Open source, minus the folklore.

Free software and open source software are programs whose source code you may read, run, change and share, under a licence that says so in writing. The licence is the whole trick: without one, code on the internet is just code you are not allowed to touch.

'Free software' comes from the GNU project and is about freedom, not price: run it, study it, change it, share it. 'Open source' was popularised in 1998 and is defined by the Open Source Initiative's Open Source Definition. The two lists of approved licences overlap heavily. The arguments about which name is better do not.

None of this means 'no rules'. It means the rules are public and you can read them, which is more than most terms of service can claim.

• GNU GPL (v2 / v3)

  Strong copyleft

  If you distribute a modified or combined work, it must be under the GPL too, source included.

  You may

  Use, modify, sell and redistribute, including commercially.

  You must

  Pass on the same freedoms and offer the source of what you distribute. Keep notices. GPLv3 adds an explicit patent grant and anti-lockdown terms.

  Watch out

  Linux itself is GPL-2.0 only. GPLv2 and GPLv3 code cannot always be combined.

  Read the licence text

• GNU LGPL

  Weak copyleft

  A GPL variant for libraries: other programs may use the library without becoming LGPL.

  You may

  Link the library from software under any licence, including proprietary.

  You must

  Changes to the library itself stay LGPL, and users must be able to replace the library with their own modified version.

  Watch out

  Static linking makes the replace-the-library requirement harder to meet.

  Read the licence text

• GNU AGPL v3

  Network copyleft

  The GPL, plus a clause for software run as a network service.

  You may

  Everything the GPL allows.

  You must

  If users interact with your modified version over a network, you must offer them its source.

  Watch out

  Many companies ban AGPL dependencies outright. This is the point.

  Read the licence text

• MIT

  Permissive

  Do nearly anything, keep the copyright notice.

  You may

  Use, copy, modify, merge, publish, distribute, sublicense and sell.

  You must

  Include the copyright and permission notice in copies.

  Watch out

  No patent grant, and no warranty: it is provided 'as is'.

  Read the licence text

• BSD (2-clause / 3-clause)

  Permissive

  Like MIT, from the Berkeley tradition.

  You may

  Use, modify and redistribute in source or binary form, including in proprietary products.

  You must

  Keep the notice and disclaimer. The 3-clause version also forbids using the authors' names to endorse derived products without permission.

  Watch out

  The old 4-clause 'advertising' version is a different, GPL-incompatible licence.

  Read the licence text

• Apache License 2.0

  Permissive

  Permissive, with an explicit patent grant.

  You may

  Use, modify and redistribute, including commercially, and a licence to the contributors' patents covering their contribution.

  You must

  Keep notices, state significant changes, and include the licence. The patent grant ends for anyone who sues over patents in the work.

  Watch out

  Apache 2.0 is compatible with GPLv3, but not with GPLv2.

  Read the licence text

• Mozilla Public License 2.0

  File-level copyleft

  Copyleft per file: change an MPL file and that file stays MPL.

  You may

  Combine MPL files with code under other licences in a larger work.

  You must

  Make the source of modified MPL files available under the MPL.

  Watch out

  Copyleft stops at the file boundary, so the rest of your project can stay as it is.

  Read the licence text

1. Pick something you already use. You know where it hurts. Documentation typos and unclear error messages are valid first contributions.
2. Read the project's rules. README, CONTRIBUTING, CODE_OF_CONDUCT and the licence. Five minutes here saves a rejected pull request.
3. Find or open an issue. Look for labels such as 'good first issue'. For anything bigger than a typo, comment first so nobody builds the same thing twice.
4. Fork, branch, build, run the tests. Confirm it works before you change it, so you know whose fault a failure is.
5. Make one small change. One concern per pull request. Follow the existing style and commit message conventions, even the ones you disagree with.
6. Explain what, why and how you tested. Link the issue. Include screenshots for visible changes. Reviewers are reading this between other things.
7. Check for sign-off requirements. Some projects ask for a Developer Certificate of Origin sign-off (git commit -s) or a contributor licence agreement.
8. Take review gracefully. Expect requests for changes, and sometimes a polite no. It is about the code and the project's direction, not you.

• Hello everybody out there using minix - I'm doing a (free) operating system (just a hobby, won't be big and professional like gnu) for 386(486) AT clones.

  The post to comp.os.minix that started it. He had considered the name Freax; the FTP administrator, Ari Lemmke, put it in a directory called linux instead.

  Source: Wikipedia: History of Linux

• I still maintain the point that designing a monolithic kernel in 1991 is a fundamental error.

  From the 'LINUX is obsolete' thread. Linux has since been a monolithic kernel on most of the world's servers, phones and supercomputers.

  Source: O'Reilly, Open Sources: the Tanenbaum-Torvalds debate

• Starting this Thanksgiving I am going to write a complete Unix-compatible software system called GNU…

  The GNU announcement. Its own kernel, the Hurd, took long enough that Debian GNU/Hurd 2025 arrived in the 2020s, described by Debian as 'not an official Debian release'.

  Source: GNU Project: initial announcement | Debian GNU/Hurd news

• Talk is cheap. Show me the code.

  The unofficial motto of every code review since.

  Source: LKML archive | Wikiquote

• WE DO NOT BREAK USERSPACE!

  The kernel's most firmly enforced rule: changes must not break programs that already run. Capital letters were considered necessary.

  Source: lore.kernel.org

• Nvidia has been the single worst company we've ever dealt with.

  Said at Aalto University. He followed it with a gesture and a word we will not reproduce here. Driver support has improved since; opinions, perhaps less so.

  Source: Wikiquote: Linus Torvalds

• Only wimps use tape backup: _real_ men just upload their important stuff on ftp…

  A backup strategy that depends on strangers mirroring your files. Cheaper than tape, and it worked out for exactly one person.

  Source: LKML hypermail archive

• Linux is a cancer that attaches itself in an intellectual property sense to everything it touches

  Said to the Chicago Sun-Times about the GPL. Microsoft has since shipped a Linux kernel inside Windows (WSL).

  Source: The Register

• quite a sick place to be in

  His description of the Linux kernel community, as reported by The Register, during the height of the systemd arguments.

  Source: The Register

• the random number generator in Debian's openssl package is predictable

  A Debian-specific change had weakened OpenSSL's random number generator, so keys generated with it could be guessed (CVE-2008-0166). Patching upstream code is fine; patching it without upstream is how you get a funeral.

  Source: Debian security-announce

• rm -rf /usr /lib/nvidia-current/xorg/xorg

  One stray space. The author meant /usr/lib/nvidia-current/xorg/xorg, and instead asked for /usr to be removed. The commit that fixed it apologised.

  Source: Bumblebee commit a047be8

• Moved ~/.local/share/steam. Ran steam. It deleted everything on system owned by user.

  The line blamed in the report was rm -rf "$STEAMROOT/"*. Quote your variables, and never trust one you have not checked.

  Source: GitHub: steam-for-linux issue 3671

• The upstream xz repository and the xz tarballs have been backdoored.

  He noticed because SSH logins were using more CPU than expected, and kept pulling the thread. One engineer with a profiler may have saved a very large number of servers.

  Source: oss-security mailing list

• I know now this was not OK and I am truly sorry.

  His apology for how he had treated people on the kernel mailing list, announced together with a break from maintainership. Included because the community is allowed to grow up too.

  Source: LWN.net