Skip to content

Latest commit

 

History

History
1252 lines (900 loc) · 101 KB

FAQ_with_ToC__generated.md

File metadata and controls

1252 lines (900 loc) · 101 KB
Raw

Freenode programming channel FAQ

This page is intended to be global for the whole ##programming channel. So feel free to clone and reuse.

The Questions

How do I talk on the channel?

In order to talk on ##programming, connect your IRC client to the irc.freenode.net server, type «/join ##programming» (without the angle brackets) and press enter. It is a good idea to register your nickname with the Freenode network (see /msg NickServ HELP). This URL may work for you: irc://irc.freenode.net/##programming .

Freenode also provides a web interface here: http://webchat.freenode.net/ . For more information about viable IRC clients see:

What is considered on-topic on ##programming? What is considered off-topic?

Refer to the channel rules for more information.

How do I address someone to grab his/her attention?

Just say their nickname followed by a colon (":") or a comma (","). Many IRC clients will allow you to do that by typing the first letters of the nickname followed by pressing the tab key (so-called "tab-completion"). Please don't generally private message someone using the /msg or /query commands because that way other people cannot view and contribute to what you are saying.

I tried joining a different channel for help, but I could not - why?

Some channels people try to join to seek help with their technology (notably ##java) require registration at the Freenode services. See Freenode’s knowledge base for more information about user/nickname registration.

Note that you are still welcome to ask questions related to specific technologies on ##programming, but otherwise you'll probably get more help on the topic-specific channels.

What is the channel for $TOPIC/$TECHNOLOGY?

You can search the channels list by following the instructions in the command /msg alis help .

Why do ##programming and other channels have two "#" whereas other channels have only one?

To quote from this post about Freenode: «The name of a “Primary Channel” starts with a single hash (“#”) and is discussion forum for a specific project or community (like Alfresco, or your local LUG). “About channels” start with a double hash and host conversations about a topic or subject (for example: ##security). »

Why do some users on the channel have +v (voice, a + sign, a blue circle, etc.)?

These users are voiced so in case the channel is under attack, and it will become restricted, then they and only they will be able to talk here. Normally, it should not matter, as when there is no major abuse, everyone can talk whether +v'ed or not.

Why should I learn how to program properly?

To quote MJD «Because you can't just make shit up and expect the computer to magically know what you mean, Retardo!». Trying to whip up working code without a fundamental understanding of programming may cause any of:

  • Compiler errors
  • Buggy code
  • Security vulnerabilities
  • Data loss
  • Loss of time or money or worse.
  • Wasting the time of the knowledgable people in the channel by being a help vampire.

So please learn programming properly .

How should I ask a question on the channel? Can I?

The first guideline is that you should not “ask to ask” by saying “May I ask a question?” or asking if anyone worked with that certain technology. Just ask for help directly. Also see https://dontasktoask.com/ .

The second guideline is that you should probably read or skim the guide “Writing the Perfect Question” (or the older and somewhat deficient in style “How To Ask Questions The Smart Way” ) and try to follow their wisdom, which involves doing some prior due diligence.

Otherwise, also see the next question for how to share the offending code with the channel.

I have a problem with my code, how should I share it for people to review?

First of all, your code should exhibit the problem you are having and should be a ''self-contained'', and ''reproducing'' example (see “The Short, Self Contained, Correct, Example” page). If the code is too large, please consider reducing it to a more minimal example that still exhibits the problem (see the bisection method ).

Please don't share non-runnable pieces and fragments of your code, see “Show Us The Whole Code” .

Once you have that, you have several options to share it. If you are coding a web page, you can try sharing your code using jsfiddle . Otherwise, if your code is self-contained in one file, you can use a paste site such as ideone or paste.debian.net. Please don't floodpaste the code to the channel because this is slower and more annoying than using a paste site and may get you kicked or devoiced out of the channel.

If you have more than one file in the project, then you should put it in a self-contained version control repository on a code sharing site such as GitHub , Bitbucket , or GitLab , so we can easily clone or checkout it.

For more guidelines, see https://github.com/shlomif/how-to-share-code-online .

I do not see anyone talking on the channel - is it dead?

The channel is still active, but may grow silent for several hours at a time, especially during the weekends and other holidays. Stick around and you should see some conversation eventually.

Which Integrated Development Environment (IDE) or Text Editor Would You Recommend?

We get asked this question a lot. The answer is a bit long. First of all, it is likely that the following editing mechanisms will not be adequate for editing code:

  • Rich-text editors including, but not limited to, Microsoft Windows WordPad (formerly known as Windows Write, we were informed that it has a text editing mode, but see below), word processors such as Microsoft Word, OpenOffice.org/LibreOffice Writer, WordPerfect, AbiWord, etc. and WYSIWYG (= "What you see is what you get") HTML editors. These are not suitable for editing code.
  • Limited text editors such as Microsoft Windows Notepad (see http://www.notepad.org/ for a very fun link about it ), or Pico which are aimed for casual editing of text documents. More capable text editors such as Notepad++, Notepad2 and GNU Nano, should be OK.

These editors may still be useful if that's all you can use, but if you can afford to install and/or use a more capable text editor, it is, by all means, a good idea.

Otherwise, asking the channel about the most recommend text editor will end up in people giving you their own preferences (which may have involved many years of trying different solutions, and dedicating some time to studying their existing preference and getting accustomed to it). Some links to get you started:

Note that many decent text editors and IDEs are open-source, free-of-charge, and cross-platform (Windows, Mac OS X and Linux/UNIX). As a result, there is no good excuse not to download one, install it, and use it.

I want to learn programming? Which language should I start with?

Different people will give different answers when asked that. Most people on ##programming believe that something like the Python programming language would be a good start. Python is a popular, useful, open source, and cross-platform (runs fine on Windows, Linux and Mac OS X and on many other less common platforms) language, that is not too hard to learn and get your feet wet with. Some resources to learn Python are:

Here are some resources for learning other programming languages, whether as a first language or after knowing some other languages:

What amount of due diligence should I do before asking for help?

