Datasets:

IndonesiaAI

/

cleaned-data-split-0

like 1

I am a LaTeX newbie. I am curious to hear your recommendations about the best places to learn it. More precisely, I am going to start writing my doctoral thesis in it. Our school already has a template but I would like to get the basics first before diving into that template. With that, I plan to continue writing my journal articles in LaTeX instead of MS Word (which is, as you may relate, sometimes painfully unfriendly to scholars when it comes to formatting).

2019/12/23

My go to is [Wiki](https://en.wikibooks.org/wiki/LaTeX). It literally has everything you need. It is laid out pretty great where you can both learn, and use it as a reference. I use it often. Also, the TeX - LaTeX StackExchange is an excellent place to get answers to specific problems. Once you have an appropriate template, this will give you everything you need to format your thesis. Good luck! Also, don't bog yourself down with the details (in my opinion). Understanding exactly what the LaTeX code is doing is going down a rabbit hole. Just focus on understanding what you need to do what you want.

I suggest that you just start using LaTeX for small projects, perhaps at [sharelatex](https://www.google.com/search?client=ubuntu&channel=fs&q=sharelatex&ie=utf-8&oe=utf-8) or [overleaf](https://www.overleaf.com/). If you're really brand new to it, search online for *latex tutorial*. Once you are underway, learn features as you need them. Ask questions on this site. For a large project like your thesis, check out the [workflow tag](https://tex.stackexchange.com/questions/tagged/workflow). You will find strategies for working on one chapter at a time by setting up a good file structure and naming conventions. @Shinobii is right to warn you not to try to understand too much about LaTeX innards. You will probably never have to read what's in the university thesis template.

I am a LaTeX newbie. I am curious to hear your recommendations about the best places to learn it. More precisely, I am going to start writing my doctoral thesis in it. Our school already has a template but I would like to get the basics first before diving into that template. With that, I plan to continue writing my journal articles in LaTeX instead of MS Word (which is, as you may relate, sometimes painfully unfriendly to scholars when it comes to formatting).

2019/12/23

My go to is [Wiki](https://en.wikibooks.org/wiki/LaTeX). It literally has everything you need. It is laid out pretty great where you can both learn, and use it as a reference. I use it often. Also, the TeX - LaTeX StackExchange is an excellent place to get answers to specific problems. Once you have an appropriate template, this will give you everything you need to format your thesis. Good luck! Also, don't bog yourself down with the details (in my opinion). Understanding exactly what the LaTeX code is doing is going down a rabbit hole. Just focus on understanding what you need to do what you want.

Check out the IITB LaTeX course at edX

I am a LaTeX newbie. I am curious to hear your recommendations about the best places to learn it. More precisely, I am going to start writing my doctoral thesis in it. Our school already has a template but I would like to get the basics first before diving into that template. With that, I plan to continue writing my journal articles in LaTeX instead of MS Word (which is, as you may relate, sometimes painfully unfriendly to scholars when it comes to formatting).

2019/12/23

Little known fact: arXiv is great for this. Look for papers in your field on arXiv, and instead of downloading the pdf, download the source file (changing the file type if needed, so that it will open). Not only will you learn how to make fancy tables, pictures, etc... --- you will also often get to read all the comments that the authors neglected to remove before submitting their paper.

I suggest that you just start using LaTeX for small projects, perhaps at [sharelatex](https://www.google.com/search?client=ubuntu&channel=fs&q=sharelatex&ie=utf-8&oe=utf-8) or [overleaf](https://www.overleaf.com/). If you're really brand new to it, search online for *latex tutorial*. Once you are underway, learn features as you need them. Ask questions on this site. For a large project like your thesis, check out the [workflow tag](https://tex.stackexchange.com/questions/tagged/workflow). You will find strategies for working on one chapter at a time by setting up a good file structure and naming conventions. @Shinobii is right to warn you not to try to understand too much about LaTeX innards. You will probably never have to read what's in the university thesis template.

I am a LaTeX newbie. I am curious to hear your recommendations about the best places to learn it. More precisely, I am going to start writing my doctoral thesis in it. Our school already has a template but I would like to get the basics first before diving into that template. With that, I plan to continue writing my journal articles in LaTeX instead of MS Word (which is, as you may relate, sometimes painfully unfriendly to scholars when it comes to formatting).

2019/12/23

I suggest that you just start using LaTeX for small projects, perhaps at [sharelatex](https://www.google.com/search?client=ubuntu&channel=fs&q=sharelatex&ie=utf-8&oe=utf-8) or [overleaf](https://www.overleaf.com/). If you're really brand new to it, search online for *latex tutorial*. Once you are underway, learn features as you need them. Ask questions on this site. For a large project like your thesis, check out the [workflow tag](https://tex.stackexchange.com/questions/tagged/workflow). You will find strategies for working on one chapter at a time by setting up a good file structure and naming conventions. @Shinobii is right to warn you not to try to understand too much about LaTeX innards. You will probably never have to read what's in the university thesis template.

Check out the IITB LaTeX course at edX

I am a LaTeX newbie. I am curious to hear your recommendations about the best places to learn it. More precisely, I am going to start writing my doctoral thesis in it. Our school already has a template but I would like to get the basics first before diving into that template. With that, I plan to continue writing my journal articles in LaTeX instead of MS Word (which is, as you may relate, sometimes painfully unfriendly to scholars when it comes to formatting).

2019/12/23

Little known fact: arXiv is great for this. Look for papers in your field on arXiv, and instead of downloading the pdf, download the source file (changing the file type if needed, so that it will open). Not only will you learn how to make fancy tables, pictures, etc... --- you will also often get to read all the comments that the authors neglected to remove before submitting their paper.

Check out the IITB LaTeX course at edX

As a previous result of a bad TFS project management, several tasks has been created in the wrong work item. Now I need to move several tasks to different work items. Is there an easy way to do it? So far, I have to edit each task, remove the previos link to primary element and create a new one, but this is taking a lot of my time.

2014/11/14

I suspect that the easiest way to do it would be from Excel. Create a Tree-based query that shows everything, then move the child records in Excel using simple cut and insert cut cells. Excel will then allow you to publish the new structure in one go. If you need to move items up to a higher or lower level, place the Title Field in the column representing the level. [See this little video I captured to show how it is done.](https://www.youtube.com/watch?v=86-8cb4LhS8&hd=1)

MS Project is extremely good with modifying hierarchies of work items. The steps are exactly the same as setting it up in Excel, but project inherently handles parent/child relationships, giving them a drag-and-drop interaction. jessehouwing's Excel answer will be easier if you have never worked with project before. **Updated** jesshouwing's comments are correct. Especially about the shivers.

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

I have exactly the same situation at work (with CVS instead of subversion and the rest of the team using KDevelop but that's no big deal). Just start a new Qt Gui project using the Qt - Eclipse integration features and then remove all the auto generated files. Now using the "Team" features of eclipse and choose to share your project, enter the path to the repository and you 're good to go.

Second nikolavp - Checkout, and mark the option to use the new project wizard, then select Qt project. I've done this (with ganymede) and it successfully finds everything and builds correctly.

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

OK, I've been playing around with this idea, and it has some merit. I can switch to the "SVN Project Exploring" perspective (which I hadn't noticed before), and do a checkout from the head of the sub-project I want. I get a nice SVN-linked copy of the tree in my Eclipse workspace for editing. Eclipse even "understands" the classes, and can do completion on methods and such. However, I still can't get Eclipse to understand that the project is a "QT Gui" project, such that I could view the properties, and control the linking of the various Qt libraries and the like. By extension, it also doesn't understand how to build my project, like it would be able to do if I had created an empty Qt Gui project from scratch. How do I get this part working?

The only way I could get this to work was to check out the project with eclipse and then copy over the .project and .cdtproject files from another Qt-project. Then do a refresh on the project. This is a horrible hack but it gets you started. You might need to define another builder for 'make'.

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

I would create a new QT project in eclipse, then switch perspectives to subclipse and simply do a SVN checkout into the new eclipse project. You should be good to go.

OK, I've been playing around with this idea, and it has some merit. I can switch to the "SVN Project Exploring" perspective (which I hadn't noticed before), and do a checkout from the head of the sub-project I want. I get a nice SVN-linked copy of the tree in my Eclipse workspace for editing. Eclipse even "understands" the classes, and can do completion on methods and such. However, I still can't get Eclipse to understand that the project is a "QT Gui" project, such that I could view the properties, and control the linking of the various Qt libraries and the like. By extension, it also doesn't understand how to build my project, like it would be able to do if I had created an empty Qt Gui project from scratch. How do I get this part working?

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

I would create a new QT project in eclipse, then switch perspectives to subclipse and simply do a SVN checkout into the new eclipse project. You should be good to go.

Checkout the project. It will ask you some options like if you want to start with a blank project, or want to use the tree to make a new project. Choose the latter and you should be ok :). It seems to work for me with Ganymed and subversive(not sure about subclipse and i don't remember.) :)

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

OK, I've been playing around with this idea, and it has some merit. I can switch to the "SVN Project Exploring" perspective (which I hadn't noticed before), and do a checkout from the head of the sub-project I want. I get a nice SVN-linked copy of the tree in my Eclipse workspace for editing. Eclipse even "understands" the classes, and can do completion on methods and such. However, I still can't get Eclipse to understand that the project is a "QT Gui" project, such that I could view the properties, and control the linking of the various Qt libraries and the like. By extension, it also doesn't understand how to build my project, like it would be able to do if I had created an empty Qt Gui project from scratch. How do I get this part working?

My solution: 1. go to the svn-view and add the repository location for your project 2. check out the project some temporary location with svn or any client you like 3. choose 'File->Import...' and say 'Qt->Qt project' 4. browse to the location of the \*.pro file, select and hit the OK-Button 5. you are in the game with an appropriate Qt-project and Subversion Access for that project

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

I have exactly the same situation at work (with CVS instead of subversion and the rest of the team using KDevelop but that's no big deal). Just start a new Qt Gui project using the Qt - Eclipse integration features and then remove all the auto generated files. Now using the "Team" features of eclipse and choose to share your project, enter the path to the repository and you 're good to go.

I would say the same as the last one, but instead of the two first steps I would set up the Qt-Eclipse integration: [Qt-Eclipse integration](http://www.qtsoftware.com/developer/eclipse-integration) before looking for the \*.pro file.

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

OK, I've been playing around with this idea, and it has some merit. I can switch to the "SVN Project Exploring" perspective (which I hadn't noticed before), and do a checkout from the head of the sub-project I want. I get a nice SVN-linked copy of the tree in my Eclipse workspace for editing. Eclipse even "understands" the classes, and can do completion on methods and such. However, I still can't get Eclipse to understand that the project is a "QT Gui" project, such that I could view the properties, and control the linking of the various Qt libraries and the like. By extension, it also doesn't understand how to build my project, like it would be able to do if I had created an empty Qt Gui project from scratch. How do I get this part working?

I would say the same as the last one, but instead of the two first steps I would set up the Qt-Eclipse integration: [Qt-Eclipse integration](http://www.qtsoftware.com/developer/eclipse-integration) before looking for the \*.pro file.

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

OK, I've been playing around with this idea, and it has some merit. I can switch to the "SVN Project Exploring" perspective (which I hadn't noticed before), and do a checkout from the head of the sub-project I want. I get a nice SVN-linked copy of the tree in my Eclipse workspace for editing. Eclipse even "understands" the classes, and can do completion on methods and such. However, I still can't get Eclipse to understand that the project is a "QT Gui" project, such that I could view the properties, and control the linking of the various Qt libraries and the like. By extension, it also doesn't understand how to build my project, like it would be able to do if I had created an empty Qt Gui project from scratch. How do I get this part working?

Checkout the project. It will ask you some options like if you want to start with a blank project, or want to use the tree to make a new project. Choose the latter and you should be ok :). It seems to work for me with Ganymed and subversive(not sure about subclipse and i don't remember.) :)

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

OK, I've been playing around with this idea, and it has some merit. I can switch to the "SVN Project Exploring" perspective (which I hadn't noticed before), and do a checkout from the head of the sub-project I want. I get a nice SVN-linked copy of the tree in my Eclipse workspace for editing. Eclipse even "understands" the classes, and can do completion on methods and such. However, I still can't get Eclipse to understand that the project is a "QT Gui" project, such that I could view the properties, and control the linking of the various Qt libraries and the like. By extension, it also doesn't understand how to build my project, like it would be able to do if I had created an empty Qt Gui project from scratch. How do I get this part working?

Second nikolavp - Checkout, and mark the option to use the new project wizard, then select Qt project. I've done this (with ganymede) and it successfully finds everything and builds correctly.

Most of the work being done at my company is Qt-based C++, and it's all checked into a Subversion repository. Until now, all work on the codebase has been done purely with nano, or perhaps Kate. Being new here, I would like to take advantage of setting up Eclipse -properly- to edit my local copy of the tree. I have the CDT "version" of Eclipse, and the Qt integration, and the Subclipse module. At this point, though, I don't know what to do. Do I "import" the projects into an Eclipse-controlled workspace? Do I edit them in place? Nothing I've tried to do gets Eclipse to recognize that the "project" is a Qt application, so that I can get the integration working.

2008/09/16

I would create a new QT project in eclipse, then switch perspectives to subclipse and simply do a SVN checkout into the new eclipse project. You should be good to go.

My solution: 1. go to the svn-view and add the repository location for your project 2. check out the project some temporary location with svn or any client you like 3. choose 'File->Import...' and say 'Qt->Qt project' 4. browse to the location of the \*.pro file, select and hit the OK-Button 5. you are in the game with an appropriate Qt-project and Subversion Access for that project

i am running the airflow pipeline but codes looks seems good but actually i'm getting the airflow.exceptions.AirflowException: Cycle detected in DAG. Faulty task: can u please help to resolve this issue

2017/08/17

This can happen due to duplicate task\_id'a in multiple tasks.

Without the code, it's kind of hard to help you. However, this means that you have a loop in your DAG. Generally, thie error happens when one of your task has a downstream task whose own downstream chain includes it again (A calls B calls C calls D calls A again, for example). That's not permitted by Airflow (and DAGs in general).

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

### Use a mixture to modify the water's properties You could use an antifreeze/water mixture that you put in a bottle. You will have a much lower freezing and higher boiling point that just water. This will allow you to keep your frozen packs at a lower temperature than plain ice for longer. ### Safety Concerns You will also need to be **extremely careful with antifreeze around food and/or drinks as it is toxic to almost all mammals.** Other creatures may also have toxic reactions to the antifreeze. Insure that if you use a recycled water bottle that it is **extremely difficult** (*aim for impossible*) for you and/or **children** to get into it. ### Options You could also try using alcohol instead of water. It is the same principle as mixing antifreeze with water. It is also much safer than antifreeze, but you will still need to modify the bottle to prevent child access.

The "R" value of the cooler insulation as well as the seal around the lid are important in retaining the cold. The size of the cooler is also important because the ratio of ice to the product you are trying to keep cool is also a factor. You can get pink RV anti-freeze used to winterize drinking water pipes and tanks in boats and motor homes It is I would assume non-toxic if accidentally ingested. Ice packs made from anti-freeze would probably cool faster to a lower temperature but not for a longer time.

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

### Use a mixture to modify the water's properties You could use an antifreeze/water mixture that you put in a bottle. You will have a much lower freezing and higher boiling point that just water. This will allow you to keep your frozen packs at a lower temperature than plain ice for longer. ### Safety Concerns You will also need to be **extremely careful with antifreeze around food and/or drinks as it is toxic to almost all mammals.** Other creatures may also have toxic reactions to the antifreeze. Insure that if you use a recycled water bottle that it is **extremely difficult** (*aim for impossible*) for you and/or **children** to get into it. ### Options You could also try using alcohol instead of water. It is the same principle as mixing antifreeze with water. It is also much safer than antifreeze, but you will still need to modify the bottle to prevent child access.

Place you whole cooler in extra isolation, like a good sleeping bag or several old blankets. This works best if you do not have to open the cooler much or at all. You can take out the beers all at once and keep them cool by rolling them, individualy, in a big towel. That allows the cooler to stay at a steady temperature.

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

Insulate your cooler more effectively. -------------------------------------- You didn't mention what type of cooler you are using, so I am guessing that you have a standard injection molded plastic cooler. If you have Styrofoam, then this answer won't work for you. Drill a hole into the injection molded plastic at one corner so that your hole makes it between the layers of plastic. On the opposite corner, drill another hole so that will allow air to escape. Buy a can of insulating spray foam at your favorite hardware store. Following the instructions on the can, spray the foam into either of the holes you drilled previously. Once you have foam flowing out the other hole, you will want to rotate the cooler to get the foam in as many places internally as possible. Use the quick curing foam so you don't have to tape the holes closed. Let it sit as per the instructions on the can. If you find that you have missed sections of the cooler, you can always drill more holes and repeat the steps as necessary.

The "R" value of the cooler insulation as well as the seal around the lid are important in retaining the cold. The size of the cooler is also important because the ratio of ice to the product you are trying to keep cool is also a factor. You can get pink RV anti-freeze used to winterize drinking water pipes and tanks in boats and motor homes It is I would assume non-toxic if accidentally ingested. Ice packs made from anti-freeze would probably cool faster to a lower temperature but not for a longer time.

