Slashdot Mirror

← Back to Stories (view on slashdot.org)

The Myth of the Isolated Kernel Hacker

Posted by CmdrTaco on from the get-outta-yer-mom's-basement dept.

Ant writes "The Linux Foundation's report (PDF) on who writes Linux — "... Linux isn't written by lonely nerds hiding out in their parents' basements. It's written by people working for major companies — many of them businesses that you probably don't associate with Linux. To be exact, while 18.2% of Linux is written by people who aren't working for a company, and 7.6% is created by programmers who don't give a company affiliation, everything else is written by someone who's getting paid to create Linux. From top to bottom, of the companies that have contributed more than 1% of the current Linux kernel, the list looks like this: ..."

13 of 282 comments (clear)

  1. the list Before a karma whore can... by Anonymous Coward · · Score: 1, Informative

    Here's the list: 1. Red Hat: 12.3% 2. IBM: 7.6% 3. Novell: 7.6% 4. Intel: 5.3% 5. Independent consultant: 2.5% 6. Oracle: 2.4% 7. Linux Foundation: 1.6% 8. SGI 1.6% 9. Parallels 1.3% 10. Renesas Technology: 1.3% 11. Academia: 1.2% 12. Fujitsu: 1.1% 13. MontaVista: 1.1% 14. MIPS Technologies: 1.1% 15. Analog Devices: 1.0% 16. HP: 1.0%

    1. Re:the list Before a karma whore can... by millwall · · Score: 5, Informative

      At least attempt to format the list, mate:

            1. Red Hat: 12.3%
            2. IBM: 7.6%
            3. Novell: 7.6%
            4. Intel: 5.3%
            5. Independent consultant: 2.5%
            6. Oracle: 2.4%
            7. Linux Foundation: 1.6%
            8. SGI 1.6%
            9. Parallels 1.3%
          10. Renesas Technology: 1.3%
          11. Academia: 1.2%
          12. Fujitsu: 1.1%
          13. MontaVista: 1.1%
          14. MIPS Technologies: 1.1%
          15. Analog Devices: 1.0%
          16. HP: 1.0%

    2. Re:the list Before a karma whore can... by asdir · · Score: 2, Informative

      Further down in the article they write about some code contributed by Volkswagen and some GPS-company.
      In other words: RTFA

    3. Re:the list Before a karma whore can... by digitalunity · · Score: 2, Informative

      It's not in the public domain. For any patch of sufficient size, the copyrights still remain with each individual developer(I'm ignoring derivative works, as that's beyond the scope of a slashdot comment).

      So let's say IBM contributes a new driver. IBM keeps the copyright to the code, but gives license to use to all Linux users via the GPL. The key here is that because IBM still has that copyright, they are free to release that same driver as a closed source product as well, whereas Sun or Oracle couldn't take IBM's driver and sell a closed source version of it.

      Microsoft has relased a few open source applications but I'm 99% sure they were PR stunts and they have zero interest in being active contributors to Linux success. I wouldn't expect to see Apple anywhere on the list either considering OSX is primarily BSD based.

      --
      You can't legislate goodness. Let each to his own destiny, by will of his freely made choices.
    4. Re:the list Before a karma whore can... by osu-neko · · Score: 2, Informative

      Actually, it's accurate, but your thinking appears to be a bit fuzzy. If you want to compare individual things, the largest single contributor is Red Hat at 12.3%. If you want to group things up, then yes, unaffiliated individuals constitute 18.2%, but corporate-affiliated individuals constitute 74.2%. What isn't a sensible comparison is grouping on one side but not the other. To say the largest contributor is unaffiliated individuals makes no more sense than saying the largest contributor is companies that contribute less than 1% (together they contribute about a quarter of the contributions, which is larger than 18.2%).

      --
      "Convictions are more dangerous enemies of truth than lies."
    5. Re:the list Before a karma whore can... by digitalunity · · Score: 2, Informative

      It's obviously relevant.

      CANBUS is short for Controller Area Network. It allows different automotive sensors to communicate with subsystem modules such as anti-lock braking, powertrain control modules, and engine control units. This is a good case for where a company that uses Linux, but whose product is NOT information technology specific might want to be involved in the development of Linux.

      Real-time linux has had a big push from integrated systems development, where it's either "on-time" or it fails. Automotive is just another sector that can benefit from this. Any company that wants the most out of Linux is well served by having one or two full time developers whose sole purpose is to contribute to Linux and other open-source projects they rely on to be sure they continue to meet their requirements.

      Already, too many companies use Linux for the core products of their business but contribute nothing back and will constantly be chasing their tail to keep their own modifications synchronized with new kernel releases. VW did the right thing and all you can think of to do is troll.

      --
      You can't legislate goodness. Let each to his own destiny, by will of his freely made choices.
  2. Re:You know what company is shamefully absent? by tristanreid · · Score: 2, Informative

    FUD much?
    Since you're currently at +5 Insightful, I have to point out that they're actually on the list, the poster above cut it off at 1%, they're .8%.

    Also from TFA, there's another list of companies that do sign-off patches. Google is at 10.5% on that list, behind only Red Hat, above Novell, Intel, and IBM.

    To put it in perspective, the list doesn't include Linus on the list of contributors (he doesn't make the cut), but it does list him on the sign-off patches list.

    Just FYI,

    -t.

  3. Re:You know what company is shamefully absent? by sapphire+wyvern · · Score: 5, Informative

    They're a little bit further down.