Make sure you know how to program in at least one programming language (see the previous question ) and understand the fundamentals of the general topic with which you are having a problem, by having read a book, a good tutorial, a video, an online course, etc. Otherwise, you might ask the wrong questions, or misguided ones, and generally waste the time of the more experienced people on the channel who are trying to help you.

How do I learn how to program?

See the previous question.

How should I do binary Input/Output properly?

Some people would naïvely assume that writing a C struct directly to a file would be a good solution for outputting it to disk. That is far from being the case, however. The representation of a struct in memory is subject to many system-dependent factors, such as:

  • Endianness - also referred to as “byte sex”, this determines whether bytes are stored in memory with the lowest byte first or the highest one.
  • Padding - for efficiency reasons, compilers tend to place empty spaces between the fields of structs. The amount of such space is not standardised and varies between systems, compilers (and even compilation options can affect it).
  • Integral data types and pointers can vary in their nominal size. I.e: an "int" can be 16-bit, 32-bit, or 64-bit, a "long" can be 32-bit or 64-bit, etc., which is system-dependent.

The O’Reilly book Porting UNIX Software (available for purchase, or a free download under a Creative Commons licence, on the link), goes into further depth about these and many other UNIX portability issues.

In any case, the best way to mitigate this problem portably and safely, is to write (and later read) every field (and sometimes every byte within) to the disk or network. You should be able to find some mechanisms inside portable libraries to do that.

What are some best practices in programming that I should adopt?

First of all, we should note that many best practices may still be controversial to a small or large extent, despite being very common in the industry. We also don't expect a beginning programmer who writes some simple programs to immediately employ all the best practices that go into developing mission-critical software on which lives are dependent.

But here goes with a grain of salt:

  1. Use a good version control system. For more about version control systems and the motivation behind them see the “Better SCM” Site and a related section from “Perl for Perl Newbies” (which is not exclusively limited to Perl) and Sawyer X’s talk about source code management systems.
  • Please don't give the usual cliches that you don't need a version control system if you’re only one person or that the project is not large enough to warrant it. You probably do.
  1. Write Automated Tests. See “What you should know about automated software testing”, the wikipedia Test-driven development entry and other resources.
  2. Use a standard build/configuration system for the language you use. Simple makefiles may be adequate for small projects, especially those written in C or C++. If you need to handle dependencies, build-time configuration variables and other complications, you should look at alternatives for C or the community-based alternatives for your programming language of choice (e.g: setuptools for Python; Dist::Zilla for Perl; RubyGems for Ruby).
  3. Deploy a local or hosted Continuous Integration (CI) service and make sure that failing builds or tests are remedied.
  4. Set up a bug tracker and use it to manage the issues and tasks' list. Some code-sharing sites provide hosted web-based issue trackers free-of-charge.
  5. Implement Reproducible Builds ( Wikipedia page ).
  6. Use code refactoring to get rid of code smells such as duplicate code or magic constants. Also see "Rub a dub dub".
  7. If possible, make use of linters, autoformatters, and validators and incorporate them into the test suite and CI.

For more information consult these links:

  1. The essay “What Makes Software High Quality?”
  2. “Awesome Guidelines” - a curated list of high quality coding style conventions and standards.
  3. The Perl Beginners Site “Perl Elements to avoid” and its spin-off “C and C++ Elements to Avoid” (both of which have a lot of good advice related to other programming languages).
  4. The essay “Clean Code JavaScript” on GitHub.
  5. The tongue-in-cheek essay “How To Write Unmaintainable Code” .
  6. A discussion on Dave Cross’ Perl Hacks blog titled “You Must Hate Version Control Systems”.
  7. The Daily WTF is a web site dedicated to posts about awful, but funny, programming practices.
  8. The Joel Test and similar resources on the Joel on Software site.
  9. “5 Tips to Increase your Code Quality”
  10. “Awesome Linters” and “Software Quality Enhancement Tools and Links”
  11. “100 Tips on Software Developer Productivity” - by Ari Koponen.

What is the best programming language?

We sometimes get asked this question, and the answer is simple: there is none. Different languages have different strengths, weaknesses and often trade offs. You may wish to read the blog post “Why Your Programming Language Must Suck” for a short coverage of some of the important trade-offs a language design will have, and why not everyone will be happy with your choices.

Even if we settle on a certain amount of preferences, then you may run into other factors that influence which language you may prefer:

  1. Should the program run as quickly as possible? Or do you want to optimise for the developers’ time?
  2. Is low memory consumption critical?
  3. Should the language be portable to various operating systems and/or CPUs? Or is it OK if it only runs somewhere specific?
  4. Should the language have an open-source implementation?
  5. Should the language have a large user base, and provide quick and useful support?
  6. Should bugs be addressed and fixed quickly in the implementations?
  7. Should the language have any particular built-in features that you desire?
  8. Should the language have an extensive library of extensions and APIs that can be reused?

Etc. If you tell us what your needs are, we can help you with a recommendation.

Note that some of the channel members happen to like a certain programming language in particular and tend to constantly recommend it to other people. Furthermore, you may run into language wars where one or more languages are bashed as awful languages by participants in the channel. That does not mean these languages are necessarily good or bad or suitable for what you wish to do.

What is the best [insert type of product with some alternatives]? What is the best and [insert adjective here]?

We also get asked this a lot and the answer is similar to the previous question: best in which respect? The various alternatives may have functionality trade-offs, and even if the factors are not at odds with one another, there can be remaining issues due to lack of developer time, policy, or design. Please tell us what exactly is the most important to you.

Note that "best and [insert adjective here]" is equally puzzling. Do you want the product with this property or do you also seek some other qualities?