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

Insulate your cooler more effectively. -------------------------------------- You didn't mention what type of cooler you are using, so I am guessing that you have a standard injection molded plastic cooler. If you have Styrofoam, then this answer won't work for you. Drill a hole into the injection molded plastic at one corner so that your hole makes it between the layers of plastic. On the opposite corner, drill another hole so that will allow air to escape. Buy a can of insulating spray foam at your favorite hardware store. Following the instructions on the can, spray the foam into either of the holes you drilled previously. Once you have foam flowing out the other hole, you will want to rotate the cooler to get the foam in as many places internally as possible. Use the quick curing foam so you don't have to tape the holes closed. Let it sit as per the instructions on the can. If you find that you have missed sections of the cooler, you can always drill more holes and repeat the steps as necessary.

Place you whole cooler in extra isolation, like a good sleeping bag or several old blankets. This works best if you do not have to open the cooler much or at all. You can take out the beers all at once and keep them cool by rolling them, individualy, in a big towel. That allows the cooler to stay at a steady temperature.

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

I've never done this myself, but I've camped with people who have used dry ice to keep things cold/frozen over the course of four days. Dry ice is much colder than regular ice, so you may actually have to insulate things you don't want to freeze from it. Keep it at the bottom of the bag, or maybe move it to the top if it isn't cooling as much. It's a bad idea to hold dry ice with your hands, so think about plastic bags and tongs. I would also say that dry ice is a "slower" coolant, so try not to open your cooler as much.

The "R" value of the cooler insulation as well as the seal around the lid are important in retaining the cold. The size of the cooler is also important because the ratio of ice to the product you are trying to keep cool is also a factor. You can get pink RV anti-freeze used to winterize drinking water pipes and tanks in boats and motor homes It is I would assume non-toxic if accidentally ingested. Ice packs made from anti-freeze would probably cool faster to a lower temperature but not for a longer time.

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

I've never done this myself, but I've camped with people who have used dry ice to keep things cold/frozen over the course of four days. Dry ice is much colder than regular ice, so you may actually have to insulate things you don't want to freeze from it. Keep it at the bottom of the bag, or maybe move it to the top if it isn't cooling as much. It's a bad idea to hold dry ice with your hands, so think about plastic bags and tongs. I would also say that dry ice is a "slower" coolant, so try not to open your cooler as much.

Place you whole cooler in extra isolation, like a good sleeping bag or several old blankets. This works best if you do not have to open the cooler much or at all. You can take out the beers all at once and keep them cool by rolling them, individualy, in a big towel. That allows the cooler to stay at a steady temperature.

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

This may seem to be a simple approach but here it goes. 2 options: 1. 2 separate coolers, one for beers one for food. If you don't open the food cooler till the morning the cooling will last much longer. 2. The easier option: place your beers in the freezer as well not long enough for them to freeze but maybe 5 hours they would drop to -4 > -7 degrees, but not freeze. Therefore bottles or cans are not damaged but cold enough that they will double the life of your cooler as they will continue to cool all the time they are in there.

The "R" value of the cooler insulation as well as the seal around the lid are important in retaining the cold. The size of the cooler is also important because the ratio of ice to the product you are trying to keep cool is also a factor. You can get pink RV anti-freeze used to winterize drinking water pipes and tanks in boats and motor homes It is I would assume non-toxic if accidentally ingested. Ice packs made from anti-freeze would probably cool faster to a lower temperature but not for a longer time.

I'd like to make some extra icepacks to put in my coolbag. The 2 commercial icepacks it comes with don't maintain the temperature for long enough, so I've tried supplementing them with plastic bottles of water that I've frozen. This works reasonably well but are there other containers that would work better and is there anything I could add to the water?

2015/07/27

This may seem to be a simple approach but here it goes. 2 options: 1. 2 separate coolers, one for beers one for food. If you don't open the food cooler till the morning the cooling will last much longer. 2. The easier option: place your beers in the freezer as well not long enough for them to freeze but maybe 5 hours they would drop to -4 > -7 degrees, but not freeze. Therefore bottles or cans are not damaged but cold enough that they will double the life of your cooler as they will continue to cool all the time they are in there.

Place you whole cooler in extra isolation, like a good sleeping bag or several old blankets. This works best if you do not have to open the cooler much or at all. You can take out the beers all at once and keep them cool by rolling them, individualy, in a big towel. That allows the cooler to stay at a steady temperature.

> > **Possible Duplicate:** > > [Where does this concept of “favor composition over inheritance” come from?](https://softwareengineering.stackexchange.com/questions/65179/where-does-this-concept-of-favor-composition-over-inheritance-come-from) > > > I have colleagues at work who claim that "Inheritance is an anti-pattern" and want to use composition systematically instead, except in (rare, according to them) cases where inheritance is really the best way to go. I want to suggest an alternative where we continue using inheritance, but it is strictly forbidden (enforced by code reviews) to use anything but public members of base classes in derived classes. For a case where we don't need to swap components of a class at runtime (static inheritance), would that be equivalent enough to composition? Or am I forgetting some other important aspect of composition?

2012/11/12

You're sort of looking at it backwards. Composition isn't preferred because of some unseen benefit of composition, it's preferred because it avoids the drawbacks of inheritance. Inheritance adds the significant benefit of substitutability. However, that benefit comes with the significant cost of tight coupling. Inheritance is the strongest coupling relationship possible. If you don't need the substitutability, the coupling cost isn't worth it. Your restricted version of inheritance addresses its other cost, a loss of encapsulation, but does nothing about the much more significant cost of coupling.

I think the difference between the two things is in their intent. Is the class you are writing really meant to be used as replacement / specialisation of its Base class or is it just using some functionality? If the First Case is true you should go with inheritance. If the latter is true go for composition. In my experience using inheritance where composition would be appropriate creates a more complex, less understandable and meanigful codebase.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

Personally, I feel that the beauty of the badges is in their names. Some are very creative and others are well, some may say abstract. As much as part of me is curious/excited to see how the graphical badges might turn out, I feel that it will dilute the names' impact. Having say that, maybe someone can come out with a couple 'prototype' graphical badges that can swing my vote. :P UPDATE: Added dboarman's recommendation [some badges from dboarman http://imgcash4.imageshack.us/img197/8514/badgeideas.png?0.6349713410095544](http://imgcash4.imageshack.us/img197/8514/badgeideas.png?0.6349713410095544)

For some reason, this reminded me of [a scene from *Time Bandits*](http://www.youtube.com/watch?v=jmS_hMnFNVM#t=1m12s). View last few seconds starting from about 1:12.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

You can see graphical badges on some of the Stack Exchange sites -- epic advice and VFX overflow come to mind. While I fully support this kind of customization for SE, I am not sure it really fits well with the highly-text oriented, no frills nature of the trilogy at the moment.

For some reason, this reminded me of [a scene from *Time Bandits*](http://www.youtube.com/watch?v=jmS_hMnFNVM#t=1m12s). View last few seconds starting from about 1:12.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

It doesn't sound real useful even for visual effect, since the badges are only seen when you look at your profile page or the badge page.

For some reason, this reminded me of [a scene from *Time Bandits*](http://www.youtube.com/watch?v=jmS_hMnFNVM#t=1m12s). View last few seconds starting from about 1:12.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

My eyes are already bleeding

You can see graphical badges on some of the Stack Exchange sites -- epic advice and VFX overflow come to mind. While I fully support this kind of customization for SE, I am not sure it really fits well with the highly-text oriented, no frills nature of the trilogy at the moment.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

Personally, I feel that the beauty of the badges is in their names. Some are very creative and others are well, some may say abstract. As much as part of me is curious/excited to see how the graphical badges might turn out, I feel that it will dilute the names' impact. Having say that, maybe someone can come out with a couple 'prototype' graphical badges that can swing my vote. :P UPDATE: Added dboarman's recommendation [some badges from dboarman http://imgcash4.imageshack.us/img197/8514/badgeideas.png?0.6349713410095544](http://imgcash4.imageshack.us/img197/8514/badgeideas.png?0.6349713410095544)

You can see graphical badges on some of the Stack Exchange sites -- epic advice and VFX overflow come to mind. While I fully support this kind of customization for SE, I am not sure it really fits well with the highly-text oriented, no frills nature of the trilogy at the moment.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

Personally, I feel that the beauty of the badges is in their names. Some are very creative and others are well, some may say abstract. As much as part of me is curious/excited to see how the graphical badges might turn out, I feel that it will dilute the names' impact. Having say that, maybe someone can come out with a couple 'prototype' graphical badges that can swing my vote. :P UPDATE: Added dboarman's recommendation [some badges from dboarman http://imgcash4.imageshack.us/img197/8514/badgeideas.png?0.6349713410095544](http://imgcash4.imageshack.us/img197/8514/badgeideas.png?0.6349713410095544)

It doesn't sound real useful even for visual effect, since the badges are only seen when you look at your profile page or the badge page.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

We do have a very large community here. Of programmers, system administrators and computer power users. I'm not sure I'd count too heavily on this community for engaging, attractive, consistent (from a design language perspective) graphical icons. Now, if you want a bunch of awesome freehand MS Paint art, you are totally in the right place. Freehand circles and arrows FTW!

For some reason, this reminded me of [a scene from *Time Bandits*](http://www.youtube.com/watch?v=jmS_hMnFNVM#t=1m12s). View last few seconds starting from about 1:12.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

You can see graphical badges on some of the Stack Exchange sites -- epic advice and VFX overflow come to mind. While I fully support this kind of customization for SE, I am not sure it really fits well with the highly-text oriented, no frills nature of the trilogy at the moment.

It doesn't sound real useful even for visual effect, since the badges are only seen when you look at your profile page or the badge page.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

My eyes are already bleeding

It doesn't sound real useful even for visual effect, since the badges are only seen when you look at your profile page or the badge page.

Well, the faq does say no question is too small and I haven't found this question asked elsewhere so... I'd like to suggest the introduction of icons for the badges, as seen in Team Fortress 2, Day of Defeat: Source or even Kongregate. Imagine your favourite badge transformed with a colourful and amusing image! With a community as large and diverse as this one, we ought to be able to come up with some decent badge icons. What say ye? EDIT: How about them apples? (Just a sample of what's possible..) <http://www.freeimagehosting.net/uploads/708d33b019.jpg> Commentor <http://www.freeimagehosting.net/uploads/e367b04bbe.jpg> Epic

2010/02/04

My eyes are already bleeding

We do have a very large community here. Of programmers, system administrators and computer power users. I'm not sure I'd count too heavily on this community for engaging, attractive, consistent (from a design language perspective) graphical icons. Now, if you want a bunch of awesome freehand MS Paint art, you are totally in the right place. Freehand circles and arrows FTW!

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

Well I responded to some posts above that had incorrect information, but I figured I should just post my own response to better explain. Anonymous is made up of basically 2 subgroups: 1. Skiddies (script kiddies) and newbies who have only the most basic security knowledge, and just sit in their IRC and basically be the pwns for the attack. These are the people that the FBI was knocking down their doors. 2. Anonymous core leadership, a group with some hacking knowledge that owned hbgary, but also got owned recently by ninja hack squad. You won't be able to trace this subgroup unless you are a security guru. **How do they hide their tracks?** Like a previous answerers mentioned, 1. **Through proxy servers** like Tor 2. **by compromising boxes and launching attacks from those boxes** (basically masquerading as that person's IP), or 3. **by using a VPN that's in a foreign country** and keeps no logs. With the VPN, all your traffic is relayed through it so wherever you connect it can only track back the IP addy to the VPN itself and no further (unless the VPN is keeping logs in which case you shouldn't use it anyways). Hope this helps clarify a bit.

Maybe you should read [this PDF](http://cryptome.org/0003/anonymous-barr.pdf). They are not so anonymous. The LOIC tool used for DDOS, leaks the original IP of the person using it. You can use the browser (JavaScript) version of the same tool, maybe hiding behind Tor. HBGary Federal exposed their names and addresses in that PDF. That is why they attacked his site, email, wipe his iPad, took over his twitter etc.... Search the #hbgary hashtag on twitter for more info on that.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

In addition to the answers that have already been given, another reason it is so hard to catch **anonymous** is because **anonymous** can be anyone, literally. I mean this in two ways. First, hackers can use a combination of malware, spyware, and bots to access and use/loop through other peoples computers anywhere in the world; thus, making any computer, theoretically, a point from which **anonymous** can work. Secondly, true to the name **anonymous**, any hacker, anywhere, using any method or style, using any random pattern of activity, can make their attack and call themselves **anonymous**. Thus, it is extremely difficult for a government/authority to track activity by pattern or style or signature, because it is always changing due to the varied nature of the attacks since it can, as I said before, literally be coming from anyone. Essentially, **Anonymous** is not one person... **Anonymous** is not one group... **Anonymous** is anywhere and everywhere... **Anonymous** could be everyone or no one... Unfortunately, that is the nature, uniqueness, and genius of the name.

Maybe you should read [this PDF](http://cryptome.org/0003/anonymous-barr.pdf). They are not so anonymous. The LOIC tool used for DDOS, leaks the original IP of the person using it. You can use the browser (JavaScript) version of the same tool, maybe hiding behind Tor. HBGary Federal exposed their names and addresses in that PDF. That is why they attacked his site, email, wipe his iPad, took over his twitter etc.... Search the #hbgary hashtag on twitter for more info on that.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

One of the most important aspects of an attack like this is covering your tracks. There are lots of different ways to do this, as it depends on the technology. To address your specific questions: When they DDoS: If the flood was coming from their own machines, then it would be fairly easy to track them. The problem lies in the fact that they aren't using their own machines. They are either a) taking control of someone elses without permission, or b) getting someone to do it on their behalf. The latter is what happened with the Wikileaks attacks. People signed up to to do it. Things start getting hinky when servers are in countries that don't generally respond to requests for logs. If the company that is being attacked is in the US, it's fairly easy to get a court order if the attack can be proven to originate in the States. What happens if it's a US target, but the attack is originating in Russia or China? The same thing goes for purchase records. As for being scared... there are quite a few of these sorts of groups out there. Most of them are (I don't want to say harmless, but...) harmless. In this particular case, someone poked the bear and the bear got pissed. EDIT: Not that I condone their actions, blah blah blah.

In addition to the answers that have already been given, another reason it is so hard to catch **anonymous** is because **anonymous** can be anyone, literally. I mean this in two ways. First, hackers can use a combination of malware, spyware, and bots to access and use/loop through other peoples computers anywhere in the world; thus, making any computer, theoretically, a point from which **anonymous** can work. Secondly, true to the name **anonymous**, any hacker, anywhere, using any method or style, using any random pattern of activity, can make their attack and call themselves **anonymous**. Thus, it is extremely difficult for a government/authority to track activity by pattern or style or signature, because it is always changing due to the varied nature of the attacks since it can, as I said before, literally be coming from anyone. Essentially, **Anonymous** is not one person... **Anonymous** is not one group... **Anonymous** is anywhere and everywhere... **Anonymous** could be everyone or no one... Unfortunately, that is the nature, uniqueness, and genius of the name.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

One of the most important aspects of an attack like this is covering your tracks. There are lots of different ways to do this, as it depends on the technology. To address your specific questions: When they DDoS: If the flood was coming from their own machines, then it would be fairly easy to track them. The problem lies in the fact that they aren't using their own machines. They are either a) taking control of someone elses without permission, or b) getting someone to do it on their behalf. The latter is what happened with the Wikileaks attacks. People signed up to to do it. Things start getting hinky when servers are in countries that don't generally respond to requests for logs. If the company that is being attacked is in the US, it's fairly easy to get a court order if the attack can be proven to originate in the States. What happens if it's a US target, but the attack is originating in Russia or China? The same thing goes for purchase records. As for being scared... there are quite a few of these sorts of groups out there. Most of them are (I don't want to say harmless, but...) harmless. In this particular case, someone poked the bear and the bear got pissed. EDIT: Not that I condone their actions, blah blah blah.

[Here is an article](http://www.scientificamerican.com/article.cfm?id=tracking-cyber-hackers) asking (and answering) just that very question from the Scientific American site posted this month. The short answer to the question is spoofing of source addresses and the use of proxies.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

The thing about a DDoS is that you use *other people's* IPs, not your own. It's relatively simple to become untraceable on the Internet -- just route your traffic though a host that is not keeping traffic logs. As someone who frequently has to try to track these people down, I can tell you what an impossible nightmare it is. Here's the pattern I frequently see: 1. Select a relatively recent exploit in some web software package (e.g. joomla extension). 2. Use google to find an appropriately vulnerable attack target 3. From some location that can't be traced to you (e.g. coffee shop), execute the attack to gain control over the vulnerable server, but don't do anything else that would draw attention to yourself. (bonus points, fix the vulnerablity so no one comes in behind you). Delete any logs that might trace back to your presumed location. 4. Repeat the above, relaying your traffic through the previously compromised server. Repeat again several times until you're removed multiple steps from the machine that will be behaving as your proxy. Ideally these servers should be located in countries like China, India, Brasil, Mexico, etc., where datacenter techs tend to be *uncooperative* toward investigations, and should all be located in different countries as to create jurisdiction and communication nightmares for the people trying to track you. Congratulations, you're now anonymous on the Internet. It's a bit like Tor, except none of the nodes know they're participating. Usually these attackers set up and use backdoors on servers for which no logs or records are kept (since the backdoor presumably doesn't exist). Once the attacker disconnects, that link becomes permanently untraceable. One hop drops your chances of detection dramatically. Two hops makes detection almost impossible. Three hops and it's not even worth the effort.

