Forgot your password?
typodupeerror
Linux Business The Almighty Buck Linux

75% of Linux Code Now Written By Paid Developers 368

Posted by timothy
from the rest-kept-in-locked-cellar dept.
i_want_you_to_throw_ writes "During a presentation at Linux.conf.au 2010 in Wellington, LWN.net founder and kernel contributor Jonathan Corbet offered an analysis of the code contributed to the Linux kernel between December 24 2008 and January 10 2010. The Linux world makes much of its community roots, but when it comes to developing the kernel of the operating system, it's less a case of 'volunteers ahoy!' and more a case of 'where's my pay?'" It's not clear from the article why anyone should perceive a contradiction between having high ideals and getting paid to do something you enjoy.
This discussion has been archived. No new comments can be posted.

75% of Linux Code Now Written By Paid Developers

Comments Filter:
  • How much does a line of code cost?
    • by Meshach (578918) on Thursday January 21, 2010 @07:08PM (#30853572)

      How much does a line of code cost?

      Cost-per-line is a patently bad way to compute the worth of code or value of a coder. Knowing what to code is more important then just writing the code. Features implemented or bugs fixed is probably a better measure.

      • Re: (Score:3, Insightful)

        From most of the Linux advocates I hear commenting on slashdot, there AREN'T bugs or missing features in Linux. So why the developers? ;)

        No, I'm serious. You'd think Windows couldn't stand on its own legs for two minutes before crashing due to the amount of bugs in the code, and you'd think Linux had no bugs whatsoever.

        I like Linux and have no problem with devs getting paid to work on it. Sound slike a good idea to me; in fact, it sound slike how almost every single product in the world is made, pretty m

        • by horza (87255)

          From most of the Linux advocates I hear commenting on slashdot, there AREN'T bugs or missing features in Linux.

          That is patently a lie.

          Phillip.

          • From most of the Linux advocates I hear commenting on slashdot, there AREN'T bugs or missing features in Linux.

            That is patently a lie.

            Phillip.

            I can't tell what level of indirection applies to the word "that" so:

            Semi-whoosh!

        • by Patch86 (1465427) on Thursday January 21, 2010 @08:11PM (#30854396)

          Value for money, my friend. My Windows and my Linux machines have, lets be honest, a relatively similar number of problems. Windows suffers from the most outright bugs, but then Linux can still sometimes throw a hardware or compatibility wobbly, and sometimes does suffer the occasional deeper problem.

          The difference is that one of them is distributed free over the internet, and the other cost me £150 and still delights in harassing how "genuine" I am every time I visit the developer's website.

          You tend to be far more forgiving when something is both free (beer) and, feels like it belongs to you instead of some distant oligarchy.

          • Re: (Score:3, Funny)

            by BitZtream (692029)

            You tend to be far more forgiving when something is both free (beer) and, feels like it belongs to you instead of some distant oligarchy.

            Only if what your doing is of little to no value. The Windows tax is trivial compared to what I use my Windows machines for. The genuine advantage thing used to bother me, but then I grew up and just realized that I got so much use out of the OS that the price wasn't really that bad. I'd spend more in 2 days getting drunk off my ass on the weekend when I lived in Orland

      • How much does a line of code cost?

        Cost-per-line is a patently bad way to compute the worth of code or value of a coder.

        Of course, which is probably why the question mentioned nothing about worth or coder value. Knowing that a line of code costs $1 (or whatever) on average, you can then start looking at various modules within the Linux kernel and project how much it would cost for an average developer to reimplement (on a BSD project, for example) based on the number of lines in that module.

        To get the relative value of a coder, you could take a sampling over a large enough period of time and then say that, on average, this

        • by shaitand (626655)

          'Of course, which is probably why the question mentioned nothing about worth or coder value'

          And yet, you outline a system that values of coder in a given area of expertise based on the number of lines of code produced.

          Even within 'a given area of expertise' the number of lines of code, or average number of lines of code produced by a developer is a meaningless number beyond determining if the developer is working at all.

          An unproductive and useless developer can churn out tons of code that doesn't actually d

          • Re: (Score:3, Funny)

            by drinkypoo (153816)

            And yet, you outline a system that values of coder in a given area of expertise based on the number of lines of code produced.

            LOC is never going to tell you the whole story, but in a system with some decent code review it will give you at least a pointer.

      • by creimer (824291)

        I have programmer friends who work in jobs where they have to produce a certain amount of code. (I'm assuming that comment lines don't count, which is why comments are non-existent in some software projects.) The underlying metric would be how much does a line of code cost. I'm not saying this is a good or bad metric. When I saw the summary, I thought it was missing some information.