I need/am forced to work with a certain technology (e.g: programming language or framework) which I hate. What should I do?

You may wish to read this post about Stoicism and Peace of Mind or the Tao Te Ching. Namely, you can say to yourself: "OK, this is not ideal, but I cannot avoid it, so I'll try to tolerate it, and make the best of it, and maybe even grow to like it.".

What are good resources (including books) to learn about algorithms?

Here are some resources for learning about algorithms:

There's also a channel called "#algorithms" on Freenode and, naturally, you can discuss algorithms on ##programming, as long as you perform due diligence.

I have an awesome idea for a great project (probably a game) and I want people to help me…

First of all, we appreciate your enthusiasm, but you should realise that most of us have better things to do than help you (even if they are unemployed), and so you should offer money (and we don't mean a share of the profits.). If not, you are quite unlikely to recruit anyone.

If you cannot offer money, please consider learning the necessary skills yourself (see the above questions) and starting work on your own. If you make your project open-source and publicise it, then other people may eventually join you, and help you if they find it interesting.

What is the difference between “scripting” languages such as Perl, PHP, Python or Ruby and “Industrial-strength” Languages such as C, C++, Java and C#?

Many people have issues with calling dynamic languages, with dynamic typing and those that are of somewhat more open-source nature such as Perl, PHP, Python and Ruby “scripting” languages. See:

The problem with calling them that is that while they allow for getting small and simple things done quickly and easily, large scale projects (100,000s to millions of lines of code) have been written in them and are still maintained without any show-stopping difficulties. That is not to say that these languages do not have their downsides, but they still do not preclude general-purpose, and/or large-scale programming.

Furthermore, if we consider Perl/PHP/Python/Ruby/Tcl/Lua/etc. as “scripting languages” and C, C++, Java and C# as “industrial strength” where does Common Lisp fit in? Where does Scheme? Where do Haskell and OCaml? What about the various dialects of Basic? Fortran? COBOL? It seems that the classification of languages as “scripting” vs “industrial strength” is more of that of public perception and marketing as opposed to their actual qualities.

On the other hand, we do not wish to claim here that C, C++, Java and/or C# are necessarily bad languages, or that you should not study them or refuse job offers that involve writing code using them.

I need to learn about various base systems, like hexadecimal (base-16), binary (base-2). Where can I find information about them?

See:

One thing to note is that numbers are the same regardless of their representation. E.g: the number represented by the base ten digits "16" is the same Platonic ideal as the hexadecimal representation "0x10", the binary representation "0b10000", or the base 7 representation "22", the English name "sixteen" (or the count of apples in this: "🍏🍏🍏🍏🍏🍏🍏🍏🍏🍏🍏🍏🍏🍏🍏🍏"). Therefore, it is inaccurate to speak of "binary numbers" or "hexadecimal numbers".

I want to buy some interesting books to read. Can you recommend me some?

Here are some lists of recommended books by some of our regulars:

In addition, here are some recommended lists of books by some notable software engineers and writers:

How do I learn how to use GDB (= the GNU debugger)?

We are often approached by people who tell us their C or C++ (or whatever) program crashes or otherwise misbehaves and then ask us why. If that's the case for you, you should realise that you should learn how to use an interactive debugger in order to debug your program, and you should learn how to do so as soon as possible.

One such debugger is the open source command-line debugger, "gdb", which can be used to debug programs in C, C++ and some other languages. Here is some information on learning to use it:

Also note that gdb has some GUI frontends that may make it easier or more convenient to use.

I feel like programming something, but I don't know what. Can you suggest some good ideas for programs?

For some lists of ideas for what to program see:

In addition, note that you may wish to consider contributing to open source projects that you find useful or interesting. For more information about that see:

Does anyone know of a program to do [insert task here]?

Such questions are tolerated on the channel, but anyway:

Note that very often these questions describe a task that is relatively quickly doable in Python/Perl/insert-productive-language-of-choice so either learn Python or whatever, or alternatively hire a programmer to write it for you.

I want to contribute to an open source project, but how? Which one can I contribute to?

First of all, it is good that you wish to contribute and good luck. Here are some resources to get you started:

If you still need ideas for which open source projects to contribute to then ask yourself - which open source projects are you using? (Either directly or indirectly.) Which ones are you interested in? If that fails perhaps try using GitHub search, DuckDuckGo, Google, the English Wikipedia search or other search resources.

What is "open source" software? Does it mean only that its source is available?

No. According to the open source definition an open source software licence must fulfill more than just the availability of the source code, including free redistribution and allowing to modify the code and distribute changes. Read the link on the wikipedia page for more information.

What is FizzBuzz?

In short, FizzBuzz is a simple programming task, which most decent programmers are expected to be able to easily write, but, reportedly, most interviewed software development candidates cannot write it. It's a kind of a litmus test for programmers.

Namely, in FizzBuzz one has to write a program that prints the natural numbers from 1 to 100 on one line each, but if the number is divisible by 3, print "Fizz" instead, and if it is divisible by 5 print "Buzz" instead; if it is divisible by both, print "FizzBuzz".

For more information, see:

How are C, C++, and C# (C-sharp) related? How are they different?

  • The C Programming Language was created by Dennis Ritchie and other workers, and was used to write most of UNIX. It is a kind of "mid-level" language because while it is above assembly, it is still lower-level than most other high-level languages.
  • C++ originated as some extensions to C to facilitate object-oriented programming (OOP) there, and has since developed many more extensions. C code can usually be compiled using a C++ compiler as is, but such code is no longer considered idiomatic C++. C and C++ can also call each other's routines, with relative ease.
  • Java is a programming language whose syntax is superficially reminiscent of C, but which has otherwise different syntax, semantics (including garbage collection) and a completely different standard library.
  • C# was inspired a lot by Java, and is very similar (but not compatible), and is not compatible with either C or C++. It is just sharing a name containing "C" for presumably marketing and branding purposes.
  • Managed C++ is a C++-like language for the .NET framework (which is also used to host C#), which isn't entirely compatible with C++.