Maybe you should read [this PDF](http://cryptome.org/0003/anonymous-barr.pdf). They are not so anonymous. The LOIC tool used for DDOS, leaks the original IP of the person using it. You can use the browser (JavaScript) version of the same tool, maybe hiding behind Tor. HBGary Federal exposed their names and addresses in that PDF. That is why they attacked his site, email, wipe his iPad, took over his twitter etc.... Search the #hbgary hashtag on twitter for more info on that.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

There are NUMEROUS ways for a hacker to cover their tracks.. Here is one very generalized example: A hacker can compromise a third party machine and use it to do attacks on the hackers behalf. Because the system is compromised, the hacker can delete/modify logs. A hacker can also piggyback machines, such as, log into machine A, from machine A log into machine B, from machine B log into machine C, from machine C attack machine D, then cleanse the logs for machines C, B, then A making tracking the hacker more difficult. This doesn't even take into account hacked internet accounts (so even if traced back they point to a different person), open proxies, etc etc etc.. I know the above isn't flawless, but like I said this is just a VERY VERY general example. There are many many ways to cover your tracks. That said, what makes you so sure certain 3 letter agencies don't already know who many of them are, but don't make a move on them so that those individuals can lead them to others? I'm sure others will chime in who can explain more thoroughly, but I think the ultimate lesson to be learned is to concern yourself less with specific hackers and hacking groups, and more with your own security. The fact that their latest claim to fame originated from something as TRIVIAL to fix as a SQL Injection vulnerability (which is nothing new, very well documented and understood) is a huge discredit to the unnamed "security firm" who was hacked. *rant over*

Several post discuss the technical difficulties in finding the persons behind these groups. It is not at all easy to backtrack their activity when using many machines to create a sense of anonymity. Another very important aspect is that the police, the intelligence communities around the globe and the different counties legislation is not really constructed to handle these situations. So if you find a server in one country that has been used to hop to a server in another country it takes too long to go through the proper channels to get the local police to get hold of the information. Even if you do the information such as logs are not always kept for longer periods of time. It's easy to unlawfully hop around the Internet, but much much slower to hop around the Internet in a lawful way. This is a very prohibiting factor when trying to find these groups.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

Maybe you should read [this PDF](http://cryptome.org/0003/anonymous-barr.pdf). They are not so anonymous. The LOIC tool used for DDOS, leaks the original IP of the person using it. You can use the browser (JavaScript) version of the same tool, maybe hiding behind Tor. HBGary Federal exposed their names and addresses in that PDF. That is why they attacked his site, email, wipe his iPad, took over his twitter etc.... Search the #hbgary hashtag on twitter for more info on that.

**Hackers can be caught, Anonymous cannot.** Anonymous is such a lose collective that it is not materially hurt by law enforcement striking out at its individual hackers. However, it does respond violently against any organization that attempts to do so. This means * Its very hard to strike down Anonymous just by catching its members. * Anonymous will make life hard on anyone who tries. All Anonymous has to do is continue to be "not worth the effort" to go after its members en masse and it will continue to be free. However, they play a dangerous game. If the public ever decides they are a sufficient nuisance, then it will suddenly be worth the cost to track down and catch its members, enduring the counter-hacks by Anonymous as they go.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

From some experience with law enforcement and forensics, I can say one of the biggest issues is that ISPs really don't want to have to track users. Once they get beyond a certain level of management they lose 'common carrier' status and become liable for an awful lot of what their customers may do. Also, many countries do not want to pass on information to another country - especially countries which may be opposed to western culture or western interference. And it is extremely easy to hide almost anything on the internet. Regarding your three points: * **Server should have IP addresses** - No - this is simple to spoof or erase * **Private server** - Not likely, although possible - but it wouldn't be their credit card used * **ISP's trace** - Not going to happen - it doesn't affect ISP's negatively, and is way too difficult **update** It might happen after all - <http://blogs.forbes.com/andygreenberg/2011/03/18/ex-anonymous-hackers-plan-to-out-groups-members/>

There is one thing that hasn't been mentioned yet: the human factor. These groups do not have a hierarchy as such, instead they form around a set of ideas. Most of the time, the only idea in common is "the governments are wrong, we must do justice by hacking", which is probably a feeling that's only getting stronger, with the current pressure the US gov (itself pressured by corporations) is putting in other countries below the covers to pass draconian laws against free speech that could harm the aforementioned corporations. So the great appeal here, especially by Anonymous, is that if you have the knowledge and hate the government (who doesn't?), you can join them by yourself and on your own account and risk. In order to see where this thinking comes from, I recommend the movie/comic novel "V for Vendetta", from which they took that mask you see so often. Some groups, of course, have much less heroic intentions. LulzSec was "all for the lulz". The bottom line is that yes, they might get a few members of each group, but more will show up.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

My answer pokes at the original question. What makes you think that they don't get caught? The CIA and DoD found Osama bin Laden. Typical means include OSINT, TECHINT, and HUMINT. Forensics can be done on Tor. Secure deletion tools such as sdelete, BCWipe, and DBAN are not perfect. Encryption tools such as GPG and Truecrypt are not perfect. Online communications was perhaps Osama bin Laden's biggest strength (he had couriers that traveled to far away cyber-cafes using email on USB flash drives) and Anonymous/LulzSec's biggest weakness. They use unencrypted IRC usually. You think they'd at least be using OTR through Tor with an SSL proxy to the IM communications server(s) instead of a cleartext traffic through an exit node. Their common use of utilities such as Havij and sqlmap could certainly backfire. Perhaps there is a client-side vulnerability in the Python VM. Perhaps there is a client-side buffer overflow in Havij. Perhaps there are backdoors in either. Because of the political nature of these groups, there will be internal issues. I saw some news lately that 1 in 4 hackers are informants for the FBI. It's not "difficult" to "catch" anyone. Another person on these forums suggested that I watch a video from a Defcon presentation where the presenter tracks down a Nigerian scammer using the advanced transform capabilities in Maltego. The OSINT capabilities of Maltego and the i2 Group Analyst's Notebook are fairly limitless. A little hint; a little OPSEC mistake -- and a reversal occurs: the hunter is now being hunted.

[Here is an article](http://www.scientificamerican.com/article.cfm?id=tracking-cyber-hackers) asking (and answering) just that very question from the Scientific American site posted this month. The short answer to the question is spoofing of source addresses and the use of proxies.

I'm not security literate, and if I was, I probably wouldn't be asking this question. As a regular tech news follower, I'm really surprised by the [outrage](http://en.wikipedia.org/wiki/Anonymous_%28group%29#Activities) of [Anonymous (hacker group)](http://en.wikipedia.org/wiki/Anonymous_%28group%29), but as a critical thinker, I'm unable to control my curiosity to dig out how exactly they are doing this? Frankly, this group really scares me. One thing that I don't understand is how they haven't been caught yet. Their IP addresses should be traceable when they DDOS, even if they spoof it or go through a proxy. * The server with which they are spoofing should have recorded the IPs of these guys in its logs. If the govt. ask the company (which owns the server) don't they give the logs? * Even if it is a private server owned by these guys, doesn't IANA (or whoever the organization is) have the address & credit card details of the guy who bought & registered the server? * Even if they don't have that, can't the ISPs trace back to the place these packets originated? I know, if it was as simple as I said, the government would have caught them already. So how exactly are they able to escape? PS: If you feel there are any resources that would enlighten me, I'll be glad to read them. [Update - this is equally appropriate when referring to the [Lulzsec group](http://en.wikipedia.org/wiki/LulzSec), so have added a quick link to the Wikipedia page on them]

2011/02/21

Well I responded to some posts above that had incorrect information, but I figured I should just post my own response to better explain. Anonymous is made up of basically 2 subgroups: 1. Skiddies (script kiddies) and newbies who have only the most basic security knowledge, and just sit in their IRC and basically be the pwns for the attack. These are the people that the FBI was knocking down their doors. 2. Anonymous core leadership, a group with some hacking knowledge that owned hbgary, but also got owned recently by ninja hack squad. You won't be able to trace this subgroup unless you are a security guru. **How do they hide their tracks?** Like a previous answerers mentioned, 1. **Through proxy servers** like Tor 2. **by compromising boxes and launching attacks from those boxes** (basically masquerading as that person's IP), or 3. **by using a VPN that's in a foreign country** and keeps no logs. With the VPN, all your traffic is relayed through it so wherever you connect it can only track back the IP addy to the VPN itself and no further (unless the VPN is keeping logs in which case you shouldn't use it anyways). Hope this helps clarify a bit.

[Here is an article](http://www.scientificamerican.com/article.cfm?id=tracking-cyber-hackers) asking (and answering) just that very question from the Scientific American site posted this month. The short answer to the question is spoofing of source addresses and the use of proxies.

Ven. Members of the Sangha (of Bhikkhus), Ven. fellows, valued Upasaka and Upasika, dear readers and interesed, This question is intellectual and theoretical, or literary, but also/merely inviting self-reflection; but of course it can also be answered with words of the Buddha in "should-form" (in the later case, exchange "you" with "a serious follower of the Buddha, a person in general, to find peace", how ever you wish and feel obligated or loyal to). So, about loyalty, * In regard of what are you able to claim being loyal? * Whom are you loyal to? * How far goes your loyalty? * Where and how does it end? Also, finally: * What must one be loyal to, to find, to reach the highest aim all Buddha-following seekers are after? * What are you loyal and/or are you not loyal to, so that you have't found final peace already, or you highest desired goal? * What is an Arahat loyal to? Has loyality, then, an end (no more required)? Maybe you try to give a loyal answer. Much joy and inside in giving a benefical answer, at least for youself.

2017/08/19

Generally Buddhist take refuge in Buddha, Dhamma, and Sangha. In that regard, I would say I support Sangha. Because they protect Dhamma. Which helps to attain Buddhahood. Generally, Sangha means Ariya Sangha. But here I meant any person who can guide me in the right direction.

If I ever knew , I forgotten many times before remembering. I forget and get caught up into the world. I suppose I am loyal to the world but more and more I am loyal to the naked truth of the Dhamma. On the frontlines of the world is were learning is, right in the defilement! I don't know if I need a monastery. I don't always need to aim for a tranquil state of mind.

Ven. Members of the Sangha (of Bhikkhus), Ven. fellows, valued Upasaka and Upasika, dear readers and interesed, This question is intellectual and theoretical, or literary, but also/merely inviting self-reflection; but of course it can also be answered with words of the Buddha in "should-form" (in the later case, exchange "you" with "a serious follower of the Buddha, a person in general, to find peace", how ever you wish and feel obligated or loyal to). So, about loyalty, * In regard of what are you able to claim being loyal? * Whom are you loyal to? * How far goes your loyalty? * Where and how does it end? Also, finally: * What must one be loyal to, to find, to reach the highest aim all Buddha-following seekers are after? * What are you loyal and/or are you not loyal to, so that you have't found final peace already, or you highest desired goal? * What is an Arahat loyal to? Has loyality, then, an end (no more required)? Maybe you try to give a loyal answer. Much joy and inside in giving a benefical answer, at least for youself.

2017/08/19

> > **self-reflecting...** > > > *Atta-manasikara-self-reflecting*. > > **In regard of what are... able to claim being loyal?** > > > Anatta ([SN 22.59](http://www.accesstoinsight.org/tipitaka/sn/sn22/sn22.059.mend.html)). Sunnata ([SN 35.85](http://www.accesstoinsight.org/tipitaka/sn/sn35/sn35.085.than.html)). Lokuttara dhamma ([SN 20.7](http://www.accesstoinsight.org/tipitaka/sn/sn20/sn20.007.than.html)). > > **Whom are... loyal to?** > > > The ignorance-created idea of "who" is properly negated in [SN 12.12](http://www.accesstoinsight.org/tipitaka/sn/sn12/sn12.012.nypo.html) > > **How far goes... loyalty?** > > > Anatta. Sunnata. Nibbana. Vimutti > > **Where and how does it end?** > > > [MN 29](https://suttacentral.net/en/mn29) & 30 state the holy life has *ceto-vimutti* (liberation of mind) as its final end (*pariyosāna*).

If I ever knew , I forgotten many times before remembering. I forget and get caught up into the world. I suppose I am loyal to the world but more and more I am loyal to the naked truth of the Dhamma. On the frontlines of the world is were learning is, right in the defilement! I don't know if I need a monastery. I don't always need to aim for a tranquil state of mind.

Ven. Members of the Sangha (of Bhikkhus), Ven. fellows, valued Upasaka and Upasika, dear readers and interesed, This question is intellectual and theoretical, or literary, but also/merely inviting self-reflection; but of course it can also be answered with words of the Buddha in "should-form" (in the later case, exchange "you" with "a serious follower of the Buddha, a person in general, to find peace", how ever you wish and feel obligated or loyal to). So, about loyalty, * In regard of what are you able to claim being loyal? * Whom are you loyal to? * How far goes your loyalty? * Where and how does it end? Also, finally: * What must one be loyal to, to find, to reach the highest aim all Buddha-following seekers are after? * What are you loyal and/or are you not loyal to, so that you have't found final peace already, or you highest desired goal? * What is an Arahat loyal to? Has loyality, then, an end (no more required)? Maybe you try to give a loyal answer. Much joy and inside in giving a benefical answer, at least for youself.

2017/08/19

Everyone who can give me the whole 6 benefits: 1)own benefit, 2)other benefits, 3)public benefit, 4)this life benefit, 5)next life benefit, and 6)nibbāna benefit. **Everyone** mean I try listen everyone. **The whole** mean I trust in no separate benefits. Infinity, until I will parinibbāna.

If I ever knew , I forgotten many times before remembering. I forget and get caught up into the world. I suppose I am loyal to the world but more and more I am loyal to the naked truth of the Dhamma. On the frontlines of the world is were learning is, right in the defilement! I don't know if I need a monastery. I don't always need to aim for a tranquil state of mind.

I am designing a simple 16 bit adder circuit in Digilent's Xilinx Spartan 6 FPGA. The Verilog design accepts two 16 bit inputs A and B and returns the 16 bit sum C = A+B. I am ignoring carry in and carry out. I want to send A and B from PC to FPGA using serial port as well send the sum C from FPGA to PC using serial port. I am not sure how to do this. I have googled but could not find something simple. Can i send A and B as decimals like A = 5, B = 3? or do i have to send it as ASCII? How do i distinguish A from B if A = 255 and B = 512 (i.e., both A and B have multiple digits). How does FPGA deal with ASCII. Some pointers or explanations will be really appreciated

2014/10/14

I would suggest either writing a UART module from scratch or finding one online. Then all you would need to do is write a wrapper that interfaces the UART to your registers. Here is one possible open source Verilog UART module that I wrote a while ago: <https://github.com/alexforencich/verilog-uart> This particular module uses the AXI stream interface, so it should not be very difficult to interface with your design. The AXI stream interface has three signals: tdata, tvalid, and tready. tvalid indicates that tdata contains valid data, and tready indicates that the sink is ready to receive the data. Data bytes are transferred only when tvalid and tready are both high. What I would recommend doing is defining a simple serial protocol that supports some basic framing. Say, to send A and B to the FPGA, you would send some sort of start indication (e.g. 0 or perhaps ASCII S or W), then the MSB of A, then the LSB of A, then the MSB of B, then the LSB of B. Then you can write a state machine that looks for the start indication, then loads the next four bytes into the appropriate registers. Then once the operation is complete, the state machine can send the result back to the computer. If you want to use ASCII instead of binary, you can do that too, but it's a bit more complicated to convert everything. I would recommend using hex if you want something human readable as it is much easier to divide by 16 than it is to divide by 10 (bit shift/bit slice instead of an actual division operation).

Ken Chapman (Xilinx employee) provides a simple and lightweight UART module as an external I/O device for his 8-bit processor (PicoBlaze). These UART modules (RX and TX) can be used on every Xilinx FPGA. Each module has a simple byte interface. This processor and it's periphery comes with a lot of documentation and examples :) => Search "PicoBlaze" on the Xilinx website. For your requested transmission of 2x 16 bit, you will need a simple statemachine which receives 4 bytes in 4 cycles e.g. coded as A0, A1, B0, B1. I would advice to send all bytes already binary encoded to the FPGA, so there is no ascii or BCD to binary decoding need.

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

VirtualBox can run VMWare VMDK images, which gives you a lot more options. (There are also ways to convert VMDK virtual hard drives to native VirtualBox format.) You can almost certainly find what you want here: <http://www.vmware.com/appliances/> Why not intstall the VirtualBox support software and just use shared folders and the console to develop in Linux, instead of ssh?

[Linux from Scratch](http://www.linuxfromscratch.org/) I just finished doing just what you want to do, to have a base very minimal Linux OS for dev work. I did the install by hand because I wanted to, but they have automated installers for it. Final disk use is around 700MB. Normally I'd just use [Gentoo](http://www.gentoo.org/) and call it a day, but I wanted even more minimal, and it worked well.