        There are some PHBs who would argued that Linux was crap because no one can tell them this metric even though the software

    • varies with the line.

      base cost is the the coderss hourly wage. if a coder earning $50/hour codes 10 lines in an hour, cost is $5/line.

      now, for difficult lines, if he takes half an hour to finish the line, the cost for that one is $25.

      of course, theres more people involved in the process than just one coder.

      but the basics are the same.

      cost per line = (sum of all employees wages * time spent to code X lines ) / number of lines writen

      this will give you an average for the whole project.

      • by tsm_sf (545316)
        base cost is the the coderss hourly wage. if a coder earning $50/hour codes 10 lines in an hour, cost is $5/line.

        Doesn't this seem like a useless metric, though? You can't measure functionality by lines of code in a program (although, at some point, you could probably draw an inverse relationship). What purpose does this data serve?
    • Re: (Score:2, Funny)

      by jabberw0k (62554)

      How much does a line ... cost?

      First one's free?

  • by OeLeWaPpErKe (412765) on Thursday January 21, 2010 @07:05PM (#30853518) Homepage

    How many paid kernel developers does microsoft have ? How many does Sun have ? I can't find any straight numbers on the web.

    A thought strikes me, though. It seems unlikely to be more than a few dozen each, at most.

    • Far too few, but there is a maximum scrum in the Cathedral.

      The question goes to the Cathedral and the Bazaar dichotomy and the Brooks "Mythical Man Month" about OS 360, and how you count the dev team, kernel core or that and associated userland.

      DEC: Tops 10 2 x lead + 6 mostly

      AT&T Bell Labs: 3 x lead + 20

      VAX-VMS: 1-2 lead + 40 (inc RSX-11 drivers)

      WNT+: initially 1-4 lead + 20, clone VMS

      Linux: initially 1 lead + 0, now 1 + 25 leads +4000

      The 4000 number says it all. Rob Gingell, who used to run SUN's Sola
  • by Rantastic (583764) on Thursday January 21, 2010 @07:07PM (#30853556) Journal

    The Linux world makes much of its community roots, but when it comes to developing the kernel of the operating system, it's less a case of 'volunteers ahoy!' and more a case of 'where's my pay?'

    Since when does community == volunteers?

    That large, well funded corporations are now contributing members of the linux community is a Good Thing.

    • by nine-times (778537) <nine.times@gmail.com> on Thursday January 21, 2010 @07:17PM (#30853708) Homepage

      Good point. In a sense, you could say that these companies are "volunteers". They're each a group of people who are "scratching their own itch" and donating their resulting work back to the rest of the community.

      And even if 75% of Linux code is contributed by these companies, that still leaves 25% which isn't. If you think about it, that's actually kind of impressive. You have all these huge companies paying very good developers to build a robust professional-level kernel-- heavyweight companies like Intel, Oracle, IBM, Novell, and Redhat-- and still 25% of the code comes from individual unpaid developers?

      • by The_Wilschon (782534) on Thursday January 21, 2010 @07:30PM (#30853900) Homepage
        You might say that these companies have discovered that there is significant value (enough to pay some developers) in the existence/availability of good code, even if they do not derive any value from the sale of that code.
        • Right. Lots of the "free" stuff in the Internet age are cases where people have found an alternate way to make money from it. Google provides some great services by selling ad space. IBM and Intel are alway listed as big Linux code contributors, but they don't sell Linux. They sell hardware, and they want a good software to run on their hardware so that they can sell more hardware. IBM also sells lots of "e-business" services and they develop custom solutions, but they need good operating systems to ru

    • by BitterOak (537666) on Thursday January 21, 2010 @07:22PM (#30853794)

      Since when does community == volunteers?

      That large, well funded corporations are now contributing members of the linux community is a Good Thing.

      Exactly! What's great about Linux is that it's free, not that its developers are unpaid!

  • I rely upon Linux for my business. If something isn't all it should be, or developments don't happen as fast as they could, I'm gratified to know that money is changing hands and somebody might get canned and replaced by another, better professional.

    If Linux wants to sit at the adults' table -- and it clearly has the depth and breadth of functionality to do so -- then there needs to be the kind of professional accountability in its developers that only a paycheck can engender.

    • by horza (87255) on Thursday January 21, 2010 @08:16PM (#30854456) Homepage

      If Linux wants to sit at the adults' table -- and it clearly has the depth and breadth of functionality to do so -- then there needs to be the kind of professional accountability in its developers that only a paycheck can engender.

      Billions lost on failed UK IT projects by the 'adults' with developers receiving very fat paycheques shows it guarantees neither success of the project nor accountability within it.

      Phillip.

