&,5{[0Cq= ʎ| xѶJ0
@xTcUA|E]fsJ/pWn[O̿*z2b40o'_AVa?Ƹ}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_mWh}[>a?ƀ_: /m6Sހ>8?!S
^#wOr~||/oI᲏ދk%W q 3<?3iw5?_ωĖ:e}CZԭFdltEb&ݺ~Ͽ5/YnxoP >t  ͎ i#6C2
|pn7vc#㗀<1 #N a-xUQQۖ8s/J7 Yi)$xQCfbG$I?0lq g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
g g
gMc}ᯄ+I9CVF2A@Ysg
? (h+0ب\k-{qf_=*'nТ ( ( ( ( ( ( ( ( ( ( ( ( ( ( (  S)[PѶ_r_*z+'_0WHa_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WG~1o_}eCE} >'
_0WU5{oH7:k76I g
zPVXKM@f:ڍe:6g>K&Kǭˤ> ḧzɍ{EƋjޓg="H_/~V`Q?n[O̿*z3Hv_٪

uތCIA'9L(#(((((((((((((((* -Om@Fǜ袀 (h+0ب\k-{qf_=e^iw4c8J ?JaEqEP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
Tu
j?j6<KO@
#@Vo[:N+> Ӭ<eQ_:)mHLG_xX~<n$FMi"L:}̫ukF %U&ݹ5wjD&KER2 @2k䟍<ѼKf;pm?ԴxI` SZ~'ֿZ~'ֿxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkhxkj7@/E񂴶 2xGXA(@,jG<@@,Wϊ-t-B-7WVv:tW8GuNs6tk%xc6~+%o*G! &G4O)NǩVQDY#utpYNA`bqܷřTf#myя(0?)vQ@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q׿ה-//=P~/3Qށᯅ#Fi|@)ccosslyΎH<\q|Dg3ESEeG=扮'NJۘlm;f̒@N[="Io)**3_#|k5,D$=o GY5 5ۖ8s/JM,4ƢKy;iG@_Vv&i>&]5@JD7+ ~x+)@eN@>_PmK6#˰d
q_%|osf/J|+;)_L;sdzsJ9OCR}ex4 #n<e1">_wª3Y↣7f'`2Pi8kl}5 y`DeO4N-Ѕ?*WGſ']*=0B]Bw_[^®tk|[ Uo!xO Uo/ Wuʺ?-Ѕ?*WGſ']*B]Bw_[^®tk|[ Uo!xO Uo/ Wuʺ?-Ѕ?*WGſ']*B]Bw_[^®tk|[ Uo!xO Uo/ Wuʺ?-Ѕ?*WGſ']*B]Bw_[^®tk|[ Uo!xO Uo/ Wuʺ?-Ѕ?*WGſ']*B]Bw_[^®tk|[ Uo!xO Uo/ Wuʺ?-Ѕ?*WGſ']*B]Bw_[^®tk|[ Uo!xO Uo/ Wuʺ?-Ѕ?*WGſ']*B]Bw_[^®tk|[ Uo!xO Uo/ Wuʺ?-Ѕ?*WGſ']*B]Bw_[^®tk|[ Uo!xO UM_UEmf:xSaB4 dh,} %Bռ?kbޏg~t #s k hE0n> YYx_;|&/|%hߋJ=t!f't#I;'fȘZ<}%x6Ð7l/4>Yoo ƢfQlQ&eɏ-`}/k}mwaY Lr}='۰[<`s)z";kZxRn y!cp<Oă]RX$ vn%%:sһK?x:KH$VK۩ xrJ\j?yD+z(Cd#_uo|=[o.MCkm
?u m?7G@&5:{]#_to|=[o.MCkm
?u m?7G@&5:{]#_to|=[o.MCkm
?u m?7G@&5:{]#_to|=[o.MCkm
?u m?7G@&5:{]#_to|=[o.MCkm
?u m?7G@&5:{]#_to|=[o.MCkm
?u m?7G@&5:{]#_to|=[o.MCkm
?u m?7G@&5:{]#_to|=[o.MCkm
?u mx躄QxAw{IUu;vfcI@Rǜ袀>pEG7ƭO⬺}y"3Bӛ
o,\H
'9v@Uډ;DQU#`Wʟ9#J4,Us{HoH?uGQ]o?S2tpi@!E;5#4{U r92=#Hi4h[I1Zd6ѕe0g?ď!|rjw?`"Y<Cc!xt+oCpu׃4Wa ,M͜<~JjS4 GtqfCVK2A Muy|%O hL=FnJkP8w_gF+J/Х"D_ROyUGSiCR7hn?|ֹzGL4e- K_~O ]<]s/;H驕cM&qe|G_ć?4'Z͇<1Uf4CqK]ۿov;Q'>jKu(gmeIavh.YR ?*/kKq\Qk]E/&|Ni&.;qpBlgMĚKHe$\p?ZRPol_QVsK堭Qlny
Cm2x֭OkEŪvr=ҖoӚ l#Wtܛו=iL%eҕ}Qm .fe=#%I=PUKFus$[kSroTP#;-bg'L_Y(;/sʶqk9|ennRlt0g/n?ZxKSqw Eo%
j&&soMcQJY8J-Y7IGcӔ"O#MVC#zIK#>˪k[~)K1_r,G 2K  Kt#U}fӋapV]s_Z¦:_K MCJ}mv4ɷzӞn^~a^zW?+ts.=zѽ8ϰ]w ܼu.A8~tY Š((((((
my\ ($j{?  ]"PPGI!m$rycAyr
>c=mAjp0y~^%Ynj" _w}g c+JUeFj7FuiF9SJ \8mj; n
,m !»# o W95M\k=A~t8XCIN@GZֿKmܿԨ}s~f~/OƣiPl E}nagqXZG_
~2U׮#X[Kfls"2žH"a^MN7Dr!xgaLvGb3X3ct2kgrٛXb_~5:UaуIxΒE}n4#~I 0}On`Vkh(ѾxA}j?/4))hi֗_`сǿ!V8N
'# onc 5mQ7Ӽ
|e bHG t`[wRV$+1i 4tf`
.#?wR 5R'>ulSseX|D^i ~m&2~I*@1N7w5IIͨ Z%A[0U>V&?PN|vߌWg%𐁦  Kcߪ͸;;lj+/~3Q%
4.BE_.ӃkmN_6;/66Mʡc *vo6ZH6i'hDs ~l%Е^OFVZվ"x4!>ȑ);72F>lwȤ?Ao'ݣFM7H{_'/Neck(wJ?E<.fܿ6 fhNiVrY 3ُw0Á֪Y~5O e> zɰ/H~ÒvwgZww[e`g*Ħa@@ޒ׍N_P[߯{?g|!us#i67HDyPwT
||<5>!xV g femv$vqYmY/!Jn߭5|[&8m;IK*_M` Z'nQC.ҵ+e}%T ]4J%k?//oiy5į<~!U7Ho3M$1A'xVoxݼcoiCE΀~w9 XpJ[?ɯT^i%ռ9g'O ]~Z?GŴ7I0sF?_#>&xO D1ٺAƔLIT7NF ckMYoO$$8*0|;hl|j|m(gMɿM~\zg<Օ|5 =bUi86Ċ~%䝜c9{w{w塯V3k^z|u(͵;]}k_  rOIqcP-/-xwM?l,u{ Rqh[Q}2M;8sf=:? k1Pjv m2 _#l@'92o81} ]__țY-KޛWƳnIͦh%lX0j#^Fva%{5EnGmذѷ[ vI#8F"j?ΖZFѧi>9 `$ߜg87}ޟW?O;q#ߍHm\71 "F9|bVk[}cAQդ]7GCi~$*MؤK?r~N_rA{k]ﵿ=JŮÐg &cuK,xK]MG97-em?63!*iey5){/OZ?Tkk-SÏS.Uۆ6"9?o_eƍO/Ƌ_iJçX~Щ}cy$d ;Z/ϸ{%{v.]|ou+{7Cq9Ӵroky?zFqC @Snquep@ƇyFNv$n{^m}u(Y"7|?Kj4Jiz@[{MS9F&I6q}[-;_6|9d9:դdsށ0U<9时kK/מtSwyhX[[𭄟|n,:X;G71faN1_<3b_#5}el5[ai ^lfT69HH@}_p+Y~ |Gf⟍>xx7GoUc]g<Q:oxoJ?_`w{wM}u3!EFK]>y>&xeurZ@it
ݼϺ|̅.K"?G:v!W"~˴椪ey[nn-1- #֭VFEP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
\7^9xXhzZOtZ3y 9$a\ Wp/ieżڌOټcc-܅ !(; }ڀ<Cᗆ4t 2>&m<9A}9ums $T(,B/¹џimKO^ᬮﴭ{m;͗K32y4cmP-kR{[SvO<YFK7
4
/bM#uZ0oZ7 7qJ[^dCLᖾ k^xB{!ڄFv;זߴ {X3@Ǐ55 8oB͵pOiI=է\B..-Uib̀H(KGV᫏ZƋ>$<˛MD*Qc @@#v^"o KTbm-mݭU6?vI!x1>jt~:N SE84 ^XDcθiG)%d]=X0 FAꖷuVo*Og|gj ,A@QGگ5xLJ9unu 'P4HI"-pG9@5oN։} Zo*N񝨊Y^y_g>2Ӵ/j׈ߊOv6nid#5ؐMs? 8GV]7^4{kע,lHxMN:~#`O¨n|cy4w1Ms^lv"C%PmR."hCFSІA%-~/_mi1Ou_æE}{t18&Y!ө #k^/wÿ|Csi<:k%pl*Afh]+ᮙ7|# Cԑ"mu"irrA)@G8t/i|20>c-"+>`)F:TUsXkY閲\2F܁O
 imc~+|K7EAuxuԺ[/=mEEs t`
##> r K@ſgOog~a u6F`W{ E b
^qoYz?/,̾ZY6dYcη Llr2Z{7}F]c Q݈\ >?&4x7I嶖lSדP-呢}CFY"Pw l~ԼOy-k%gk%jYlPI@ůz_ěM:-5>=oyG™pJ53xR"L[/>GX᷷dIUTw<2F7¯麶=DXnq H9IcfF0Qu?z-Zci5ԁK$.Rq@Ÿƞ ύ436zs[ K35,QQ.p>\esk( {:{柤i;h"^IU rkw55i\xWHۦXn^8&/1m
qPZUt ꖞ4t%o4d+I3$?UXk"mt3O..Xa5fwbh߉?<-7%uXtZXSZ`a]4急kƛ jr{WX<25w}"]*:-*,L  eYXFY{_-k~/x$x~ ҵxN-SX(< [U6am'eBv
|%$z}'׺ֳ=뗖~'txCyTYyVEWTLb^7PúGY`r0vb aA ԭ,<i]Nωjx$h4V?* F1Ad\Ec'c |$:/_,ǀ?s_k6Fbq;lD9$N׌):߀-/jvxo-|yX4\02F+|=mĺ$wAh[D; BI*C͇CFzO~ Rբƺj| ޴-So
x-z;ۤдյX騠2 2Q* *}zsh:4WVaUUa 1` 5&k6'j
:A^ԾxxCM-kC[[ܵHcXh8'K9o~~*Ş'/QRMF܈ eXnXOxt6'm'OY,ۉFƊ *NW%[?GfG7`V/ F\ ZwƍCOXLF.)J,.H@?}x2[N;X Ǡ'4
lj> |27'og\#hZqh}gʻiUt`ym0I;; ׍alW'ht RIP΋+7#Gp/?o<.c K5Ą奕p $|ou / u{KtD %w 2J!ٱ
7t=G6k jx2Tkk C`ki"1<Щ91)|iu鶷mwxj;FI3:`}#:ߏ~1ğPZ/ Vn5xef6t4˦ghHฌG 6GѾ_<9'ĭI}WFQų >\_&M'c7./ŚiE]]2EU $ Wگo,>&|Woos$yZM "8Df`fIoh$K!W>`l͜Sƞ2f-j=>$ [,R(;Wo
EmKޛiFIkwֲⴇΛ͑
w
fg-߈#^w+-Z?5'FI|4 _^6Gmm E H8B(
v|.ŗp]ko$hٛ+(ݺ|igO0W  Uf-Wↁj*g
g⏅> Xk9:]Y)7qf[c`RN1|)
fxᆬ eZoWT-\ۉ#" m*l&@IJ-xm ='_F ߳n-gm7Uy<n_&Pf #HXp> ~ ~[xP57.59>m(N}N],rI2#m2ƪ͸7?Q^w_~iW m /#>>}G-?:߈^,yohE>](!{4LcK#>d
;H¯^o |F#߆Gz-4kk#j%1h7*}'EqC7':Lf=㴔GmǏDaH>5mw\֭-;vdiA4aKnp9^
<-|QxZV[=Ƣr̬EzBKqC_O¶[xY^D;/apSay'h tEB"U =|xZ/;xݼ?9;__'YEyrȾK[pC m%>3t?X nVګ#o%RXq=<#c?zOxD"մeMP0?`|'xO<# LjKkZk 2%rg>( 8v<;3?OiãZo^Q(7`r >UNştӮ/]x.l yީ:Ve>0oAIQDDPp/#/©MuA:KF. /pgeh +W#.?g4߇:Χh]2h(>Pd
 PGqFD*4~>_E)ŷ4']o×;w
h;ѰD4
|&ڟ5cJ#ơ$Z伊&Dt2@+ԟ(幓R
0[c沩a
|Uƻ_ VwΠm6!InŸ`Jvb'I߀6W)k xa嶟OV}
&r"H'hf ]@|XND?(H#{|$`W_y?xç0im[bcKL RM4!HC2
S*ŏh"|4wӥ\xZgBTqdz۬BeFA$g n_qP>ſBw&<ͯ7H oxOxAC4Y Bvj3ু|UoF_m|$qĻR[ˆ^t@Bąbp
~;F|Amhinj̒?kd4QG (<_)S~ ^7iN >,3Mɨ9 XGg /hZzծln/nlnF6u-#1b7 רTW^;:Xuar'\)'C4 ~*xa4 \zU^MF ~e> 0-R?<xoVqڦ>]!}JfI;1$zi|<52<QDH:"#|_=VUf#6.bWPҞ(r/崈
==Ὕ{ho4ڳ1!f`
Sz/_|k|ee*vb@Dc=K0 I xLχ[Ś?(knϦ|/^]ZrZLQ5IY\#8އ):aoXhwa1\XZR#2~4ǁ'yxź]RbԈI6a~/lonl׼% hwV$#7-Y<LF
 ?g !3xo-rφ뛽WbKkM;]\Fw䪾Mztk7z[NmVh?[쑡v
%d8.K(eb S< Kj:.XZŮlghس)-V_-bɌ?x_?4- #z5 hTu;fb{9 Ie`K}EP
\7>ᨼ5 K+4ۻkӯXCuo"h
r? ~xY⬶xË>YͪxI* ZH`?[^+NdEHQ6Ҫp,BI<7~;O3?SiwXLs\YHhdfݍK~fiSDϥxvΡkUQUifb@$VKS)f
ZZA0Bqơ P8(fUsC*e 0AP3o|W iq%imk%pvG>]zѸYu;*SGocM;\*' < ?>#j!ёWLˁ-(5YcL6z1E10ƱDPP8
P xE_5ZU\˂cQ9'[:ozڅ
q/fD$R13哆$PU+᥶Ikmg_D/-8tsP @[$E~d;[ 
TԵs#(|Cڤv [XrQhP)h {k{`d(a$
0-%8P?W3Zuۓ[sr7T (Ǡֵ>3T!QOvZ7i`0mr2FQb295@yM:pi<#5/ -ղu]$1&:K>6-v/SO_jNF^F9=s_!? u/Ox+\-ܾRd=SN ͑G#j$$y9
>
?t_[RxZZX51x"}dX+Ҽm <Ğ*ZKh$w8Ha%i%8DRN+&oɩD&/۷7w{C,:2yCLY r }@"<;ec.sL9oȖCuvѼQK>@:O;|ӧxOԮDKq!v ;8<_GWxIq z|ځ % K lR3,R$2
߃bjW~l~ 叅u T%&Y7]ݑ?v9,Et~0?kT-jI\zZK*Z$QF>'xWËZ LxE>.-JN?6ď" >2Dl }@ZT7uX/ 6H`"D!RU˂Y=[Fx;Vږ`Ay kYoYKzq k 1FUF\"X|E1,g{y -*G7]ʬ0EutP
T]cIЬQֵ;K X{Ԟ~
GX߉m]] wiq.ʉCHx&f+ 75=S?/vںN-ͯIw<)SIFH`U @Q@Q@Q@Q@Q5?Ú5"Yi]\o%'eKw2l /'ÐxRFZKyifIl줧P
~۾:xƟ 3h WJik<[iX.BE^8dntfKw}_KrLܥ% @͐wh ('wm-OBW1%[ûtp`P9/_5/ x=7~Lt1mk٣<'>[V~57Ga\%'wX v &<\xz--䲇VA4ʌ+ 7# ib׼,¾iݠZSrgY27<R,
7_~:n6z&i˦oflռ6[ۻ];!|קxR𮣣>0o jZ mcC#0c>2

t~%\pq3ර?Xx'UDt^[5RʍLda7O n_6Gd{ /wsAܿ͋P|$cK4w@Bo gh'Y#}>-eIepjX2P2 qB۶1^#@|Ke|S붾--.专\5y1@" V?s/h?jg!?8e/4Pr. IH҇! q jUQpP
EP
Ey/_ZxzjvĶWCQW-.
8f6qi~!d&3Cj/)PX[hg =9k@t˻A^:f r-J>Q[|WS9QӼmmP}gԭR#(\42um·2 `` ( ;wK_x}㯆 {6ORDڥKρ߾J֍7Z|J|!w28=4̪T.#{-P
E# T21u/zo]K?wSǣxk:s\1ݦk67,XANU

c-+i<+㻭:$ͤ}nO5"lAC
H
gg=vܴ4 $ 
n8'Aa,3KLJuo_xV$]L!]Ol #3u狾<3<h= bi"fi9dW`h$p
|C W/,$4E:O DQX7m2Grx5k?7u->o Oծ<Ԅ{f[ch7F%U¸b
￳xGPk>eƼMcOdhBK d$N^,1hX)rp(࿎Zßx] ݷOKI|?dSc H&x]Df;ϛL\k]hm =b%"9%Ng۴?oobt=?Im>%E{V}͟3 ( ~47>kU:d1%X.a'3Ĉ%2~w*/0w hvsM/Rӕ ЩzgX%ˉ$F_{{x- jQ@UPP
EP
EP s@OϡG4ַ$H
nF U`
ڀ(( Z((((|G/ ^x-.'X#9U SЎEW7,UZ 8Y ̪HwaO|LO~YQ$S6v9veV-}>iRnt Cx[_ԴSK$$Cu8l 15*@ 8P@Q@Cuik} "Uru%He8|'>#Xx+,ƥaXRWVt#$r\=r2dLs#y W  |e}?z3M3_G =ÝFo]\o<$Is0*|w :wƻx-"/i៵[-,F ,!!$l((((((((((((*iZu:}3p3prFӾh-?W<~)½tMW.% ulw`~gGA|= hxv]&)eQ24}DT9>\k1ݚT`Q@Q@Q@W KUkaN#}X@h"YѼHW h2/:ς7X/F`J3*#&((( endstream endobj 71 0 obj(http://www.shirky.com/writings/ontology_images/del_user.jpg) endobj 72 0 obj(R8":ݓXzҳ0) endobj 73 0 obj<> endobj 74 0 obj<> endobj 75 0 obj<>stream ExifMM*
z







(

1
2
<

File created with CoreGraphicsH
H
QuickTime 6.5.22005:04:19 22:07:25Mac OS X 10.3.9 AppleMark ۄ       
  



















 

}
!1AQa"q2#BR$3br %&'()*456789:CDEFGHIJSTUVWXYZcdefghijstuvwxyz

w
!1AQaq"2B #3Rbr $4%&'()*56789:CDEFGHIJSTUVWXYZcdefghijstuvwxyz
 
!


?S ( ( ( (
F~h|O~-Y`!a},s#(pp0h{kdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?Hibeٓokdsa(Z?!?ᖿfO7?HibeٓoP ( ( ( (0߀Dw<5@Y 5֭d3S]?z.#K'%ai&uA*)Ub!A@]C֚7qCŷLeWP<3vIq.3j迴!ThxD?j?P
|G?!ThxD?j?P
|G?!ThxD?j?P
|G?!ThxD?j?P
|G?!ThxD?j?P
|G?z ((((+1R f*[,I@/ڿ㟉/ ir5ʟު/f6d"w,n3nlto_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@O'ltq_-/O@>Zx@Ѵҭ隼ښ:y嵷(hV#{P
EP
EP
EP
E!.ӇRk] Z7<8e8? <jjb᥎T,w8_EB7ը⹔OZΩSR|;tGwE|aEP
EP
EP
EP
EP
EP
EP
EP
EP
Et ?{ k`((((?s~
pz}y*6 j|7Ij~4Үn7b͒D]C 2Gx\4@ZD(@ck~~3lFK62Io,ͰM1SH#@C^&g?G%Լfv1ijY
NQ f[{Ž߈o%p;igȊX1GE gQ}h(5'Ӈ/E
gQ}h(4YA_?}|, /~f__ ??w3@/E
gQ}h(4YA_?}|, /~f__ ??w3@/E
gQ}h(4YA_?}|, /~f__ ??w3@/E
gQ}h(4YA_?}|, /~f__ ??w3@/E
gQ}h(4YA_?}|, /z_Ox:熯eKVkr9q^@Q@Q@Q@Q@yb',>xONfk{[uxo}Bxe2C2I1MT?P>[TH䳞[IED$+D8~B3|;š4 5V xC3mAK1=KI$I4
~u\xVT̾"qھ"*b|#G)M=_4yW|M=_4yWtG7
M=_4yWtG7
M=_4yWtG7
M=_4yWtG7
M=_4yWtG7
M=_4yWtG7
M=_4yWtG7
M=_4yWtG7
M=_4yWtG7
M=_4yWt ǚv-ש~ ( ( ( ( OT
ľ$𥾕~<.>Nk 9h;iO]y"֭u|:tvH{-\d Nq,m/i>jZޣ6d:3$yꦀ>RnmŎ6-@W#+KU_K^!* ߖ`;{aE[_O_#~e?>e?____________________________________________________________________________M`*8=%X~J:((((+!ѯ֙H'#sI* dq^º煾^ȱԒ1VAdeI%DK" q@&Coi~*>Um +2UtxtO\R,o`r߄WaE[_O_#[ nJ_[uoZQ/- -([ nJ?
EVEB۫Ҁ@~!mտQiG ?(_[uoZQ/- -([ nJ?
EVEB۫Ҁ@~!mտQiG ?(_[uoZQ/- -([ nJ?
EVEB۫Ҁ@~!mտQiG ?(_[uoZQ/- -([ nJ?
EVEB۫Ҁ@~!mտQiG ?(_[uoZQ/- -([ nJ?
EVEB۫Ҁ='|uZ:|A$U4[Μl>95TQ@Q@Q@Q@WxNo?"xNO (cF>5iuS_1E\WaE[_O_#.5 Wn3qSFA{>F92MahB((((((((('z|Q@Q@Q@Q@WxNo?"xNOw>y|=6L]=ռ"Bw|៉|64_um4W oe=+~yԏs?kܾ]'#Z{5$$X7>~€>OĞ/|YB WơqFy,ffiDHU9PW9 4[]uyt/ \벴o+Ǭ k"/'b?}G/kYZ\Ю1^ͺujxvr[ 5=^9YR5M>%# ,f*]+/jm#+hG=_:W$̟N|H'J?6DOWΕ4m#+hG=_:W$
ď"ztIoD_@&?9ҿ&M$sM|H'J?6DOWΕ4m#+hG=_:W$
ď"ztIoD_@&?9ҿ&M$sM|H'J?6DOWΕ4m#+hG=_:W$
ď"ztIoD_@&?9ҿ&M$sM|H'J?6DOWΕ4m#+hG=_:W$
ď"ztIoD_@&?9ҿ&M$sM|H'J?6DOWΕ4m#+hG=_:W$~Ϻx^VWv&U;Gg {P
EP
EP
EP
E!.Ӈ鬪Q2d@|>tBZxIj-c\Lm'۰ )ooEJ*jXMZ]TWQ7UV1>H q_[
+wf*]+/tW̟NP
EP
EP
EP
EP
EP
EP
EP
EP
Et ?{ k`((((?s~
pz}dx> MdkM}Bue"4
z'yxjñ/5=nHKVy4%֙%dgij3H=kK:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)(K:/)+%[ox^WoC]wy_`w.ݿ6vZލxF5{qqc[KisHMp'dچ%^Zߦ-ˎ#:޵[G9c][ |čYfbOrIu;+(Hx -%TD 3a[Fy&v,+ȣ4V_Sn8?+E!?ʦ:nf>'WΔYŦ}*Ө$ 7JaӬ?—/LeO4?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ ?Z߄ [xX[{xG$F
>|
_O_7^@HQ@֚$uEwXAހ]TXm 袀 ( ( OT
$Y]Qfcu$
|t-wR_<=P۟2ē*ħ)Ep1u. ĺ$%~dek. VV+R;E|2|]bo"/'b?}G/cDŽe mxh2l+]u8f Ֆ) 0;oyJ2.p}y>g\ɱ-w+x yi$@$NZZ8imܑ^\rL+o=9fR4rQ_ZAymN6J<THѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?ѽ?to Ěקw<_iKb6+him)_eֽ^E-cWAo4"FD)) د?|-6O4SK MN#EzʹK8*F`r_o \|{8)izWVjoswHC\Z!Ku
υn4MjQ$=>Lede(mlɳ ;77Iwoo@, ?&YM#/?g|F_q4 h}$
;77Iwoo@, ?&5$ >Ojݻ|;
gxAӼQxwWIQXh H+<t/υ|U}mz(>-kĶ+8<~dd?2tW_x7^ {Hb\agy'|/9
|5G|WBFOZ?&.#iW ]NO5&/Y:hʩHǪ05V1>H_A|;;fA)HeSb{zԞWv}WVt92מ)+]u1-B3@$wɸx?ok<+H?hgrF *G#~f9z/͞~_"\OO<˻y**c<qU=#NҦѬ-Vpf(͵=2I=jrb̬E?XF֏cmustye3ڮO&*(*Gw?r(`49G()?G,M?.Q ?g iw?rQ ?SO YB]Цs(49G()?G,M?.Q ?g iw?rQ ?SO YB]Цs(49G()?G,M?.Q ?g iw?rQ ?SO YB]Цs(49G()?G,M?.Q ?g iw?rQ ?SO YB]Цs(go2׭<5 (ntIeQ,na48y2~؍ (s'ؑ~@Q@Q@Q@S/߅D?<=@IS]4i$`bpI khX޹}_8jrG4k,R+rP|?j꧂bȊX1Ee?Z/e*yAe5]N5w>;4?z?/.'Hmo롩?e2?`cv
ъ k|}&mƼ 8K)MS=,uI7E=6oOlwOd%jZj; }k,V25y_k՞m^sVŶ g3Z CV=3] 4dz((((((G^@e/ɴ%A|
xG67~/ c?034Sȱ*Ig`:` <xG>AEq= ,P.\EQNT*⏂#xOR6Apqcsfdq42'&Sr6851WO
iL:N*ٵW)nk 5/NgI_vֻ4+E[O82Z1u*ŠP((No"xN?2<:,z;̖0k!A</U'u:żx߇|H6 k;#A!f]7e?{>jA9XdGO`d|P^=Wu&o;_47aJ =Ki&ux5eѦ+
"*b|#G)iBxuk#~=i>"Mq|GO3>-m9O3^,F멑y ,oŵ`IJlvҷmZuO_ǬڒۮVzW 6yq>E=Fƣyw'r4AWtmAu $U}JS{mGSw˕Byk,/Yh[a#ƚ{/{d,>k.-vQW&N'kXq\\8}?s3KZWʼn;12Ei>t?o|@I?pt{pM,gpp;+[~܌A(̟sۥli4u˨ٕVkUmٳ7 Wfg4֖,M`q=o/ɔt}J9Җ˹b_|Wog_e]|bu]̏sQ綝ͽ\/nE'_E~/_ ?&WhWhWhWhWhWhWhWhWhWhWhWhWhWhWhWhWhWhWhWh?g7oS&}*ѢlbΜ
'/M/M?__E5xtShiy #X0y?~w]7^=m\j%m&Ė/jrĻ `nrY ? U[Rn&֓+Dd7n&,\lq ,~!ԗ]R|5!.ڈ(by#`8X6v1~9'6&!5%H܆k5խ Mt_~!mEzyjy?"_?)*=_[kG*=_[k^1*=_[kG*=_[k@*=_[kG*=_[k@*=_[kG*=_[k@*=_[kWF¿߇Ǭx^&mnSٷP1Gn7as }]EPǚ|?j꧂bȊX1Ee?Z/e*yAe5]N5w>;4?z?/.'Hmo롩?e2?`c#ι ~WD_;?7?K"\?<_ij+\һs/,pߓ9ߏ+q?OmFu?A_7o˪b3 2cFX3߇ibO5EO_#^]A]4w d M}ti;klҔ9J+&YUY_3n_x_^ᶸils3^eynRD$)y^ױ\ۦK_{sKڥ֥?ڛsBw?Z;XZk g IQ#TG[3ԲjN?.
85W%\/ʙ99rG
́xmkrx\Ntw2p Z?iJ ʶ5?ֆ>"u!7ȯi6>R›ۅ?ޥ8>wШp멮E}B{YD'^MwIsm%k(mO\Q z8AGu'N9l@1Agiڃvwi+(Ңj*яb牥NhV]溶 ONM4!Q^ 4 .*6ܽΗ/z|
_O_7^@Wx/K: hz?!*z(8EԾS += ( _9 ؉8xz></5/ nZ5ڑ㕉VFfT1H {<7<*cDTj"8/8whz_x"'ZVȗIFb_5NFv–#Coy}\>6&Mrn34x3㟜$\o!-#s?.?fOIŜ-%po3đ dpEuz5 Xoekm{ko/"
Oj mn-:E{ɺvU8Gy*-rk߯/ 'G7~xB+wI)Q] Sy%Xp@"|Ru冃|oۚ< &Rߪ롒`(._*N{I!EOV4"BD/+Q ~{7X֟+V}L#Ms'~ݾߠn:|sk)b׬I!g1q֨xWvsM[1ir򹰙*J$oӋ˪ūS߿B?k'^p{X*~Sy=j->⟀ ؽOt7Nڧ(x53 saI'.U,39\[ϧ ռ'{$H iʻg|Pԭ?Z|sjz9UJiMM#,Q3,@ɭ=nUҾMCHRUl5(ivϖֶ?eJB=R  j
|[qk$Iuq-.^Tw)e>
xr:XJ8ye$y{mmb!ZݯKmzsi_a/4T,
70*|Ty$ ĴX,աe>lҬQ ໨l+8q˥r`jK12Fw-kݪjs^BmKGS7Oe{2 'Xgy$r332S= [WS95j~;PӼ' pf۩i#T^R 4Bo-IR(e%M8=WE\ބ8cR|[ksOZ\Mfۧf}Ss-S^b^21 SϔmO
ԊnOڍ?|N+]ږʜmGvwRMlF-^[M ^U=oɪk%Q $L9eOnk|hԭ?
|RvQj:W60G
[ $t#ԭQŮX+>vZ|[֝x@ ⶱHrt;MGÃ*Γ|U/< |ޡl}h%م!NFvF#N )Aױx| J854Ndmr0:2r5=#8?Z2j⏁-]:[[5ﴙ[7B0/2
dp
5{Xy=Ɵ3Ű@ڇS8eOLEZmӖƵ0ϨSҨHz/xOރ ʥ䖲5ʆGAT1+*fݏq]%s̿e6#h?o6?>hݷu;Kʹm+a;9eVNVbp8aׂshZdžtе&Ty/&K/XaBbAd%F@%Dko°m=a N3|a˝MoQvd>/xG χE3MͱmĿj[H @t%1n^9_'3^&׆unWW6?v[ + ;@-τU _* †
s/CcjxK_߇^}%u/<چ;[vt4
E
WRfZ\oxDl<ԶvZZ@ 5 `
ڀEv5S#or768X'hHR2L1FeYjdE*@8$v(
+jb`d$ !]X0e
}GGcP`3
`<h
#8XED@UF$x՚2J2TN bEe8LqS?*Y#h)Q]aA`Zv8ϥ8Fpz*
Q$QK̍_c\#ހLXGyV5&70:dtIQ#dzJU@`
ֆ'%x՞<b
+="8 *}OXvM6dd(ĦaB$ 323@EPF 0 "RHm7 lpOO,1$*F'9]J:0AP"$HƁQUP01DeYjdU(B (
+jY$?RI'
,QL&]r 22 S
LBS?aP7m g$̿e6#h?o69Ol%1x{Q׵*;k\E#o2OLmJx8྘>#/yyi ͔Ki-G/;~7@wع>q6jeiֶg٭"ڋw|ov~=Sal\xK1rjyl㯽m4=Wc qף[Xg'0㧸@ѿczYjy?"_?).@{(ѿcxغ7
?F=Gbh]? P