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

Turnkey linux has a number of preconfigured server images that will work in virtualbox <http://www.turnkeylinux.org/> <http://www.turnkeylinux.org/docs/installation-appliances-virtualbox>

VirtualBox can run VMWare VMDK images, which gives you a lot more options. (There are also ways to convert VMDK virtual hard drives to native VirtualBox format.) You can almost certainly find what you want here: <http://www.vmware.com/appliances/> Why not intstall the VirtualBox support software and just use shared folders and the console to develop in Linux, instead of ssh?

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

> > If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking. > > > I don't believe you'll easily find anything prebuilt that matches your requirements. I'd suggest [Ubuntu Server](http://www.ubuntu.com/server) to build your own. Install it without anything extra, then run aptitude to get these packages: * **openssh-server** — SSH server * **screen** or **tmux** — per your choice * **vim** or another editor — the vim-tiny package omits many features * **gcc**, **g++**, **gdb**, etc. * **make**, **automake**, **autoconf**, etc. * **manpages-dev**, **manpages-posix-dev** — useful * **mercurial**, **git-core**, **bzr** — you'll want at least one Plus any other packages you want. The list above should already meet and exceed the requirements you've mentioned. (Use "/" in aptitude to search using a regex, "n" and "N" for the next and previous matches. Use "+" to mark for installation, which also marks dependencies, then "g" to preview the selection and "g" on the preview to download and install. You'll need to run aptitude via sudo.)

If I understand your question correctly, you are wanting a VBox image that is a pre-made linux distro that you will run, but ssh into to communicate, to do some non-gui development? But you do not have a lot of RAM? I guess my first concern is that even with just running a CLI interface, won't VBox take up your RAM anyways? Do you have enough to run vbox in the first place (which I believe is the amount needed for you host system + the amount needed for your virtual instance)? Second, though no pre-made distros come to mind (bearing, I have done no research), it seems given your unique circumstances it might be to your benefit to download something like [Debian-minimal](http://www.debian.org/distrib/netinst) and install the core system and the utilities you need on top of that. It will require a one-time setup, but once it's done, you can just reuse your own custom image. At this point in linux's development, I am all for package-managed systems, even when space is a concern. I know this doesn't directly answer the question at hand, but I hope it helps a little. Cheers!

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

Turnkey linux has a number of preconfigured server images that will work in virtualbox <http://www.turnkeylinux.org/> <http://www.turnkeylinux.org/docs/installation-appliances-virtualbox>

[Linux from Scratch](http://www.linuxfromscratch.org/) I just finished doing just what you want to do, to have a base very minimal Linux OS for dev work. I did the install by hand because I wanted to, but they have automated installers for it. Final disk use is around 700MB. Normally I'd just use [Gentoo](http://www.gentoo.org/) and call it a day, but I wanted even more minimal, and it worked well.

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

VirtualBox can run VMWare VMDK images, which gives you a lot more options. (There are also ways to convert VMDK virtual hard drives to native VirtualBox format.) You can almost certainly find what you want here: <http://www.vmware.com/appliances/> Why not intstall the VirtualBox support software and just use shared folders and the console to develop in Linux, instead of ssh?

If I understand your question correctly, you are wanting a VBox image that is a pre-made linux distro that you will run, but ssh into to communicate, to do some non-gui development? But you do not have a lot of RAM? I guess my first concern is that even with just running a CLI interface, won't VBox take up your RAM anyways? Do you have enough to run vbox in the first place (which I believe is the amount needed for you host system + the amount needed for your virtual instance)? Second, though no pre-made distros come to mind (bearing, I have done no research), it seems given your unique circumstances it might be to your benefit to download something like [Debian-minimal](http://www.debian.org/distrib/netinst) and install the core system and the utilities you need on top of that. It will require a one-time setup, but once it's done, you can just reuse your own custom image. At this point in linux's development, I am all for package-managed systems, even when space is a concern. I know this doesn't directly answer the question at hand, but I hope it helps a little. Cheers!

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

You can try [virtualboximages.com](http://virtualboximages.com/) - they have plenty of distros, but they charge a token sum ($.99-3.99) per download - which, i assume, goes to cover their bandwidth costs.

[Linux from Scratch](http://www.linuxfromscratch.org/) I just finished doing just what you want to do, to have a base very minimal Linux OS for dev work. I did the install by hand because I wanted to, but they have automated installers for it. Final disk use is around 700MB. Normally I'd just use [Gentoo](http://www.gentoo.org/) and call it a day, but I wanted even more minimal, and it worked well.

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

You can try [virtualboximages.com](http://virtualboximages.com/) - they have plenty of distros, but they charge a token sum ($.99-3.99) per download - which, i assume, goes to cover their bandwidth costs.

If I understand your question correctly, you are wanting a VBox image that is a pre-made linux distro that you will run, but ssh into to communicate, to do some non-gui development? But you do not have a lot of RAM? I guess my first concern is that even with just running a CLI interface, won't VBox take up your RAM anyways? Do you have enough to run vbox in the first place (which I believe is the amount needed for you host system + the amount needed for your virtual instance)? Second, though no pre-made distros come to mind (bearing, I have done no research), it seems given your unique circumstances it might be to your benefit to download something like [Debian-minimal](http://www.debian.org/distrib/netinst) and install the core system and the utilities you need on top of that. It will require a one-time setup, but once it's done, you can just reuse your own custom image. At this point in linux's development, I am all for package-managed systems, even when space is a concern. I know this doesn't directly answer the question at hand, but I hope it helps a little. Cheers!

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

Turnkey linux has a number of preconfigured server images that will work in virtualbox <http://www.turnkeylinux.org/> <http://www.turnkeylinux.org/docs/installation-appliances-virtualbox>

You can try [virtualboximages.com](http://virtualboximages.com/) - they have plenty of distros, but they charge a token sum ($.99-3.99) per download - which, i assume, goes to cover their bandwidth costs.

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

Turnkey linux has a number of preconfigured server images that will work in virtualbox <http://www.turnkeylinux.org/> <http://www.turnkeylinux.org/docs/installation-appliances-virtualbox>

If I understand your question correctly, you are wanting a VBox image that is a pre-made linux distro that you will run, but ssh into to communicate, to do some non-gui development? But you do not have a lot of RAM? I guess my first concern is that even with just running a CLI interface, won't VBox take up your RAM anyways? Do you have enough to run vbox in the first place (which I believe is the amount needed for you host system + the amount needed for your virtual instance)? Second, though no pre-made distros come to mind (bearing, I have done no research), it seems given your unique circumstances it might be to your benefit to download something like [Debian-minimal](http://www.debian.org/distrib/netinst) and install the core system and the utilities you need on top of that. It will require a one-time setup, but once it's done, you can just reuse your own custom image. At this point in linux's development, I am all for package-managed systems, even when space is a concern. I know this doesn't directly answer the question at hand, but I hope it helps a little. Cheers!

Long story short, I am trying to find a pre-made Linux development environment for VirtualBox so I don't have to worry about installing a distro, getting all packages installed, and so forth, since last time I did that it took me a full day to get something that remotely did what I wanted. I've got a few semi-weird requirements, which doesn't help... * Preferably SSH only, no slick GUI etc installed, as the laptop host is not endowed with a lot of HD space or RAM. (Which is also why I cannot dual boot, so please do not suggest that option.) I intend to do pretty much all using PuTTy, as it is a workflow I'm already accustomed to. * Basic gcc, automake etc are all I need. I really don't care about the distro itself, or the packaging system, or anything else really. I just want to download it, set up some mounts, and be ready to go. If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking.

2010/10/15

Turnkey linux has a number of preconfigured server images that will work in virtualbox <http://www.turnkeylinux.org/> <http://www.turnkeylinux.org/docs/installation-appliances-virtualbox>

> > If such a thing doesn't exist, I'd still appreciate suggestions of the best distro to use that will get me to my requirements fastest without a lot of tweaking. > > > I don't believe you'll easily find anything prebuilt that matches your requirements. I'd suggest [Ubuntu Server](http://www.ubuntu.com/server) to build your own. Install it without anything extra, then run aptitude to get these packages: * **openssh-server** — SSH server * **screen** or **tmux** — per your choice * **vim** or another editor — the vim-tiny package omits many features * **gcc**, **g++**, **gdb**, etc. * **make**, **automake**, **autoconf**, etc. * **manpages-dev**, **manpages-posix-dev** — useful * **mercurial**, **git-core**, **bzr** — you'll want at least one Plus any other packages you want. The list above should already meet and exceed the requirements you've mentioned. (Use "/" in aptitude to search using a regex, "n" and "N" for the next and previous matches. Use "+" to mark for installation, which also marks dependencies, then "g" to preview the selection and "g" on the preview to download and install. You'll need to run aptitude via sudo.)

I've read everywhere that you can't have more than 4 partitions because of GPT on intel Macs. But what happens if you make more than 4? On my iMac I have EFI, Macintosh HD, Windows, Linux, and Linux swap partitions and I am able to boot from all three operating systems with rEFIt. So, I have 5 partitions, so why does it work? I made the partitions with Snow Leopard's Disk Utility by the way.

2010/07/05

Intel-based Macs use the GUID Partition Table (GPT) by default. GPT in turn supports up to 128 partitions by default (that value can be increased if necessary, although most partition tools don't enable you to do so). Thus, there's no problem with having more than four partitions on an Intel-based Mac. The limitation you've heard about is a distortion of the limitation on hybrid MBRs, which are a dangerous and standards-violating hybridization of GPT with the older Master Boot Record (MBR) partitioning system used on most PCs. In a hybrid MBR, up to *three* of the GPT's partitions are duplicated in an MBR data structure. MBR is limited to four *primary* partitions, and in a hybrid MBR the fourth primary partition is occupied by a special partition that identifies the disk as being a GPT disk. This fourth partition is often mistaken for an MBR-side duplicate of the EFI System Partition (ESP) that's present on most GPT disks, but it's not that. Apple uses hybrid MBRs to enable Windows to dual-boot with OS X on Macs. Windows favors the MBR data structures, so it sees the disk as being an MBR disk, whereas OS X favors GPT data structures, so it sees the disk as being a GPT disk. (Linux, like OS X, sees a hybrid MBR as a GPT disk.) A hybrid MBR doesn't limit the number of GPT partitions you may have, but it does limit the number of partitions that the Windows installation can see, to no more than three. Note that extended partitions and Extended Boot Records (EBRs) have nothing to do with hybrid MBRs -- or at least, they shouldn't! In the MBR scheme, extended partitions serve as placeholders for logical partitions, which are defined by EBRs. Using this scheme, an MBR disk can support a huge number of partitions -- theoretically about half as many as there are sectors on the disk, although practical limits are much lower than that. Disks with hybrid MBRs don't use extended partitions, though, because maintaining consistency between the GPT and MBR sides of the disk -- already a challenging enough task with regular hybrid MBRs -- would become much tougher.

It's not because of GPT. It's because of MBR. There are many problems with MBR and GPT compatibility which limit partition number to 4. If you aren't using MBR at all, which could be happening, everything should work fine. New versions of windows support GPT, so there is a chance that MBR isn't being used. If you are using MBR, then I have no idea what's happening.

I've read everywhere that you can't have more than 4 partitions because of GPT on intel Macs. But what happens if you make more than 4? On my iMac I have EFI, Macintosh HD, Windows, Linux, and Linux swap partitions and I am able to boot from all three operating systems with rEFIt. So, I have 5 partitions, so why does it work? I made the partitions with Snow Leopard's Disk Utility by the way.

2010/07/05

Intel-based Macs use the GUID Partition Table (GPT) by default. GPT in turn supports up to 128 partitions by default (that value can be increased if necessary, although most partition tools don't enable you to do so). Thus, there's no problem with having more than four partitions on an Intel-based Mac. The limitation you've heard about is a distortion of the limitation on hybrid MBRs, which are a dangerous and standards-violating hybridization of GPT with the older Master Boot Record (MBR) partitioning system used on most PCs. In a hybrid MBR, up to *three* of the GPT's partitions are duplicated in an MBR data structure. MBR is limited to four *primary* partitions, and in a hybrid MBR the fourth primary partition is occupied by a special partition that identifies the disk as being a GPT disk. This fourth partition is often mistaken for an MBR-side duplicate of the EFI System Partition (ESP) that's present on most GPT disks, but it's not that. Apple uses hybrid MBRs to enable Windows to dual-boot with OS X on Macs. Windows favors the MBR data structures, so it sees the disk as being an MBR disk, whereas OS X favors GPT data structures, so it sees the disk as being a GPT disk. (Linux, like OS X, sees a hybrid MBR as a GPT disk.) A hybrid MBR doesn't limit the number of GPT partitions you may have, but it does limit the number of partitions that the Windows installation can see, to no more than three. Note that extended partitions and Extended Boot Records (EBRs) have nothing to do with hybrid MBRs -- or at least, they shouldn't! In the MBR scheme, extended partitions serve as placeholders for logical partitions, which are defined by EBRs. Using this scheme, an MBR disk can support a huge number of partitions -- theoretically about half as many as there are sectors on the disk, although practical limits are much lower than that. Disks with hybrid MBRs don't use extended partitions, though, because maintaining consistency between the GPT and MBR sides of the disk -- already a challenging enough task with regular hybrid MBRs -- would become much tougher.

I found [this article](http://www.rickycampbell.com/the-intel-mac-partitioning-system-efi-and-gpt/) which explains the partitioning scheme. Essentially, if I understand the article correctly, there is an “Extended Boot Record” on one of the partitions on one of the primary four partitions which describes the extra "extended" partitions. Some legacy operating systems cannot see them. Linux can see the extended partitions after the kernel is loaded. I presume you're able to boot properly because the Linux partition is on one of the four "primary" partitions.

I've read everywhere that you can't have more than 4 partitions because of GPT on intel Macs. But what happens if you make more than 4? On my iMac I have EFI, Macintosh HD, Windows, Linux, and Linux swap partitions and I am able to boot from all three operating systems with rEFIt. So, I have 5 partitions, so why does it work? I made the partitions with Snow Leopard's Disk Utility by the way.

2010/07/05

Intel-based Macs use the GUID Partition Table (GPT) by default. GPT in turn supports up to 128 partitions by default (that value can be increased if necessary, although most partition tools don't enable you to do so). Thus, there's no problem with having more than four partitions on an Intel-based Mac. The limitation you've heard about is a distortion of the limitation on hybrid MBRs, which are a dangerous and standards-violating hybridization of GPT with the older Master Boot Record (MBR) partitioning system used on most PCs. In a hybrid MBR, up to *three* of the GPT's partitions are duplicated in an MBR data structure. MBR is limited to four *primary* partitions, and in a hybrid MBR the fourth primary partition is occupied by a special partition that identifies the disk as being a GPT disk. This fourth partition is often mistaken for an MBR-side duplicate of the EFI System Partition (ESP) that's present on most GPT disks, but it's not that. Apple uses hybrid MBRs to enable Windows to dual-boot with OS X on Macs. Windows favors the MBR data structures, so it sees the disk as being an MBR disk, whereas OS X favors GPT data structures, so it sees the disk as being a GPT disk. (Linux, like OS X, sees a hybrid MBR as a GPT disk.) A hybrid MBR doesn't limit the number of GPT partitions you may have, but it does limit the number of partitions that the Windows installation can see, to no more than three. Note that extended partitions and Extended Boot Records (EBRs) have nothing to do with hybrid MBRs -- or at least, they shouldn't! In the MBR scheme, extended partitions serve as placeholders for logical partitions, which are defined by EBRs. Using this scheme, an MBR disk can support a huge number of partitions -- theoretically about half as many as there are sectors on the disk, although practical limits are much lower than that. Disks with hybrid MBRs don't use extended partitions, though, because maintaining consistency between the GPT and MBR sides of the disk -- already a challenging enough task with regular hybrid MBRs -- would become much tougher.

While Windows 7 x86\_64 supports GPT, it only supports UEFI or BIOS. Windows 7 does not support "Apple EFI" which is an amalgam of Intel EFI 1.1, and some bits of Apple's own and some bits of UEFI 2.x. It is not strictly a UEFI 2.x implementation which is what all non-Mac OS operating systems require. So this in effect limits Bootcamp (or rEFIt) users to using the CSM (BIOS), and thus require an MBR since BIOS only understands MBRs. MBR=4 partitions, but 1 of which is reserved when it's a hybrid MBR, as mentioned it's to indicate that the entire drive is GPT so that you have some warning from an MBR-only aware partitioning application. So in effect hybrid MBR is limited to 3 partitions. In theory you can choose which 3 of the potentially 128 GPT partitions you want exported to the hybrid MBR. But most tools don't let you do this. The only one I know that does is gdisk. I agree that hybrid MBR is standards-violating and is a bastardization. My preference would be for Apple to issue a fully standards compliant UEFI 2.x firmware for all machines sold in the past few years. In particular it's annoying as balls that brand new machines sold today still are based on Intel EFI 1.1, not UEFI 2.x.