As a result, the term “the C family of languages” is confusing and should be avoided, in part because other languages not mentioned above may share similar syntax to C.

I am facing a number of problems while using floating-point data types (e.g: "float" or "double"). Is there a explanation for this?

Computers use binary-encoded numbers because they’re faster at dealing with them, and because for most calculations, a tiny error in the 17th decimal place doesn’t matter at all since the numbers you work with aren’t round (or that precise) anyway. Moreover, because they use a binary floating point, they cannot represent a number like 0.1, 0.3, etc. accurately at all.

When the code is compiled, the floating point number is rounded to nearest number in the format, which results in rounding error even before the calculations. Furthermore, associativity in addition may result in different results: a + (b + c) != (a + b) + c and it is possible to get weird results like a + b == a if a’s exponent is much larger than b’s.

For more information, refer to these resources:

How do I hide/obscure/“encrypt” my source code to prevent end-users from learning how it works?

If you want to distribute some code you wrote and prevent end-users from learning its inner working, there are a few things you need to be aware of.

The first is that there are decompilers and that it is often possible for humans to do reverse engineering in order to determine the inner workings of a program to duplicate them. We were told that there's a code-obfuscation technique that can usually prevent that, but its downside is that the resultant code is much larger and much slower (by a factor of about 100,000), so it'll likely be unacceptable to you.

The second is that some popular approaches for "hiding" the source code are known to be ineffective - see this post about distributing CPython's .pyc files for an example, as well as this 2002 thread about obscuring Python code.

The third is that a mostly failproof way to avoid revealing your code is to set up a world-wide-web service (or a different networked service) for it and have the local end-user´s code query it using the network stack.

Another point is that you should consider trusting your users instead. Shlomi Fish (“rindolf”) has written a cautionary tale about the fact that BitKeeper’s parent company, BitMover, has done a series of licence changes to BitKeeper over the years to prevent abuses from users who were unlikely to become paying customers out of sheer paranoia in “What BitMover Got Wrong?”, and you should better avoid it. Freenode is dedicated to discussing collaborative and free/open projects and you are likely to receive a lot of heat for not trusting your end users.

Finally, note that Copyrights or software patents may provide adequate protection into people misappropriating the working of your code, although they may be controversial (especially software patents). For more information, see Eric Sink’s essay “Five Things Every Micro-ISV Should Know About Intellectual Property” .

Who is the world’s best programmer?

We got asked several times who is world’s best programmer, so let’s try to answer it once and for all. The first question we need to ask you is: “Why does it matter?” If you knew who the best programmer in the world is, do you think you can afford to hire him? Moreover, as a post by Ovid notes, studies found that the best programmers are only 5 times more productive than the worst, which is consistent with productivity variation in other fields. Furthermore, a contractor who did both web development and Linux kernel development noted that despite common belief, “web development is harder than Linux kernel development”.

Finally note that some software developers were able to contribute a lot to the world of computing while writing only sub-optimal code. See the post “Why some Great Hackers Write Inelegant Code”.

The most important thing you can do is constantly try to improve your programming skills, and learn from your fellow programmers, and from good and bad code people wrote out there. One way to achieve that is to be present on software development's forums, including ##programming .

I printed something but don't see it on the console yet! My output is out-of-order? Help!

You may be suffering from buffering. First of all, go over Mark Jason Dominus’ “Suffering from Buffering?” page which is somewhat focused on these issues for Perl 5. Then you may wish to:

  1. In C, make a judicious use of fflush(), or use locks for synchronisation.
  2. In C++, use std::flush.
  3. For Python, see this Stackoverflow thread
  4. Use similar routines for your programming language of choice.

What do “NP-Complete”, “NP-hard”, etc. mean?

Try referring to these links:

In addition, the English wikipedia has a list of problems that were shown to be NP-complete.

What do you think about interpreted vs. compiled languages?

First of all note that “interpreted” vs. “compiled” is an implementation detail, and a programming language may have more than one implementation.

Morever, it is not always that straightforward - there are also language implementations that compile the raw text source code into p-code, bytecode, abstract syntax tree, or some other intermediate representation (IR) and then interpret that; there is just-in-time compilation (JIT); some compilers compile to different high-level languages (see for example the List of languages that compile to JS on CoffeeScript's wiki as well as the Wikipedia page about "Source-to-source compilers"); some packagers package a runtime of a language with a higher level code (which to be honest most compilers do even if they just bind to libc).

There may also be some other approaches.

Note that one should not use the fact that an implementation is interpreted against it, because it may not be slower than compiled code all the time, and because it may still be fast enough for many uses given today’s processors’ speeds.

Despite some common belief, C has some interpreters as mentioned on stackoverflow, while some dynamically-typed languages such as Common Lisp, Scheme, or Python have compilers.

Moreover, note that interpreters are often easier to implement and port than compilers, which may justify the poorer performance.

As another anecdote, one can think of most CPUs as interpreters for their machine code, which are implemented using logic gates and electronic components.

How is JavaScript related to Java? What is the difference between them?

JavaScript (also see its Wikipedia entry) is a programming language first introduced in version 2.0 of the browser Netscape Navigator which also introduced support for browser-hosted Java applets written in the Java programming language.

According to this blog post titled “Popularity” it was designed so its syntax will resemble Java (which ruled out the use of other popular open source languages at the time) and was named “JavaScript” for marketing reasons. Despite that, Java and JavaScript codes are not compatible with one another.

For more information, see:

How can I learn Web Development? (HTML/etc.)?

First of all, you may wish to go over MDN's "How the Web works", “The Unix and Internet Fundamentals HOWTO”, or Mozilla Hacks' "Web Demystified" video series to gain some basic understanding of how the Web works over the Internet.