?to4@.@{*VV߆>g }Tjx穠(3՞I/rI3OqɗQq=!5|(xOPuO_xGK$B.bO]ѾQ `"-oK/
sg7c`~ 6*)
f|~ zyV&g,9ilPFug,Q*^ ~=F~-{_XҬtOA( g4
'g>MS=GºtAHIQi0op7uT$5|Rѵ|^+riӖ;QP9 ė~/^-CNOSyi"0\Rhy!dW<7'>}rCT!-U%?,02Fu}DSִ旕k*=:q%h~ i|^#ԠwnѤ;ljX:3xuƾGFflBƥF0۷,[5
G^(|ç7 $8@@Oz='7+kIބ{iJ-MdܗޯWFc(@P__X閒WZ[By8}Yր<|YˍKt{s ?隅澴II1
dpFh
k$|^IN]ych,+2G5 o2];.Qx'UћJJgЕSdoPWƺ4y]WX[{--+c-+F뵕y;9>F/h됓88. 6 Iž;-`-4nd{fztrXExw2|A~%Xh,Ěń֯i0錦Fe$
~X!PuG8Ux o.N[_Y7Ja$"6Fh&V$KnXKqA @(Mn,5 ?QK-ޡYJ夅VM++oOčcG]WEKT,wEQ$Ir|hBLRoBIZ(<_iKb6+h ؃v[aU?GW?r^^#/%gȺJaExP
EP
T '7oѴ'h6nkg>NKo@!W7w-E._*#Tgyf|UVbTr<[#G^kMkVo;25+(.1bꁶr&*~4=Nt|[(R;Q*Nh
cٛ
;"B4`
?3(͗v'<_+2Ҭw!iU~*m >AOm/PhaF[;Ip]rmP(J4$a
۳*)f 2I(|fψF>Y[,ObYB\R^f1WnA;?|3ۯIt:M.[so;?
,@=&YKk}>VI %#“H)hܣ)8P[m|${7>"xGNnͅixHͣ|3ym`n^Ӂ@|a?5(xL'X\ZHy?<ڝmyk|ns}h->w:/9F
lR4\{1ϭfxGInY FdG' bx>vXq%[+˧̸euqAx{׀|>#tkuV!Z@[
3NHPp3tWk3Bq('|_4z75X5;IoRI,d u%DI5`["Hy{P({/:,u6.m!)r` K,cs]x/z:?xyT ,G1NNWch|mâ>)L{kyKIӮ"G7 g$#8`@ߋ'<&+w .-R"L s߄FSeu MBǑ.lim gv~tm/úUXgaeoKcE >+|M>h^s} ,an\7"vAG .|GxS\X.qmkx4hd_.H\: Y(<-žм?iQ\
$ʈݎ(kni異v8' 4
į(BjVy*ť_Awz!%qp*V9V8 ? ~h^n4[OunNslSΕR $[
Uv:
u^7VWOl;?KՙwCotCw.m)*0|!E7Vgm>}=eyLW]kkY Vǽk;=>#V񳴥b@I(2~؍ O?xM xI#2[@l;L%f69mx= ŷ֖Ockwіw є= ?5MzK6𮅨j=>C|*'6kOCm[-+_}m4=Wc sZV=3XOxSvSF+t]7%?芏*(8EԾSK*UY]SªCEGV_E|T*?(QUG%?芏*?.DTUeQ wO" /"K*UYWVMo?xĞGMahըٷ ׮:zgjv2ٞ"I|0c,@ξ ԭYh|GjVdV[jRJe+=I_1sD4D^?⧎[vB'DH-.V0ʖ/bIQrѴ]#Ú]~h\,QD03_t i:ޭpڮf`$@d9
|#ǟM/R|>$IT3xBa+*'o F]ec%z׼=3zlֶW7Ws<3` V嶝g5(-idsT Pk3վ2_i~5iZٱӯĬyy.`N u+M޼{9-[mhIg~i6 ikjo}'OI a3 y1P |+tîK{ۍ=~ţQ,!LeC#&~$ׯdP%s(f@}#L״˝Y6x&@"=Ef74? XkrYyB/,Ke8t
&k=Xе]\$ K |/^+{uuKAht n KdF
$56𫸫C`{r?zkZռl&{{!K*>,$EE)VM_J|Q=^DK[N&
eˁ.`BuY (<_iKb6+hq
Wx/K: hz?!*z(8EԾS += ( _9 ؉8xz> (
<3FUՆCzU]F<=hkiDE*w3ƾ8>1'KM=[y.(Xe3*:'A< xX5kᯈSVmᴾHR)ž)A$r8 *RO*L#nx3Hk{;)5{`Vϸd +zI#6WTD:@<Ss|c<k'4) lÙcEӣlG.G6k[.Rm"W4)!Ô XkEP
EP
EG<[B$jY@9'W~ wǿbNe?K{k5s¶=P(?χ6iwi ]BH~B^Kuodo3q$Ut
~'_U}Y=폝xVW! -΃t2 T
if*qi5yw=[E\_%M-@EP
EPS^GG:^i:.Е ("+&RE*GR0AS0kIN]kXYiǨjrlmT<FX.v)jh^T(Ի9$W|Axմi/,r.E06lo f#"Mu_ Y;ԚW7Z}Й,L^T.iTHY"Q@Q@Q@kឳKhZ%rwĖ^1[ Z ?ƛVMZ(۴(v%b\
׍~_5
IoHV̊ĉˈQ%ے qN(Y5}Yj62✿98*1d ( (9뚆۵ׄ&ӯNQcBpx̑Hܨ[eoDu%j:R!n$& <mr EP
EP
EP
EZLӵXVS"W*Oʡ!`FAhR
EP
EP
^D_l5x*+M
{]b[h1xu1%wzZ?utsmrxj+kMI{yn#쮒H@,
V8"8
%Qh
P
EP
Ey&mMW
E?/ KN1~8~?ʽ,F5~_J<u/Šd((No"xN(Ȓc=h EסJl+8ك$>0o^Yڿh+, (Tf0Ҫb<*(((((((?? '.Ķ[o"[yRAh 8uj6_AZ((/M/M㋿
_]pڭ1]7|fXwgVᶢ)m@ qw_^=U;MNm uau}E'.3^3/<[aU?GWU_j=|~F,F5~_J<u/ʗyhmd<~F*_ߑʗyhm<~F*_ߑʗyj՗#(O}#EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP~m? Fzmۧk.rڵE]yzP,Vɭ<VHFhm
7|#ݵ}j{mMŃ[d"L(3KNዸاjE>ែ]?MԃJ᯵K"0~^Nv1~9/~d8ܨKl2!?o.yMzyjy?"_?)(߅.]S?F-r$׌{! 7o <&Q \4[B5?Io.yM~йwOh˿j@(߅.]S䚃Lc|?ޟƫh^E:70e0dQtk 2:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&?:~a&U7 }JY -?~ǏWs >1<3OizÿY B5¿ԏJ+vfݸٖ0sC@#ߴ߆nu/hh{ Υ ~W~Կo7:44=M]aR憏Gi+ ?_
7EwK?0sC@#ߴ߆nu/hh{ Υ ~W~Կo7:44=M]aR憏Gi+ ?_
7EwK?0sC@#ߴ߆nu/hh{ Υ ?_iKb6+h ؃v[aU?GW?r^^#/%gȺJaExP
EP
EP
EP
EP
EP
EP
EP
EP
^W<-i.l^|47QZ5vtF8Uފ71` UON+KzO44g~n6s4fߗ\+oWDGoi?MZ,cB1~93^n׊].#ppkgFK?_3]K0gF)7o&s__h?}|hgF)7o&__h?}|hgF)7o&__h?}|hgF)7o&__h?}|hgF)7o&__h?}|hgF)7o&__h?}|hgF)7o&__h?}|hgF)7o&__h?}|hgF)7o&__h?}|hgFmAxzN B[a$rϣ? 9o؍/8x6 ( ( ( ( (̿e6#h?o6 u_;WJInHNI(|Q3@| M-s ooLVo=Fy~d"L"lk~."|v / g;SƇq?+I쿺~Ҿs?f,F5~_J<u/W?/hUK}d>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/hUK}>s?f\Ŀ٠W?/k.>$gbCrҀ+N[ǟ#xKN#M((((/M/M+^ xM LW/ԡn~_+ZWgK.^NkSiK`!' _8+n}J4_UL"T2?^+,T^F3={&fqkwjLnP@jYEttHS%,tP[DY#:tzYzhw:-C
^Vn?Y6ѣ&2|޸rRN%5hZ뫎M<3N󽟠oĶ=T(G_"E3Bvժ^VqЭ^w2iгXF&>c1Z7ڮlK+Jez8Gޟ _݇1M,8da0*Nm/ITgdڝ+s-MMo<7PUqJXz>1JueB2֭y`ִg<9NsZPcYogXOsZKXp|w2;*R#>d0Ho7&8ҲyWxnʽ'˙Z[ykx-̸#\s 4S%EFSj}=.r96b?fT^nqxݟLW|kl(9ӷ:j@HT#N;w ?F++tcOkG_-_ $g%p<Wݮx⶷}GlKEӠH W{'B&MhZ$D*qq{36FNūYVWCxDO
8*a%gg7>9|AQKt[t4߰en<cXReZo-<:8t}}G-xZӦIc1hffH/;E'v_#?GKaN֚ciiT ^/gܩ~ڮi 㿎;]4H*t3& c&Og_4M'ǎc]- Mx>cox<c0rOV(Qa{-4 -*ugZ+ޝ~Z]]mw+ĈU4C[ Wv5<_ @ .qYCZjF2_ߡtjJi:?n};TM(E㏎9/.>b'ٳ/;ſ.:_W[4|-P؛}]n#F4|'>IlAB*߸]-(W!/}7ޛ7샥5߉n3wHseAa+kտ>;g 88Zy`dmGk,:Я){NK|e 3>ڻK]Ғ< mGnDXOr*sps|mһof .Rq[OSМiQҝuz[n-?ejWWǏfu Zǡ, WiAe쟫iV#:[/<&$f\G$6pC9Br
Z>r_i{ߵޛd=27|Hyd9f4k^ս~^*hGώ9Ri61_:^wy^Xn۷cS2G$˩Д׍/|=xKռWykd^wnBZD7"hg I94WUrW~Ue0ZX^odӇk ( ( ( ( (̿e6#h?o6 (
4: ,
ԓX^<}is}XYyP\CèYYOFV ߢ<(5XjWQ鲱fmacg+*|%A !W>@q?-OM޹hVoug$dYe%e:Fã(HFҾ,p=yo=E5XĆN\ IWz%AuO:u4tX]5g);s$F@5  7grju;Xt8Li沝,c@`x~(_ϩ*fn66ȑܤ^d؍#}Bޢ1lgmGQҴr^y+o I#*ʊH c+׼ :_Cm?OԵVrTMBqQ@GqlY˨ç[+! aw6R+J /|?#zbT[KH7r 1ڠQT"5 c&Ueu<RP/⟉|&)񆗧=DE,-+ؐnm{yeX:H28 #KX+< >nKGXӶ
sT < '4
ac[je7vwŽ$hC#JAjPxN[ǟ#xKN#M((((/M/MM? 'W]5],۽%Bʗ6u0_8li<Da >oJK=[[o"Tw\y`}䄉 R5T7
WA3סWZnONOϨjr[6kmj˽
Uw2#(^;χOKηgY^=ơiɯO/w l@~1MY}O)&}3PYPky"BقB5FvW_6x_Fqa\"F[q uK *
?0x5H't_LRDǹ
,d
hS^ O-#NnDhbۨ$FUkgM ;T
3k>%okM7K]Z pDHObRkF[*:o kR\YKe纶K%)[3hȻp=G(|5gx'\o#ZN [n%
aX
>xsx[kJ6jƝovmbl +ѬRHx-y7 5gZ
-V{cHHc b!+0 |Ѿ)WLy'; S[>MFZUv$IRhe
yowt7GXIu.vr+) /ʜU?~'kNZqWP ȁ9f@\Q1$1j(U$p9<QCGӯm|'q342W6j|QY0r>c;{뿴omSZ7z([H 8wܐ[2a8t+}>b!̑mf8I=Ms<YzʺEvpDcsi.|gsZ> V1noMGS%w#y:L+}ӅFa7OI7n#8 _|A5^Ja[pd0;Efb
97yGO-Rk.4,IFh[n6Mmp;o*]Cx/:G-CBF[2*-rY1€8o|a?vmm.Eup:$| <]|9i7 Ph> [N%s))${9FQT1+>| i#mo.-ʀw1Œ*(
*6:ܽB+Gh3$|W$F*p٤b)c 3ylB} h1kzc=2H$$hE*F*Rw<r<_pzmP
EP
EP
EP
EP~m? FzmeĚ?4mrAng 34QYG'_5|COzuOKmG\v}OSEs;i"y$4G 4D,}Rw+B+/!m湶CѠ@N8%z]NGuImgK,xFOX0hx?Ϯx{ Y^x!5C[c$Zs+ӧ=@s >X`w0 ~ ~F//rEލ(Mqqyx,u 
/ i9 fKkg[;0hSWS Ҽ_W/Usj'[")-ld347`h"9&vPE}^GD 𷇴FB5+ImM.3b|-ܨıHS a9|!]g76Aoz,-u^W95w巖UET`0 h:.twDy!d+yk>+zM[Y.uk 7]ڲd EtBIf>Y絴.K;V U=Nk,k3YO27 8X u/qY R_X;D6Ly1*ǁNt m/L%"PjUP: &]ÚYvӢ%ȍJ粗 xg~23&Gwu.c^/#p1FF)iUQB
jxce쭯nney ]I,Iʔ}J6]{t@_ZͦϨ{{oYZC"BgGd!#b1$~sT45=;Fξ{cLey2mocgY}=emXNMNtĿC} 
? [.dGdk]2~׽E88$Fq@$ –>4jtof4ٖ0Ӽs8,7jxBBKx'O|-GC3$ffwwlY<8ZO<)xZmNT<諸bFK
#nƯh^6+km#E6K"kPv%m*$ƪ xcDv9al]geX**x^ қ[VeYXs(dozho|O~-PRifGɺq,m-hX x_᎗&i~}ss2Ƃ8QnJG"U^9$(2?Fvyiᲀ򐻤G+o>6T tKOmCAGHDyWq7;HIXUU>I(`2 S#` N=ƾ6|?/]_4}"Yt; sFnlCImۻl y~I#+0h)9G~мC ٛM:x-ב G* *+f (<r<_pzmP
EP
EP
EP
EP~m? Fz.ag\隅ky xSt⺆ [Kw3J`!eYP~U0$eP
^cDZ?ÆҴ.Q:ڵQuPb\Tu`C@ 8|6ox קY<+AŭuWR,Uwg#" (SѾO~~YM'%1C-h<
xs?xz[\ne#<XO3:V! ('=">^+a1IY ] 0 Ocž&񯉓>*:]u% KO/nUmBW LL1iikai VѬ0H@ :
@Q@Q@ks߀Om6l񴰇ӻ";x؃,+#<|Gύ5?hpxr[a֣[ #&1iHT}-:U2Oӭ#%$ 2O'/x:6 'moO
cMwI`[k8^y'1:.MlQ@R5+D.u}J9dXck$QK wwwKcѴ 9MORWmy=h򯘳ەto(E&_~q#1j/Z%^dHQJƻd@8
@Q@e9o؍/8x6 ( ( ( ( (̿e6#h?o6 ( ( ( ( |q]֓G<@̥d2RU0nniֺFg)oo*"@UU
Vh (9/ <%&K!{;ɭEberG)2G\>|*Ev-<UW\(](AaW@Q@Q@e9o؍/8x6 ( ( ( ( (̿f6#h?o6 ( ( ( ( ( ( ( ( ( ( (<r<_pzmP
EP
EP
EP
EPo4kZQ@Q@Q@Q@Q@Q@Q@Q@Q@Q@Q@
goF?P
EP
EP
EP
EP endstream endobj 76 0 obj(http://www.shirky.com/writings/ontology_images/del_urls.jpg) endobj 77 0 obj(7zupw>΍) endobj 78 0 obj<> endobj 79 0 obj<> endobj 80 0 obj<> endobj 81 0 obj<>>> endobj 82 0 obj<> endobj 83 0 obj<> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <>1<>]/P 82 0 R/S/Article/T()/Pg 84 0 R>> endobj 84 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 85 0 obj<> endobj 86 0 obj[83 0 R 83 0 R] endobj 87 0 obj<>/A<>/Border[0 0 0]>> endobj 88 0 obj[87 0 R 239 0 R 240 0 R] endobj 89 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 90 0 obj[83 0 R] endobj 91 0 obj<> endobj 92 0 obj[89 0 R/XYZ 0 653.395 null] endobj 93 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 94 0 obj[93 0 R 96 0 R] endobj 95 0 obj[89 0 R/XYZ 0 606.141 null] endobj 96 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 97 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 98 0 obj[83 0 R] endobj 99 0 obj[97 0 R/XYZ 0 261.749 null] endobj 100 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 101 0 obj[100 0 R] endobj 102 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 103 0 obj[83 0 R] endobj 104 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 105 0 obj[83 0 R] endobj 106 0 obj[104 0 R/XYZ 0 567.16 null] endobj 107 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 108 0 obj[107 0 R] endobj 109 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 110 0 obj[83 0 R 83 0 R] endobj 111 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 112 0 obj[83 0 R 83 0 R] endobj 113 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 114 0 obj[83 0 R] endobj 115 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 116 0 obj[83 0 R] endobj 117 0 obj[115 0 R/XYZ 0 536.629 null] endobj 118 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 119 0 obj[118 0 R] endobj 120 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 121 0 obj[83 0 R] endobj 122 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 123 0 obj<> endobj 124 0 obj<> endobj 125 0 obj<> endobj 126 0 obj<> endobj 127 0 obj<> endobj 128 0 obj<> endobj 129 0 obj<> endobj 130 0 obj<> endobj 132 0 obj[83 0 R] endobj 133 0 obj[122 0 R/XYZ 0 46.3703 null] endobj 134 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 135 0 obj[134 0 R] endobj 136 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 138 0 obj[83 0 R] endobj 139 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 140 0 obj<> endobj 141 0 obj<> endobj 143 0 obj[83 0 R] endobj 144 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 146 0 obj[83 0 R] endobj 147 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 148 0 obj<> endobj 149 0 obj<> endobj 151 0 obj[83 0 R] endobj 152 0 obj[147 0 R/XYZ 0 322.911 null] endobj 153 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 154 0 obj[153 0 R] endobj 155 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 156 0 obj<> endobj 158 0 obj[83 0 R] endobj 159 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 161 0 obj[83 0 R] endobj 162 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 164 0 obj[83 0 R] endobj 165 0 obj[162 0 R/XYZ 0 47.9529 null] endobj 166 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 167 0 obj[166 0 R] endobj 168 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 170 0 obj[83 0 R] endobj 171 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 173 0 obj[83 0 R] endobj 174 0 obj[171 0 R/XYZ 0 259.15 null] endobj 175 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 176 0 obj[175 0 R] endobj 177 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 178 0 obj<> endobj 180 0 obj[83 0 R] endobj 181 0 obj[177 0 R/XYZ 0 120.085 null] endobj 182 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 183 0 obj[182 0 R 184 0 R] endobj 184 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 185 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 187 0 obj[83 0 R] endobj 188 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 190 0 obj[83 0 R] endobj 191 0 obj[188 0 R/XYZ 0 220.039 null] endobj 192 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 193 0 obj[192 0 R] endobj 194 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 196 0 obj[83 0 R] endobj 197 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 198 0 obj<> endobj 200 0 obj[83 0 R] endobj 201 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 202 0 obj<> endobj 204 0 obj[83 0 R] endobj 205 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 207 0 obj[83 0 R] endobj 208 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 210 0 obj[83 0 R] endobj 211 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 213 0 obj[83 0 R] endobj 214 0 obj[211 0 R/XYZ 0 551.227 null] endobj 215 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 216 0 obj[215 0 R] endobj 217 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 219 0 obj[83 0 R] endobj 220 0 obj<>/Font<>/ProcSet[/PDF/Text/ImageC]>>/Type/Page>> endobj 221 0 obj<> endobj 223 0 obj[83 0 R] endobj 224 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 226 0 obj[83 0 R] endobj 227 0 obj[224 0 R/XYZ 0 109.193 null] endobj 228 0 obj<>/A<>/PA<>/Border[0 0 0]>> endobj 229 0 obj[228 0 R] endobj 230 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 232 0 obj[83 0 R] endobj 233 0 obj<>/ProcSet[/PDF/Text]>>/Type/Page>> endobj 235 0 obj[83 0 R 83 0 R] endobj 236 0 obj<>/A<>/Border[0 0 0]>> endobj 237 0 obj[236 0 R 241 0 R 242 0 R] endobj 239 0 obj<>/A<>/Border[0 0 0]>> endobj 240 0 obj<>/A<>/Border[0 0 0]>> endobj 241 0 obj<>/A<>/Border[0 0 0]>> endobj 242 0 obj<>/A<>/Border[0 0 0]>> endobj 251 0 obj<> endobj 252 0 obj<> endobj 253 0 obj<> endobj 254 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 720.579 m 564.194 720.579 l 563.314 719.699 l 31.686 719.699 l h f 30.806 720.579 m 30.806 33.843 l 31.686 33.843 l 31.686 719.699 l h f 564.194 720.579 m 564.194 33.843 l 563.314 33.843 l 563.314 719.699 l h f EMC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.8 0.8 0.8 RG 0.88 w 10 M 0 j 0 J [20 5]15.488 d 40.488 651.045 m 554.512 651.045 l S EMC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.925 0.925 0.925 rg 40.488 472.644 514.024 161.458 re f EMC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 458.572 794.954 m 458.572 742.143 l 459.453 742.143 l 459.453 794.954 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 1 RG 0.951 w []0 d 30.806 774.124 m 391.972 774.124 l S 0 0 1 rg BT /TT0 1 Tf 19.804 0 0 19.804 30.8062 776.9762 Tm (Clay Shirky's Writings About the Internet)Tj 0.4 0.4 0.4 rg 13.2027 0 0 13.2027 30.8062 757.4437 Tm (Economics & Culture, Media & Community)Tj 0 0 0 rg /TT1 1 Tf 19.804 0 0 19.804 40.4882 692.4241 Tm (Ontology is Overrated: Categories, Links, and )Tj 0 -1.375 TD (Tags )Tj /TT2 1 Tf 13.2027 0 0 13.2027 44.8891 612.9862 Tm (This piece is based on two talks I gave in the spring of 2005 -- one at \ the )Tj 0 -2.313 TD (O'Reilly ETech conference in March, entitled "Ontology Is Overrated", an\ d )Tj T* (one at the IMCExpo in April entitled "Folksonomies & Tags: The rise of u\ ser-)Tj T* (developed classification." The written version is a heavily edited )Tj T* (concatenation of those two talks. )Tj /TT0 1 Tf -0.333 -3.892 Td (Today I want to talk about categorization, and I want to convince you th\ at a lot of what )Tj 0 -2.313 TD (we think we know about categorization is wrong. In particular, I want to\ convince you )Tj T* (that many of the ways we're attempting to apply categorization to the el\ ectronic world )Tj T* (are actually a bad fit, because we've adopted habits of mind that are le\ ft over from )Tj T* (earlier strategies. )Tj 0 -3.579 TD (I also want to convince you that what we're seeing when we see the Web i\ s actually a )Tj 0 -2.313 TD (radical break with previous categorization strategies, rather than an ex\ tension of them. )Tj T* (The second part of the talk is more speculative, because it is often the\ case that old )Tj 0 -2.313 TD (systems get broken before people know what's going to take their place. \ \(Anyone )Tj T* (watching the music industry can see this at work today.\) That's what I \ think is )Tj T* (happening with categorization. )Tj 0 -3.579 TD (What I think is coming instead are much more organic ways of organizing \ information )Tj 0 -2.313 TD (than our current categorization schemes allow, based on two units -- the\ link, which can )Tj ET EMC /Article <>BDC Q q 0 18 595 806 re W* n 0 0 1 RG 0.634 w []0 d 480.929 772.002 m 573.876 772.002 l S 0 0 1 rg BT /TT3 1 Tf 13.2027 0 0 13.2027 480.929 773.9027 Tm (clay@shirky.)Tj ET 543.074 754.398 m 573.876 754.398 l S BT /TT3 1 Tf 13.2027 0 0 13.2027 543.0739 756.2992 Tm (com)Tj 0 0 0 rg ( )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(1 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 255 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 792.2431 Tm (point to anything, and the tag, which is a way of attaching labels to li\ nks. The strategy )Tj 0 -2.313 TD (of tagging -- free-form labeling, without regard to categorical constrai\ nts -- seems like a )Tj T* (recipe for disaster, but as the Web has shown us, you can extract a surp\ rising amount of )Tj T* (value from big messy data sets. )Tj ET 0 0 0 RG 0.634 w 10 M 0 j 0 J []0 d 40.488 651.494 m 323.566 651.494 l S BT /TT1 1 Tf 13.2027 0 0 13.2027 40.4882 653.3952 Tm (PART I: Classification and Its Discontents)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 326.748 651.494 m 335.238 651.494 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 326.7483 653.3952 Tm (#)Tj 0 0 0 rg ( )Tj /TT1 1 Tf -21.682 -3.579 Td (Q: What is Ontology? A: It Depends on What the Meaning of "Is" Is.)Tj /TT0 1 Tf ( )Tj ET 497.921 604.239 m 506.41 604.239 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 497.9208 606.1406 Tm (#)Tj 0 0 0 rg ( )Tj -34.647 -3.579 Td (I need to provide some quick definitions, starting with ontology. It is \ a rich irony that )Tj 0 -2.312 TD (the word "ontology", which has to do with making clear and explicit stat\ ements about )Tj 0 -2.313 TD (entities in a particular domain, has so many conflicting definitions. I'\ ll offer two general )Tj 0 -2.313 TD (ones. )Tj 0 -3.579 TD (The main thread of ontology in the philosophical sense is the study of e\ ntities and their )Tj 0 -2.313 TD (relations. The question ontology asks is: What kinds of things exist or \ can exist in the )Tj T* (world, and what manner of relations can those things have to each other?\ Ontology is )Tj T* (less concerned with what is than with what is possible. )Tj 0 -3.579 TD (The knowledge management and AI communities have a related definition --\ they've )Tj 0 -2.313 TD (taken the word "ontology" and applied it more directly to their problem.\ The sense of )Tj T* (ontology there is something like "an explicit specification of a concept\ ualization." )Tj 0 -3.579 TD (The common thread between the two definitions is essence, "Is-ness." In \ a particular )Tj 0 -2.312 TD (domain, what kinds of things can we say exist in that domain, and how ca\ n we say those )Tj 0 -2.313 TD (things relate to each other? )Tj 0 -3.579 TD (The other pair of terms I need to define are categorization and classifi\ cation. These are )Tj 0 -2.313 TD (the act of organizing a collection of entities, whether things or concep\ ts, into related )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(2 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 256 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 786.6095 Tm (groups. Though there are some field-by-field distinctions, the terms are\ in the main )Tj 0 -2.313 TD (used interchangeably. )Tj 0 -3.579 TD (And then there's ontological classification or categorization, which is \ organizing a set of )Tj 0 -2.313 TD (entities into groups, based on their essences and possible relations. A \ library catalog, )Tj T* (for example, assumes that for any new book, its logical place already ex\ ists within the )Tj T* (system, even before the book was published. That strategy of designing c\ ategories to )Tj T* (cover possible cases in advance is what I'm primarily concerned with, be\ cause it is both )Tj T* (widely used and badly overrated in terms of its value in the digital wor\ ld. )Tj 0 -3.579 TD (Now, anyone who deals with categorization for a living will tell you the\ y can never get a )Tj 0 -2.313 TD (perfect system. In working classification systems, success is not "Did w\ e get the ideal )Tj T* (arrangement?" but rather "How close did we come, and on what measures?" \ The idea of )Tj 0 -2.312 TD (a perfect scheme is simply a Platonic ideal. However, I want to argue th\ at even the )Tj 0 -2.313 TD (ontological )Tj /TT1 1 Tf (ideal)Tj /TT0 1 Tf ( is a mistake. Even using theoretical perfection as a measure of )Tj 0 -2.312 TD (practical success leads to misapplication of resources. )Tj 0 -3.579 TD (Now, to the problems of classification. )Tj /TT2 1 Tf T* (Cleaving Nature at the Joints)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 239.848 259.847 m 248.338 259.847 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 239.8484 261.7486 Tm (#)Tj 0 0 0 rg ( )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(3 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 257 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n q 371.434906 0 0 235.0074005 43.1286926 581.9511871 cm /Im0 Do Q 0 0 0 rg BT /TT0 1 Tf 12.3225 0 0 12.3225 414.5636 699.4549 Tm ( )Tj -23.4 -10.848 Td ([ The Periodic Table of the Elements ])Tj 13.2027 0 0 13.2027 40.4882 515.1748 Tm (The periodic table of the elements is my vote for "Best. Classification.\ Evar." It turns out )Tj 0 -2.313 TD (that by organizing elements by the number of protons in the nucleus, you\ get all of this )Tj T* (fantastic value, both descriptive and predictive value. And because what\ you're doing is )Tj T* (organizing )Tj /TT1 1 Tf (things)Tj /TT0 1 Tf (, the periodic table is as close to making assertions about essence as i\ t )Tj 0 -2.312 TD (is physically possible to get. This is a really powerful scheme, almost \ perfect. Almost. )Tj 0 -3.579 TD (All the way over in the right-hand column, the pink column, are noble ga\ ses. Now noble )Tj 0 -2.313 TD (gas is an odd category, because helium is no more a gas than mercury is \ a liquid. )Tj T* (Helium is not fundamentally a gas, it's just a gas at most temperatures,\ but the people )Tj T* (studying it at the time didn't know that, because they weren't able to m\ ake it cold )Tj T* (enough to see that helium, like everything else, has different states of\ matter. Lacking )Tj T* (the right measurements, they assumed that gaseousness was an essential a\ spect -- )Tj T* (literally, part of the essence -- of those elements. )Tj 0 -3.579 TD (Even in a nearly perfect categorization scheme, there are these kinds of\ context errors, )Tj 0 -2.313 TD (where people are placing something that is merely true at room temperatu\ re, and is )Tj T* (absolutely unrelated to essence, right in the center of the categorizati\ on. And the )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(4 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 258 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (category 'Noble Gas' has stayed there from the day they added it, becaus\ e we've all just )Tj 0 -2.313 TD (gotten used to that anomaly as a frozen accident. )Tj 0 -3.579 TD (If it's impossible to create a completely coherent categorization, even \ when you're doing )Tj 0 -2.313 TD (something as physically related to essence as chemistry, imagine the pro\ blems faced by )Tj T* (anyone who's dealing with a domain where essence is even less obvious. )Tj 0 -3.579 TD (Which brings me to the subject of libraries. )Tj /TT1 1 Tf T* (Of Cards and Catalogs)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 192.926 565.259 m 201.415 565.259 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 192.9261 567.1603 Tm (#)Tj 0 0 0 rg ( )Tj -11.546 -3.579 Td (The periodic table gets my vote for the best categorization scheme ever,\ but libraries )Tj 0 -2.312 TD (have the best-known categorization schemes. The experience of the librar\ y catalog is )Tj 0 -2.313 TD (probably what people know best as a high-order categorized view of the w\ orld, and )Tj 0 -2.313 TD (those cataloging systems contain all kinds of odd mappings between the c\ ategories and )Tj 0 -2.313 TD (the world they describe. )Tj 0 -3.579 TD (Here's the first top-level category in the Soviet library system: )Tj /T1_1 1 Tf 2.667 -3.535 Td (A: Marxism-Leninism)Tj /T1_2 1 Tf 0 -2.313 TD (A1: Classic works of Marxism-Leninism)Tj T* (A3: Life and work of C.Marx, F.Engels, V.I.Lenin)Tj T* (A5: Marxism-Leninism Philosophy)Tj T* (A6: Marxist-Leninist Political Economics)Tj T* (A7/8: Scientific Communism)Tj /TT0 1 Tf -2.667 -3.623 Td (Some of those categories are starting to look a little bit dated. )Tj 0 -3.579 TD (Or, my favorite -- this is the Dewey Decimal System's categorization for\ religions of the )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(5 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 259 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (world, which is the 200 category. )Tj /T1_1 1 Tf 2.667 -3.535 Td (Dewey, 200: Religion)Tj /T1_2 1 Tf 0 -2.313 TD (210 Natural theology)Tj 0 -2.313 TD (220 Bible)Tj T* (230 Christian theology)Tj T* (240 Christian moral & devotional theology)Tj T* (250 Christian orders & local church)Tj T* (260 Christian social theology)Tj T* (270 Christian church history)Tj T* (280 Christian sects & denominations)Tj T* (290 Other religions)Tj /TT0 1 Tf -2.667 -3.623 Td (How much is this not the categorization you want in the 21st century? )Tj 0 -3.579 TD (This kind of bias is rife in categorization systems. Here's the Library \ of Congress' )Tj 0 -2.313 TD (categorization of History. These are all the top-level categories -- all\ of these things are )Tj 0 -2.313 TD (presented as being co-equal. )Tj /TT1 1 Tf 12.3225 0 0 12.3225 82.7367 268.886 Tm (D: History \(general\))Tj /T1_2 1 Tf 0 -3.469 TD (DA: Great Britain)Tj 0 -2.513 TD (DB: Austria)Tj 0 -2.513 TD (DC: France)Tj T* (DD: Germany)Tj T* (DE: Mediterranean)Tj T* (DF: Greece)Tj 0 -2.513 TD (DG: Italy)Tj ET EMC /Article <>BDC BT /T1_2 1 Tf 12.3225 0 0 12.3225 220.7485 226.1356 Tm (DK: Former Soviet Union)Tj 0 -2.513 TD (DL: Scandinavia)Tj T* (DP: Iberian Peninsula)Tj T* (DQ: Switzerland)Tj /T1_1 1 Tf T* (DR: Balkan Peninsula)Tj T* (DS: Asia)Tj T* (DT: Africa)Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(6 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 260 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /T1_1 1 Tf 12.3225 0 0 12.3225 82.7367 792.4523 Tm (DH: Low Countries)Tj 0 -2.513 TD (DJ: Netherlands)Tj ET EMC /Article <>BDC BT /T1_1 1 Tf 12.3225 0 0 12.3225 220.7485 792.4523 Tm (DU: Oceania)Tj T* (DX: Gypsies)Tj /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 689.2074 Tm (I'd like to call your attention to the ones in bold: The Balkan Peninsul\ a. Asia. Africa. )Tj 0 -3.579 TD (And just, you know, to review the geography: )Tj ET q 385.517746 0 0 205.0813599 43.1286926 403.4176178 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 428.6465 505.9583 Tm ( )Tj -20.663 -9.634 Td ([ Spot the difference? ])Tj 13.2027 0 0 13.2027 40.4882 336.6413 Tm (Yet, for all the oddity of placing the Balkan Peninsula and Asia in the \ same level, this is )Tj 0 -2.313 TD (harder to laugh off than the Dewey example, because it's so puzzling. Th\ e Library of )Tj T* (Congress -- no slouches in the thinking department, founded by Thomas Je\ fferson -- )Tj T* (has a staff of people who do nothing but think about categorization all \ day long. So )Tj T* (what's being optimized here? It's not geography. It's not population. It\ 's not regional )Tj 0 -2.313 TD (GDP. )Tj 0 -3.579 TD (What's being optimized is number of books on the shelf. That's what the \ categorization )Tj 0 -2.313 TD (scheme is categorizing. It's tempting to think that the classification s\ chemes that )Tj T* (libraries have optimized for in the past can be extended in an uncomplic\ ated way into )Tj T* (the digital world. This badly underestimates, in my view, the degree to \ which what )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(7 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 261 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 797.7217 Tm (libraries have historically been managing is an entirely different probl\ em. )Tj 0 -3.579 TD (The musculature of the Library of Congress categorization scheme looks l\ ike it's about )Tj 0 -2.312 TD (concepts. It is organized into non-overlapping categories that get more \ detailed at lower )Tj 0 -2.313 TD (and lower levels -- any concept is supposed to fit in one category and i\ n no other )Tj 0 -2.313 TD (categories. But every now and again, the skeleton pokes through, and the\ skeleton, the )Tj 0 -2.313 TD (supporting structure around which the system is really built, is designe\ d to minimize )Tj T* (seek time on shelves. )Tj 0 -3.579 TD (The essence of a book isn't the ideas it contains. The essence of a book\ is "book." )Tj 0 -2.313 TD (Thinking that library catalogs exist to organize concepts confuses the c\ ontainer for the )Tj T* (thing contained. )Tj 0 -3.579 TD (The categorization scheme is a response to physical constraints on stora\ ge, and to )Tj 0 -2.313 TD (people's inability to keep the location of more than a few hundred thing\ s in their mind )Tj T* (at once. Once you own more than a few hundred books, you have to organiz\ e them )Tj T* (somehow. \(My mother, who was a reference librarian, said she wanted to \ reshelve the )Tj 0 -2.313 TD (entire University library by color, because students would come in and s\ ay "I'm looking )Tj T* (for a sociology book. It's green..."\) But however you do it, the frailt\ y of human memory )Tj T* (and the physical fact of books make some sort of organizational scheme a\ requirement, )Tj T* (and hierarchy is a good way to manage physical objects. )Tj 0 -3.579 TD (The "Balkans/Asia" kind of imbalance is simply a byproduct of physical c\ onstraints. It )Tj 0 -2.313 TD (isn't the ideas in a book that have to be in one place -- a book can be \ about several )Tj T* (things at once. It is the book itself, the physical fact of the bound ob\ ject, that has to be )Tj T* (one place, and if it's one place, it can't also be in another place. And\ this in turn means )Tj T* (that a book has to be declared to be )Tj /TT1 1 Tf (about)Tj /TT0 1 Tf ( some main thing. A book which is equally )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(8 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 262 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (about two things breaks the 'be in one place' requirement, so each book \ needs to be )Tj 0 -2.313 TD (declared to about one thing more than others, regardless of its actual c\ ontents. )Tj 0 -3.579 TD (People have been freaking out about the virtuality of data for decades, \ and you'd think )Tj 0 -2.313 TD (we'd have internalized the obvious truth: there is no shelf. In the digi\ tal world, there is )Tj T* (no physical constraint that's forcing this kind of organization on us an\ y longer. We can )Tj T* (do without it, and you'd think we'd have learned that lesson by now. )Tj 0 -3.579 TD (And yet. )Tj /TT1 1 Tf T* (The Parable of the Ontologist, or, "There Is No Shelf")Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 402.4 534.728 m 410.889 534.728 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 402.3996 536.6292 Tm (#)Tj 0 0 0 rg ( )Tj -27.412 -3.579 Td (A little over ten years ago, a couple of guys out of Stanford launched a\ service called )Tj 0 -2.312 TD (Yahoo that offered a list of things available on the Web. It was the fir\ st really significant )Tj 0 -2.313 TD (attempt to bring order to the Web. As the Web expanded, the Yahoo list g\ rew into a )Tj 0 -2.313 TD (hierarchy with categories. As the Web expanded more they realized that, \ to maintain )Tj 0 -2.313 TD (the value in the directory, they were going to have to systematize, so t\ hey hired a )Tj T* (professional ontologist, and they developed their now-familiar top-level\ categories, )Tj T* (which go to subcategories, each subcategory contains links to still othe\ r subcategories, )Tj T* (and so on. Now we have this ontologically managed list of what's out the\ re. )Tj 0 -3.579 TD (Here we are in one of Yahoo's top-level categories, Entertainment. )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(9 de 34\)28/01/\ 2008 19:04:54)Tj ET EMC endstream endobj 263 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n q 448.0103607 0 0 272.8550262 43.1286926 541.0399475 cm /Im0 Do Q 0 0 0 rg BT /TT0 1 Tf 12.3225 0 0 12.3225 491.1391 677.4675 Tm ( )Tj -25.968 -12.384 Td ([ Yahoo's Entertainment Category ])Tj 13.2027 0 0 13.2027 40.4882 474.2636 Tm (You can see what the sub-categories of Entertainment are, whether or not\ there are new )Tj 0 -2.313 TD (additions, and how many links roll up under those sub-categories. Except\ , in the case of )Tj T* (Books and Literature, that sub-category doesn't tell you how many links \ roll up under )Tj T* (it. Books and Literature doesn't end with a number of links, but with an\ "@" sign. That )Tj T* ("@" sign is telling you that the category of Books and Literature isn't \ 'really' in the )Tj 0 -2.313 TD (category Entertainment. Yahoo is saying "We've put this link here for yo\ ur convenience, )Tj 0 -2.312 TD (but that's only to take you to where Books and Literature 'really' are."\ To which one can )Tj 0 -2.313 TD (only respond -- "What's real?" )Tj 0 -3.579 TD (Yahoo is saying "We understand better than you how the world is organize\ d, because )Tj 0 -2.313 TD (we are trained professionals. So if you mistakenly think that Books and \ Literature are )Tj T* (entertainment, we'll put a little flag up so we can set you right, but t\ o see those links, )Tj T* (you have to 'go' to where they 'are'." \(My fingers are going to fall of\ f from all the air )Tj T* (quotes.\) When you go to Literature -- which is part of Humanities, not \ Entertainment -- )Tj T* (you are told, similarly, that booksellers are not 'really' there. Becaus\ e they are a )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(10 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 264 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (commercial service, booksellers are 'really' in Business. )Tj ET q 450.6508942 0 0 223.565094 43.1286926 543.4984283 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 493.7796 655.281 Tm ( )Tj -23.538 -10.384 Td ([ 'Literature' on Yahoo ])Tj 13.2027 0 0 13.2027 40.4882 476.7221 Tm (Look what's happened here. Yahoo, faced with the possibility that they c\ ould organize )Tj 0 -2.313 TD (things with no physical constraints, )Tj /TT1 1 Tf (added the shelf back)Tj /TT0 1 Tf (. They couldn't imagine )Tj 0 -2.312 TD (organization without the constraints of the shelf, so they added it back\ . It is perfectly )Tj 0 -2.313 TD (possible for any number of links to be in any number of places in a hier\ archy, or in )Tj 0 -2.313 TD (many hierarchies, or in no hierarchy at all. But Yahoo decided to privil\ ege one way of )Tj T* (organizing links over all others, because they wanted to make assertions\ about what is )Tj T* ("real." )Tj 0 -3.579 TD (The charitable explanation for this is that they thought of this kind of\ a priori )Tj 0 -2.313 TD (organization as their job, and as something their users would value. The\ uncharitable )Tj T* (explanation is that they thought there was business value in determining\ the view the )Tj T* (user would have to adopt to use the system. Both of those explanations m\ ay have been )Tj T* (true at different times and in different measures, but the effect was to\ override the )Tj T* (users' sense of where things ought to be, and to insist on the Yahoo vie\ w instead. )Tj /TT2 1 Tf 0 -3.579 TD (File Systems and Hierarchy)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 230.118 44.469 m 238.607 44.469 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 230.118 46.3703 Tm (#)Tj 0 0 0 rg ( )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(11 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 265 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 782.2311 Tm (It's easy to see how the Yahoo hierarchy maps to technological constrain\ ts as well as )Tj 0 -2.313 TD (physical ones. The constraints in the Yahoo directory describes both a l\ ibrary )Tj T* (categorization scheme and, obviously, a file system -- the file system i\ s both a powerful )Tj T* (tool and a powerful metaphor, and we're all so used to it, it seems natu\ ral. )Tj ET q 330.9467468 0 0 193.6390533 43.1286926 463.544754 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 374.0754 560.3643 Tm ( )Tj -16.266 -9.169 Td ([ Hierarchy ])Tj 13.2027 0 0 13.2027 40.4882 396.7684 Tm (There's a top level, and subdirectories roll up under that. Subdirectori\ es contain files or )Tj T* (further subdirectories and so on, all the way down. Both librarians and \ computer )Tj T* (scientists hit the same next idea, which is "You know, it wouldn't hurt \ to add a few )Tj T* (secondary links in here" -- symbolic links, aliases, shortcuts, whatever\ you want to call )Tj T* (them. )Tj ET q 344.1494141 0 0 193.6390533 43.1286926 47.5508575 cm /Im1 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 387.2781 144.3704 Tm ( )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(12 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 266 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 12.3225 0 0 12.3225 179.271 814.4959 Tm ([ Plus Links ])Tj 13.2027 0 0 13.2027 40.4882 763.8899 Tm (The Library of Congress has something similar in its second-order catego\ rization -- )Tj 0 -2.313 TD ("This book is mainly about the Balkans, but it's also about art, or it's\ mainly about art, )Tj T* (but it's also about the Balkans." Most hierarchical attempts to subdivid\ e the world use )Tj T* (some system like this. )Tj 0 -3.579 TD (Then, in the early 90s, one of the things that Berners-Lee showed us is \ that you could )Tj 0 -2.313 TD (have a lot of links. You don't have to have just a few links, you could \ have a whole lot of )Tj 0 -2.313 TD (links. )Tj ET q 346.7899475 0 0 206.8417206 43.1286926 323.6839905 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 389.9186 427.1048 Tm ( )Tj -18.62 -9.705 Td ([ Plus Lots of Links ])Tj 13.2027 0 0 13.2027 40.4882 256.9076 Tm (This is where Yahoo got off the boat. They said, "Get out of here with t\ hat crazy talk. A )Tj 0 -2.313 TD (URL can only appear in three places. That's the Yahoo rule." They did th\ at in part )Tj T* (because they didn't want to get spammed, since they were doing a commerc\ ial )Tj T* (directory, so they put an upper limit on the number of symbolic links th\ at could go into )Tj T* (their view of the world. They missed the end of this progression, which \ is that, if you've )Tj 0 -2.313 TD (got enough links, you don't need the hierarchy anymore. There is no shel\ f. There is no )Tj T* (file system. The links alone are enough. )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(13 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 267 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n q 337.1079865 0 0 205.9615326 43.1286926 607.8432007 cm /Im0 Do Q 0 0 0 rg BT /TT0 1 Tf 12.3225 0 0 12.3225 380.2367 710.824 Tm ( )Tj -22.151 -9.669 Td ([ Just Links \(There Is No Filesystem\) ])Tj 13.2027 0 0 13.2027 40.4882 541.0668 Tm (One reason Google was adopted so quickly when it came along is that Goog\ le )Tj 0 -2.313 TD (understood there is no shelf, and that there is no file system. Google c\ an decide what )Tj T* (goes with what )Tj /TT1 1 Tf (after)Tj /TT0 1 Tf ( hearing from the user, rather than trying to predict in advance )Tj 0 -2.312 TD (what it is you need to know. )Tj 0 -3.579 TD (Let's say I need every Web page with the word "obstreperous" and "Minnes\ ota" in it. )Tj 0 -2.313 TD (You can't ask a cataloguer in advance to say "Well, that's going to be a\ useful category, )Tj T* (we should encode that in advance." Instead, what the cataloguer is going\ to say is, )Tj T* ("Obstreperous plus Minnesota! Forget it, we're not going to optimize for\ one-offs like )Tj T* (that." Google, on the other hand, says, "Who cares? We're not going to t\ ell the user )Tj T* (what to do, because the link structure is more complex than we can read,\ except in )Tj T* (response to a user query." )Tj 0 -3.579 TD (Browse versus search is a radical increase in the trust we put in link i\ nfrastructure, and )Tj 0 -2.313 TD (in the degree of power derived from that link structure. Browse says the\ people making )Tj 0 -2.313 TD (the ontology, the people doing the categorization, have the responsibili\ ty to organize )Tj T* (the world in advance. Given this requirement, the views of the cataloger\ s necessarily )Tj T* (override the user's needs and the user's view of the world. If you want \ something that )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(14 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 268 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (hasn't been categorized in the way you think about it, you're out of luc\ k. )Tj 0 -3.579 TD (The search paradigm says the reverse. It says nobody gets to tell you in\ advance what it )Tj 0 -2.312 TD (is you need. Search says that, at the moment that you are looking for it\ , we will do our )Tj 0 -2.313 TD (best to service it based on this link structure, because we believe we c\ an build a world )Tj 0 -2.313 TD (where we don't need the hierarchy to coexist with the link structure. )Tj 0 -3.579 TD (A lot of the conversation that's going on now about categorization start\ s at a second )Tj 0 -2.313 TD (step -- "Since categorization is a good way to organize the world, we sh\ ould..." But the )Tj T* (first step is to ask the critical question: Is categorization a good ide\ a? We can see, from )Tj T* (the Yahoo versus Google example, that there are a number of cases where \ you get )Tj T* (significant value out of )Tj /TT1 1 Tf (not)Tj /TT0 1 Tf ( categorizing. Even Google adopted DMOZ, the open source )Tj 0 -2.312 TD (version of the Yahoo directory, and later they downgraded its presence o\ n the site, )Tj 0 -2.312 TD (because almost no one was using it. When people were offered search and \ )Tj 0 -2.313 TD (categorization side-by-side, fewer and fewer people were using categoriz\ ation to find )Tj T* (things. )Tj /TT2 1 Tf 0 -3.579 TD (When Does Ontological Classification Work Well?)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 382.345 321.01 m 390.834 321.01 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 382.3447 322.9111 Tm (#)Tj 0 0 0 rg ( )Tj -25.893 -3.579 Td (Ontological classification works well in some places, of course. You nee\ d a card catalog )Tj 0 -2.312 TD (if you are managing a physical library. You need a hierarchy to manage a\ file system. So )Tj 0 -2.313 TD (what you want to know, when thinking about how to organize anything, is \ whether that )Tj 0 -2.313 TD (kind of classification is a good strategy. )Tj 0 -3.579 TD (Here is a partial list of characteristics that help make it work: )Tj /TT2 1 Tf T* (Domain to be Organized)Tj /TT0 1 Tf ( )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 42.2995 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 42.2995 Tm (Small corpus )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(15 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 269 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 794.8837 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 794.8837 Tm (Formal categories )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 764.3525 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 764.3525 Tm (Stable entities )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 733.8214 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 733.8214 Tm (Restricted entities )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 703.2902 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 703.2902 Tm (Clear edges )Tj -2.667 -3.579 Td (This is all the domain-specific stuff that you would like to be true if \ you're trying to )Tj 0 -2.313 TD (classify cleanly. The periodic table of the elements has all of these th\ ings -- there are )Tj 0 -2.313 TD (only a hundred or so elements; the categories are simple and derivable; \ protons don't )Tj T* (change because of political circumstances; only elements can be classifi\ ed, not )Tj T* (molecules; there are no blended elements; and so on. The more of those c\ haracteristics )Tj T* (that are true, the better a fit ontology is likely to be. )Tj 0 -3.579 TD (The other key question, besides the characteristics of the domain itself\ , is "What are the )Tj 0 -2.313 TD (participants like?" Here are some things that, if true, help make ontolo\ gy a workable )Tj 0 -2.312 TD (classification strategy: )Tj /TT1 1 Tf 0 -3.579 TD (Participants)Tj /TT0 1 Tf ( )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 300.554 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 300.554 Tm (Expert catalogers )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 270.0228 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 270.0228 Tm (Authoritative source of judgment )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 239.4917 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 239.4917 Tm (Coordinated users )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 208.9605 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 208.9605 Tm (Expert users )Tj -2.667 -3.579 Td (DSM-IV, the 4th version of the psychiatrists' Diagnostic and Statistical\ Manual, is a )Tj 0 -2.312 TD (classic example of an classification scheme that works because of these \ characteristics. )Tj 0 -2.313 TD (DSM IV allows psychiatrists all over the US, in theory, to make the same\ judgment )Tj 0 -2.313 TD (about a mental illness, when presented with the same list of symptoms. T\ here is an )Tj 0 -2.313 TD (authoritative source for DSM-IV, the American Psychiatric Association. T\ he APA gets to )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(16 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 270 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 792.1656 Tm (say what symptoms add up to psychosis. They have both expert cataloguers\ and expert )Tj 0 -2.313 TD (users. The amount of 'people infrastructure' that's hidden in a working \ system like DSM )Tj T* (IV is a big part of what makes this sort of categorization work. )Tj 0 -3.579 TD (This 'people infrastructure' is very expensive, though. One of the probl\ em users have )Tj 0 -2.313 TD (with categories is that when we do head-to-head tests -- we describe som\ ething and )Tj T* (then we ask users to guess how we described it -- there's a very poor ma\ tch. Users have )Tj T* (a terrifically hard time guessing how something they want will have been\ categorized in )Tj T* (advance, unless they have been educated about those categories in advanc\ e as well, and )Tj T* (the bigger the user base, the more work that user education is. )Tj 0 -3.579 TD (You can also turn that list around. You can say "Here are some character\ istics where )Tj 0 -2.313 TD (ontological classification doesn't work well": )Tj /TT1 1 Tf 0 -3.579 TD (Domain)Tj /TT0 1 Tf ( )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 358.8982 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 358.8982 Tm (Large corpus )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 328.3671 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 328.3671 Tm (No formal categories )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 297.8359 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 297.8359 Tm (Unstable entities )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 267.3047 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 267.3047 Tm (Unrestricted entities )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 236.7736 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 236.7736 Tm (No clear edges )Tj /TT1 1 Tf -2.667 -3.579 Td (Participants)Tj /TT0 1 Tf ( )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 142.2645 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 142.2645 Tm (Uncoordinated users )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 111.7334 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 111.7334 Tm (Amateur users )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 81.2022 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 81.2022 Tm (Naive catalogers )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 50.671 Tm (l)Tj /TT0 1 Tf ( )Tj 13.2027 0 0 13.2027 75.6953 50.671 Tm (No Authority )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(17 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 271 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 786.5319 Tm (If you've got a large, ill-defined corpus, if you've got naive users, if\ your cataloguers )Tj 0 -2.313 TD (aren't expert, if there's no one to say authoritatively what's going on,\ then ontology is )Tj T* (going to be a bad strategy. )Tj 0 -3.579 TD (The list of factors making ontology a bad fit is, also, an almost perfec\ t description of the )Tj 0 -2.313 TD (Web -- largest corpus, most naive users, no global authority, and so on.\ The more you )Tj T* (push in the direction of scale, spread, fluidity, flexibility, the harde\ r it becomes to )Tj T* (handle the expense of starting a cataloguing system and the hassle of ma\ intaining it, to )Tj T* (say nothing of the amount of force you have to get to exert over users t\ o get them to )Tj T* (drop their own world view in favor of yours. )Tj 0 -3.579 TD (The reason we know SUVs are a light truck instead of a car is that the G\ overnment says )Tj 0 -2.313 TD (they're a light truck. This is voodoo categorization, where acting on th\ e model changes )Tj T* (the world -- when the Government says an SUV is a truck, it )Tj /TT1 1 Tf (is)Tj /TT0 1 Tf ( a truck, by definition. )Tj 0 -2.312 TD (Much of the appeal of categorization comes from this sort of voodoo, whe\ re the people )Tj 0 -2.313 TD (doing the categorizing believe, even if only unconciously, that naming t\ he world )Tj T* (changes it. Unfortunately, most of the world is not actually amenable to\ voodoo )Tj T* (categorization. )Tj 0 -3.579 TD (The reason we don't know whether or not )Tj /TT1 1 Tf (Buffy, The Vampire Slayer)Tj /TT0 1 Tf ( is science fiction, )Tj 0 -2.312 TD (for example, is because there's no one who can say definitively yes or n\ o. In )Tj T* (environments where there's no authority and no force that can be applied\ to the user, )Tj 0 -2.313 TD (it's very difficult to support the voodoo style of organization. Merely \ naming the world )Tj T* (creates no actual change, either in the world, or in the minds of potent\ ial users who )Tj T* (don't understand the system. )Tj /TT2 1 Tf 0 -3.579 TD (Mind Reading)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 139.165 46.052 m 147.654 46.052 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 139.1649 47.9529 Tm (#)Tj 0 0 0 rg ( )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(18 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 272 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 783.8138 Tm (One of the biggest problems with categorizing things in advance is that \ it forces the )Tj 0 -2.313 TD (categorizers to take on two jobs that have historically been quite hard:\ mind reading, )Tj T* (and fortune telling. It forces categorizers to guess what their users ar\ e thinking, and to )Tj T* (make predictions about the future. )Tj 0 -3.579 TD (The mind-reading aspect shows up in conversations about controlled vocab\ ularies. )Tj 0 -2.313 TD (Whenever users are allowed to label or tag things, someone always says "\ Hey, I know! )Tj 0 -2.313 TD (Let's make a thesaurus, so that if you tag something 'Mac' and I tag it \ 'Apple' and )Tj T* (somebody else tags it 'OSX', we all end up looking at the same thing!" T\ hey point to the )Tj T* (signal loss from the fact that users, although they use these three diff\ erent labels, are )Tj T* (talking about the same thing. )Tj 0 -3.579 TD (The assumption is that we both can and should read people's minds, that \ we can )Tj 0 -2.313 TD (understand what they meant when they used a particular label, and, under\ standing )Tj T* (that, we can start to restrict those labels, or at least map them easily\ onto one another. )Tj 0 -3.579 TD (This looks relatively simple with the Apple/Mac/OSX example, but when we\ start to )Tj 0 -2.313 TD (expand to other groups of related words, like movies, film, and cinema, \ the case for the )Tj T* (thesaurus becomes much less clear. I learned this from Brad Fitzpatrick'\ s design for )Tj 0 -2.312 TD (LiveJournal, which allows user to list their own interests. LiveJournal \ makes absolutely )Tj 0 -2.313 TD (no attempt to enforce solidarity or a thesaurus or a minimal set of term\ s, no check-box, )Tj T* (no drop-box, just free-text typing. Some people say they're interested i\ n movies. Some )Tj T* (people say they're interested in film. Some people say they're intereste\ d in cinema. )Tj 0 -3.579 TD (The cataloguers first reaction to that is, "Oh my god, that means you wo\ n't be )Tj 0 -2.313 TD (introducing the movies people to the cinema people!" To which the obviou\ s answer is )Tj T* ("Good. The movie people don't )Tj /TT1 1 Tf (want)Tj /TT0 1 Tf ( to hang out with the cinema people." Those terms )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(19 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 273 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 797.8191 Tm (actually encode different things, and the assertion that restricting voc\ abularies )Tj 0 -2.313 TD (improves signal assumes that that there's no signal in the difference it\ self, and no value )Tj T* (in protecting the user from too many matches. )Tj 0 -3.579 TD (When we get to really contested terms like queer/gay/homosexual, by this\ point, all the )Tj 0 -2.313 TD (signal loss is in the collapse, not in the expansion. "Oh, the people ta\ lking about 'queer )Tj T* (politics' and the people talking about 'the homosexual agenda', they're \ really talking )Tj T* (about the same thing." Oh no they're not. If you think the movies and ci\ nema people )Tj T* (were going to have a fight, wait til you get the queer politics and homo\ sexual agenda )Tj T* (people in the same room. )Tj 0 -3.579 TD (You can't do it. You can't collapse these categorizations without some s\ ignal loss. The )Tj 0 -2.312 TD (problem is, because the cataloguers assume their classification should h\ ave force on the )Tj 0 -2.313 TD (world, they underestimate the difficulty of understanding what users are\ thinking, and )Tj T* (they overestimate the amount to which users will agree, either with one \ another or with )Tj 0 -2.313 TD (the catalogers, about the best way to categorize. They also underestimat\ e the loss from )Tj T* (erasing difference of expression, and they overestimate loss from the la\ ck of a )Tj T* (thesaurus. )Tj /TT1 1 Tf 0 -3.579 TD (Fortune Telling)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 149.054 257.249 m 157.543 257.249 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 149.0537 259.1504 Tm (#)Tj 0 0 0 rg ( )Tj -8.223 -3.579 Td (The other big problem is that predicting the future turns out to be hard\ , and yet any )Tj 0 -2.312 TD (classification system meant to be stable over time puts the categorizer \ in the position of )Tj 0 -2.313 TD (fortune teller. )Tj 0 -3.579 TD (Alert readers will be able to spot the difference between Sentence A and\ Sentence B. )Tj /T1_1 1 Tf 2.667 -3.535 Td (A: "I love you.")Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(20 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 274 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /T1_1 1 Tf 13.2027 0 0 13.2027 75.6953 801.0995 Tm (B: "I will always love you.")Tj /TT0 1 Tf -2.667 -3.623 Td (Woe betide the person who utters Sentence B when what they mean is Sente\ nce A. )Tj 0 -2.313 TD (Sentence A is a statement. Sentence B is a prediction. )Tj 0 -3.579 TD (But this is the ontological dilemma. Consider the following statements: \ )Tj /T1_1 1 Tf 2.667 -3.535 Td (A: "This is a book about Dresden.")Tj 0 -2.312 TD (B: "This is a book about Dresden, )Tj 0 -2.313 TD ( and it goes in the category 'East Germany'." )Tj /TT0 1 Tf -2.667 -3.623 Td (That second sentence seems so obvious, but East Germany actually turned \ out to be an )Tj 0 -2.313 TD (unstable category. Cities are real. They are real, physical facts. Count\ ries are social )Tj T* (fictions. It is much easier for a country to disappear than for a city t\ o disappear, so )Tj T* (when you're saying that the small thing is contained by the large thing,\ you're actually )Tj T* (mixing radically different kinds of entities. We pretend that 'country' \ refers to a )Tj 0 -2.313 TD (physical area the same way 'city' does, but it's not true, as we know fr\ om places like the )Tj T* (former Yugoslavia. )Tj 0 -3.579 TD (There is a top-level category, you may have seen it earlier in the Libra\ ry of Congress )Tj 0 -2.313 TD (scheme, called Former Soviet Union. The best they were able to do was ju\ st tack )Tj T* ("former" onto that entire zone that they'd previously categorized as the\ Soviet Union. )Tj T* (Not because that's what they thought was true about the world, but becau\ se they don't )Tj T* (have the staff to reshelve all the books. That's the constraint. )Tj ET 0 0 0 RG 0.634 w 10 M 0 j 0 J []0 d 40.488 118.184 m 509.618 118.184 l S BT /TT1 1 Tf 13.2027 0 0 13.2027 40.4882 120.0851 Tm (Part II: The Only Group That Can Categorize Everything Is Everybody)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 512.8 118.184 m 521.29 118.184 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 512.8002 120.0851 Tm (#)Tj 0 0 0 rg ( )Tj /TT1 1 Tf -35.774 -3.579 Td ("My God. It's full of links!")Tj /TT0 1 Tf ( )Tj ET 223.622 70.929 m 232.112 70.929 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 223.6223 72.8306 Tm (#)Tj 0 0 0 rg ( )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(21 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 275 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (When we reexamine categorization without assuming the physical constrain\ t either of )Tj 0 -2.313 TD (hierarchy on disk or of hierarchy in the physical world, we get very dif\ ferent answers. )Tj T* (Let's say you wanted to merge two libraries -- mine and the Library of C\ ongress's. \(You )Tj T* (can tell it's the Library of Congress on the right, because they have a \ few more books )Tj T* (than I do.\) )Tj ET q 312.4630127 0 0 130.2662659 43.1286926 514.6726074 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 355.5917 579.8058 Tm ( )Tj -21.564 -6.598 Td ([ Two Categorized Collections of Books ])Tj 13.2027 0 0 13.2027 40.4882 447.8963 Tm (So, how do we do this? Do I have to sit down with the Librarian of Congr\ ess and say, )Tj T* ("Well, in my world, )Tj /TT1 1 Tf (Python In A Nutshell)Tj /TT0 1 Tf ( is a reference work, and I keep all of my books )Tj 0 -2.312 TD (on creativity together." Do we have to hash out the difference between m\ y )Tj 0 -2.313 TD (categorization scheme and theirs before the Library of Congress is able \ to take my )Tj 0 -2.313 TD (books? )Tj 0 -3.579 TD (No, of course we don't have to do anything of the sort. They're able to \ take my books in )Tj 0 -2.313 TD (while ignoring my categories, because all my books have ISBN numbers, In\ ternational )Tj T* (Standard Book Numbers. They're not merging at the category level. They'r\ e merging at )Tj T* (the globally unique item level. My entities, my uniquely labeled books, \ go into Library )Tj T* (of Congress scheme trivially. The presence of unique labels means that m\ erging )Tj T* (libraries doesn't require merging categorization schemes. )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(22 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 276 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n q 311.58284 0 0 127.6257324 43.1286926 686.6360931 cm /Im0 Do Q 0 0 0 rg BT /TT0 1 Tf 12.3225 0 0 12.3225 354.7115 750.449 Tm ( )Tj -16.187 -6.491 Td ([ Merge ISBNs ])Tj 13.2027 0 0 13.2027 40.4882 619.8575 Tm (Now imagine a world where )Tj /TT1 1 Tf (everything)Tj /TT0 1 Tf ( can have a unique identifier. This should be )Tj 0 -2.312 TD (easy, since that's the world we currently live in -- the URL gives us a \ way to create a )Tj 0 -2.313 TD (globally unique ID for anything we need to point to. Sometimes the point\ ers are direct, )Tj T* (as when a URL points to the contents of a Web page. Sometimes they are i\ ndirect, as )Tj 0 -2.313 TD (when you use an Amazon link to point to a book. Sometimes there are laye\ rs of )Tj T* (indirection, as when you use a URI, a uniform resource identifier, to na\ me something )Tj T* (whose location is indeterminate. But the basic scheme gives us ways to c\ reate a globally )Tj T* (unique identifier for anything. )Tj 0 -3.579 TD (And once you can do that, anyone can label those pointers, can tag those\ URLs, in ways )Tj 0 -2.313 TD (that make them more valuable, and all without requiring top-down organiz\ ation )Tj T* (schemes. And this -- an explosion in free-form labeling of links, follow\ ed by all sorts of )Tj T* (ways of grabbing value from those labels -- is what I think is happening\ now. )Tj /TT2 1 Tf 0 -3.579 TD (Great Minds Don't Think Alike)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 250.648 218.138 m 259.137 218.138 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 250.6481 220.0391 Tm (#)Tj 0 0 0 rg ( )Tj -15.918 -3.579 Td (Here is del.icio.us, Joshua Shachter's social bookmarking service. It's \ for people who )Tj 0 -2.312 TD (are keeping track of their URLs for themselves, but who are willing to s\ hare globally a )Tj 0 -2.313 TD (view of what they're doing, creating an aggregate view of all users' boo\ kmarks, as well )Tj 0 -2.313 TD (as a personal view for each user. )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(23 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 277 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n q 476.1760406 0 0 290.4585724 43.1286926 523.5606537 cm /Im0 Do Q 0 0 0 rg BT /TT0 1 Tf 12.3225 0 0 12.3225 519.3047 668.7899 Tm ( )Tj -25.26 -13.098 Td ([ Front Page of del.icio.us ])Tj 13.2027 0 0 13.2027 40.4882 456.7843 Tm (As you can see here, the characteristics of a del.icio.us entry are a li\ nk, an optional )Tj 0 -2.313 TD (extended description, and a set of tags, which are words or phrases user\ s attach to a )Tj T* (link. Each user who adds a link to the system can give it a set of tags \ -- some do, some )Tj T* (don't. Attached to each link on the home page are the tags, the username\ of the person )Tj T* (who added it, the number of other people who have added that same link, \ and the time. )Tj 0 -3.579 TD (Tags are simply labels for URLs, selected to help the user in later retr\ ieval of those )Tj 0 -2.313 TD (URLs. Tags have the additional effect of grouping related URLs together.\ There is no )Tj T* (fixed set of categories or officially approved choices. You can use word\ s, acronyms, )Tj T* (numbers, whatever makes sense to you, without regard for anyone else's n\ eeds, )Tj T* (interests, or requirements. )Tj 0 -3.579 TD (The addition of a few simple labels hardly seems so momentous, but the s\ urprise here, )Tj 0 -2.312 TD (as so often with the Web, is the surprise of simplicity. Tags are import\ ant mainly for )Tj 0 -2.313 TD (what they leave out. By forgoing formal classification, tags enable a hu\ ge amount of )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(24 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 278 0 obj<> endobj 279 0 obj<> endobj 280 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 287.817 l 31.686 288.697 l 31.686 816.959 l h f 564.194 816.959 m 564.194 287.817 l 563.314 288.697 l 563.314 816.959 l h f 30.806 287.817 m 564.194 287.817 l 563.314 288.697 l 31.686 288.697 l h f EMC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.925 0.925 0.925 rg 40.488 297.499 514.024 39.333 re f EMC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 458.572 271.313 m 458.572 218.503 l 459.453 218.503 l 459.453 271.313 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 814.7201 Tm (you may like these links.'" )Tj 0 -3.579 TD (What it's doing instead is a lot simpler: "A lot of users tagging things\ foobar are also )Tj 0 -2.312 TD (tagging them frobnitz. I'll tell the user foobar and frobnitz are relate\ d." It's up to the )Tj 0 -2.313 TD (user to decide whether or not that recommendation is useful -- del.icio.\ us has no idea )Tj 0 -2.313 TD (what the tags )Tj /TT1 1 Tf (mean)Tj /TT0 1 Tf (. The tag overlap is in the system, but the tag semantics are in the )Tj 0 -2.312 TD (users. This is not a way to inject linguistic meaning into the machine. \ )Tj 0 -3.579 TD (It's all dependent on human context. This is what we're starting to see \ with del.icio.us, )Tj 0 -2.313 TD (with Flickr, with systems that are allowing for and aggregating tags. Th\ e signal benefit )Tj T* (of these systems is that they don't recreate the structured, hierarchica\ l categorization so )Tj T* (often forced onto us by our physical systems. Instead, we're dealing wit\ h a significant )Tj T* (break -- by letting users tag URLs and then aggregating those tags, we'r\ e going to be )Tj 0 -2.312 TD (able to build alternate organizational systems, systems that, like the W\ eb itself, do a )Tj 0 -2.313 TD (better job of letting individuals create value for one another, often wi\ thout realizing it. )Tj 0 -3.579 TD (Thank you very much. )Tj /TT2 1 Tf 0.333 -3.933 Td (Thanks to Alicia Cervini for invaluable editorial help.)Tj /TT3 1 Tf ( )Tj ET 0 0 1 RG 0.951 w 10 M 0 j 0 J []0 d 30.806 245.863 m 391.972 245.863 l S 0 0 1 rg BT /TT0 1 Tf 19.804 0 0 19.804 30.8062 248.7145 Tm (Clay Shirky's Writings About the Internet)Tj 0.4 0.4 0.4 rg 13.2027 0 0 13.2027 30.8062 229.182 Tm (Economics & Culture, Media & Community, Open Source)Tj ET EMC /Article <>BDC 0.634 w 480.929 248.361 m 573.876 248.361 l S 0 0 1 rg BT /TT4 1 Tf 13.2027 0 0 13.2027 480.929 250.262 Tm (clay@shirky.)Tj ET 543.074 230.757 m 573.876 230.757 l S BT /TT4 1 Tf 13.2027 0 0 13.2027 543.0739 232.6585 Tm (com)Tj 0 0 0 rg ( )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(34 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 281 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (user-produced organizational value, at vanishingly small cost. )Tj 0 -3.579 TD (There's a useful comparison here between gopher and the Web, where gophe\ r was )Tj 0 -2.312 TD (better organized, better mapped to existing institutional practices, and\ utterly unfit to )Tj 0 -2.313 TD (work at internet scale. The Web, by contrast, was and is a complete mess\ , with only one )Tj 0 -2.313 TD (brand of pointer, the URL, and no mechanism for global organization or r\ esources. The )Tj 0 -2.313 TD (Web is mainly notable for two things -- the way it ignored most of the t\ heories of )Tj T* (hypertext and rich metadata, and how much better it works than any of th\ e proposed )Tj T* (alternatives. \(The Yahoo/Google strategies I mentioned earlier also spl\ it along those )Tj T* (lines.\) )Tj 0 -3.579 TD (With those changes afoot, here are some of the things that I think are c\ oming, as )Tj 0 -2.313 TD (advantages of tagging systems: )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 414.5045 Tm (l)Tj /TT0 1 Tf ( )Tj /TT1 1 Tf 13.2027 0 0 13.2027 75.6953 414.5045 Tm (Market Logic)Tj /TT0 1 Tf ( - As we get used to the lack of physical constraints, as we )Tj 0 -2.313 TD (internalize the fact that there is no shelf and there is no disk, we're \ moving )Tj T* (towards market logic, where you deal with individual motivation, but gro\ up )Tj T* (value. )Tj 0 -3.579 TD (As Schachter says of del.icio.us, "Each individual categorization scheme\ is worth )Tj 0 -2.313 TD (less than a professional categorization scheme. But there are many, many\ more of )Tj 0 -2.313 TD (them." If you find a way to make it valuable to individuals to tag their\ stuff, you'll )Tj T* (generate a lot more data about any given object than if you pay a profes\ sional to )Tj T* (tag it once and only once. And if you can find any way to create value f\ rom )Tj T* (combining myriad amateur classifications over time, they will come to be\ more )Tj T* (valuable than professional categorization schemes, particularly with reg\ ards to )Tj T* (robustness and cost of creation. )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(25 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 282 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 75.6953 797.7993 Tm (The other essential value of market logic is that individual differences\ don't have )Tj 0 -2.313 TD (to be homogenized. Look for the word 'queer' in almost any top-level )Tj T* (categorization. You will not find it, even though, as an organizing prin\ ciple for a )Tj T* (large group of people, that word matters enormously. Users don't get to \ )Tj T* (participate those kind of discussions around traditional categorization \ schemes, )Tj 0 -2.313 TD (but with tagging, anyone is free to use the words he or she thinks are a\ ppropriate, )Tj 0 -2.312 TD (without having to agree with anyone else about how something "should" be\ )Tj 0 -2.313 TD (tagged. Market logic allows many distinct points of view to co-exist, be\ cause it )Tj 0 -2.313 TD (allows individuals to preserve their point of view, even in the face of \ general )Tj T* (disagreement. )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 475.7643 Tm (l)Tj /TT0 1 Tf ( )Tj /TT1 1 Tf 13.2027 0 0 13.2027 75.6953 475.7643 Tm (User and Time are Core Attributes)Tj /TT0 1 Tf ( - This is absolutely essential. The )Tj 0 -2.313 TD (attitude of the Yahoo ontologist and her staff was -- "We are Yahoo We d\ o not )Tj T* (have biases. This is just how the world is. The world is organized into \ a dozen )Tj T* (categories." You don't know who those people were, where they came from,\ what )Tj T* (their background was, what their political biases might be. )Tj 0 -3.579 TD (Here, because you can derive 'this is who this link is was tagged by' an\ d 'this is )Tj 0 -2.313 TD (when it was tagged, you can start to do inclusion and exclusion around p\ eople )Tj T* (and time, not just tags. You can start to do grouping. You can start to \ do decay. )Tj T* ("Roll up tags from just this group of users, I'd like to see what they a\ re talking )Tj T* (about" or "Give me all tags with this signature, but anything that's mor\ e than a )Tj T* (week old or a year old." )Tj 0 -3.579 TD (This is group tagging -- not the entire population, and not just me. It'\ s like Unix )Tj 0 -2.313 TD (permissions -- right now we've got tags for user and world, and this is \ the base on )Tj T* (which we will be inventing group tags. We're going to start to be able t\ o subset )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(26 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 283 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 75.6953 797.9968 Tm (our categorization schemes. Instead of having massive categorizations an\ d then )Tj 0 -2.313 TD (specialty categorization, we're going to have a spectrum between them, b\ ased on )Tj T* (the size and make-up of various tagging groups. )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 689.6799 Tm (l)Tj /TT0 1 Tf ( )Tj /TT1 1 Tf 13.2027 0 0 13.2027 75.6953 689.6799 Tm (Signal Loss from Expression)Tj /TT0 1 Tf ( - The signal loss in traditional categorization )Tj T* (schemes comes from compressing things into a restricted number of catego\ ries. )Tj T* (With tagging, when there is signal loss, it comes from people not having\ any )Tj T* (commonality in talking about things. The loss is from the multiplicity o\ f points of )Tj T* (view, rather than from compression around a single point of view. But in\ a world )Tj 0 -2.313 TD (where enough points of view are likely to provide some commonality, the \ )Tj T* (aggregate signal loss falls with scale in tagging systems, while it grow\ s with scale )Tj 0 -2.312 TD (in systems with single points of view. )Tj 0 -3.579 TD (The solution to this sort of signal loss is growth. Well-managed, well-g\ roomed )Tj 0 -2.313 TD (organizational schemes get worse with scale, both because the costs of s\ upporting )Tj T* (such schemes at large volumes are prohibitive, and, as I noted earlier, \ scaling )Tj T* (over time is also a serious problem. Tagging, by contrast, gets better w\ ith scale. )Tj T* (With a multiplicity of points of view the question isn't "Is everyone ta\ gging any )Tj T* (given link 'correctly'", but rather "Is anyone tagging it the way I do?"\ As long as at )Tj T* (least one other person tags something they way you would, you'll find it\ -- using a )Tj 0 -2.312 TD (thesaurus to force everyone's tags into tighter synchrony would actually\ worsen )Tj 0 -2.313 TD (the noise you'll get with your signal. If there is no shelf, then even )Tj /TT2 1 Tf (imagining)Tj /TT0 1 Tf ( that )Tj 0 -2.312 TD (there is one right way to organize things is an error. )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 106.6723 Tm (l)Tj /TT0 1 Tf ( )Tj /TT1 1 Tf 13.2027 0 0 13.2027 75.6953 106.6723 Tm (The Filtering is Done Post Hoc)Tj /TT0 1 Tf ( - There's an analogy here with every )Tj 0 -2.313 TD (journalist who has ever looked at the Web and said "Well, it needs an ed\ itor." The )Tj T* (Web has an editor, it's everybody. In a world where publishing is expens\ ive, the )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(27 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 284 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 75.6953 798.1942 Tm (act of publishing is also a statement of quality -- the filter comes bef\ ore the )Tj 0 -2.313 TD (publication. In a world where publishing is cheap, putting something out\ there )Tj T* (says nothing about its quality. It's what happens after it gets publishe\ d that )Tj T* (matters. If people don't point to it, other people won't read it. But th\ e idea that )Tj T* (the filtering is )Tj /TT1 1 Tf (after)Tj /TT0 1 Tf ( the publishing is incredibly foreign to journalists. )Tj 0 -3.579 TD (Similarly, the idea that the categorization is done after things are tag\ ged is )Tj 0 -2.313 TD (incredibly foreign to cataloguers. Much of the expense of existing catal\ ogue )Tj T* (systems is in trying to prevent one-off categories. With tagging, what y\ ou say is )Tj T* ("As long as a lot of people are tagging any given link, the rare tags ca\ n be used or )Tj T* (ignored, as the user likes. We won't even have to expend the cost to pre\ vent )Tj T* (people from using them. We'll just help other users ignore them if they \ want to." )Tj 0 -3.579 TD (Again, scale comes to the rescue of the system in a way that would simpl\ y break )Tj 0 -2.313 TD (traditional cataloging schemes. The existence of an odd or unusual tag i\ s a )Tj T* (problem if it's the only way a given link has been tagged, or if there i\ s no way for a )Tj 0 -2.313 TD (user to avoid that tag. Once a link has been tagged more than once, thou\ gh, users )Tj T* (can view or ignore the odd tags as it suits them, and the decision about\ which tags )Tj T* (to use comes after the links have been tagged, not before. )Tj /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 228.9944 Tm (l)Tj /TT0 1 Tf ( )Tj /TT2 1 Tf 13.2027 0 0 13.2027 75.6953 228.9944 Tm (Merged from URLs, Not Categories)Tj /TT0 1 Tf ( - You don't merge tagging schemes at )Tj 0 -2.313 TD (the category level and then see what the contents are. As with the 'merg\ ing )Tj T* (ISBNs' idea, you merge individual contents, because we now have URLs as \ )Tj T* (unique handles. You merge from the URLs, and then try and derive somethi\ ng )Tj T* (about the categorization from there. This allows for partial, incomplete\ , or )Tj 0 -2.313 TD (probabilistic merges that are better fits to uncertain environments -- s\ uch as the )Tj T* (real world -- than rigid classification schemes. )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(28 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 285 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /T1_1 1 Tf 6.6013 0 0 6.6013 62.519 781.6684 Tm (l)Tj /TT0 1 Tf ( )Tj /TT1 1 Tf 13.2027 0 0 13.2027 75.6953 781.6684 Tm (Merges are Probabilistic, not Binary)Tj /TT0 1 Tf ( - Merges create partial overlap )Tj 0 -2.313 TD (between tags, rather than defining tags as synonyms. Instead of saying t\ hat any )Tj T* (given tag "is" or "is not" the same as another tag, del.icio.us is able \ to recommend )Tj T* (related tags by saying "A lot of people who tagged this 'Mac' also tagge\ d it 'OSX'." )Tj T* (We move from a binary choice between saying two tags are the same or dif\ ferent )Tj 0 -2.313 TD (to the Venn diagram option of "kind of is/somewhat is/sort of is/overlap\ s to this )Tj 0 -2.312 TD (degree". That is a really profound change. )Tj /TT1 1 Tf -2.667 -3.579 Td (Tag Distributions on del.icio.us)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 256.629 549.326 m 265.118 549.326 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 256.629 551.2269 Tm (#)Tj 0 0 0 rg ( )Tj -16.371 -3.579 Td (Here's something showing what I mean about the breakdown of binary categ\ orization. )Tj ET q 484.9778137 0 0 323.0251465 43.1286926 147.4933472 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 528.1065 309.0059 Tm ( )Tj -23.25 -14.419 Td ([ Tags per user ] )Tj 13.2027 0 0 13.2027 40.4882 80.717 Tm (This is a chart based on a small sample of links from the del.icio.us fr\ ont page, taken )Tj 0 -2.313 TD (during a 2-hour window. The X axis is the 64 users who posted links duri\ ng that period. )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(29 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 286 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (The Y axis is the total number of discrete kinds of tags that those user\ s have ever used )Tj 0 -2.313 TD (in their history on del.icio.us. )Tj 0 -3.579 TD (The chart shows a great variability in tagging strategies among the vari\ ous users. The )Tj 0 -2.313 TD (user all the way to the left has an enormous number of unique tags, almo\ st 600 of )Tj T* (them. Then there's this group of people who are not quite power taggers \ but who tag )Tj T* (quite a bit, and of course to the right of them there's the characterist\ ic long tail of )Tj T* (people who use many fewer tags than the power taggers. \(Because this is\ a two-hour )Tj T* (snapshot, it has a natural bias towards frequent del.icio.us users. I'm \ trying to get a )Tj T* (larger data set. My guess is the tail goes out quite a bit further than \ this.\) But this is )Tj T* (what organization looks like when you turn it over to the users -- many \ different )Tj T* (strategies, each of which works in its own context, but which can also b\ e merged. )Tj ET q 508.7425995 0 0 273.7352142 43.1286926 171.2933655 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 551.8713 308.161 Tm ( )Tj -25.662 -12.419 Td ( [ A single user's tags ] )Tj 13.2027 0 0 13.2027 40.4882 104.517 Tm (This is a single user's tags. From here, you can tell something about th\ is person -- he or )Tj 0 -2.313 TD (she is obviously a Flash programmer -- the commonest tag here is Flash, \ followed by a )Tj T* (number of other frequently used tags mainly related to programming. Like\ the front )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(30 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 287 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 796.0389 Tm (page, this distribution has the organic signature. Experts don't catalog\ this way; experts )Tj 0 -2.313 TD (who learn how to catalogue produce much more consistent labeling. Here, \ it's whatever )Tj T* (the user thought would help them remember the link later. )Tj 0 -3.579 TD (You can see there's a tag "to_read". A professional cataloguer would loo\ k at this tag in )Tj 0 -2.313 TD (horror -- "This is context-dependent and temporary." Well, so was the ca\ tegory "East )Tj T* (Germany." Once you expand your time scale to include the actual life of \ the )Tj T* (categorization scheme itself, you recognize that the distinction between\ temporary and )Tj T* (permanent is awfully vague. There isn't in fact a binary condition of a \ tag that can or )Tj T* (cannot survive any kind of long-term examination. )Tj ET q 462.0932007 0 0 345.9097595 43.1286926 155.7026672 cm /Im0 Do Q BT /TT0 1 Tf 12.3225 0 0 12.3225 505.2219 328.6575 Tm ( )Tj -28.95 -15.348 Td ([ Different tag 'signatures' for different URLs ])Tj 13.2027 0 0 13.2027 40.4882 88.9263 Tm (Then there's this set of graphs. This is to me in a way the most interes\ ting and least well )Tj 0 -2.313 TD (understood part of the del.icio.us right now -- these are two different \ URLs and the tags )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(31 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 288 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 800.5174 Tm (that a whole group of users applied to them. The graph at the bottom lef\ t refers to a site )Tj 0 -2.313 TD (for downloading old versions of programs that are no longer supported. Y\ ou can see )Tj T* (here that there is broad communal consensus. 140 people tagged this Soft\ ware. Then, )Tj T* (the next commonest tag, with only 20 occurrences, is Windows, then Old, \ then )Tj T* (Download, and so forth. For this URL, there's a core consensus -- this l\ ink is about )Tj 0 -2.313 TD (software -- and after that one bit of commonality, there is a really sha\ rp, clear fall off in )Tj T* (tags. )Tj 0 -3.579 TD (The graph at the upper right, by contrast, shows the tags for a page det\ ailing how to )Tj 0 -2.313 TD (embed standing searches in Gmail. You can see the tags -- Gmail, Firefox\ , Search, )Tj T* (Javascript, GreaseMonkey -- this is a much smearier distribution, with a\ much less )Tj T* (sharp fall-off. The consensus view is that this link is about more kinds\ of things than )Tj T* (the software download link is, or, rather, occupies more contexts for de\ l.icio.us users )Tj T* (than the software download link does. )Tj 0 -3.579 TD (Looking at this sort of data, we can start to say, of particular URLs, t\ hat the users )Tj 0 -2.313 TD (tagging this URL either did or did not center around a certain core tags\ , with this )Tj T* (degree of certainty, and, thanks to the time stamps, we can even start t\ o understand )Tj T* (how the distribution of a URLs tags changes over time. It was 5 years be\ tween the )Tj T* (spread of the link and Google's figuring out how to use whole collection\ s of links to )Tj T* (create additional value. We're early in the use of tags, so we don't yet\ have large, long-)Tj T* (lived data sets to look at, but they are being built up quickly, and we'\ re just figuring out )Tj T* (how to extract novel value from whole collections of tags. )Tj /TT1 1 Tf 0 -3.579 TD (Organization Goes Organic)Tj /TT0 1 Tf ( )Tj ET 0 0 1 RG 0.634 w 10 M 0 j 0 J []0 d 226.91 107.292 m 235.399 107.292 l S 0 0 1 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 226.9098 109.1929 Tm (#)Tj 0 0 0 rg ( )Tj -14.12 -3.579 Td (We are moving away from binary categorization -- books either are or are\ not )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(32 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 289 0 obj<>stream /Artifact <>BDC 0 0 0 rg 0 i BT /T1_0 1 Tf 0 Tc 0 Tw 0 Ts 100 Tz 0 Tr 7.9216 0 0 7.9216 18 830.9884 Tm (Shirky: Ontology is Overrated -- Categories, Links, and Tags)Tj ET EMC /WebCaptureBG BMC /WebCaptureFN <>BDC q 0 18 595 806 re W* n /Artifact <>BDC Q 0.6 0.6 0.6 rg 30.806 816.959 m 30.806 33.843 l 31.686 33.843 l 31.686 816.959 l h f 564.194 816.959 m 564.194 33.843 l 563.314 33.843 l 563.314 816.959 l h f EMC EMC EMC /Article <>BDC q 0 18 595 806 re W* n 0 0 0 rg BT /TT0 1 Tf 13.2027 0 0 13.2027 40.4882 814.5226 Tm (entertainment -- and into this probabilistic world, where N% of users th\ ink books are )Tj 0 -2.313 TD (entertainment. It may well be that within Yahoo, there was a big debate \ about whether )Tj T* (or not books are entertainment. But they either had no way of reflecting\ that debate or )Tj T* (they decided not to expose it to the users. What instead happened was it\ became an all-)Tj T* (or-nothing categorization, "This is entertainment, this is not entertain\ ment." We're )Tj 0 -2.313 TD (moving away from that sort of absolute declaration, and towards being ab\ le to roll up )Tj 0 -2.312 TD (this kind of value by observing how people handle it in practice. )Tj 0 -3.579 TD (It comes down ultimately to a question of philosophy. Does the world mak\ e sense or do )Tj 0 -2.313 TD (we make sense of the world? If you believe the world makes sense, then a\ nyone who )Tj T* (tries to make sense of the world differently than you is presenting you \ with a situation )Tj T* (that needs to be reconciled formally, because if you get it wrong, you'r\ e getting it wrong )Tj T* (about the real world. )Tj 0 -3.579 TD (If, on the other hand, you believe that we make sense of the world, if w\ e are, from a )Tj 0 -2.313 TD (bunch of different points of view, applying some kind of sense to the wo\ rld, then you )Tj T* (don't privilege one top level of sense-making over the other. What you d\ o instead is you )Tj T* (try to find ways that the individual sense-making can roll up to somethi\ ng which is of )Tj T* (value in aggregate, but you do it without an ontological goal. You do it\ without a goal of )Tj T* (explicitly getting to or even closely matching some theoretically perfec\ t view of the )Tj T* (world. )Tj 0 -3.579 TD (Critically, the semantics here are in the users, not in the system. This\ is not a way to get )Tj 0 -2.313 TD (computers to understand things. When del.icio.us is recommending tags to\ me, the )Tj T* (system is not saying, "I know that OSX is an operating system. Therefore\ , I can use )Tj T* (predicate logic to come up with recommendations -- users run software, s\ oftware runs )Tj T* (on operating systems, OSX is a type of operating system -- and then say \ 'Here Mr. User, )Tj ET EMC /Artifact <>BDC Q 0 0 0 rg BT /T1_0 1 Tf 7.9216 0 0 7.9216 18 7.9884 Tm (http://www.shirky.com/writings/ontology_overrated.html \(33 de 34\)28/01\ /2008 19:04:54)Tj ET EMC endstream endobj 290 0 obj(Shirky: Ontology is Overrated -- Categories, Links, and Tags) endobj 291 0 obj<> endobj 292 0 obj<> endobj 293 0 obj<> endobj 294 0 obj<> endobj 295 0 obj[292 0 R] endobj 296 0 obj(http://www.shirky.com/writings/ontology_overrated.html) endobj 297 0 obj(=&hIk H\() endobj 298 0 obj<> endobj 299 0 obj<> endobj 300 0 obj($0~r_c-r[) endobj 301 0 obj 1 endobj 302 0 obj 1 endobj 303 0 obj 1 endobj 304 0 obj 1 endobj 305 0 obj 1 endobj 306 0 obj 1 endobj 307 0 obj 1 endobj 308 0 obj 1 endobj 309 0 obj 1 endobj 310 0 obj 1 endobj 311 0 obj 1 endobj 312 0 obj 1 endobj 313 0 obj 1 endobj 314 0 obj 1 endobj 315 0 obj<> endobj 316 0 obj<> endobj 317 0 obj<>stream 2008-01-28T19:04:54+01:00 2008-01-28T19:04:51+01:00 2008-01-28T19:04:54+01:00 application/pdf Shirky: Ontology is Overrated -- Categories, Links, and Tags uuid:a5934fd0-2fbf-41cc-b8c8-916614de598a uuid:50b400d3-495e-4054-a30d-9333e3ba8be1 Acrobat Web Capture 7.0 endstream endobj xref 0 318 0000000004 65535 f 0000000016 00000 n 0000000146 00000 n 0000000239 00000 n 0000000006 00000 f 0000000431 00000 n 0000000131 00001 f 0000000499 00000 n 0000058453 00000 n 0000058529 00000 n 0000058562 00000 n 0000058648 00000 n 0000058698 00000 n 0000058754 00000 n 0000058992 00000 n 0000059230 00000 n 0000108993 00000 n 0000109065 00000 n 0000109099 00000 n 0000109187 00000 n 0000109238 00000 n 0000157121 00000 n 0000157203 00000 n 0000157238 00000 n 0000157326 00000 n 0000157377 00000 n 0000196021 00000 n 0000196095 00000 n 0000196129 00000 n 0000196217 00000 n 0000196268 00000 n 0000227093 00000 n 0000227171 00000 n 0000227205 00000 n 0000227293 00000 n 0000227344 00000 n 0000263620 00000 n 0000263704 00000 n 0000263738 00000 n 0000263826 00000 n 0000263877 00000 n 0000313121 00000 n 0000313210 00000 n 0000313245 00000 n 0000313333 00000 n 0000313384 00000 n 0000357778 00000 n 0000357857 00000 n 0000357892 00000 n 0000357980 00000 n 0000358031 00000 n 0000370619 00000 n 0000370699 00000 n 0000370733 00000 n 0000370821 00000 n 0000370872 00000 n 0000383624 00000 n 0000383709 00000 n 0000383743 00000 n 0000383831 00000 n 0000383882 00000 n 0000450280 00000 n 0000450352 00000 n 0000450386 00000 n 0000450474 00000 n 0000450525 00000 n 0000487080 00000 n 0000487158 00000 n 0000487192 00000 n 0000487280 00000 n 0000487331 00000 n 0000526550 00000 n 0000526627 00000 n 0000526661 00000 n 0000526749 00000 n 0000526800 00000 n 0000571453 00000 n 0000571530 00000 n 0000571564 00000 n 0000571652 00000 n 0000571703 00000 n 0000571807 00000 n 0000571852 00000 n 0000571901 00000 n 0000573084 00000 n 0000573331 00000 n 0000573718 00000 n 0000573749 00000 n 0000573900 00000 n 0000573940 00000 n 0000574163 00000 n 0000574187 00000 n 0000574868 00000 n 0000574911 00000 n 0000575198 00000 n 0000575229 00000 n 0000575272 00000 n 0000575523 00000 n 0000575759 00000 n 0000575783 00000 n 0000575826 00000 n 0000576105 00000 n 0000576131 00000 n 0000576370 00000 n 0000576395 00000 n 0000576646 00000 n 0000576671 00000 n 0000576715 00000 n 0000576978 00000 n 0000577004 00000 n 0000577240 00000 n 0000577272 00000 n 0000577513 00000 n 0000577545 00000 n 0000577755 00000 n 0000577780 00000 n 0000578005 00000 n 0000578030 00000 n 0000578075 00000 n 0000578344 00000 n 0000578370 00000 n 0000578598 00000 n 0000578623 00000 n 0000578891 00000 n 0000578990 00000 n 0000579086 00000 n 0000580292 00000 n 0000580515 00000 n 0000581731 00000 n 0000581960 00000 n 0000583172 00000 n 0000000137 00001 f 0000583396 00000 n 0000583421 00000 n 0000583466 00000 n 0000583739 00000 n 0000583765 00000 n 0000000142 00001 f 0000584005 00000 n 0000584030 00000 n 0000584259 00000 n 0000585472 00000 n 0000000145 00001 f 0000585704 00000 n 0000585729 00000 n 0000000150 00001 f 0000585970 00000 n 0000585995 00000 n 0000586246 00000 n 0000586338 00000 n 0000000157 00001 f 0000586425 00000 n 0000586450 00000 n 0000586495 00000 n 0000586798 00000 n 0000586824 00000 n 0000587048 00000 n 0000000160 00001 f 0000587147 00000 n 0000587172 00000 n 0000000163 00001 f 0000587396 00000 n 0000587421 00000 n 0000000169 00001 f 0000587659 00000 n 0000587684 00000 n 0000587729 00000 n 0000587973 00000 n 0000587999 00000 n 0000000172 00001 f 0000588210 00000 n 0000588235 00000 n 0000000179 00001 f 0000588474 00000 n 0000588499 00000 n 0000588543 00000 n 0000588794 00000 n 0000588820 00000 n 0000589059 00000 n 0000000186 00001 f 0000589158 00000 n 0000589183 00000 n 0000589228 00000 n 0000589473 00000 n 0000589507 00000 n 0000589732 00000 n 0000000189 00001 f 0000589973 00000 n 0000589998 00000 n 0000000195 00001 f 0000590266 00000 n 0000590291 00000 n 0000590336 00000 n 0000590613 00000 n 0000590639 00000 n 0000000199 00001 f 0000590868 00000 n 0000590893 00000 n 0000591117 00000 n 0000000203 00001 f 0000591184 00000 n 0000591209 00000 n 0000591433 00000 n 0000000206 00001 f 0000591532 00000 n 0000591557 00000 n 0000000209 00001 f 0000591793 00000 n 0000591818 00000 n 0000000212 00001 f 0000592054 00000 n 0000592079 00000 n 0000000218 00001 f 0000592348 00000 n 0000592373 00000 n 0000592418 00000 n 0000592703 00000 n 0000592729 00000 n 0000000222 00001 f 0000592958 00000 n 0000592983 00000 n 0000593212 00000 n 0000000225 00001 f 0000593343 00000 n 0000593368 00000 n 0000000231 00001 f 0000593594 00000 n 0000593619 00000 n 0000593664 00000 n 0000593934 00000 n 0000593960 00000 n 0000000234 00001 f 0000594159 00000 n 0000594184 00000 n 0000000238 00001 f 0000594446 00000 n 0000594478 00000 n 0000594631 00000 n 0000000243 00001 f 0000594673 00000 n 0000594826 00000 n 0000594979 00000 n 0000595132 00000 n 0000000244 00001 f 0000000245 00001 f 0000000246 00001 f 0000000247 00001 f 0000000248 00001 f 0000000249 00001 f 0000000250 00001 f 0000000000 00001 f 0000595285 00000 n 0000595376 00000 n 0000596596 00000 n 0000596826 00000 n 0000600707 00000 n 0000604098 00000 n 0000606698 00000 n 0000609345 00000 n 0000612103 00000 n 0000614410 00000 n 0000616896 00000 n 0000620046 00000 n 0000622778 00000 n 0000625287 00000 n 0000628008 00000 n 0000630093 00000 n 0000632552 00000 n 0000635202 00000 n 0000638419 00000 n 0000641862 00000 n 0000645176 00000 n 0000648395 00000 n 0000651592 00000 n 0000654618 00000 n 0000657758 00000 n 0000660378 00000 n 0000663261 00000 n 0000665638 00000 n 0000666857 00000 n 0000667091 00000 n 0000670529 00000 n 0000673568 00000 n 0000676766 00000 n 0000680179 00000 n 0000683472 00000 n 0000685909 00000 n 0000688422 00000 n 0000690666 00000 n 0000693877 00000 n 0000697082 00000 n 0000697161 00000 n 0000697199 00000 n 0000697290 00000 n 0000697321 00000 n 0000697425 00000 n 0000697451 00000 n 0000697524 00000 n 0000697560 00000 n 0000697938 00000 n 0000697991 00000 n 0000698026 00000 n 0000698045 00000 n 0000698064 00000 n 0000698083 00000 n 0000698102 00000 n 0000698121 00000 n 0000698140 00000 n 0000698159 00000 n 0000698178 00000 n 0000698197 00000 n 0000698216 00000 n 0000698235 00000 n 0000698254 00000 n 0000698273 00000 n 0000698292 00000 n 0000698379 00000 n 0000698521 00000 n trailer < <1F19F8A9E73D414F8E8E319B283E7471>]>> startxref 702033 %%EOF