Usually one constructs a graph and then asks questions about the adjacency matrix's (or some close relative like the [Laplacian](http://en.wikipedia.org/wiki/Laplacian_matrix)) eigenvalue decomposition (also called the [spectra of a graph](http://en.wikipedia.org/wiki/Spectral_graph_theory)). But what about the reverse problem? Given $n$ eigenvalues, can one (efficiently) find a graph that has this spectra? I suspect that in general this is hard to do (and might be equivalent to GI) but what if you relax some of the conditions a bit? What if you make conditions that there is no multiplicity of eigenvalues? What about allowing graphs that have "close" spectra by some distance metric? Any references or ideas would be welcome. **EDIT**: As Suresh points out, if you allow undirected weighted graphs with self loops, this problem becomes pretty trivial. I was hoping to get answers on the set of undirected, unweighted simple graphs but I would be happy with simple unweighted directed graphs as well.

2010/12/11

Cvetcovic et all in Section 3.3 of "Recent results in the theory of graph spectra" [goes over](http://yaroslavvb.com/upload/save/cstheory-cvetcovic-spectra.pdf) algorithms for constructing graphs given spectrum in some special cases

One other obstacle in defining your question is that the are isospectral (same eigenvalues) but non-isomorphic graphs. So given a list of eigenvalues in such a case, which graph do you want? Maybe you just want an algorithm to return one random element of the set of such non-isomorphic graphs?

Usually one constructs a graph and then asks questions about the adjacency matrix's (or some close relative like the [Laplacian](http://en.wikipedia.org/wiki/Laplacian_matrix)) eigenvalue decomposition (also called the [spectra of a graph](http://en.wikipedia.org/wiki/Spectral_graph_theory)). But what about the reverse problem? Given $n$ eigenvalues, can one (efficiently) find a graph that has this spectra? I suspect that in general this is hard to do (and might be equivalent to GI) but what if you relax some of the conditions a bit? What if you make conditions that there is no multiplicity of eigenvalues? What about allowing graphs that have "close" spectra by some distance metric? Any references or ideas would be welcome. **EDIT**: As Suresh points out, if you allow undirected weighted graphs with self loops, this problem becomes pretty trivial. I was hoping to get answers on the set of undirected, unweighted simple graphs but I would be happy with simple unweighted directed graphs as well.

2010/12/11

Even asking whether a graph with a given spectrum exists is a tough question. This is witnessed by the [open problem](https://en.wikipedia.org/wiki/Moore_graph) of determining whether there exists a graph of girth 5 diameter 2 and order 3250 whose spectrum (if it exists) is known.

One other obstacle in defining your question is that the are isospectral (same eigenvalues) but non-isomorphic graphs. So given a list of eigenvalues in such a case, which graph do you want? Maybe you just want an algorithm to return one random element of the set of such non-isomorphic graphs?

I have a Vauxhall Corsa "C" (2003)(EU) which is a standard 16v 1.2 petrol, the engine is chain driven and sits at 50k. I don't think it was well looked after (service wise) in previous lifes. Recently it sounds a bit more like a diesel when cold, but not too loud -and a common issue with the Corsa, however at the moment when the engine hits ~80-90 degrees it starts to "tick" loudly, to the point I can hear it in the cabin. Observations ------------ * When I accelerate the ticking gets faster, if I depress the clutch and throttle, it gets faster, once the ticking starts it doesn't go away till the engine is cool. It will be audible when Idle and parked too. * "slight" loss of power, hardly noticable not anything serious though, and no misfiring or overheating. Taking ~10min to get to temperature in 20-35mph traffic. * prior to the ticking starting, when moving off in 1st, it sounds a bit "throaty" like an exhaust rattle but being in the car I can't work out where this is. No noticable smoke or anything from the exhaust as far as I can tell. * Plenty (almost full) oil and coolant. Oil is fairly dark and theres a bit of gunk around from a bit of moisture (Not HGF, a common mistake when diagnosing things with this engine!) It's from doing short journeys and the build up of condensation. Multiple owner forums confirmed this. Conditions: ----------- * Engine is warm and has typically being driven for ~15min * Starting in cold weather ~0-4 degrees c. * Used for short journeys (~6 miles each way) Notes ----- * When I start the engine, it goes to about 2K RPM, sounds rough for about half a second then it's fine. This has been diagnosed as normal though for this car. * When I move off when the engine is cold, I get slight revving for the first couple of seconds, again this looks fairly normal for this engine looking on owner forums. Thoughts? --------- My first though would be that it could be the oil, as it gets warmer the vicosity changes and it's not as effective? I thought it could be the timing chain, but it doesn't happen from cold. My third thought would be tappets. I'm not brilliant with mechanics so not a clue. I recently put some Wynns Hydraulic Lifter fluid into the oil with no success. Sounds the same. The car *did* have some gasket work done and I'd be suprised if they did a full flush of fluids. Tried to include as much info as I thought was relevant, but if theres anything else let me know EDIT: Conclusion? ----------------- It got a bit worse when the engine got to 85-86ish and the ticking was getting louder, it was also happening on *some* cold starts for 30seconds. It went in for a full check hoping it was the Ex. Man. gasket, and they found that the **timing chain** had a small amount of slack on it, and on an unrelated note the oil pressure switch needed replacing, the old oil in the system probably didn't help the longevity of the chain. **Update** Months on and the car still does the same thing. I have checked the oil, changed it, had the chain done and still after a 15-20 min drive it will tick away to itself. I think its the tappets, and occasionally the sound will go away again before returning later in my journey.

2015/02/06

Corsa's (actually Vauxhalls/Opels in general) sometimes have noisy hydraulic lifters. I drove mine for 60 thousand kilometers with this noise before I upgraded to a Turbo Coupe and gave the car to my mom. No problems so far. The Coupe had the same stupid tick. Again, I drove it for 30 thousand kilometers before trading it in for a Subaru Forester XT. Only problem I had with it was the turbo. Nothing wrong with the lifters or valves. If you click on my profile and check out my questions, you'll actually see one [where I posted a video of a loud ticking coming from my engine bay in the Coupe](https://mechanics.stackexchange.com/questions/7539/astra-turbo-sounding-like-a-diesel). I assume this is the same sound you hear. (The video makes it out to be a bit harsher than it actually is). These noises are usually a valve that's stuck for whatever reason. It's something you can have looked at when the car is in for a major service or something, but it's my experience, and there is considerable consensus, that it's not a serious problem that you need to immediately spend your Valentine's Day money on.

I think this is a dual overhead camshaft engine. In that case, one of your lifters or springs may be going bad. This would slow down the return of the lifter as the cam moves away from it. that would create a momentary space and then a click when the lifter contacts the cam again. Just a guess. Your mechanic probably has a modified stethoscope that he (or she) can use to pinpoint which cylinder is clicking. I had a rocker arm in a 351W ford that was slightly loose because it was not thick enough at the mounting hole and creating a click. Added a .020" SS washer and fixed that. No rockers (probably) in the DOHC.

I have a Vauxhall Corsa "C" (2003)(EU) which is a standard 16v 1.2 petrol, the engine is chain driven and sits at 50k. I don't think it was well looked after (service wise) in previous lifes. Recently it sounds a bit more like a diesel when cold, but not too loud -and a common issue with the Corsa, however at the moment when the engine hits ~80-90 degrees it starts to "tick" loudly, to the point I can hear it in the cabin. Observations ------------ * When I accelerate the ticking gets faster, if I depress the clutch and throttle, it gets faster, once the ticking starts it doesn't go away till the engine is cool. It will be audible when Idle and parked too. * "slight" loss of power, hardly noticable not anything serious though, and no misfiring or overheating. Taking ~10min to get to temperature in 20-35mph traffic. * prior to the ticking starting, when moving off in 1st, it sounds a bit "throaty" like an exhaust rattle but being in the car I can't work out where this is. No noticable smoke or anything from the exhaust as far as I can tell. * Plenty (almost full) oil and coolant. Oil is fairly dark and theres a bit of gunk around from a bit of moisture (Not HGF, a common mistake when diagnosing things with this engine!) It's from doing short journeys and the build up of condensation. Multiple owner forums confirmed this. Conditions: ----------- * Engine is warm and has typically being driven for ~15min * Starting in cold weather ~0-4 degrees c. * Used for short journeys (~6 miles each way) Notes ----- * When I start the engine, it goes to about 2K RPM, sounds rough for about half a second then it's fine. This has been diagnosed as normal though for this car. * When I move off when the engine is cold, I get slight revving for the first couple of seconds, again this looks fairly normal for this engine looking on owner forums. Thoughts? --------- My first though would be that it could be the oil, as it gets warmer the vicosity changes and it's not as effective? I thought it could be the timing chain, but it doesn't happen from cold. My third thought would be tappets. I'm not brilliant with mechanics so not a clue. I recently put some Wynns Hydraulic Lifter fluid into the oil with no success. Sounds the same. The car *did* have some gasket work done and I'd be suprised if they did a full flush of fluids. Tried to include as much info as I thought was relevant, but if theres anything else let me know EDIT: Conclusion? ----------------- It got a bit worse when the engine got to 85-86ish and the ticking was getting louder, it was also happening on *some* cold starts for 30seconds. It went in for a full check hoping it was the Ex. Man. gasket, and they found that the **timing chain** had a small amount of slack on it, and on an unrelated note the oil pressure switch needed replacing, the old oil in the system probably didn't help the longevity of the chain. **Update** Months on and the car still does the same thing. I have checked the oil, changed it, had the chain done and still after a 15-20 min drive it will tick away to itself. I think its the tappets, and occasionally the sound will go away again before returning later in my journey.

2015/02/06

I think this is a dual overhead camshaft engine. In that case, one of your lifters or springs may be going bad. This would slow down the return of the lifter as the cam moves away from it. that would create a momentary space and then a click when the lifter contacts the cam again. Just a guess. Your mechanic probably has a modified stethoscope that he (or she) can use to pinpoint which cylinder is clicking. I had a rocker arm in a 351W ford that was slightly loose because it was not thick enough at the mounting hole and creating a click. Added a .020" SS washer and fixed that. No rockers (probably) in the DOHC.

I'm going with a leaking exhaust manifold for the ticking. Try tightening the manifold bolts a little. They will have a torque specification but hand tight will have a noticeable affect on the ticking if the manifold is your problem. Go back later and torque them properly. As far as the car running at 2000 RPM on start up, that is normal for most modern cars. My 2011 Nissan starts at about 1500 and slowly ends at 900. -JMR(Auto Hobbyist / Diesel Mech. 8 Years)

I have a Vauxhall Corsa "C" (2003)(EU) which is a standard 16v 1.2 petrol, the engine is chain driven and sits at 50k. I don't think it was well looked after (service wise) in previous lifes. Recently it sounds a bit more like a diesel when cold, but not too loud -and a common issue with the Corsa, however at the moment when the engine hits ~80-90 degrees it starts to "tick" loudly, to the point I can hear it in the cabin. Observations ------------ * When I accelerate the ticking gets faster, if I depress the clutch and throttle, it gets faster, once the ticking starts it doesn't go away till the engine is cool. It will be audible when Idle and parked too. * "slight" loss of power, hardly noticable not anything serious though, and no misfiring or overheating. Taking ~10min to get to temperature in 20-35mph traffic. * prior to the ticking starting, when moving off in 1st, it sounds a bit "throaty" like an exhaust rattle but being in the car I can't work out where this is. No noticable smoke or anything from the exhaust as far as I can tell. * Plenty (almost full) oil and coolant. Oil is fairly dark and theres a bit of gunk around from a bit of moisture (Not HGF, a common mistake when diagnosing things with this engine!) It's from doing short journeys and the build up of condensation. Multiple owner forums confirmed this. Conditions: ----------- * Engine is warm and has typically being driven for ~15min * Starting in cold weather ~0-4 degrees c. * Used for short journeys (~6 miles each way) Notes ----- * When I start the engine, it goes to about 2K RPM, sounds rough for about half a second then it's fine. This has been diagnosed as normal though for this car. * When I move off when the engine is cold, I get slight revving for the first couple of seconds, again this looks fairly normal for this engine looking on owner forums. Thoughts? --------- My first though would be that it could be the oil, as it gets warmer the vicosity changes and it's not as effective? I thought it could be the timing chain, but it doesn't happen from cold. My third thought would be tappets. I'm not brilliant with mechanics so not a clue. I recently put some Wynns Hydraulic Lifter fluid into the oil with no success. Sounds the same. The car *did* have some gasket work done and I'd be suprised if they did a full flush of fluids. Tried to include as much info as I thought was relevant, but if theres anything else let me know EDIT: Conclusion? ----------------- It got a bit worse when the engine got to 85-86ish and the ticking was getting louder, it was also happening on *some* cold starts for 30seconds. It went in for a full check hoping it was the Ex. Man. gasket, and they found that the **timing chain** had a small amount of slack on it, and on an unrelated note the oil pressure switch needed replacing, the old oil in the system probably didn't help the longevity of the chain. **Update** Months on and the car still does the same thing. I have checked the oil, changed it, had the chain done and still after a 15-20 min drive it will tick away to itself. I think its the tappets, and occasionally the sound will go away again before returning later in my journey.

2015/02/06

Corsa's (actually Vauxhalls/Opels in general) sometimes have noisy hydraulic lifters. I drove mine for 60 thousand kilometers with this noise before I upgraded to a Turbo Coupe and gave the car to my mom. No problems so far. The Coupe had the same stupid tick. Again, I drove it for 30 thousand kilometers before trading it in for a Subaru Forester XT. Only problem I had with it was the turbo. Nothing wrong with the lifters or valves. If you click on my profile and check out my questions, you'll actually see one [where I posted a video of a loud ticking coming from my engine bay in the Coupe](https://mechanics.stackexchange.com/questions/7539/astra-turbo-sounding-like-a-diesel). I assume this is the same sound you hear. (The video makes it out to be a bit harsher than it actually is). These noises are usually a valve that's stuck for whatever reason. It's something you can have looked at when the car is in for a major service or something, but it's my experience, and there is considerable consensus, that it's not a serious problem that you need to immediately spend your Valentine's Day money on.

I'm going with a leaking exhaust manifold for the ticking. Try tightening the manifold bolts a little. They will have a torque specification but hand tight will have a noticeable affect on the ticking if the manifold is your problem. Go back later and torque them properly. As far as the car running at 2000 RPM on start up, that is normal for most modern cars. My 2011 Nissan starts at about 1500 and slowly ends at 900. -JMR(Auto Hobbyist / Diesel Mech. 8 Years)

I have a Vauxhall Corsa "C" (2003)(EU) which is a standard 16v 1.2 petrol, the engine is chain driven and sits at 50k. I don't think it was well looked after (service wise) in previous lifes. Recently it sounds a bit more like a diesel when cold, but not too loud -and a common issue with the Corsa, however at the moment when the engine hits ~80-90 degrees it starts to "tick" loudly, to the point I can hear it in the cabin. Observations ------------ * When I accelerate the ticking gets faster, if I depress the clutch and throttle, it gets faster, once the ticking starts it doesn't go away till the engine is cool. It will be audible when Idle and parked too. * "slight" loss of power, hardly noticable not anything serious though, and no misfiring or overheating. Taking ~10min to get to temperature in 20-35mph traffic. * prior to the ticking starting, when moving off in 1st, it sounds a bit "throaty" like an exhaust rattle but being in the car I can't work out where this is. No noticable smoke or anything from the exhaust as far as I can tell. * Plenty (almost full) oil and coolant. Oil is fairly dark and theres a bit of gunk around from a bit of moisture (Not HGF, a common mistake when diagnosing things with this engine!) It's from doing short journeys and the build up of condensation. Multiple owner forums confirmed this. Conditions: ----------- * Engine is warm and has typically being driven for ~15min * Starting in cold weather ~0-4 degrees c. * Used for short journeys (~6 miles each way) Notes ----- * When I start the engine, it goes to about 2K RPM, sounds rough for about half a second then it's fine. This has been diagnosed as normal though for this car. * When I move off when the engine is cold, I get slight revving for the first couple of seconds, again this looks fairly normal for this engine looking on owner forums. Thoughts? --------- My first though would be that it could be the oil, as it gets warmer the vicosity changes and it's not as effective? I thought it could be the timing chain, but it doesn't happen from cold. My third thought would be tappets. I'm not brilliant with mechanics so not a clue. I recently put some Wynns Hydraulic Lifter fluid into the oil with no success. Sounds the same. The car *did* have some gasket work done and I'd be suprised if they did a full flush of fluids. Tried to include as much info as I thought was relevant, but if theres anything else let me know EDIT: Conclusion? ----------------- It got a bit worse when the engine got to 85-86ish and the ticking was getting louder, it was also happening on *some* cold starts for 30seconds. It went in for a full check hoping it was the Ex. Man. gasket, and they found that the **timing chain** had a small amount of slack on it, and on an unrelated note the oil pressure switch needed replacing, the old oil in the system probably didn't help the longevity of the chain. **Update** Months on and the car still does the same thing. I have checked the oil, changed it, had the chain done and still after a 15-20 min drive it will tick away to itself. I think its the tappets, and occasionally the sound will go away again before returning later in my journey.

2015/02/06

Corsa's (actually Vauxhalls/Opels in general) sometimes have noisy hydraulic lifters. I drove mine for 60 thousand kilometers with this noise before I upgraded to a Turbo Coupe and gave the car to my mom. No problems so far. The Coupe had the same stupid tick. Again, I drove it for 30 thousand kilometers before trading it in for a Subaru Forester XT. Only problem I had with it was the turbo. Nothing wrong with the lifters or valves. If you click on my profile and check out my questions, you'll actually see one [where I posted a video of a loud ticking coming from my engine bay in the Coupe](https://mechanics.stackexchange.com/questions/7539/astra-turbo-sounding-like-a-diesel). I assume this is the same sound you hear. (The video makes it out to be a bit harsher than it actually is). These noises are usually a valve that's stuck for whatever reason. It's something you can have looked at when the car is in for a major service or something, but it's my experience, and there is considerable consensus, that it's not a serious problem that you need to immediately spend your Valentine's Day money on.