    The next two rows on the list in TFA are as follows:

    17: Freescale 1,375 0.9%
    18: Google 1,261 0.9%

    I'm not sure why the parent decided to stop where they did.

    These rankings are based on number of kernel changes submitted broken down by employer.

    However it seems that Google employees are making a significant contribution to Linux project management and quality processes though: Red Hat employees sign off on over 36.4% of changes, which is the highest proportion of sign-offs in the hands of a single company, but Google has second place in that table with 10.5% of all sign-offs. It looks like several Google employees are filling the roles of subsystem maintainers - they may not write as much code as some other companies but they are still contributing some senior people.

    Interesting stuff!

  4. Re:GPL good for business by the_womble · · Score: 4, Informative

    whether you can start your own for-profit software business if you license your software under the GPL.

    Depends on what the alternatives are and what your business model is. Assuming we narrow down the choices to the two best known open source licences (others are broadly similar to one of the other) and proprietary licensing:

    • Want to do all the development yourself, distribution yourself, not using GPL licensed libraries, want to make your money from license sales? Proprietary.
    • Want to accept community contributions and sell a proprietary version, happy for competitors to use your code, do not need GPL libraries, want others to redistribute? BSD or GPL and persuade contributors to sign over the copyrights, or cleanly separate open and proprietary components and LGPL.
    • Do not want to sell a proprietary version, want to use GPL licensed libraries, want to accept outside contributions, want to prevent people from reselling your code without contributing back, want other to distribute? GPL
    • Doing all the development yourself, but want to use GPL libraries, want others to redistribute, want to prevent competitors reselling your code without contributing? GPL
    • Want to re-assure users that the software will still be around if you go bust? GPL or BSD. Additionally want to stop competitors reselling a proprietary version? GPL.

    Obviously this does not cover anything like all the possibilities, but I just want to make the point that there are business reasons for every choice.

  5. Re:shocking by bfields · · Score: 4, Informative

    "I doubt they contribute to the kernel itself at all."

    They do, see below--just not as much as some others.

    $ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
    $ cd linux-2.6
    $ git shortlog --author="@canonical.com" --author="@ubuntu.com"  --since="6 months ago"
    Andy Whitcroft (12):
          checkpatch: make in_atomic ok in the core
          checkpatch: do not warn about -p0 patches when checking files
          checkpatch: correctly handle type spacing in the face of modifiers
          checkpatch: pointer type star may have modifiers following
          checkpatch: a modifier is not an identifier at the end of a type
          checkpatch: extend attribute testing to all modifiers
          checkpatch: add __ref as a sparse modifier
          checkpatch: version 0.28
          Input: synaptics - ensure we reset the device on resume
          suspend: switch the Asus Pundit P1-AH2 to old ACPI sleep ordering
          mmc: add MODALIAS linkage for MMC/SD devices
          acer-wmi: Cleanup the failure cleanup handling

    Colin Watson (1):
          parisc: expose 32/64-bit capabilities in cpuinfo

    Leann Ogasawara (1):
          x86: add Dell XPS710 reboot quirk

    Luke Yelavich (1):
          ALSA: hda - add another MacBook Pro 3,1 SSID

    Scott James Remnant (13):
          [SCSI] ch: Add scsi type modalias
          sbus: Auto-load openprom module when device opened.
          netfilter: auto-load ip6_queue module when socket opened
          netfilter: auto-load ip_queue module when socket opened
          [MTD] Auto-load mtdchar module when device opened.
          [MTD] Auto-load nftl module when device opened.
          V4L/DVB (10947): Auto-load videodev module when device opened.
          floppy: provide a PNP device table in the module.
          applicom: Auto-load applicom module when device opened.
          cyclades: Auto-load cyclades module when device opened.
          specialix: Auto-load specialix module when device opened.
          usb: Auto-load cdc_acm module when device opened.
          riscom8: Auto-load riscom8 module when device opened.

  6. Re:Where is the missing 24.1%? by slyn · · Score: 3, Informative

    Individuals and anonymous contributors make up for the remaining 24% according to the youtube link posted in the first comment thread.

  7. Re:shocking by AvitarX · · Score: 2, Informative

    Didn't they have a lot to do with upstart too?

    I mean that has to count for something as others are starting to use it.

    And I would think pulling everything together into a cohesive package has some value too, but maybe it doesn't.

    --
    Wow, sent an e-mail as suggested when clicking on "use classic" banner, and got a fast response that addressed my msg
  8. Re:shocking by bfields · · Score: 2, Informative

    $ git shortlog -s --since="6 months ago" --author="@debian.org"
              2 Andres Salomon
              2 Bastian Blank
            65 Moritz Muehlenhoff
              1 Ron Lee
              1 dann frazier

    (And googling around suggests Mortiz Muehlenoff probably isn't a canonical employee.) Sure, people could also be using personal addresses, etc., though note the addresses come from git metadata, not necessarily "From:" addresses. (So, e.g., in my case, I use an email address from some personal domain on all my mail, but it's so easy to use my work address on git metadata (and forward it to my regular address so mail to it still gets through), that I can't see why I shouldn't do that minimum to give some credit to the employer that has me working on this stuff.)

    Also, the same since-6-months search for @redhat.com gets over 1400 commits....

    I like debian and ubuntu (I'm posting this from a ubuntu machine), but I think the general impression that they're not one of the larger kernel contributors is correct.