Clearview AI Offered Free Facial Recognition Trials to Police Around the World

Law enforcement agencies and government organizations from 24 countries outside the United States used a controversial facial recognition technology called Clearview AI, according to internal company data reviewed by BuzzFeed News. That data, which runs up until February 2020, shows that police departments, prosecutors’ offices, universities, and interior ministries from around the world ran nearly 14,000 searches with Clearview AI’s software. At many law enforcement agencies from Canada to Finland, officers used the software without their higher-ups’ knowledge or permission. After receiving questions from BuzzFeed News, some organizations admitted that the technology had been used without leadership oversight. In March, a BuzzFeed News investigation based on Clearview AI’s own internal data showed how the New York–based startup distributed its facial recognition tool, by marketing free trials for its mobile app or desktop software, to thousands of officers and employees at more than 1,800 US taxpayer-funded entities. Clearview claims its software is more accurate than other facial recognition technologies because it is trained on a database of more than 3 billion images scraped from websites and social media platforms, including Facebook, Instagram, LinkedIn, and Twitter. Law enforcement officers using Clearview can take a photo of a suspect or person of interest, run it through the software, and receive possible matches for that individual within seconds. Clearview has claimed that its app is 100% accurate in documents provided to law enforcement officials, but BuzzFeed News has seen the software misidentify people, highlighting a larger concern with facial recognition technologies. Based on new reporting and data reviewed by BuzzFeed News, Clearview AI took its controversial US marketing playbook around the world, offering free trials to employees at law enforcement agencies in countries including Australia, Brazil, and the United Kingdom. To accompany this story, BuzzFeed News has created a searchable table of 88 international government-affiliated and taxpayer-funded agencies and organizations listed in Clearview’s data as having employees who used or tested the company’s facial recognition service before February 2020, according to Clearview’s data. Some of those entities were in countries where the use of Clearview has since been deemed “unlawful.” Following an investigation, Canada’s data privacy commissioner ruled in February 2021 that Clearview had “violated federal and provincial privacy laws”; it recommended the company stop offering its services to Canadian clients, stop collecting images of Canadians, and delete all previously collected images and biometrics of people in the country. In the European Union, authorities are assessing whether the use of Clearview violated the General Data Protection Regulation (GDPR), a set of broad online privacy laws that requires companies processing personal data to obtain people’s informed consent. The Dutch Data Protection Authority told BuzzFeed News that it’s “unlikely” that police agencies’ use of Clearview was lawful, while France’s National Commission for Informatics and Freedoms said that it has received “several complaints” about Clearview that are “currently being investigated.” One regulator in Hamburg has already deemed the company’s practices illegal under the GDPR and asked it to delete information on a German citizen. Despite Clearview being used in at least two dozen other countries, CEO Hoan Ton-That insists the company’s key market is the US. “While there has been tremendous demand for our service from around the world, Clearview AI is primarily focused on providing our service to law enforcement and government agencies in the United States,” he said in a statement to BuzzFeed News. “Other countries have expressed a dire need for our technology because they know it can help investigate crimes, such as, money laundering, financial fraud, romance scams, human trafficking, and crimes against children, which know no borders.” In the same statement, Ton-That alleged there are “inaccuracies contained in BuzzFeed’s assertions.” He declined to explain what those might be and did not answer a detailed list of questions based on reporting for this story. Clearview AI has created a powerful facial recognition tool and marketed it to police departments and government agencies. The company has never disclosed the entities that have used its facial recognition software, but a confidential source provided BuzzFeed News with data that appeared to be a list of agencies and companies whose employees have tested or actively used its technology. Using that data, along with public records and interviews, we have created a searchable database of internationally based taxpayer-funded entities, including law enforcement agencies, prosecutor’s offices, universities, and interior ministries. We have included only those agencies for which the data shows that at least one associated individual ran at least one facial recognition scan as of February 2020. The database has limitations. Clearview has neither verified nor disputed the underlying data, which The data begins in 2018 and ends in February 2020, so it does not account for any activity after that time or for any additional organizations that may have started using Clearview after February 2020. Not all searches corresponded to an investigation, and some agencies told us that their employees had merely run test searches to see how well the technology worked. BuzzFeed News created search ranges based on data that showed how many times individuals at a given organization ran photos through Clearview. We found inaccuracies in the data, including organizations with misspelled or incomplete names, and we moved to correct those issues when they could be confirmed. If we were not able to confirm the existence of an entity, we removed it. BuzzFeed News gave every agency or organization in this database the opportunity to comment on whether it had used Clearview’s technology and whether the software had led to any arrests. Of the 88 entities in this database: 36 said they had employees who used or tried Clearview AI. Officials at 9 of those organizations said they were unaware that their employees had signed up for free trials until questions from BuzzFeed News or our reporting partners prompted them to look. Officials at another 3 entities at first denied their employees had used Clearview but later determined that some of them had. 10 entities declined to answer questions as to whether their employees had used Clearview. 12 organizations denied any use of Clearview. 30 organizations did not respond to requests for comment. Responses from the agencies, including whether they denied using Clearview’s technology or did not respond to requests for comment, are included in the table. Just because an agency appears on the list does not mean BuzzFeed News was able to confirm that it actually used the tool or that its officials approved its employees’ use of Clearview. By searching this database, you affirm that you understand its limitations. According to a 2019 internal document first reported by BuzzFeed News, Clearview had planned to pursue “rapid international expansion” into at least 22 countries. But by February 2020, the company’s strategy appeared to have shifted. “Clearview is focused on doing business in the USA and Canada,” Ton-That told BuzzFeed News at that time. Two weeks later, in an interview on PBS, he clarified that Clearview would never sell its technology to countries that “are very adverse to the US,” before naming China, Russia, Iran, and North Korea. Since that time, Clearview has become the subject of media scrutiny and multiple government investigations. In July, following earlier reporting from BuzzFeed News that showed that private companies and public organizations had run Clearview searches in Great Britain and Australia, privacy commissioners in those countries opened a joint inquiry into the company for its use of personal data. The investigation is ongoing, according to the UK's Information Commissioner’s Office, which told BuzzFeed News that “no further comment will be made until it is concluded.” Canadian authorities also moved to regulate Clearview after the Toronto Star, in partnership with BuzzFeed News, reported on the widespread use of the company’s software in the country. In February 2020, federal and local Canadian privacy commissioners launched an investigation into Clearview, and concluded that it represented a “clear violation of the privacy rights of Canadians.” Earlier this year, those bodies officially declared Clearview’s practices in the country illegal and recommended that the company stop offering its technology to Canadian clients. Clearview disagreed with the findings of the investigation and did not demonstrate a willingness to follow the other recommendations, according to the Office of the Privacy Commissioner of Canada. Prior to that declaration, employees from at least 41 entities within the Canadian government — the most of any country outside the US — were listed in internal data as having used Clearview. Those agencies ranged from police departments in midsize cities like Timmins, a 41,000-person city where officers ran more than 120 searches, to major metropolitan law enforcement agencies like the Toronto Police Service, which is listed in the data as having run more than 3,400 searches as of February 2020. A spokesperson for the Timmins Police Service acknowledged that the department had used Clearview but said no arrests were ever made on the basis of a search with the technology. The Toronto Police Service did not respond to multiple requests for comment. Clearview’s data show that usage was not limited to police departments. The public prosecutions office at the Saskatchewan Ministry of Justice ran more than 70 searches with the software. A spokesperson initially said that employees had not used Clearview but changed her response after a series of follow-up questions. “The Crown has not used Clearview AI to support a prosecution.” “After review, we have identified standalone instances where ministry staff did use a trial version of this software,” Margherita Vittorelli, a ministry spokesperson, said. “The Crown has not used Clearview AI to support a prosecution. Given the concerns around the use of this technology, ministry staff have been instructed not to use Clearview AI’s software at this time.” Some Canadian law enforcement agencies suspended or discontinued their use of Clearview AI not long after the initial trial period or stopped using it in response to the government investigation. One detective with the Niagara Regional Police Service’s Technological Crimes Unit conducted more than 650 searches on a free trial of the software, according to the data. “Once concerns surfaced with the Privacy Commissioner, the usage of the software was terminated,” department spokesperson Stephanie Sabourin told BuzzFeed News. She said the detective used the software in the course of an undisclosed investigation without the knowledge of senior officers or the police chief. The Royal Canadian Mounted Police was among the very few international agencies that had contracted with Clearview and paid to use its software. The agency, which ran more than 450 searches, said in February 2020 that it used the software in 15 cases involving online child sexual exploitation, resulting in the rescue of two children. In June, however, the Office of the Privacy Commissioner in Canada found that RCMP’s use of Clearview violated the country’s privacy laws. The office also found that Clearview had “violated Canada’s federal private sector privacy law by creating a databank of more than three billion images scraped from internet websites without users’ consent.” The RCMP disputed that conclusion. The Canadian Civil Liberties Association, a nonprofit group, said that Clearview had facilitated “unaccountable police experimentation” within Canada. “Clearview AI’s business model, which scoops up photos of billions of ordinary people from across the internet and puts them in a perpetual police lineup, is a form of mass surveillance that is unlawful and unacceptable in our democratic, rights-respecting nation,” Brenda McPhail, director of the CCLA’s privacy, technology, and surveillance program, told BuzzFeed News. Like a number of American law enforcement agencies, some international agencies told BuzzFeed News that they couldn’t discuss their use of Clearview. For instance, Brazil’s Public Ministry of Pernambuco, which is listed as having run more than 100 searches, said that it “does not provide information on matters of institutional security.” But data reviewed by BuzzFeed News shows that individuals at nine Brazilian law enforcement agencies, including the country’s federal police, are listed as having used Clearview, cumulatively running more than 1,250 searches as of February 2020. All declined to comment or did not respond to requests for comment. The UK’s National Crime Agency, which ran more than 500 searches, according to the data, declined to comment on its investigative techniques; a spokesperson told BuzzFeed News in early 2020 that the organization “deploys numerous specialist capabilities to track down online offenders who cause serious harm to members of the public.” Employees at the country’s Metropolitan Police Service ran more than 150 searches on Clearview, according to internal data. When asked about the department's use of the service, the police force declined to comment. Documents reviewed by BuzzFeed News also show that Clearview had a fledgling presence in Middle Eastern countries known for repressive governments and human rights concerns. In Saudi Arabia, individuals at the Artificial Intelligence Center of Advanced Studies (also known as Thakaa) ran at least 10 searches with Clearview. In the United Arab Emirates, people associated with Mubadala Investment Company, a sovereign wealth fund in the capital of Abu Dhabi, ran more than 100 searches, according to internal data. Thakaa did not respond to multiple requests for comment. A Mubadala spokesperson told BuzzFeed News that the company does not use the software at any of its facilities. Data revealed that individuals at four different Australian agencies tried or actively used Clearview, including the Australian Federal Police (more than 100 searches) and Victoria Police (more than 10 searches), where a spokesperson told BuzzFeed News that the technology was “deemed unsuitable” after an initial exploration. “Between 2 December 2019 and 22 January 2020, members of the AFP-led Australian Centre to Counter Child Exploitation (ACCCE) registered for a free trial of the Clearview AI facial recognition tool and conducted a limited pilot of the system in order to ascertain its suitability in combating child exploitation and abuse,” Katie Casling, an AFP spokesperson, said in a statement. The Queensland Police Service and its homicide investigations unit ran more than 1,000 searches as of February 2020, based on data reviewed by BuzzFeed News. The department did not respond to requests for comment. Clearview marketed its facial recognition system across Europe by offering free trials at police conferences, where it was often presented as a tool to help find predators and victims of child sex abuse. In October 2019, law enforcement officers from 21 different nations and Interpol gathered at Europol’s European Cybercrime Centre in the Hague in the Netherlands to comb through millions of image and video files of victims intercepted in their home countries as part of a child abuse Victim Identification Taskforce. At the gathering, outside participants who were not Europol staff members presented Clearview AI as a tool that might help in their investigations. After the two-week conference, which included specialists from Belgium, France, and Spain, some officers appear to have taken back home what they had learned and began using Clearview. “The police authority did not know and had not approved the use.” A Europol spokesperson told BuzzFeed News that it did not endorse the use of Clearview, but confirmed that “external participants presented the tool during an event hosted by Europol.” The spokesperson declined to identify the participants. “Clearview AI was used during a short test period by a few employees within the Police Authority, including in connection with a course arranged by Europol. The police authority did not know and had not approved the use,” a spokesperson for the Swedish Police Authority told BuzzFeed News in a statement. In February 2021, the Swedish Data Protection Authority concluded an investigation into the police agency’s use of Clearview and fined it $290,000 for violating the Swedish Criminal Data Act. Leadership at Finland’s National Bureau of Investigation only learned about employees’ use of Clearview after being contacted by BuzzFeed News for this story. After initially denying any usage of the facial recognition software, a spokesperson reversed course a few weeks later, confirming that officers had used the software to run nearly 120 searches. “The unit tested a US service called Clearview AI for the identification of possible victims of sexual abuse to control the increased workload of the unit by means of artificial intelligence and automation,” Mikko Rauhamaa, a senior detective superintendent with Finland’s National Bureau of Investigation, said in a statement. Questions from BuzzFeed News prompted the NBI to inform Finland’s Data Protection Ombudsman of a possible data breach, triggering a further investigation. In a statement to the ombudsman, the NBI said its employees had learned of Clearview at a 2019 Europol event, where it was recommended for use in cases of child sexual exploitation. The NBI has since ceased using Clearview. Data reviewed by BuzzFeed News shows that by early 2020, Clearview had made its way across Europe. Italy’s state police, Polizia di Stato, ran more than 130 searches, according to data, though the agency did not respond to a request for comment. A spokesperson for France’s Ministry of the Interior told BuzzFeed News that they had no information on Clearview, despite internal data listing employees associated with the office as having run more than 400 searches. “INTERPOL’s Crimes Against Children unit uses a range of technologies in its work to identify victims of online child sexual abuse,” a spokesperson for the international police force based in Lyon, France, told BuzzFeed News when asked about the agency’s more than 300 searches. “A small number of officers have used a 30-day free trial account to test the Clearview software. There is no formal relationship between INTERPOL and Clearview, and this software is not used by INTERPOL in its daily work." Child sex abuse typically warrants the use of powerful tools in order to save the victims or track down the perpetrators. But Jake Wiener, a law fellow at the Electronic Privacy Information Center, said that many tools already exist in order to fight this type of crime, and, unlike Clearview, they don’t involve an unsanctioned mass collection of the photos that billions of people post to platforms like Instagram and Facebook. “If police simply want to identify victims of child trafficking, there are robust databases and methods that already exist,” he said. “They don’t need Clearview AI to do this.” Since early 2020, regulators in Canada, France, Sweden, Australia, the UK, and Finland have opened investigations into their government agencies’ use of Clearview. Some privacy experts believe Clearview violated the EU’s data privacy laws, known as the GDPR. To be sure, the GDPR includes some exemptions for law enforcement. It explicitly notes that “covert investigations or video surveillance” can be carried out “for the purposes of the prevention, investigation, detection, or prosecution of criminal offences or the execution of criminal penalties, including the safeguarding against and the prevention of threats to public security…” But in June 2020, the European Data Protection Board, the independent body that oversees the application of the GDPR, issued guidance that “the use of a service such as Clearview AI by law enforcement authorities in the European Union would, as it stands, likely not be consistent with the EU data protection regime.” This January, the Hamburg Commissioner for Data Protection and Freedom of Information in Germany — a country where agencies had no known use of Clearview as of February 2020, according to data — went one step further; it deemed that Clearview itself was in violation of the GDPR and ordered the company to delete biometric information associated with an individual who had filed an earlier complaint. In his response to questions from BuzzFeed News, Ton-That said Clearview has “voluntarily processed” requests from people within the European Union to have their personal information deleted from the company’s databases. He also noted that Clearview does not have contracts with any EU customers “and is not currently available in the EU.” He declined to specify when Clearview stopped being available in the EU. Christoph Schmon, the international policy director for the Electronic Frontier Foundation, told BuzzFeed News that the GDPR adds a new level of complexity for European police officers who had used Clearview. Under the GDPR, police can’t use personal or biometric data unless doing so is “necessary to protect the vital interests” of a person. But if law enforcement agencies aren’t aware they have officers using Clearview, it's impossible to make such evaluations. “If authorities have basically not known that their staff tried Clearview — that I find quite astonishing and quite unbelievable, to be honest,” he said. “It’s the job of law enforcement authorities to know the circumstances that they can produce citizen data and an even higher responsibility to be held accountable for any misuse of citizen data.” "If authorities have basically not known that their staff tried Clearview — that I find quite astonishing." Many experts and civil rights groups have argued that there should be a ban on governmental use of facial recognition. Regardless of whether a facial recognition software is accurate, groups like the Algorithmic Justice League argue that without regulation and proper oversight it can cause overpolicing or false arrests. “Our general stance is that facial recognition tech is problematic, so governments should never use it,” Schmon said. Not only is there a high chance that police officers will misuse facial recognition, he said, but the technology tends to misidentify people of color at higher rates than it does white people. Schmon also noted that facial recognition tools don’t provide facts. They provide a probability that a person matches an image. “Even if the probabilities were engineered correctly, it may still reflect biases,” he said. “They are not neutral.” Clearview did not answer questions about its claims of accuracy. In a March statement to BuzzFeed News, Ton-That said, “As a person of mixed race, ensuring that Clearview AI is non-biased is of great importance to me.” He added, “Based on independent testing and the fact that there have been no reported wrongful arrests related to the use of Clearview AI, we are meeting that standard.” Despite being investigated and, in some cases banned around the world, Clearview’s executives appear to have already begun laying the groundwork for further expansion. The company recently raised $30 million, according to the New York Times, and it has made a number of new hires. Last August, cofounders Ton-That and Richard Schwartz, along with other Clearview executives, appeared on registration papers for companies called Standard International Technologies in Panama and Singapore. In a deposition for an ongoing lawsuit in the US this year, Clearview executive Thomas Mulcaire shed some light on the purpose of those companies. While the subsidiary companies do not yet have any clients, he said, the Panama entity was set up to “potentially transact with law enforcement agencies in Latin America and the Caribbean that would want to use Clearview software.” Mulcaire also said the newly formed Singapore company could do business with Asian law enforcement agencies. In a statement, Ton-That stopped short of confirming those intentions but provided no other explanation for the move. “Clearview AI has set up two international entities that have not conducted any business,” he said. ● CONTRIBUTED REPORTING: Ken Bensinger, Salvador Hernandez, Brianna Sacks, Pranav Dixit, Logan McDonald, John Paczkowski, Mat Honan, Jeremy Singer-Vine, Ben King, Emily Ashton, Hannah Ryan

Top YouTube Content Ideas for Your Channel

Blog PinePage is a safe haven, where talent of all genres can build their online presence, engage with their communities, and monetize their projects without the fear of expulsion It was created as a natural consequence of the alarming recent trend of multiple creatives being banned by many mainstream platforms. We believe that progress is achieved through dialogue and want to be a part of crafting a world where people can express themselves no matter how controversial their views may be. As a result, we take pride in being a Free Speech friendly space and uphold a Zero Ban Policy. If you're tired of big tech corporates’ restrictive policies, we urge you to build your own platform on PinePage and take back control: Your content. Your community. Your rules Blog New post by AlecIThink by kenjizuko by Pritanshu by Shaya by Arthur by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage by PinePage

Private DNS with MagicDNS

Using Tailscale and MagicDNS to create a private DNS setup Too much email? RSS Twitter

More than 8% of American adults are millionaires

Odds are you know at least one millionaire. More than 8% of adults in the U.S. have enough assets to fit the definition, according to the Global Wealth Report 2020 by Credit Suisse. That works out to more than 20 million Americans. Chris Hogan, radio host and author of the book "Everyday Millionaires," surveyed more than 10,000 of those wealthy individuals to figure out their secret to success. Hogan quickly learned that most millionaires aren't the jet-setters you see on television. "These are regular, hardworking, everyday people. They're not flashy," he said. Most of them accumulated their wealth over time by making wise decisions, according to Hogan. And having a particular mindset almost universally contributed to their success, Hogan said. He found that around 97% of millionaires surveyed believed they were in control of their own destiny. That is much higher than the 55% of the general population Hogan found to hold the same opinion. Check out this video to see the other beliefs and practices that more than 90% of millionaires apply to their money. strong 'Predictably Irrational' author says this is what investors should be doing during the pandemic Coronavirus forced this couple into a 27-day quarantine amid their honeymoon cruise How to prepare for a family member with Covid-19 Disclosure: NBCUniversal and Comcast Ventures are investors in Acorns .

New in Scala 3

The exciting new version of Scala 3 brings many improvements and new features. Here we provide you with a quick overview of the most important changes. If you want to dig deeper, there are a few references at your disposal: Scala 3 is a complete overhaul of the Scala language. At its core, many aspects of the type-system have been changed to be more principled. While this also brings exciting new features along (like union types), first and foremost, it means that the type-system gets (even) less in your way and for instance type-inference and overload resolution are much improved. Besides many (minor) cleanups, the Scala 3 syntax offers the following improvements: One underlying core concept of Scala was (and still is to some degree) to provide users with a small set of powerful features that can be combined to great (and sometimes even unforeseen) expressivity. For example, the feature of implicits has been used to model contextual abstraction, to express type-level computation, model type-classes, perform implicit coercions, encode extension methods, and many more. Learning from these use cases, Scala 3 takes a slightly different approach and focuses on intent rather than mechanism. Instead of offering one very powerful feature, Scala 3 offers multiple tailored language features, allowing programmers to directly express their intent: Abstracting over contextual information. Using clauses allow programmers to abstract over information that is available in the calling context and should be passed implicitly. As an improvement over Scala 2 implicits, using clauses can be specified by type, freeing function signatures from term variable names that are never explicitly referred to. Providing Type-class instances. Given instances allow programmers to define the canonical value of a certain type. This makes programming with type-classes more straightforward without leaking implementation details. Retroactively extending classes. In Scala 2, extension methods had to be encoded using implicit conversions or implicit classes. In contrast, in Scala 3 extension methods are now directly built into the language, leading to better error messages and improved type inference. strong. Implicit conversions have been redesigned from the ground up as instances of a type-class Conversion. Higher-order contextual abstractions. The all-new feature of context functions makes contextual abstractions a first-class citizen. They are an important tool for library authors and allow to express concise domain specific languages. Actionable feedback from the compiler. In case an implicit parameter cannot be resolved by the compiler, it now provides import suggestions that may fix the problem. Besides greatly improved type inference, the Scala 3 type system also offers many new features, giving you powerful tools to statically express invariants in the types: strong. Enums have been redesigned to blend well with case classes and form the new standard to express algebraic data types. Opaque Types. Hide implementation details behind opaque type aliases without paying for it in performance! Opaque types supersede value classes and allow you to set up an abstraction barrier without causing additional boxing overhead. strong. Basing the type system on new foundations led to the introduction of new type system features: instances of intersection types, like A & B, are instances of both A and of B. Instances of union types, like A | B, are instances of either A or B. Both constructs allow programmers to flexibly express type constraints outside the inheritance hierarchy. strong. Scala 2 already allowed return types to depend on (value) arguments. In Scala 3 it is now possible to abstract over this pattern and express dependent function types. In the type type F = (e: Entry) => e.Key the result type depends on the argument! strong. Like with dependent function types, Scala 2 supported methods that allow type parameters, but did not allow programmers to abstract over those methods. In Scala 3, polymorphic function types like [A] => List[A] => List[A] can abstract over functions that take type arguments in addition to their value arguments. Type lambdas. What needed to be expressed using a compiler plugin in Scala 2 is now a first-class feature in Scala 3: Type lambdas are type level functions that can be passed as (higher-kinded) type arguments without requiring an auxiliary type definition. Match types. Instead of encoding type-level computation using implicit resolution, Scala 3 offers direct support for matching on types. Integrating type-level computation into the type checker enables improved error messages and removes the need for complicated encodings. Scala has always been at the frontier between functional programming and object-oriented programming – and Scala 3 pushes boundaries in both directions! The above-mentioned type system changes and the redesign of contextual abstractions make functional programming easier than before. At the same time, the following novel features enable well-structured object-oriented designs and support best practices. While macros in Scala 2 were an experimental feature only, Scala 3 comes with a powerful arsenal of tools for metaprogramming. The macro tutorial contains detailed information on the different facilities. In particular, Scala 3 offers the following features for metaprogramming: If you want to learn more about metaprogramming in Scala 3, we invite you to take our tutorial.

Scientists use big data to sway elections and predict riots in the 1960s

Skip to main content p COMMENT 16 September 2020 Scientists use big data to sway elections and predict riots — welcome to the 1960s A cold-war-era corporation targeted voters and presaged many of today’s big-data controversies. Jill Lepore 0 Jill Lepore Jill Lepore is professor of American history at Harvard University in Cambridge, Massachusetts, a staff writer at The New Yorker, host of podcast The Last Archive, and author of If Then: How the Simulmatics Corporation Invented the Future (2020). Protests against racism in Detroit, Michigan, and many other US cities in 1967 prompted attempts to forecast future demonstrations. Credit: -/AFP via Getty Access options Rent or buy this article Get just this article for as long as you need it $39.95 Prices may be subject to local taxes which are calculated during checkout Nature 585, 348-350 (2020) doi: https://doi.org/10.1038/d41586-020-02607-8 References Duhigg, C. ‘Did Uber Steal Google’s Intellectual Property?’ The New Yorker (22 October 2018). Lepore, J. ‘The Disruption Machine’ The New Yorker (23 June 2014). Lepore, J. ‘Politics and the New Machine’ The New Yorker (16 November 2015). Mesthene, E. G. (ed.) Toward the Year 2018 (Foreign Policy Association, 1968). Google Scholar de Sola Pool, I. Backgr. 10, 111–122 (1966). Article Google Scholar Download references The long road to fairer algorithms Don’t ask if artificial intelligence is good or fair, ask how it shifts power Data — from objects to assets Reform predictive policing Subjects History Human behaviour Computer science Policy Latest on: History India cuts periodic table and evolution from school textbooks — experts are baffled News 31 MAY 23 From the archive: foods of the future and cryptography secrets News & Views 30 MAY 23 From the archive: aggressive anemones, and Louis Pasteur’s birthday News & Views 23 MAY 23 Human behaviour Expertise increases planning depth in human gameplay Article 31 MAY 23 Users choose to engage with more partisan news than they are exposed to on Google Search Article 24 MAY 23 p Correspondence 23 MAY 23 Computer science Towards quantum machine learning Spotlight 24 MAY 23 Quantum computers: what are they good for? Spotlight 24 MAY 23 Rewriting the quantum-computer blueprint Outlook 24 MAY 23 Jobs Postdoctoral Research Fellow at the Dalian Institute of Chemical Physics Professor/Associate Professor/Assistant Professor/Senior Lecturer/Lecturer Professor/Associate Professor/Assistant Professor/Senior Lecturer/Lecturer Faculty Positions at SUSTech Department of Biomedical Engineering Postdoctoral Fellows/Research scientists Close banner Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily. Email address I agree my information will be processed in accordance with the Nature and Springer Nature Limited Privacy Policy. Close banner p Sign up for Nature Briefing

High-performance computers are under siege by a new backdoor called Kobalos

High-performance computer networks, some belonging to the world’s most prominent organizations, are under attack by a newly discovered backdoor that gives hackers the ability to remotely execute commands of their choice, researchers said on Tuesday. Kobalos, as researchers from security firm Eset have named the malware, is a backdoor that runs on Linux, FreeBSD, and Solaris, and code artifacts suggest it may have once run on AIX and the ancient Windows 3.11 and Windows 95 platforms. The backdoor was released into the wild no later than 2019, and the group behind it was active throughout last year. While the Kobalos design is complex, its functionalities are limited and almost entirely related to covert backdoor access. Once fully deployed, the malware gives access to the file system of the compromised system and enables access to a remote terminal that gives the attackers the ability to run arbitrary commands. In one mode, the malware acts as a passive implant that opens a TCP port on an infected machine and waits for an incoming connection from an attacker. A separate mode allows the malware to convert servers into command-and-control servers that other Kobalos-infected devices connect to. Infected machines can also be used as proxies that connect to other servers compromised with Kobalos. These proxies can be chained so that the operators can use multiple Kobalos-compromised machines to reach their final targets. The figure below shows an overview of the Kobalos features: To maintain stealth, Kobalos encrypts communications with infected machines using two 16-byte keys that are generated and then encrypted with a password-protected RSA-512 private key. All inbound and outbound traffic from then on is RC4-encrypted using the two keys. The malware uses a complex obfuscation mechanism that makes third-party analysis difficult. Those infected with the malware include a university, an end-point security company, government agencies, and a large ISP, among others. One high-performance computer compromised had no less than 512 gigabytes of RAM and almost a petabyte of storage. Eset said the number of victims was measured in the tens. The number comes from an Internet scan that measures behavior that occurs when a connection is established with a compromised host from a specific source port. The image below shows that the victims were located in the United States, Europe, and Asia: The robustness of the malware, combined with the small number of prominent targets, demonstrates that Kobalos is the work of an advanced team of hackers, particularly in the less-traveled path of non-Windows-based malware. “The numerous well-implemented features and the network evasion techniques show the attackers behind Kobalos are much more knowledgeable than the typical malware author targeting Linux and other non-Windows systems,” Eset researchers Marc-Etienne M.Léveillé and Ignacio Sanmillan wrote in a report. “Their targets, being quite high-profile, also show that the objective of the Kobalos operators isn’t to compromise as many systems as possible. Its small footprint and network evasion techniques may explain why it went undetected until we approached victims with the results of our Internet-wide scan.” So far, it’s not clear how Kobalos is getting installed. A component that steals credentials that administrators used to log in to machines using the SSH protocol is one possibility, but it's unlikely it's the sole means of infection. It's also unclear precisely what the Kobalos operators are doing with the malware. There were no signs that compromised systems were used to mine cryptocurrency or carry out other compute-intensive tasks. “The intent of the authors of this malware is still unknown,” they wrote. “We have not found any clues to indicate whether they steal confidential information, pursue monetary gain, or are after something else.”

Partying and Traveling Around Eastern Europe as a Freelancer

Traveling and partying often go hand in hand, and people who work as freelancers usually manage their own time and their income. That’s why so many freelancers choose to explore some more unusual destinations with a rich culture. Working as a freelance writer, vlogger or in some other creative line of work can open doors to many possibilities regarding traveling. And oftentimes you get to write or vlog about your experiences in that particular country and its nightlife. You get to explore all the secrets of that wonderful city and of course, feel the real atmosphere when the night falls down. But many young people often forget that constant traveling, partying a poor sleep schedule, poor eating habits, etc. can have a toll on your health. Even though you might not have any particular fitness goal in mind, keeping your body healthy is extremely important, especially while you are young. But with just a few tips you can keep on traveling, partying and stay healthy! Eastern Europe has among many young travelers gained a cult status. Maybe it’s because of the communist regime the majority of countries in Eastern Europe had or because of their particular style of music and culture and the way they are hospitable towards strangers. But one thing’s for sure, countries in Eastern Europe are perfect for people who love rustic towns filled with magnificent architecture; who like to drink cheap drinks and experience the party of their lives! One of the hottest party spots today is definitely Bucharest, Romania! Not only will you get the chance to take some of the coolest Slavic aesthetic photos, but you will also get to experience some of the most beautiful nature and of course a wild nightlife. Bucharest is known for being a city with the best bachelor party selection. It gained the status of the best bachelor city for South-East Europe. One thing for sure, Bucharest will inspire you as a writer, vlogger, blogger, etc. because of its unique brutalist architecture, wonderful and warm people, superb cuisine, and of course – Dracula’s Castle that you simply must visit! p You may like:Do You Need Insurance Before Planning a Trip to the Philippines? Another fantastic town know for its wild nightlife is in Belgrade, Serbia! You can easily take a plane from Bucharest to Belgrade and will be there in just a couple of hours! That’s what’s so great about Easter Europe you can be anywhere in a very short amount of time. Belgrade is known for having one of the craziest clubs perfect for people who like to start their party late at night. You will find all sorts of clubs from Serbian Folk music clubs to electronic underground clubs and jazz bars. You will definitely find something to your taste and will most definitely have an unforgettable night! And let’s not forget the beautiful town Budapest, Hungary! You can get to Budapest from Belgrade by plane, train or bus depending on how much time you want to save and how much money you want to spend on your ticket. Budapest is one of the most historically rich cities in Eastern Europe and can definitely say that they have one of the best nightlife scenes! They also like to start partying around midnight and don’t stop until the sun comes up! Won’t that make a fantastic photo for your next blog? One of the best ways you can decrease your chances of temporarily getting sick or god forbid to have any major health issues in the future when traveling is to eat healthy food, sleep, walk and take care of yourself both physically and mentally! We know that eating healthy while traveling can sometimes be a difficult task but it can be managed pretty easily. The most important thing when it comes to eating healthy meals while traveling is to plan every meal in advance! Well, you don’t need to get all obsessive about it, but you should aim to have an overall idea of meals you will prepare for yourself. This will also help you to avoid going straight to KFC or Mc Donald’s! Also, make sure you absolutely take care of your body and especially your hands! Many freelance bloggers, writers, and people who generally type a lot can experience tendonitis which is an inflammation of the tendon. It can occur in any part of the body, but it usually affects hands which leads to pain that will in return prevent you from writing, video editing, etc. If it happens to you, you know what they say “Better be safe than sorry”, so consider investing time and money into reviewing what is the best hand brace for tendonitis and save yourself from possible injuries. p You may like:The Northern Lights - The Most Spectacular Natural Event in the World People often talk about taking care of your body as a general rule for healthy living, but often forget to mention some of the problems freelancers can experience one of them being inflammation of the tendon for example. Freelancers also often sit a lot so while you’re on your journey trough Eastern Europe try to walk as much as you can and stretch as soon as you get up! This will help you prevent sore muscles and will save you from back pain which occurs form all that sitting. One of the best advice you can choose to follow when traveling, partying and in life generally is to enjoy every moment of it! Try to stay present and don’t let the outer noise disturb your inner peace and joy. By being present and listening to what your body is telling you, you will much easily soak in the scenery of the new town, you will learn how to savor your food and will know when to take a break from working while traveling. Being mindful of your body, surroundings, and your inner voice can lead you to have balance in everything you do, which is extremely important when leading an action-packed lifestyle. Being young and wild and free can have its benefits of enormous and unmeasurable life experiences that can’t compare to anything else And of course, don’t forget to drink a lot of water! Jokes aside, but always keep in mind that having a positive mindset and a healthy body means that you can continue to explore the world, meet new people and lead a long and fulfilling life! An author of Namaste UI, published several articles focused on blogging, business, web design & development, e-commerce, finance, health, lifestyle, marketing, social media, SEO, travel. For any types of queries, contact us on info[at]namasteui.com.

New residents near Fukushima nuclear plant can get 2M yen

JR Namie Station, upper right, in Fukushima Prefecture is surrounded by vacant land in March 2019, after homes and shops damaged by the 2011 nuclear disaster were demolished. (Asahi Shimbun file photo) The government will pay up to 2 million yen ($19,300) to families that move to areas around the crippled Fukushima No. 1 nuclear power plant, an unprecedented offer for recovery from the 2011 disaster. Under the program that will start in fiscal 2021, the Reconstruction Agency will provide an additional amount of up to 4 million yen to those who start new businesses in 12 cities, towns and villages where residents had been ordered to evacuate from after the triple meltdown at the plant. Eleven of those municipalities had come under the central government’s evacuation order, while in the remaining municipality, Hironomachi, residents were ordered to leave by the town government. Katsuei Hirasawa, the reconstruction minister, said on Dec. 17 that his agency is focused on repopulating those areas because only around 20 percent of residents have returned there even after the evacuation orders were lifted. One requirement is that the families must live in the locations for at least five years. The agency will provide 1.2 million yen to families that relocate to the 12 areas from other parts of Fukushima Prefecture and 2 million yen to those from other prefectures. The amount is 800,000 yen for single-person households that relocate from other areas of the prefecture and 1.2 million yen for those from outside the prefecture. The agency’s goal is to have 300 people moving into the areas in the first year. It will start accepting applications for the program as early as next summer. The agency will also provide a subsidy to Fukushima Prefecture and local municipalities to promote their town development and entice more people to relocate to the target areas. It expects 5 billion yen will be needed to carry out the program for fiscal 2021. In fiscal 2019, the central government started a program to provide up to 1 million yen to people who move out of Tokyo’s 23 wards as part of its efforts to promote local revitalization. The upper limit of money given to those eligible for the Reconstruction Agency’s new program for Fukushima Prefecture is double the amount set under the government’s local revitalization program.

Connections by James Burke (1978)

Connections explores an Alternative View of Change (the subtitle of the series) that rejects the conventional linear and teleological view of historical progress. Burke contends that one cannot consider the development of any particular piece of the modern world in isolation. Rather, the entire gestalt of the modern world is the result of a web of interconnected events, each one consisting of a person or group acting for reasons of their own (e.g., profit, curiosity, religious) motivations with no concept of the final, modern result of what either their or their contemporaries' actions finally led to. The interplay of the results of these isolated events is what drives history and innovation, and is also the main focus of the series and its sequels. To demonstrate this view, Burke begins each episode with a particular event or innovation in the past (usually Ancient or Medieval times) and traces the path from that event through a series of seemingly unrelated connections to a fundamental and essential aspect of the modern world. For example, The Long Chain episode traces the invention of plastics from the development of the fluyt, a type of Dutch cargo ship. 1. The Trigger Effect details the world’s present dependence on complex technological networks through a detailed narrative of New York City and the power blackout of 1965. 2. Death in the Morning examines the standardization of precious metal with the touchstone in the ancient world. 3. Distant Voices suggests that telecommunications exist because Normans had stirrups for horse riding which in turn led them to further advancements in warfare. 4. Faith in Numbers examines the transition from the Middle Ages to the Renaissance from the perspective of how commercialism, climate change and the Black Death influenced cultural development. 5. The Wheel of Fortune traces astrological knowledge in ancient Greek manuscripts from Baghdad’s founder, Caliph Al-Mansur, via the Muslim monastery/medical school at Gundeshapur, to the medieval Church’s need for alarm clocks (the water horologium and the verge and foliot clock). 6. Thunder in the Skies implicates the Little Ice Age (ca. 1250-1300 AD) in the invention of the chimney, as well as knitting, buttons, wainscoting, wall tapestries, wall plastering, glass windows, and the practice of privacy for sleeping and sex. 7. The Long Chain traces the invention of the Fluyt freighter in Holland in the 1500s. Voyages were insured by Edward Lloyd (Lloyd's of London) if the ships hulls were covered in pitch and tar which came from the colonies until the American Revolution in 1776. 8. Eat, Drink and Be Merry begins with plastic, the plastic credit card and the concept of credit then leaps back in time to to the Dukes of Burgundy, which was the first state to use credit. 9. Countdown connects the invention of the movie projector to improvements in castle fortifications caused by the invention and use of the cannon. 10. Yesterday, Tomorrow and You. A bit of a recap: change causes more change. Start with the plow, you get craftsmen, civilization, irrigation, pottery and writing, mathematics, a calendar to predict floods, empires, and a modern world where change happens so rapidly you can't keep up. 1. Revolutions - What do all these things have in common: 3 grandfathers' lifetimes, 2 revolutions, 1750 Cornwall tin mines, water in mines, pumps, steam engines, Watt's copier, carbon paper, matches, phosphorus fertilizer, trains and gene pool mixing... 2. Sentimental Journeys - What do these have in common: Freud, lifestyle crisis, electric shock therapy, hypnotherapy, magnetism, frenology, penology, physiology, synthetic dyes, the Bunsen burner, absorption, Fraunhofer lines, astronomical telescopes, chromatic aberrations, and surveying? 3. Getting it Together - James Burke explains the relationship between hot air balloons and laughing gas, and goes on to surgery, hydraulic water gardens, hydraulic rams, tunneling through the Alps, the Orient Express, nitroglycerin, heart attacks & headaches, aspirin, carbolic acid, disinfectant, Mabach-Gottlieb Daimler-Mercedes, carburetors, and helicopters. 4. Whodunit? - This episode starts with a billiard ball and ends with a billiard ball. Along the way, Burke examines Georgius Agricola's De Re Metallica, how mining supported war, the role of money, the Spanish Armada, large ships, problems posed by a wood shortage, glass making, coal, plate glass, mirrors, the sextant, barometers... 5. Something for Nothing - How do shuttle landings start with the vacuum which was forbidden by the Church? Burke takes us on an adventure with barometers, weather forecasting, muddy and blacktop roads, rain runoff, sewage, a cholera epidemic, hygiene, plumbing, ceramics, vacuum pumps, compressed air drills, tunnels in the Alps... 6. Echoes of the Past - The past in this case starts with the tea in Dutch-ruled India, examines the Japanese tea ceremony, Zen Buddhism, porcelain, the architecture of Florence, Delftware, Wedgwood, Free Masons, secret codes, radio-telephones, cosmic background radiation and - finally – radio astronomy. 7. Photo Finish - Another series of discoveries examined by Burke which include Eastman's film Kodak Brownie, the disappearing elephant scare of 1867, billiard balls, celluloid as a substitute for ivory, false teeth that explode, gun cotton, double shot sound of a bullet, Mach's shock wave, aerodynamics, nuclear bombs... 8. Separate Ways - Burke shows how to get from sugar to atomic weapons by two totally independent paths. The first involves African slaves, Abolitionist societies, Sampson Lloyd II, wire, suspension bridges, galvanized wire, settlement of the wild West, barbed wire, canned corn, and cadmium. 9. High Times - The connection between polyethylene and Big Ben is a few degrees of separation, so let's recount them: polyethylene, radar, soap, artificial dyes, color perception, tapestries, far East goods, fake lacquer furniture, search for shorter route to Japan, Hudson in Greenland, the discovery of plentiful whales... 10. Deja Vu - James Burke provides evidence that history does repeat itself by examining the likes of black and white movies, Conquistadors, Peruvian Incas, small pox, settlements that look like Spain's cities, the gold abundance ending up in Belgium, Antwerp, colony exploitation, the practice of burying treasure to avoid pirates... 11. New Harmony - A dream of utopia is followed from microchips to Singapore, from the transistor to its most important element, germanium, to Ming Vases and cobalt fakes, which contribute to the blue in blue tiles used in special Islamic places, and Mosaics in Byzantium, the donation of Constantine, Portuguese navigation by stars... 12. Hot Pickle - Burke starts out in a spice market in Istanbul where you can find hot pickle, recounts the retaking of Istanbul by the Turks in 1453, follows the trail of pepper and tea and opium and the exploitation of addicts, moves to the jungles of Java, then to zoos, the use of canaries as carbon monoxide detectors... 13. The Big Spin - is a California lottery which is basically gambling. From here Burke takes us through Alexander Flemming's chance discovery of penicillin, to Vierschoft's observation that contaminated water is related to health, to Schliemann's search for City of Troy, the theft of discovered treasure, and to Vierschoft's criminology. 14. Bright Ideas - Gin and tonic was invented to combat Malaria in British colonies like Java, which leads us to Geneva where cleanliness is an obsession. Here tonic water was sealed with a disposable bottle cap, and razors became disposable, leading us to Huntsman's steel, invaluable for making clock springs and chronometers. 15. Making Waves - a permanent wave in ladies' hair is aided by curlers, and this leads us to explore borax, taking us to Switzerland, Johan Sutter's scam, and the saw mill, and that means the discovery of gold leading to the 1848 California gold rush. 16. Routes - Jethro Tull, a sick English lawyer, recuperates sipping wine and contributes the hoe to help fix farming problems. Farm production is not going so well in France, either. 17. One Word - The one word that changed everything was "filioque" but we must make a trip to Constantinople, visit the Renaissance, meet Aldus Manutius of Venice, explore abbreviations, learn about Italic print, which resulted in an overload of books, requiring the development of a cataloging system. 18. Sign Here - Murphy's Law says you need insurance from Lloyd's of London, so pack your bags to study international law and protect yourself from piracy by calculating the probability. You better study Pascal's math for that, but you might find yourself jailed for free thinking. 19. Better Than the Real Thing - starts in the 1890's with bicycles and bloomers and then takes a look at boots, zippers, sewing machines, and infinitesimal difference. Speaking of small, we look at microscopic germs, Polarized light, sugar, coal, iron, micro-bubbles, the spectroscope, night vision... 20. Flexible Response - is a whimsical look at the myth of the English longbow, Robin Hood, sheep, the need to drain land with windmills, the effect of compound interest, decimal fractions, increased productivity, the Erie Canal, railroads, telegraphs, department stores, Quaker Oats, X-ray diagnostics... 1. Feedback - Electronic agents on the Internet and wartime guns use feedback techniques discovered in the first place by Claude Bernard, whose vivisection experiments kick off animal rights movements called humane societies that really start out as lifeboat crews rescuing people from all the shipwrecks happening because of all the extra ships out there... 2. What's in a Name? - Remember the cornflakes from last episode? Thanks to the fact that corncobs make adhesives to bond Carborundum, otherwise known as silicon carbide, to grinding wheels used to grind light-bulbs. 3. Drop the Apple - At the Smithsonian, we learn of electric crystals that help Pierre and Marie Curie discover what they call radium, and then Langevin uses the piezo-electric crystal to develop sonar that helps save liberty ships (from German U-boats) put together with welding techniques using acetylene made with carbon arcs... 4. An Invisible Object - Black holes in space, seen by the Hubble Telescope, brought into space with hydrazine fuel, which was a by-product of fungicidal French vines, fueled by quarantine conventions and money orders, American Express and Buffalo Bill, Vaudeville and French battles, Joan of Arc and the Inquisition... 5. Life is No Picnic - Instant coffee gets off the ground in World War II and Jeeps lead to nylons and stocking machines smashed by Luddites, who were defended by Byron, who meets John Galt in Turkey, avoiding the same blockade that inspires the "Star-Spangled Banner,"... 6. Elementary Stuff - Alfred Russel Wallace, who studied beetles, Oliver Joseph Lodge and telegraphy, a radio designed by Reginald Fessenden, which was used by banana growers, studied by Augustin Pyramus de Candolle, who got the Swiss to use stamps on postcards with cartoons of Gothic Houses of parliament, which in turn had been inspired by Johann Gottfried Herder's Romantic movement... 7. A Special Place - Professor Sir Alec Jeffries of Leicester University in England develops DNA profiling and schlieren photography used by Theodore von Karman to study aerodynamics and Anthony Fokker's airborne machine guns and the Red Baron and geography and Romantic ideas that start in Italy... 8. Fire from the Sky - Thanks to Continental Drift and Alfred Wegener's passion for mirages, magic images from the sister of King Arthur, whose chivalry supposedly triggers the medieval courtly love answer to adultery, which were in turn inspired by the free love ideas of the mystical Cathars, who lived next to the mystical cabalists... 9. Hit the Water - Thanks to napalm, made with palm oil, also used for margarine, stiffened with a process using kieselguhr that comes from plankton living in currents studied by Ballot bbefore observing the Doppler Effect that caused Fizeau to measure the speed of light speed. Fizeau's father-in-law's friend, Prosper Mérimée, who wrote "Carmen"... 10. In Touch - Starting from an attempt for cheaper fusion power using superconductivity, which was discovered by Onnes, with liquid gas provided by Cailletet, who carried out experiments on a tower built by Eiffel, who also built the Statue of Liberty with its famous poem by the Jewish activist Emma Lazarus...

Skeleton: Responsive CSS Boilerplate

You should use Skeleton if you're embarking on a smaller project or just don't feel like you need all the utility of larger frameworks. Skeleton only styles a handful of standard HTML elements and includes a grid, but that's often more than enough to get started. In fact, this site is built on Skeleton and has ~200 lines of custom CSS (half of which is the docking navigation). Love Skeleton and want to Tweet it, share it, or star it? Well, I appreciate that <3 The grid is a 12-column fluid grid with a max width of 960px, that shrinks with the browser/device at smaller sizes. The max width can be changed with one line of CSS and all columns will resize accordingly. The syntax is simple and it makes coding responsive much easier. Go ahead, resize the browser. One Eleven Two Ten Three Nine Four Eight Five Seven Six Six Seven Five Eight Four Nine Three Ten Two Eleven One One Eleven Two Ten 1/3 2/3 1/2 1/2 Type is all set with the rems, so font-sizes and spacial relationships can be responsively sized based on a single <html> font-size property. Out of the box, Skeleton never changes the <html> font-size, but it's there in case you need it for your project. All measurements are still base 10 though so, an <h1> with 5.0remfont-size just means 50px. strong is Raleway served by Google, set at 15rem (15px) over a 1.6 line height (24px). Other type basics like anchors, strong, emphasis, and underline are all obviously included. Headings create a family of distinct sizes each with specific letter-spacing, line-height, and margins. Headingp <h1> Headingp <h2> Headingp <h3> Headingp <h4> Heading 24rem <h5> Heading 15rem <h6> Heading Heading Heading Heading Heading Heading The base type is 15px over 1.6 line height (24px) Bolded Italicized Colored Underlined Forms are a huge pain, but hopefully these styles make it a bit easier. All inputs, select, and buttons are normalized for a common height cross-browser so inputs can be stacked or placed alongside each other. Your email Reason for contacting Questions Admiration Can I get your number? Message Send a copy to yourself Your email Reason for contacting Questions Admiration Can I get your number? Message Send a copy to yourself Unordered lists have basic styles They use the circle list style Nested lists styled to feel right Can nest either type of list into the other Just more list items mama san Ordered lists also have basic styles They use the decimal list style Ordered and unordered can be nested Can nest either type of list into the other Last list item just for the fun Item 1 Item 2 Item 2.1 Item 2.2 Item 3 Code styling is kept basic – just wrap anything in a <code> and it will appear like this. For blocks of code, wrap a <code> with a <pre>. .some-class { background-color: red; } .some-class { background-color: red; } Be sure to use properly formed table markup with <thead> and <tbody> when building a table. Name Age Sex Location Dave Gamache 26 Male San Francisco Dwayne Johnson 42 Male Hayward Name Age Sex Location Dave Gamache 26 Male San Francisco Dwayne Johnson 42 Male Hayward Skeleton uses media queries to serve its scalable grid, but also has a list of queries for convenience of styling your site across devices. The queries are mobile-first, meaning they target min-width. Mobile-first queries are how Skeleton's grid is built and is the preferrable method of organizing CSS. It means all styles outside of a query apply to all devices, then larger devices are targeted for enhancement. This prevents small devices from having to parse tons of unused CSS. The sizes for the queries are: Desktop HD: 1200px Desktop: 1000px Tablet: 750px Phablet: 550px Mobile: 400px /* Mobile first queries */ /* Larger than mobile */ @media (min-width: 400px) {} /* Larger than phablet */ @media (min-width: 550px) {} /* Larger than tablet */ @media (min-width: 750px) {} /* Larger than desktop */ @media (min-width: 1000px) {} /* Larger than Desktop HD */ @media (min-width: 1200px) {} Skeleton has a number of small utility classes that act as easy-to-use helpers. Sometimes it's better to use a utility class than create a whole new class just to float an element. /* Utility Classes */ /* Make element full width */ .u-full-width { width: 100%; box-sizing: border-box; } /* Make sure elements don't run outside containers (great for images in columns) */ .u-max-full-width { max-width: 100%; box-sizing: border-box; } /* Float either direction */ .u-pull-right { float: right; } .u-pull-left { float: left; } /* Clear a float */ .u-cf { content: ""; display: table; clear: both; } This template is an example of how easy it can be to create a landing page with just the Skeleton grid and a few custom styles. The entire demo is ~150 lines of CSS including comments (most of which is positioning the phones at the top). More examples will be added to help anyone get started or more familiar with how Skeleton works. The goal is education. If you're more interested in real, live examples of Skeleton sites, I'll be creating a "Built on Skeleton" list soon!