I had similar ticking noise when accelerating after changing my engine oil on Toyota Corolla 1NZE. I came to find out that the oil does not reach to the most parts of the engine and it was caused by a poor performance of the oil strainer. The strainer was blocked by some bad oil left overs. Get the straigner cleaned and everything will be OK.

I have a Vauxhall Corsa "C" (2003)(EU) which is a standard 16v 1.2 petrol, the engine is chain driven and sits at 50k. I don't think it was well looked after (service wise) in previous lifes. Recently it sounds a bit more like a diesel when cold, but not too loud -and a common issue with the Corsa, however at the moment when the engine hits ~80-90 degrees it starts to "tick" loudly, to the point I can hear it in the cabin. Observations ------------ * When I accelerate the ticking gets faster, if I depress the clutch and throttle, it gets faster, once the ticking starts it doesn't go away till the engine is cool. It will be audible when Idle and parked too. * "slight" loss of power, hardly noticable not anything serious though, and no misfiring or overheating. Taking ~10min to get to temperature in 20-35mph traffic. * prior to the ticking starting, when moving off in 1st, it sounds a bit "throaty" like an exhaust rattle but being in the car I can't work out where this is. No noticable smoke or anything from the exhaust as far as I can tell. * Plenty (almost full) oil and coolant. Oil is fairly dark and theres a bit of gunk around from a bit of moisture (Not HGF, a common mistake when diagnosing things with this engine!) It's from doing short journeys and the build up of condensation. Multiple owner forums confirmed this. Conditions: ----------- * Engine is warm and has typically being driven for ~15min * Starting in cold weather ~0-4 degrees c. * Used for short journeys (~6 miles each way) Notes ----- * When I start the engine, it goes to about 2K RPM, sounds rough for about half a second then it's fine. This has been diagnosed as normal though for this car. * When I move off when the engine is cold, I get slight revving for the first couple of seconds, again this looks fairly normal for this engine looking on owner forums. Thoughts? --------- My first though would be that it could be the oil, as it gets warmer the vicosity changes and it's not as effective? I thought it could be the timing chain, but it doesn't happen from cold. My third thought would be tappets. I'm not brilliant with mechanics so not a clue. I recently put some Wynns Hydraulic Lifter fluid into the oil with no success. Sounds the same. The car *did* have some gasket work done and I'd be suprised if they did a full flush of fluids. Tried to include as much info as I thought was relevant, but if theres anything else let me know EDIT: Conclusion? ----------------- It got a bit worse when the engine got to 85-86ish and the ticking was getting louder, it was also happening on *some* cold starts for 30seconds. It went in for a full check hoping it was the Ex. Man. gasket, and they found that the **timing chain** had a small amount of slack on it, and on an unrelated note the oil pressure switch needed replacing, the old oil in the system probably didn't help the longevity of the chain. **Update** Months on and the car still does the same thing. I have checked the oil, changed it, had the chain done and still after a 15-20 min drive it will tick away to itself. I think its the tappets, and occasionally the sound will go away again before returning later in my journey.

2015/02/06

I had similar ticking noise when accelerating after changing my engine oil on Toyota Corolla 1NZE. I came to find out that the oil does not reach to the most parts of the engine and it was caused by a poor performance of the oil strainer. The strainer was blocked by some bad oil left overs. Get the straigner cleaned and everything will be OK.

I'm going with a leaking exhaust manifold for the ticking. Try tightening the manifold bolts a little. They will have a torque specification but hand tight will have a noticeable affect on the ticking if the manifold is your problem. Go back later and torque them properly. As far as the car running at 2000 RPM on start up, that is normal for most modern cars. My 2011 Nissan starts at about 1500 and slowly ends at 900. -JMR(Auto Hobbyist / Diesel Mech. 8 Years)

I am working on [this code](https://github.com/nfmcclure/tensorflow_cookbook/blob/master/09_Recurrent_Neural_Networks/02_Implementing_RNN_for_Spam_Prediction/02_implementing_rnn.py) for spam detection using recurrent neural networks. Question 1. I am wondering whether this field (using RNNs for email spam detection) worths more researches or it is a closed research field. Question 2. What is the oldest published paper in this field? Quesiton 3. What are the pros and cons of using RNNs for email spam detection over other classification methods?

2017/06/10

There's a strong sentiment towards the idea that medical diagnosis is largely [Abductive Reasoning](https://en.wikipedia.org/wiki/Abductive_reasoning). See [this presentation](https://blogs.kent.ac.uk/jonw/files/2015/04/Aliseda2012.pdf) for additional details. One approach to automated abductive reasoning is [parsimonious covering theory](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2244953/). If you want a relatively in-depth look at all of this, check out the book [Abductive Inference Models for Diagnostic Problem-Solving](http://rads.stackoverflow.com/amzn/click/0387973435) . Expert systems have also been shown to work very well for medical diagnosis. As far back as the 1970's, systems like [MYCIN](https://en.wikipedia.org/wiki/Mycin) could beat human experts in terms of diagnosis and treatment plans. From what I can tell, there doesn't seem to be any reason in principle to think that AI/ML can't be used across the board for medical diagnosis, mental health or otherwise.

You may infer some mental problems using text processing. Any text written by an individual contains a lot of clues about their mental state, probably including any health problems. However, I believe pursuing this would prove to be difficult due to lack of ground truth. Not many people makes it public that they have mental problems and without a reliable database it is quite difficult to make research on the field.

I am working on [this code](https://github.com/nfmcclure/tensorflow_cookbook/blob/master/09_Recurrent_Neural_Networks/02_Implementing_RNN_for_Spam_Prediction/02_implementing_rnn.py) for spam detection using recurrent neural networks. Question 1. I am wondering whether this field (using RNNs for email spam detection) worths more researches or it is a closed research field. Question 2. What is the oldest published paper in this field? Quesiton 3. What are the pros and cons of using RNNs for email spam detection over other classification methods?

2017/06/10

There's a strong sentiment towards the idea that medical diagnosis is largely [Abductive Reasoning](https://en.wikipedia.org/wiki/Abductive_reasoning). See [this presentation](https://blogs.kent.ac.uk/jonw/files/2015/04/Aliseda2012.pdf) for additional details. One approach to automated abductive reasoning is [parsimonious covering theory](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2244953/). If you want a relatively in-depth look at all of this, check out the book [Abductive Inference Models for Diagnostic Problem-Solving](http://rads.stackoverflow.com/amzn/click/0387973435) . Expert systems have also been shown to work very well for medical diagnosis. As far back as the 1970's, systems like [MYCIN](https://en.wikipedia.org/wiki/Mycin) could beat human experts in terms of diagnosis and treatment plans. From what I can tell, there doesn't seem to be any reason in principle to think that AI/ML can't be used across the board for medical diagnosis, mental health or otherwise.

It's already being done, and apparently with very good results. See: [Predicting Risk of Suicide Attempts Over Time Through Machine Learning](http://journals.sagepub.com/doi/abs/10.1177/2167702617691560?journalCode=cpxa&), Walsh, Ribiero, Franklin Here is the abstract from the paper: > > Traditional approaches to the prediction of suicide attempts have limited the accuracy and scale of risk detection for these dangerous behaviors. We sought to overcome these limitations by applying machine learning to electronic health records within a large medical database. Participants were 5,167 adult patients with a claim code for self-injury (i.e., ICD-9, E95x); expert review of records determined that 3,250 patients made a suicide attempt (i.e., cases), and 1,917 patients engaged in self-injury that was nonsuicidal, accidental, or nonverifiable (i.e., controls). We developed machine learning algorithms that accurately predicted future suicide attempts (AUC = 0.84, precision = 0.79, recall = 0.95, Brier score = 0.14). Moreover, accuracy improved from 720 days to 7 days before the suicide attempt, and predictor importance shifted across time. These findings represent a step toward accurate and scalable risk detection and provide insight into how suicide attempt risk shifts over time. > > > This has gotten national press attention quite recently. Here are a couple of prior articles on the result and endeavor: [Artificial Intelligence is Learning to Predict and Prevent Suicide](https://www.wired.com/2017/03/artificial-intelligence-learning-predict-prevent-suicide/) (Wired) [Artificial intelligence can now predict suicide with remarkable accuracy](https://qz.com/1001968/artificial-intelligence-can-now-predict-suicide-with-remarkable-accuracy/) (Quartz) Is strongly suspect ML can be used in an array of application related to human mental health. For instance, it is entirely possible a app would be able to discern the mood swings in people with bipolar disorder based on activity or lack thereof.

I am working on [this code](https://github.com/nfmcclure/tensorflow_cookbook/blob/master/09_Recurrent_Neural_Networks/02_Implementing_RNN_for_Spam_Prediction/02_implementing_rnn.py) for spam detection using recurrent neural networks. Question 1. I am wondering whether this field (using RNNs for email spam detection) worths more researches or it is a closed research field. Question 2. What is the oldest published paper in this field? Quesiton 3. What are the pros and cons of using RNNs for email spam detection over other classification methods?

2017/06/10

It's already being done, and apparently with very good results. See: [Predicting Risk of Suicide Attempts Over Time Through Machine Learning](http://journals.sagepub.com/doi/abs/10.1177/2167702617691560?journalCode=cpxa&), Walsh, Ribiero, Franklin Here is the abstract from the paper: > > Traditional approaches to the prediction of suicide attempts have limited the accuracy and scale of risk detection for these dangerous behaviors. We sought to overcome these limitations by applying machine learning to electronic health records within a large medical database. Participants were 5,167 adult patients with a claim code for self-injury (i.e., ICD-9, E95x); expert review of records determined that 3,250 patients made a suicide attempt (i.e., cases), and 1,917 patients engaged in self-injury that was nonsuicidal, accidental, or nonverifiable (i.e., controls). We developed machine learning algorithms that accurately predicted future suicide attempts (AUC = 0.84, precision = 0.79, recall = 0.95, Brier score = 0.14). Moreover, accuracy improved from 720 days to 7 days before the suicide attempt, and predictor importance shifted across time. These findings represent a step toward accurate and scalable risk detection and provide insight into how suicide attempt risk shifts over time. > > > This has gotten national press attention quite recently. Here are a couple of prior articles on the result and endeavor: [Artificial Intelligence is Learning to Predict and Prevent Suicide](https://www.wired.com/2017/03/artificial-intelligence-learning-predict-prevent-suicide/) (Wired) [Artificial intelligence can now predict suicide with remarkable accuracy](https://qz.com/1001968/artificial-intelligence-can-now-predict-suicide-with-remarkable-accuracy/) (Quartz) Is strongly suspect ML can be used in an array of application related to human mental health. For instance, it is entirely possible a app would be able to discern the mood swings in people with bipolar disorder based on activity or lack thereof.

You may infer some mental problems using text processing. Any text written by an individual contains a lot of clues about their mental state, probably including any health problems. However, I believe pursuing this would prove to be difficult due to lack of ground truth. Not many people makes it public that they have mental problems and without a reliable database it is quite difficult to make research on the field.

Assuming that there are no known complete graph invariants in the spirit of [Harrsion's question](https://mathoverflow.net/questions/11631/complete-graph-invariants) that do not depend on any labelling (see [graph property](http://en.wikipedia.org/wiki/Graph_invariant) at Wikipedia), I wonder if there are graph invariants that are * almost complete, i.e. discriminating almost all finite graphs up to isomorphism * almost complete in a weaker sense, i.e. discriminating all finite graphs except for a small, but finite fraction (the smaller the fraction the greater the invariant's discriminating power) * probably complete, i.e. not proven to be incomplete yet (e.g. by counterexamples) Can anyone provide examples or references?

2010/04/22

I'm not sure I understand the question well enough to know whether this is a reasonable answer. But a standard observation concerning the graph isomorphism problem is that there is an approach that feels as though it almost works: work out the eigenvalues and eigenvectors of the adjacency matrix. There is a small technical problem that you can't do this exactly, and a more fundamental problem that if you have eigenvalues of multiplicity greater than 1 then you don't distinguish all graphs. But I would guess that almost all graphs have adjacency matrices with distinct eigenvalues. Can anyone confirm this? And if that is the case, does it answer your question?

Though painfully slow to compute, I've yet to find a counterexample to the first invariant proposed in [this](http://arxiv.org/abs/math/0503316) paper by Mehendale: a vector of integers counting, for each n, the number of labeled subgraphs of G which are trees on n vertices. Update: I've since discovered several counterexamples, perhaps the simplest being S3 U P2 U P2 and P4 U P3 U P1 both of which yield the signature (8,5,3,1)

Which is the real one? What is the other one? How did I get it? 

2014/09/14

TL;DR: keep the one that looks like a bag, it is now a universal app. The cart one was the iPad specific one. Looks like the one of the left (with the shopping cart icon) is the old one. Launching it now directs me to the new one, but it didn't when I was trying to figure this out before. Here's what I get when launching the old one now. EDIT-20140917: It appears that Apple may have had a separate iPad app that they have since deprecated and made their iPhone/iPod Touch app a universal app and that's why this app is now defunct.

The old iPad Store APP is obsolete and has been updated and replaced with this new one.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

Your compassion is admirable but is not unique. Other answers have provided more detailed answers about grading, content prep and presentation, students 'earning' their grade and that is all true. Responsibility vs Accountability -------------------------------- With my answer I want to add to the discussion the point that the economic situation of a student may seem dire it is still the students responsibility and not yours. Them making you aware of their situation (even innocently) can have the effect of shifting that responsibility unfairly to one who is not accountable for it (you). It's Not Life or Death, Until It Is ----------------------------------- When I was in college there was a professor who would comment about students complaining about difficult tests or mounds of homework by recalling how different it was in the 1970's when a male student failing out of college made them eligible for the draft with a likely outcome of being sent to Vietnam; which carried with it a very real risk to their health or lifespan. He wasn't saying that he failed people in order to send them to be drafted, he was saying the motivation of students facing such risks were much higher in those days. It wasn't him that decided who was born male, who would be drafted, who would be given a gun and sent out into the jungle. Guilt, party of 2 ----------------- Did the student confer with you before choosing their major? Did they ask your opinion about which school to attend? Did they check with you before taking out those loans? You are neither responsible nor accountable for their choices. Advice ------ While I would not ask you to stop feeling compassion, you should not allow their predicament to turn into guilt on your part. Should you turn your back on these students? Absolutely not, but what then? Either add something in your syllabus or to your first lecture something to effect of, "I know some of you may be relying on financial aid (loans, scholarships) which carry an eligibility component which may be affected by how well you do in this class. If that applies to you, I would suggest you get with me early in the semester so I can assist however I can to point you to resources to improve your chances of success in this class." To effect, you are saying "I can point you to water but I can't make you drink". That is the most that can be reasonably be asked of you since you can't do the work for them or make them attend class, etc.

I generally design my exams so that even a student with relatively limited understanding of the course can pass. I would really only consider a student who has achieved the highest grade (> 80%) to have a good understanding of the material, and I often see just-passing exams (50-60%) where I think that the student really had no idea, but somehow managed to scrape together enough marks to get over the line. Armed with that viewpoint, when I see a student who fails, I think they must have really deserved it! So I don't feel too bad about failing them; they need to put in some effort to fail my course!

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

**You are responsible for teaching the students to the best of your ability, and to judge their capacities to use what they have learned**. That judgment is made based on their grades. So you have several things to think about here. 1. **Are you teaching the best you can?** Teaching does not mean "downloading facts", as I'm sure you're aware. It means "transferring knowledge, skills, and attitudes". That "transfer" part is the important bit — transfer means that the student is able to reproduce and use what they've learned. Is your teaching enhancing this transfer? This is a tough nut to crack — how do you know? Are you planning your assessments so that you can really tease out the nuances and to see which students really understand, or are they just assessments because you need to assign a grade somehow? Your institution might have a center for teaching how to teach, and if you feel that you aren't teaching your best class then start there. Otherwise, lots of books and resources exist, which I'm sure we can all provide. 2. **Are you assessing fairly?** Fairly doesn't mean easily. It means that you are creating assessments that actually test understanding and that a student with reasonable ability will be able to succeed at. It also means to understand their context. It's easy to make a "really good" assessment that everyone fails because they also have three projects and two midterms in their other courses. Are your expectations clearly communicated, and are you ensuring that you only assess what you've asked for? (that doesn't mean that you can't expect students to go above and beyond, just that you need to tell them you expect them to) 3. **Are you assessing accurately?** I'm distinguishing this from "fair", but you can treat "fair" and "accurate" as two sides of the same coin. Accurate means that your assessments are set up so that appropriate weight is given to appropriate topics, and that your tests actually enable students to display their understanding and capacities, rather than whether they memorized the example or found the answer on stack exchange. Creating fair assessments is challenging, but there is a lot of research and resources available. 4. **Are you giving every student the chance to seek help?** I often find that if students are slipping through the cracks, setting up a regular meeting with them to keep them on track can do wonders. However, I am in a job in which I'm required to work with students like this, so it's easy for me to do. If you are a busy research professor who is teaching two courses per semester while juggling other things, it's a lot harder. Ultimately, the final exam is not when a student should find out they failed the course. They should know that they are on a bad path long before then, and should have opportunities to get on track. If you are doing these things, then **you** are not causing them financial ruin. It's similarly not fair to say that the students are causing this — you don't know their context and can't make the judgment. Perhaps they went to a bad high school that just didn't prepare them, or perhaps they are always on the train to another city because their parents are sick and they can't attend classes. It is not your responsibility to help them in this way unless you are capable of providing everyone the same help. Which brings me to the most unfortunate reality of post-secondary education: **Not everyone can make it**. For whatever reason, some students simply will not demonstrate that their abilities are up to the standard that has been set. Notice the wording I used there — I didn't say that they don't have those abilities, but that they will not **demonstrate** that they have those abilities. Provided you are assessing them fairly/accurately, teaching the best you can, and giving the help they pay for, then you are providing them with every opportunity to demonstrate those abilities. If they are unable to do so, then it would be unethical to let them pass regardless of the reason.