After that, read HTML Dog which hosts some quality tutorials about HTML, CSS, and JavaScript.

After that, perhaps using a static site generator will be good enough for you, and static sites are more secure and require less maintenance than sites with server-side scripting. If you need something to run on the server, then you have a lot of choice for server-side technologies and often web frameworks and APIs for them.

Also see the post “Four Ways to Build Web Apps” by Tom Hummel

Where can I get help with writing my own operating system or kernel?

First of all, Freenode has a channel called #osdev and it is possible that you may get better help there.

Otherwise, we should distinguish between two different cases. The first is that you're just writing the kernel as a learning exercise and do not intend it to be production-ready. In that case, you may go forward with whatever design you have in mind.

The other case is that you want your operating system to be popular. In that case, you should probably make it POSIX-compatible in order to avoid what Joel on Software calls “Chicken and Egg Problems”. You should also consider contributing to one or more of the existing open source operating systems out there, because that way your efforts will be much more likely to amount to something.

Is {software dev paradigm, e.g: Object-Oriented Programming, Functional Programming, Structured Programming} bad/good/better/worse?

We had several people join to try to provoke a discussion about whether Object-Oriented Programming or a different software paradigm was bad or had alternatives. These discussions were rarely constructive and usually were pointless, so you should avoid sparking them.

If you are working on a project that is under your control, you can try to use programming languages which avoid this particular paradigm. If you need to use it for your day job, you can either look for a different job and/or realise that “Reality to be commanded must be obeyed.” and all that jazz, and try to cope with what you have.

Otherwise, it is unlikely that adopting a certain programming paradigm, or alternatively ditching one completely, will magically solve all the world's software development problems, and provoking heated debates about it on ##programming is even less likely to make that happen.

Why are people still using C when there are languages like Rust or Pony?

First of all note that we do not mean to imply that languages like Rust or Pony are not without their merit, or do not provide any significant advantages in comparison to C or C++.

That put aside, there are some reasons why C may still be popular:

  1. Due to legacy and inertia and the fact C has been around for many years before Rust was introduced, there is a lot of C code out there, and it still needs to be maintained.
  2. Rust reportedly compiles extremely slowly, which may be prohibitive especially for large projects, and similar complaints were voiced against C++’s compilation speeds.
  3. Rust and Pony may be quite more complex than C.
  4. Some people may still like C, and they are still allowed to use it.
  5. Rust appears to have become a religion-of-sorts, which deters some people from using it.

If you like Rust or Pony or whatever, and wish to promote them, then it would be a better idea to actively contribute to and improve their core or ecosystem, rather than ranting about how much their older alternatives suck on online forums. Like a wise man once said: “After all is said and done, a hell of a lot more is said than done.”, and you should try to keep the ratio smaller.

How do I factor a large integer number into its prime components quickly enough? (Project Euler problem 3)

You can peruse the sources of mini-factor which is quite fast and licensed under the permissive Expat licence, those of factor from GNU coreutils which is under the strong copyleft GPLv3 licence, and those of FreeBSD’s factor which is under a permissive open source licence.

Here are some tips in English:

  1. Make sure to divide the number by the found factor (repeatedly as long as the remainder is 0) so it becomes smaller.
  2. This way, you don't need to check the factors for primality.
  3. If a number is a composite (= non-prime), then its smallest prime factor is less or equal to its square root. So if you have checked all numbers up to its root and they didn't divide it, then you know it is prime.
  4. Consider skipping common composite factors such as products of 2 or 3.

How can I write code to perform operations on web-sites for me? (that otherwise should be done manually)

This is called web automation and can be conveniently done in several languages:

Note that if the site provides a WWW-based API, then you should use it, and that automating client-side JavaScript may prove to be tricky. You should also avoid overloading a site or going against its Terms of Service.

Some technologies that allow automating using client-side JavaScript are:

Where can I learn to write and/or read regular expressions?

Here are some resources:

I have problems understanding recursion. Can you help?

What does "indentation" mean?

Indentation means that the contents of every block are promoted from their containing environment by using a shift of some space. This makes the code easier to read and follow.

Code without indentation is harder to read and so should be avoided. The Wikipedia article lists several styles - pick one and follow it.

Also consider using an autoformatter: awesome-linters .

I shared my code (using a paste-site/etc.) and people on the channel complained that my indentation is bad / off / inconsistent , but it looks fine in my text editor. Why?

This is usually caused by mismatch of hard tabs vs whitespace expansions or because you are using a different tab expansion step than the paste-site or that of other channel members' local editor (e.g: 4 spaces, 8 spaces, 2 spaces, 3 spaces). Please make sure your indentation is consistent either by expanding all tabs to spaces, or by using a prettyprinter / autoformatter for your programming language.

I want to release my code - which (open source) licence should I use?

First of all note that IANAL (I am not a lawyer) and all that jazz. Some resources to get you started are:

  1. GitHub’s choosealicense which provides the most popular and recommended choices.
  2. Understand the distinction between permissive licences, weak copyleft, strong copyleft, and proprietary software licences.
  3. The licences list of the Free Software Foundation is comprehensive, but opinionated.
  4. The OSI list of popular licences.

Please avoid:

  1. Creating your own licence - this risks licence proliferation and may very well be legally unsound.
  2. Not specifying a licence - this risks the software being considered as All Rights Reserved by default which may taint and encumber the future work of software developers who read it.
  3. Choosing an unpopular or obscure licence.
  4. Choosing a licence that is incompatible with the GPL.
  5. If you want to use the public domain as a non-licence, please use CC0 .

What do Continuous Integration (CI) services such as Travis-CI, Jenkins, or AppVeyor provide?

First of all, you may wish to peruse the wikipedia entry on Continuous Integration. A CI service runs the build and tests on pristine environments (= such that dependencies need to be explicitly installed) on each version control commit. For example, it may be configured to install some dependencies, and then run ./configure && make && make test. If one of the steps fails, then the problem is reported, and the developers need to restore the build to order.