My Project Got 800 Stars in Two Days on GitHub

In this article, you will discover how my open-source JavaScript project got 800 stars on GitHub within two days of publishing it. My project is called KOAN (github.com/soygul/koan), and I created it to preserve my general knowledge in Koa and Angular frameworks, as well as Node.js and MongoDB, by making a ready-to-use project template. Upon submitting it to a couple of JavaScript newsletters for review, it exploded and got 800 stars in its first two days of existence! So how did this happen? Did I get lucky? Or did I build something special that people wanted or needed? Or did I simply game the GitHub? Well, read on. You can find the video version of this article on YouTube: https://www.youtube.com/watch?v=vOQyo64WBAA Video has live demonstration of my project KOAN. If you want to read the comments or leave a comment, do so under the YouTube video. If you want to contribute to the article, make a pull request on GitHub. My open-source project KOAN that is the focus of this video: JavaScript Newsletter Issue that featured my project: Relevant articles referred to in this article: First things first. My project, KOAN, is a full-stack web framework template which you can use as a starting point for your JavaScript web apps. On the backend, it uses Koa framework plus Node.js and MongoDB. On the frontend, it uses Angular and WebSockets for real-time interactions between logged-in users. A web app boilerplate with built-in WebSocket support is what got the attention of JavaScript newsletters in the first place. Also, few open-source projects offer a complete working sample of their projects, so creating a Facebook clone with KOAN was a great demonstration of its capabilities. Now let me demonstrate the project to you, and then we can investigate why it was successful. Let’s start with the KOAN’s GitHub project page. It is an open-source project with the MIT license. It is a 7-year-old project, and I didn’t update it much in the last five years, so it has been losing stars since then, almost back to day 2! If we check out the README file, you can see that it is very clean and well structured. It starts with introducing the technology stack, continues with the live example and getting started instructions. Getting started is easy, you clone the repo, do npm install && npm start, and you have a basic working Facebook clone on your computer. You can see that I even have a “Deploy to Heroku” button for one-click sample deployments of this project for people who want a quick start. Finally, the readme provides the rest of the configuration, testing, and licensing information. If you check out the source code, you can see that it is equally well documented. As a wise man once said, code comments should tell you the why, and the code itself should tell you how. Note: Following section is the excerpt from the video demonstration. You can find the link to it in the resources section above. Now let’s go ahead and check the live sample. You can see that I created a Facebook clone as an example usage of KOAN template. Yet again, the very first page of the demo website starts by introducing the project: “KOAN (Koa, Angular, Node): Full-stack JavaScript Web development boilerplate.”. It also introduces the technology stack: Let’s log in and see how it goes. Default credentials are pre-filled, so I will just use them. Logins with other social accounts is a feature that I implemented as a bonus. After login, you can see that it is a clone of Facebook in its early days. You can type comments or create new posts. You can also check your profile page, notifications, and messages. Remember when I said that the project was using WebSockets. The green dot near your name indicates that you are connected to the Node.js backend using a WebSocket connection. Let me open another window and log in as another user. When I submit a comment on one tab, it appears on the other in real-time, thanks to the persistent WebSocket connection to the backend. I won’t go into the details, but this very basic feature alone made this project very valuable to many starters who are looking to implement a real-time web app with WebSockets. By the way, if you are interested, I have an entire article demonstrating my open-source project workflow using Git, GitHub, and Docker. You can find the link to it in the resources section. Now you have seen a successful project with a successful execution, which resulted in a ton of recognition (and stars, forks etc.) on GitHub. So how did this all get together? Let me summarize it: Don’t forget that the primary objective of open-sourcing software is to be useful to the open-source community. The recognition (stars etc.) that comes with it is just a bonus. When the community starts using your software, they will begin to contribute back, which will benefit you in return, and everybody else. Take being useful as your goal, and you will get recognition eventually. Just like this article! If you found this article useful, share it with someone who you think might also find it useful. I will showcase more of my successful software projects in the future. If you want to see them, follow me on my socials. If you have a project that you want me to check out, let me know. I will review them and maybe even feature them in upcoming articles. And that is it for this writeup, I will see you on the next one.

Chinese company uses facial recognition to police teen gamers

Chinese company uses facial recognition to police teen gamers We’re sorry, this feature is currently unavailable. We’re working to restore it. Please try again later. Dismiss Our network Open Navigation Menu The Sydney Morning Herald Subscribe The Sydney Morning Herald Advertisement World North America Inside China This was published 1 year ago Chinese company uses facial recognition to police teen gamers By and pAmy Chang Chien July 9, 2021 — 11.51am Save Log in, register or subscribe to save articles for later. Save articles for later Add articles to your saved list and come back to them any time. Got it Normal text size Larger text size Very large text size Hong Kong: For almost every video game restriction, children and teenagers will find a way around it. But the room to manoeuvre is shrinking in China, where underage players are required to log on using their real names and identification numbers as part of countrywide regulations aimed at limiting screen time and keeping internet addiction in check. In 2019, the country imposed a cyber curfew banning those under 18 from playing games between 10pm and 8am. Recognising that wily teenagers might try to use their parents’ devices or identities to circumvent the restrictions, the Chinese internet conglomerate Tencent said this week that it would close the loophole by deploying facial recognition technology in its video games. Beijing introduced an e-game curfew in 2019. “Children, put your phones away and go to sleep,” Tencent said in a statement Tuesday when it officially introduced the new feature, which is called “Midnight Patrol”. The wider rollout set off a debate on Chinese internet platforms about the benefits and privacy risks of the technology. Some were in favour of the controls, saying they would combat adolescent internet addiction, but they also questioned how the data would be relayed to the authorities. Others said Tencent was assuming an overly paternalistic role. “This type of thing ought to be done by the parents,” a user named Qian Mo Chanter wrote on Zhihu forum. “Control the kid and save the game.” Thousands of internet users complained about the tightening controls and the shrinking space for anonymity in cyberspace. A hashtag on Weibo, a microblogging platform, reminded gamers to make sure they were fully dressed in case the camera captured more than their faces. Loading Xu Minghao, a 24-year-old programmer in the northern city of Qingdao, said that he would delete any video games that required facial recognition, citing privacy concerns. “I don’t trust any of this software,” he wrote on Zhihu. Advertisement Privacy concerns were widely discussed when the real-name registration requirement for minors was introduced in 2019. Describing facial recognition technology as a double-edged sword, the China Security and Protection Industry Association, a government-linked trade group, said in a paper published last year that the mass collection of personal data could result in security breaches. Facial recognition is widely used in China, like here, where Beijing commuters have their face scanned at the railway station. Tencent said that it began testing facial recognition technology in April to verify the ages of avid night-time players and has since used it in 60 of its games. In June, it prompted an average of 5.8 million users a day to show their faces while logging in, blocking more than 90 per cent of those who rejected or failed facial verification from accessing their accounts. Facial recognition technology is commonly used in China to streamline daily activities as well as regulate public behaviour. Hotels use it when checking in guests while banks use it to verify payments. The state uses it to track down criminal suspects. One city has even deployed the technology to shame its residents out of the habit of wearing pyjamas in public. In the case of video games, the government has long blamed them for contributing to youthful nearsightedness, sleep deprivation and low academic performance. The 2019 regulations also limited how much time and money underage users could spend playing video games. In another sign of China’s moves to control online behaviour, Tencent’s popular WeChat messagingservice has deleted accounts on LGBT topics run by university students and nongovernment groups. Wechat sent account holders a notice they violated rules but gave no details, according to the founder of an LGBT group, who asked not to be identified further out of fear of possible official retaliation. She said dozens of accounts were shut down, all at about 10 pm on Tuesday. It wasn’t clear whether the step was ordered by Chinese authorities, but it comes as the ruling party tightens political controls and tries to silence groups that might criticize its rule. The Communist Party decriminalized homosexuality in 1997, but gay, lesbian, bisexual, transsexual and other sexual minorities still face discrimination. While there is more public discussion of such issues, some LGBT activities have been blocked by authorities. The New York Times, AP Save Log in, register or subscribe to save articles for later. Inside China China Most Viewed in World Loading

Alfred Anaya Put Secret Compartments in Cars. The DEA Put Him in Prison (2013)

Alfred Anaya took pride in his generous service guarantee. Though his stereo installation business, Valley Custom Audio Fanatics, was just a one-man operation based out of his San Fernando, California, home, he offered all of his clients a lifetime warranty: If there was ever any problem with his handiwork, he would fix it for the cost of parts alone—no questions asked. Anaya's customers typically took advantage of this deal when their fiendishly loud subwoofers blew out or their fiberglass speaker boxes developed hairline cracks. But in late January 2009, a man whom Anaya knew only as Esteban called for help with a more exotic product: a hidden compartment that Anaya had installed in his Ford F-150 pickup truck. Over the years, these secret stash spots—or traps, as they're known in automotive slang—have become a popular luxury item among the wealthy and shady alike. This particular compartment was located behind the truck's backseat, which Anaya had rigged with a set of hydraulic cylinders linked to the vehicle's electrical system. The only way to make the seat slide forward and reveal its secret was by pressing and holding four switches simultaneously: two for the power door locks and two for the windows. Esteban said the seat was no longer responding to the switch combination and that no amount of jiggling could make it budge. He pleaded with Anaya to take a look. Anaya was unsettled by this request, for he had suspicions about the nature of Esteban's work. There is nothing intrinsically illegal about building traps, which are commonly used to hide everything from pricey jewelry to legal handguns. But the activity runs afoul of California law if an installer knows for certain that his compartment will be used to transport drugs. The maximum penalty is three years in prison. Anaya thus thought it wise to deviate from his standard no-questions-asked policy before agreeing to honor his warranty. "There's nothing in there I shouldn't know about, is there?" he asked. Esteban assured him that he needn't worry. Esteban drove the F-150 to Anaya's modest ranch-style house and parked by the back porch. A friend of his, who introduced himself as Cesar, followed right behind in a black Honda Ridgeline truck. The 37-year-old Anaya, a boyishly handsome man whose neck and arms are covered with tattoos of dice and Japanese art, tested the switches that controlled the truck's trap. He heard the hydraulics whirr to life, but the seat stayed firmly in place. He would have to use brute force. Anaya punched a precise hole through the upholstery with his 24-volt Makita drill, probing for the screws that anchored the seat to the hydraulics. After a few moments he heard a loud pop as the drill seemed to puncture something soft. When he finally managed to remove the backseat, he saw what he had hit: a wad of cash about 4 inches thick. The whole compartment was overflowing with such bundles, several of which spilled onto the truck's floor. Esteban had jammed the trap by stuffing it with too much cash—over $800,000 in total. Anaya stumbled back from the truck's cab, livid. "Get it out of here," he growled at Esteban. "I don't want to know about this. I don't want any problems." Esteban Magallon Maldanado and Cesar Bonilla Montiel scrambled to haul armfuls of money from the F-150 to the Ridgeline's trunk. They wanted to stay in Anaya's good graces, because men with his skills are extremely valuable in the narcotics trade. To distribute product from wholesalers to retailers, drug-trafficking organizations need vehicles equipped with well-disguised traps so that loads aren't routinely seized while in transit. The word in the California underworld was that no one built more elegant traps than Anaya, a perfectionist who made sure his hiding spots were invisible to even the most expert eyes. Maldanado and Montiel, key players in a smuggling ring that was sending large quantities of cocaine and methamphetamine to the Midwest, were eager to use his services again. Once all of the money had been moved to the Ridgeline, Anaya, now feeling calmer, agreed to fix the F-150’s trap for $1,500—a third of what he had originally charged to install it. He even offered to improve the compartment by adding another switch—the one that reclined the driver's seat—to the unlocking sequence. A grateful Maldanado then asked Anaya if he could install a trap in the Ridgeline too. The Honda truck already had one, but it was the work of a rank amateur—just a crude hole sawed into the base of the trunk. Maldanado wanted an electronic trap like the F-150’s, and he offered to leave a cash deposit so Anaya could buy the necessary hydraulics. Anaya, who was deeply in debt to numerous creditors, decided to accept the job. He hadn't totally forgiven Maldanado for failing to warn him about the money jammed in the trap, but he figured that he was still adhering to the letter of the law. The fact was that he hadn't seen any drugs, and there had been no discussion of how Maldanado had earned his small fortune. Given those circumstances, Anaya assumed that he was immune from legal trouble in connection with his meticulous creations. He was, after all, just an installer. Unlocking a Stash Spot The artisans who build hidden compartments in cars are secretive about their work. Their clients, who range from tycoons to gangsters, use these "traps" to stash items that are either tremendously valuable or tremendously illicit (or both). The most sought-after traps, like this one in the dashboard of a 2012 Honda Accord, can be opened only by following an elaborate set of steps.—B.I.K. Illustration: Paul Pope 1. Sit in driver's seat. The trap is connected to a pressure sensor under the driver's seat; someone must be sitting in the seat before the compartment can be opened. Illustration: Paul Pope 2. Close all doors. The stash spot won't open unless all the doors are closed—which would rarely be the case during a typical roadside search by law enforcement officers. Illustration: Paul Pope 3. Turn on defroster. To continue the unlatching sequence, you must activate the rear defroster while simultaneously pushing two window switches on the driver's door. Illustration: Paul Pope 4. Swipe card. A magnet is hidden behind an air-conditioning vent. A magnetic card must be swiped across the vent to complete the sequence that unlocks the trap. Illustration: Paul Pope 5. Retrieve contraband. A pair of hydraulic cylinders open the hatch for the secret compartment, which is located in the void where the passenger-side airbag should be. Illustration: Paul Pope When he was 8 years old, Alfred Anaya destroyed his mother's vacuum cleaner in the pursuit of knowledge. "I took it apart because I wanted to find the motor inside," he recalls. "I was so young, I thought the motor would work all by itself even after I took it out. I didn't realize it needed to be plugged in to go." His mother was upset but hardly surprised to discover her ruined vacuum, for she knew all about her youngest son's rabid curiosity. Alfred was forever disassembling Sony Walkmans or clock radios so he could fill his favorite junk drawer with circuit boards, which thrilled him with their intricacy. Anaya idolized his father, Gabriel, a hardworking cement mason who had emigrated from Mexico. Before he even hit adolescence, Alfred started skipping school to help his father pour concrete at shopping malls. He used discarded materials from these construction sites to build labyrinthine clubhouses in the backyard of his family's San Fernando home. By furtively borrowing his dad's circular saw, he outfitted his structures with pulley-operated drawbridges, camouflaged trapdoors, and secret rooms where he could snuggle with girls. In his mid-teens, Anaya developed an obsession with cars. He saved up $500 to buy a wrecked 1963 Volkswagen Beetle, which he lovingly restored by hand. After dropping out of school in the 11th grade, he started to hang out at a local stereo shop, Super Sound Electronics. He swept floors and washed customers' cars for free, just so he could peer over the shoulders of the shop's installers as they molded speaker boxes and snaked cables through trunk walls. Once he cajoled Super Sound's owner into taking him on as an apprentice, Anaya quickly established himself as the shop's rising star. Years of reading blueprints with his dad had given him a knack for visualizing how best to meld stereo components into a car's natural contours. "When you customize cars, you got to have an imagination, you got to be able to see the way it's going to look when you're done putting in this outrageous sound system," says Tony Cardone, a childhood friend of Anaya's who also became a stereo installer. "That's one thing Alfred has always been really good at." Anaya excelled at fabricating candy-colored subwoofer enclosures with voluptuous curves; he often achieved his desired shape by stretching fleece pajamas atop wooden frames, then pouring on molten resin that stiffened as it cooled. Anaya also learned that sometimes the best approach was to conceal his work. "Sound always sounds best when you have no idea where it's coming from," he says. "You want people to feel like they're listening to magic." To cater to clients who preferred stealth over flash, Anaya taught himself to build speaker boxes that fit into the irregularly shaped voids behind door panels and back seats. That skill came in handy when customers started asking for traps, presumably as places to hide their weapons, cash, or weed from both cops and robbers. Anaya was happy to provide this service, which appealed to his innate sense of mischief. The first trap he ever saw, designed by one of his Super Sound mentors, was carved into a dashboard, with a door hinged on a power antenna that could be extended or retracted via remote control. Anaya ached to build similarly ingenious compartments that would dazzle his fellow gearheads, who adore innovations that seem plucked from the world of James Bond. "Blowing everyone's mind, that's what's so rewarding about what we do—the feedback and adrenaline you get from that," Anaya says. "I wanted my compartments to be more sophisticated than anybody else's." By 2002, Anaya had become one of the most sought-after installers in Southern California, with a client list that included rappers, pro basketball players, and porn stars. Mobile Electronics honored him as one of the top 100 installers in the nation, and his systems were later featured in the bikini-laden pages of magazines like Lowrider and Lug. Anaya capitalized on his fame to open his own shop, Valley Custom Audio Fanatics, in a San Fernando storefront. A year later, shortly before he married a woman named Aimee Basham, Anaya persuaded an investor to help him move into larger quarters in nearby North Hollywood. Excited by the opportunity, Anaya spent a month making the new shop's centerpiece, a 12-foot-long fiberglass display case fashioned to resemble an alien's spine. His dad, Gabriel, who was suffering from terminal colon cancer, visited the store shortly before its grand opening. As he unpacked crates of gear, Alfred spotted his withered father sitting on a speaker box, beaming with pride over all his son had accomplished. "Maybe the greatest memory I have," Anaya says. But that happy moment was soon overshadowed by Valley Custom Audio's financial woes. Like many people blessed with formidable creative talent, Anaya was a horrendous manager of both time and money. He took on too many projects and failed to keep track of expenditures. Stressed by the burdens of business ownership, he began to drink too much, downing beer after beer as he struggled to finish cars that were weeks behind schedule. His personal finances became a mess too, thanks to a crushing mortgage and his lavish spending on motorcycles, strip clubs, and camping trips with his two young sons. (One of the boys was from a relationship prior to his marriage to Basham.) In 2007, Anaya was forced to move the failing business to his home—much to the annoyance of Basham, who hated the constant din of generators out by the garage. But Anaya's troubles persisted: Shady customers stiffed him for thousands, yet he kept buying the finest Rockford Fosgate subwoofers and Snap-on tools with his overburdened credit cards. The only bright spot for Valley Custom Audio was its burgeoning trade in traps. Anaya didn't advertise this service, but satisfied customers referred their friends. He charged $4,000 to $5,000 per compartment, far more than he earned from the typical stereo installation. Best of all, these customers paid on time, and they paid in cash. By the end of 2008, trap building represented about 70 percent of Anaya's workload. He knew full well that he was flirting with danger—he could certainly guess how some of his traps might be used. But he also thought that California's law on hidden compartments, one of the very few in the nation, offered clear guidance: Building a trap was illegal only if it was done with the "intent to store, conceal, smuggle, or transport a controlled substance." Based on his consultations with fellow installers, Anaya believed he would cross that line only if a client specifically mentioned drugs. So Anaya adopted a policy similar to the one used by shops that sell bongs: He would turn away anyone who used drug-related lingo when ordering a trap. As long as a customer was discreet, Anaya saw no problem with taking their money. The forefather of modern trap making was a French mechanic who went by the name of Claude Marceau (possibly a pseudonym). According to a 1973 Justice Department report, Marceau personally welded 160 pounds of heroin into the frame of a Lancia limousine that was shipped to the US in 1970—a key triumph for the fabled French Connection, the international smuggling ring immortalized in film. Traps like Marceau's may be difficult to detect, but they require significant time and expertise to operate. The only way to load and unload one of these "dumb" compartments is by taking a car apart, piece by piece. That makes economic sense for multinational organizations like the French Connection, which infrequently transport massive amounts of narcotics between continents. But domestic traffickers, who must ferry small shipments between cities on a regular basis, can't sacrifice an entire car every time they make a delivery. They need to be able to store and retrieve their contraband with ease and then reuse the vehicles again and again. Early drug traffickers stashed their loads in obvious places: wheel wells, spare tires, the nooks of engine blocks. Starting in the early 1980s, however, they switched to what the Drug Enforcement Administration refers to as "urban traps": medium-size compartments concealed behind electronically controlled facades. The first such stash spots were usually located in the doors of luxury sedans; trap makers, who are often moonlighting auto body specialists, would slice out the door panels and then attach them to the motors that raised and lowered the windows. They soon moved on to building traps in dashboards, seats, and roofs, with button-operated doors secured by magnetic locks. Over time, the magnets gave way to hydraulic cylinders, which made the doors harder to dislodge during police inspections. By the early 1990s, however, drug traffickers had discovered that these compartments had two major design flaws. The first was that the buttons and switches that controlled the traps' doors were aftermarket additions to the cars. This made them too easy to locate—police were being trained to look for any widgets that hadn't been installed on the assembly line. Second, opening the traps was no great challenge once a cop identified the appropriate button: The compartment's door would respond to a single press. Sometimes the police would even open traps by accident; a knee or elbow would brush against a button during a vigorous search, and a brick of cocaine would appear as if by magic. Trap makers responded to the traffickers' complaints by tapping into the internal electrical systems of cars. They began to connect their compartments to those systems with relays, electromagnetic switches that enable low-power circuits to control higher-power circuits. (Relays are the reason, for example, that the small act of turning an ignition key can start a whole engine.) Some relays won't let current flow through until several input circuits have been completed—in other words, until several separate actions have been performed. By wiring these switches into cars, trap makers could build compartments that were operated not by aftermarket buttons but by a car's own factory-installed controls. "With the relay switches, you can only have access to the compartment if you do a series of events in exactly the right sequence," says Michael Lewis, the sheriff of Wicomico County, Maryland, who became a nationally recognized expert on traps during his 22-year career as a state trooper. A typical sequence will consist of pushing a variety of switches a specific number of times: a window switch three times, a door lock four times, the rear defroster button twice. But for trap makers who are particularly adept with relays, the complexity of the unlocking sequence is limited only by their imaginations. Many rig the electronics so that the compartment won't open unless all of the vehicle's doors are closed—something that is rarely the case during a roadside search. Another tactic is to link a trap with the pressure sensor beneath the driver's seat, so that the compartment can't be opened unless someone is sitting behind the wheel. In recent years, trap makers have competed to see who can dream up the most elaborate opening tricks. The acknowledged masters of this art are the Dominican-born installers of the Bronx, many of whom work out of auto body shops on Jerome Avenue—a gritty strip that DEA agents call the Silicon Valley of trap making. "The Dominicans started doing voice activation about six years ago," says Lewis, who teaches classes in trap recognition to law-enforcement agencies nationwide. "I have videotape of a Dominican trap—you have to activate cruise control, pull one window up while you pull another window down, and you speak. And when you speak, you complete a circuit and activate the compartment. It's pretty badass." But the ultimate measure of a compartment's worth is not how hard it is to open but how hard it is to find. A cop may never be able to guess the event sequence that opens a trap's door, but that obstacle is irrelevant if the compartment's existence is betrayed by faulty craftsmanship—a stray wire poking out from under a seat cushion or a haphazard bead of metal bonding. If there is any visual hint that a car contains a trap, police can often get a warrant to tear it apart. And even the most well-fortified compartment cannot withstand the incursions of drills and saws. Alfred Anaya attracted a loyal clientele because his compartments were immaculate and therefore undetectable. He was meticulous to a fault, the sort of man who once painted his house 10 times because he couldn't settle for anything less than the perfect shade of white. His customers, who gambled hundreds of thousands of dollars every time they put a shipment on the road, greatly appreciated his attention to detail. If Anaya was less diligent about understanding the legal nuances of his business, that wasn't their problem. Illustration: Paul Pope Sometime in late 2008, Anaya received a call from a customer who lived in the San Diego area. The man wanted him to fix a malfunctioning trap located in Tijuana. Anaya was scared to venture across the border; as much as he hated to renege on his warranty, he refused to go to Mexico. Anaya thought he had protected himself by turning down the job, but the damage had been done the moment he answered the phone. This particular customer was the target of a DEA investigation, and agents had eavesdropped on their conversation. The DEA decided to tap Anaya's phone too, in an effort to identify other drug traffickers who were having traps built by Valley Custom Audio. Shortly after that tap went live on January 30, 2009, agents heard Anaya tell Esteban Magallon Maldanado that he had finished repairing the Ford F-150—the truck with the trap that had been jammed with cash. Maldanado and his partner, Cesar Bonilla Montiel, picked up the vehicle at once, for they had an important delivery to make: Their associates in Kansas City, Kansas, were expecting a shipment of 6 kilos of cocaine and 5 pounds of methamphetamine. Running drugs from Southern California to Kansas was a highly profitable endeavor for Maldanado and Montiel. The two men frequented underground cockfights, where they would arrange to purchase cocaine and meth from a pair of high-level Mexican wholesalers they knew only as Suki and Gordito. They would then hire drivers to transport the product to Kansas City, where further distribution was handled by a brash twentysomething dealer named Curtis Crow. On this particular trip in February 2009, Maldanado and Montiel hired a cocaine addict named Jaime Rodriguez to drive the F-150 to Kansas City. Rodriguez was nearly at the end of the 1,600-mile journey when the Kansas Highway Patrol pulled him over for speeding. A suspicious officer sent the vehicle to be searched by a K-9 unit at a Topeka garage. The dog indicated the possible presence of drugs, so a trooper went over every inch of the truck by hand. But try as he might, he could not locate the trap behind the backseat. Rodriguez was eventually allowed to drive away with more than 18 pounds of drugs still hidden in the truck. There could be no greater testament to Anaya's value to the business, though Anaya himself knew nothing of this near miss. Over the next several weeks, Maldanado and Montiel paid Anaya to build traps in three more vehicles: the Honda Ridgeline that they had dropped off while getting the F-150 fixed, a 2007 Toyota Camry, and a 2008 Toyota Sequoia. The Ridgeline made a run to Kansas in March, while the Sequoia and the Camry were part of a convoy in April. Those trips brought Crow another 9 kilograms of cocaine and 9 pounds of meth. But the cars Anaya had worked on were rapidly losing their powers to deceive. On April 5, for example, the California Highway Patrol stopped the Sequoia and found the trap with ease, seizing more than $106,000 in cash. On April 24, the CHP stopped the Camry and again found the trap. This one contained 2 pounds of meth. The tap on Anaya's phone, combined with surveillance of his house, was giving the DEA all the intelligence it needed to frustrate his clients. Obviously unaware of the DEA's scrutiny, Maldanado and Montiel feared that Anaya had become a snitch. They cut off all contact with the trap maker and got rid of any vehicles he had touched. But despite these precautions, the California-to-Kansas operation was too reckless to elude the authorities for long. Crow, in particular, was wildly incautious: He robbed fellow dealers, hired friends with drug habits, and got high on his own supply. After the DEA traced a phone call that Montiel had made to the house where Crow stored his drugs, it was only a matter of time before the organization was crushed. The inevitable end came in September 2009, after a driver who had been caught with 8 kilos of cocaine agreed to cooperate with the DEA. Virtually all of the ring's participants were immediately rounded up, save for Maldanado, who went on the lam. (He was finally captured in Riverside, California, in March 2012.) Anaya, of course, did not hear a word about these arrests; he hadn't been contacted by either Maldanado or Montiel since the spring. He was now busy dealing with two personal crises: a mounting pile of debt that totaled nearly $55,000, not including his underwater mortgage, and the dissolution of his marriage to Basham, who had become fed up with his workaholism and carousing and filed for divorce. On November 18, as Anaya drove his Ford F-350 through a Home Depot parking lot, he noticed a dark sedan that seemed to be shadowing him in an adjacent aisle. He thought the car might belong to friends. But when the sedan stopped in front of him, the men who got out were strangers to Anaya. They identified themselves as DEA agents and ordered him out of his truck. "You know why we're here," one agent said to Anaya, who was bewildered to be in handcuffs for the first time in his life. "Your compartments." The agents took Anaya to the DEA's office in downtown Los Angeles, where they questioned him at length. Anaya spoke freely about his traps, estimating that he had built 15 over the past year. He even boasted about his perfectionism, stressing that he was always careful to conceal his wire harnesses. The agents told Anaya that he could avoid any potential legal complications by doing them a big favor: They wanted him to outfit his clients' cars with GPS trackers and miniature cameras, so the DEA could build cases against suspected traffickers. They told him to take a few days to mull over the offer, then they released him from custody. The next day, a dazed Anaya drove to his father's grave to meditate on the choice before him. The epiphany he had while kneeling by the headstone wasn't comforting. "I had a feeling that no matter what decision I made, something bad was going to happen," Anaya says. "But I couldn't do anything that would put my family in danger." And while he felt he could handle jail time, he worried that any trafficker big enough to interest the DEA would have no compunctions about killing his children, nieces, and nephews. That made the decision clear. When Anaya told the DEA that he was too frightened to become an informant, the agents made a new, more enticing proposition: They would set up Valley Custom Audio in a deluxe storefront, complete with every piece of equipment that Anaya desired. They wouldn't ask him to place any surveillance gadgets in cars, but the shop would be bugged from floor to ceiling. Once again, Anaya refused. On December 10, Anaya was arrested and subsequently charged in Los Angeles Superior Court for "false compartment activity." He was initially denied bail, in part because an illegal assault rifle and a bulletproof vest had been discovered in his house during a police search. ("Y'know, hey, I like to shoot guns," Anaya says unapologetically; he has two large pistols tattooed on his chest.) His lawyer advised him that, given his totally clean criminal record, he was unlikely to spend much time behind bars for such a minor offense. But in March 2010, Anaya received grim and surprising news: The federal government was taking over the case, and it was going to prosecute him in Kansas—a state he had never set foot in. Prosecutions of trap makers are exceedingly rare. There is no federal law against building hidden compartments, even if they're made with the sole intent of smuggling drugs. The Justice Department occasionally goes after trap makers for violating statutes that ban the sale of drug paraphernalia, but these are difficult cases to make; they require hard evidence, such as an audio recording, that proves the defendant was explicitly told how his compartment would be used. Anaya was never caught on tape discussing drugs. But the prosecutors in Kansas went after Anaya for a much graver crime than selling paraphernalia: They indicted him as a full-fledged conspirator in the California-to-Kansas trafficking operation. Even though he had never seen or touched any drugs and had been shunned as an informant after building just four traps in exchange for less than $20,000, Anaya faced the exact same charge as Maldanado, Montiel, and Crow. This aggressive legal stratagem was almost without precedent. The only similar case on record was that of Frank Rodriguez Torres, a New York trap maker who was extradited to North Carolina in 1998. He was sentenced to five years in prison. By the time Anaya was placed in custody in Kansas in April 2010, virtually all of the case's 23 defendants were scrambling to cut deals. But Anaya resisted his court-appointed lawyer's advice to plead guilty; he still couldn't fathom how building traps made him a drug trafficker, and he was confident that a jury would sympathize with his plight. When the trial started on January 25, 2011, the lead prosecutor, an assistant US attorney named Sheri McCracken, argued that Anaya was one of the main reasons the smuggling ring had evolved into a multimillion-dollar enterprise. The organization "moved up in the world when they met Mr. Anaya," she told the jury. "He built supreme compartments, and because he did that, drug hauling became easier ... But for Mr. Anaya's compartment building, lots of loads would be lost." The primary evidence against Anaya was the testimony of Montiel, who had agreed to cooperate with the government. While in jail in Kansas, he had initially signed an affidavit stating that Anaya had nothing to do with the conspiracy. But he later recanted, claiming that Anaya had enlisted a fellow inmate to intimidate him into signing the document. (Anaya denies this charge.) On the witness stand, Montiel vividly described the incident with the F-150's broken trap, when Anaya had glimpsed more than $800,000 in cash. The prosecutor contended that seeing such a large sum was tantamount to seeing drugs, since Anaya surely must have deduced the source of the money. Montiel also shared a potentially damning anecdote regarding the negotiations over the Honda Ridgeline's trap. "We asked him to build us a hidden compartment for 10 kilos," he testified. "I remember we had problems because he asked, 'Well, what's a kilo like?' I remember I saw a brick on the ground, and I said, 'It's a little bit bigger than this. I need you to do it for 10.'" This was the only evidence that directly linked Anaya to drugs. But it was unrecorded and uncorroborated, and Anaya's attorney made some headway by painting Montiel as a man who would say anything to reduce his own sentence. (Anaya points out—correctly—that his San Fernando home contains no brick.) McCracken's case may have been largely circumstantial, but she did an effective job of portraying Anaya as a man who enjoyed the perks of drug trafficking. She spoke of his "expensive motorcycles and four-wheel bikes to go on the sand," his collection of guns, and his vast array of Snap-on tools. On several occasions, she mentioned that he had a backyard pool "custom built with his name in the bottom of it in marble." Anaya's lawyer tried to explain that all of these supposed extravagances had been bought on credit and that his client was on the brink of bankruptcy. The name by his pool—not in it, as McCracken had claimed—was an $8 DIY project hacked together from grinding concrete and artfully applied stain. But the jury bought into McCracken's narrative; it convicted Anaya on all counts. At his sentencing on January 4, 2012, a visibly nervous Anaya addressed the court for the first time, expressing his feelings of regret and confusion: I built these compartments just like any other business that I had, doing stereo business, customizing needs to people's needs in their vehicles, and I admit there was probably some irresponsibility of building these things, but I was only—I just figured it would be, like, as long as I didn't know what was going on—and don't want to know—there was no law against it ... If I had known there was a law against it, I wouldn't be here. If there was a law that says these compartments are illegal to build, I would not build them. If I had known this was going to happen to me, I wouldn't have done it. McCracken took no pity on him. "He makes the drug world work," she told the judge. "He is equivalent to what I consider somewhat of a genius that takes cocaine and molds them into shapes so that they can be moved in plain sight ... I don't feel bad at all today. In fact, this is a pleasure. And Mr. Anaya says that he's part of this big group of people that puts in compartments. He's part of this secret society, I guess. Well, I hope he tells a friend, because we're coming for them." The judge agreed with McCracken's harsh assessment. He sentenced Anaya to 292 months in federal prison—more than 24 years—with no possibility of parole. Curtis Crow and Cesar Bonilla Montiel, the men at the top of the organization, received sentences half that length. A common hacker refrain is that technology is always morally neutral. The culture's libertarian ethos holds that creators shouldn't be faulted if someone uses their gadget or hunk of code to cause harm; the people who build things are under no obligation to meddle in the affairs of the adults who consume their wares. But Alfred Anaya's case makes clear that the government rejects that permissive worldview. The technically savvy are on notice that they must be very careful about whom they deal with, since calculated ignorance of illegal activity is not an acceptable excuse. But at what point does a failure to be nosy edge into criminal conduct? In light of what happened to Anaya, that question is nearly impossible to answer. "What's troubling a lot of people is that this conviction seems to impose a new sort of liability on people that create state-of-the-art technology," says Branden Bell, an attorney in Olathe, Kansas, who is handling Anaya's appeal. "The logic goes that because he suspected his customers of doing something, he had a duty to ask. But that is a duty that is written nowhere in the law." The challenge for anyone who creates technology is to guess when, exactly, they should turn their back on paying customers. Take, for example, a manufacturer of robot kits for hobbyists. If someone uses those robots to patrol a smuggling route or help protect a meth lab so that traffickers can better evade law enforcement, how will prosecutors determine whether the company acted criminally? If it accepted payment in crumpled $20 bills and thus should have known it was dealing with gangsters? If the customer picked up the merchandise in an overly flashy car? The law offers scant guidance, but prosecutors have tremendous leeway to pursue conspiracy charges whenever they see fit. And as 3D printers enable the unfettered production of sophisticated objects, those prosecutors will be tempted to make examples out of people who are careless about their clients. Anaya can attest to the great sorrows of becoming such an example. When I visited him at the Victorville Federal Correctional Complex, on the sun-cooked edge of California's Mojave Desert, he was still coming to grips with the desolation of prison life. His ex-wife, Aimee Basham, with whom he recently reconciled, brings the family to visit at least once a month. But Anaya is anguished by the prison's restrictions on personal contact with his children; he can scarcely believe that his youngest son will never again sit on his lap. And he bemoans the financial disaster that has befallen his family in his absence—ING Direct foreclosed on the house, and his other creditors are hounding Basham for tens of thousands of dollars in unpaid bills. Above all, Anaya seems baffled that he will likely spend the next two decades in prison for doing something that isn't specifically forbidden by federal law. "If it takes me never building another compartment again for me to get out of here, that's what I'm willing to do," he says. "But I think I should be able to." As he waits for his appeal to be heard, Anaya is trying to earn money to help support his family. He applied to work as a mechanic in Victorville's motor pool but was rejected as a security risk. He instead started his own business, fixing his fellow inmates' radios. Years ago his childhood junk drawer was filled with circuit boards. Today, his prison locker overflows with spare parts.

Where to Find Interesting Podcasts and Listen

1 Where to Find Interesting Podcasts 2 How to Listen to Podcasts p Listening to Podcasts on Android p Features of Google Podcasts App p Listening to Podcasts on iPhone and iPad p Listening to Podcasts on PC p Listening to Podcasts on the Web 3 Similar posts: Here, you can learn about podcasts, how to listen to podcasts and where to find interesting podcasts that worth your time. Where to Find Interesting Podcasts Podcasts are audio presentations on the internet. They are usually made available by means of download or stream. You can easily find one free on the web through your browser or via iTunes or other popular audio streaming websites and apps. Podcasts can span across different topics and niches. It can be a show where you can have a host and guest speaker, or a fix/ explanation of an idea or a comedy just to make listeners laugh. Just like you read on blogs and see videos on vlogs, a podcast is just an equivalent of those. Only that you can only listen to it. Also, See: Amazingly, anyone can make or listen to a podcast as long as they have a recording device and a good internet connection that can be used to send the recording online. Podcasts are made in audio mp3 format types and are distributed online with the use of an RSS feed. Once a listener subscribes to a podcast, they are automatically downloaded to their device and can readily listen to them. Podcasts can be found on iTunes or the Apple Podcasts app for people using a Mac, iPhone, iPad, Apple TV, or any other apple hardware device. Besides, you can find podcasts on Spotify, a strong competitor of iTunes in this regard, Google Podcasts for free trending podcasts, SoundCloud and other audio streaming platforms. p RELATED:Treasure Seeking Seelie in Genshin Impact Podcasts help to communicate thoughts, information, and experiences to listeners just by listening to some voices. No video whatsoever on the screen. Simply play and start listening. It is that easy! How to Listen to Podcasts Listening to a podcast is very easy. You can either use your Android or iOS device or directly from the web if you find it more convenient. Here is how: Listening to Podcasts on Android If you use an Android device, there are tons of apps that can be used. One of the most reliable in this category is the Google Podcasts app which can be used to discover and listen to trending podcasts for free. Features of Google Podcasts App You can install the Google Podcasts app and start listening for free. Other podcasts apps for Android are Spotify, Apple Music, and several others on the Google Play store. Listening to Podcasts on iPhone and iPad To listen to podcasts on an iPhone or iPad, you only need the Apple Podcasts app. You can have access to all the podcasts on iTunes without browsing the entire library. By specifying your interest, Apple Podcasts will let you listen to audio content on iTunes by helping you find the shows you really want and subscribe to.  Apple Podcasts can be installed for free on all versions of iOS if it hasn’t been installed by default already. p RELATED:How to Fix Poke Transporter Not Working Listening to Podcasts on PC To get to listen to podcasts on macOS and Windows 10, you can choose iTunes, the perfect app for listening to podcasts. iTunes app will help you to find recommendations, top episodes, and top podcasts around. iTunes is often called the home of podcasts because it can get them to come from a lot of sources. Listening to Podcasts on the Web The likes of Spotify, SoundCloud and other related websites will let you listen to a podcast using your web browser. Simply enter the website URL and start playing. However, some of the websites require you to have a free account to use their services and you might have some interrupting your listening with audio adverts. Also, See: Found this post helpful? Kindly share with friends. About Latest Posts Latest posts by Goodness (see all) Why is my TikTok post not showing up? - June 3, 2023 How to Fix Twitter Not Showing Tweets - June 3, 2023 How to Fix Favorite Sounds Not Showing on TikTok - June 3, 2023

Apify starts a web scraping academy

Apify Documentation Mold our tools any way you want to scrape websites or automate repetitive tasks. Find the solution to your task here or use the search box above.

What the Hell Is Water?

Welcome to all our new subscribers! I write a weekly newsletter about anything, as long as it’s interesting. span click here to share on Twitter to help us grow! There are these two young fish swimming along and they happen to meet an older fish swimming the other way, who nods at them and says “Morning, boys. How’s the water?” And the two young fish swim on for a bit, and then eventually one of them looks over at the other and goes “What the hell is water?” Taking inspiration from the continued interest in the speech excepts over the last few weeks, today I am pulling from David Foster Wallace ’s This is Water . David Foster Wallace was born in 1962 in Ithaca, New York, to a philosophy professor and English teacher, though he was raised in Illinois (and was actually a regionally ranked tennis player there). He became a professional writer and teacher, known for his satirical and sometimes darkly humorous review of American consumption-based culture. His first novel, The Broom of the System , was written while an undergraduate at Amherst College. Arguably his most famous novel, Infinite Jest , was published in 1996. In addition to writing, Wallace taught creative writing at Emerson College, Illinois State University, and Pomona College. Suffering from depression since his early 20s, Wallace dealt with bouts of alcoholism, drug use, and stays in psychiatric hospitals on drug detox programs. He took his own life in 2008. Wallace delivered this speech, now known as This is Water , as a commencement address in 2005 at Kenyon College . It appears to be the only speech Wallace gave outlining his views on how to approach life generally. Time magazine has ranked  This Is Water  among the best commencement speeches ever delivered, and in 2009 it was compiled and augmented  into a short book . In the speech, Wallace digs into higher education and what the true value of it might be. He suggests that education’s end-game purpose is not to teach students facts and figures or fill their heads with knowledge — but instead, it brings the most value to students by teaching them how to consciously choose how and what to think. The topics Wallace worked through in this speech were further explored in his final novel,  The Pale King , which was published in 2011, uncompleted, following his suicide in 2008. Yes, the audio of the speech can be found in the SoundCloud links below. The entire speech transcript can be found on my main website here . Audio, Part 1: Audio, Part 2: If you have made it this far, please take a moment to share the article with someone that might find it interesting — I appreciate your support. The point of the fish story is merely that  the most obvious, important realities are often the ones that are hardest to see and talk about . Stated as an English sentence, of course, this is just a banal platitude, but the fact is that in the day to day trenches of adult existence, banal platitudes can have a life or death importance, or so I wish to suggest to you on this dry and lovely morning… The point here is that I think this is one part of what teaching me how to think is really supposed to mean. To be just a little less arrogant. To have just a little critical awareness about myself and my certainties.  Because a huge percentage of the stuff that I tend to be automatically certain of is, it turns out, totally wrong and deluded. I have learned this the hard way, as I predict you graduates will, too. As I’m sure you guys know by now, it is extremely difficult to stay alert and attentive, instead of getting hypnotized by the constant monologue inside your own head (which may be happening right now). Twenty years after my own graduation, I have come gradually to understand that the liberal arts cliché about teaching you how to think is actually shorthand for a much deeper, more serious idea:  p learning how to exercise some control over how and what you think The point is that petty, frustrating crap like this is exactly where the work of choosing is gonna come in. Because the traffic jams and crowded aisles and long checkout lines give me time to think, and  if I don’t make a conscious decision about how to think and what to pay attention to, I’m gonna be pissed and miserable every time I have to shop because my natural default setting is the certainty that situations like this are really all about me.  About my hungriness and my fatigue and my desire to just get home, and it’s going to seem for all the world like everybody else is just in my way. Or, of course, if I’m in a more socially conscious liberal arts form of my default setting, I can spend time in the end-of-the-day traffic being disgusted about all the huge, stupid, lane-blocking SUV’s and Hummers and V-12 pickup trucks, burning their wasteful, selfish, 40-gallon tanks of gas, and I can dwell on the fact that the patriotic or religious bumper-stickers always seem to be on the biggest, most disgustingly selfish vehicles, driven by the ugliest – [as the crowd starts to clap] this is an example of how NOT to think, though – most disgustingly selfish vehicles, driven by the ugliest, most inconsiderate and aggressive drivers… You get the idea. p tends to be so easy and automatic that it doesn’t have to be a choice . It is my natural default setting. It’s the automatic way that I experience the boring, frustrating, crowded parts of adult life when I’m operating on the automatic, unconscious belief that I am the center of the world and that my immediate needs and feelings are what should determine the world’s priorities. The thing is that, of course, there are totally different ways to think about these kinds of situations. In this traffic, all these vehicles stopped and idling in my way, it’s not impossible that some of these people in SUV’s have been in horrible auto accidents in the past, and now find driving so terrifying that their therapist has all but ordered them to get a huge, heavy SUV so they can feel safe enough to drive. Or that the Hummer that just cut me off is maybe being driven by a father whose little child is hurt or sick in the seat next to him, and he’s trying to get this kid to the hospital, and he’s in a bigger, more legitimate hurry than I am. It is actually I who am in his way. Or I can choose to force myself to consider the likelihood that everyone else in the supermarket’s checkout line is just as bored and frustrated as I am, and that some of these people probably have harder, more tedious, and painful lives than I do. Again, please don’t think that I’m giving you moral advice, or that I’m saying you are supposed to think this way, or that anyone expects you to just automatically do it. span Because it’s hard. It takes will and effort p But most days, if you’re aware enough to give yourself a choice, you can choose to look differently at this fat, dead-eyed, over-made-up lady who just screamed at her kid in the checkout line. Maybe she’s not usually like this. Maybe she’s been up three straight nights holding the hand of a husband who is dying of bone cancer. Or maybe this very lady is the low-wage clerk at the motor vehicle department, who just yesterday helped your spouse resolve a horrific, infuriating, red-tape problem through some small act of bureaucratic kindness. Of course, none of this is likely, but it’s also not impossible. It just depends on what you want to consider. If you’re automatically sure that you know what reality is, and you are operating on your default setting, then you, like me, probably won’t consider possibilities that aren’t annoying and miserable. But if you really learn how to pay attention, then you will know there are other options.  It will actually be within your power to experience a crowded, hot, slow, consumer-hell type situation as not only meaningful but sacred, on fire with the same force that made the stars: love, fellowship, the mystical oneness of all things deep down. Not that that mystical stuff is necessarily true.  The only thing that’s capital-T True is that you get to decide how you’re gonna try to see it. p You get to consciously decide what has meaning and what doesn’t . You get to decide what to worship. If you worship money and things, if they are where you tap real meaning in life, then you will never have enough, never feel you have enough. It’s the truth. Worship your body and beauty and sexual allure and you will always feel ugly. And when time and age start showing, you will die a million deaths before they finally grieve you.  p It’s been codified as myths, proverbs, clichés, epigrams, parables; the skeleton of every great story . The whole trick is keeping the truth upfront in daily consciousness. Worship power, you will end up feeling weak and afraid, and you will need ever more power over others to numb you to your own fear. Worship your intellect, being seen as smart, you will end up feeling stupid, a fraud, always on the verge of being found out.  But the insidious thing about these forms of worship is not that they’re evil or sinful, it’s that they’re unconscious. They are default settings. They’re the kind of worship you just gradually slip into, day after day, getting more and more selective about what you see and how you measure value without ever being fully aware that that’s what you’re doing. This kind of freedom has much to recommend it. But of course, there are all different kinds of freedom, and the kind that is most precious you will not hear much talk about in the great outside world of wanting and achieving.  The really important kind of freedom involves attention and awareness and discipline and being able truly to care about other people and to sacrifice for them over and over in myriad petty, unsexy ways every day. That is real freedom. That is being educated, and understanding how to think. The alternative is unconsciousness, the default setting, the rat race, the constant gnawing sense of having had, and lost, some infinite thing… It is about the real value of a real education, which has almost nothing to do with knowledge, and everything to do with simple awareness ; awareness of what is so real and essential, so hidden in plain sight all around us, all the time, that we have to keep reminding ourselves over and over: It is unimaginably hard to do this, to stay conscious and alive in the adult world day in and day out. Which means yet another grand cliché turns out to be true: your education really is the job of a lifetime. And it commences, now. If you made it all the way through the article, please take a moment to share it with someone that might find it interesting, or consider becoming a supporter of the newsletter by being a paying subscriber — I appreciate your support and interest in the newsletter very much. Take care and have a great rest of the week, But What For? Writing about anything, as long as it’s interesting

Jerry Lee Lewis' marriage to a 13-year-old wrecked his career (2008)

Among teenagers of a musical bent, there was much anticipation 50 years ago this week. Jerry Lee Lewis, an American rock and roll singer with long blond hair who played a frenetic boogie woogie piano while standing up, and often with one foot on the keyboard, was on his way to Britain for a six-week tour. This may not seem like a big deal today, as rock musicians criss-cross the Atlantic all the time, but in May 1958 it was thrilling. Whole lotta trouble goin' on: Jerry Lee Lewis with child bride Myra To us, that first generation of rock fans, this guy was the real thing. And that was important, because, having been completely overlooked by Elvis Presley who’d never come to Britain (and who was by then in the U.S. Army, anyway), there was a feeling that we were getting everything second-hand and missing all the fun. True, we’d had a couple of would-be early rock stars of our own, but they were limp counterfeits like Tommy Steele, who already seemed to have one eye on becoming the dreaded all-round entertainers. Jerry Lee Lewis, however, or, "The Killer", as he was known, had enjoyed two classic worldwide hits with Whole Lotta Shakin’ Goin’ On and Great Balls Of Fire, and had even appeared in a Hollywood rock film, High School Confidential. Nor was he middle-aged like Bill Haley. He was young and vital. Could he possibly live up to his advance billing, those of us who bought the music papers wondered, as we read about him on our way to school. Would he be the wild man of the Louisiana swamps we’d been led to believe? No sooner had he landed at Heathrow than we had our answer, in no small part due to the inquiries of a Daily Mail reporter called Paul Tanfield. Meeting the star at the airport, Tanfield noticed that there was a very young girl in The Killer’s party. Tanfield asked whom she might be. "I’m Myra," answered the girl. "Jerry’s wife." Tanfield was astonished. "And how old is Myra?" he asked Jerry Lee. 'The Killer': Jerry Lee Lewis' nickname "Fifteen," the singer replied, obviously thinking that sounded suitably mature. It wasn’t. Despite Lewis’s assertions that Myra was "a grown woman", as far as Britain was concerned, she was below the age of consent. The headlines the next day were not good for the star’s first day in Britain. But they were about to get much worse when it was quickly discovered that Lewis, 22 at the time of the wedding, had been lying. Myra wasn’t 15. She was 13, and, therefore, absolutely not a "grown woman". What’s more, she was the singer’s first cousin once removed. And if that wasn’t enough, it was also revealed that he may have been bigamously married to her, since he hadn’t yet become divorced from his second wife, whom he’d married at 17, having wed his first wife at 14. If you’re becoming confused, think how we must have felt back in 1958 as the hillbilly courting behaviour of some citizens of America’s Deep South unfolded in our newspapers. We'd heard about the phenomenon of the child bride in fiction from the Tennessee Williams’ play and the film Baby Doll. But buttoned-up, respectable, repressed Fifties Britain had never come across the real thing before. With Jerry Lee, the Louisiana swamps had exceeding all expectations in what they had thrown up. Goodness gracious, as the man himself was wont to sing. This furore soon was great balls of fire! In this way began one of the most extraordinary episodes in the history of rock music — and, let’s face it, there have been quite a few. Right from the beginning, rock and roll music had been soaked in scandal, perhaps not too surprisingly when it’s remembered that the actual words "rock and roll" had been, in black American nightclubs, a euphemism for sexual activity long before they became associated with music. Legendary: Their marraige portrayed in 1989 film Great Balls Of Fire, with Dennis Quaid as Jerry and Winona Ryder as Myra So, when the music swept the world a couple of years earlier, teachers, preachers, parents and pundits alike had been quick to fulminate against  the youthful, on-stage gyrations of Elvis Presley, describing them as obscene, and to read into the lyrics of rock songs a lewd carnality which was probably accurate but being missed by most young fans. Up to this point, however, most of the outrage against rock had happened in America. Now, as Jerry Lee Lewis and Myra arrived in London, a storm of outrage erupted here, too. And instantly the fashionable Westbury Hotel in London’s Mayfair, into which The Killer’s retinue was booked, found itself besieged by competing armies of fans, the Press, police and outraged citizens. To start with, Lewis seemed to find it difficult to understand what all the fuss was about. In fact, initially he was quite pleased with all the publicity he was getting. While, for her part, Myra was happy watching children’s television in their suite, chirpily telling anyone who would listen that although her husband had given her a red Cadillac, what she really wanted was a wedding ring. Were this to happen today, any star would instantly surround himself with a legion of publicists who would do their utmost to put a positive gloss on the situation — not the easiest of tasks, I have to admit. Come to think of it, just about impossible. But those were less sophisticated times when it came to media manipulation. The best thing to do, Jerry Lee decided, was to get on with his tour as if nothing had happened, and, since he maintained he was a God-fearing country boy, to ask the good Lord for help. Consequently, it is said, he and his whole entourage fell down on their knees and prayed for a full hour before he took the stage at the Gaumont State, Kilburn, North London. For some reason, God doesn’t seem to have been listening — but then in the Southern states where Lewis came from, many people believed that rock and roll was the Devil’s music. Whatever the reason, nothing stopped The Killer, dressed in what was described witheringly in one newspaper as a "custard-coloured jacket", making his British debut to a half-full theatre with a performance that was repeatedly interrupted by whistles and boos and cries of "cradle snatcher" from the audience. 2005: Jerry in Essen, Germany, while touring Europe with Little Richard and Chuck Berry Off stage, things were getting much, much worse. On learning of Myra’s age, the police had turned up at the Westbury Hotel to interview the star and his bride, after which their notes were passed on to the Director of Public Prosecutions to see if any British laws had been broken. Meanwhile, in the House of Commons, the Home Office minister, Iain Macleod, was called upon to answer questions from MPs. Jerry Lee thought he could struggle on and win the fans round. By now, however, the posh Westbury Hotel had had enough. The star was asked to leave. Desperately, Lewis and his manager tried to explain that it wasn’t that unusual for girls of 13 to marry in Mississippi, and that the marriage to Myra couldn’t have been bigamous, because at the time of Jerry Lee’s second marriage he’d still been married to his first wife. Thus the second marriage had been null and void, and as he was now divorced from the first wife, everything was fine and dandy! Neither the newspaper reporters nor the Rank and Grade organisations, in whose theatres the Jerry Lee concerts were to have taken place, were convinced. After only three appearances, the tour was cancelled, and Jerry Lee and Myra, his managers and hangers-on, were back on a plane to America. A little less than nine months later, Myra gave birth to a boy. The maker of some classic rock hits he might have been, but The Killer’s career never properly recovered. He became a musical pariah. And after disc jockeys around the world refused to play his records, he never had another big hit. From $10,000-a-night shows, he was reduced to earning $100 a night. Myra divorced him in 1970, after 12 years of marriage when she was all of 25, became an estate agent and wrote her autobiography, Great Balls Of Fire, which was filmed with Dennis Quaid as Jerry Lee and Winona Ryder as Myra. The scandal of 1958 proved, however, to have lasting effects in quite different ways. It may have been coincidental, but very quickly attempts were made in America to clean up the image of rock and roll. Payola investigations were begun and several famous disc jockeys were revealed as having taken bribes to play records. And when the mighty Elvis himself fell in love with a 14-year-old girl, Priscilla Beaulieu, the following year, steps were taken to make sure that not a word of scandal leaked out. As for us here in Britain, within a few months, we’d come up with our own pop star, someone whose reputation was, and would remain, cleaner than clean. His name was Cliff Richard. One thing, however, couldn’t be denied. Although the affair had ruined the career of Jerry Lee Lewis, it had also made him very famous, infamous, actually. And as the Fifties rolled into the Sixties, rock Svengalis-would soon see that the right kind of scandal, carefully managed and well publicised, could work wonders for the careers of rock stars. Five years later, Andrew Loog Oldham, the young manager of the Rolling Stones, would give a masterclass in how this could be done. While the nicely-turned out Beatles began to find fame by sticking carefully, in public, anyway, to the goody-goody script neatly mapped out for them by their manager Brian Epstein, Oldham did everything he could to grab outrageous headlines for the five, gurning, rebellious Rolling Stones. Stunt followed stunt, from urinating in public, to singing more blatantly than anyone else about sex. If there was a rule to be broken, the Stones broke it, and in the process built legends for themselves as the bad boys of rock and roll. Indeed, by the mid-Sixties it had got to the point that just about anything could be believed about them, whether true or not. There never was a Mars Bar at that party with Marianne Faithfull down at Keith Richards’ house in 1967, but anyone who had followed their careers in the newspapers believed there was, and the band didn’t mind at all. Confrontational in the extreme, they milked scandal about themselves for all it was worth. Of course, as with Jerry Lee Lewis and every other rock attraction, there were always a lot of girls involved, though none as young as Myra Lewis — at least, not until, having left the band, 47-year-old bass player Bill Wyman fell for 13-year-old Mandy Smith. He married her when she was 18. By the Seventies, outrageous behaviour had become synonymous with rock music, as groups vied with each other for publicity. Some set their amplifiers on fire on stage while others drove cars or pushed grand pianos into swimming pools. It was all about creating controversy, getting headlines, and nothing to do with music. Thus the punk group the Sex Pistols swore on television, Ozzy Osbourne was alleged to have bitten the head off a bat and Madonna disgracefully mimed having sex on Top Of The Pops. And so it goes on, as every new generation of stars struggles to be noticed in the rush. Sometimes, of course, publicity isn’t sought, as both Michael Jackson and Phil Spector have recently found in lurid and tragic circumstances. But, believe me, the bigger the headlines about rock music the greater the stepping stones to stardom. Quite what Jerry Lee Lewis thinks about the behaviour of some of today’s musicians would be worth knowing. Today, at 73, after suffering from bouts of alcoholism and depression, he still tours. Appreciated by some stalwart fans as one of the pioneers of rock and roll, he is remembered by most of us, if at all, for that week in London 50 years ago when his bizarre marital life shocked the nation.

New Tutorial: Using PHP Composer in the WordPress Ecosystem

This website use security service to prevent unauthorized access and online attack. Your action might trigger the security solution. Please contact the site administrator to resolve this issue.

Apple’s Long Journey to the M1 Pro Chip

Steven Sinofsky p Follow Published in Learning By Shipping p 10 min read p Oct 19, 2021 -- 6 Listen Share This twitter thread explores the journey from the original Mac to today’s M1 Pro-based MacBook and the history behind such a mind-blowing innovation. Apple’s M1 Pro/Max is the second step in a major change in computing. What might be seen as an evolution from iPhone/ARM is really part of an Apple story that began in 1991 with PowerPC. And what a story of innovation 💡 1/ [Quick thoughts] 2/ If you studied Computer Science in the 80s then you were deep into the raging debate of RISC v. CISC. And what a debate it was. Out of that debate emerged an implementation at IBM, the POWER processor/instruction set. And an SV company MIPS. 3/ PowerPC was a huge investment from IBM — an effort to regain end-to-end control of computing, starting with workstations. They had no software platform really (though Unix was all the rage for workstations and OS/2 all the hope) so the big bet was on Windows NT. 4/ An unlikely partnership was formed between IBM, Apple, Microsoft. This followed the breakup of the alliance known as AIM between Apple, IBM, and Motorola. Recall, Moto was the provider for all Apple’s chips going back to the start of Macintosh. 5/ Apple relied on Motorola including the launch of the PowerBook. But even in the latest 68030 Mac was woefully far behind Intel PCs. While the 68030 was a fantastic chip, Mac software didn’t fully exploit it and WinTel was iterating rapidly to the 486 [6 months later than ‘030 Macs and was significantly more powerful]. Many at Apple, such as @gassee firmly believed Apple needed to gain full control of computing, but becoming a Chip maker cost billions. 6/ This is really key in terms of understanding the present. Apple was essentially left hanging by a partner for chips, when their core deliverable to customers was a computer. That seemed an impossible situation. 7/ Over in x86 land, compute speeds and megahertz kept rising with Moore’s Law. Intel’s investment and excellence in manufacturing made it all but impossible for anyone to catch up, especially making a similar CISC product. 8/ Apple moving to x86 would put them in direct competition with everyone else. That seemed the opposite of the right way to go, much as licensing Macintosh System seems to be wrong. 9/ This led to a bet on PowerPC and the release of the first Power Macintosh 6100 in *March 1994*. The project was a very bumpy ride with IBM but also an enormously difficult software project. 10/ The Mac System needed to be ported to the PowerPC chipset, including all the hardware and peripherals. It also meant the whole ecosystem needed to move as well, but that ecosystem wasn’t healthy and had started to slow down Mac and favor Windows. 11/ To ease transition, power of PowerPC was used for 68k emulation. This allowed lagging apps (Photoshop, which took something like 2 years to release as a native app) to at least run on new hardware, but performance was marginal at best. BUT Apple learned a great deal in this transition — tooling (like Universal), emulation, and compilers. 12/ The results were lukewarm when it came to performance, especially versus x86. In the meantime, developers were investing heavily in x86 and Intel was investing billions in scale. The gap at launch would get worse, not better over time. 13/ There was little Apple could do for the next few years. Essentially the price/performance gap would continue to become even more bleak. Then Steve Jobs returned. 14/ Of course first came the most famous PowerPC Macintosh of all time. By focusing on consumers and the internet, the weakness of PowerPC could be ignored. The industrial design was a hit too. The G3 then G4. 15/ In the meantime, Apple began to use ARM processors for the iPod. ARM was a RISC design, but had only achieved success in small devices and peripherals. Intel and many other companies were all ARM architecture licensees. This is how Apple got into chips. No one worried. Note I should have mentioned the Newton here as it was clearly the first ARM device from Apple, though it was another generation of people and from what I gather there was not a lot of continuity when it got to iPod unlike the OS evolution. 16/ The end-to-end control of iPod was exactly what was needed to bring together hardware, software, and internet services. First some custom chip features, then more, then more…Of course as we know now, that also marked the resurgence of Apple. 17/ Rebuilding Macintosh began as well. By this time NeXTStep was already ported to Intel. The work to get the rest of Mac working involved a lot of knowledge about chipset transitions from NeXT and Apple history. Note: Many senior Apple engineers worked at NeXT, and several of them worked at Apple previously. 18/ There were no choices for chipsets for “computers” at the time, but of course Apple was clearly going to use ARM for the iPhone (and iPad) project which itself has a long and interesting story (out of context here). 18.1/ Was asked: Apple using Intel for phones, how Intel famously declined. I’ve heard first-hand from both sides and honestly not sure how serious Apple was or how uncommitted Intel was. Also, within Apple it was a debate of sorts, which seems crazy given what I saw on Windows. 19/ At the 2005 WWDC, jobs announced that Apple would be transitioning to Intel. Many said this is something that should have happened long ago. It really was the only option (IBM was not going to invest “Intel money”) 20/ So while the phone was going on and ultimately released on ARM in 2007, the first Intel Mac was released in January 2006. That was essentially a PowerBook but with Intel inside. If this looks familiar it is because that’s exactly how Apple did the M1. This is a huge learning point — the DTK 2.0 (the first one was for the transition to x86) was essentially an iPhone inside a Mac Mini chassis. It was proof of concept. Then later the first MacBook Air and MacBook Pro with M1 were also chassis look-alikes. 21/ In 2008, proving it could innovate and even do with with Intel, Apple introduced the MacBook Air. It was the first laptop on a new Intel chipset (though everyone would eventually have this for Windows, no one built a MacBook Air-like PC). Until Ultrabooks there was no PC. 22/ A huge part of the Intel transition was bringing up hardware, seeding developers, universal binary, and emulating PPC instructions. If this all sounds familiar it is because it is 2.0 of a process that was done in the 68k->PPC transition. It was much smoother. Same process, just iterated. And apps came on board as native much faster. Photoshop, Office, etc. all came much sooner. 23/ As Apple progressed with ARM it became clear that the breadth of the ecosystem was a benefit for ARM but was holding Apple back in terms of innovating. iPhone success gave Apple the resources to realize its chipset dream. Thus the System on Chip from Apple began. Note. In 2009 Apple acquired PA Semiconductor. I should have included this because it was a huge step, though many at the time were quite puzzled by it. This story below shows how confused most were. Apple of course was not. 24/ The M1 chip was a realization of all the iPad and iPhone work (sensors, OS port, security, power management, graphics, and more). The M1 not only aimed at fixing what ailed Intel, but also PPC. It was learning from the past decade+. 25/ When you look at M1 Pro/Max today it is tempting to think of this in terms of performance, but performance per watt AND integrated graphics AND integrated memory AND integrated application processors is innovation in an entirely different direction. Just the beginning. 26/ Here’s a thing about “Laptops versus Phones”. The Phone is the computer for everyone around the world now. Laptops (and desktops) are specialized devices for work. About 400M people really use/need laptops for work. That’s what M1 is for and why Apple does not need to stress about pricing like it did in the 1990s — essential tools for highly paid information workers are worth the money. 27/ The number of laptops won’t grow, but it is likely Apple will continue to take share from Windows (as will Chromebooks). At 275M units a year, laptops are big but serving this base of 400M. Phones serve everyone including them. That’s where software innovation is for masses. 28/ That’s why this transition to M1 is so fascinating. Back when Apple went on its own with the A-series chips, one could easily be concerned that they would end up in the same place as PPC — not enough volume to win against Qualcomm/Samsung doing their ARM designs. 29/ Apple, by virtue of being vertically integrated, raced ahead. Because of their units and revenue/R&D investment they are in a globally unique perspective. Today, Qualcomm is closer to Intel than it is to Apple Si/Ax/Mx. 30/ Even though Intel serves Windows, ARM/QC/Sam serve Android, they all must serve myriad of OEMs. It appears as though that point of openness (v say at the s/w and service level) is a real constraint on innovation. Partners don’t all want to make the same device, for example. 31/ At the same time, Apple itself has developed a whole family of operating systems. These are not just related, but vastly similar enabling a huge ecosystem opportunity. THIS was Microsoft’s vision going back to early 1990s. Slide from 1992 PDC. 31.1/ This is the slide that goes with 31 that was left out/ The Microsoft history of Scalable Windows. This is the first slide from 1992 PDC. Over the next few years it would grow at the low end (Windows CE, Wallet) and high end (High Performance Computing w/NT and Itanium [sic]). 32/ The M1x capabilities of shared memory, SoC that isn’t just smaller but has so many aux functions, Pro Res, super fast SSD, even multiple TB ports — all these things require deeply integrated software (from the chipset to the experience). 33/ Some might say only 1% of people (of the 400M) even need these. 2 things. 34/ Across phone, tablet, tv, watch, speaker, laptop, desktop, there’s a platform of capabilities are unmatched even in one device group, let alone the whole spectrum. It would be like if Ferrari also made mass market passenger cars and electric bikes. Unprecedented situation. 35/ I’d be remiss if I didn’t say these weren’t perfect devices. Why no ethernet in the giant power brick like iMac? Why no cellular option? Yeah FaceID doesn’t fit but soon? Oh and people harping about the notch, STOP! PS/ This thread is about context, not the whole story. There are a lot more details, even books, about this particular transition. Here’s a deep overview of PowerPC to show what apple was after, and challenges.