Being put on academic probation was one of the best things that ever happened to me. I switched to community college, took art, music, and math classes, learned who I **really** was instead of who I **thought** I was, and eventually graduated with a Bachelor's and have had a lovely career doing things that are truly meaningful to me. The bad grades I "earned" in my first two years at a university were trying to tell me several things: I was not at the right school. I was not in the right major. I was not in the right place in my life mentally and emotionally. But the grades alone didn't help me understand. The letter I got that said I was on probation was what I needed to get to understand that I was doing the wrong things. If professors had just passed me because they felt for me, I would probably be far less happy with my life right now. I would argue that students are paying you to give them the feedback they need on their work and views, they are not paying you for credits and/or a degree.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

Being put on academic probation was one of the best things that ever happened to me. I switched to community college, took art, music, and math classes, learned who I **really** was instead of who I **thought** I was, and eventually graduated with a Bachelor's and have had a lovely career doing things that are truly meaningful to me. The bad grades I "earned" in my first two years at a university were trying to tell me several things: I was not at the right school. I was not in the right major. I was not in the right place in my life mentally and emotionally. But the grades alone didn't help me understand. The letter I got that said I was on probation was what I needed to get to understand that I was doing the wrong things. If professors had just passed me because they felt for me, I would probably be far less happy with my life right now. I would argue that students are paying you to give them the feedback they need on their work and views, they are not paying you for credits and/or a degree.

Here in Germany universities are quite strict. After my daughter failed the very last maths exam three times, she was out. However, strict does not mean heartless or planless. Her uni had an arrangement with the tech unis; one of them agreed to accept the exams she had already passed and enrol her in a similar course, for the third year. She was not alone in this. This way all that time and investment was not lost. You could organise a similar scheme with nearby polys or tech colleges, to take up your students, who didn't quite manage, but are still worthy of education. So everybody wins.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

The other answers are great, but I want to stress one thing not really mentioned in other answers: **if you are concerned, talk to them before they fail**. Don't single them out for help because again it won't be fair, but still: tell them explicitly that if they continue performing as they have, they are going to fail the class. Point them to the various resources that are available (e.g. undergraduate tutors, office hours). If they still fail, as the proverb goes, you can lead a horse to water, but you can't make it drink.

Here in Germany universities are quite strict. After my daughter failed the very last maths exam three times, she was out. However, strict does not mean heartless or planless. Her uni had an arrangement with the tech unis; one of them agreed to accept the exams she had already passed and enrol her in a similar course, for the third year. She was not alone in this. This way all that time and investment was not lost. You could organise a similar scheme with nearby polys or tech colleges, to take up your students, who didn't quite manage, but are still worthy of education. So everybody wins.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

Take the example of a medical student. Do you want to pass someone who does not have the necessary knowledge to treat patients correctly? It is your duty to make sure that only the ones who know what they are doing will pass. This may be less strict in other subjects but the principle is the same. --- EDIT --- Another example where this becomes clear would be an airplane engineer or pilot that does not have the necessary knowledge (thanks to Mike's comment below!).

Here's an added perspective: If you start passing students for non-academic reasons (or really even just start spending mental energy considering that), there is *no lower bound* to that concern. Wherever you decide to set your threshold, there will always be students further down academically. There will always be some students to be "ruined", and the more your institution lowers standards, the further-down your population will drift over time, in like response to the lowered expectations. This is notably coming from my position as a faculty member at a (large, urban, northeast US) community college. Nationally, [community colleges only have a 22% graduation rate after 3 years](https://www.communitycollegereview.com/blog/the-catch-22-of-community-college-graduation-rates) (our is somewhat higher than that). I've routinely witnessed courses from remedial level up with 50%-60% failure rates. (In contrast, 4-year colleges only have 60% graduation rates after six years.) Many of our students cannot read or write at an elementary level, handle the simplest elementary arithmetic, have emotional/intellectual disabilities, etc. There is no threshold we could possibly set that would serve to pass all or most of these students. Example 1: I've had students ask whether they were guaranteed a passing grade as long as they physically attended every class session, and were incredulous when the answer was "no". (Apparently that's fairly common in some courses now.) Example 2: A few years ago we had a university-wide remedial algebra exam for all students (e.g, at the 8th-9th grade level). I attended a central planning meeting where someone asserted something like, "Our goal was to make an exam that no one could possibly fail. We have not succeeded, because 50% of the students are still failing." (This was given as a positive argument for further reducing the standard of the exam.) Ultimately this was found to be an impossible endeavor, so the college has now abandoned the exam and the basic-algebra requirement entirely. Example 3: In light of budget and enrollment pressures, among the university's new endeavors is to more widely expand advertising and enrollment to even more severely learning-disabled and intellectually-disabled prospective students. The OP identifies a keenly-felt and significant problem. But granted my perspective, I might suggest that there is *no solution* to this problem. No matter where you set the threshold or cutoff, there are more (many more) students lined up further down (arbitrarily further down) the skill and intellectual ladder hoping for the same judgement. It's a cycle that has no hypothetical end.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

Take the example of a medical student. Do you want to pass someone who does not have the necessary knowledge to treat patients correctly? It is your duty to make sure that only the ones who know what they are doing will pass. This may be less strict in other subjects but the principle is the same. --- EDIT --- Another example where this becomes clear would be an airplane engineer or pilot that does not have the necessary knowledge (thanks to Mike's comment below!).

All things in life can cause you economic ruin these days. House mortgages, car loans, hacking, frauds. It is basically a society based on debt-based servitude where the banks are our overlords. You are likely just one well-meaning person in this huge mess and likely cannot affect this fact. As far as I know you also can not affect what these students future employers will base their decisions on. It is not your moral obligation (nor is it even within your power) to try and steer where these students end up. It is the privilege of these debt overlords to do. Your job is to teach and to grade. Do your job, peasant.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

Some of the discussion surrounding this question demonstrates that people are viewing it through two different implicit lenses, and I'd like to make them explicit. * **Considered as an act in isolation** A single decision to give a student a "kindly" mark has quite diffuse negative consequences. As long as we assume that the rest of the student's record is accurate, the harm to students who have legitimately passed the course, to the institution awarding the credential, and to the organizations that will later trust that credential as a mark of suitability are all very modest. Down in the weeds, really. * **Considered as a pattern** If we assume that "give the poor kid a break" is a social norm that could be applied over and over again to a single student, then the picture is different. Some subset of students will be awarded credentials that don't mean what they say on the label, their subsequent failure in work environments will drag down the reputation of the institution and with it the value of degrees *earned* by students who didn't get a bunch of soft passes. And the damage starts even earlier than that, because professors in second year and later courses will have classes coming in less prepared than they should be and will have to give time over to remedial explanations and hand-holding to the detriment of progress in their course. Several times during my career I've agonized over the kind of decision facing John, and I'm hugely sympathetic. I feel for students who have gotten themselves into a financial bind they don't have the where-with-all to haul themselves out of. But by the time a student is standing in my office explaining that if they don't pass my class they won't graduate and, and, and ... they have already have years of warning signs. That can't be on one professors' head. More than once I discretely (no names) sounded out some of my colleagues about a student only to come away reasonably convinced that the subject had already been given a break; probably more than once. --- As an aside I'm convinced that a significant amount of blame for the binds the students get into in the US lies with a financing system that "just grew that way" through a serious of short-sighted and frankly stupid decisions made by politicians who were personally isolated from the consequences. Both major parties have been vastly wrongheaded in their own ways, but the damage is worse because of they ways they have compromised; the old joke about bipartisanship leading to stupid-evil legislation applies.

Going to college in a country with high tuition fees is a risk. A potentially very huge risk. I think it is important to establish a scheme that makes sure that weak students notice this immediately. This means that in the very first year, it should become clear to the students whether they will make it or not.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

I think you are trying to assume too much responsibility. The way you grade your students (assuming no malicious intent on your part) is not what is (possibly) causing their ruin. What might be causing it (besides, perhaps, some bad decisions of their own) is the system which forces them to go into debt or does not provide them with sufficient opportunities to study properly (e.g. by having them go to bad schools, or forcing them to work long hours in a part-time job). You are not directly responsible for any of this. Indirectly, you can try to change the system. I can think of a couple of ways to do it: 1. You can support (or even start) political initiatives which aim to change what you consider to be unfair. 2. If you think your job is part of the problem, you could quit, and look for a different job which would have less of a bad impact (e.g. a worse paid job at a place with lower fees). 3. You could do something to actively sabotage the system. I think your suggestion (passing students who don't really merit a passing grade) falls under the sabotage category. Indeed, by passing students regardless of their merit, you undermine the whole system --- the more teachers at your institution do that, the less value a passing grade will have. At the logical conclusion, the whole higher education system would collapse, and something better could come up from the ruins. Maybe (but probably not). Either way, in the interim, this would harm all the *other* students (who do merit a passing grade) by diminishing their accomplishments, other people (when the "passed" ones get a job they are not competent for, perhaps instead of someone more qualified), and possibly the ones you wanted to help (by pushing them to an ill-chosen career path). I think ideally, the first solution I have mentioned should be the best, and sabotage should only be used as a last resort.

All things in life can cause you economic ruin these days. House mortgages, car loans, hacking, frauds. It is basically a society based on debt-based servitude where the banks are our overlords. You are likely just one well-meaning person in this huge mess and likely cannot affect this fact. As far as I know you also can not affect what these students future employers will base their decisions on. It is not your moral obligation (nor is it even within your power) to try and steer where these students end up. It is the privilege of these debt overlords to do. Your job is to teach and to grade. Do your job, peasant.

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

Something other answers did not yet address - by giving unfairly good grades to undeserving students, ***you are dramatically, and unfairly, penalizing good students***. Students - in large part - choose to attend a university based on its academic reputation. Perhaps, paying premium. If your lax grading methods graduate unfit students instead of failing them, this will materially affect the successful student's reputations as graduates of your school, since the employers or graduate schools would have no way of knowing if someone graduated your program because they were a good student, OR, because you took pity on them. So, everyone will be tarred with the bad reputation and that would negatively affect their lives and careers.

I generally design my exams so that even a student with relatively limited understanding of the course can pass. I would really only consider a student who has achieved the highest grade (> 80%) to have a good understanding of the material, and I often see just-passing exams (50-60%) where I think that the student really had no idea, but somehow managed to scrape together enough marks to get over the line. Armed with that viewpoint, when I see a student who fails, I think they must have really deserved it! So I don't feel too bad about failing them; they need to put in some effort to fail my course!

The issue has crept up on me slowly over the last several years. I am increasingly aware of the massive debt that many of my students are taking on, debt which is far beyond the sort of debt that I incurred as an undergraduate in the 1980s. Because of this, in recent semesters I have found it somewhat difficult to fail students. Instead of simply asking myself "does this student deserve to fail this class?", I find myself asking "does this student deserve to have their life ruined?" In many cases (e.g. students who are already on academic probation) this is not much of an exaggeration. It is a very bad situation to find yourself in your early 20s with no college degree but $30,000 in debt. In some cases, I am aware that a decision of mine might be a contributing cause of a student ending up in just such a situation. I can no longer regard a failing grade as a relatively minor matter (like a speeding ticket). How do professors reconcile their de jure role as guardians of academic integrity with their de facto role of being (at least in part) responsible for their students' economic future?

2019/05/15

Take the example of a medical student. Do you want to pass someone who does not have the necessary knowledge to treat patients correctly? It is your duty to make sure that only the ones who know what they are doing will pass. This may be less strict in other subjects but the principle is the same. --- EDIT --- Another example where this becomes clear would be an airplane engineer or pilot that does not have the necessary knowledge (thanks to Mike's comment below!).

You're not failing the students. Assuming you performed your teaching job well, and you're grading them fairly, the students are failing themselves. They didn't study well enough to pass the class, or maybe they just don't have a talent for this material. Giving someone a passing grade when they haven't earned it is not fair to them, and it dilutes the value of passing grades for all the other students. If they need to use what you're teaching them in their career, they're not going to be as successful. The passing grade you gave them doesn't actually make them competent. When you go to college, you're not buying a degree. No matter how much debt you get yourself into, you're not entitled to the diploma, you still have to do the work and pass the classes. And as a teacher, it's your responsibility to make sure they've done that before passing them.

For Category theory there is a well known and quite active research mailing list <https://www.mta.ca/~cat-dist/> Do similar mailing lists or usenet groups exist for other branches of mathematics? For example in metric geometry, conformal geometry,... Or do most research related discussions now take place on MO?

2017/04/18

Here is a list of some mailing list for specific areas of mathematics. (This is posted as a [community wiki](https://meta.mathoverflow.net/tags/community-wiki/info), if you are aware of some other mailings lists, do not hesitate to add them to the list. And also if you have some additional useful information related to some of the lists.) * FOM -- Foundations of Mathematics - [basic info](http://www.cs.nyu.edu/mailman/listinfo/fom), [archive](http://cs.nyu.edu/pipermail/fom/). This mailing list has been mentioned quite often [in MO posts](https://www.google.com/search?q=pipermail+fom+site:mathoverflow.net) and [elsewhere](https://www.google.com/search?q=pipermail+fom), including citations in some [books](http://books.google.com/books?q=pipermail+fom) and [papers](https://scholar.google.com/scholar?hl=en&q=pipermail+fom). * Banach -- Banach Space Theory News - [basic info](https://www.mathdept.okstate.edu/cgi-bin/mailman/listinfo/banach), [archive](https://www.mathdept.okstate.edu/pipermail/banach/) * ALGTOP-L -- Algebraic Topology Discussion Group - [basic info](https://lists.lehigh.edu/mailman/listinfo/algtop-l), [archive](https://lists.lehigh.edu/pipermail/algtop-l/) * Categories List (category theory) - [basic info](https://www.mta.ca/~cat-dist/), [archive until 2009](https://www.mta.ca/~cat-dist/#archives) * ~~Real Analysis List~~ - [basic info](http://www.math.louisville.edu/~lee/rae/rafaq.html) (now very likely defunct - see [GEdgar's comment](https://meta.mathoverflow.net/questions/3214/mailing-lists-usenet-groups-for-research-math/3219?noredirect=1#comment14271_3219))

Sci.math, sci.math.research, alt.sci.(whatever) are some of the newsgroups that dealt with mathematics and various branches. You can find histories of them and other information on them with web searches. The unmoderated groups became spam wastelands, especially after Google instituted their version called Google Groups. I left sci.math for this reason, and only looked sparingly at sci.math.research, which was moderated. When Kevin Buzzard posted in s.m.r. about MathOverflow back in 2010, I tried it, and have not looked back since. If you frequent some older web pages (using the Wayback Machine, for example), you will find references to mailing lists and other resources. It seems that blogs and forums and YouTube are the current means of idea exchange; (as I believe Gerry Myerson might say) I will be surprised if you find many active mailing lists. Gerhard "The Technologies, They Are A-Changing" Paseman, 2017.04.18.

After reading [an article on REST](http://www.ibm.com/developerworks/java/library/j-grails09168/index.html) ("Restful Grails"), I have gotten the impression that it is not possible to truly conform to a REST style in a service that demands a lot of parameters. Is this so? All the examples I have seen so far seem to imply that true REST style services are "parameterless". Using parameters would be RPC-ish and not truly RESTful. To be more specific, say we have a service that returns graph data for stock prices, and this service needs to know the start date, end date, the currency, stock name, and whatever else might be applicable. In any case, at least 4-5 parameters are needed to retrieve the information needed. I would imagine the URL to be something like this : /stocks/YAHOO?startDate="2008-09-01"&endDate=... ("YAHOO" is here a made-up stock name). Would this really be REST or is this more RPC-like, what the author of the aforementioned article calls "GETful" (i.e. just low ceremony rpc)?

2011/03/21

Feel free to use as many parameters as you need to identify the resource you wish to access. REST doesn't care.