While one can set up local CI services (using e.g Jenkins), there are hosted services such as GitHub Actions, GitLab Pipelines, Travis-CI or AppVeyor. They provide a service, which: 1. open source projects can use free of charge 2. one can pay to use for non-public-source-projects to hopefully ease the maintenance burden.

How do I debug memory errors (e.g: crashes, memory leaks, double-freeing) in my C / etc. programs?

There are some tools that can help:

  1. valgrind - an open source memory debugger.
  2. AddressSanitizer - another open source tool supported by the GCC and Clang compilers.
  3. Insure++
  4. PurifyPlus
  5. BoundsChecker - Windows-only.

How do I remove duplicates from a list of items? How do I make the items unique?

The best way is to use a Set or a dictionary/map/associative-array ADTs . For more insights consult:

How is this symbol/character/punctuation mark called in English?

The following resources may help:

How do I find the size and/or length of an array in C?

See this Stackoverflow question .

Is it true that macOS / Mac OS X is superior to Linux because it is a "Certified Unix"?

First of all, see the wikipedia page for when OSes are considered certified to use the UNIX® trademark.

In the context of this question, we are not trying to disuade you from using macOS, just to point out that being "Certified UNIX®" is not too relevant and should not be used as an argument against using Linux. Namely:

  1. Certified Unix® is reportedly given to particular OSes running on specific hardware. This is harder for Linux and other open source Unix-like OSes to achieve because they support heterogeneous hardware (including x86 and its peripherals).
  2. Two OSes based on the Linux kernel and the GNU/etc. runtime - namely EulerOS and Inspur K-UX - were certified as Unix. See the wikipedia page for details.
  3. At a time, Windows NT by Microsoft was certified as Unix as well.
  4. Being certified as such does not mean that the system will behave exactly as a Unix system is expected to behave. For example, the /etc directory in macOS is reportedly mostly neutered.
  5. Most people and organisations do not care too much about the Unix certification and will be happy to use a sane non-certified Unix-like operating system. The certification as Unix does not buy you much.

Is a certain software technology good just because I heard that NASA uses it?

NASA is large (and has a large budget, and a large scope of concern) and has been around for many decades, and so uses many different technologies, some of them may not be too recommended in the general case. Moreover, like many other organisations (and individuals), its developers, of software of otherwise, often experiment with technological trends or ideas, that turn out to be fads, or even long-term technical debts or liabilities.

While the Python homepage used to proclaim that "NASA uses Python", with a photo of an astronaut, and it is technically true, it was largely marketing built on common ignorance.

Among the other technologies that NASA has used are:

  1. Fortran
  2. COBOL
  3. VAX
  4. Forth
  5. Assembly
  6. Perl 5
  7. Windows, MS Word, and MS Outlook

Note that we do not mean to imply that Python or whatever are bad just because NASA uses them, just that you should not use that fact as an argument in their favour.

Can a copyright owner change the licence of an open source project if they dislike someone's abuse of it?

First of all, note that IANAL (I am not a lawyer) and all that jazz.

A copyright holder may change the licence as they please, and permissive licences also allow other parties to sublicense the code under different licences. However, the original version of the code may be forked and maintained under the original licence by other parties.

For more information, see:

Should I avoid using an open source project just because it is primarily developed by a company or individual that I dislike?

First of all, note that IANAL (I am not a lawyer) and all that jazz.

In essence, the FOSS nature of the licence protects you from claims of misuse by the copyright holder, but note that some open source licences (e.g the AGPLv3) are quite restrictive, which may not be acceptable for your use cases.

An open source project can also be legally forked and maintained independently if all else fails.

So it is a good idea to avoid guilt by association in this case assuming the relevant project is of decent quality.

I was recommended a Unix tool. How can I set up a Unix-like subsystem on my system?

If you already have Linux, one of the BSDs, or macOS, then they are usually suitable for that out of the box. If you are using MS Windows, then look into:

  1. cygwin - an open source Unix-like environment that runs on several versions of Win32/Win64.
  2. WSL by Microsoft for Windows 10.
  3. Install GNU/Linux or similar on a virtual machine emulator such as VirtualBox.

For more information, see “How to start contributing to or using Open Source Software”.

What is language X good for?

More often than not, it is asked about a general purpose programming language which can be used for a wide array of purposes. Note however, that sometimes they may not be too suitable for certain tasks for various reasons.

How can I run / program .NET and ASP.NET applications on Linux or macOS?

You can try using one or more of:

  • .NET Core - open source by Microsoft and other contributors
  • Mono - also open source

Note that some APIs such as WPF were not open sourced yet.

What can I use to document my code? Can you recommend a documentation tool/system/framework ?

Here are some resources to get you started:

  1. rindolf wrote a page titled “Choice of Document Formats”.
  2. Wikipedia has a page about Lightweight markup languages with a comparison.
  3. Doxygen is an open source documentation tool that supports many programming languages.
  4. DocBook and TEI are open XML-based standards for encoding semantic documents and translating them, but with complex tooling.
    1. AsciiDoctor can be converted to DocBook.
  5. Many languages have a common standard for documentation such as perl Pod or python's RST.
  6. http://pandoc.org/ can convert between formats, but the conversion is not always flawfree.

Which Linux Distribution / OS / version can you recommend?

See this link with an introduction and some recommended alternatives.

How can I learn how to use the Git version control system?

The “awesome-git” document on GitHub has a list of tutorials.

What do people mean when they say “s/pattern/replacement/”?

This means replace (or "substitute") the pattern with the replacement in a previously said text.

This notation is derived from the Unix ed editor, which in turn was inherited by sed, vi (and vim) and Perl.

For more information see:

What does “FOSS” mean?

Free and open source software .

How do developers of open source software make money from it?

First of all see "A handy guide to financial support for FOSS".