Uber and Lyft lose appeal in case classifying California drivers as employees

Missed the GamesBeat Summit excitement? Don't worry! Tune in now to catch all of the live and virtual sessions here. (Reuters) — A California appeals court on Thursday unanimously ruled against ride-hailing companies Uber and Lyft, saying they must reclassify their drivers in the state as employees. While the ruling does not take effect before a November 3 company-sponsored ballot measure that will give voters the chance to decide on the future status of gig workers, it narrows the companies’ options should their ballot measure fail. The case emerged after California implemented a law known as AB5 that is aimed at reclassifying ride-hailing, food delivery, and other app-based workers as employees entitled to benefits such as unemployment insurance and minimum wage. California in May sued Uber and Lyft for not complying with AB5. A California judge in August ordered the companies to reclassify their drivers as employees, a ruling the companies appealed under the threat of leaving the state altogether. The appeals court on Thursday upheld the ruling. The judges said in a 74-page ruling that Uber’s and Lyft’s misclassification caused irreparable harm to drivers, who as independent contractors miss out on employee benefits. Remedying those harms more strongly served the public interest than “protecting Uber, Lyft, their shareholders, and all of those who have come to rely on the advantages of online ride-sharing,” the ruling said. Lyft and Uber in a statement said they were considering all legal options, including an appeal. “This ruling makes it more urgent than ever for voters to stand with drivers and vote yes on Prop. 22,” Lyft said, referring to the November 3 ballot measure, which would repeal AB5 and provide drivers with more limited benefits. “Today’s ruling means that if the voters don’t say Yes on Proposition 22, rideshare drivers will be prevented from continuing to work as independent contractors, putting hundreds of thousands of Californians out of work and likely shutting down ride-sharing throughout much of the state,” Uber said. (Reporting by Kanishka Singh in Bangaluru and Tina Bellon in New York. Editing by Daniel Wallis.) VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.

Developer won’t get hit by a bus, they’ll get hired by Netflix

Your developer won’t get hit by a bus. They’ll get hired by Netflix! August 3, 2020 You may have heard of the “bus factor.” The strategy and processes designed to cover for the sudden loss of a teammate. The story goes: you want more than one person in your business to have domain knowledge because they might get hit by a bus! Your business would lose all that experience and expertise in one tragic accident. The premise in isolation sounds ridiculous, but mitigating the “bus factor” is serious business. Everything gets a lot easier if you select the right software and framework, primarily if you use Ruby on Rails. Rails itself is a full-stack framework that has a best practice for every piece of a web application. If you are committed to doing things “the Rails way,” you cut the total cost of ownership by A LOT. “Staying on the Rails” makes any new developer productive almost immediately (and will keep them “out of the weeds”). Sticking to the standards evangelized by the community and adding proper documentation (possibly in the form of useful tests) makes onboarding a breeze, which creates more productive developers, which makes hiring more accessible, which reduces your “bus factor.” The only guaranteed way to save money is to spend it on reducing your “bus factor”. By investing in making the code of your application more straightforward and easier to understand, you make it easier to add developers and reduce the future occurrence of bugs. You’ll make working on your code more enjoyable (reducing the likelihood your developers will want to leave), and possibly most importantly: reduce the barriers (time/money) to building new features. If you haven’t been keeping an eye on this, and find yourself in a position to hire someone, then these tasks will be their first job every time you hire. Thankfully, I have no personal experience with a colleague getting hit by a bus, but I know a few that have been hired by Netflix. Remember to look both ways before crossing a street and never underestimate the amount of money the FAANGs have to poach your talent. It is effortless to lose good talent. When (not if) your developer leaves to make $500k+ a year in the valley, you’ll be in a much better position with a lower bus factor to quickly fill the gap and keep moving forward. ‍ ‍ p The Scammer's Playbook: AI, Voice Cloning, and the Timeless Art of Deception As AI continues to evolve, so too does the concern that we'll be effortlessly swindled out of our personal and financial information by cyber tricksters armed with voice cloning technology. p Sustainability In Tech: The Next Frontier Sustainability is a concept that has been around for centuries. In modern times, the word sustainability has become synonymous with environmental causes. While that association is valid and important, there are other forms of sustainability that need conscious consideration. p The Robots Are Going To Take Your Jobs… And Make Them Better: Don’t Believe The Hype, AI Will Increase Jobs Conspiracy theorists crow that AI is going to take away all of our jobs and make the human worker obsolete, eventually leading to war and the decline of civilization as we know it. The robots will rule with a heavy mechanized hand, and the average person will be out on the street selling pencils from a cup for scraps of bread. A scary prospect indeed. p Understanding Ruby on Rails An easy breakdown of understanding Ruby on Rails.

Quantifying Racial Segregation by State and City

I look to a day when people will not be judged by the color of their skin, but by the content of their character. One of the disappointing aspects of race discussions in the media today, is that they predominantly focus on anecdotes and sensational examples. Stories are wonderful at illustrating truths, but are a horrible way of proving them. Is there a more quantifiable way of evaluating race relations in America today? One objective metric that comes to mind, is rates of interracial marriage, normalized by society’s level of diversity. This single metric encompasses a wealth of different information – including housing/education/career segregation, societal attitudes towards embracing diversity, and close social ties that bridge racial divides. And most conveniently, interracial marriages and their resulting children are explicitly surveyed by a number of sources, such as the US Census. This makes it far easier to quantify them, as compared to more nebulous signals such as “friendships” and “social circles”. One major disclaimer – it is certainly not my intention to embarrass anyone who is married to a partner of the same race. Indeed, the vast majority of people, regardless of race, marry someone of the same race. Often purely by coincidence. And other times because of people’s understandable desire to be with someone who shares their culture and world view – something often mentioned by people of color as well. The goal of this study is not to “call out” any individuals, but simply to examine wider societal trends. The best data source I could find was the US Census – specifically the American Community Survey. This survey provides racial demographic information, broken down by state and city. Unfortunately it doesn’t explicitly give us statistics for interracial marriage. But it does give us a different statistic which we can use as a proxy – the percentage of people who are biracial, broken down by each race-combination. To be sure, this is a flawed proxy. Not all couples have the same number of children. Not to mention that the percentage of biracial individuals, is a lagging indicator of today’s rate of interracial marriages. These drawbacks make it very hard to use the metrics below in any absolute sense. But they are still of great value when used in a relative context. The methodology we use to quantify racial integration is simple. In a world of complete racial integration, if X% of society is Black, then each person would have a X% likelihood of marrying someone who is Black. This implies that in a society that is: – X% White, – Y% Black, we would expect 2*X*Y/100 % of individuals in the subsequent generation to be White-Black biracial. Given the above, we can then score each society’s level of integration by computing the actual:expected biracial population ratio. For example, in a hypothetical colorblind society consisting of 90 White and 10 Black citizens, each person would have a 10% chance of marrying a Black person. This implies that of the resulting 50 marriages, 9 of them (ie, 18%) would be White-Black. We would then expect 18% of children in the following generation to be White-Black biracial. If the actual biracial demographic is 1%, our racial integration score would be 1/18 ~= 0.055 ~= 5.5% The above methodology has been coded into a software application, which parses the Census data and computes all the scores. Both the code and the dataset can be found here, and is readily available for you to modify and/or run directly. The above methodology has many weaknesses of course. Chief among them, it doesn’t take into account age. Ideally, we would be looking only at the 20-40 demographic, in order to best estimate the expected fraction of interracial marriages. And we should also be looking only at the 0-10 demographic, in order to estimate the actual fraction of recent marriages that are interracial. Or even better, we should be measuring directly the fraction of recent marriages that are interracial, without having to rely on any other proxy. In the absence of more precise data, we are forced to make do with the above approximations instead. More fundamentally, in order to truly quantify segregation, we should be looking at all social ties – not just marriages. It is possible for an African-American couple to be perfectly integrated into society, with friends, neighbors and colleagues of all backgrounds, and this methodology fails to capture that. Lastly, this study uses marriage as a proxy for social integration. But arguably, we should be looking at other measures such as wellbeing, societal attitudes, and levels of discrimination experienced. Just because someone is in an interracial marriage, does not mean that they feel comfortable and welcomed by the wider society around them. And the converse is true as well. Given the above limitations, this study should definitely be supplemented with other studies that explore this topic from different angles. The absolute scores computed are of minimal use on their own. But when used in a relative manner to compare different communities, it gives us some very objective, quantifiable and data-driven insights. To be honest, I was shocked by how much the states differed from one another. When looking at the states with the lowest scores, we see extremely low levels of interracial marriage. A mere 0.7% of people in Mississippi identify as White-Black biracial. Despite the fact that 58% of Mississippi is White and 38% is Black. In contrast, New Hampshire has the exact same biracial demographic, despite the fact that less than 2% of people in New Hampshire are Black. Looking at the states with the lowest scores, it seems to validate commonly held beliefs about the South. Of the 11 states with the lowest scores, every single one is a Southern state. States like Mississippi, Alabama and Louisiana have some of the highest Black demographic in the country – but some of the smallest White-Black biracial demographic. Of little surprise, the state that is most integrated, is Hawaii. The same state where Barack Obama grew up. And the same state which many say is closest to being post-racial. But the other results do hold quite a few surprises. Many states commonly thought to be progressive and diverse – New York, New Jersey, Illinois – are actually below-average. And California – commonly thought to be the most liberal and diverse state of all – is not even in the top-15. When looking at the states with the highest scores, what we actually see are the states that no one talks about. New Hampshire, Oregon, South Dakota, Maine, Washington, Utah, Vermont – these are the states with the highest scores in the mainland. Oregon, for example, has 1 White-Black biracial person for every 3 Black people in the state. Looking closer, the reason becomes clear. There seems to be a strong relationship between diversity and segregation. Ie, the more diverse a state is, the more segregated it is as well. Except for Puerto Rico, the highest scoring states all have extremely small Black demographics – less than 4%. The US as a whole has a 12.7% Black demographic. Every single state with an above-average score, has a smaller Black demographic. And except for Texas (which has a surprisingly low score), every state with a below-average score, has a larger Black demographic. It is worth noting though, that the above isn’t a universal rule. Texas has half the Black demographic as Delaware, but still scores lower. And Ohio has a four times larger Black demographic than North Dakota, but still scores higher. On the face of it, this contradicts prevalent assumptions that more diversity is associated with less racism. The above data implies that greater diversity is actually associated with more segregation. There are many possible explanations for this. One possible explanation is amicable self-segregation. People of color often report a preference for socializing with (and especially marrying) someone of similar background. If true, this means that when given the opportunity, people of color are more likely to socialize with (and marry) someone of the same race. However, when living in communities where almost everyone is White, they would be forced to integrate into wider society. An alternative, and less innocuous, explanation is the Racial Threat Theory – when living in a more diverse society, the dominant group is more likely to feel threatened, and to engage in prejudicial behavior. Racial threat theory proposes that a higher population of members of a minority race results in the dominant race imposing higher levels of social control on the subordinate race, which, according to this hypothesis, occurs as a result of the dominant race fearing the subordinate race’s political, economic, or criminal threat. Political scientist V. O. Key found that white voters in the U.S. South turned out at higher rates and voted more for conservative politicians in areas with high levels of African-Americans; Kety argued that whites felt threatened by African-Americans, thus becoming more politically motivated. Research has shown a strong association between the size of a state’s nonwhite prison population and the likelihood of that state enacting a felon disenfranchisement law, which supports a link between racial threat and the passage of such laws. A 2016 study by Harvard University political scientist Ryan Enos, which relied on a quasi-experimental design, found that when public housing projects in Chicago were removed over the period 2000-2004, turnout among white voters decreased substantially and white voters were less likely to vote for conservatives. This has recently even been posited as a major factor in the Capitol Riots: Counties with the most significant declines in the non-Hispanic white population are the most likely to produce insurrectionists. This finding held true, Mr. Pape determined, even when controlling for population size, distance to Washington, unemployment rate and urban or rural location. Further research is needed to determine which of the above best explains the segregation seen in states that are very diverse. It is commonly thought that cities, by virtue of their progressive populace, are more racially integrated than suburbs and rural areas. This may even explain why states like New York and Illinois have particularly low scores, despite the reputation of their flagship cities. Does our data validate this hypothesis? Looking at the above, we see a couple of data points that strongly dispute our hypothesis. Contrary to popular belief, it appears that large urban cities are even more racially segregated, compared to the state they are in. This can potentially be explained by the same relationship we found earlier – major cities tend to have larger Black demographics, and we’ve seen that segregation tends to be (but not always) more prevalent in more diverse communities. Ames (Iowa) exemplifies this trend. With a Black demographic that makes up a mere 1.3% of its population, and 1.8% more who identify as White-Black biracial, Ames trounces every major city when it comes to White-Black integration. The above numbers are all specific to White-Black race relations. But what about other races like Asians? How do they compare, and do the above patterns hold true there as well? Once again, Hawaii scores at the very top of the list, cementing its status as the most racially integrated state of all. With a White demographic of 24% and an Asian demographic of 38%, it even bucks the diversity-segregation trend that we’ve seen elsewhere – demonstrating decisively that diverse communities can indeed overcome segregation. And looking at the other states, we see that Asians have been integrated into mainstream society far more so than African-Americans. The US as a whole clocks in at 9.9 – almost double the 5.5 score that we saw earlier for White-Black integration. But the biggest surprise of all, are the states that we see at the bottom of the list – New Jersey and New York. Even California, boasting the largest Asian population, is significantly below average. In our earlier White-Black rankings, Mississippi Alabama and Louisiana were at the bottom, and had reputations for poor racial relations. However, New Jersey, New York and California, are hardly seen in the same light. One possible explanation is the correlation we saw earlier between diversity and segregation. But it doesn’t explain everything. California has a much larger Asian demographic than both New Jersey and New York, but is significantly more integrated. There are also a number of other states with outsized Asian demographics, such as Washington and Nevada, which score above average. Conversely, there are a number of states with below-average Asian demographics, such as Pennsylvania and South Dakota, which score significantly below average. With regard to White-Asian race relations, the correlation between diversity and segregation does seem to exist, but isn’t nearly as strong as what we saw earlier. A different explanation, is the recent influx of Asian immigrants, who have disproportionately settled down in New York, New Jersey and California. Given their recent arrival, it is expected that they would swell the Asian demographic in those states. However, since the biracial demographic is a lagging indicator, it would take many decades for it to catch up. Especially since many Asian immigrants have moved here with a spouse in tow. In order to truly quantify White-Asian integration in states disproportionately impacted by immigration, we would need to exclude immigrants from our study, or gather other more precise data. If any of the above sounds like a depressing indictment of our race relations, there is still much reason for hope. We need only compare today’s scores, to the same scores from 8 years prior. Most states, especially the ones with large Black populations, have seen tremendous progress over the past 8 years alone Looking at the nation as a whole, our level of integration has grown from 3.7 -> 5.5. A whopping 49% increase in less than a decade. We are making tremendous strides every decade, with every new generation. By the end of the 21st century, we may truly find ourselves living in a post-racial world. Share: h3 Loading...

Launching Steve.Ai – Fastest way to create Videos in seconds

The Only AI Tool To Create Animation Videos Using Text With our patented AI technology, you can make professional videos in MINUTES. See the MAGIC happen as the AI picks the right creative media assets for your Video. Sign Up For Free Trusted by leading brands across the world Video Making As Easy As Uploading Content Out of the box Content Creative Automation AI Engine Collaborate and Publish Solutions For Today’s Video Creation Needs Whether you are a beginner, an expert, or a professional video creator, we offer solutions to help you reach newer heights. Go Text to Video in a flash! Simply paste your text in the Script editor and produce engaging videos. See the AI picking the most relevant assets without breaking a sweat. Instantly Repurpose Blog to Video Move as fast as your customers. Convert your blogs to one or many bite-sized videos in seconds across channels. Paste the URL and see the AI work like a Humming bird’s wing-beat. Scale the power of Audio to Video Repurpose your audio files by converting them into thumb-stopping videos. Extract the text, build context, and convert to videos like Eminem’s Rap. Super Intent Assist the AI in choosing highily accurate assets for your videos with Keywords. Re-generate videos with a new context everytime. Ratings From Our Happy Customers 4.6 out of 5 5 out of 5 4.3 out of 5 Get Started For Free! AI Avatar Speak Make presentation videos and bring life to your videos with perfect lip-syncing AI Animated Avatars Choose from +10 templates and +100 professionals to animals to kids avatars Massive Assets Collection 100+ Million Assets library with everything in the world to pick from. Plus-size & Animal Characters, Epidemic Music Customize & Stay on Brand Leverage the advanced video editing suite to customize the videos, upload your own assets, and integrate custom branding. Nothing But Beautiful Produce stunning 4K quality videos in Minutes. With Steve.AI's latest cloud technology, you don't need costly machines to produce quality videos. Templates For Every Concept You don’t need any design skills to tell a captivating story Marketing Videos Explore Explainer Videos Explore People & Culture Explore Commercials Explore Educational Explore YouTube Explore Infographics Explore Thought Leadership Explore Greetings/Wishes Explore Try it for yourself No credit card required Personalize Videos For Who You Are Marketing Team News & Editorial Social Media Agencies Corporate Communication Enterprise Our Users Love Us "Wow.. Imagine having your Ai Marketing content creator? This is is exactly it" Ritch C User and Beginner-friendly: As soon as you click on ‘create’ to begin crafting your video, a helpful tutorial pops up to guide you through the entire process. The process of creating a video is intuitive and easy to do. Annalie You can easily get access to the stock video footage from Pexels and Pixabay to add variety to the templates offered. It makes the whole editing process a lot easier. Possible Jerry Steve.ai is easy to Use for anyone it Saves endless hours on time-consuming tasks like cutting scenes together or color correcting. this revolutionary AI technology automatically turns scripts into videos within minutes VEMPALI PAWAN KUMAR Become A Superstar Video Creator with Steve.AI Free forever, no credit card required. Get Started For Free!

Susan Kare Explains Macintosh UI Ergonomics on the Computer Chronicles (1984)

Your browser isn’t supported anymore. Update it to get the best YouTube experience and our latest features. Learn more

The Apollo Guidance Computer: A Kinder, Gentler Introduction

The Apollo spacecraft used for was really two different spacecraft, the and the .  The CM was used to get the three astronauts to the moon, and back again.  The LM was used to land two of the astronauts on the moon while the third astronaut remained in the CM, in orbit around the moon. Apollo 15 CSM(Command and Service Modules) Apollo 16 LM Each of the spacecraft needed to be able to navigate through space, with or without the assistance of the astronauts, and therefore needed to have a "guidance system".  The guidance system was developed by MIT's Instrumentation Lab, now an independent company known as the . An important part of the guidance system was the Apollo Guidance Computer—or just "AGC" for short.  On any given Apollo mission, there were two AGCs, one for the Command Module, and one for the Lunar Module.  The two AGCs were identical and interchangeable, but they ran different software because the tasks the spacecraft had to perform were different. Moreover, the software run by the AGC evolved over time, so that the AGC software used in later missions like Apollo 17 differed somewhat from that of earlier missions like Apollo 8. Considered just as a computer, the AGC was severely under-powered by any more-modern standards.  The basic characteristics of the AGC were: It is occasionally quipped—with perhaps greater wit than insight—that the AGC was more like a calculator than a computer.  But to say this is to grossly underestimate the AGC's sophistication. For example, the AGC was , so that it could seemingly run multiple programs simultaneously. Another important part of the guidance system was the Display/Keyboard unit—or just "DSKY" for short.  The AGC by itself was simply a box with electrical connections, without any built-in way for the astronaut to access it.  The DSKY provided the astronaut with an interface by which to access the AGC. The Lunar Module had a single DSKY, positioned between the two astronauts where it could be operated by either of them. The Command Module actually had two DSKYs.  One of the CM's DSKYs was only the main control panel, while the other was positioned near the optical equipment used to mark the positions of stars or other landmarks. The DSKY as a substantial piece of equipment in its own right: Perhaps the most important part of the guidance system was the Inertial Measurement Unit—or just "IMU" for short.  The IMU continuously kept track of the acceleration and rotation of the spacecraft, and reported this information back to the AGC.  By mathematically processing this data, the AGC could know on a moment-by-moment basis the orientation and position of the spacecraft. (Fabrizio Bernardini has contributed this alternate introduction to the AGC, from a poster presentation he made, and then was kind enough to give use both the poster, in Italian, and an English translation.  The words and photos are his, and I'll simply present them as-is.  Don't be surprised if there's some duplication of material.  Thanks, Fabrizio!) The first contract assigned within the scope of the Apollo program, a little after the decision by president Kennedy to reach the Moon before the decade was out, has been that for the computerized guidance and navigation system. Requirements for the system were clear and simple, but apparently unreachable at the time: build a small autonomous system, able to steer a crew from the Earth to the Moon and bring it back safely to the Earth. A giant leap for technology When in 1962 MIT's Instrumentation Laboratory began, under the direction of the legendary Dr. Charles Stark Draper (a pioneer of inertial navigation systems), the development of a computer to bring people from the Earth to the Moon, computers where in their childhood. They were giant electronic-tubes devices, big as rooms or even building floors, controlled by means of of perforated tapes or cards, with insignificant memory capacity compared to the personal computers that would be born at the end of the 1970's. With an engineering courage that has few precedents, designers (historical characters like Hal Laning Jr, Eldon Hall, Ramon Alonso, Albert Hopkins and many others) elected to use for the first time in history a newly born electronic component: the integrated circuit. Still far away from the microprocessor or microcontrollers of the 1980's, the only functions available were simple logic ones. Using thousands of these logic gates, it was possible to implement what would be called today the central processing unit of the computer and all support circuitry. For memory, since it was still impractical to build it with integrated circuits, the core technology (where small rings were used to store single bits) was adopted. being reasonable easy to miniaturize. Using the same technology, the program memory was made of tens of thousands of Permalloy-wire cores, hand woven with thin copper wire to represent the 1's and 0's of the computer software. These wire were to be woven by specialized workers at least one month before the launch of a mission. These critical decisions, taken in the era of the "dinosaur computers", were found to be winners in the following years, just seven, that saw men land on the Moon. The onboard computer developed for the Apollo program (and installed into both the Command Module and the Lunar Module, but with different software for each vehicle) was an outstanding technological success and established rules for the development of future computers both for aerospace and other applications. It also opened the way to new applications of integrated circuits (it has been the first big user on an industrial scale) and being at the time the smallest computer in the world, helped push computers into other sectors of technology. AGC, CMC, LGC? The computer and the navigation system developed for the Apollo program were used both on board the Command (and Servce) Module, or CSM, and on board the Lunar Module, or LM. To distinguish them, the one in the Command Module was said CMC (Command Module Computer), that in the Lunar Module was said LGC (Lunar-module Guidance Computer). The software was different for the two computers, and each "edition" had an independent development path, with shared components. It was in common the management of the various programs, as well as the management of the inertial platform and that of the interface toward the crew (called DSKY, for Display and Keyboard). The CMC software was different for the presence of the Saturn V rocket monitoring and control program (never used in its controlling functions as the Saturn computer always performed flawlessly). It also contained functions to steer during the trans-lunar trajectory maneuver, during orbital flight and of course for the re-entry into the Earth's atmosphere. The LGC software was different for the descent and landing programs, and for the ascent program and the subsequent lunar orbit insertion. Another part in common was devoted to the rendezvous maneuvers, as both vehicles could perform as the active part in the complex process of rejoining each other in orbit. Obviously, the software of both vehicles was different in the management of propulsion devices, and the navigation sensor which, while in the CSM were a sextant and a telescope, in the LM they were a single optical device with a special behavior (due to the reduced size of the LM and the need to keep weight within precise limits). Display and keyboard The most visible component of the whole AGC was the panel used by the crew to interact with the computer. In an era in which "speak" to a computer using a video terminal was still a futuristic advancement, the unit called Display and Keyboard, or DSKY, represented a drastic achievement in the relationship between man and machine. For the first time the famous seven-digits numerical displays were introduced and for the first time it was possible to tell the computer, while it was working, what to do, being also possible to interact with it to change parameters, or to monitor numerical values of different kind (which by itself was a real novelty in an aerospace world still dominated by gauges). System limitations imposed, of course, a number of restrictions and ingenuous solutions even if using the language developed for the purpose of user interaction with the computer required patience and training to ensure efficiency and confidence. The DSKY's language was based in two-digits numeric codes, used to represent Verbs (that is actions) and Nouns (that is the object of an action). In the figure it is shown part of a checklist (from the Apollo A17 Flight Crew G&N Dictionary) that summarized all the Verb/Noun codes for the Lunar Module computer. Verb and Noun values had a prominent display on the DSKY as well as the active program (or major mode, related usually to a flight phase), which was also assigned a two-digits numerical value and displayed on the top right corner of the unit. The Verb and Noun keys were used to set the Verb/Noun combination of interest, and telling the computer to accept it by pressing the Enter key (e.g,, a typical sequence could have been: Verb 16 Noun 65 Enter). The Clear key was used to correct user typing errors, while the Reset key was used to cancel error warnings from the computer. The Pro key (where Pro stands for Proceed) was sued to confirm critical operations (like the execution of a maneuver or the starting and stopping of the computer). The Key Release key was used to release the control of the DSKY to other routines that required its use, that is when the corresponding indicator light was flashing to grab the attention of the operator. The latter behavior was another novelty as many programs could run inside the AGC at the same time. Why "guidance"? The real task assigned to MIT's Instrumentation Laboratory (now Draper Laboratories) was the design and implementation of a guidance system for Apollo, that is a system able to keep a notion of the state of the vehicle in every instant and provide guidance directions for the execution of specific flight phases (for instance, landing on the Moon). The whole AGC is therefore a wider system, not limited only to the computer, that provided sensors, actuators, interfaces for the crew and the communications system (to exchange data with the control center). For the navigation sensors, designers chose an inertial navigation system based on a "three-axis inertial platform" able to maintain an given orientation (used as reference) independently by the motions of the vehicle. On the "platform", three accelerometers were mounted to measure accelerations and thus update the vehicle state during propulsive phases. In the free-fall phases, when the vehicle moved along orbits dictated by Earth/Moon gravities, the state was updated mathematically by applying Newton's and Kepler's equation. Verification of the state, to correct it in case of possible deviations from the intended path, were done by means of a sextant mounted with the "platform". By observing stars and the angle between stars and the horizon (or a landmark) crew were able to navigate in the Earth/Moon system like sailors did for centuries before the advent of of modern radio navigation systems. The mathematics behind the whole guidance and navigation system was worked out by prof. Richard Battin, at the time one of the most prominent experts in the sector, who also went to initially lead the software development task for AGC. In the course of the project, the AGC was added other main functions, like that to act as go-between for the manual control of the vehicle. By all means that was the first "fly-by-wire" guidance system in history, and the first "digital autopilot". This and other additional utilities for secondary tasks demonstrated the versatility of computerized system with respect to more classical approaches. AGC Software As in all computerized systems, the physical part of the computer (circuits, logic gates and memory) is useless without a program to manage its operation. Software became soon one of the principal aspects of each Apollo mission and hundreds of programmers (at a time when the word "software" was still unknown but to a few insiders) worked to produce programs of very high quality, and with lots of ingenuity, having to overcome not only the intricacies of the space guidance and navigation equations, but also the limitations imposed by the inflexible hardware architecture. Even in this case, methods and technologies were developed for this specific purposes to soon become part of the required knowledge in the aerospace world. Software development was not even worthy a mention in the original contract, as it was considered expected, at the time (but often it is still the same today), that programming the computer was a side task in the whole project. In truth, at the peak of the Apollo program, in 1969, software workers were approximately as much as the hardware workers. But being the "software culture" not widely diffused, everybody looked with suspicion at this intangible but fundamental, "thing". Two were the programming languages available: one was the proper central processing unit language, or assembly language, and the other was a higher-level language, that was to be interpreted at execution time within the computer by a program called, exactly, Interpreter. This second language enabled to save memory speeding up the programming of complex mathematical operations, at the expense of a slower execution speed. During the course of software development, in addition to performance limitations, programmers had to fight against memory limitations. Uncountable tricks were used by programmers to save memory words, often to the detriment of code readability. In addition, to worsen the situation, it was required to adopt particular techniques to protect critical functions from a sudden computer restart. By keeping tab on the state of programs under execution it was possible to ensure that in case of serious problems, the computer were able to restart and resume programs from the place they were at the moment of the trouble, without causing issued with critical functions, like the propulsive phases. The experience acquired with the Apollo program, that continued after the lunar missions with the Skylab flights and the Apollo-Soyuz mission, was not lost. The performance of the software and the reliability of the hardware made the AGC the logical choice for the first computer controlled flight, or "fly-by-wire", of an airplane (a modified F-8 fighter). The experience of programmers was subsequently transferred to the Space Shuttle program, where an evolution of the AGC software, and the methods used to implement it, contributed to the making of a space vehicle entirely controlled by the on-board computer. Myths and legends Is it true that the Apollo computer had less power than a digital pocket calculator? No it is not true. The performances of the Apollo Guidance Computer were comparable to that of the central processing unit of computers like the Apple II, the Commodore 64 and the ZX Spectrum, the personal computer that opened the way for the digital revolution at the end ot the 1970's. The big difference was that the AGC was highly reliable and had many more input/output ports to be able to control, and receive data from, sensors and devices within the Command Module and the Lunar Module. In addition, it was able to execute programs in real-time, that is which operated continuously performing more tasks at the same time and guaranteeing that high priority tasks were performed first when required. Its architecture is not much different from that of modern microcontrollers used today for the most disparate tasks, with the difference that the latter are at least 10 to 20 times faster than the AGC and enormously smallest (being integrated in a single chip). For this reason programming and using the AGC is not very far from the world in which control systems and embedded systems designers and hobbyists work today. Is it true that during the landing of Apollo 11, Neil Armstrong had to take control from the computer because of the alarms it generated? Stupid media remarks often cite "software errors" that generated alarms during Apollo 11 descent to the Moon (the most critical phase of the mission). In truth, the robustness of AGC software, in terms of good design and reliability, saved the mission when other issues, external to the computer, were causing an overload in it its operation (and was thus correctly alerting the crew of the condition). The LGC kept working correctly, but toward the end of the descent, Armstrong took manual control (as everybody else who followed him on this task) because the landing point targeted by the computer was too dangerous for a safe landing. Was AGC the first space computer? No, AGC wasn't the first space computer or the first computer who flew on a crewed vehicle. Gemini spacecrafts had also an onboard computer, made with transistors, and able to assist the crew (but not totally control the vehicle) in the propulsive phases, during 'rendezvous' and during re-entry. But that computer was not an essential part of the mission, while AGC was born to provide assistance in completing the mission without ground support.In addition to that, other digital computers were developed for the Titan rockets, and of course for the Saturn, but they were limited to a single task and were much bulkier and heavier. The Lunar Module also had a second onboard computer, the AGC, Abort Guidance System, much simpler and able to ensure the climb back to orbit and a rendezvous with the CSM in case of troubles with the primary guidance and navigation system. Did the AGC have an operating system? In a certain sense it had, in the sense today commonly considered for computers for control applications. The AGC operating system was made of an executive and a manager of queue for waiting programs. The executive provided, in turn and basing on priority, a chance to each program to execute its own instructions and also managed possible interruptions generated by random external events. This way more than a single program could be active at the same time, even if the allocation of "machine time" was not based by a rigid scheduling (preferred technique at the time). but was was asynchronous and linked to the correct design of each function. Also, it was available a separate process to verify the well-being of the computer, one for the management of the DSKY and one for managing external devices. And another part of this "operating" was devoted to the higher level language used for mathematical operations that did not require a real-time response. The Virtual AGC project provides a which simulates the AGC, the DSKY, and some other portions of the guidance system.  In other words, if the virtual machine—which we call —is given the same software which was originally run by the real AGCs, and is fed the same input signals encountered by the real AGCs during Apollo missions, then it will responds in the same way as the real AGCs did.  The Virtual AGC software is free of charge, so that it can be studied or modified. The video clip above (courtesy of Dean Koska and YouTube) illustrates Virtual AGC running on a Palm Centro—which isn't supported directly from our .  But it's one of the great advantages of that you can take it an adapt it to your wishes without being at the mercy of the creators of the software. After you install the Virtual AGC software you can run the simulation by clicking its desktop icon : While there are all manner of runtime options you can select in this screen, such as which Apollo mission you which to fly, it's simplest as a newcomer to just click the Novice button and then to click Run.  In the screenshot below, the DSKY is displaying the time since startup (00000 hours, 00003 minutes, 00270 hundreds of a second), though that's not something it will do automatically when you start it up. In this simplest of configurations, you are simulating theAGC of the Lunar Module of Apollo 13, and you are providedwith a simulated AGC and simulated DSKY.  You can entercommands at the DSKY keypad, and the AGC will report theresults of its actions on the DSKY's display. The first thing you need to know when operating the AGC is that it contains a collection of programs, each identified by a two-digit number.  The most-basic program is program 00, usually known as "pooh" (as in "Winnie the Pooh") or P00.  In fact, in the screenshot above, the AGC is running P00, as you can see by looking at the "PROG" area in the upper right-hand corner of the DSKY.  The available programs differ somewhat from mission to mission, and the picture below (taken from reference cards supplied to the astronauts) is a sampling of some of the programs available for the Apollo 15 Command Module. From the astronaut's perspective, all operations of the computer are controlled by a quirky system of "verbs" and "nouns". and aren't words, as you might suppose, but are instead two-digit numbers.  What makes them "verbs" and "nouns" is that they are somewhat analogous to verbs and nouns in natural languages like English, in which the verb describes the action to be taken and the noun describes the data to which the action is applied.  So most commands are applied by hitting a key sequence like the following on the DSKY: VERB NOUN ENTRVERB ENTR For example, a command like V06N36E (VERB 0 6 NOUN 3 6 ENTR) would mean to perform action 06 on data 36.  The exact meanings of the verbs-numbers and noun-numbers differ from mission to mission—i.e., from AGC software version to version.  The pictures below show a sampling of the available verbs and nouns from the astronaut's reference cards for the Apollo 15 Command Module, from which you can see that verb 06 means to perform a decimal display and noun 36 means "time of AGC clock", so V06N36E should display the current time from the AGC's clock, if that happens to be a legal action for the current program. There are also some exceptions, such as getting the AGC to run the P00 program.  You do that with a key sequence like V37E00E, in which NOUN is replaced by ENTR for some reason. How did this strange verb/noun system get put into place? One of the original AGC developers, Ram\F3n Alonso, provides an explanation.  Apparently, nobody had yet arrived at any kind of software requirements for the AGC's user interface when the desire arose within the Instrumentation Laboratory to set up a demo guidance-computer unit with which to impress visitors to the lab.  Of course, this demo would have to something, if it was going to be at all impressive, and to do something it would need some software. In short order, some of the coders threw together a demo program, inventing and using the verb/noun user-interface concept, but without any idea that the verb/noun concept would somehow survive into the flight software.  As time passed, and more and more people became familiar with the demo, nobody got around to inventing an improvement for the user interface, so the coders simply built it into the flight software without any specific requirements to do so.  Many objections to the verb/noun system were received by the developers, such as "it's not scientific", "it's not dignified", or even "astronauts won't understand it".  Even though the coders of the demo hadn't seriously intended the verb/noun interface to be used in any permanent way, it became a kind of devilish game to counter these objections with arguments as to why the interface was really a good one.  In the end, the coders won and the verbs and nouns stayed.  Nevertheless, the following Shakespearean quote (from ) is embedded within the AGC source code:  "It will be proved to thy face that thou hast men about thee that usually talk of a noun and a verb, and such abominable words as no Christian ear can endure to hear." you can perform to see the DSKY actually do something. You may notice that on the Virtual AGC main display screen there are a number of different Apollo missions listed, but that not all of them are selectable.  That's because Virtual AGC strictly simulates the original AGC computer hardware, and that requires having the original software that ran on the AGC.  But we don't have copies of all of the different software versions used for the different missions!  Collecting this software is one of the principal goals of the Virtual AGC project, so if you happen to know a collector, a museum, or any of the original AGC developers who have printouts of AGC software, !  (We don't want the printouts themselves, just copies of the information in them, and we're willing to travel there to do the copying work ourselves.) Here's a brief rundown on the known AGC software: Mission AGC Software by  Name and Version Command Module Lunar Module Apollo 1 Sunspot build TBD - Apollo 4 Solarium build 54 - Apollo 5 - Sunburst build TBD Apollo 6 Solarium build 55 - Apollo 7 Sundisk build TBD - Apollo 8 Colossus build 237 - Apollo 9 Colossus build 249 Sundance build 306 Apollo 10 Comanche build 45? Luminary build 69 Apollo 11 Comanche build 55 Luminary build 99 Apollo 12 Comanche build TBD Luminary build 116 Apollo 13 Comanche build 67 Luminary build 131 Apollo 14 Comanche build 72 Luminary build 178 Apollo 15 Artemis build 72 Luminary build 210 Apollo 16 Artemis build 72 Luminary build 210 Apollo 17 Artemis build 72 Luminary build 210 Skylab 2 Skylab 3 Skylab 4 Skylark build 48? - Apollo-Soyuz Skylark build TBD - "What would you call a collection of programs, routines, subroutines, procedures, major modes, jobs and tasks, that all have to exert the utmost courtesy in sharing the resources of a single spacecraft computer? The physical form of fixed memory gave us our answer: a rope. With no tapes, disks, or any form of backing store, every scrap of coded logic had to be built into one rope memory, 24K words in a Block I AGC, for each Apollo mission. Ropes had names of up to 8 characters, and early ones were named in honor of Apollo\92s role in ancient mythology, as driver of the phaeton carrying the Sun across the sky: ECLIPSE (begun on the day of a solar eclipse that occurred 6 years to the day before the first Moon landing), SUNRISE, CORONA, SOLARIUM, and SUNSPOT. In Block II development, there was a maverick called RETREAD, then AURORA, SUNDIAL, SUNDISK, SUNBURST, and SUNDANCE. Finally, Fred Martin, our transplanted New Yorker who steadfastly refused to pledge allegiance to the Red Sox, persuaded us that it was more important to give them names that linked them to their vehicles: COLOSSUS for the Command Module; LUMINARY for the LM." — Hugh Blair-Smith (A description of the software-naming convention thatdidn't quite make it into the final version of AGCdeveloper Hugh Blair-Smith's book, .)Left Brains for the Right Stuff These programs, though having different names, are not really independent programs.  There is a complex family heritage, and a lot of overlap between them.  So if you understand one of the programs, you will understand 99% of what's in any of the others. The AGC software is mostly written in .  "Assembly language" is a very simple language in which each instruction—usually, each line of software source code—represents a single primitive CPU operation.  Since each type of CPU provides different primitive operations, the assembly language for each type of CPU differs from that of every other type of CPU.  The AGC assembly language, being based on the custom AGC CPU, thus differs in detail from every other type of assembly language but shares certain customary elements.  The specific assembly language of the AGC was referred to as "basic"— to be confused with the famous later computer language BASIC which is many people's first introduction to simple computer programming.  Here's a short excerpt from the Apollo 11 Lunar Module software to give the flavor of it: L POSMAX Q # CREATING OVERFLOW AND Q-1 IN Q BBANK SUPERBNK L Q # SAVE DELTA WAITEXIT # WAITEXIT. WAITEXIT # IF TWIDDLING, THE TS SKIPS TO HERE 0 # PICK UP 2CADR OF TASK. -1 WAITADR # BBCON WILL REMAIN IN L WAITBB # ENTRY FROM FIXDELAY AND VARDELAY. BBANK WAIT2 The things at the left-hand side (, , ) are "program labels", and are used to provide names for different blocks of source code.  The next column to the right contains the actual CPU instructions (, , , , etc.) while the column to the right of that (, , , etc.) specifies the data on which the CPU instructions are supposed to operate.  For example, at the very bottom you'll find ""; when the program reaches that point, the instruction causes it to jump to program label . Finally, anything preceded by the '#' symbol is just a comment added for explanatory purposes by the programmer, but not affecting program execution in any way. You can learn all about AGC basic assembly language, if you so desire, in the .  The Virtual AGC software provides an assembler, which is a program that converts assembly-language source code to executable code which can actually be run on the AGC CPU.  So with Virtual AGC, you can actually create and run your own AGC programs. As it happens, the functionality which the AGC needed to provide was so complex that it wouldn't have fit within the amount of physical memory provided by the AGC if it were written entirely in basic assembly language.  One thing that was done to get around this was to provide also a higher-level computer language referred to as "interpretive", in which each interpretive instruction represents a large number of basic assembly-language instructions.  Interpretive thus has the advantage of cutting down on memory usage, but typically has the disadvantage of running much more slowly, and therefore had to be used with care.  Within any given AGC program like Luminary, basic and interpretive code is intermixed. Here's a sample of interpretive code: INTPRET RTB QUITFLAG NOINT LOADTIME TDEC1 INTSTALL CALL NODOFLAG SETIFLGS STATEUP 60000 B-28 EXIT STATEFLG TC PHASCHNG # KILL INTEGRATION UNTIL NEXT P00. Some of the things here are the same as in the basic assembly language, such as the program labels and the program comments.  As with basic assembly language, the next column to the right of the program labels ( , , etc.) contains instructions—though mostly interpretive instructions rather than basic instructions. In this sample, "" and "" are basic assembly-language code.  In fact, "" actually means to begin interpreting the lines that follow as interpretive code, and every block of intepretive code must somehow be preceded by a basic instruction like that.  Near the end, you see the interpretive instruction and this means to stop working with interpretive code and to resume working with basic assembly language. In other respects, though, interpretive is very weird in comparison to basic assembly language.  For example, the column to the right of the instruction column contains mostly names of variables operated on by the instructions ... but can also contain more instructions, because intepretive instructions can be packed two to a word of memory.  Also, many variables don't seem to be preceded by instructions, because a single interpretive instruction may operate on the lines below it as well as on the column to the right of it.  For example, in the two lines of code " / " near the end of the sample, the and instructions are grouped together because they are packed into a single word of memory, but the instruction actually operates on the variable .  Oh, those wacky AGC developers!  How they must have chuckled when they dreamed up that scheme. You can learn all about AGC interpretive language, if you so desire, in the . The AGC was not the only computer in the Lunar Module.  As a backup to the main guidance system, there was a completely separate guidance system called the Abort Guidance System (AGS) developed by Aerospace, now a division of Northrop Grumman.  As the name implies, the AGS was intended to be used only in the case of an aborted landing, and its basic function was simply to get the LM into an orbit from which it would be possible for the CM to rendezvous.  Fortunately, this functionality never had to be used in a real mission, but that does not alter the significance of the system. Naturally, an important part of the AGS was a computer, and this computer was called the Abort Electronics Assembly—or just "AEA" for short.  Since the AGC and AEA were developed independently by unrelated groups, there is no commonality between them in terms either of hardware or of software. There was essentially no interaction between the AGC and AEA, except that the spacecraft's "state vector"—its current position, velocity, and orientation—could be transferred. Here are the basic characteristics of the AEA: Just as the AGC had its DSKY for interfacing to the astronaut, the AEA had its own Data Entry and Display Assembly—or "DEDA" for short. The astronaut interface was very simple.  You could do one of two things: In other words, to usefully interact with the AEA, you had to know the numerical values of the memory addresses that had any significant purposes.  But don't be fooled by the simplicity of the interface into believing that the AEA was a mere calculator.  Though its functionality was much less than that of the AGC, it still had to perform very significant guidance functions on demand, and to have very sophisticated software. Virtual AGC provides the ability to simulate the AEA and DEDA, and you can activate this ability by selecting the "LM Abort Computer (AEA)" option in the Virtual AGC main window.  In the screenshot below, the DEDA is displaying the contents of location 377, which is the number of interations of the units self-test which have been executed (and passed). As with the AGC software, the AEA software evolved over time and we have only a subset of the software versions known to have been created, so !  We have the versions for and probably for . The AEA was programmed in its own assembly language, which was completely different from that of the AGC.  The Virtual AGC software includes an AEA assembler, so you can create and run your own AEA software, if you'd like to do so. You can learn more about the AEA, including the definition of its assembly language, at the . It's natural to think that the AGC might have steered and otherwise controlled the Saturn V rocket that hurled the Command Module and Lunar Module onto its moon-bound trajectory, or the Saturn IB rocket that was used for Earth-orbital missions like Apollo 7, but in fact it isn't true.  The AGC was able to provide limited guidance-system backup for some of the Saturn burns, but the main responsibility for guiding the Saturn rested with yet another onboard computer, the Launch Vehicle Digital Computer (or LVDC for short).  The astronauts were able to monitor various sensor readings from the Saturn's guidance system on their displays, but neither they nor the AGC normally had any control over the rocket. Unlike the AGC and AGS, the LVDC was not installed in the Command Module or Lunar Module, but was instead installed in the Saturn itself, in a non-propulsive stage called the IU (Instrumentation Unit).  The IU perched above the final propulsive stage, the S-IVB, and therefore was the last part of the Saturn to be discarded by the Command Module.  The picture at right illustrates this for a Saturn IB rocket, but the Saturn V is very similar ... only much bigger! There's a lot more information about the LVDC on the , but there's no denying that at present we know far less about the LVDC than about any of the other Apollo flight computers. This web page has provided a brief introduction to the AGC & friends, and to Virtual AGC, but if you'd like to learn more about the AGC without taking the trouble to become a full expert on it, you might want to read the AGC's . If that's not enough for you, you can learn a more by looking at the , and of course that's what we think you do!  In fact, the main Virtual AGC website contains every AGC-related scrap of data, documentation, and software we've been able to find, so it can be somewhat intimidating.  We're sorry about that, but training for an Apollo moon landing isn't easy! This page is available under the Creative Commons No Rights Reserved License Last modified by Ronald Burkey on 2017-01-11.

LeetCode – Unique Paths

LeetCode - Unique Paths Problem statement A robot is located at the top-left corner of a m x n grid (marked 'Start' in the diagram below). The robot can only move either down or right at any point in time. The robot is trying to reach the bottom-right corner of the grid (marked 'Finish' in the diagram below). How many possible unique paths are there? Problem statement taken from: https://leetcode.com/problems/unique-paths Example 1: Input: m = 3, n = 7Output: 28 Example 2: Input: m = 3, n = 2Output: 3 Explanation: From the top-left corner, there are a total of 3 ways to reach the bottom-right corner: 1. Right -> Down -> Down 2. Down -> Down -> Right 3. Down -> Right -> Down Example 3: Input: m = 7, n = 3Output: 28 Example 4: Input: m = 3, n = 3Output: 6 Constraints: - 1 <= m, n <= 100- It's guaranteed that the answer will be less than or equal to 2 * 10^9 Explanation Brute force approach As per the problem statement the robot can move either down or right. We can use recursion to find the count. Let numberOfPaths(m, n) represent the counts of path to reach row number m and column number n in the grid. numberOfPaths(m, n) in C++ can be recursively written as following. int numberOfPaths ( int m , int n ) { if ( m == 1 || n == 1 ) return 1 ; return numberOfPaths ( m - 1 , n ) + numberOfPaths ( m , n - 1 ) ; } The time complexity of the above solution is exponential. There are many overlapping sub-problems and hence we can use dynamic programming approach to avoid re-computing overlapping sub-problems. Dynamic programming approach We can avoid re-computing the overlapping sub-problems by constructing a temporary 2D array count[][] in bottom up manner using the above recursive approach. int numberOfPaths ( int m , int n ) { // create a 2D array to store results of sub-problems int count [ m ] [ n ] ; // count of paths to reach any cell in first column is 1 for ( int i = 0 ; i < m ; i ++ ) count [ i ] [ 0 ] = 1 ; // count of paths to reach any cell in first row is 1 for ( int j = 0 ; j < n ; j ++ ) count [ 0 ] [ j ] = 1 ; for ( int i = 1 ; i < m ; i ++ ) { for ( int j = 1 ; j < n ; j ++ ) count [ i ] [ j ] = count [ i - 1 ] [ j ] + count [ i ] [ j - 1 ] ; } return count [ m - 1 ] [ n - 1 ] ; } The time complexity of the above program is O(mn). The space complexity is O(mn). We can reduce the space more by O(n) where n is column size. int numberOfPaths ( int m , int n ) { int count [ n ] = { 1 } ; count [ 0 ] = 1 ; for ( int i = 0 ; i < m ; i ++ ) { for ( int j = 1 ; j < n ; j ++ ) { count [ j ] += count [ j - 1 ] ; } } return count [ n - 1 ] ; } Combinatorics approach We have to calculate m+n-2 C n-1 here which will be (m+n-2)! / (n-1)! (m-1)! Let's check the algorithm on how to compute the above formula: - set paths = 1 - loop for i = n; i < m + n - 1; i++ - set paths = paths * i - update paths = paths / (i - n + 1) - return paths C++ solution class Solution { public : int uniquePaths ( int m , int n ) { long int paths = 1 ; for ( int i = n ; i < m + n - 1 ; i ++ ) { paths *= i ; paths /= ( i - n + 1 ) ; } return int ( paths ) ; } } ; Golang solution func uniquePaths ( m int , n int ) int { paths := 1 for i := n ; i < m + n - 1 ; i ++ { paths *= i paths /= ( i - n + 1 ) } return paths } Javascript solution var uniquePaths = function ( m , n ) { let paths = 1 ; for ( let i = n ; i < m + n - 1 ; i ++ ) { paths *= i ; paths /= ( i - n + 1 ) ; } return paths ; } ; Dry Run Let's dry-run our algorithm to see how the solution works. Input: m = 3, n = 7 Step 1: set paths = 1 Step 2: loop for i = n; i < m + n - 1 i = 7 7 < 7 + 3 - 1 7 < 9 7 < 9 true paths = paths * i paths = 1 * 7 = 7 paths = paths / (i - n + 1) = 7 / (7 - 7 + 1) = 7 / 1 = 7 i++ i = 8 Step 3: loop for i < m + n - 1 8 < 8 + 3 - 1 8 < 9 8 < 9 true paths = paths * i paths = 7 * 8 = 56 paths = paths / (i - n + 1) = 56 / (8 - 7 + 1) = 56 / 2 = 28 i++ i = 9 Step 4: loop for i < m + n - 1 9 < 8 + 3 - 1 9 < 9 false Step 5: return paths So we return answer as 28. Share this post!

React-Smooth-Scroll-Hook:Easily use ScrollTo in React App

{{ message }} ron0115/react-smooth-scroll-hook You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.

Making the whole web better, one canvas at a time

Making the whole web better, one canvas at a time. One can have an entire career on the web and never write a single canvas.getContext('2d'), so "Why should I care about this new OffscreenCanvas thing?" is a decent question for many. In this post, I'll tell you why I'm certain that it will matter to you, in real ways. How relevant is canvas? As a user, you know from lived experience that <video> on the web is pretty popular. It isn't remotely niche. However, many developers I talk to think that <canvas> is. The sentiment seems to be something like... I can see how it is useful if you want to make a photo editor or something, but... It's not really a thing I've ever added to a site or think I experience much... It's kind of niche, right? What's interesting though, is that in reality, <canvas>'s prevalence in the the HTTPArchive isn't so far behind <video> (63rd/70th most popular elements respectively). It's considerably more widely used than many other standard HTML elements. Amazing, right? I mean, how could that even be?! The short answer is, it's just harder to recognize. A great example of this is maps. As a user, you recognize maps. You know they are common and popular. But what perhaps you don't recognize that it's on a canvas. As a developer, there is a fair chance you have included a <canvas> somewhere without even realizing it. But again, since it is harder to recognize "ah this is a canvas" we don't idenitfy it the way we do video. Think about it: We include videos similarly all the time - not by directly including a <video> but via an abtraction - maybe it is a custom element or an iframe. Still, as a user you still clearly idenitfy it, so in your mind, as a developer you count it. If canvas is niche, it is only so in the sense of who has to worry about those details. So let's talk about why you'll care, even if you don't directly use the API... The trouble with canvas... Unfortunately, <canvas> itself has a fundamental flaw. Let me show you... Canvas (old) This video is made by Andreas Hocevar using a common mapping library, on some fairly powerful hardware. You'll note how janky it gets - what you also can't tell from the video is that user interactions are temporarily interrupted on and off as rendering tries to keep up. The interface feels a little broken and frustrating. For whom the bell tolls For as bad as the video above is, as is the case on all performance related things, it's tempting to kind of shrug it off and think "Well, I don't know.. it's pretty usable, still - and hardware will catch up". For all of the various appeals that have been made over the years to get us to care more about performance ("What about the fact that the majority of people use hardware less powerful than yours?" or "What about the fact that you're losing potential customers and users?" etc), we haven't moved that ball as meaningfully as we'd like. But,W I'd like to add one more to the list of things to think about here... Ask not for whom the performance bell tolls, because increasingly: It tolls for you. While we've been busy talking about phones and computers, something interesting happened: Billions of new devices using embedded web rendering engines appeared. TVs, game consoles, GPS systems, audio systems, infotainment systems in cars, planes and trains, kiosks, point of sale, digital signage, refridgerators, cooking appliances, ereaders, etc.. They're all using web engines. Interstingly, if you own a high-end computer or phone, you're similarly more likely to enounter even more of these, as a user. Embedded systems are generally way less powered than the universal devices we talk about often when they're brand new -- and their replacement rate is way slower. So, while that moderately uncomfortable jank on your new iPhone still seems pretty bearable, it might translate to just a few (or even 1) FPS on your embedded device. Zoiks! In other words, increasingly, that person that all of the other talks ask you to consider and empathize with... is you. Enter: OffscreenCanvas OffscreenCanvas is a solution to this. It's API surface is really small: It has a constructor, and a getContext('2d') method. Unlike the canvas element itself, however, it is neatly decoupled from the DOM. It can be used in a worker - in fact, they are tranferrable - you can pass them between windows and workers via postMessage. The existing DOM <canvas> API itself adds a .transferControlToOffscreen which will (explcitly) give you one back, and is in charge of painting in this rectangle. If you are one of the many people who don't program against canvases yourself, don't worry about the details... Instead, let me show you what that means. The practical upshot of simply decoupling this is pretty clear, even on good hardware, as you can see in this demo... OffscreenCanvas based maps Using OffscreenCanvas, user interactions are not blocked - the rendering is way more fluid and the interface is able to feel smooth and responsive. A Unique Opportunity Canvas is also pretty unique in the history of the web because it began as unusually low level. That has its pros and its cons - but one positive thing is that the fact that most people use it by abstraction presents an intersting opportunity. We can radically improve things for pretty much all real users, through the actions of comparatively group of people who directly write things against the actual canvas APIs. Your own work can realize this, in most cases, without any changes to your code. Potentially without you even knowing. Nice. New super powers, same great taste There's a knock on effect here too that might be hard to notice at first. OffscreenCanvas doesn't create a whole new API to do its work - it's basically the same canvas context. And so are Houdini Custom Paint worklets. In fact, it's pretty hard to not see the relationship between painting on a canvas in a worker, and painting on a canvas in a worklet - right? They are effectively the same idea. There is minimal new platform "stuff" but we gain whole new superpowers and a clearer architecture. To me, this seems great. What's more, while breaking off control and decoupling the main thread is a kind of easy win for performance and an intersting super power on it's own, we actually get more than that: In the case of Houdini we are suddenly able to tap into all of the rest of the CSS infrastructure and use this to brainstorm, explore and test and polyfill interesting new paint ideas before we talk about standardizing them. Amazing! That's really good for both standards and users. Really interestingly though: In the case of OffscreenCanvas, we now suddenly have the ability to parallelize tasks and throw more hardware at highly parallelizable problems. Maps are also an example of that, but they aren't the only one. My colleague Chris Lord recently gave a talk in which he gave a great demo visualizing an interactive and animated Mandlebrot Set (below). If you're unfamilliar with why this is impressive: A fractal is a self repeating geometric pattern, and they can be pretty intense to visualize. Even harder to make explorable in a UI. At 1080p resolution, and 250 iterations, that's about half a billion complex equations per rendered frame. Fortunately, they are also an example of a highly parallelizable problem, so they make for a nice demo of a thing that was just totally impossible with web technology yesterday, suddenly becomming possible with this new superpower. OffscreenCanvas super powers! A video of a talk from a recent WebKit Contributors meeting, showing impressive rendering. It should be time jumped, but on the chance that that fails, you can skip to about the 5 minute mark to see the demo. What other doors will this open, and what will we see come from it? It will be super exciting to see!