Why would you think it is not possible? Google uses REST for their charts api, and they take alot of params: <http://chart.apis.google.com/chart?cht=bvg&chs=350x300&chd=t:20,35,10&chxr=1,0,40&chds=0,40&chco=FF0000|FFA000|00FF00&chbh=65,0,35&chxt=x,y,x&chxl=0:|High|Medium|Low|2:||Task+Priority||&chxs=2,000000,12&chtt=Tasks+on+my+To+Do+list&chts=000000,20&chg=0,25,5,5>

After reading [an article on REST](http://www.ibm.com/developerworks/java/library/j-grails09168/index.html) ("Restful Grails"), I have gotten the impression that it is not possible to truly conform to a REST style in a service that demands a lot of parameters. Is this so? All the examples I have seen so far seem to imply that true REST style services are "parameterless". Using parameters would be RPC-ish and not truly RESTful. To be more specific, say we have a service that returns graph data for stock prices, and this service needs to know the start date, end date, the currency, stock name, and whatever else might be applicable. In any case, at least 4-5 parameters are needed to retrieve the information needed. I would imagine the URL to be something like this : /stocks/YAHOO?startDate="2008-09-01"&endDate=... ("YAHOO" is here a made-up stock name). Would this really be REST or is this more RPC-like, what the author of the aforementioned article calls "GETful" (i.e. just low ceremony rpc)?

2011/03/21

You can see the querystring as a filter on the resource you are GETing. Here, your resource is the stock prices of yahoo. Doing a GET on that resource give you all the available data, or the most recents. The query string filter the prices you want. Content negociation allow you to change the representation, e.g. a png graph, a csv file, and so on. To add a price, simply POST a representation (e.g. CSV) to the same resource. The "restfulness" is not realy in the URL itself, since URIs are obscures to client, but in the way you interact with resources themselves identified by their URI

Why would you think it is not possible? Google uses REST for their charts api, and they take alot of params: <http://chart.apis.google.com/chart?cht=bvg&chs=350x300&chd=t:20,35,10&chxr=1,0,40&chds=0,40&chco=FF0000|FFA000|00FF00&chbh=65,0,35&chxt=x,y,x&chxl=0:|High|Medium|Low|2:||Task+Priority||&chxs=2,000000,12&chtt=Tasks+on+my+To+Do+list&chts=000000,20&chg=0,25,5,5>

After reading [an article on REST](http://www.ibm.com/developerworks/java/library/j-grails09168/index.html) ("Restful Grails"), I have gotten the impression that it is not possible to truly conform to a REST style in a service that demands a lot of parameters. Is this so? All the examples I have seen so far seem to imply that true REST style services are "parameterless". Using parameters would be RPC-ish and not truly RESTful. To be more specific, say we have a service that returns graph data for stock prices, and this service needs to know the start date, end date, the currency, stock name, and whatever else might be applicable. In any case, at least 4-5 parameters are needed to retrieve the information needed. I would imagine the URL to be something like this : /stocks/YAHOO?startDate="2008-09-01"&endDate=... ("YAHOO" is here a made-up stock name). Would this really be REST or is this more RPC-like, what the author of the aforementioned article calls "GETful" (i.e. just low ceremony rpc)?

2011/03/21

You can see the querystring as a filter on the resource you are GETing. Here, your resource is the stock prices of yahoo. Doing a GET on that resource give you all the available data, or the most recents. The query string filter the prices you want. Content negociation allow you to change the representation, e.g. a png graph, a csv file, and so on. To add a price, simply POST a representation (e.g. CSV) to the same resource. The "restfulness" is not realy in the URL itself, since URIs are obscures to client, but in the way you interact with resources themselves identified by their URI

Feel free to use as many parameters as you need to identify the resource you wish to access. REST doesn't care.

Unfortunately the printer being used to print envelopes uses a LPR port which can only be attached to an old machine running Windows 98. The rest of the systems in the network are running Windows 7 and need to have the ability to send print jobs to Windows 98 print server. Are there any alternatives? Unfortunately Linux is not an option as there are no known drivers for the printer unless this does not matter. If there any no alternatives, what should be secured on the network and the print server so that it is not susceptible to viruses, unauthorized access, etc? The network does not run off a domain as they only have 5 computers.

2011/10/16

Get a [USB-to-parallel](http://www.google.com/search?q=usb%20to%20parallel%20adapter) adapter, it's around $7. Connect this printer with this adapter to any computer running Windows. HP LaserJet drivers are available for modern Windows versions too.  This is what one vendor says about their adapter: > > Add a DB25 parallel port to your desktop or laptop PC through USB. The ICUSB1284D25 6 ft USB to DB25 parallel printer adapter cable turns an available USB port on a host PC into a DB25 female parallel port - allowing you to connect a DB25 printer to the computer as if the necessary parallel port was built-on. A cost-effective and reliable solution, the USB to DB25 adapter saves the expense of replacing a parallel printer for the sake of USB capability. > > >

D-Link makes a parallel print server that plugs directly into the parallel port on the printer. You plug a network cable into it, and access it as a standard TCP/IP LPR port. We've been using one on our old hp LaserJet 4100 series printer that was used as the mainline printer back when printer built-in networking was a little over priced. We purchased the LPT Port print server to get the printer off the Domain Server to quit wasting server clock cycles on spooling a shared printer when the workstations could do a better job of it and communicate directly to the printer. The lj-4100 outlasted the hp LJ4250dtn with built-in print server that replaced it and is still plugging away over in Accounts Payable with this nifty D-Link device plugged into its parallel port. I think the current iteration is D-Link DP-301P or something similar.

Context ======= I'm working on very complex enterprise solution. We have this table that shows list of let's call them contracts. Every contract (row) can contain one or more customers (and one or more products). Problem ======= I need to think of these use cases: 1. People want to copy & paste customer IDs. 2. People should be able to remove all the customers at once. 3. People should still have an overview what customers were selected. 4. People should be able to browse customers and pick those they want to. (Since we are in the able the solution needs to be compact) My proposal =========== [](https://i.stack.imgur.com/HLqCS.png) - Clicking on the "CD-0005" would trigger the browse window - **Missing** copy & paste functionality [](https://i.stack.imgur.com/vctwb.png) - Looks super ugly - Hard to provide overview on all the items (having tooltip on input hover is somehow strange) **What do you think? How can this problem be solved?**

2018/02/06

1. You can add a copy link in front of each customer ID on hover. [](https://i.stack.imgur.com/LMt7G.jpg) 2. Add filters on table headers 3. Highlight the selected rows or make customer ID a badge when selected 4. Making customer ID badge would make it easier to identify the ones needed and switch between states. I loved how "Clear Move" handle the tables. You can use it as inspiration. [](https://i.stack.imgur.com/zmTlD.jpg)

Why don't you add a context menu (invoked by right mouse click) containing "Copy" and "Copy all" items? To know which customer ID you are about to copy a hover (on mouse over above each ID) helps. The balloon closes on mouse out event. It can be programmatically challenging to add a context menu on a balloon but once you get over it, the ease of use compensates for your effort.

Context ======= I'm working on very complex enterprise solution. We have this table that shows list of let's call them contracts. Every contract (row) can contain one or more customers (and one or more products). Problem ======= I need to think of these use cases: 1. People want to copy & paste customer IDs. 2. People should be able to remove all the customers at once. 3. People should still have an overview what customers were selected. 4. People should be able to browse customers and pick those they want to. (Since we are in the able the solution needs to be compact) My proposal =========== [](https://i.stack.imgur.com/HLqCS.png) - Clicking on the "CD-0005" would trigger the browse window - **Missing** copy & paste functionality [](https://i.stack.imgur.com/vctwb.png) - Looks super ugly - Hard to provide overview on all the items (having tooltip on input hover is somehow strange) **What do you think? How can this problem be solved?**

2018/02/06

1. You can add a copy link in front of each customer ID on hover. [](https://i.stack.imgur.com/LMt7G.jpg) 2. Add filters on table headers 3. Highlight the selected rows or make customer ID a badge when selected 4. Making customer ID badge would make it easier to identify the ones needed and switch between states. I loved how "Clear Move" handle the tables. You can use it as inspiration. [](https://i.stack.imgur.com/zmTlD.jpg)

Usman Mani already proposed an elegant solution, however for anyone seeing this after... I had similar questions and stumbled upon a very useful article about building tables for reusability. <https://uxdesign.cc/designing-tables-for-reusability-490a3760533> TLDR: It's ok to not show all of the information in a row, do some usability testing and find the 2-3 most important data points a user needs in a table row and then hide the rest. Utilizing font awesome illustrate that there is more data if the user does require it.

Context ======= I'm working on very complex enterprise solution. We have this table that shows list of let's call them contracts. Every contract (row) can contain one or more customers (and one or more products). Problem ======= I need to think of these use cases: 1. People want to copy & paste customer IDs. 2. People should be able to remove all the customers at once. 3. People should still have an overview what customers were selected. 4. People should be able to browse customers and pick those they want to. (Since we are in the able the solution needs to be compact) My proposal =========== [](https://i.stack.imgur.com/HLqCS.png) - Clicking on the "CD-0005" would trigger the browse window - **Missing** copy & paste functionality [](https://i.stack.imgur.com/vctwb.png) - Looks super ugly - Hard to provide overview on all the items (having tooltip on input hover is somehow strange) **What do you think? How can this problem be solved?**

2018/02/06

1. You can add a copy link in front of each customer ID on hover. [](https://i.stack.imgur.com/LMt7G.jpg) 2. Add filters on table headers 3. Highlight the selected rows or make customer ID a badge when selected 4. Making customer ID badge would make it easier to identify the ones needed and switch between states. I loved how "Clear Move" handle the tables. You can use it as inspiration. [](https://i.stack.imgur.com/zmTlD.jpg)

An easier and cleaner way would be to provide an Edit icon for each Contract row and open the whole thing in a modal overlay, provided the form does not run too long.

My wife doesn't work anymore and she has a 401(k) of about $80,000 from an old job. I have been kicking around the idea of taking that money out or rolling into an IRA so that I can have better access to this money. Given the interest rates I want to purchase a new home and would like to leverage this money as a down payment if it makes sense. Thanks for the help.

2011/03/04

Unless that 401K has very low [expense ratios](http://www.investopedia.com/terms/e/expenseratio.asp) on its funds, you should roll it into an IRA and choose funds with low expense ratios. After rolling it over you should *not* take the 10% penalty and use it to purchase a home. Unless you use that home as an income property, it is unlikely to provide you more than a 1% inflation-adjusted rate of return given historical data. The S&P 500 is about 4% adjusted for inflation. And that money currently in your 401(k) is for your retirement - your future. Don't borrow against your future. Let compound interest do its work on that money. The value of a house is in the rent you aren't paying to live somewhere and there are a lot of costs to consider. That doesn't mean don't buy. It just means buy wisely. If you are currently maxing out your 401(k), you may consider cutting back to save for your down payment. Other than that I wouldn't touch retirement money unless it was a dire financial emergency.

It's not your money. What does your wife think of this? You know, the withdrawal is subject to full tax at your marginal rate as well as a 10% penalty. That's quite a price to pay, don't do it.

My wife doesn't work anymore and she has a 401(k) of about $80,000 from an old job. I have been kicking around the idea of taking that money out or rolling into an IRA so that I can have better access to this money. Given the interest rates I want to purchase a new home and would like to leverage this money as a down payment if it makes sense. Thanks for the help.

2011/03/04

Don't borrow from your future to live in the present.

It's not your money. What does your wife think of this? You know, the withdrawal is subject to full tax at your marginal rate as well as a 10% penalty. That's quite a price to pay, don't do it.

I found this answer here regarding graphic design: <https://graphicdesign.stackexchange.com/questions/265/font-face-loaded-on-windows-look-really-bad-which-fonts-are-you-using-that-rend> This is exactly what my fonts are doing, but I'm trying to find out if there's a way to prevent this using html or css or anything web-based. I'm using "platin" as my font. Do I just need to find a different font? Any other thoughts on the topic?

2013/09/06

An answer from [this similar question](https://stackoverflow.com/questions/11225654/xcode-how-to-connect-xib-to-viewcontroller-class/18003928#18003928): "Here's a more step-by-step way to associate your new UIViewController and .xib. Select File's Owner under Placeholders on the left pane of IB. In the Property Inspector(right pane of IB), select the third tab and edit "Class" under "Custom Class" to be the name of your new UIViewController subclass. Then ctrl-click or right-click on File's Owner in the left pane and draw a line to your top level View in the Objects portion of the left pane. Select the 'view' outlet and you're done. You should now be able to set up other outlets and actions. You're ready to instantiate your view controller in code and use initWithNibName and your nib name to load it."

If it's a UITableViewController, simply create a new XIB file via Xcode (File -> New -> iOS -> UserInterface -> View) and then add set the file's owner to your subclassed UITableViewController. You'll likely want to re-do how the user interface looks -- in terms of dropping objects like the table view and buttons or whatever else -- into the XIB's view. It'll certainly save a lot of time versus trying to debug programmatically creating and adding subviews and actions. And once that's in place, you can then make IBOutlets and IBActions to your heart's content.

I'm doing a total kitchen renovation with solid hardwood flooring. Do I go hardwood all the way under the cabinets to the wall or stop just past the toe kick and finish the rest with ply of the same thickness?

2011/12/27

You can do either, but in my opinion it's better to use plywood as the base because: * it's a lot cheaper than hardwood flooring. * if you need to replace the hardwood flooring at some point in the future, it will be a lot easier to remove just the hardwoods: if you run planks under the cabinets, you'll have to cut them off where they go under the toe kick which will be a lot trickier with the cabinets in place.

I ran mine all the way to the wall. Turned out to be a good thing because I made a cabinet change which pushed everything down six inches.

I'm doing a total kitchen renovation with solid hardwood flooring. Do I go hardwood all the way under the cabinets to the wall or stop just past the toe kick and finish the rest with ply of the same thickness?

2011/12/27

In your situation, using real 3/4" hardwood flooring should be a lifetime floor. I cannot see any reason to remove it as any other type of flooring in the future could be installed right over it. Since it is much more likely to have some type of cabinet upgrade rather than actually needing to remove the hardwood, I would install the hardwood wall to wall and install cabinets over it. The actual extra amount of hardwood used under the cabinet area is minimal, and will make installation easier and faster than working around the cabinet footprint.

I ran mine all the way to the wall. Turned out to be a good thing because I made a cabinet change which pushed everything down six inches.

I've been doing C# with XNA for a year or so now, and I'm pretty comfortable with 2D games. But after some reading, I'm worried about XNA's future since it isn't supported in Windows 8 and stuff like that. So I've been considering switching to Unity 3D? What are the benefits of Unity over XNA/C# and it is worth the move? if not, why? I'm also open to suggestions of other languages and engines. I'm currently going through school and considering game development as a career, so I would like something which won't die in a year or so (as far as we can tell) and will give me skills I need. Also consider that I have previous programming knowledge with C#. Thanks, David.

2012/11/17

**XNA** XNA still works on Windows 8. The issue is that they are not supporting XNA based games in Windows 8 Modern UI. XNA still works for Windows 8 desktop games. The terminology is extremely confusing. XNA will either get a serious overhaul when the new XBox console is released or something brand new will be designed. **Language** If you want to create games for Windows 8 Modern UI, such as Cut The Rope, etc, you'll need to use C++. The last time I saw C++ was the only supported language that could interop with DirectX and Windows 8 mode. All the other features of Windows 8 are available with C#. **Unity3d** If you want to make video games you should pick Unity3d, or an equivalent gaming engine and framework. The problem a lot of video games creators get into is trying to design yet another game engine. This has been done to the point of them becoming commodities. Focus on the game, not the engine. Unity3d knowledge will be far more value than creating simple games with XNA. You should still understand 3D theory though.

**[MonoGame](http://monogame.codeplex.com/)** is a free XNA-compatible library that allows you to make games in C# for WinRT, Windows Store apps and Windows Phone 8, and on top of the [Xamarin](http://www.xamarin.com) tools also for iOS and Android. This works on top of SharpDX, the optimized managed wrapper for DirectX, so your game runs pretty much at the same speed as with XNA. Since MonoGame is open source, actively developed and targets all the current platforms, you probably need not worry about it being obsolete soon. It gives you a great way to keep your C#, XNA-based codebase to target Windows RT and the Windows Store now, and potentially extend to the other mobile platforms if you buy the Xamarin tools. So no reason to move your C# stuff to C++. Unity3D on the other hand is a totally different game framework, and would not be an easy port. But you can then target many more platforms (particularly consoles, and now Flash too), and still script most of the game in C#. But it's quite a different tool to learn.

I am performing 10-folds cross-validation to evaluate the performances of a series of models (variable selection + regression) with R. I created manually the folds with this [code](https://stats.stackexchange.com/questions/61090/how-to-split-a-data-set-to-do-10-fold-cross-validation). At the moment I'm performing first variable selection, then hyperparameters tuning through cv, and finally testing the performance with RMSE and MAE for all the models, but I have a doubt. Is it correct to "use" the same fold for all the models? Or should I do a separate cv for each model?

2018/10/29

I recommend trying both (more than once), and exploring any differences. In my experience, using the same set of folds for all models or using a new set of folds for each model doesn't make any material difference. Post if you find different! Regarding "I'm performing first variable selection, then hyperparameters tuning through cv", maybe watch <https://www.youtube.com/watch?reload=9&v=S06JpVoNaA0> to be sure you are not introducing any bias.

If you want to evaluate the performance of different models i.e. Model Benchmarking, it is necessary to keep the input environment same i.e. any external input like CV (number of folds). While you can tune the model-specific parameters to optimize the model.