Otherwise, note that copies of FOSS can be copied and distributed at a very low cost, like other digital files. As a result, some developers opt to provide their gratis digital works (including FOSS) free of charge, without expecting to be reimbursed, simply because they do not get penalised for every downloaded copy.

How can I calculate the top / bottom number of items of a certain property from a stream of them?

You can try using an efficient implementation of the Priority queue ADT. Some implementations of it are:

What does "bisecting" mean?

There is a wikipedia page that details the VCS technique, but it has a more general utility.

As the old joke goes: "How do you find a lion in the desert? You split the desert into two halfs, and the lion is in either one. Then you split the half with the lion into two quarters, and so on until you find the lion-sized fraction of the area with the lion." You can apply this to many problems, e.g by gradually moving away or temporarily removing configuration files or directives until you pinpoint the problem.

How can I overwrite text in a terminal or write a console-based UI?

  • The de-facto standard for writing TUIs is curses which has some open source implementations such as ncurses and PDCurses, but may be somewhat baroque in nature.
  • libtickit and others are newer and possibly more modern alternatives with incompatible APIs.
  • There are bindings for other languages like curses.py
  • For simple tasks, printing no more than the terminal width (typically 80 characters) and then printing "\r" and overwriting may be adequate.
  • There are libraries for line input such as readline (GPLed) or libedit (BSDLed).

How much time / competitive programming problems / books / online courses / etc. will it take before I am employable / a "master programmer" / etc.?

How many roads must a man walk down, before you can call him a man? "The answer, my friend, is blowing in the wind".

Somewhat more seriously now, here are some thoughts:

There is a How to Teach Yourself C++ in 21 days parody comics and a Teach Yourself Programming in Ten Years essay (but note that there was a post claiming that the study it was based on was about becoming an exceptional artisan, and that it takes far less time to merely become proficient).

In "Lord Palmerston on Programming", Joel on Software notes that "Becoming proficient, really proficient, in just one programming world takes years. Sure, lots of bright teenagers learn Delphi one week and Python the next week and Perl the next week and think they are proficient. Yet they don’t have the foggiest clue how much they’re missing.".

You can probably get an entry-level or junior position software development job with relatively little or no experience, but likely one with a relatively low salary.

Some books like "The Pragmatic Programmer" or "Coder to Developer" aim to teach inexperienced developers, part of the wisdom gained by experience.

It is a perceived fact among software developers that some developers write more elegant code with only six months of experience, than many developers with ten years of experience and Ph.D.'s in Computer Science.

There is an interesting quote by Henry Ford about "experts":

“ None of our men are "experts." We have most unfortunately found it necessary to get rid of a man as soon as he thinks himself an expert -- because no one ever considers himself expert if he really knows his job. A man who knows a job sees so much more to be done than he has done, that he is always pressing forward and never gives up an instant of thought to how good and how efficient he is. Thinking always ahead, thinking always of trying to do more, brings a state of mind in which nothing is impossible. The moment one gets into the "expert" state of mind a great number of things become impossible. — From Henry Ford Sr., "My Life and Work," p. 86 (1922) ”

According to the Coding Horror blog most interviewed job candidates cannot program FizzBuzz.

Also note that being proficient and experienced may not automatically get you accepted into job openings because there are other factors that employers may find important.

The learn X in Y minutes site provides short introductions to computing topics which, although potentially useful, are likely not enough to become truly proficient.

To sum up, there is no clear cut answer and magical moment during which you transition from a "newbie" into a "super duper rock star ninja wizard awesomesauce hotstuff" dev. However, the more you learn and try, the better off you'll be with the right attitude.

What do you mean by "generating makefiles" or by "generating code" in general?

One can write code in some programming language that will emit plaintext, which in turn is valid code or markup of the same programming language or a different one that can be passed to its backend for execution. This is called "code generation".

For more information, see:

  1. GNU Autogen
  2. Text Processing Tools
  3. Metaprogramming

Will a change I would like to do break some functionality?

As the aphorism goes: The difference between theory and practice is that in theory, there is no difference between theory and practice, while in practice, there is.. There is usually a risk, however small, that a change will break some functionality. With good tooling (such as Version control (git-scm / etc.) , Virtual machines ( KVM , VMware, Xen, VirtualBox, etc. ) and Docker / etc. ), and having a comprehensive automated tests' suite. it should be relatively easy to revert a change which introduced regressions, and you should do adequate testing.

A change may have to be avoided due to being estimated as too time or money consuming, or as having too little gain. However, promising changes should be attempted because:

  1. "No guts - no glory."
  2. What does "if it ain't broke, don't fix it" really mean?
  3. If you never change anything, your project won't progress.

What does Object-Oriented Programming ( OOP ) mean?

When people say OOP they can mean one of several things:

  • A catch-all phrase for the paradigm.
  • Like Alan Kay noted: Actually, I made up the term object-oriented, and I can tell you I did not have C++ in mind. An important thing here is—I have many of the same feelings about Smalltalk
  • The "OOP/OOD" approach taken by C++ and later Java and C# and other languages, which was inspired by Simula.

As a result, the term has become somewhat muddled. Also note that if the language is Turing-complete, then usually one can implement OOP-like systems such as GObject or libcello for C.

What does Functional Programming ( FP ) mean?

Like OOP, FP can mean several things:

  1. The concept of programming without mutation or side effects. This was modelled in Lambda Calculus, and can often also be done in imperative languages.
  2. A language that is "marketed" as a functional language, which includes non-purely functional languages such as Scheme or https://en.wikipedia.org/wiki/Standard_ML and excludes others which have functional programming capabilities such as Perl, Ruby, or JavaScript.
  3. The polymorphism implementation in such statically-typed languages such as Haskell, Standard ML, or OCaml.
  4. A functional code, i.e: one that has external quality: is mostly bug free, has acceptable performance, etc. as opposed to the elegance of the code, and other internal quality aspects.