Zeroth-Principles Thinking

Bryan Johnson p Follow Published in Future Literacy p 7 min read p Mar 9, 2021 -- 15 Listen Share How does one come up with a truly original idea? One that reshapes society for centuries. Learning about the discovery of the number zero, and its philosophical representation — nothingness — changed my orientation with transformational thinking. My first thought after reading Zero: A Biography of a Dangerous Idea was that I wanted to build a sturdy mental model for exploration of analogously difficult-to-see concepts. I wanted to find capital-Z Zeros. It was intuitively what I had been trying to do for years after selling my company Braintree Venmo but I just didn’t have words or concepts to make concrete what I was consciously experiencing. The question I was trying to answer once I sold the company and could explore anew: What one thing can I do to help humanity thrive most? What is missing from the zeitgeist but is essential? Silicon Valley was full of “first principles” companies and ideas from engineers and designers. But every “first principles”-derived possibility I evaluated felt like it fell short of what was expansively possible and what humanity needed right now for its future existence. Thinking outside the box can lead to productive insights, but that wasn’t enough; I wanted to call the box itself into question. Now, I think about Zeros everyday as I run Kernel and OS Fund, two companies trying to usher in the future of what it means to be human. Each is simultaneously enabled by first-principles thinking and fueled by Zero discovery time-locked to meaningful horizons in all of our futures — 2025, 2050, and 2500. Zero discovery is difficult to express with the words and concepts at our disposal today, but I know the emotional profile of finding a Zero. In the professional analysis of one of AlphaGo’s games against the best human Go player in the world, one observer remarked that the moves were like watching the AI play “Go from an alternate dimension.” That’s the profile. That’s what it feels like to witness Zero discovery. Since hearing that, I keep coming back to this “from an alternate dimension” idea. To the trained eye, AlphaGo seemed to play with new ideas or strategies that were either entirely new or ancient enough to have been abandoned by modern professionals. The AI did not take our human assumptions as a given and played with new conceptual primitives lurking inside the structure of the game that not even experts could see. The moves already existed. A human could have played them, if they knew to. They were always possible. But something about the human brain made us never see them, even after we iterated for thousands of years of Go culture and strategy. Why not? AlphaGo found zeroes in the game of Go, but Go is a finite, complete-information game. My primary interest: How do we find the best moves — those ”from an alternate dimension” — for the future of intelligent existence? No previous generation has ever had the opportunity that we do right now: to look out over our expected lifetime and see the real possibility of evolving as/into entirely novel forms of conscious experience. An existence so transformed that our current selves, and today’s realities, will appear historically primitive and uninteresting. What might these be? Zeros by definition cannot be defined and are hard to imagine. Some that come to my mind: What if we weren’t motivated by social status, others approval or wealth accumulation? What if we felt no tribal proclivities? What if we primarily cared about harmonious play, of a type we don’t even have words or concepts for today, with all things around us? (What [enter your most unintuitive, counterintuitive and unlikely assertions] are you thinking of that I can’t see?) After all, our modern world would appear all those things, and equally foreign, to our ancestors. These ruminations even a few decades ago could have easily been dismissed as practically impossible or as a waste of time for anyone actually wanting to accomplish something. (As we build our Autonomous Selves, our mental energies will be freed up to explore these frontiers.) Hunting for Zeros is relevant for everything currently top of mind, from how we address climate change, how we adapt to technological disruption, how we build AI, how we govern ourselves, and how we evolve and motivate our reasons to keep going, to keep existing. Recently, I’ve condensed these ideas and have started practicing “Zeroth-principles” thinking, which aims to uncover transformative new elements of thought. It is a twin with “first-principles” thinking. Zeroth-principles thinking is about building blocks, or the structure of all things, whereas first-principles thinking is about system laws, or how things interact. First-principles thinking sets goals in known terms and then pursues them, inventing and learning new things as necessary, but it rarely uncovers brand new conceptual primitives of the kind which the world needs. When people say, “From first principles…” to start their argument, what they mean is that they wish to assume as few things as possible from within a given frame. Innovation in space exploration is a good example of the good that can come from first principles thinking alone. Our agenda in space today is valuable, difficult, and we don’t have all the answers for what’s out there in the universe, but we do have a good sense of the questions and the building blocks. Few new building blocks have been discovered for space travel in a long time — we’re still working out the implications and laws of the same basic concepts from the 1970s and, arguably, all the way back from Newton. Sherlock Holmes is a classic first-principled thinker, achieving success as a detective with his own brand of careful logic: “When you’ve eliminated the impossible, whatever remains, however improbable, must be true.” In this, Holmes — trapped in a book, after all — can presume a closed universe of objects where, if X and Y are the only things that exist, not-X is the same as Y. The deductions of Dirk Gently, however, the “holistic detective” created by Douglas Adams as a contrast to Sherlock Holmes, have a different approach: “I don’t like to eliminate the impossible”. This is Zeroth-Principles thinking. Seeing through the blind spots. In a murder investigation, Holmes wouldn’t entertain the possibility, or even the hypothesis, that time travel could play a role. Dirk Gently, though, wouldn’t rule it out. In part, because how does he know what is and isn’t possible yet? The two fictional detection styles illustrate a fundamental distinction in logic. What would it mean to say, “From zero principles…”? The future of human existence is not deducible from first-principles thinking alone. We need Zero Explorers. If we can’t fully model where we are going or what we can aspire to, it’s challenging to create practical plans which have clear goals and which galvanize large-scale cooperation. I think that is a feature, not a bug. Our co-evolutionary future with AI will introduce a record-breaking number of Zero-like building blocks, which will in turn level up our aspirations. When brain interfaces allow for the real-time pairing of one’s mind with AI, we may experience a Cambrian Explosion-like emergence of Zeros. Our subconscious is to our thinking as deep learning is to AI which, when harnessed, will be capable of producing thoughts that seem like they are “from another dimension.” By definition, you can’t “get to Zero” or introduce a novel element into thinking directly, so you have to try and stress first-principles efforts to the extremes until there are no marginal gains left. How can humans, the system architects of intelligence, increase the speed of Zero discoveries? First, some more examples from tech and science: Or, consider the historical example of European geometry before and after the introduction of the zero (the original, literal number/concept). Everything Europeans thought they understood about math had to be rebuilt from the ground up to include the concept of zero once it was introduced. It’s not that Euclid’s Elements from the Greek era suddenly became obsolete or worthless. The power of zero was that it allowed existing mathematics to be rebuilt for a more powerful understanding of geometry that included the concept of zero. Which led to Cartesian geometry — including the idea of the origin of the coordinate system on the plane at (0, 0) — and, suddenly, geometry and algebra could be connected conceptually via the decimal system. Before Europeans learned about zero, Descartes could not have come up with Cartesian geometry. All anyone had ever tried was to do better and better Euclidean geometry. We need more such insights. In a world where we are continuously expanding our spheres of understanding, each Zeroth-principle insight can potentially unlock a set of more expansive spheres. This is bigger than just an exponential effect, of the kind popularized by Ray Kurzweil, the “up to the right” / hockey stick / knee-of-the-curve style stuff. Zeros are game changers. The graph is not just exponential — the units change. The graph reorganizes its axes. Dimensions are added to accommodate ideas “from another dimension.” That’s what Zeros get you. You don’t know what big goals are worth pursuing until you’ve uncovered the new building blocks and understood their properties. Here’s the big question, the Zero Explorer mantra: What concepts are hiding in plain sight but can’t be seen by anyone? How do we, with our limited human cognition, start thinking “from an alternate dimension”? Instead of aspiring to be number one, be a zero.

Board Game Price Breakdown

(anovac) Patron Badge for 2011 through 2022

How one musician took on the world’s biggest TV network over copyright–and won

You’ve heard Kerry Muzzey’s work (Bandcamp, Spotify), even if you haven’t heard of him. The 50-year-old classical music composer from Joliet, Illinois, who now lives in Los Angeles, produces haunting orchestral scores that soundtrack some of the most poignant moments in film and television. When Finn Hudson kissed Rachel Berry for the first time on TV’s Glee, it was Muzzey’s stripped-back piano playing in the background. Some of his works have been choreographed and performed on So You Think You Can Dance?, too. The use of Muzzey’s music across pop culture has no doubt brought the veteran composer some success and acclaim. And around 2012, he decided to see for himself, searching for his name on YouTube. Muzzey recalls the site’s algorithm surfaced 20 or 40 videos. The majority were fan compilations that teenagers obsessed with Glee had painstakingly put together to memorialize their two favorite characters' love story—and they were all soundtracked to the full version of Muzzey’s music. “It was really kind of cool and validating, especially for someone who was a complete independent, to have a kid finding a piece of instrumental music, which is the most uncool kind of music for a kid to find, and to make a tribute montage using it,” he explains. “It was stuff nobody would have a problem with.” No one had a problem, that is, until a few years later when Muzzey gained access to YouTube’s ContentID system, the platform’s automated copyright tracker. That handful of Glee fan videos just scratched the surface. Unbeknownst to the composer, waiting beyond a YouTube search for his name was a seeming subindustry that consistently used Muzzey’s music without his knowledge. ContentID surfaced roughly 20,000 videos for Muzzey in the first month—200 or 400 more got flagged every single day. Many of these works weren’t from amateur obsessives tinkering around with video editing software. Some were annoying but smallscale, like professional wedding videographers who had decided Muzzey’s music was the perfect backing track for a bride’s big day, but they didn’t want to pay for the rights. “Things like that were mildly angering,” he says. But other video makers clearly should’ve had no issue shelling for a license. There were projects from ad agencies producing spots to hawk bottled water, hotel chains, and car commercials. Yet the thing that annoyed Muzzey the most were the pages upon pages of full TV episodes uploaded to YouTube that returned a positive hit through the ContentID system. They came from all over Asia: Vietnam, South Korea, Malaysia, the Philippines, and Thailand. China was one of the biggest offenders. “It was overwhelming,” Muzzey tells Ars about the hours spent poring over the results. The list of offending work was growing so fast that by the time Muzzey had looked at a couple of videos to ensure YouTube’s copyright infringement system wasn’t misfiring, another page of 25 search results had been appended to the end. A never-ending stream of videos from across the world was evidently co-opting Muzzey’s work. And the list of infringers eventually included one of the world’s biggest TV networks. Copyright infringement on the Internet is as old as the Internet itself. Lax rules and the free spirit ethos that embodied the early days of the Web made it seem almost acceptable to share illicitly obtained copies of materials with fellow users, and every digitally connected generation since has its own memories of getting files and footage illegally. For early Internet users, IRC channels were one main method; later browsers will still swell with pride when Kazaa or Limewire are mentioned. Today, those wanting free access to pirated material are spoilt for choice: they can gain access to free books through Lib-Gen; movies and TV shows from sites like Putlocker, illegal streaming services, or through torrents; and Internet users in 2021 can even educate themselves for free.  against the tyranny of paid-for access to academic papers means it’s possible to get almost any research paper you could want without handing over a single penny. As the coronavirus whips up a perfect storm of people stuck at home because movie theaters and concert venues are closed, coupled with less disposable income because of the mass ranks of unemployment as a result of the pandemic, piracy is on the rise. There was a within the first month of lockdown, from February 2020 to March 2020, according to Muso, a company that tracks online piracy. So even today, the Internet is built on a remix and republish ethos—a mantra that has laid waste to copyright and the ability for copyright holders to lay claim to their work. The reason we all log on to YouTube now and not any number of its oddly named competitors like Revver or Vimeo is in part because of the site’s willingness to look the other way about copyrighted material uploaded onto the platform. Notably, a Saturday Night Live skit called Lazy Sunday was one of YouTube’s earliest viral successes. Of course, YouTube has attempted to clean up its act since. It’s now as much a site for professional production companies to post full TV shows, documentaries, and music videos as it is an online repository for hobbyists with video cameras. ContentID has been praised by those who own the rights to works and lambasted by those who think the commercialism of the site has robbed it of its creative streak. Though Alphabet, parent company of YouTube and Google, doesn’t break down copyright removal requests on YouTube specifically, across Google it received requests to take down content from more than 220,000 individual copyright owners in 2019. (YouTube declined to comment for this story.) While this copyright system is strong in theory, in practice there are loopholes. “Power asymmetries mean YouTube is not really incentivized to care about an appropriate resolution when problems crop up for individual musicians,” explains Kevin Erickson, director of the Future of Music Coalition, a Washington, DC-based lobby group campaigning for musicians’ rights. That asymmetry means there are still people—companies, even—who get around the system and who think copyright doesn’t work for them. Surprisingly, global entertainment behemoth China Central Television stands firmly within this copyright-antagonistic group. China Central Television (CCTV) is a network of dozens of TV channels that broadcasts video content to more than a billion people inside China. As you’d expect from a monolithic media outlet in the centrally controlled Communist state, it’s an arm of the government. The public service broadcaster is nominally like the UK’s BBC or America’s PBS, but in reality it’s closely connected to the Chinese state. And CCTV has repeatedly shown it’s more than happy to breach copyright. Among the huge numbers of results that Muzzey found when he ran his regular ContentID searches were scores of hits from TV shows broadcast on CCTV. In all, he found 17 TV programs and movies emanating from CCTV that used some of his music. Some of those videos weren’t posted by official CCTV channels; instead they came from third-party uploaders wanting to share episodes of their favorite shows more freely. Muzzey was infuriated: he had previously pursued a large TV production company based in China through a lawyer years ago because some of his music ended up being used on their programs—unlicensed. The case was pursued by the Chinese arm of the law firm and ended up in an out-of-court settlement that netted Muzzey precisely zero dollars. “I didn’t realize how expensive it was to use a law firm like that,” he says. But he had thought that would be the end of his need to pursue legal action in China. It turned out it wasn’t. Through the third-party uploads, Muzzey was able to match back the use of his music to certain TV shows, then to specific episodes. That led him to the official uploads that CCTV would occasionally post themselves on YouTube. “I found quite a few usages,” he says. “My music was in anything from silly reality TV shows to dating shows, but then also scripted dramas and movies.” He started using YouTube’s built-in copyright claim system to strike each one of the videos under the Digital Millennium Copyright Act (DMCA); most of the videos had at least a million views already. (A copyright strike under the DMCA is treated differently to a ContentID claim on YouTube: ContentID claims can result in the offending material being taken down, or any profits made from it redirected to the rightful copyright owner. DMCA strikes result in warnings against the channel owners.) At the same time, Muzzey also laid strikes against a handful of other Chinese TV networks. That angered the TV network. Muzzey hadn’t considered the chain of events that was set off every time he pressed a button to claim copyright over music used in a YouTube video illegally. It triggers an alert that gets passed through YouTube to the uploader, warning them that someone believes they’ve acted against the law. YouTube cautions uploaders alleged to have breached copyright rules through a three-strike system. Three strikes, and your channel is terminated. For rank-and-file YouTubers, the strikes can pile up quickly. But, according to Muzzey, YouTube “do[es] acknowledge that if it’s a highly monetized YouTube partner, or one of their premium uploaders, that uploader gets something of a grace period to resolve the problem.” YouTube provides seven days to members of its YouTube Partner Program to rectify the situation, during which time the channel remains live. In this instance, the scale of the strikes Muzzey was filing against CCTV blew through any grace period to resolve the problem. It was clear there had been a pattern of persistent, unrepentant copyright infringement of Muzzey’s works—alongside potentially hundreds of infringement incidents against other creators. The sheer volume of copyright strikes ultimately compelled CCTV to engage with Muzzey.

Monads to Machine Code (2016)

Just-in-time or JIT compilation is compilation done by dynamically generating executable code. It’s a common technique used in many language runtimes to generate optimized code for hot code paths as well ahead of time compilation for various tasks. So let’s build a small LLVM-like intermediate language and JIT execution engine in Haskell. This will only function with modern Linux and x86-64 architecture, although in principle this will work on any platform if you can make the appropriate FFI calls to mmap and mprotect syscalls on your respective platform. Source code is available here. The x86-x64 instruction set is the 64-bit revision of x86 instruction set which was first developed for the Intel 8086 CPU family. The base types which hardware operates over are integers and floating point types. Let us just consider the integral types for now, these come in four major varieties: On the Intel architecture numbers are represented little endian meaning lower significant bytes are stored in lower memory addresses. The whole memory representation for a value is partitioned into high bits and low bits. For example the hexadecimal number 0xc0ffee as a DWORD is stored in memory as: In Haskell unboxed integral machine types are provided by the Data.Word module. Pointers are simply literal addresses to main memory, where the underlying access and storage are managed by the Linux kernel. To model this abstractly in Haskell we’ll create a datatype containing the possible values we can operate over. To convert from Haskell types into byte sequences we’ll use the binary library to convert integral types into little endian arrays of bytes. For example given a hexadecimal literal this will expand it out into an array of it’s bit components. The x64 architecture contains sixteen general purpose 64-bit registers capable of storing a quadword. They major ones are labeled rax, rbx, rcx, rdx, rbp, rsi, rdi, and rsp. Each of the registers is given a specific index (r), which will be used in the binary representation of specific instructions that operate over these registers. Each of these registers can be addressed as a smaller register containing a subset of the lower bits. The 32-bit register of rax is eax. These are shown in the table below. These smaller registers are given specific names with modified prefixes. In Haskell we model this a sum type for each of the 64-bit registers. Consider just the 64-bit registers for now. The index for each register is defined by a simple pattern match case expression. Monads are an algebraic structure with two functions (bind) and (return) and three laws. The compiler will desugar do-blocks of the form into a canonical form involving generic bind and return statements. Monad is implemented as a typeclass indexed by a parameter m, that when instantiated with a typeclass instances replaces the bind and return functions with a specific implementation of the two functions (like State or Reader). The State monad is an instance of Monad with several functions for composing stateful logic. For example a little state machine that holds a single Int value would be written like the following. More common would be to have the state variable s be a record with multiple fields that can be modified. For managing our JIT memory we’ll create a struct with the several fields. This will be composed into our X86 monad which will hold the JIT memory as we assemble individual machine instructions and the pointer and memory offsets for the sequence of assembled instructions. To start creating the JIT we first need to create a block of memory with executable permissions. Inside of C runtime we can get the flags needed to be passed to the various mmap syscall to create the necessary memory block. Then we simply allocate a given block of memory off the Haskell heap via mmap with the executable flags. Haskell pointers can be passed to our JIT’d code by simply casting them into their respective addresses on the Haskell heap. For example if we want allocate a null-terminated character array and pass a pointer to it’s memory to our JIT’d code we can write down a asciz to synthesize this memory from a Haskell string and grab the heap pointer. For C functions we simply use the dynamic linker to grab the function pointer the given symbol in the memory space of the process. The Haskell runtime links against glibc’s stdio.h and math.h so these symbols will all be floating around in memory. When we’ve compiled our byte vector of machine code we’ll copy into executable memory. Then we’ll use the FFI to synthesize a function pointer to the memory and invoke it. Before we start manually populating our executable code with instructions, let’s look at assembly form of what we’ll write and create a small little DSL in Haskell make this process closer to the problem domain. Assembly is the intermediate human readable representation of machine code. Both clang and gcc are capable of dumping out this representation before compilation. For example for the following C program takes two integers passed in registers, multiplies them respectively and adds the result. Internally the C compiler is condensing the Destructuring the expressions into a linear sequence instructions storing the intermediate results in scratch registers and writing the end computed result to return register. It then selects appropriate machine instruction for each of the abstract operations. We can output the assembly to a file add.s. We’ll use the Intel Syntax which puts the destination operand before other operands. The alternate AT&T syntax reverses this convention. The generated code will resemble the following. Notice that there are two kinds of statements: directives and instructions. Directive are prefixed with a period while instructions are an operation together with a list operands. Statements of instructions are grouped into labeled blocks are suffixed with a colon for example f: is the label containing the logic for the function f. The assembler will then turn this sequence of instructions into either an executable or an object file containing the generated machine code. To disassemble the output we can use objdump to view the hex sequence of machine instructions and the offsets within the file. The compiled program in memory is then a contiguous array of bytes, which is evaluated by moving the instruction pointer at the start address. Instructions consist of two parts, an opcode and a set of operands which specify labels, registers, or addresses to memory which the CPU will execute over for the give instruction. We’ll consider a subset of interesting operations which operate over integers and manipulate the call stack. To add to the JIT memory we’ll simply modify the state by appending an instruction to the _mach field and adjusting the memory offset pointer by the length of instructions added. Registers are identified as lowercase (i.e. rbp, rsp). In our expression builder we’ll simply write down several functions which construct a register value type from an underlying Reg value. Immediate operands are direct references to constants (literals or memory addresses) or labels. For example: For immediate values we simply push the array of bytes for the number directly on the byte sequence. The full instruction set for x86 is vast and including AVX, SSE and other specialized intrinsics there is an extraordinary amount of complexity and quirky specifications. Each of these abstract instructions can have multiple opcodes for each type of operands it may take. For example the mov instruction for register to register movement has opcode 0x89 while moving immediate data into a register has opcode 0xC7. The reference for the most common operations the x86asm.net site has a very useful quick reference. For the full set of possible instructions on your modern Intel processor refer to the 1500 page Intel Software Developer’s Manual. To lookup the numerical opcodes for a given instructions, we use a specific naming conventions for the operands. For opcodes that operate over a set of possible operands, these are demarcated with a slash, in the form r8/r16/r32. For our limited set of instructions there are two types of opcodes. On top of this opcodes may have an additional prefixes which modify the sizes of arguments involved. These were added to allow 32-bit compatibility in the transition between 32-bit and 64-bit systems and preserve the underlying opcodes of the 32-bit system. For instance the following mov instructions all operate over registers and perform the same action but over different sizes. But translate into different opcodes depending on size. Ok, let’s look at the full structure of an instruction. It consists of several parts. The sizes of these parts depend on the size and type of the opcode as well as prefix modifiers. The header fixes the first four bits to be constant 0b0100 while the next four bits indicate the pretense of W/R/X/B extensions. The W bit modifies the operation width. The R, X and B fields extend the register encodings. The Mod-Reg-R/M byte determines the instruction’s operands and the addressing modes. These are several variants of addressing modes. In the case of mod = 00, 01 or 10 For example given the following instruction that uses register direct mode and specifies the register operand in r/m. Scale Index Base Scale is the factor by which index is multipled before being added to base to specify the address of the operand. Scale can have value of 1, 2, 4, or 8. If scale is not specified, the default value is 1. Both the index and base refer to register in the usual index scheme. Moving forward we’ll create several functions mapping to X86 monadic operators which assemble instructions in the state monad. Let’s do some simple arithmetic logic first. Each of these functions takes in some set of operands given by the algebraic datatype Val and pattern matches on the values to figure out which x86 opcode to use and how to render the values to bytes. ret The simplest cases is simply the return function which takes no operands and is a 1-bit opcode. add <r64> <imm32> Add for immediate values extends the operand with a REX.W flag to handle 64-bit immediate data. add <r64> <r64> Register to register add uses the REX.W flag in the same manor but passes the source register in the ModRM.reg field using register direct mode. We do bitwise or over the mode 0xc0 and then shift 3 bits to specify the register in register index in the reg bits. mov <r64>, <r64> Same logic applies for the mov instruction for both the register-to-register and immediate data cases. mov <r64>, <imm32> inc <r64>, dec <r64> The inc and dec functions are slightly different in that they share the same opcode but modify the ModRM bit to specify the operation. Putting everything together we’ll JIT our function and call it from Haskell. And running it we get the result. Now let’s write some logic that uses control flow and jumps between labeled blocks of instructions. Consider the factorial function that takes the value to compute in the rcx register and computes the result my repeatedly multiply the rax until reaching one. To do this we create a block .factor and use the loop instruction. Let’s look at the machine code for this assembly. Notice that the loop instruction takes a relative address in memory fc (i.e. go back 4 instructions) as it’s operand. So let’s create a label function which simply reaches into the monad and grabs the current pointer location in the JIT memory that we’re at. When given an memory address, the loop instruction then simply emits the instruction simply emits the 0xE2 opcode and calculates the delta of the source and destination and the emits it’s value as the immediate data for the instruction. Now we’ll create the symbolic representation of this factorial assembly our in Haskell DSL and parameterize it by a Haskell integer to compute. Putting everything together we’ll JIT our function and call it from Haskell. And running it we get the result. Final task is to be able to call out of the JIT into either Haskell runtime or a given C function pointer. To do this we need to look at the calling convention for moving in out of other logic and setting up the registers so we can hand them off to another subroutine and restore then when we jump back. In the 64 bit System V ABI calling convention, the first 5 arguments get passed in registers in order rdi, rsi, rdx rcx, r8, and r9. Subsequent arguments get passed on the stack. For our call function we simply compute the delta of our given position and the address of the function we want to jump into. Before and after we call the function we are responsible for handling it’s arguments and the push and popping the stack frame on entry and exit. On entry we call the function prologue and on exit we call the epilogue, in between lies the function logic. So for example let’s call out to the libc printf function passing a string pointer to it from inside our JIT. To do this we use dlsym to grab the symbol reference and then pass it as an address to the call instruction after pushing the string pointer on the argument stack. Putting everything together we invoke it by grabbing the printf address and passing a pointer to Haskell string using our asciz function. Running it we get our friendly greeting by reaching outside the JIT. So that’s basic JIT compilation in a nutshell.

GCP Tau VM Instances: Better Performance, Price-Performance Than Graviton2 M6g

Announced earlier this year for Google Cloud was a new family of virtual machines called Tau VMs. The initial T2D instances are powered by AMD EPYC 7003 "Milan" processors to deliver leading performance and are also positioned to deliver great value in going up against the likes of Amazon's Graviton2 instances. Tau VM instances are now available as a preview and Google has provided us with gratis access to the new instance types for benchmarking. Over a variety of tests we are seeing T2D providing over ~47% higher performance when compared to Graviton2 with M6g instances. While Tau T2D VM instances cost about 10% more than the M6g for on-demand pricing, the price-performance TCO savings Tau T2D VM deliver are still significant. The Tau VM instances were born out of Google and AMD collaborating to come up with a new instance type that could offer leading performance, price, and streamlined integration. The T2D instances are designed for web services, container workloads, large Java applications, and other scale-out workloads. Google's performance and value claims for Tau VMs from their June announcement. Back in the June announcement of Tau VMs, Google cited up to 56% higher absolute performance and up to 42% higher price-performance over the general purpose VMs over other public cloud providers. The Tau VM family currently consists of instances up to 60 vCPUs and 4GB of system memory per vCPU. Google's original announcement notes that T2D VMs should be available as a preview in Q3'2021, which has begun happening this month. Tau VM instances are powered exclusively by AMD EPYC 7003 series processors. Google Cloud was kind enough to provide us gratis access to the T2D instances early for carrying out benchmarks. For this initial T2D benchmarking is a look at the T2D performance against Amazon's M6g Graviton2-based instances, which appear to be the principal competition for this new class of virtual machines. The T2D VM instances tested were using an AMD EPYC 7B13 processor. For the purposes of this initial Tau VM TD2 performance comparison, benchmarks were carried out against equal sized Amazon M6g instances. The t2d-standard-8 with 8 vCPUs was tested against the m6g.2xlarge, which is eight cores on Graviton2. Both of these instances offered 32GB of RAM. For a larger size and additional comparison was the t2d-standard-32 against the m6g.8xlarge, both having 32 vCPUs and 128GB of system memory. Across all these virtual machines tested, Debian 10 with the Linux 4.19 kernel and other default software toolchain components were used for testing in their out-of-the-box configuration as deployed by each of these public cloud providers. When it comes to the on-demand pricing, the t2d-standard-8 in the US central region used is at $0.337968 per hour while the on-demand EC2 m6g.2xlarge pricing was $0.308 per hour. For the t2d-standard-32 pricing, the on-demand rate was $1.351872 per hour while the on-demand m6g.8xlarge was at $1.232. From the t2d-standard-1 up through the t2d-standard-60, the M6g instances did cost ~9% less than the similarly-sized T2D instance, but we'll see with these benchmarks and the performance-per-dollar metrics, which public cloud provider was delivering better value.

Nature journals announce first open-access agreement

Skip to main content p NEWS 20 October 2020 Nature journals announce first open-access agreement The arrangement will allow some researchers in Germany to publish openly — but critics say it comes with a high price. Richard Van Noorden Richard Van Noorden Access options Rent or buy this article Get just this article for as long as you need it $39.95 Prices may be subject to local taxes which are calculated during checkout doi: https://doi.org/10.1038/d41586-020-02959-1 Additional reporting by Holly Else. Open-access Plan S to allow publishing in any journal Nature to join open-access Plan S, publisher says Ambitious open-access Plan S delayed to let research community adapt Funders flesh out details of Europe’s bold open-access plan Subjects Publishing Latest on: Publishing EU council’s ‘no pay’ publishing model draws mixed response News 02 JUN 23 The world’s first pollinating frog and more — May’s best science images News 02 JUN 23 Preprint clubs: why it takes a village to do peer review Nature Index 01 JUN 23 Jobs Postdoctoral Research Fellow at the Dalian Institute of Chemical Physics Professor/Associate Professor/Assistant Professor/Senior Lecturer/Lecturer Professor/Associate Professor/Assistant Professor/Senior Lecturer/Lecturer Faculty Positions at SUSTech Department of Biomedical Engineering Postdoctoral Fellows/Research scientists Close banner Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily. Email address I agree my information will be processed in accordance with the Nature and Springer Nature Limited Privacy Policy. Close banner p Sign up for Nature Briefing

Top Machine Learning Algorithms

Muhammad Furqan Ul Haq p Follow Published in Edopedia p 4 min read p Jan 20, 2021 -- Listen Share Photo by Possessed Photography on Unsplash Introduction Ever since the third industrial revolution, there is a massive increase in data. So, the need for organizing and managing data becomes imperative. We can process this data through machine learning algorithms to do some predictions. For example:- Movies Recommendation Weather Forecasting Medical diagnosis Predicting Sports Outcomes Better Inventory Planning And then one question arises “What is Machine Learning”? Basically, a machine works according to the instructions given to it. But, if a machine starts learning from its past data and operates according to it then this is called Machine Learning. Machine learning is a subset of Artificial intelligence and it uses computer algorithms and statistical models that permit your device to learn and adapt changes from past events without being explicitly programmed. Meaning that we train our algorithms with the help of previous data. Machine learning focuses on the development of computer programs that can change when exposed to new data. The accuracy of the result depends on the amount of data and better model. You know what! Machine Learning is the future. It seems like a game-changer for every field of life. Just imagine a world where intelligent machines are working for us. That’s why Machine Learning Engineers are in high demand at the moment. It means that you can easily get a high-paying job at big tech companies. So, in my opinion, everyone should learn it to boost their career. By the way, I and 739,000+ students learned it from this “Machine Learning Course” which is taught by some of the best data scientists out there. You could also try it if you don’t want to spend years on YouTube tutorials!!! Anyways, today we’ll have a look at the available machine learning algorithms. There are basically three types of machine learning algorithms. Supervised learning Unsupervised learning Reinforcement learning Supervised learning Supervised learning is the basic subset of machine learning. Newbie machine learners began their practice with supervised learning algorithms. In supervised learning, you can classify and process your data using a labeled data set. It means that you first train your model using labeled data before using it. Labeled data contains both the input and correct output. You must train your model with labeled data until it gives the right result. Supervised learning is categorized into two main categories. Classification Regression Classification It is a supervised learning technique, and in classification problems, your output is classes or categories such as predicting the color of the car. Regression In a regression problem, your output is a real number such as age or weight. List of common Supervised Learning Algorithms:- Nearest Neighbors Naive Bayes Decision Trees Support Vector Machine Neural Networks Linear discriminant analysis Unsupervised learning In unsupervised learning, data does not have any label and we don’t know what these data points do. Here, the machine uses unlabeled data and allows the algorithms to discover the unknown patterns and information in data that was previously undetected. The task of the machine is to categorize or classify data. Unsupervised learning is computationally complex and is less accurate as compared to supervised learning. Types of unsupervised learning Clustering Association Clustering Clustering is a type of unsupervised learning where you find the patterns of data on which you are working. Association Association is a type of unsupervised learning where we have to find the dependencies of one data item on another data item. List of common Unsupervised Learning Algorithms:- K-means for Clustering Problems Principal Components Analysis Singular value decomposition Independent Components Analysis Reinforcement learning Reinforcement learning is a subfield of machine learning. It uses a reward-based learning algorithm. It is derived from supervised learning and the difference between both is that in supervised learning we already know that what will be the output but in Reinforcement learning, we don’t have any knowledge about the output. Reinforcement learning enables the agent to interact with the environment by performing actions and its learning is based on the trial and error method. It does not use any predefined data. Common approaches for implementing Reinforcement Learning Algorithms:- Value-Based Policy-based Model-Based I hope that you now have some know-how about Machine Learning and its different algorithms. In the near future, I’ll be sharing a step by step guide on how to implement these algorithms in real-life projects. So, don’t forget to follow me! Also, check out my web development blog to read more interesting topics like this. https://www.edopedia.com/blog/ Disclosure: I only recommend products I would use myself and all opinions expressed here are my own. This post contains affiliate links. If you use these links to buy something then I may earn a small commission.

Chinese actor Zhao Wei has been deleted from video platforms there

Skip to main content ABC News Homepage Loading News Home ABC News Homepage Chinese actor Zhao Wei has been deleted from video platforms there – here's what we know Share

China Blocks Signal

Signal is an encrypted messaging app that claims to be more private than WhatsApp and Facebook : GETTY IMAGES p Encrypted messaging app Signal appears to have been blocked in mainland China, the latest foreign social media service to stop working in a country where the government tightly controls the flow of information. Signal was growing in popularity within China, including among human rights lawyers and dissidents, as a secure way to communicate.  But the app's increasing popularity with the Chinese public is believed to have worried the authorities, prompting government censors to block the platform. China also recently blocked Clubhouse, an invitation-only audio chat platform, after Chinese joined global discussions on a number of topics deemed sensitive by the authorities, such as human rights abuses of Uighur Muslims and pro-democracy protests in Hong Kong. Chinese government censors – dubbed the ‘Great Firewall’ – tightly control news and information in mainland China. Foreign news websites including the New York Times, messaging apps like WhatsApp, and social media sites such as Twitter and Facebook are already all blocked in China. Chinese state surveillance was able to identify and punish Wuhan whistleblower Dr Li Wenliang for alerting colleagues to a coronavirus in 2019 : GETTY IMAGES p Google and all its related services, such as Drive and Gmail, are also blocked in China. Blocking foreign websites, apps and programs forces users back to Chinese platforms – state news outlets and social media platforms like WeChat, all of which are monitored by Chinese authorities. This kind of surveillance is how Chinese police were able to identify and punish late Wuhan whistleblower Dr Li Wenliang for alerting colleagues to a coronavirus circulating in late 2019. Chinese censors also routinely delete online content deemed ‘sensitive,’ such as complaints about the government’s handling of the coronavirus pandemic. Tech-savvy users in China can use virtual private networks (VPNs) to circumvent government censors to use these platforms and access censored content. But Chinese authorities are increasingly cracking down on the use of VPNs, which are illegal in the country. People have been arrested and slapped with multi-year sentences for using VPNs. Only Chinese state-approved VPNs are allowed, which likely still exposes users to surveillance. Chinese government officials and state media outlets often scale the ‘Great Firewall’ likely with the help of VPNs to post on Twitter and Facebook. China has scored the lowest for six consecutive years according to an annual ranking of internet freedom by the US-based NGO Freedom House, falling behind Iran, Syria and Cuba. Signal did not immediately respond to a request for comment.

Show HN: Tilemaker – DIY vector tiles from OpenStreetMap data

Create the vector tiles on your home machine or using cloud CPU. Upload the file to your site or app. And that's it. No database to maintain; no contract to pay; no restriction on commercial use. tilemaker is a single executable that takes OpenStreetMap data and makes it into vector tiles. It's supremely customisable, but if you just want off-the-shelf tiles in a standard style, tilemaker comes bundled with the files to do that too. The tiles are in the industry-standard Mapbox Vector Tiles format, but you don't need a Mapbox contract. Use the open-source MapLibre GL library to render your tiles in-browser, in iOS apps or on Android.

Who Is Moral Enough to Teach Morals to Machines?

Teaching machines right from wrong might prove to be an impossible task The Unlikely Techie p Follow Published in Digital Diplomacy p 3 min read p Feb 26, 2021 -- 7 Share Dal Spooner never trusted robots. He resented them for their cold logic based solely on Asimov’s principles — a rather mathematical approach to moral principles. The box office hit I, Robot highlighted the grey areas of what it means to be human, or, in reverse…

Employees in Iceland Start Working Four Days a Week

p Employees in Iceland Start Working Four Days a Week Researchers announced an “overwhelming success” of four-day week trials in Iceland Francis Scialabba p Sherry Qin Along with the Northern Lights and the possibility of glimpsing the elusive Huldufólk, there might be one more reason to move to Iceland: in June, 86% of Iceland’s working population were currently, or on the path to, working four-day workweeks with no reduction in pay. The stat comes from a recent study that tracked 2,500 employees whose workweek was shortened to 35–36 hours over five years. Researchers found that a four-day week with the same pay improved workers’ well-being and productivity. Big picture: The explosion of remote work during the pandemic has led to a rethinking of work-life balance, and more countries and even some companies in the US are experimenting with a shorter workweek. Buffer, a social media software company, has let its 89 employees work four days a week since May 2020. The crowdfunding platform Kickstarter will test a four-day week for its 95 employees starting in 2022. Both Spain and Japan are piloting a 32-hour work week over three years in response to changing perceptions of work during Covid. Bottom line: Experts said the “overwhelming success” of the trial in Iceland could signal that a four-day workweek could gain traction far from the Arctic Circle. Become smarter in just 5 minutes Get the daily email that makes reading the news actually enjoyable. Stay informed and entertained, for free.

National+Identification+Number:+Here+is+how+to+check+using+a+USSD+code

Insights National Identification Number: Here is how to check using a USSD code · January 13, 2020 · 2 min read strong: You can now link up to 7 phone numbers with your NIN using the NIMC mobile app. Here's how. The National Identity Management Commission (NIMC), created in 2010, was mandated by the NIMC Act No. 23 of 2007 to register people the Act covers and give each person a unique National Identification Number (NIN). According to the NIMC, the essence of each individual having an NIN is part of a measure to create a national identity database and to put a stop to both double identity and identity fraud. Upon registration, individuals will be provided with an NIN and eventually a national ID card. Advertisement In 2018, the NIMC revealed plans to complete issuing electronic identity cards (e-ID) to 70 million Nigerians by the end of December 2019. However, with many people complaining about not having the e-ID after the registration, it appears that the commission has not met its target. It's been 6 years since I registered for national ID card but NIMC still claim the card ain't ready yet 😅. A whole 6 years ain't enough time to print card? — E-zrael Ani (@EzraelAni) January 9, 2020 It's more than 4 years since I registered with NIMC for my National ID and it's not ready yet. A four years old child has started walking and even working. — Adewale Adetona (@iSlimfit) January 9, 2020 The day registered for Tax Identification Number (TIN) was the day I got it They just took my photograph and fingerprints and in less than 20 minutes the card was ready Why can't NIMC adopt this same method? I mean it's been 5 years now and my National ID card isn't ready How? — CllayBaba (@Cllaytus) January 9, 2020 In addition, the commission is also expected to provide citizens, as well as others legally residing within the country, with a General Multi-Purpose Card (GMPC). The GMPC is embedded with a payment chip to enable users to withdraw cash from an automated teller machine (ATM). Consequently, the commission partnered with Chams consortiums, OneSecureCard consortium, SecureID, and Iris Technologies to provide the data capture services and the GMPC. How to check your NIN using USSD code If you have enrolled for the national ID card scheme, you can get your NIN by dialling *346# from the phone number associated with your voter's card, drivers licence, or BVN registration. Ensure you have at least ₦20 airtime on your line and then dial the code. Then you either write or take a screenshot when the NIN is displayed on the screen as it will not be displayed more than once. Advertisement At the moment, the USSD code works on MTN, Airtel, Glo and 9mobile. In other news, the Joint Admission Matriculation Board (JAMB) has suspended the use of the NIN to register for UTME and direct entry exams until 2021. The registrar of the board, Professor Is-haq Oloyede, said the decision was a result of the technical issues that affected the registration system at the NIMC. NIN IS SUSPENDED FOR 2020 REGISTRATION. CANDIDATES ARE TO ONLY SEND THEIR NAMES TO 55019 TO CREATE PROFILE. NO NEED FOR NIN FOR 2020 REGISTRATION. — JAMB (@JAMBHQ) January 11, 2020 Intending applicants are to send their names to 55019 to create an applicant's profile on any Nigerian network (MTN, GLO, 9mobile, and Airtel). Subscribe To Techpoint Digest This is A daily 5-minute roundup of happenings in African and global tech, sent directly to your email inbox, between 5 a.m. and 7 a.m (WAT) every week day! Digest Subscription Full Name Email Subscribe Give it a try, you can unsubscribe anytime.  Privacy Policy.

Styling child component from a parent in Angular

Angular provides a modular design that encourages the developer to create separate components with its own logic and styles. This approach has many advantages, but it can cause some problems to solve. In this post, I'd like to show how to solve a problem with styling inheritance in Angular. Let's create few components and apply styles to them. p < class parent__container = "" > p < > p p > p p > .parent__container { display : flex; justify-content : center; align-items : center; background-color : peru; width : 300px; height : 300px; } p < class child__container = "" > p p > .child__container { background-color : green; width : 150px; height : 150px; } This is how it looks. A Very simple markup and the result. Now, imagine the situation where we want to style the child component basing on the action in the parent. A new css class is added to the parent component, and based on this class we want to change the styling of the container that's inside it. .parent__container { display : flex; justify-content : center; align-items : center; background-color : peru; width : 300px; height : 300px; } .parent__container.alert .child__container { background-color : darkred; } The inner div should change the background-color property to darkred now. But it does not. If we inspect the elements with the developer tools we notice that styles to the child__container class are not applied. This is when the encapsulation of the property comes in. By default, all Angular components styles are encapsulated. This means that they apply only to the component itself. If we try to style the css classes that are outside the component, they won't be applied. The simplest solution for this problem is to set the encapsulation property to ViewEncapsulation.None in the component. import { Component, OnInit, ViewEncapsulation } from '@angular/core' ; @Component ( { selector : 'app-parent' , templateUrl : './parent.component.html' , styleUrls : [ './parent.component.scss' ] , encapsulation : ViewEncapsulation.None } ) export class ParentComponent implements OnInit { constructor ( ) { } ngOnInit ( ) : void { } } Now the styles are no more encapsulated and the result of our manipulation is exactly what we wanted: However, this solution has a serious downside. The styles from the parent component crossed component boundaries and are global now. If there are more elements with the same classes, the styles will be applied to these elements. This may cause unexpected behavior and we should use this solution carefully. Luckily, there's a better solution to this problem. Angular provides the ::ng-deep pseudo-class. Using it will disable the encapsulation for that particular rule. If we use any selector with this pseudo-class, it will become a global style. But, compared to the previous solution, only the selector and its descendants will be applied in the global scope. Here's how to use it in our example: ::ng-deep .parent__container { display : flex; justify-content : center; align-items : center; background-color : peru; width : 300px; height : 300px; } .parent__container.alert .child__container { background-color : darkred; } That's it. The ::ng-deep selector will target every element inside the parent__container element. Using it along with the BEM css class naming convention in your project should be enough to prevent the styles "leaking" from the outside of the component. The last solution in our case is to put the styles to styles.css file in the src directory of the Angular project. It is possible to spread css styles outside of the Angular component. However, it may cause some unexpected styling issues so try to reduce the usage of this approach. If there's a need to apply styles for the descendant elements, use the ::ng-deep pseudo-class.

A step-by-step analysis of a new version of Darkside Ransomware

Darkside ransomware is the malware family responsible for the Colonial Pipeline attack on May 7 2021 as described at https://www.zdnet.com/article/darkside-the-ransomware-group-responsible-for-colonial-pipeline-cyberattack-explained/. The binary contains an encrypted configuration that will be decrypted using a custom algorithm, which reveals a 22-byte buffer that describes different actions performed by the malware. These actions include: checking the system language and avoiding to encrypt Russian language machines, deleting Shadow copies, wiping Recycle Bin, ignore specific files, directories and file extensions, killing specific processes, deleting specific services, etc. The ransomware can perform privilege escalation using the CMSTPLUA COM interface and achieves persistence by installing itself as a service. The files are encrypted using the custom Salsa20 implementation, with the Salsa20 matrix being encrypted by the public RSA key hard-coded in the binary. Darkside uses multithreading with I/O completion ports to communicate between the main thread and the worker threads responsible for file encryptions. It’s important to mention that the process generates a random Salsa20 matrix using the RDRAND and RDSEED instructions, as opposed to earlier versions that use the RtlRandomEx function. Analyst: @GeeksCyber The malware comes with an encrypted configuration that is decrypted using a custom algorithm: Figure 1 The custom decryption algorithm consists of 4 subtraction operations by 0x10101010 each time and then some addition operations, as shown below: Figure 2 For each DLL to be loaded, there is a hash function that is applied to the DLL name, and the 4-byte result is compared to hardcoded values: Figure 3 For example, the following value corresponds to kernel32.dll: Figure 4 The following DLLs are expected to be loaded: ntdll, kernel32, advapi32, user32, gdi32, ole32, oleaut32, shell32, shlwapi, wininet, netapi32, wtsapi32, activeds, userenv, mpr, rstrtmgr. The process retrieves the address of multiple export functions based on similar hash values computed using the same algorithm: Figure 5 The decrypted configuration is presented below and is composed of the RSA-1024 exponent (0x010001 = 65537), 0x80-byte RSA-1024 modulus, victim UID, 22 configurations bytes (will be detailed further on) and the aPLib-compressed configuration: Figure 6 The binary uses an aPLib-decompression algorithm to decrypt different strings. The following list represents the directories to avoid in the encryption process: Figure 7 The following files will be ignored by the ransomware: Figure 8 If the file’s extension belongs to the following list, then the file will not be encrypted by the process: Figure 9 The binary intends to delete folders that contain the word “backup” in their name: Figure 10 A feature not used by the malware would use the following strings decompressed as the other ones (our guess is that the actor would try to kill the SQL-related processes in order to encrypt databases): Figure 11 The following processes will not be terminated by the file: Figure 12 If a process name contains any of the following strings, it will be killed by the binary: Figure 13 There is also a list of services to be stopped and deleted, as shown in the figure below: Figure 14 The list of C2 servers is also obtained using the same algorithm: Figure 15 The process reveals a message that will be utilized to set a custom wallpaper that contains important instructions for the victim: Figure 16 The content of the ransom note is also written in the process memory, as shown in figure 17: Figure 17 The following table describes the actions that the malware takes depending on the configuration decrypted above: The malware uses the NtQueryInstallUILanguage and NtQueryDefaultUILanguage APIs to determine the language of the system and compares the result with 0x419 (Russian language identifier). If there is a match between these two values, then the malware exits: Figure 18 There is a call to the RegCreateKeyExW function, which is supposed to create (or open if it already exists) the “Software\Microsoft\Cryptography” registry key, as follows: Figure 19 The malware extracts the “MachineGuid” value from the above registry key, as presented in the next figure: Figure 20 Figure 21 A custom hashing algorithm that generates 8 lowercase hexadecimal characters is implemented by the process (the “MachineGuid” value is the input, and the algorithm applies 8 times): Figure 22 Figure 23 The value computed above (let’s call it RansomPseudoValue) will be used in the following constructions: The binary uses the SHTestTokenMembership API to verify if the user belongs to the Administrators groups (0x220 = 544 in decimal): Figure 24 We’ll split the analysis into 3 different parts depending on the user’s privileges: low level privileges, administrative privileges, and SYSTEM privileges. Low Level privileges The malware attempts a UAC bypass that uses the CMSTPLUA COM interface as described at https://gist.github.com/api0cradle/d4aaef39db0d845627d819b2b6b30512. It utilizes ZwOpenProcessToken to open the access token associated with the process (0x8 = TOKEN_QUERY – required to query an access token): Figure 25 The NtQueryInformationToken function is used to get the group accounts associated with the token (0x2 = TokenGroups) and it checks if the administrators group can be found in the TOKEN_GROUPS structure: Figure 26 There is a call to the CoInitialize routine in order to initialize the COM library on the current thread, as highlighted in figure 27: Figure 27 As presented so far, the binary uses a lot of lower level APIs (from ntdll). It allocates a new memory area using the ZwAllocateVirtualMemory API (0x3000 = MEM_COMMIT | MEM_RESERVE and 0x4 = PAGE_READWRITE): Figure 28 We have encountered a call to an undocumented API function called LdrEnumerateLoadedModules: Figure 29 The file executes CoGetObject with the object name as Elevation:Administrator!new:{3E5FC7F9-9A51-4367-9063-A120244FBEC7}, as highlighted below: Figure 30 Basically, it will relaunch the malware with SYSTEM privileges: Figure 31 Figure 32 As in the first case, the binary uses ZwOpenProcessToken to open the access token associated with the process (0x8 = TOKEN_QUERY – required to query an access token): Figure 33 The NtQueryInformationToken API is utilized to retrieve the token’s user account (0x1 = TokenUser): Figure 34 The malicious process uses LookupAccountSidW to obtain the name of the account associated with the SID provided as the input, as shown in figure 35: Figure 35 There are 3 different comparison operations that compare the domain name (the name of the computer in our case) with “NT AUTHORITY”, “AUTORITE NT” and “NT-AUTORITAT” (basically, it tries to determine if the user account is SYSTEM or not): Figure 36 The OpenSCManagerW routine is utilized to establish a connection to the service control manager: Figure 37 The process tries to open a service called <RansomPseudoValue> (which doesn’t exist at this time): Figure 38 Because the service doesn’t exist, it will be created by the malware for persistence purposes, as shown in the following pictures: Figure 39 Figure 40 The newly created service is started, and the binary launches itself as a service: Figure 41 The malicious binary can run with no arguments, one, two, or three arguments (these cases will be described later on). As we can see below, it uses CommandLineToArgvW to obtain pointers to the command line arguments (argv[0] is the executable name) + the number of arguments: Figure 42 The WTSQueryUserToken API is utilized to obtain the primary access token of the logged-on user specified by session 1: Figure 43 OpenWindowStationW is used to open the “Winsta0” windows station (the interactive window station), 0x40000 – WRITE_DAC – modify the DACL in the security descriptor for the object: Figure 44 The DACL (discretionary access control list) of the “Winsta0” windows station is modified by calling the NtSetSecurityObject routine with the 0x4 = DACL_SECURITY_INFORMATION parameter: Figure 45 There is a call to OpenDesktopW that is utilized to open the “Default” desktop object with the argument 0x40081 = WRITE_DAC | DESKTOP_WRITEOBJECTS | DESKTOP_READOBJECTS, as follows: Figure 46 The DACL of the “Default” desktop object is modified by calling the NtSetSecurityObject function with the 0x4 = DACL_SECURITY_INFORMATION parameter: Figure 47 The malware creates a mutex called “Global\4787658f1cc4202b8a15e05dd0323fde” (this value has been computed before this operation and represents a custom “hash” value of the malware), which makes sure that there is only one instance of the ransomware running at a time (if the mutex already exists, then the malware quits): Figure 48 Figure 49 The ransomware forces the system not to enter sleep mode and not to turn off the display while the process is running, one of the parameters being 0x80000001 = ES_CONTINUOUS | ES_SYSTEM_REQUIRED: Figure 50 The file changes the privilege to SE_PRIVILEGE_ENABLED in order to enable the token’s privileges (note the TOKEN_PRIVILEGES structure) by a function call to ZwAdjustPrivilegesToken: Figure 51 The CreateThread API is used to create a new thread, as described in the next figure: Figure 52 A list of valid drives on the system is extracted using the GetLogicalDriveStringsW routine: Figure 53 The ransomware is looking for DRIVE_REMOVABLE (0x2) and DRIVE_FIXED (0x3) drives, as highlighted in figure 54: Figure 54 All files and directories from Recycle Bin are deleted by the process. It starts to enumerate via a FindFirstFileExW API call: Figure 55 As presented below, the files are deleted using the DeleteFileW function, and the directories are removed using the RemoveDirectoryW routine: Figure 56 The binary uses COM objects and WMI commands to delete volume shadow copies. It calls the CoCreateInstance function to create a single object of the class IWbemLocator with the CLSID {dc12a687-737f-11cf-884d-00aa004b2e24} (Ref. https://forum.powerbasic.com/forum/user-to-user-discussions/source-code/25222-wmi-wrapper-functions): Figure 57 There is also a new IWbemContext interface with the CLSID {44aca674-e8fc-11d0-a07c-00c04fb68820} (Ref. https://docs.microsoft.com/en-us/openspecs/windows_protocols/ms-wmi/3485541f-6950-4e6d-98cb-1ed4bb143441) created via a CoCreateInstance function call: Figure 58 Using the IWbemLocator object, the process calls the ConnectServer API to connect to the local “ROOT\CIMV2” namespace and retrieves a pointer to a IWbemServices object, as follows: Figure 59 There is a call to CoSetProxyBlanket performed by the ransomware, as described in the next figure (0xA = RPC_C_AUTHN_WINNT – NTLMSSP, 0x3 = RPC_C_AUTHN_LEVEL_CALL and 0x3 = RPC_C_IMP_LEVEL_IMPERSONATE): Figure 60 The process executes the following SQL query “SELECT * FROM Win32_ShadowCopy” to obtain an enumerator of all shadow copies, and then it deletes each of the shadow copy objects via the DeleteInstance method: Figure 61 A list of all services and their status is retrieved by calling the EnumServicesStatusExW function (0x30 = SERVICE_WIN32, 0x3 = SERVICE_STATE_ALL): Figure 62 Each service name is compared to the list that was decrypted at the beginning of the analysis: Figure 63 The malware opens the targeted services by calling the OpenServiceW routine (0x10020 = DELETE | SERVICE_STOP): Figure 64 Every targeted service is stopped and deleted using ControlService and DeleteService, as displayed in figure 65: Figure 65 The NtQuerySystemInformation API returns an array of SYSTEM_PROCESS_INFORMATION structures (one for each process running on the system, 0x5 = SystemProcessInformation): Figure 66 Each process name is compared to the list that was decrypted in the beginning, as displayed below: Figure 67 For every targeted process, the binary opens the process and terminates it and all of its threads: Figure 68 The binary creates an ico file called <RansomPseudoValue>.ico, as displayed below: Figure 69 A new registry key called <RansomPseudoValue> is created using the RegCreateKeyExW function, as shown in figure 70: Figure 70 The DefaultIcon subkey is created, and it specifies the path for the newly created ico file: Figure 71 The malware calls the SHChangeNotify routine to notify the shell to update its icon cache (0x08000000 = SHCNE_ASSOCCHANGED, 0x1000 = SHCNF_FLUSH): Figure 72 A new file called %PROGRAMDATA%\<RansomPseudoValue>.BMP is created using the CreateFileW function: Figure 73 Moving forward, there is a registry key opened by calling the RegCreateKeyExW API, as shown in the next picture: Figure 74 The “WallPaper” value inside the registry key is changed to the location of the newly created BMP file: Figure 75 After all of these activities, the Desktop has been changed to the following image: Figure 76 Thread activity – sub_4095AB The thread starts by decrypting the following information: Figure 77 The version of the Darkside ransomware is also decrypted and represents the latest version analyzed in the wild (2.1.2.3): Figure 78 Another JSON structure is decrypted by the binary and will be used to collect data about the local machine: Figure 79 One more time, the process checks the type of the drives and is looking for DRIVE_REMOVABLE (0x2), DRIVE_FIXED (0x3) and DRIVE_REMOTE (0x4): Figure 80 The GetDiskFreeSpaceExW function is used to retrieve information about the targeted drives, such as the total amount of space and the total amount of free space: Figure 81 NtDuplicateToken is utilized to duplicate an existing token and to obtain a handle to a new access token (0xC = TOKEN_DUPLICATE | TOKEN_IMPERSONATE | TOKEN_QUERY and 0x2 = TokenImpersonation): Figure 82 The thread’s impersonation token is changed via a call to the ZwSetInformationThread routine, as shown in figure 83 (0x5 = ThreadImpersonationToken): Figure 83 The ransomware retrieves the username associated with the current thread, as well as the NetBIOS name of the local machine: Figure 84 Figure 85 The current language of the machine is retrieved from the “LocaleName” value, as presented below: Figure 86 NetGetJoinInformation is used to get the join status information for the local computer: Figure 87 The product name of Windows can be extracted by querying the “ProductName” value and the Windows product ID can be extracted by querying the “ProductId” value, as shown in the following pictures: Figure 88 Figure 89 The malware constructs the following JSON, which contains data to be exfiltrated to the C2 server: Figure 90 The final data looks like in the following JSON form: Figure 91 The data from above is encrypted by a custom encryption algorithm: Figure 92 Figure 93 The result of the encryption operation is base64-encoded, as shown below: Figure 94 Figure 95 The following function is used to generate 2 random 4-byte values that will be utilized in the network communications. It uses instructions such as RDRAND and RDSEED to generate random numbers (if these are supported), but we’ll provide a deeper understanding of it when we discuss file encryption (it’s also used to generate the Salsa20 matrix): Figure 96 The parameters of the network request have the following structure: random_number1=base64(encryptionresult)&random_number2=victim_uid: Figure 97 The InternetOpenW function is called using a user agent decrypted by the malware as a parameter: Figure 98 InternetConnectW is utilized to connect to one of the C2 servers (baroquetees[.]com) on port 443: Figure 99 The process creates an HTTP request handle using the HttpOpenRequestW routine, as shown in figure 100: Figure 100 There is also a call to the InternetSetOptionW API that is used to set the security flags for the handle (0x1f = INTERNET_OPTION_SECURITY_FLAGS): Figure 101 The binary sends the POST request to the C2 server using HttpSendRequestW: Figure 102 Figure 103 The status code returned by the server is retrieved using the HttpQueryInfoW API (0x13 = HTTP_QUERY_STATUS_CODE): Figure 104 Interestingly, the ransomware doesn’t expect a 200 status code but a 500 (Internal Server Error). If the status code isn’t 500, then the process repeats the steps described so far using the second C2 server, rumahsia[.]com: Figure 105 Figure 106 This last idea concludes our analysis of this thread. We continue to analyze the main thread. The binary enumerates the volumes available on the machine and uses the CreateFileW routine to open them: Figure 107 DeviceIoControl is utilized to get information about the type, size, and nature of a disk partition (0x70048 = IOCTL_DISK_GET_PARTITION_INFO_EX): Figure 108 A new thread is created by the file using CreateThread: Figure 109 Thread activity – sub_407558 The only action the thread does is using the GetLogicalDriveStringsW API to retrieve the valid drives on the local machine: Figure 110 If a volume doesn’t have a drive letter associated with it, then the ransomware does that using the SetVolumeMountPointW API, as highlighted in the following picture: Figure 111 The malicious process targets the following types of drives – DRIVE_REMOVABLE (0x2), DRIVE_FIXED (0x3) and DRIVE_REMOTE (0x4): Figure 112 The CreateFileMappingW function is used to create a named file mapping object (name “Local\\job0-<Process Id>” means the object is created in the session namespace): Figure 113 The binary maps a view of the file mapping into the address space of the process by calling the MapViewOfFile routine (0xf001f = FILE_MAP_ALL_ACCESS): Figure 114 A named event object called “Local\\job0-<Process Id>-Event” is created by the binary: Figure 115 The ransomware launches itself with 3 parameters, and the new process will execute the encryption operations: Figure 116 OpenMutexW is utilized to open a named mutex called “Global\\T-job0-<Process Id>” (which doesn’t exist at this time) – 0x100000 = SYNCHRONIZE: Figure 117 The event object created earlier is opened by calling the OpenEventW API (0x1f0003 = EVENT_ALL_ACCESS), as displayed in figure 118: Figure 118 The file creates an I/O completion port that isn’t associated with a file handle, which will be used by the main thread to send data that will be encrypted to worker threads: Figure 119 Two different threads that will take care of the files’ encryption are created using the CreateThread routine: Figure 120 The ransom note README<RansomPseudoValue>.TXT is created and populated in every directory the malware encrypts: Figure 121 The process doesn’t encrypt some certain files, as displayed in the next figure: Figure 122 A list of file extensions decrypted at the beginning of the execution is also excluded from the encryption process: Figure 123 Every targeted file is opened and read using the CreateFileW and ReadFile functions: Figure 124 Figure 125 The file extension is changed to also include <RansomPseudoValue>, as shown below: Figure 126 There is a second function call to CreateIoCompletionPort that associates the existing I/O completion port with the FileHandle parameter: Figure 127 The RSA public exponent and the RSA modulus will be used in the encryption process of the Salsa20 matrix, as we’ll describe later on: Figure 128 The ransomware checks to see if the RDRAND and RDSEED instructions are supported by the processor. If that’s the case, it will use one of them to generate 56 random bytes, and 8 NULL bytes are added to the resulting buffer (Salsa20 matrix -> custom Salsa20 implementation). If none of these are supported, the malware uses the rdtsc instruction to generate deterministic timestamps that will provide a 64-byte Salsa20 matrix: Figure 129 Figure 130 The thread poses a custom implementation of the RSA-1024 algorithm (it doesn’t rely on Windows APIs). Basically, the data d will produce a ciphertext = (d^exponent)%modulus. The raw modulus calculation is performed using addition and subtraction and part of the implementation is presented in the following figures: Figure 131 Figure 132 The Salsa20 matrix is encrypted using the custom RSA implementation, as shown in figure 133: Figure 133 There is a custom “hash” function applied to the above encryption result, which produces a 16-byte output: Figure 134 The file content that will be encrypted is appended to the buffer that will be sent to the worker threads: Figure 135 The Salsa20 matrix is also added to the buffer, and it will be utilized by the worker threads to encrypt the files: Figure 136 Thread activity – sub_405E7C (File encryption) The file content is encrypted using a custom Salsa20 implementation and the ciphertext overwrites the plaintext in the buffer: Figure 137 A snippet of the custom implementation is presented below: Figure 138 The encrypted content is written to the initial file, followed by the encrypted Salsa20 matrix and the hash value, as displayed in the following figures: Figure 139 Figure 140 This last idea concludes our analysis of this thread. We continue to analyze the main thread. If the current directory contains “backup”, then the malware deletes it: Figure 141 The main thread sends the buffer described above (which includes file content to be encrypted etc.) to the worker threads by calling the PostQueuedCompletionStatus routine: Figure 142 We’ve also identified a function that we believe it’s used to propagate the malware to domain controllers (we didn’t have one in our environment). It calls functions such as DsGetDcNameW, DsGetDcOpenW and DsGetDcNextW: Figure 143 Darkside enumerates all network shares using the NetShareEnum API and encrypts each one of them by the main encryption routine described so far: Figure 144 Thread activity – sub_4096A4 The following JSON is decrypted by the thread: Figure 145 The file opens the following registry key by calling RegCreateKeyExW: Figure 146 The Product ID is retrieved again by calling the RegQueryValueExW function: Figure 147 The machine GUID is extracted from the registry and represents a unique identifier for the machine: Figure 148 Figure 149 After the encryption finishes, the malware sends encryption statistics to the C2 server, such as: victim ID, uid, number of encrypted files, size of encrypted files, number of skipped files and elapsed time. The final JSON structure looks like the following: Figure 150 As already described so far regarding the C2 communication, the buffer is encrypted with a custom algorithm and base64-encoded. The request sent to the C2 server is presented in the next picture: Figure 151 If the self deletion feature would be enabled, Darkside would delete itself using ShellExecuteW: Figure 152 Figure 153 As we specified at the beginning of the analysis, the binary can run with different parameters: A particular case is handled by the ransomware differently when it deals with a shortcut file (.lnk file). Basically, the binary wants to extract the full path to the file from this link. It calls the CoCreateInstance API with the CLSID of {000214F9-0000-0000-C000-000000000046} (IShellLinkW interface): Figure 154 Unfortunately, Scylla didn’t help us here and it couldn’t provide us the methods. We’ve found that the next 2 function calls are used to extract the path of the file/directory: Figure 155 Figure 156 The file extracted above is encrypted as usual: Figure 157 MSDN: https://docs.microsoft.com/en-us/windows/win32/api/ Fakenet: https://github.com/fireeye/flare-fakenet-ng Any.run: https://any.run/report/0a0c225f0e5ee941a79f2b7701f1285e4975a2859eb4d025d96d9e366e81abb9/e7a712f5-961a-45b4-a7e5-a0f7196113a5 VirusTotal: https://www.virustotal.com/gui/file/0a0c225f0e5ee941a79f2b7701f1285e4975a2859eb4d025d96d9e366e81abb9/detection Analysis of Darkside Ransomware v1.8.6.2: https://chuongdong.com/reverse%20engineering/2021/05/06/DarksideRansomware/ Fireeye report: https://www.fireeye.com/blog/threat-research/2021/05/shining-a-light-on-darkside-ransomware-operations.html https://gist.github.com/api0cradle/d4aaef39db0d845627d819b2b6b30512 https://forum.powerbasic.com/forum/user-to-user-discussions/source-code/25222-wmi-wrapper-functions https://docs.microsoft.com/en-us/openspecs/windows_protocols/ms-wmi/3485541f-6950-4e6d-98cb-1ed4bb143441 Created files: README<RansomPseudoValue>.TXT, %PROGRAMDATA%\<RansomPseudoValue>.BMP, %PROGRAMDATA%\<RansomPseudoValue>.ico Service Name: <RansomPseudoValue>, Service display name: <RansomPseudoValue> User-Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:79.0) Gecko/20100101 Firefox/80.0 (prone to False Positives)

Show HN: Create a clickable map from your photos

{{ message }} esteininger/Map-My-Experience You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.