Why do people keep speaking against [language]?

Because it is impossible to please everyone or as Bjarne Stroustrup noted: There are only two kinds of languages: the ones people complain about and the ones nobody uses.

Furthermore, rindolf identified several tradeoffs in programming language design which means no perfect language can ever exist, and for every chosen language design, some people will dislike it or complain about it.

So some people will always complain, and while their opinions may have merit, it does not imply the language is not otherwise decent and usable. One cannot afford to completely rewrite the whole codebase in the trendiest language every month or even every year (although gradual conversions are a less suicidal option), so many languages will linger-on due to legacy.

How can I handle several different tasks / events (e.g: replying networking requests) at the same time?

You likely want an event-driven programming approach using an events / asynchronous library or framework. Some examples for those are:

  1. libuv and the older libevent for C.
  2. Twisted (which may be somewhat baroque), and asyncio for Python (e.g. the pydle IRC library for asyncio).
  3. Some alternatives for Perl are listed on this page.
  4. Node.js is inherently event-driven.

There are other approaches for multitasking including multithreading, multiprocessing, distributed computing, as well as hybrid approaches.

How can I make [insert application or library here] from scratch?

We appreciate your enthusiasm, and often writing small-scoped systems' components "from scratch" is a useful exercise, for learning as well as for proof of implementation and reproducibility.

Tweet: “My friend Botje from freenode noted that some people on its #osdev chatroom, have implemented kernels from scratch, which supported enough of POSIX to bootstrap a GNU ( https://gnu.org ) system [as I suspected]. Now try this with CSS 2.1…”.

However, see:

  1. Carl Sagan: “If you wish to make an apple pie from scratch, you must first invent the universe,”
  2. “I thought using loops was cheating…” on reddit.
  3. "It’s not a Fooware - it's an operating system!"

Different operating systems, programming languages, and runtimes provide different facilities out of the box. For example, cpython3 has a bundled HTTP server. You are probably going to rely on prior art when implementing your task and that is OK.

Merging these 2 sides-of-the-coin, we get (from fortune-mod):

A young man wrote to Mozart and said:

Q: "Herr Mozart, I am thinking of writing symphonies. Can you give me any
   suggestions as to how to get started?"
A: "A symphony is a very complex musical form, perhaps you should begin with
   some simple lieder and work your way up to a symphony."
Q: "But Herr Mozart, you were writing symphonies when you were 8 years old."
A: "But I never asked anybody how."

How hard is it to implement my own compiler?

That depends on the source, target, and implementation languages. A compiler from brainfuck to C can be written in, say, Python in a few hours. On the other hand, an optimising compiler from C++17 to x86-64 machine code will likely take months or more to implement even by very competent programmers.

How can I achieve this task using Unix shell?

Often many programming tasks outgrow the Unix shells' functionality. In that case use python or similar.

How do I create an installable .exe or .msi out of my Python/etc. code?

First of all note that these solutions are not effective ways to hide the code's inner workings from end-users.

Otherwise, you can try using:

Will Artificial Intelligence (“AI”) eventually eliminate the need for paid human programmers' workers?

Many people asked us that, and these discussions have not been too fruitful.

First of all, note that a far more efficient way for you to earn money than working as a grunt software developer, is to request it from people (whether online or in reallife) in exchange for sessions of "amateur philosophy / comedy". These have also been called "education and entertainment", "geeking out", or even "love / knowledge".

You can think of that as "begging for donations with style".

Pro tip: "originality" is overrated: see "Commercial Real Person Fan Fiction (RPFs), crossovers and parodies as 2021 geek/hacker imperatives for revitalising the film industry".

Most of the best salaried employees are geeky, and will not work on projects they consider unimportant even for obscene wages. Geeks are also known as "amateurs": workers who love their work.

If you still want a job (say because you are bored), then there are many non-software-development openings.

As a case study, if you need money, then "rindolf" (see his contact info) encourages you to:

  1. Lookup his bank credentials on GitHub, and —
  2. Extract 25,000 US dollars , which should allow you to support yourself for many days.

(There are similar publicly-viewable lists on GitHub with bank credentials of many rich celebrities.)

Which music do you listen to while programming?

A lot of people ask us that. Different software developers have different musical tastes, and may find different music suitable to listen to while programming. Some of them may find it hard to concentrate while listening to music.

That put aside, here are some recommendations from our regulars:

Is it true that ##programming is more about Anime than programming? (Just like #anime is more about programming than Anime)

That's not really true. While we sometimes discuss popular culture and entertainment (including but not limited to Anime), most of our conversations are about programming, or related computer help. It's not our fault that your favourite Anime-themed channel discusses programming a lot, and that may be caused by the fact that the IRC medium and Freenode in particular are somewhat more "techie"-oriented. So don't expect to join ##programming and hear many quality discussions about Anime.

About this FAQ

This FAQ seems useful. Why isn't it in the channel's topic?

Only channel operators can edit the /topic and none of them have seen it fit to put it there. Some of the maintainers of this FAQ approve of it being there (and the licencing terms allow that.), but none of them are channel ops. So there you go.

Does this FAQ reflect the opinions of the channel as a whole?

##programming is a collective and often different members of it have different opinions, so there is often no unified opinion. As the old joke goes: "Ask N programmers what they think about X, and you'll get at least N+1 opinions." The contributors to this document tried to be as informative and non-bigotted as possible, but there probably is some bias.

If you do not approve of parts of this document, you can:

  1. Send a pull request or a patch.
  2. Fork it and maintain your own version (which you are entitled to do after accepting some preconditions such as giving attribution).
  3. Write a similar document of your own.

Licence/License

The textual content of this document is made available under the Creative Commons Attribution (CC-by) 4.0 International licence (or at your option - any later version). Whatever code is given here is made available under the MIT/Expat License. Share and enjoy!