What are the best free SEO tools for Nonprofits

SEO tools are incredibly important for your growth and understanding of relevant keywords for your target group. Although intent remains one of the most important factors to retain your audience’s attention. If you run on a tight marketing budget it is hard to create brand awareness through paid advertisement, let alone buy one of those shiny marketing tools. We compiled a list of free SEO tools for you so you don’t have to go look for them. 1. Ubersuggest Awareness probably tops your list of priorities when it comes to your generating awareness for your cause. To create more awareness you are probably posting blogs, articles or text throughout your website. However, with the amount of content on the internet, it is hard to sift through all of them just to find yours. Neil Patel’s ‘Ubersuggest’ has made this easier. This free-to-use SEO tool is able to analyze your website within seconds. With a user-friendly UI, you can navigate easily through the menu’s to carefully read what you need to work on through a site audit. There is no requirement to have an account which makes this tool easy and quick to use. Any website can be looked up with this tool. You might want to do some research on your competitors to see what kind of keywords they are using 😉 Check out this article to learn more about how to do proper keyword research. 2. Answerthepublic This beautifully designed website can help you with the visitor’s keywords search behavior. Based on your desired keyword it is able to generate a web of long-tail keywords. It is mainly looking at the web for questions that are related to the specific keyword. Based on the questions that are asked it is able to find relevant keywords around your main subject. Unfortunately, the system is not able to make suggestions on the desired language. However, if you are able to afford their subscription then this option will open up for you. 3. Google Search Console Besides Google Analytics, the Search Console from Google is a great tool to get rich SEO data. Unlike any other tools on the market, it is able to retrieve the data directly from Google. Google’s other product: Analytics has more limitations when it comes to analyzing your SEO performance. This is where the Search Console complements the analytics tool. With the rank tracker, you are able to check your site rankings on various SERPs. Besides the pages that are indexed, there is also a list of pages that aren’t getting indexed. The Index Coverage report gives you several reasons as to why you are page isn’t getting indexed. This will help you to get your pages back on track. By the way. If you want to learn more about SEO, check out this resource: How to learn SEO [newby-friendly guide]. 4. Moz’s Link explorer There are many different factors to get a high rank within Google search. One of them is based on the value of your link. If you have lots of inbound links from a website with a low domain ranking it will most likely impact your ranking. Moz’s link explorer gives you insight regarding your domain and page authority. You might wonder what the difference is between these two.  Domain authority mainly measures the predictive ranking strength of the entire domains or subdomains whilst Page authority is only looking at the strength of individual pages. Whilst the tool can help you improve SEO of your nonprofit it is, however, being used as a comparative metric than an absolute score. This means there isn’t really a good or bad score, you just want to do better than your competitor within the SERP. It is important to note that you need an account to make use of Moz’s link explorer. Creating an account is free. Another metric commonly used is Trust Flow, which gives you the scores of external links to your website. It basically shows you the trustworthiness of your website. The trust flow can be found on Majestic after you have registered for a free account (which gives you access to make three search queries a day). 5. Yoast WordPress plugin If your website is running on WordPress then Yoast would be an excellent addition. The Yoast WordPress plugin helps you proofreading your website’s content. By using a key phrase the plugin is able to recognize the density of the keyword within your text. Based on this information it determines if the use of that specific keyword is sufficient. Besides the SEO help, it also rates your content on readability. A few tips are given to improve your text if necessary. The tool is easy to use and can improve your SEO and readability on your website significantly. 6. BuzzSumo When you are creating SEO-optimized content the focus isn’t on just ‘gaming the system’. Your content also needs to appeal to your target group. Buzzsumo is a great tool to stay up to date on the latest and greatest of anything related to your industry. Buzzsumo’s tool can help you analyze past content as well as the influencer connected to specific keywords. Know what content is most read on the subject and adjust your content towards the target group. 7. SimilarWeb SimilarWeb is a free web analytics tool you can use after registration. With an easy-to-find UI, you can use this tool to analyze your traffic, referrals, social as an extra besides Google Analytics. However, I have mainly used this as a comparative tool. The aforementioned categories can be used to put your website next to your direct competitor for analytic purposes. Using data from your competitor next to yours give you the right insights to what you been missing and can improve upon. 8. Agency Dashboard p  is an all-in-one reporting platform where you can analyse the performance of your marketing campaigns in a single dashboard. You can track following: a. Organic Search I.e SEO progress of your website. b. Ads (Google, FB and Insta Ads) c. Organic Social (Facebook, Insta and many more) p Check out Agency Dashboard. Author Bio: Michael Vuong creates beautiful content for GlobalOwls and her partners and helps with Social Media strategies. See what he has written on GlobalOwls. More Tools to Help you reach your Goals Faster Top Social Media Marketing Tools for Nonprofits Top Amazon Keyword Research Tools Top Team Collaboration Tools Nonprofits Top SEO Tools Nonprofits Top Article Rewriting Tools Top Email Marketing Tools Nonprofits Top Social Media Monitoring Tools Nonprofits Best Online Marketing Tools Top Tools to Measure Customer Experience Top Tools to Manager Your Subscriber List Top Nonprofit Fundraising Tools Top Tools to Manage Remote Teams Top B2B Lead Generation Tools Top Online Marketing Tools for Nonprofits Top Tools to Write Quality Articles Faster Top Email Automation Tools Top Social Media Analyses Tools Top Chatbot Tools to Improve Customer Service Top Social Media Tools to Improve Customer Relationships Top AI Content Marketing Tools p p p p p p p 8. Agency Dashboard p More Tools to Help you reach your Goals Faster You have a story to tell. We want to help.

VideoRey – Video Marketing Tool

VideoRey - Marketing video maker Videomaker,marketing,advertisement,videoeditor reviews 0 Do you use VideoRey - Marketing video maker? What is VideoRey - Marketing video maker? VideoRey is the marketing video maker for business and brand promotion. Create promo video & social media marketing video in 5 minutes using the ready made templates. Feature: Add text and image on video Image background removal Resize to different resolution Recent launches

I've Recorded My Drug Use in an Excel Sheet for Nine Years

This piece was originally published on VICE France . The first time I did drugs recreationally was nine years ago, on the 21st of November, 2011. I’m certain of this fact, because I’ve taken note of every single time I’ve done drugs since that first experience. As a clubber, drugs have helped me to shed inhibitions and overcome trauma. They’ve allowed me to explore my senses, and what some people might call alternative dimensions. Yes, they can be dangerous, but mostly – and for the majority of people – they’re just fun. I’m neither ashamed nor proud of my drug use, I’m just stating the facts. Advertisement In 2012, I left France to live in Berlin for a year, which is when I started going to clubs and experimenting with drugs. I was young, naïve and a bit overexcited, so I began making lists of the clubs I’d been to and the music I liked, worried that one day I might forget these precious memories. Initially, I wrote them down in a notebook, but as my drug use became more frequent I moved the data to an Excel file. Fetish Photos of Ravers in Both Their Fetish Gear and Day Job Clothes 01.22.20 My method was simple: each time I used a psychoactive substance recreationally (besides weed – which I don’t like – and alcohol), I noted the date and the drug, plus where I was and who I was with. If I was at a party spanning two days, I’d count it as two events. I took it so seriously that I’d text myself in the middle of the night to remember what I’d taken. For that reason, I think my data is pretty reliable – I’d say around 97 percent accurate . The author's Excel table. My project fits within a larger trend of people using data to understand themselves – for instance, apps that measure your sleeping pattern, count your steps or track your calories. After nine years, I’m left with a matrix of 226 rows and 13 columns, which is both fascinating and a little pointless. Using my data, I can confirm I’ve taken ecstasy 81 times in my life, MDMA 74 times, ketamine 45 times and 3-MMC (or metaphedrone) seven times. I also know that I’ve tried LSD, 2C-B, 4-MFP, mephedrone and dexamphetamines (ADHD medication). I can plot charts and see the statistics laid out in front of me. Advertisement For example: Pie chart showing the author's drug use by type of drug. "Champis" is "shrooms", "Autres" is "others". On average, I’ve done drugs every 14-and-a-half days since I started tracking my use. That sounds like a lot, but the figure is inflated by festivals and parties spanning multiple days. I was in Germany 22.6 percent of the time, and in the Netherlands another 20.4 percent. MDMA and ecstasy have been my favourites, but my consumption changed a lot by season – I used psychedelics mostly at summer festivals, while my cocaine consumption peaked in the winter. Since I noted who I was with, I can also delve back into a surprising number of memories that I would have otherwise forgotten. Drugs Tripping On Mushrooms with My Mum, in Photos 10.22.20 Although my data is pretty reliable, there’s one major flaw. If I were to do this again, I wouldn’t just log when I did drugs, but also the quantities. When I started this project, I was more concerned about how often I was using rather than if I felt the effects or not. That means I’ve conflated the times I scraped a few crumbs of ecstasy from the bottom of a baggy with wild New Year’s Eve parties where I indulged a lot more. It might seem unimportant, but it does skew my data. For example, if you just saw the graph plotting my consumption over time, you’d think I used more drugs in 2019 than in 2012, but that’s not the case. Even though I still take a sip from a bottle of MDMA-water from time to time, I rarely do big quantities these days. It’s a bit like my alcohol consumption – I used to drink a lot in high school, but now I'm in my thirties I’ll have a drink nearly every day, but nothing more. The frequency has increased, the quantity hasn’t. In this sense, my data is misleading. Advertisement I thought about this a lot and, short of bringing a micro-scale with me on a night out, I think I’d use a “TripAdvisor system” – I’d rate my drug use on a scale of one to five stars, one for a small dose with little to no effects, five for a 12-hour trip with multiple doses. Frequency of consumption of different drugs by season. Graphic by the author. Life 'I Became a Drug Dealer During Lockdown to Support My Family' 09.10.20 Between the rows of data, I can also see the shape of my own life. There’s a first peak of intensity in 2012, marking my Berlin “Summer of Love”; there’s a period of stability and low consumption between 2013 and 2015, followed by a calm phase between 2016 and 2017, when I was in a serious relationship. And then, in the summer of 2018, there’s the end of my relationship and an issue with workplace harassment, which led to me going to three festivals back-to-back and on a long, drug-fuelled holiday in Berlin. The pandemic and the end of partying as we know it is also visible on the charts, starting in May of 2020. Seeing the data has helped me understand my own drug use. In these particularly stressful times, when many people are struggling and doing drugs to cope, I can only encourage them to develop their own tools and take an honest look at their consumption. Being mindful of when I used helped me space out doses and keep my habits casual. The silence surrounding drug use makes people hide their addictions and contributes to dangerous behaviour. Maybe that’s why I’ve kept recording my consumption for all these years – so I can’t lie about it, to others or myself.

Mutt2 (2020)

So Mutt 2.0 got released, so I figured it was time to take it for another spin. The gmail web client has gotten much slower and less pleasant over the years, and hopefully the rough edges on Mutt + imap have been sanded down. There are a ton of instructions floating around, and I got a little scared about “well what if this wasn’t 2.0 compatible”, so I figured I’d write down the steps again with ‘Mutt 2.0’ at the top for SEO purposes and maybe help the next poor soul. Anway, first, go get a mutt-specific password here- https://security.google.com/settings/security/apppasswords Then go plug it into From here, you should be able to type mutt, and see your inbox. Luckily, Mutt is not breaking-change happy, so all of the 2 year old instructions in the wiki still work- Anyway… it still has the problems it used to. If you open up a large folder, like ‘all mail’, it blocks until it downloads headers for everything. Over 90k emails in my case, at ‘maybe go get a cup of coffee and come back’ speeds. Now, maybe I should have a smaller inbox, and get rid of mailing list archives from 2004, but I’m not sure I want to do a bunch of maybe-scary data deletion & organization just for a mail client that doesn’t do windowing. I also don’t want to sign up for syncing emails to a maildir again, it really hasn’t worked out in the past. So it’s not going to be a mutt year for me, still :/

Ancient civilization: Cracking the Indus script (2016)

Andrew Robinson reflects on the most tantalizing of all the undeciphered scripts — that used in the civilization of the Indus valley in the third millennium bc. The mysterious Indus unicorn on a roughly 4,000-year-old sealstone, found at the Mohenjo-daro site. Credit: Robert Harding/Corbis The Indus civilization flourished for half a millennium from about 2600 bc to 1900 bc. Then it mysteriously declined and vanished from view. It remained invisible for almost 4,000 years until its ruins were discovered by accident in the 1920s by British and Indian archaeologists. Following almost a century of excavation, it is today regarded as a civilization worthy of comparison with those of ancient Egypt and Mesopotamia, as the beginning of Indian civilization and possibly as the origin of Hinduism. More than a thousand Indus settlements covered at least 800,000 square kilometres of what is now Pakistan and northwestern India. It was the most extensive urban culture of its period, with a population of perhaps 1 million and a vigorous maritime export trade to the Gulf and cities such as Ur in Mesopotamia, where objects inscribed with Indus signs have been discovered. Astonishingly, the culture has left no archaeological evidence of armies or warfare. Most Indus settlements were villages; some were towns, and at least five were substantial cities (see 'Where unicorns roamed'). The two largest, Mohenjo-daro — a World Heritage Site listed by the United Nations — located near the Indus river, and Harappa, by one of the tributaries, boasted street planning and house drainage worthy of the twentieth century ad. They hosted the world's first known toilets, along with complex stone weights, elaborately drilled gemstone necklaces and exquisitely carved seal stones featuring one of the world's stubbornly undeciphered scripts. Follow the script The Indus script is made up of partially pictographic signs and human and animal motifs including a puzzling 'unicorn'. These are inscribed on miniature steatite (soapstone) seal stones, terracotta tablets and occasionally on metal. The designs are “little masterpieces of controlled realism, with a monumental strength in one sense out of all proportion to their size and in another entirely related to it”, wrote the best-known excavator of the Indus civilization, Mortimer Wheeler, in 1968 1 . Once seen, the seal stones are never forgotten. I became smitten in the late 1980s when tasked to research the Indus script by a leading documentary producer. He hoped to entice the world's code-crackers with a substantial public prize. In the end, neither competition nor documentary got off the ground. But for me, important seeds were sown. More than 100 attempts at decipherment have been published by professional scholars and others since the 1920s. Now — as a result of increased collaboration between archaeologists, linguists and experts in the digital humanities — it looks possible that the Indus script may yield some of its secrets. Since the discovery of the Rosetta Stone in Egypt in 1799, and the consequent decipherment of the Egyptian hieroglyphs beginning in the 1820s, epigraphers have learnt how to read an encouraging number of once-enigmatic ancient scripts. For example, the Brahmi script from India was 'cracked' in the 1830s; cuneiform scripts (characterized by wedge-shaped impressions in clay) from Mesopotamia in the second half of the nineteenth century; the Linear B script from Greece in the 1950s; and the Mayan glyphs from Central America in the late twentieth century. Several important scripts still have scholars scratching their heads: for example, Linear A, Etruscan from Italy, Rongorongo from Easter Island, the signs on the Phaistos Disc from the Greek island of Crete and, of course, the Indus script. In 1932, Flinders Petrie — the most celebrated Egyptologist of his day — proposed an Indus decipherment on the basis of the supposed similarity of its pictographic principles to those of Egyptian hieroglyphs. In 1983, Indus excavator Walter Fairservis at the American Museum of Natural History in New York City, claimed in Scientific American 2 that he could read the signs in a form of ancient Dravidian: the language family from southern India that includes Tamil. In 1987, Assyriologist James Kinnier Wilson at the University of Cambridge, UK, published an 'Indo-Sumerian' decipherment, based on a comparison of the Indus signs with similar-looking ones in cuneiform accounting tablets from Mesopotamia. Three problems In the 1990s and after, many Indian authors — including some academics — have claimed that the Indus script can be read in a form of early Sanskrit, the ancestral language of most north Indian languages including Hindi. In doing so, they support the controversial views of India's Hindu nationalist politicians that there has been a continuous, Sanskrit-speaking, Indian identity since the third millennium bc. Whatever their differences, all Indus researchers agree that there is no consensus on the meaning of the script. There are three main problems. First, no firm information is available about its underlying language. Was this an ancestor of Sanskrit or Dravidian, or of some other Indian language family, such as Munda, or was it a language that has disappeared? Linear B was deciphered because the tablets turned out to be in an archaic form of Greek; Mayan glyphs because Mayan languages are still spoken. Second, no names of Indus rulers or personages are known from myths or historical records: no equivalents of Rameses or Ptolemy, who were known to hieroglyphic decipherers from records of ancient Egypt available in Greek. Third, there is, as yet, no Indus bilingual inscription comparable to the Rosetta Stone (written in Egyptian and Greek). It is conceivable that such a treasure may exist in Mesopotamia, given its trade links with the Indus civilization. The Mayan decipherment started in 1876 using a sixteenth-century Spanish manuscript that recorded a discussion in colonial Yucatan between a Spanish priest and a Yucatec Mayan-speaking elder about ancient Mayan writing. Mohenjo-daro existed at the same time as the civilizations of ancient Egypt, Mesopotamia and Crete. Credit: Ancient Art and Architecture Collection/Bridgeman Images What we know Indus scholars have achieved much in recent decades. A superb three-volume photographic corpus 3 of Indus inscriptions, edited by the indefatigable Asko Parpola, an Indologist at the University of Helsinki, was published between 1987 and 2010 with the support of the United Nations Educational, Scientific and Cultural Organization; a fourth and final volume is still to come. The direction of writing — chiefly right to left — has been established by analysis of the positioning of groups of characters in many differing inscriptions. The segmentation of texts containing repeated sequences of characters, syntactic structures, the numeral system and the measuring system are partly understood. Views vary on how many signs there are in the Indus script. In 1982, archaeologist Shikaripura Ranganatha Rao published a Sanskrit-based decipherment with just 62 signs 4 . Parpola put 5 the number at about 425 in 1994 — an estimate supported by the leading Indus script researcher in India, Iravatham Mahadevan. At the other extreme is a high estimate 6 of 676 signs, published this year by archaeologist and epigrapher Bryan Wells. Nevertheless, almost every researcher accepts that the script contains too many signs to be either an alphabet or a syllabary (in which signs represent syllables), like Linear B. It is probably a logo-syllabic script — such as Sumerian cuneiform or Mayan glyphs — that is, a mixture of hundreds of logographic signs representing words and concepts, such as &, £ and %, and a much smaller subset representing syllables. As for the language, the balance of evidence favours a proto-Dravidian language, not Sanskrit. Many scholars have proposed plausible Dravidian meanings for a few groups of characters based on Old Tamil, although none of these 'translations' has gained universal acceptance. No firm information is available about its underlying language. A minority of researchers query whether the Indus script was capable of expressing a spoken language, mainly because of the brevity of inscriptions. The carvings average five characters per text, and the longest has only 26. In 2004, historian Steve Farmer, computational linguist Richard Sproat (now a research scientist at Google) and Sanskrit researcher Michael Witzel at Harvard University caused a stir with a joint paper 7 comparing the Indus script with a system of non-phonetic symbols akin to those of medieval European heraldry or the Neolithic Vinča culture from central and southeastern Europe 8 . This theory seems unlikely, for various reasons. Notably, sequential ordering and an agreed direction of writing are universal features of writing systems. Such rules are not crucial in symbolic systems. Moreover, the Indus civilization must have been well aware through its trade links of how cuneiform functioned as a full writing system. Nevertheless, the brevity of Indus texts may indeed suggest that it represented only limited aspects of an Indus language. This is true of the earliest, proto-cuneiform, writing on clay tablets from Mesopotamia, around 3300 bc, where the symbols record only calculations with various products (such as barley) and the names of officials. Digital approach The dissident paper has stimulated some fresh approaches. Wells — a vehement believer that the Indus script is a full writing system — working with the geoinformation scientist Andreas Fuls at the Technical University of Berlin, has created the first, publicly available, electronic corpus of Indus texts (see www.archaeoastronomie.de). Although not complete, it includes all the texts from the US-led Harappa Archaeological Research Project. A group led by computer scientist Rajesh Rao at the University of Washington in Seattle has demonstrated the potential of a digital approach. The team has calculated the conditional entropies — that is, the amount of randomness in the choice of a token (character or word) given a preceding token — in natural-language scripts, such as Sumerian cuneiform and the English alphabet, and in non-linguistic systems, such as the computer programming language Fortran and human DNA. The conditional entropies of the Indus script seem to be most similar to those of Sumerian cuneiform. “Our results increase the probability that the script represents language,” the Rao group has written 9 . Sproat strongly disagrees 10 . On the ground in Pakistan and India, more inscriptions continue to be discovered — although not, as yet, any texts longer than 26 characters. Unfortunately, less than 10% of the known Indus sites have been excavated. The difficulty — apart from funding — is the politically troubled nature of the region. Many of the most promising unexcavated sites lie in the Pakistani desert region of Cholistan near the tense border with India. One such is the city of Ganweriwala, discovered in the 1970s and apparently comparable in size with Mohenjo-daro and Harappa. If these sites, and some others within Pakistan and India, were to be excavated, there seems a reasonable prospect of a widely accepted, if incomplete, decipherment of the Indus script. It took more than a century to decipher the less challenging Mayan script, following several false starts, hiatuses and extensive excavation throughout the twentieth century. Indus-script decipherers have been on the much barer trail — older by two millennia — for less than a century, and excavation of Indus sites in Pakistan has stagnated in recent decades.

Show HN: Explore jobs available on twitter with Twitter jobs board

Use below filters to narrow down your job search Skills Enter comma separated values for multiple skills. City By user Search p p p p p p p p p p p p p p p p p p p p p p p p Status Page Showing 10 out of 4812 jobs posted today. Found 147,485+ jobs in this month We're hiring! Click to apply: Financial Solutions Specialist I - https://t.co/febGbomaEF #NorthernCambria, PA #Banking https://twitter.com/jobshelpers/status/1664694852075003907 Dollar General is hiring in #Sutherland, NE! Click the link in our bio to apply: LEAD SALES ASSOCIATE-PT in SUTHERLAND, NE S15980 #Retail https://twitter.com/cybersecjobsio/status/1664694938557378560 Yessir, we’re hiring too 💯 https://t.co/sWXoWg52jO https://t.co/FvGN5PMFfb https://twitter.com/tmj_buf_eng/status/1664695002428067842 Want to work at Dobbs Peterbilt? We're hiring in Sumner, WA! Click for details: https://t.co/ibBRxn6e6W #JobSearch https://twitter.com/tmj_mi_facmgmt/status/1664695040302776321 SS&C Technologies is hiring in #Denver, CO! Click the link in our bio to apply: Senior Legal Counsel - Hybrid https://twitter.com/tmj_ash_nursing/status/1664695229885345792 🤠 Aldrin is hiring a remote Senior DevOps Engineer #Aldrin #remotework #remotejob #workfromhome #Linux #BASH #BashShell #Web #HTTP #Networking #Security #Grafana #CICD https://t.co/QWlmfJDF6X https://twitter.com/telecomllc/status/1664695314320613376 We're hiring! Click to apply: Retail Sales Consultant - AT&T - https://t.co/tj7lMwlk9G #SalesJobs https://twitter.com/tmj_pa_banking/status/1664695337951592450 UF Health Shands Hospital in Gainesville, FL is hiring Registered Nurse, Pediatrics https://t.co/XcIrjYO79E #Nursing #Nurses #RN @UFHealth https://t.co/7k7N6hBYSh https://twitter.com/tmj_nes_jobs/status/1664695379663769604 We're hiring! Click to apply: Speech Language Pathologist (SLP) - https://t.co/1XA8YmFKJv TN #Nashville, #MusicCity https://twitter.com/adunmaradan_/status/1664695442465247232 We’re hiring throughout North America! From Food Safety Team Members to Management roles, we provide the support and training you will need to succeed. Find your new career at PSSI and apply today! https://t.co/9vA653w9g7 #Florence, KY #FoodSafetyJobs https://twitter.com/redraidering/status/1664695484697702401 Load more jobs

Benzene detected in many sunscreen products

strong – Valisure LLC has tested and detected high levels of benzene, a known human carcinogen, in several brands and batches of sunscreen, which are considered drug products by the Food and Drug Administration (FDA), as well as in after-sun care products, which are generally regulated by FDA as cosmetics. Benzene is known to cause cancer in humans according to the U.S. Centers for Disease Control and Prevention , the U.S. Department of Health and Human Services, the World Health Organization, and other regulatory agencies. The National Institute for Occupational Safety and Health (NIOSH) defines benzene as a carcinogen and lists “inhalation, skin absorption, ingestion, skin and/or eye contact” as exposure routes. 27% of samples tested by Valisure contained detectable benzene and some batches contained up to three times the conditionally restricted FDA concentration limit of 2 parts per million (ppm). Valisure is asking for a recall of the contaminated batches and requesting FDA better define limits for benzene contamination in drug and cosmetic products. It is important to note that not all sunscreen products contain benzene and that uncontaminated products are available, should continue to be used, and are important for protecting against potentially harmful solar radiation. Valisure’s FDA Citizen Petition: Complete Valisure’s FDA Citizen Petition on Sun Care Products (All attachments and other resources linked below.) Valisure is accepting sunscreen products for analysis at no cost: Sunscreen Crowdsourcing Study Link ConsumerLab.com has published a review for its members of the Valisure data suggesting products to avoid and which might be safest with regard to the benzene results. Link to review FDA currently recognizes the serious danger of benzene and lists it as a “ Class 1 solvent ” that “should not be employed in the manufacture of drug substances, excipients, and drug products because of their unacceptable toxicity…However, if their use is unavoidable in order to produce a drug product with a significant therapeutic advance, then their levels should be restricted” and benzene is restricted to 2 ppm for these special circumstances. Being that many of the tested sunscreen and after-sun care products did not contain detectable levels of benzene, it does not appear that benzene use is unavoidable for their manufacture and considering the long history of widespread use of these products, it also does not appear that they currently constitute a significant therapeutic advance; therefore, any significant detection of benzene should be deemed unacceptable. The toxicity of benzene in humans has been well established for over 120 years. The hematotoxicity of benzene has been described as early as 1897. A study from 1939 strongon benzene stated that “exposure over a long period of time to any concentration of benzene greater than zero is not safe,” which is a comment reiterated in a 2010 review of benzene research specifically stating “There is probably no safe level of exposure to benzene, and all exposures constitute some risk in a linear, if not supralinear, and additive fashion.”  Benzene is specifically associated with blood cancers such as leukemia, making absorption through the skin particularly concerning as there have been multiple studies by FDA researchers showing that chemicals in sunscreen products are found in the blood at high levels after application to the skin. “Benzene is one of the most studied and concerning human carcinogens known to science. Its association with forming blood cancers in humans has been shown in numerous studies at trace levels of parts per million and below. The presence of this known human carcinogen in products widely recommended for the prevention of skin cancer and that are regularly used by adults and children is very troubling,” said David Light, Founder and CEO of Valisure. The FDA has clearly determined that benzene should not be used in standard pharmaceutical production at all because of its unacceptable toxicity; however, there currently does not exist any established exposure limit for benzene. The 2ppm concentration limit only applies in special circumstances, which do not include sunscreen manufacturing. Therefore, in addition to recalls, Valisure is also petitioning the FDA to create a concentration limit for standard drug products, including sunscreen, and to also set a daily exposure limit. Public concern regarding contamination of major drug products has significantly increased following a string of medication recalls over the past three years. Most of these recalls of drugs like valsartan, ranitidine, and metformin have been due to the “nitrosamine” class of carcinogens, specifically N-Nitrosodimethylamine (NDMA), which is a Group 2 “probable human carcinogen.” Although NDMA has not been directly linked to cancer in humans, it has both a concentration limit, which ranges between 0.3 – 3.0 ppm for -sartan medications, and a total daily intake limit of 96 nanograms (ng), which is kept constant for all drug products Table of limits set for NDMA and benzene in drug products: *** FDA states that benzene should not be employed in the manufacture of drug substances because of its unacceptable toxicity but does not specifically define limits. Under “special circumstances,” where benzene is unavoidable for the manufacture of a drug with a “significant therapeutic advance,” the 2 ppm limit is similar to that of NDMA. If a similar daily exposure limit is set for benzene of 96 ng, then using the most contaminated sunscreen product Valisure identified of 6.26 ppm equates to approximately 695,800 ng of benzene or 7,248 times the NDMA limit. “There is not a safe level of benzene that can exist in sunscreen products,” stated Dr. Christopher Bunick, MD, PhD, Associate Professor of Dermatology at Yale University. “Even benzene at 0.1 ppm in a sunscreen could expose people to excessively high nanogram amounts of benzene.” Further support for this concern is seen in a 2019 study on sunscreen ingredients where FDA researchers stated , “Understanding the extent of systemic exposure of [sunscreen] products is important, as even a low percentage of systemic absorption could represent a significant systemic exposure.” This study also found that significant amounts of the ingredients in sunscreens are absorbed through the skin and were detected in the blood at high levels.  Additionally, a study by Health Canada’s Bureau of Chemical Hazards has shown that the application of sunscreen specifically increases the absorption rate of benzene through the skin. Absorption through skin and into the blood is particularly concerning regarding benzene, which is associated specifically with blood cancers. Decades of research has shown that regular exposure to benzene at concentrations of 1 ppm, or potentially even lower, have been clearly associated with the development of cancers of blood tissues, such as leukemia. This research has largely focused on exposure through inhalation, though absorption of benzene through skin has also been well established. The National Institute for Occupational Safety and Health (NIOSH) recommends protective equipment be worn by workers expecting to be exposed to benzene at concentrations of 0.1 ppm and defines “skin absorption” as an exposure route. Valisure’s March 24, 2021 Citizen Petition on benzene contamination in hand sanitizer , and the recent recalls of contaminated hand sanitizer products due to the presence of benzene, further underscores the necessity to better regulate benzene and its apparent prevalence in the drug and consumer product supply chains. Beyond the significant concern for public health, there is also evidence that both sunscreen products and benzene pose a serious risk to the environment, marine ecosystems, and United States waterways. The National Oceanic and Atmospheric Administration (“NOAA”) has published reports and infographics intended to educate consumers regarding the potential for sunscreen products to threaten corals and other marine life . Additionally, scientific papers published by NOAA have shown that benzene can be absorbed by fish and short-term exposure (48 hours) to concentrations of benzene at parts per billion levels can significantly reduce survival of certain fish eggs . Furthermore, NOAA has proposed that the use of sunscreen followed by swimming or showering may cause sunscreen chemicals to wash off and enter waterways, an area of significant concern to the Environmental Protection Agency (“EPA”) which extensively regulates benzene. Strict EPA regulations on benzene are detailed in a report authored by the Agency for Toxic Substances and Disease Registry (“ATSDR”)  which stated that, “EPA has set 5 ppb [equivalent to 0.005 ppm] as the maximum permissible level of benzene in drinking water. EPA has set a goal of 0 ppb for benzene in drinking water and in water such as rivers and lakes because benzene can cause leukemia.” Valisure analyzed 294 unique batches from 69 different companies. Significant variability from batch to batch was observed, even within a single company. Fourteen lots of sunscreen and after-sun care products from four different brands contained between 2.78 – 6.26 ppm of benzene; 26 lots from eight brands contained detectable benzene between 0.11 – 1.99 ppm; and 38 lots from 17 brands contained detectable benzene at < 0.1 ppm. Benzene was not detected in an additional 217 batches of sunscreen from 66 different brands through initial analysis of at least one sample. Benzene contamination was detected in sprays, gels, and lotions with both chemical and mineral-based formulations. Benzene is a colorless or light-yellow liquid chemical at room temperature. It has been used primarily as a solvent in the chemical and pharmaceutical industries and is well known to cause cancer in humans. Trace levels of benzene may be found in cigarette smoke, gasoline, glues, adhesives, cleaning products, and paint strippers. “Valisure’s research identifying benzene contamination in multiple over-the-counter sunscreen products is an extremely important discovery for several reasons. First, it warns people practicing sun protection and skin care that some, but not all, sunscreens have potentially hazardous benzene contamination. Second, it is important for people, especially heading into the summer months, to understand that many sunscreen products tested by Valisure did not have benzene contamination, and those products are presumably safe and should continue to be used, along with appropriate hats and sun-protective clothing, to mitigate skin cancer risk,” according to Dr. Bunick. “I believe it is critical that regulatory agencies address benzene contamination in sunscreens, and all topical medications at the manufacturing and final product level, so that all individuals feel safe using sunscreen products.” Valisure’s independent chemical testing has identified several serious drug quality issues, which have resulted in global recalls of ranitidine , a once common antacid medication,strong metformin , a top diabetes drug with roughly 90 million prescriptions written per year in the United States, and hand sanitizer products , marking the first broadly FDA announced drug product recalls for benzene. “These findings of benzene in sunscreens and after-sun care products build upon our substantiated findings of benzene in hand sanitizers that have already led to national recalls,” stated David Light. “It is unfortunately apparent that benzene contamination is a broad and very concerning issue in the American consumer product supply chain and underscores the critical need for independent testing. It is imperative for FDA to expeditiously address current regulatory gaps regarding benzene in both drug and cosmetic products, and we urge FDA and manufacturers to quickly investigate and remove contaminated products from the market.” Complete Valisure FDA Citizen Petition Documents: Full Citizen Petition – Contains lists of products where benzene was detected. Attachment A – Contains list of products where benzene was not detected. Attachment B – Resolution from the American College of Cardiology regarding independent testing. Please read about Responsible Disposal of Potentially Contaminated Products. aSunscreen products are often available in dozens of formulations from numerous companies, and it is estimated by FDA that over 11,000 sunscreen products are on market in the United States. To further assess the pervasiveness of the presence of benzene in sunscreen and after-sun care products on the U.S. market, Valisure is conducting a crowdsourcing study by inviting participants to send in their sunscreen and after-sun care products for testing. For more information and to participate in this study, please click here . bThe views and opinions contained at ConsumerLab.com and its related reports on sun care products are that of ConsumerLab.com, LLC and do not reflect the opinions and beliefs of Valisure, LLC, its members, or Dr. Christopher Bunick.

Y Combinator: Bookmarklet

Thanks to Phil Kast for writing this bookmarklet for submitting links to Hacker News. When you click on the bookmarklet, it will submit the page you're on. To install, drag this link to your browser toolbar: post to HN

The bridge that crossed an ocean (and the man who moved it)

The bridge that crossed an ocean(And the man who moved it) By Lauren Potts In 1831 the last brick was laid on London Bridge - the culmination of years of work by renowned engineer John Rennie. But its tenure as a thoroughfare in Britain’s smoky capital was just a chapter in its lifetime. Exactly 140 years later, the bridge would find a new purpose thousands of miles away in Arizona, as a tourist attraction in a fledgling desert city. Almost as soon as the deal was done, nine-year-old Michael McCulloch was marched to the principal’s office.“He said to me, ‘I just heard your grandfather bought London Bridge’.“I said, ‘the one in the song?’ And he said, ‘yeah’.“And I said, ‘OK’, then went back to class.“When I got home I asked my mom.“She said, ‘yes, RP’s bought London Bridge and he’s going to move it out to Lake Havasu City’.” ‘He was a genius’ Robert Paxton McCulloch was already a multi-millionaire by the time he turned 18. But inheriting part of his grandfather’s fortune in 1925 didn’t blunt his ambition or work ethic. Driven by a compulsion to invent, after graduating from Stanford University he started a company that created supercharged car engines. While he was still a young man McCulloch sold the business for $1m. It was the first chapter in a lucrative career that would take him from the Midwest to California, where he had successful ventures making chainsaws and boat engines. “He was a genius,” says his grandson Michael. “He was astute at building and invention; he designed engines - that was his forte. “He was a little on the eccentric side, but extremely nice. I only ever saw a very gentle person.” McCulloch married Barbara Ann Briggs, daughter of the co-founder of the engine company Briggs and Stratton McCulloch, who was known to his grandchildren as RP, never lost his temper and was fun to be around, Michael says. He was something of a character who kept odd hours, working late into the night and sleeping until noon. Later, he would order the same thing at the same branch of Marie Callender’s pie place; a distinctive sight in the outfit he wore almost every day. “It was a cream chino suit, a very thin black tie, a white shirt, and he had his trousers taken up three inches shorter. He would wear yellow socks, with white and brown saddle shoes,” says Michael. “When he passed away, I wanted a pair of those shoes because that was his trademark. I opened the closet and there were 40 pairs of shoes and 25 suits and nothing else.” Robert Paxton McCulloch was already a multi-millionaire by the time he turned 18. But inheriting part of his grandfather's fortune in 1925 didn’t blunt his ambition or work ethic. Driven by a compulsion to invent, after graduating from Stanford University he started a company that created supercharged car engines. While he was still a young man McCulloch sold the business for $1m. It was the first chapter in a lucrative career that would take him from the Midwest to California, where he had successful ventures making chainsaws and boat engines. McCulloch married Barbara Ann Briggs, daughter of the co-founder of the engine company Briggs and Stratton “He was a genius,” says his grandson Michael. “He was astute at building and invention; he designed engines - that was his forte. “He was a little on the eccentric side, but extremely nice. I only ever saw a very gentle person.” McCulloch, who was known to his grandchildren as RP, never lost his temper and was fun to be around, Michael says. He was something of a character and kept odd hours, working late into the night and sleeping until noon. Robert Jr was the eldest of the couple's four children Later, he would order the same thing at the same branch of Marie Callender’s pie place; a distinctive sight in the outfit he wore almost every day. “It was a cream chino suit, a very thin black tie, a white shirt, and he had his trousers taken up three inches shorter. He would wear yellow socks, with white and brown saddle shoes,” says Michael. “When he passed away, I wanted a pair of those shoes because that was his trademark. I opened the closet and there were 40 pairs of shoes and 25 suits and nothing else.” McCulloch was a man who fizzed with ideas and was often found tinkering at his holiday house in Palm Springs. It was a place of innovation, where the mundane became fascinating. Drawers did not hold typical domestic trappings - they slid open to reveal chilled cocktail glasses. Such quirks led to Life magazine dubbing the country club home a “push-button paradise” in 1956. The issue featured a number of McCulloch’s inventions “It was way, way before its time,” says Michael. “Everything was electronic. All the doors, all the drawers, everything opened by pushing buttons or by sliding a toggle. “There were features that would turn on the coffee [machine] or the bath or the steam room at whatever time you wanted. All the lighting was controlled by a master panel. No other house had that.” After a successful start in making racing engines, McCulloch branched out into boat engines McCulloch was constantly looking for ways to make life easier. The magazine’s front cover featured another of his creations - a human “Lazy Susan” that rotated with the sun to ensure an even tan. “He came up with a lot of ideas,” says Michael. “He created a jet pack, a two-person helicopter. He made a thing that was like a facelift [using] steam, which would tighten your skin. “It was just something he did for fun. He was a very smart, creative person that needed to keep inventing.” Lake Havasu was created after the completion of the Parker Dam in 1938, which held back the Colorado River It was this compulsion, coupled with good business acumen, that led to his next project. With nowhere to test his boat engines in land-locked eastern California, he hopped over the border into Arizona in search of water. The story goes that it was from the air that McCulloch spotted the ribbon of Havasu’s lake below, snaking through a parcel of land flanked by the Chemehuevi, Whipple and Mojave Mountains. “They noticed it was a really, really beautiful place,” says Michael. “They took a look around, offered $76 an acre and started building. “He realised there was a huge opportunity there.” ‘A man of single purpose’ The Bridge House Estates committee of the City of London Common Council had known for some time London Bridge was sinking further into the River Thames with every passing rush hour. Horse-drawn carriages had long since made way for cars and double-decker buses and, over the years, the structure had been hammered deeper into the riverbed. The obvious solution to members in 1965 was to demolish it and start again; build a new bridge, for a new era of commuters. Former newspaper and PR man Ivan Luckin had other ideas. “He thought it was all very well knocking it down, but what about its future?” says former councillor Archie Galloway. “That’s when Ivan made his move. He said to the committee, ‘we ought to sell it’. “A lot of eyebrows went up at that.” Luckin wanted to go a step further and advertise in the United States, where he felt certain someone would be interested in buying a well-known London landmark. “Someone sensibly asked what they might get for [the bridge] and Ivan is recorded as saying, ‘one million’,” says Archie. “And they said, ‘one million dollars?’ “Ivan said, ‘I’m talking about one million pounds.’ [Nearly three millions dollars at the time.] They sat up at that.” News of the sale was soon the subject of newspaper and TV reports trotting out the inevitable line that London Bridge “was falling down”. Luckin’s glossy 40-page brochure for prospective buyers promoted not only the structure itself, but the chance to own a slice of history - a bridge had crossed the Thames in this part of London since Roman times. The new bridge was opened in 1973 by the Queen It was this potential that inspired McCulloch and his business partner, Cornelius Vanderbilt “CV” Wood, who was known for designing Disneyland in California. Stories about how the pair got wind of the sale vary - according to Michael, they saw an advert on TV while on business in London. “They had been having a few drinks together and saw it and I think they probably looked at each other and thought, ‘we should take that apart, that will be our gimmick’.” Another version has Wood at a meeting about the sale of the RMS Queen Mary, at New York’s Plaza Hotel, where he supposedly asked, “anything else for sale?” Either way, Luckin’s perseverance paid off and when the ink dried on the contract of sale on 18 April 1968, he was a man vindicated - his outlandish idea matched only by an even more ludicrous one. “Ivan was immensely proud of it but never went around shouting it from the rooftops,” says Archie. “I think a lot of people thought he was a lunatic, but he was a man of single purpose. “I can’t believe anyone else would have thought of the idea.” ‘Godforsaken place’ When his parents announced they were moving to Lake Havasu City in February 1965, Rick Kingsbury was incredulous. It sounded to him the place barely existed - and he was right. “It wasn’t my idea to move here; no 13-year-old kid in his right mind would want to move to a godforsaken place like this at the time. “We had no electricity, no water, we didn’t have air conditioning. There was one phone, the post office was a closet at the liquor store and the mail went to a box 20 miles away. “Those first few months were like a big camping trip you were never going to go home from. “There was nothing here. I hated the place.” In 1965 there were about 20 homes and a trailer park For Joe Hunnicutt, the worst thing was the dizzying heat of summers inching towards 50C (122F). “I absolutely hated it. My parents took me out of Wyoming and they took me to hell. “We came in August and I stepped outside and I was like,‘are you kidding me?’ “I was eight and all I wanted to do was ride my bike. I couldn’t do that there. “My whole goal was to get out, but you’ve got to give it to them, the people that stuck it out in that godforsaken middle of nowhere.” Those who moved to Lake Havasu between 1964 and 1974 call themselves pioneers - a nod to their 19th Century forefathers who travelled westwards in wagons in search of a better life. But Rick wasn’t imbued with such spirit and it was a year before he saw the merits of living somewhere with one stop sign, two roads and 600 neighbours. The first 10 houses were built in 1964 by Alfred Anderson “I was a reluctant pioneer,” says Rick. “It was only during that second summer I realised what a paradise it was for a kid. “There was all this exploring in the desert, so there was this feeling of Tom Sawyer, of living in this wonderland. “We had this great big lake pretty much to ourselves and we were up to our necks in the cool water. “It was this huge playground and it was all ours.” By the end of 1965, the elementary school had acquired two new classes, and a grocery store had opened its doors. The real change, however, came in 1968, when news trickled back to this tiny desert “city” that its founder had made a bizarre purchase. In the beginning there was just one major crossroad “The whole economics of the place and the attitude of the town changed overnight,” says Rick. “It was like a rocket ship, it took off and went crazy. In the three years it took to build the bridge, the population tripled. “But there was almost a little regret for me. We were growing and I realised it was going to go away. “It was all going to change.” ‘A good life’ Visitors to Lake Havasu City might assume it existed only because its founder had embarked upon a curious plan to rebuild a piece of British history in the desert. In reality, McCulloch’s small community grew steadily in its fledgling years. The US census recorded that by 1970 the city had more than 4,000 residents - a substantial increase on the handful that called it home six years earlier. But even the engineer’s most faithful supporters struggled to share his vision back in 1964. Early West remembers how in the spring of that year, he and McCulloch had been drinking coffee outside the entrepreneur's new factory in Lake Havasu when, seemingly out of the blue, he said: “I see a city.” “I didn’t know what to say; there was nothing, just desert,” says Early. “I said, ‘I don’t see that city’. And Robert said, ‘I promise, you will’.” Adverts appeared in newspapers offering free trips to Lake Havasu City Lake Havasu’s growth was in part down to dynamic marketing of the American dream - the promise of a better (and sunnier) life. “RP started doing commercials and advertising in the Midwest where it was freezing,” says Michael. “He purchased a bunch of planes and started flying people out on free flights. “They were treated like kings. They were put up in a beautiful hotel for two or three days, had all their meals paid for and all they had to do was go on a tour for a couple of hours. “And it worked like a charm. They sold a lot of those homes that way.” Visitors to Lake Havasu City might assume it existed only because its founder had embarked upon a curious plan to rebuild a piece of British history in the desert. In reality, McCulloch’s small community grew steadily in its fledgling years. The US census recorded that by 1970 the city had more than 4,000 residents - a substantial increase on the handful that called it home six years earlier. But even the engineer’s most faithful supporters struggled to share his vision back in 1964. The city’s first residents were given the grand tour by the salesmen Early West remembers how in the spring of that year, he and McCulloch had been drinking coffee outside the entrepreneur's new factory in Lake Havasu when, seemingly out of the blue, he said: “I see a city.” “I didn’t know what to say; there was nothing, just desert,” says Early. “I said, ‘I don’t see that city’. And Robert said, ‘I promise, you will’.” A salesman was assigned to each couple that stepped off the plane Lake Havasu’s growth was in part down to dynamic marketing of the American dream - the promise of a better (and sunnier) life. “RP started doing commercials and advertising in the Midwest where it was freezing,” says Michael. “He purchased a bunch of planes and started flying people out on free flights. “They were treated like kings. They were put up in a beautiful hotel for two or three days, had all their meals paid for and all they had to do was go on a tour for a couple of hours. “And it worked like a charm. They sold a lot of those homes that way.” Hundreds boarded the McCulloch-branded propeller planes on chilly runways in Chicago, Boston and New York. On arrival, they were chauffeured to the Havasu Hotel, where a waterfall cascaded dramatically over its front door. “It was unbelievable,” says Evie Cistaro, who visited from Boston on the Thanksgiving weekend of 1972. “We stopped and picked up people in Pennsylvania and Chicago and when we arrived it was 75F (24C) and the sun was shining. “They wined us and dined us and showed us the land, and we bought it and signed. “It was a new beginning.” A salesman was assigned to each couple that stepped off the plane Joe Hunnicutt remembers hearing the sales pitch from his father, Ray, who sold lots for McCulloch. “They drove people around to buy property and I used to sit in the back of the Jeep with them,” he said. “I had that [pitch] memorised... some of it was true, some of it wasn’t.” Hundreds boarded the McCulloch-branded propeller planes on chilly runways in Chicago, Boston and New York. On arrival, they were chauffeured to the Havasu Hotel, where a waterfall cascaded dramatically over its front door. “It was unbelievable,” says Evie Cistaro, who visited from Boston on the Thanksgiving weekend of 1972. “We stopped and picked up people in Pennsylvania and Chicago and when we arrived it was 75F (24C) and the sun was shining. “They wined us and dined us and showed us the land, and we bought it and signed. “It was a new beginning.” Evie Cistaro said there were about 100 people on their flight Joe Hunnicutt remembers hearing the sales pitch from his father, Ray, who sold lots for McCulloch. “They drove people around to buy property and I used to sit in the back of the Jeep with them,” he said. “I had that [pitch] memorised... some of it was true, some of it wasn’t.” Betty Gunnarsson says several salesmen were waiting at the airstrip in matching blue blazers after the 13-hour flight from Long Island. “As soon as we got off the plane the pressure was on,” says the 75-year-old. “The salesman took us out on a boat and to show how clear the water was, he dipped a glass in the lake and drank it. “So we bought a bit of property. I thought, ‘it’s an investment’. “But I remember people walking up and down the aisles [on the flight back] saying, ‘who wants to buy a pile of rocks?’ “And I thought, ‘they don't see beyond’.” Evie’s salesman brother, Jay Scavusso, found the Gunnarssons their home The Gunnarssons moved three years later into the only house on an undeveloped hill. “People used to say… ‘you live way up there?’,” says Brooklyn-born Betty. “You could look out for almost a mile. The boys had bikes and they would go out into the desert and [crumble crackers] to make trails so they didn’t get lost. “The experience opened up their eyes. It was a good life.” ‘Crazy like a fox’ Before April 1968 few people knew where Lake Havasu City was, but news of the sale made international headlines and gave the tiny desert community a high profile. However, the purchase of London Bridge was also the source of much scorn. Many people assumed McCulloch had lost his mind - although others recognised there might be some cunning behind the bizarre scheme. “There were lots of articles that described RP as crazy,” says Michael. “Then that changed - to crazy like a fox. “People couldn’t conceptualise it… take this bridge in London and put it in the middle of the desert?” A rumour that McCulloch - a millionaire and successful businessman - had “bought the wrong bridge” probably didn’t help, although it was, of course, complete nonsense. McCulloch and Wood went to London to inspect their purchase “I know for a fact they knew exactly which bridge they were going to buy,” says Michael. “Do you think anyone in their right mind would look at Tower Bridge and think they could take it apart and bring it here? “But RP and CV felt that by allowing that rumour to perpetuate, it gave them continued publicity, so they let it go.” Before April 1968 few people knew where Lake Havasu City was - but news of the sale made international headlines and gave the tiny desert community a high profile. However, the purchase of London Bridge was also the source of much scorn. Many people assumed McCulloch had lost his mind - although others recognised there might be some cunning behind this bizarre scheme. “There were lots of articles that described RP as crazy,” says Michael. “Then that changed - to crazy like a fox. “People couldn’t conceptualise it... take this bridge in London and put it in the middle of the desert?” A rumour that McCulloch - a millionaire and successful businessman - had “bought the wrong bridge” probably didn’t help, though it was, of course, complete nonsense. McCulloch and Wood went to London to inspect their purchase “I know for a fact they knew exactly which bridge they were going to buy,” says Michael. “Do you think anyone in their right mind would look at Tower Bridge and think they could take it apart and bring it here? “But RP and CV felt that by allowing that rumour to perpetuate, it gave them continued publicity, so they let it go.” In the soon-to-be-home of London Bridge, the jokes were in abundance. “We started a rumour McCulloch was going to bring the Leaning Tower of Pisa and put it in the park,” laughs Rick Kingsbury. “We didn’t need a bridge,” says Bobbi Holmes. “We thought, ‘for where?’” She had watched Lake Havasu City grow from her family’s tiny resort six miles away on the California side of the lake. “We only had three TV channels and we didn’t have a local newspaper, but I remember hearing from someone that McCulloch was buying London Bridge and thinking it was just a silly rumour. “We thought it was hilarious, we thought it was a joke.” Bobbi Holmes remembers seeing the bricks piled up on the side of the road The nearest high school was in the city, so Bobbi had to take a boat across the lake. She remembers how the McCulloch salesmen would point her out as a local curiosity to those fresh off the flights, and say “there’s the girl that comes by boat”. She watched as the bricks piled up on the side of the road slowly transformed into a bridge. “I thought it was pretty bizarre at the time,” she says. After all, there was no river in the city for the exported bridge to cross. “Most people build bridges to cross a river, McCulloch built a river under a bridge.” English expat Linda Binder felt similarly when she heard the bridge she had often walked across as a child was moving. “I saw in the papers McCulloch was buying London Bridge and I thought, ‘what are those crazy Yanks doing building a bridge on sand?’ “But I think he did an extraordinary job of bringing it over and having the foresight to bring it brick by brick and rebuild it with the city around it. “I mean, who does that?” Years later, the Londoner moved to Lake Havasu where she became an Arizona state senator and, for many years, the point of contact for British dignitaries visiting the new city. Her husband, Bill, recalls how his friend CV Wood had to get permission from the President of the United States to create a channel under London Bridge. US president Lyndon B. Johnson had to be persuaded by Wood to sign off on the channel “Woody told me the [plan] was to build the bridge on dry land and dredge it out, but they applied for a permit and got denied,” says Bill. Wood was told such permission could be only be granted by the highest authority in the land, so he used his connections to secure a 10-minute audience with Lyndon Johnson, only to be told by the president: “Sorry, you can’t set this precedent. You can’t move a river.” “Woody said the pressure was really on,” says Bill. “He was a Texas boy and so was Johnson, so he started talking with this Texan accent. “He said ‘by golly, the next time someone buys London Bridge and moves it to the US, well, you’ll have to move another damn river’. “Johnson looked at Woody and said, ‘sign it’ [to an aide about the necessary paperwork] and left the room.” Wood, right, was an instrumental part in McCulloch’s plan It wasn’t the only time Wood - who was married to Hollywood actress Joanne Dru - put his famed persuasiveness into action. He is also said to have taken less than two hours to convince 16 environmental agencies opposed to the bridge to get on board. “The reality is RP was an engineer, an entrepreneur,” says Michael. “But CV, he was a showman; he could sell ice to Eskimos. “That’s why they were so good together. They made a good pair.” ‘It was complicated’ The first stone arrived in Lake Havasu on 9 July 1968, its swift delivery in part because work had begun to take it down long before McCulloch had signed a cheque for $2.46m. Alan Saines numbered the pieces as they came off the bridge in London. “There was only me numbering them and there were hundreds. If we did 20 a week, we were lucky. “It wasn’t a case of smashing it; the [stones] had to be taken down carefully and stored carefully and all the time they were building the new bridge.” Alan Saines was 17 when he got his first job, taking down the bridge Alan Simpson worked on the design of the new bridge for engineers Mott, Hay and Anderson. He said the old bridge came down in slices to make room for the new one. “The only way of doing it was to take down one side of the old bridge so there was just room for the new bridge, then move over to the other side, switch the traffic, then take down the rest of the bridge in the middle and tie [it] back together. “It was complicated. And it wasn’t until fairly late in the day that taking it all down carefully to ship halfway around the world became an additional requirement.” Dockers put the stones on to freighters bound for the US The stones were sent to Surrey Docks, where fellow Mott engineer Bernard Waterworth would consult a long-range telescopic photo of the bridge and direct some pieces on to freighters bound for Long Beach, California, via the Panama Canal. Others first went to Merrivale Quarry in Devon, where the faces were sliced off to cut down on shipping costs, before continuing their 5,000-plus mile journey. “It was an historic monument and you were aware it was unusual,” explains Bernard. “You don’t forget it - projects like that are few and far between.” About 10,600 stones arrived in Lake Havasu, making the final leg of the journey by road on flatbed trucks. Curious tourists would line the roads to watch them trundle into the storage compound, where the stones were unloaded and put in order, ready to be used by construction crews. The interior of the bridge was sold for $2.5m and made into counter-tops and headstones “We would see the trucks coming down the highway and we would sit and watch them go by,” says Norma Grzesiowski, who had just moved to nearby Desert Hills. "We were thinking, ‘what are they doing? What are they building?’” The cornerstone was lowered by Sir Gilbert Inglefield as McCulloch and Wood looked on “We decided to steal some of the stones and the four of us were in my ’53 Chevy station wagon waiting with bolt-cutters to get in the yard,” remembers Rick Kingsbury. “But those things weighed three times as much as my car did! The ground was shaking as they went by on those eight-axle trucks. “We thought... ‘well that’s a bad idea’.” About 10,600 stones arrived in Lake Havasu on flatbed trucks. Curious tourists would line the roads to watch them trundle into the storage compound, where the stones were unloaded and put in order, ready to be used by construction crews. The interior was sold for $2.5m and made into counter-tops and headstones “We would see the trucks coming down the highway and we would sit and watch them go by,” says Norma Grzesiowski, who had just moved to nearby Desert Hills. "We were thinking, ‘what are they doing? What are they building?’” The cornerstone was lowered by Sir Gilbert Inglefield as McCulloch and Wood looked on “We decided to steal some of the stones and the four of us were in my ’53 Chevy station wagon waiting with bolt-cutters to get in the yard,” remembers Rick Kingsbury. “But those things weighed three times as much as my car did! The ground was shaking as they went by on those eight-axle trucks. “We thought... ‘well that’s a bad idea’.” The cornerstone was laid on 23 September 1968 in a ceremony attended by London’s lord mayor. Then work to resurrect the bridge began, starting with the arches, which were laid across the sand. The rebuild began by calculating the angle of the arches “You just saw the skeleton going up, and then they put the face on,” says Norma, who used to sit in a boat on the lake and watch. “The builders explained it was like putting together a puzzle. It was exciting, but it was surreal.” Construction workers had to build the bridge in the sweltering Arizona heat The rest was built around the framework in reinforced concrete. To avoid the sinking fate of its 133,000-tonne predecessor, it was made hollow and finished with a veneer of stone, making it 33,000 tonnes. “It almost felt like I was in a movie,” says former labourer Roy Martin, who helped winch the stones. “People were always coming up to us and asking questions, wanting to take pictures and wanting to know if we had any extra pieces they could take home as souvenirs.” The stones were manoeuvred in by hand and secured with pins, wire and concrete. The work was slow and laborious - crews had to force the 400-500lb pieces in one at a time. On a good day they’d get through 10; the worst was one. If a stone was too damaged to use, it had to be replaced, says Harvey Robertson, who laid steps on the south side of the bridge. But the local rock didn’t match that which had been blackened by a century of London pollution. “It was a lighter colour, so they had these kerosene burners making soot,” he recalls. “Then they painted it on there to make it look dark and weathered, so people wouldn’t know they weren’t the original stones.” The dredge swung across the channel, eating out a path for the water Once the bridge was up, the dunes were sucked out of the arches to make way for the president-approved channel. The land underneath the structure was dredged, allowing London Bridge to tower above a waterway for the first time since it had left Britain. “I remember thinking at the time McCulloch had more money than sense,” says Harvey. “But once I started working on it, I thought he was pretty much a genius.” ‘Home at last’ On 10 October 1971, Lake Havasu’s London Bridge opened to great fanfare. The British were out in force, represented by a delegation from London accompanied by a troupe of red-jacketed horse-riders trotting through the desert. McCulloch, London's lord mayor Sir Peter Studd and Wood attended the dedication The English Village offered a taste of home with its red phone box, double-decker bus and pub serving Yorkshire pudding and roast beef. But the jewel was the bridge, with the British and US flags flying down both sides. Riders on horseback made a strange sight against the desert backdrop “The whole place was just buzzing,” says Norma Grzesiowski. “They had a big tent down there on the bridge and lots of events and games going on. “I remember them releasing the birds and they had a big hot-air balloon with the British flag on. It was breathtaking - such an exciting thing to be part of.” There was a party atmosphere as the anticipation built According to a report in the Lake Havasu City Herald at the time, some 50,000 people lined the roads to watch the celebrations. The gala featured a float-filled parade and the dramatic release of 30,000 balloons and 3,000 birds, to a backdrop of “home at last” which had been scrawled across the sky. The dedication day festivities were an explosion of colour Bobbi Holmes performed in the opening ceremony - an experience, she jokes, that scarred her for life. “I was a member of the pom-pom girls and we did a dance. We were carrying these pom-poms and doing this routine and my arms felt so heavy. I thought, ‘I never want to be in a parade again’. “But it was really fun and people came from all over. The English Village reminded me of Disneyland, which I later realised was because CV was behind it. “It was pretty, sparkly, clean, with beautiful flowers. It was kind of magical.” On 10 October 1971, Lake Havasu’s London Bridge opened to great fanfare. The British were out in force, represented by a delegation from London accompanied by a troupe of red-jacketed horse-riders trotting through the desert. Riders on horseback made a strange sight against the desert backdrop The English Village offered a taste of home with its red phone box, double-decker bus and pub serving Yorkshire pudding and roast beef. But the jewel was the bridge, with the British and US flags flying down both sides. Rick Kingsbury’s Polaroids showed the scale of the event “The whole place was just buzzing,” says Norma Grzesiowski. “They had a big tent down there on the bridge and lots of events and games going on. “I remember them releasing the birds and they had a big hot-air balloon with the British flag on. It was breathtaking - such an exciting thing to be part of.” According to a report in the Lake Havasu City Herald at the time, some 50,000 people lined the roads to watch the celebrations. The gala featured a float-filled parade and the dramatic release of 30,000 balloons and 3,000 birds, to a backdrop of “home at last” which had been scrawled across the sky. There was a party atmosphere as the anticipation built Bobbi Holmes performed in the opening ceremony - an experience, she jokes, that scarred her for life. “I was a member of the pom-pom girls and we did a dance. We were carrying these pom-poms and doing this routine and my arms felt so heavy. I thought, ‘I never want to be in a parade again’. “But it was really fun and people came from all over. The English Village reminded me of Disneyland, which I later realised was because CV was behind it. “It was pretty, sparkly, clean, with beautiful flowers. It was kind of magical.” “Everyone came out,” says Joni Echelberger, whose family had not long moved from Illinois. “My aunt and uncle got to have the $100 dinner in the big tent and I was so jealous because I wanted to be in there. “We were outside waiting and waiting and waiting for the fireworks and they wouldn’t go off until the celebration was done [so] they didn’t go off until midnight. “For such a small community, it was really a big deal.” A long red carpet led to the huge striped tent Joe Hunnicutt remembers a “crazy” day that was bursting with community pride. But he tells a different story about those birds, claiming McCulloch hadn’t been able to find doves as planned. Instead, he mainly settled for pigeons. “I remember standing there with my dad and he was wearing his blazer and tie, smoking a cigar, and he says, ‘You know, I think McCulloch is going to regret this particular decision.' “And since 10 October 1971, Lake Havasu City has been infested with pigeons. “They’re the joke and scourge of the town; you can’t get rid of them.” Six years later, McCulloch died from an accidental overdose of alcohol and pills. His death was keenly felt in the community he had built, where his legacy is remembered with a bronze statue and a road bearing his name. Early West, who worked for him for 40 years, describes him as a man who “thought number one about his employees” and remained down to earth. A statue of McCulloch and Wood stands at one end of the bridge “Bob was a rich man, but he was pretty common,” the 86-year-old says. “Ordinary people live on a certain level, and he didn’t get too much above that. “I never noted him putting himself above anyone else.” Early West was convinced by McCulloch to move from the entrepreneur’s LA factory to Lake Havasu Fifty years on and people are still poking fun at McCulloch’s dream. At this year’s Academy Awards, host Jimmy Kimmel made a jibe about Lake Havasu while encouraging Oscar winners to keep their acceptance speeches short. His gag actually caused a surge in interest in the resort - not that it really needed the help. According to the state tourist board, 3.65m people visited London Bridge in 2017 - making it Arizona’s third most popular attraction after the Grand Canyon and the Glen Canyon national parks. Thousands of tourists have pierced a map in the visitors’ centre with pins marking their home state McCulloch’s grandson Michael says people come because they’re "curious". It’s a word that sums up Lake Havasu perfectly. Driving in from Highway 95, the billboards scream “Visit the home of the London Bridge”. Before long, a flash of blue appears - not a desert mirage, but the waters of the Colorado River filling the city’s man-made lake up ahead. Then before you know it, the bridge appears, dominating a landscape peppered with fast food outlets, hotels and palm trees. It’s a surreal moment, even though you’ve been warned what to expect. For the 52,000 residents that call Lake Havasu City home, the bridge is just a part of everyday life, providing the only access to and from the marina and the homes perched on the first spit of land purchased by McCulloch back in 1958. And although the whole project cost about $12m, McCulloch is said to have recouped his costs before the last brick had even been laid in its new home. The English Village was recently refurbished after standing derelict for years Today, the resort thrives not just on tourism from the bridge, but from income from water sports enthusiasts and college kids on spring break. The calm lake and sunny weather have made the city a legitimate holiday destination - an outcome even its founder would probably not have predicted. That a sinking bridge in London would be the making of a city in Arizona is a testament to McCulloch. But that it became such a success is down to boldness, determination and belief - a reflection, perhaps, of the American spirit. Michael McCulloch says Lake Havasu was his grandfather's "pride and joy" “The bridge changed things immensely, I don’t know what the city would be like if it wasn’t here,” reflects Michael. “It was [supposed to be] a retirement resort but it’s turned into a sustainable economy and a lot of people are born here and have chosen not to leave. “I think RP would be amazed by that.”

Liberal Hypocrisy Is Fueling American Inequality. Here’s How. – NYT Opinion

Your browser isn’t supported anymore. Update it to get the best YouTube experience and our latest features. Learn more

New evidence shows this uranium cube is likely relic of Nazi A-bomb program

For decades, the Pacific Northwest National Laboratory (PNNL) has been home to an unusual artifact from World War II: a small cube of solid uranium metal, measuring about two inches on each side and weighing just under 2.5 kilograms. Lab lore holds that the cube was confiscated from Nazi Germany's failed nuclear reactor experiments in the 1940s, but that has never been experimentally verified. PNNL scientists are developing new nuclear forensic techniques that should help them confirm the the pedigree of this cube—and others like it—once and for all. Those methods could also eventually be used to track illicit trafficking of nuclear material. PNNL's Jon Schwantes and graduate student Brittany Robertson presented some of their initial findings this week at the fall meeting of the American Chemical Society (a hybrid virtual/in-person event). University of Maryland physicist Timothy Koeth is among the outsider collaborators in this ongoing research. He has spent over seven years tracking down these rare artifacts of Nazi Germany's nuclear research program, after receiving one as a gift. As of 2019, he and a UMD colleague, Miriam Herbert, had tracked down 10 cubes in the US: one at the Smithsonian, another at Harvard University, a handful in private collections—and of course, the PNNL cube. What makes these cubes so special is their historical significance. As we reported previously: Underpinning the Manhattan Project in the US was the fear that German scientists under Adolf Hitler's Nazi regime would beat the Allies to a nuclear bomb. The Germans had a two-year head-start, but according to Koeth, "fierce competition over finite resources, bitter interpersonal rivalries, and ineffectual scientific management" resulted in significant delays in their progress toward achieving a sustained nuclear reaction. German nuclear scientists were separated into three isolated groups based in Berlin (B), Gottow (G), and Leipzig (L). Renowned physicist Werner Heisenberg headed up the Berlin group, and as the Allied forces advanced in the winter of 1944, Heisenberg moved his team to a cave under a castle in a small town called Haigerloch—now the site of the Atomkeller Museum. That's where the group built the B-VIII reactor. It resembled an "ominous chandelier," per Koeth, because it was composed of 664 uranium cubes strung together with aircraft cable and then submerged in a tank of heavy water shielded by graphite to prevent radiation exposure. As the German scientists were racing against time, Manhattan Project lead Lieutenant General Leslie Groves kicked off a covert mission dubbed "Alsos," with the express purpose of gathering information and materials related to Germany's scientific research. When the Allied forces closed in at last, Heisenberg took apart the B-VIII experiment and buried the uranium cubes in a field, ferreting away key documentation in a latrine. (Pity Samuel Goudsmit, the poor physicist who had to dig those out.) Heisenberg himself escaped by bicycle, carrying a few cubes in a backpack. As Heisenberg himself acknowledged, the German scientists' final experiment failed because the amount of uranium in the cubes was insufficient to trigger a sustained nuclear reaction. But Heisenberg was confident that "a slight increase in its size would have been sufficient to start off the process of energy production." A model described in a 2009 paper bears that out, showing that the group would only have needed 50 percent more uranium cubes to get the design to work. If it had, our world might look very different today. The Alsos team purportedly brought the cubes confiscated from Berlin to the United States for use in the uranium processing facility at Oak Ridge. However, Koeth learned that, by April 1945, the US didn't need additional feedstock material. And there is no official record of any cubes entering the country, so most of them have never been accounted for. Ditto for the 400 or so uranium cubes that had been in use by the Gottow group, led by Kurt Diebner. According to PNNL lore, their cube was stored at DOE headquarters until 1989. That's when it was brought to the laboratory as a radiation training tool for RadCAD, a set of hands-on courses on the detection and interception of illicit trafficking in radioactive materials. The PNNL cube, like its brethren, is made of solid natural uranium metal. The cubes are only slightly radioactive and don't pose a health concern. Because uranium is so dense, it essentially shields itself. Any measured radiation comes from the surface. Nonetheless, the PNNL cube is kept in a double-plexiglass container to prevent exposure to radiation during handling and contamination of the cube from oxidization, according to Robertson. The PNNL scientists were pretty confident they had a "Heisenberg cube"; among other evidence, the cube is notched, the better to hang on the cables used in the German reactor efforts. But that evidence is largely anecdotal, per Robertson and Schwantes. The cube was analyzed back in 2002 via high-resolution gamma spectroscopy to get an estimate of its age, but those results were inconclusive. "That's typically not sensitive enough to provide an accurate age for the cube," said Schwantes. Several years ago, as the PNNL cube was being repackaged, Schwantes and a colleague shaved a few small samples off the metal for analysis. They hoped to confirm once and for all that it is one of the Heisenberg cubes—or possibly a "Diebner cube." Robertson's work—part of her doctoral thesis research—is to study those samples using her own modified analytic techniques, in conjunction with PNNL's standard nuclear forensic methods. For instance, radiochronometry is a popular method with geologists. It's commonly used to determine the age of a uranium-rich material by measuring the byproducts of the uranium's decay, namely the radioactive isotope thorium-230 and protactinium. Robertson's modified approach involves simultaneously separating the thorium and protactinium in the hope that the materials' relative concentrations will give some indication of when the cube was made. In addition, analyzing the rare-earth-element impurities could help PNNL scientists determine where the original uranium was mined. So far, initial findings have confirmed that at least one of the three cubes being tested at PNNL is natural uranium. There are also preliminary results from Robertson's analysis of the coatings the Germans applied to the cubes to keep oxidation in check. Cyanide-based coatings were used by the Berlin group, while Diebner's Gottow group used styrene-based coatings. If one could accurately measure the relevant signatures, it would enable the team to tell whether a given cube came from the Berlin or Gottow group. "As far as we know, no one else has performed this measurement," said Robertson. "And I have to be honest, I thought it was kind of a long shot. I didn't think an organic would last sitting next to uranium metal for this many decades and still be detectable." That long shot paid off. Koeth's cube was among those tested, revealing a styrene coating—a bit of a surprise, given that Koeth's historic sleuthing tracked the cube to the Berlin group. However, it turns out that Diebner sent some of his group's cubes to Heisenberg in Berlin when the latter sought more fuel for his reactor. So Koeth's cube may possibly have been used by both groups.

Who needs a teacher? Artificial intelligence designs lesson plans for itself

Unlike human students, computers don't seem to get bored or frustrated when a lesson is too easy or too hard. But just like humans, they do better when a lesson plan is "just right" for their level of skill. Coming up with the right curricula isn't easy, though, so computer scientists wondered: What if they could make machines design their own? That's what researchers have done in several new studies, creating artificial intelligence (AI) that can figure out how best to teach itself. The work could speed learning in self-driving cars and household robots, and it might even help crack previously unsolvable math problems. In one of the new experiments, an AI program tries to quickly reach a destination by navigating a 2D grid populated with solid blocks. The "agent" improves its abilities through a process called reinforcement learning, a kind of trial and error. To help it navigate increasingly complex worlds, the researchers—led by University of California (UC), Berkeley, graduate student Michael Dennis and Natasha Jaques, a research scientist at Google—considered two ways in which they could draw the maps. One method randomly distributed blocks; with it, the AI didn't learn much. Another method remembered what the AI had struggled with in the past, and maximized difficulty accordingly. But that made the worlds too hard—and sometimes even impossible—to complete. So the scientists created a setting that was just right, using a new approach they call PAIRED. First, they coupled their AI with a nearly identical one, albeit with a slightly different set of strengths, which they called the antagonist. Then, they had a third AI design worlds that were easy for the antagonist—but hard for the original protagonist. That kept the tasks just at the edge of the protagonist's ability to solve. The designer, like the two agents, uses a neural network—a program inspired by the brain's architecture—to learn its task over many trials. After training, the protagonist attempted a set of difficult mazes. If it trained using the two older methods, it solved none of the new mazes. But after training with PAIRED, it solved one in five, the team reported last month at the Conference on Neural Information Processing Systems (NeurIPS). "We were excited by how PAIRED started working pretty much out of the gate," Dennis says. In another study, presented at a NeurIPS workshop, Jaques and colleagues at Google used a version of PAIRED to teach an AI agent to fill out web forms and book a flight. Whereas a simpler teaching method led it to fail nearly every time, an AI trained with the PAIRED method succeeded about 50% of the time. The PAIRED approach is a clever way to get AI to learn, says Bart Selman, a computer scientist at Cornell University and president of the Association for the Advancement of Artificial Intelligence. Selman and his colleagues presented another approach for so-called "autocurricula" at the meeting. Their task was a game called Sokoban, in which an AI agent must push blocks to target locations. But blocks can get stuck in dead ends, so success often requires planning hundreds of steps ahead. (Imagine rearranging large furniture in a small apartment.) Their system creates a collection of simpler puzzles to train on, with fewer blocks and targets. Then, based on the recent performance of their AI, it selects puzzles that the agent only occasionally solves, effectively ratcheting the lesson plan to the right level. Sometimes, the right puzzles are hard to predict, Selman says. "The notion of what is a simpler task is not always obvious." The researchers tested their trained agent on 225 problems that no computer had ever solved. It cracked 80% of them, with about one-third of its success coming strictly from the novel training method. "That was just fun to see," Selman says. He says he now receives astounded messages from AI researchers who've been working on the problems for decades. He hopes to apply the method next to unsolved math proofs. Pieter Abbeel, a computer scientist at UC Berkeley, also showed at the meeting that autocurricula can help robots learn to manipulate objects. He says the approach could even be used for human students. "As an instructor, I think, ‘Hey, not every student needs the same homework exercise,'" Abbeel says, noting that AI could help tailor harder or easier material to a student's needs. As for AI autocurricula, he says, "I think it's going to be at the core of pretty much all reinforcement learning."

No Tech for Apartheid: 40 Groups Demand Amazon and Google Ditch Israeli Military

To donate by check, phone, or other method, see our More Ways to Give page. × A Palestinian child collects debris from the rubble of a destroyed mosque in Beit Lahia, in the northern Gaza Strip, on May 27, 2021. (Photo: Thomas Coex/AFP via Getty Images) 'No Tech for Apartheid': 40+ Groups Demand Amazon and Google Ditch Israeli Military "Courageous workers at Google and Amazon are calling on their employers to stop enabling the Israeli government's oppression of Palestinian families—and we are proud to stand with them." Oct 13, 2021 A day after hundreds of Amazon and Google workers condemned their employers for complicity in Israel's human rights violations against Palestinians, over 40 grassroots groups on Wednesday announced a campaign to amplify the efforts of activists around the world working to stop apartheid profiteers. "As the Israeli military bombed homes, clinics, and schools in Gaza and threatened to push Palestinian families from their homes in Jerusalem this past May, Amazon Web Services and Google Cloud executives signed a $1.22 billion contract to provide cloud technology to the Israeli government and military," the campaign noted. "By doing business with Israeli apartheid, Amazon and Google will make it easier for the Israeli government to surveil Palestinians and force them off their land." "Technology should be used to bring people together, not enable apartheid and ethnic cleansing." The "No Tech for Apartheid" campaign was launched by groups including the Adalah Justice Project, American Friends Service Committee (AFSC), Arab Resource and Organizing Center (AROC), the Boycott, Divestment, and Sanctions (BDS) Movement, Center for Constitutional Rights (CCR), Data for Black Lives, Fight for the Future, Jewish Voice for Peace (JVP), MPower Change, Palestine Solidarity Campaign (PSC), and World Beyond War. "We're heeding the call from hundreds of Google and Amazon workers to rise up against the contract, known as Project Nimbus," the campaign said. "Technology should be used to bring people together, not enable apartheid and ethnic cleansing. Following in the footsteps of those who fought to divest from apartheid South Africa and won, it's our responsibility to rise up in support of Palestinian freedom." International critics--including leading human rights groups, survivors of South African apartheid, and former U.S. President Jimmy Carter--have argued that Israel's ongoing ethnic cleansing of Palestinians, illegal Jews-only settlements, segregated roads, separation wall, and other policies and actions constitute a form of apartheid. The No Tech for Apartheid campaign follows an open letter signed by more than 500 Amazon and Google employees calling on the tech giants to cancel Project Nimbus, which they said "allows for further surveillance of and unlawful data collection on Palestinians, and facilitates expansion of Israel's illegal settlements on Palestinian land." Fight for the Future director Evan Greer said in a statement that "Israel's military has contracted Amazon and Google to build and fuel the technology used to oppress, occupy, and bomb Palestinians. The services Amazon and Google provide and the technology they build [power] drones, surveillance, and sophisticated weaponry for the Israeli military." "Make no mistake, if we don't get Amazon and Google to cut their ties with Israel, the stage will be set for them to become the backbone of the 21st-century military-industrial complex," she continued. "And developments made in powering Israel's war machine will be exported to militaries and police departments across the world, including America." "Make no mistake, if we don't get Amazon and Google to cut their ties with Israel, the stage will be set for them to become the backbone of the 21st-century military-industrial complex." "We're at a precipice with lives at stake," Greer added. "Nothing short of Amazon and Google ending the Project Nimbus contract is acceptable." Olivia Katbi Smith, North America coordinator for the BDS Movement, said that "Amazon and Google are helping to sustain Israeli apartheid in its repression of the Palestinian people with a massive cloud contract that will enable increased surveillance, discrimination, and displacement." "Mass as well as targeted surveillance of disenfranchised Indigenous Palestinians, who are denied basic rights and recourse to justice, is a core feature of Israel's system of repression, oppression, and colonial dispossession," she added. "This repression, which is often tested on Palestinians before being exported globally, must be challenged by us all, together." JVP executive director Stefanie Fox said that "courageous workers at Google and Amazon are calling on their employers to stop enabling the Israeli government's oppression of Palestinian families--and we are proud to stand with them." "Google and Amazon executives should listen to their employees, pull out of the Nimbus contract, and cut all ties with the Israeli military," she added. "No tech for apartheid!" Our work is licensed under Creative Commons (CC BY-NC-ND 3.0). Feel free to republish and share widely. SIGN UP FOR OUR NEWSLETTER Quality journalism. Progressive values. Direct to your inbox. Follow Us

Implement struct no copy in Golang

Today's Question: How much time do you code every day? GIVE A SHOUT Technical Article => Programming => Go Implement struct no copy in GoLang sonic0002 2020-09-04 22:24:58 18,101 0 There is some case where some of the struct in GoLang which is not meant to be copied. For example, sometimes a global configuration which should have only one version passed around the whole application and should not be copied and modified. In GoLang, there is no intuitive solution on preventing copying of struct. But there is still some way which can be leveraged to help prevent this while developing the code. The trick is to define some struct implementing sync.Locker interface and has this struct embedded in any struct not meant to be copied. // noCopy may be embedded into structs which must not be copied // after the first use. // // See https://golang.org/issues/8005#issuecomment-190753527 // for details. type noCopy struct{} // Lock is a no-op used by -copylocks checker from `go vet`. func (*noCopy) Lock() {} func (*noCopy) UnLock() {} Below is a sample snippet explaining above. package main import ( "fmt" ) type noCopy struct{} func (*noCopy) Lock() {} func (*noCopy) Unlock() {} type Demo struct { noCopy noCopy } func Copy(d Demo) { } func main() { d := Demo{} fmt.Printf("%+v", d) Copy(d) fmt.Printf("%+v", d) } Now if you run go vet on the main.go file, you will see some errors telling you that your are having some place copying the struct(Copy()). > go vet main.go # command-line-arguments .\main.go:16: Copy passes lock by value: main.Demo contains main.noCopy .\main.go:21: call of fmt.Printf copies lock value: main.Demo contains main.noCopy .\main.go:23: call of Copy copies lock value: main.Demo contains main.noCopy .\main.go:25: call of fmt.Printf copies lock value: main.Demo contains main.noCopy One thing to note here is above approach will not tell compiler to fail the compilation. If you proceed to compile and run it, it would still work. This just provides another hint to go vet which normally we would run before we build the project to check there might be potential issue in our source code. > go run main.go {noCopy:{}}{noCopy:{}} Reference: https://github.com/golang/go/issues/8005#issuecomment-190753527 p   p   p RELATED 0 COMMENT

John Deere takes a step into the future with first fully autonomous tractor

John Deere revealed its first fully autonomous tractor today at CES 2022, building off of what it says are 20 years of AI development – and it’s not a concept. The self-driving John Deere is already up and running on select farms and available for sale. Built on John Deere’s 8R tractor, the new autonomous product features a TruSet-enabled chisel plow, GPS guidance system, and other advanced technologies, including six pairs of stereo cameras that enable 360-degree obstacle detection and the calculation of distance. Images captured by the cameras are passed through a deep neural network that classifies each pixel in approximately 100 milliseconds and determines if the machine continues to move or stops, depending on if an obstacle is detected. What’s more, John Deere executives spoke of that same neural network being able to differentiate between “weeds” and “crops” on the fly, selecting certain plants for removal while leaving others behind to grow. (Note: I used quotes on “weeds” because a rose is a weed in a cornfield, you know?) The stated goal of the new John Deere tractor is to help farmers “do more with less.” At the press conference, company reps discussed how future farmers will feed a growing global population with fewer farm workers and less available farm land while, at the same time, reducing emissions. This tractor, they say, is a product that will do just that. John Deere has been buying up a number of tech companies in recent years, from the acquisition of AI startup Blue River back in 2017 and to the purchase of EV battery maker Kreisler Electric late last year. Those acquisitions helped the tractor company develop an initial autonomous prototype that became the production model, first put into service on Minnesota farmer Doug Nimz’ farm a few years ago. “It takes a while to get comfortable because … first of all, you’re just kind of amazed just watching it,” Nimz told c|net. Nimz described himself as, “very, very interested” but also a “little suspicious” of autonomous technology before using John Deere’s machine on his farm. “When I actually saw it drive … I said, ‘Well, goll [sic], this is really going to happen. This really will work.'” This tractor is “[a] way to get the job done on time, every time and do it at a high level of quality,” Jahmy Hindman, Chief Technology Officer for Deere & Co., explained. “It’s 20 years in the making.” And, as several company reps pointed out in the press conference, it’s a hedge against losses incurred from labor shortages (read: farmers don’t have to pay people anymore). The average farmer in the US is over 55 years old, and John Deere notes that fewer and fewer people are choosing to enter the agriculture business, while demand continues to grow. As John Deere continues to push its autonomous vehicle programs, its dependence on bigger, more powerful electrical systems to power those systems will grow as well. Indeed, the push toward automation is cited as one of the driving forces behind the industry’s appetite for electrification, too. As recently as 2020, Ford told Wired that “testing shows that more than 50% of a battery-electric vehicle’s range would be sucked up by the computing power demanded by self-driving software, plus the air conditioning and entertainment systems needed to keep passengers comfortable.” In a tractor? On a farm? Speed isn’t of the essence, torque is, and an electric tractor can not only keep working longer at its most efficient low rpm range, but it also has a relatively short distance to cover to get back to its “home” and charge or swap out its batteries. It’s not about miles at that point, it’s about hours – and that’s something John Deere’s latest acquisition knows a lot about. This launch is a statement of intent from John Deere. They’re going autonomous now, and they’re doing it on the backs of advanced batteries that they’ve paid good money not to share. Expect good things if you spend the $500,000+ to order one. In the meantime, you can watch the complete John Deere CES 2022 press conference for yourself, below, and let us know what you think of John Deere’s autonomous tractor in the comments. strong: c|net; John Deere, via PR Newswire. Add Electrek to your Google News feed. Stay up to date with the latest content by subscribing to Electrek on Google News. You’re reading Electrek— experts who break news about Tesla, electric vehicles, and green energy, day after day. Be sure to check out our homepage for all the latest news, and follow Electrek on Twitter, Facebook, and LinkedIn to stay in the loop. Don’t know where to start? Check out our YouTube channel for the latest reviews.

Peter+James

He’ll give his heart to the first woman who does this... His Secret Obsession Review Introduction: There are thousands of books on the market about relationships and marriage. Some will tell you how to ‘save your marriage’ while others will tell you ‘how to understand your man’. In most cases, the women who read these books and try to apply the advice fail miserably. Just looking at the divorce statistics is enough to make you realize that the information that’s so readily available out there is just not working. There is good reason for this. The books that are traditionally published have a need to sugar-coat the truths and be politically correct. With feminism threatening to emasculate men and treat them like women, dispensing any advice to the contrary will incur the wrath of the perpetual moaners and complainers that abound in society these days. No one wants to accept the fact that men and women are different. However, one self-published book, His Secret Obsession has taken the online world by storm. Written by author, James Bauer, it’s been a bestseller for years and truly exposes how a man thinks and why he acts the way he does. This is excellent relationship advice for women who wonder why their partner doesn’t seem interested in them anymore. It shows several techniques that a woman can employ to make her husband desire her all over again. It can put the sizzle in a marriage that’s about to fizzle out. It can turn a single woman who has no luck with men into a vixen that most men desire. It’s all about getting into his head… and this book will show you exactly how to do that. Let’s analyze the pros and cons of His Secret Obsession. This book has been an online bestseller for years with thousands of copies sold. There are many satisfied customers and lots of social proof that the advice with the book works wonders. It’s a proven product. What’s interesting about His Secret Obsession is that it does NOT try to be politically correct. It truly exposes how a man thinks and what you need to do to hook your man and keep him attracted to you… and it’s really not what you think. Many women feel like they’re doing so much for their man and he should be appreciative… but James Bauer’s book reveals that it’s not how much you do, but what you do that matters. This book is honest and dispenses highly effective advice. His Secret Obsession is backed by a 60-day money-back guarantee. If you feel like the book is not your cup of tea, you can always ask for a refund. The advice in this guide is for both married and single women. The married woman who wants to save her marriage or make her husband desire her again will find all the techniques she needs inside. The single woman who is dating and wondering why the men she goes out with never seem to call her back even when the date was seemingly fine will discover the answer in here too. The mantra of this book is that men and women are different. This is refreshing advice and it goes against the grain. Most of the relationship advice dispensed by supposed relationship experts in women’s magazines are absolute hogwash because they try to treat the man like a woman and it backfires. James Bauer’s approach is much more effective because it understands the man’s psyche and shows the woman how to get into his mind and make him desire her. You’ll NEVER find such truthful advice in a woman’s magazine or the usual relationship books that are on the shelves. You can only purchase this product online and will need a credit card. The book is delivered in digital format. So, you’ll either have to read it on your computer or you’ll have to print it out for easier reading. Some women may feel hesitant applying the information in this book because they feel like ‘it’s not them’. This is understandable… but it’s necessary to step out of your comfort zone if you wish to effect positive change in your life. Overcoming this mental hurdle is crucial. Consistency is required. You’ll need to follow the guide and be consistent in your efforts.

Internal Google document reveals campaign against EU lawmakers

Internal Google document reveals campaign against EU lawmakers Gain a global perspective on the US and go beyond with curated news and analysis from 600 journalists in 50+ countries covering politics, business, innovation, trends and more. Subscribe to unlock this article Try unlimited access Try full digital access and see why over 1 million readers subscribe to the FT Only$1 for 4 weeks Then $69 per month New customers only Cancel anytime during your trial Then $69 per month New customers only Cancel anytime during your trial What is included in my trial? During your trial you will have complete digital access to FT.com with everything in both of our Standard Digital and Premium Digital packages. Standard Digital includes access to a wealth of global news, analysis and expert opinion. Premium Digital includes access to our premier business column, Lex, as well as 15 curated newsletters covering key business themes with original, in-depth reporting. For a full comparison of Standard and Premium Digital, click here. Change the plan you will roll onto at any time during your trial by visiting the “Settings & Account” section. What happens at the end of my trial? If you do nothing, you will be auto-enrolled in our premium digital monthly subscription plan and retain complete access for $69 per month. For cost savings, you can change your plan at any time online in the “Settings & Account” section. If you’d like to retain your premium access and save 20%, you can opt to pay annually at the end of the trial. You may also opt to downgrade to Standard Digital, a robust journalistic offering that fulfils many user’s needs. Compare Standard and Premium Digital here. Any changes made can be done at any time and will become effective at the end of the trial period, allowing you to retain full access for 4 weeks, even if you downgrade or cancel. When can I cancel? You may change or cancel your subscription or trial at any time online. Simply log into Settings & Account and select "Cancel" on the right-hand side. You can still enjoy your subscription until the end of your current billing period. What forms of payment can I use? We support credit card, debit card and PayPal payments. Read more Explore our subscriptions Individual Find the plan that suits you best. Group Premium access for businesses and educational institutions. Check if your university or organisation offers FT membership to read for free.

Stripe Cross-border payouts are now available in beta

Transfer and pay out funds around the world. Cross-border payouts enable you to pay sellers, freelancers, content creators, and service providers in their local currencies. You can transfer funds to connected accounts in other countries with your existing platform account and charge configuration. Cross-border payouts enable platforms in the US to pay out to Express and Custom connected accounts in the following countries: Albania Algeriap Angolap Antigua & Barbuda Argentina Armenia Australia Austria Azerbaijanp Bahamasp Bahrain Bangladeshp Belgium Beninp Bhutanp Bolivia Bosnia & Herzegovina Botswanap Bruneip Bulgaria Cambodia Canada Chile Colombia Costa Rica Côte d’Ivoire Croatia* Cyprus Czech Republic* Denmark Dominican Republic Ecuador Egypt El Salvador Estonia Ethiopia Finland France Gabonp Gambia Germany Ghana Greece Guatemala Guyana Hong Kong Hungary Iceland* India Indonesia Ireland Israel Italy Jamaica Japan Jordan Kenya Kuwait Laosp Latvia Liechtenstein* Lithuania Luxembourg Macao SAR China Madagascar Malaysia Malta Mauritius Mexico Moldova Monacop Mongolia Morocco Mozambiquep Namibia Netherlands New Zealand Nigerp Nigeria North Macedonia Norway Oman Panama Paraguay Peru Philippines Poland Portugal Qatar Romania Rwanda San Marinop Saudi Arabia Senegal Serbia Singapore Slovakia Slovenia South Africa South Korea Spain Sri Lanka St. Lucia Sweden Switzerland* Taiwanp Tanzania Thailand Trinidad & Tobago Tunisia Turkey United Arab Emirates United Kingdom Uruguay Uzbekistan Vietnam * Bank accounts in countries with an asterisk (*) can only receive Euro (EUR) payouts. Note Stripe might pause payouts to countries in the preview program while any issues are resolved. This can happen without advance notice to you as the owner of the platform or the owners of your connected accounts. Note Banks in some countries have special requirements for payments received from outside their country’s borders. These countries might also have higher minimum payout amounts . We also support Crypto payouts to select countries. To view additional supported payout countries, view our Crypto Payouts docs . Cross border-specifications Cross-border payouts only work with the recipient service agreement . The onboarding specifications for cross-border payouts vary by destination country. To learn more, see: Required verification information Supported settlement currencies Bank account formats Get started Existing Connect platforms - Select the recipient service agreement at account creation. New Connect platforms - Follow one of the guides below that best fits your use case, and make sure to specify the recipient service agreement in the account creation step. Collect payments and then pay out Pay out money

What among us teaches us about byzantine failure

Besides Among Us [1] facing devops challenges as its users outgrow its server capacity, the concept of the game itself has its roots in a problem that is famously difficult to solve in the world of distributed systems. Those who have played the game before can probably skip to the next paragraph. The basic concept of the game is that you and your friends are in outer space somewhere, and are confined to a fixed map of either a base station or a spacecraft. You each have a series of tasks to complete, with the exception of the impostor, who is pretending to perform tasks. The impostor is perceived as one of your crewmates, but is actually doing their best to murder everyone on the ship without the others catching on to who it is. If there are enough players, there may be more than one imposter. This may sound this level of betrayal would damage relationships with your friends, but studies have shown [2] that games where the other players are opposition rather than working together breeds the opposite of togetherness, anyway, so Catan is probably doing just as much damage. The game is ended when your crewmates guess correctly who the impostor is during a meeting and launch them into space. A meeting can be called either when a deceased crewmember is found, or may be called a fixed number of times per crewmate (including the impostor) per game. This is the only time that the players may speak during the game, and a fixed amount of time is given to decide on a democratic majority basis. The crewmembers can also win by completing all the tasks before the impostor murders everyone. Murdered crew members can still complete tasks, but cannot communicate during meetings. Conversely, the impostor wins if every crewmember but one is dead. The real key to the game is that, during a meeting, the crewmembers (who aren’t dead) can democratically (majority) choose the member who’s the impostor. If they choose the wrong person, then the wrong person has been ejected from the game, and the impostor is that much closer to winning. This introduces the first concept that the game calls on from computer science, which is consensus. Consensus is a problem introduced when one or more components are located on a different entity, such as a computer, and those entities must communicate to accomplish a common goal [3]. To be classified as a consensus problem, entities must put forward their votes, communicate with one another, and eventually agree on a common solution. This is a common problem in distributed systems, where the components are located on different computers [4]. This problem is also found in multi-agent systems, where more than one entity is required to solve the problem due to the size or complexity of the data. In the case of Among Us, the entities are humans, and it can be argued that the humans are acting as a distributed system by passing “the truth” of what they’ve witnessed around as the messages from one “node” (person) to another. The person who saw the deceased crew member knows it was in a certain room, and another might have seen another crew member coming from that general direction on the map. The nodes have to communicate together to join their components into one coherent story for what occurred during the current portion of the game. The imposter may lie if they are at all good at the job, and this touches on the fundamental problem of unreliable data or state within consensus; if all entities were truthful, this would be a fairly straight-forward problem. In order to address this problem, systems are built with fault tolerance or resilience. Fault tolerance allows a system to continue operating with the degradation of one or more components [5]. In the case of Among Us, this would be if a crew member was mistaken for the impostor, and unceremoniously ejected. The system is somewhat fault tolerant in that, as long as there are more than three crew members at the time of voting, this mistake may rattle confidence in the non-impostors, but won’t end the game. This could be referred to as graceful degradation. One mistake won’t likely bring the entire system down, but your dead crew member can no longer communicate with the crew what they’ve seen, or will see as a ghost. To reach consensus, it’s an acceptable method to require all entities to agree on a majority value, rather than requiring every entity to produce the same value. This method may be adversely affected by an incorrect value (the impostor’s vote). However, majority vote circumnavigates the problem of the inability to move on because of the Byzantine failure inherent to this game. This type of failure occurs when the symptoms may look different to different observers, and may send contradictory data down the line [6]. It’s interesting to note here that in real-life systems, it may be a fault of the electrical components, and isn’t a problem isolated to human bad-actors. But we’re all just electricity in meat suits, so same difference. An example of a Byzantine failure in Among Us is in the case that there are enough players to have two impostors at one time, it’s a worthwhile strategy to consider voting the other impostor off the ship if it means you won’t look sus. An impostor may also call a meeting directly after killing someone and accuse others. Or, in a notable case, you may choose not to debate that you are the impostor, and tell the crew that they may vote you off if they would like with no other defense. For the sake of knowing, the other failure type of consensus is a crash failure. This is where the process stops and does not resume on fail, also called a fail-stop failure. A system that can handle Byzantine failures are called Byzantium fault tolerant. The idea of developing a Byzantine failure-based consensus was first formalized for a NASA-funded project out of SRI International in 1978 [7]. The goal of this project was to achieve fault tolerance in a system thats failure would result in loss of life — an airplane. The first investigator, who dubbed it an interactive consistency problem, had several computers communicating through pairwise messaging that were plagued by a misbehaving processor. The interactive consistency problem was renamed the Byzantium Generals Problem after a suggestion of calling it the Albanian Generals Problem [8] was understood to perhaps end in future cancellation. A later investigation in 1980 [9] outlined a proof that a system that may overcome Byzantine failure must have at least 3 times the number of non-faulty processors to faulty if signed signatures are used. You may read this as that there needs to be three times as many true crew members as impostors in our game to have a hope of winning. As an aside, the concepts of dependability and fault tolerance were hot topics during that time (1980s), and there was a whole technical committee within the IEEE (Institute of Electrical and Electronics Engineers) called the Dependable Computing and Fault-Tolerance and IFIP Working Group 10.4 on Dependable Computing and Fault Tolerance [10]. Just a very interesting tidbit. To handle a consensus problem, protocols are introduced. To be considered a protocol, the algorithm must eventually decide a value for a process. The protocol also must have integrity, or validity, which is the assertion that if a correct process decides a value, then this means a correct process must have suggested the value. The integrity may be on a spectrum of strong to weak. The integrity may be increased in a Byzantine failure type by making this value more stringent. The third requirement for a protocol is that all correct processes will agree on a single consensus. Without these attributes, the protocol breaks down, and the imposter(s) win. The first of the modern (1999) algorithms that attempted to solve the problem is called the Practical Byzantine Fault Tolerance algorithm [11]. A more recent algorithm (2018), building off of the PBFT algorithm, is the Tendermint BFT [12], which Proof of Stake (e.g. Etherium) blockchains use today due to its handling of partially asynchronous networks. Attempting to use one of the original protocols [8] proposed as “A Solution with Oral Messages” in 1982 in your next Among Us game might be fun. The protocol is summarized as follows : Several rounds of game are required to reach the solution. In the paper, both the team member and the person who called the meeting (the lieutenant and general) are assumed to potentially be the “traitor”. However, you may see a glaring hole in this plan: several rounds of this may be difficult to apply here, as the answer given to the question of “Where were you in relation to the dead team member?” is different each time; it’s nearly impossible to replicate the previous rounds’ result to see if the answers are different for each round. As it is, the paper states “…it is the traitors’ ability to lie that makes the Byantine Generals Problem so difficult.". The paper goes on to state that the problem becomes simpler with the introduction of signed messages. This is also not a viable solution in this particular usecase. The problem with designing a protocol for this particular game lies in that when voting, it’s nearly impossible to prove validity through proof of work or proof of stake, as a solution to the original paper’s suggested signed messages. Proof of stake and proof of work are both used today in applications such as cryptocurrency, as noted above. A solution that may have some bearing in Among us was proposed by a paper titled “An Efficient Algorithm for Byzantine Agreement without Authentication” [13]. Handy, right, since we have no means for authentication? It seems to add to the previous work done by adding in the idea of Low and High thresholds. The basic summary is that the low threshold of one correct process must support an assertion (number of impostors plus one), and correct processes that support the low threshold will support the high threshold (2*number of impostors + 1). The idea being here that incorrect data won’t skew results, if you ask for confirmation of a fact before it’s even considered for the high processes and asking for agreement from the rest of the processes. The Low (sus) processes won’t be able to add false claims without support from one correct process. This algorithm also ignores assertions critical to decision making that are too close to the end of the algorithm to be communicated effectively with all processes — meaning here that the impostor could not throw everyone off by giving information 5 second before the deadline to vote in the game. The protocol for this algorithm is as follows Well, hold on, step 4 still doesn’t really jive with how many rounds there are likely to be - most games end before the minimum of the fifth meeting. A strategy may be to have as many meetings as possible in a short amount of time, but overall, a useful thing to take away from the paper is not letting incorrect data be casually added by an impostor without analysis. There’s a reason why Byzantine Failure is considered one of the hardest problems in distributed systems. If anyone has an effective algorithm that they’ve applied to their Among Us games, I would love to hear it! The tweet that originally sparked my curiosity on the topic used a heartbeat protocol [14], which seems to violate the rules of not speaking between each round, but is still a very cool idea. It was implemented by calling out to other players every so often, and if a crewmember didn’t respond, they has more clues about who a likely impostor was because they knew that crew member was dead without calling a meeting [15]. Lastly, I just want to stick this white paper here on the topic of possible scenarios where it might be impossible to solve the Byzantium Generals Problem [16], because although I didn’t have a place for it in this article, it was a very fun read about how this particular problem is hard. [1] http://www.innersloth.com/gameAmongUs.php [2] https://www.alieward.com/ologies/ludology [3] https://en.wikipedia.org/wiki/Consensus_(computer_science) [4] https://en.wikipedia.org/wiki/Distributed_computing [5] https://en.wikipedia.org/wiki/Fault_tolerance [6] https://en.wikipedia.org/wiki/Byzantine_fault [7] https://www.microsoft.com/en-us/research/uploads/prod/2016/12/Design-and-Analysis-of-a-Fault-Tolerant-Computer-for-Aircraft-Control.pdf [8] https://lamport.azurewebsites.net/pubs/byz.pdf [9] https://lamport.azurewebsites.net/pubs/reaching.pdf [10] https://web.archive.org/web/20150402141319/http://www.dependability.org/ [11] http://www.pmg.lcs.mit.edu/papers/osdi99.pdf [12] https://docs.tendermint.com/master/introduction/what-is-tendermint.html [13] https://www.cs.huji.ac.il/~dolev/pubs/efficient-ba-no-auth.pdf [14] https://martinfowler.com/articles/patterns-of-distributed-systems/heartbeat.html [15] https://twitter.com/_tessr/status/1318636680895778818 [16] https://groups.csail.mit.edu/tds/papers/Lynch/jacm85.pdf

Swiping right on jabs: dating app adds Covid vaccine badges in Australia

Single Australians looking for sex or love will soon stand out from the competition if they’ve been vaccinated against Covid-19. Dating app Bumble announced it will roll out a new feature this week enabling users in Australia and New Zealand to add a “vaccination badge” to their profile if they’ve received a Covid-19 jab. Competitor Tinder says it looks forward to “making [vaccine] badges available soon” in Australia too. A Bumble representative says it will not independently verify the vaccination status of those who claim the badge, meaning the system will rely on user honesty. The dating app says it decided to launch the feature after a recent survey revealed a “45% increase in users asking potential dates if they had the vaccine or have Covid symptoms”. Without a ‘single bubble’ during lockdown, NSW are switching off an energy source I need to survive | Melanie Tait It’s not just in Australia that dating apps are encouraging users to boast about their vaccination status. In the US, the White House recently partnered with Tinder, Hinge and OKCupid to give away premium features such as free “super likes” – which increase your chance of being noticed by a potential date – to users who set their vaccination status. Last month the UK’s Department of Health and Social Care launched a similar initiative by teaming up with dating apps to offer users a range of perks, including free profile “boosts” if they add their vaccination status to their profile. Both schemes are aimed at promoting vaccine uptake in young people. Currently, just 11% of Australians are fully vaccinated against Covid-19. In late June, all Australian adults under 40 were told by prime minister Scott Morrison they could get the AstraZeneca vaccine via their GPs. However the advice from the expert Australian Technical Advisory Group on Immunisation is that the Pfizer jab is the preferred option for under-60s – and that vaccine remains in limited supply. Australian online dating website RSVP says vaccine badges are something the company has been considering, but they will wait until more Australians have been immunised before bringing the feature in. “The reason we haven’t pushed anything ahead is just looking at the amount of people that have access to vaccines,” CEO Dave Heysen explains. “So we don’t really want to discriminate when people don’t actually have access [to the vaccine] at this stage. “We’ve thought about adding it in at some stage; we just want to make sure everyone’s on a level playing ground first.” The addition of vaccine badges could make users feel comfortable about meeting up with dates in real life, says Dr Rosalie Gillett, a postdoctoral research fellow at the Queensland University of Technology, who has researched user safety on dating apps. “As dating apps often encourage in-person interactions, users might feel safer meeting with people who say they have been vaccinated against Covid-19,” Gillett says. Joanne Orlando, the author of Life Mode On, a book about navigating the digital world, agrees vaccination badges are “definitely a plus” when it comes to Covid safety. But the feature could also have another benefit: helping users match with those whose values align with their own. Covid Australia vaccine rollout tracker: total number of people and per cent vaccinated, daily vaccine doses and rate of progress “Vaccination badges give you a little bit more insight into the profile and person,” she says. “It gives you an idea of how they’re thinking about contributing to society, doing their part, and if [their vaccination status] is something they want everyone to know about and are proud about. It’s just another element to help you to consider whether or not this person is for you.” Does that mean the badges could help socially conscious singles avoid matching with anti-vaxxers? “That’s right, or if you are one, you’re not going to go for someone who’s got that stamp.” Orlando says. “When it comes to online dating, there’s such a limited amount of info straight up, so this kind of thing could really help someone. “It just adds another dimension that we don’t normally get on dating app profiles.”

The most important API metric is time to first call

API publishers among Postman’s community of more than 15 million are working toward more seamless and integrated developer experiences for their APIs. Distilled from hundreds of one-on-one discussions, I recently shared a study on increasing adoption of an API with a public workspace in Postman. One of the biggest reasons to use a public workspace is to enhance developer onboarding with a faster time to first call (TTFC), the most important metric you’ll need for a public API. If you are not investing in TTFC as your most important API metric, you are limiting the size of your potential developer base throughout your remaining adoption funnel. To understand a developer’s journey, let’s first take a look at factors influencing how much time and energy they are willing to invest in learning your technology and making it work. Urgency: Is the developer actively searching for a solution to an existing problem? Or did they hear about your technology in passing and have a mild curiosity? Constraints: Is the developer trying to meet a deadline? Or do they have unlimited time and budget to explore the possibilities? Alternatives: Is the developer required by their organization to use this solution? Or are they choosing from many providers and considering other ways to solve their problem? Developer journey to an API With that context in mind, the following stages describe the developer journey of encountering a new API: Step 1: Browse A developer browses your website and documentation to figure out what your API offers. Some people gloss over this step, preferring to learn what your tech offers interactively in the next steps. But judgments are formed at this very early stage, likely while comparing your product among alternatives. For example, if your documentation and onboarding process appears comparatively unorganized and riddled with errors, perhaps it is a reflection of your technology. Step 2: Signup Signing up for an account is a developer’s first commitment. It signals their intent to do something with your API. Frequently going hand-in-hand with the next step, signing up is required to generate an API key. Step 3: First API call Making the first API call is the first payoff a developer receives and is oftentimes when developers begin more deeply understanding how the API fits into their world. Stripe and Algolia embed interactive guides within their developer documentation to enable first API calls. Stripe and Twitter also use Postman public workspaces for interactive onboarding. Since many developers already use Postman, experiencing an API in familiar territory gets them one step closer to implementation.

Graham Greene Against the World

The last novelist who acted like he might save the world may have been Graham Greene. He belonged to a generation of writers who might not always share the same political opinions but who supported many of the same causes: defending jailed dissidents, protesting illegal wars, and challenging the unfairness (or even stupidity) of censoring great books. He wrote a novel, The Comedians, and developed its film adaptation with the intention of helping to “isolate” Haiti’s Papa Doc Duvalier, who contemplated having Greene assassinated in retaliation. At one point, Greene was so celebrated that the South African State Department asked him to negotiate the release of a kidnapped ambassador in El Salvador. (Greene eventually came to an agreement with the rebels, but the ambassador was killed anyway—for reasons that were never fully understood.) By the end of his life, he was known as an international public figure who partied with Capote, Yoko Ono, and Kissinger, as much as a writer of entertaining, always absorbing novels, short stories, screenplays, and essays. The Unquiet Englishman: A Life of Graham Greene Buy on Bookshop In the final decades of his life, Greene defied the Cold War logic of his times by befriending politicians and intellectuals from across the political spectrum—Fidel Castro in Cuba, Omar Torrijos in Panama, Jorge Luis Borges in Argentina, R.K. Narayan in India, Vaclav Havel and Josef Skvorecky in Czechoslovakia, and Pablo Neruda, when he was acting as Chile’s ambassador to France. In the early 1950s, Greene wrote open letters of support for Charlie Chaplin, who had been targeted by McCarthy; and he publicly confessed his brief Oxford-era membership in the Communist Party to challenge U.S. immigration over the McCarran Act. All over the world, people wanted to know what Greene was doing, even when they weren’t reading him. When Kim Philby was reviled as a mole for the USSR at MI6 (where he had been Greene’s drinking buddy and supervisor), Greene was the only notable British figure to visit the disgraced Philby in Russia, or attend his funeral in 1988. As Beatrice Severn declares near the end of Greene’s funniest novel, Our Man in Havana (1958): “I don’t care a damn about men who are loyal to the people who pay them, to organizations.… Would the world be in the mess it is if we were loyal to love and not to countries?” It’s hard to think of any similarly productive, commercial novelist today who speaks so vigorously against religious and political pieties. Richard Greene’s book The Unquiet Englishman is the third major biography of Graham Greene in 40 years and provides the most readable, balanced approach so far to both a complicated life and an intensely enjoyable body of work; it makes use of newly available letters, diaries and recollections concerning Greene and his closest friends. It is neither as excessively detailed as the Norman Sherry biography released in three volumes between 1989 and 2005, nor as combative as Michael Shelden’s 1994 portrait, Graham Greene: The Man Within . (Shelden was unforgiving about Greene’s loyalty to old friends such as Philby and the novelist Norman Douglas, who was charged with indecent assault.) The Greene who emerges here rarely stayed in one place for very long and was continually dissatisfied with the world that he witnessed changing convulsively around him. Born in 1904 at the Berkhamsted boarding school where his father was house master, Greene quickly grew accustomed to living secret lives, especially through the books that absorbed him, such as John Buchan’s spy stories (most famously, The Thirty-Nine Steps ), and while exploring the lost kingdoms of H. Rider Haggard ( King Solomon’s Mines and She ). Then, almost inevitably, considering the pathologies of Edwardian-era boarding schools, Greene was soon being bullied, tormented, and co-opted by his fellow schoolboys; and he later compared the cruel “finesse” of his tormentors with the major figures in geopolitical conflicts. They made him feel as if he was always betraying someone, especially his father. He grew depressed and withdrawn, and was diagnosed as manic-depressive. Perhaps to challenge his deep sense of failure, he indulged in suicidal games, and claimed to have played Russian roulette several times with a gun from the family cupboard; he even compared the “minute click” of the chamber against his ear with “a young man’s first successful experience of sex” (though later revised the story so many times that it may not have been entirely true). Even if he hadn’t come close to suicide with a gun, he was soon relentlessly and unambiguously acting out suicidal impulses through travel. Throughout his life, he set out for treacherous and dangerous places—to investigate Firestone and the British government’s involvement with forced labor in Liberia (where he nearly died of fever) to leper colonies in the Congo Basin to research A Burnt-Out Case, to Haiti in the midst of political terror, to central America, and to some of the most troubled, anti-Catholic regions of Mexico. He was quick to make new contacts with what his own government might consider unsavory people, such as members of Sinn Fein and the Free State in Northern Ireland, and to take on risky tasks, like ferrying communications between anti-Fascist groups in Germany and France. Even before he turned 20, Greene was a tough man to pin down—either on a map or as a person. The notion of “identity” was never comfortable for Greene, and whenever he sensed the shadow of a firm identity descending upon him, he went looking for the nearest way out of it. ( Ways of Escape was, appropriately, the title of his second autobiographical memoir, in 1980.) He was still an atheist when, at 20, he fell in love with a young poet and devout Catholic, Vivienne Dayrell-Browning; he subsequently accompanied her to his first Catholic service, deluged her with multitudinous letters of affection (everything about Greene was multitudinous), and, in order to marry her, converted. Although he spent the rest of his life known as a “Catholic” writer, he wore that designation loosely. He frequented brothels; he soon left his young wife and children to live in London with Dorothy Glover, who illustrated several of Greene’s books for children; and then he embarked on a long, complicated series of affairs—most notably with Catherine Walston, a married woman who provided perhaps the most both intensely happy and deeply unhappy relationship of his life. (He was unhappy when she was not with him and desperate to see other women when she was.) Both Catherine’s personality and her open marriage provided models for the relationships in one of Greene’s greatest novels, The End of the Affair, in 1951. He often maintained significant relationships with more than one woman at a time; he drank heavily and continuously and moved easily from one profession to another—as a journalist for Th e Times of London, as a film and book critic for The Spectator, as an editor at both Eyre & Spottiswoode and Bodley Head, and as a novelist producing roughly a novel every year or two. (He insisted on sitting down every day and writing at least 500 words of fiction, which he kept track of in a diary.) In 1941, through his sister, he procured a job in Sierra Leone mining intelligence for MI6, with the code name Agent 59200 (the same as Wormold’s in Our Man in Havana). When that profession grew too boring and predictable, he went to Hollywood and wrote movies. The best part about being a novelist, Greene once told a friend’s teenage son at a party, “is that you can spy on people. Everything is useful to a writer, you see, every scrap.” And as his friend and fellow spy-novelist John le Carré recalled, Greene always felt a “sense of alienation, of being an observer within society rather than a member of it.” Constantly roving, Greene wrote successful novels, short stories, screenplays, travel books, and plays from the points of view of everyone he met or could imagine: a Jewish businessman selling currant sweets; a lesbian journalist and a chorus girl in Stamboul Train; a professional killer with a harelip in This Gun for Hire; and a teenage Catholic psycho-gangster named Pinkie in Brighton Rock. And then, of course, there are the countless spies and their accomplices, such as the clumsy and deeply loving Wormold in Our Man in Havana, who signs on with MI6 and snaps blurry photos of the insides of vacuum cleaners to convince his London bosses that he has discovered a sophisticated new weapon system. Every Greene spy is an exercise in love—whether it is Wormold’s devotion to his spoiled (and awful) daughter, Milly, or Maurice Castle’s passion for his Black, South African wife in what may be Greene’s best (and certainly my favorite) novel, The Human Factor (1978). According to the morals of Greene’s fictional universe, you can betray your country—easy-peasy. But if you betray the people you love, you deserve the worst that can happen to you. Even the titles of Greene’s books testify to an almost unendurable subjective human struggle: The Man Within, It’s a Battlefield, Journey Without Maps, The Lawless Roads, Monsieur Quixote, A Burnt-Out Case, and so forth. Greene saw geographical spaces in terms of spiritual struggle; and while he often dramatized the horrors of political and military conflicts, his primary novelistic aim was to document the deeper conflicts occurring within the subjective life of his characters. After many years of increasing commercial success as a novelist, Greene’s “breakthrough” book, The Heart of the Matter (1948), achieved international fame by appealing to Catholics who found in it an expression of living in a post–World War II world where God didn’t seem to reward the faith they placed in Him. Despite his appeal to Catholics, Richard Greene (who is no relation to the novelist) writes, Graham Greene was more Manichaean than Christian: “[H]e could not take the idea of a devil very seriously, and if there was evil in the world God must be in some sense answerable for it.” If there was a God, he believed, then it was a God who “had placed human beings in unbearable circumstances.” To a large extent, Greene’s cynicism about the world—and the sufferings of humanity—came close to nihilism, or even a form of Ligotti-like anti-natal philosophy. As Scobie tells himself in one of The Heart of the Matter’s most troubling scenes: “Point me out the happy man and I will point you out either extreme egotism, evil—or else an absolute ignorance.” Life is awful, Greene often said; but when it came to art—such as his own daily production of beautiful paragraphs and scenes throughout most of his life—well, at least art could make it seem worthwhile. The least interesting character in Greene’s substantial literary output is Greene himself; for while even his most craven, fault-ridden characters—such as the unnamed “whisky priest” in The Power and the Glory —always drive compelling stories, the Greene persona in his travel books and autobiographies floats passively on the current of events without much urgency. On his deathbed, when the partner of his late years, Yvonne Cloetta, asked him if he wanted to call in his priest for last rites, Greene reportedly replied: “You decide.” They were the most appropriate last words Greene could have uttered. Nobody saw politics and power with greater human clarity than Greene; and it’s hard to read a single paragraph of his tight, controlled, angry, and always alert prose without underlining passages to reread again later. In The Lawless Roads, he wrote, “After Mexico,” where he traveled in the late thirties “I shall always associate balconies and politicians—plump men with blue chins wearing soft hats and guns on their hips. They look down from the official balcony in every city all day long with nothing to do but stare, with the expression of men keeping an eye on a good thing.” Greene saw things that readers hadn’t yet seen for themselves; but once Greene showed them the things they should see, readers couldn’t stop seeing them. Greene famously divided his works into “entertainments” and “serious fiction.” The “entertainments,” such as This Gun for Hire (1936) and The Ministry of Fear (1943), were thrillers designed to indulge smart readers, whereas the “serious” novels did not tidily resolve themselves. Often, it’s not easy to tell Greene’s two types of novels apart—they uniformly possess weight, deep texture, emotion, absurdist comedy, and compelling characters. In fact, the only thing that really distinguishes his “entertainments” from his “serious” novels is the extent to which the characters suffer. And in the serious novels, they suffer a lot. When Castle, after years of subterfuge designed to protect the wife he loves, is exiled to Russia without her, it is hard not to feel his pain and uselessness; but when Raven—the disfigured hit man of This Gun for Hire—dies, it’s hard to feel anything but relief. Only the “good” suffer in a Greene novel; that’s because, in Greene’s universe, the “bad” never feel remorse, guilt, or the capacity to love. His most unregenerate characters are innocent of self-reflection, like Alden Pyle in Greene’s prescient novel about American “do-gooders” in 1950s Vietnam, The Quiet American . A CIA op trying to save a nation of people from making their own decisions, Pyle is innocent, handsome, and wants to do well. Yet, like many devout Americans who proudly do unto others what they don’t want done unto them, Pyle causes terrible human damage. “Innocence,” the narrator, Fowler, observes of Pyle, “is like a dumb leper who has lost his bell, wandering the world meaning no harm.” As the 2002 film adaptation made perfectly clear, this deadly combination of American innocence and evil looks a lot like Brendan Fraser. No other writer moved so effortlessly between films and fiction. Greene wrote novels with a movie camera’s eye—showing his readers what his characters see, in the sequence in which they see it. (The opening of This Gun for Hire is a textbook example of perfect cinematic technique; nobody ever used the semicolon better to bite off each second of a character’s sensory experience.) At the same time, he wrote films with a novelist’s sense of human complexity, which is probably why he dismissed Alfred Hitchcock as a “clown.” Holly Martin’s problem in The Third Man is that he genuinely loves his boyhood friend Harry Lime, even though he suspects that there may be depths to Harry that he shouldn’t love at all. Harry “was a wonderful planner,” Rollo explains in the novel, but “I was always the one who got caught.” This complexity of personal affection for evil people would never have interested Hitchcock, just as Greene never succumbed to using a meaningless MacGuffin to set his stories rolling. Instead, The Third Man is driven by the real human tragedy of war profiteers in Berlin selling fake antibiotics to sick people (including children). Greene’s films are all tilting Carol Reed camera angles and swerves of light and conjunctions of street and alleyway where people appear and reappear as someone they might actually be; or their conspiracies emerge through the fog and smoke of bombed cities, as in Fritz Lang’s visually fraught version of The Ministry of Fear. For Greene (even though he loved popular fiction and film), the heart of a story wasn’t a clever technical contraption. It always concerned perilous attitudes of perception. Greene wrote about a world we still live in, where horrible crimes are committed every day by beaming politicians with good teeth, aided by their government-conscripted thugs and double agents. And more likely than not, Greene told his stories through the perceptions of people who were either bad or not so good as they (or we) might wish they were. Deeply flawed people who were, in fact, much like Greene saw himself to be. After 86 years of intellectually intense living (when he wasn’t writing books, he seemed to be reading them), Greene died in 1991 without saving the world, but after going a long way toward saving himself. He settled down with one woman in one home in Antibes, France; his depressions and suicidal impulses seemed to withdraw a little, and those who knew him said he grew less angry. And while his final, short novels grew weaker with age, they still provided many beautiful passages. For those of us who love great fiction, those passages will surely endure—at least until the world doesn’t.

An Astronomer Cancels His Own Research–Because the Results Weren’t Popular

As universities try desperately to serve two masters (knowledge production; diversity and inclusion), they will increasingly end up sanctioning speech that should be protected.

Do Repeat Yourself: Duplicated code is not always evil

Welcome to another edition of “Station Wagon Full of Tapes”. In this article I am practically complaining about the increasing obsessiveness on keeping the code DRY and why I like to sometimes duplicate code, on purpose. One of the holy principles of programming seems to be “Don’t repeat yourself”. I, for one, did not learn or hear about this before I was part of the industry. I know who coined the term . I know that they are well-skilled software engineers. However, I am sensing a trend where the idea behind itself has been through the DRY filter, and became a lesser, “simpler” version of itself. Which is all duplicate code is evil, please avoid at all costs . Today, one of the most common code review comments I notice is “do you think we can abstract this section out to a function and re-use?”. I, for one, also tend to give this feedback during reviews from time to time. There are trends in software engineering — as with anything. This specific trend about the journey of DRY code is not new, however it became more and more only about duplicated code. It is in a way gamified, the satisfaction of creating another helper method to do one thing in multiple places. Recently, I had a change of heart. Mostly because I am now part of a team that had to take over some “old” (it really isn’t that old) code and had to retrofit it to a scalable future. The code contained re-usable utilities that were only re-used for two (2) times in total. When the utility was created it did in fact avoid duplication of code. The logic that was moved to a function that had single responsibility. Time passed by, re-orgs happened, some unknown unknowns occurred, unit testing coverage needed to increase. The function now was still only used in two locations, however had so many logical branching (if statements) to accommodate the two it became a hard one to understand. Which sources my frustration with the DRY approach now. I see many examples of moving duplicated sections into common functions, however, I rarely see a code change the other way around. The single responsibility function gets more and more complicated. Instead, what if we applied some moisturizer to the dryness and moved the function back to duplicated versions. To be able to enable this new architecture and structure I’ve mentioned above; I was trying to reorganize the existing code and modules. Fighting with dependency cycles, trying to create some more utilities to abstract sections out. I was not moving forward but just circling around. Things have changed when I just started copy-pasting sections of code to each module that actually used it, and then making them private to that module only. The code became easier to follow, easier to reason, easier to test. What better value to a scalable code than it having the aforementioned attributes. It was not a beautifully structured code (can a code be beautiful?) , at least not in the 2021 DRY standards that every engineer keeps pushing towards each other. But it works, and I guarantee you that it is more flexible than it ever was for future changes, for the next engineer assigned to work on it after another team shuffling at a company. Thank you for reading “Station Wagon Full of Tapes”.