      • Re: (Score:3, Interesting)

        by RAMMS+EIN (578166)

        Billions lost on failed UK IT projects by the 'adults' with developers receiving very fat paycheques shows it guarantees neither success of the project nor accountability within it.

        That, and if you look at Unix-like systems, you will see that all of them are dead or dying, except for those that are being carried by volunteers (BSD, GNU, Darwin (mostly BSD), and perhaps OpenSolaris). If "the kind of professional accountability in its developers that only a paycheck can engender" is what a software project needs to "sit at the adults' table", then maybe sitting at the adults' table is not what you really want for your project. After all, all those adults are now either dead or on life s

    • by Zero__Kelvin (151819) on Thursday January 21, 2010 @09:30PM (#30855120) Homepage

      "If Linux wants to sit at the adults' table ..."

      Linux is the adults table. The adults all sit at it. You've heard of Google, IBM, Sun, Oracle, Novell?

  • by Mr. Slippery (47854) <tms@@@infamous...net> on Thursday January 21, 2010 @07:08PM (#30853564) Homepage

    There seems to be some assumption that "community" means "unpaid". Not at all. The Free Software community includes a whole lot of people who get paid to use software to meet the needs of employers. If meeting those needs involves improving bits of Free Software, the employer benefits from having those contributions integrated into the product.

    • by H0p313ss (811249)

      There seems to be some assumption that "community" means "unpaid". Not at all. The Free Software community includes a whole lot of people who get paid to use software to meet the needs of employers. If meeting those needs involves improving bits of Free Software, the employer benefits from having those contributions integrated into the product.

      I suspect that the provided statistic is true of a large number of *SUCESSFUL* open source projects, Apache and OpenOffice come to mind.

  • You see the same thing in academic publishing. There seems to be a sentiment that getting paid for an article would somehow compromise the objectivity of the writer. However, people contributing these articles are doing it in fields they study professionally, and it is often essential resume building work. This is not a situation I'd like to see mirrored in the computer world.
  • pay? (Score:5, Insightful)

    by gbjbaanb (229885) on Thursday January 21, 2010 @07:13PM (#30853652)

    ...it's less a case of 'volunteers ahoy!' and more a case of 'where's my pay?'"

    I'd say its more a case of "I get paid to do this? who-hoo".

  • > It's not clear from the article why anyone should perceive a contradiction between having high ideals and getting paid to do something you enjoy.

    Sure, it's cool to be able to say that you're paid to work on the Linux kernel. But how many of that paid 75% would do it for free? How many would have to do something else to put food on the table if there were not a corporation to pay them?

    What I take away from this is the fact that the Linux "community" is dominated by corporations. In many cases (but no
    • by gmack (197796)

      Considering most of them(Linus included) used to do it for free in their after work times? I'd say pretty good.

      Dominated is the wrong word.. Linus for instance refused to have a "Linux job" for years because he was afraid it would taint his decision making. Now he works for the Linux Foundation so hes guaranteed neutrality. A lot of other developers got hired because some corporation liked what they were doing but wanted it done faster so it's more a matter of Linux developers getting payed to do what th

    • Re: (Score:3, Insightful)

      by horza (87255)

      Sure, it's cool to be able to say that you're paid to work on the Linux kernel.

      Certainly is. I'd decided at the age of 8yrs old the first company I was going to work for was Acorn. And it was. My friend loves Linux and so picks jobs where he gets to play with top end Linux clusters. Previously at CERN and now a top Swiss bank. For a real techie the work is more important than the size of the pay cheque.

      But how many of that paid 75% would do it for free?

      Depends what the code being contributed is. IBM is port

      • I bet a large % of the paid developers are contributing code that is pretty useless to the home desktop user.

        I dunno. With the big contributors being Red Hat, Intel, IBM, Novell, and Oracle, I'd suspect that much of what some of them do -- particularly Red Hat and Intel and Novell -- has pretty broad impact on desktop users. (While IBM is supporting their mainframes, they do a lot of other things to, and I wouldn't be surprised if there contributions also had broad impact.)

        Note also that paid contributions

    • " What I take away from this is the fact that the Linux "community" is dominated by corporations. In many cases (but not all), for-profit corporations, all trying to compete against several other for-profit corporations named Microsoft, Apple, Google, Oracle, etc.

      Put it back. It's no good. (Among the many flaws in your analysis is your mixing of commodity and service based businesses, and the fact that Google is powered by Linux)

  • What about Google? (Score:2, Insightful)

    by netcruiser (1645001)
    So why isn't Google more involved in kernel development? I assume they use Linux extensively and hence make billions from using it. Do no evil, do no good?
    • Re: (Score:2, Funny)

      by igadget78 (1698420)

      So why isn't Google more involved in kernel development? I assume they use Linux extensively and hence make billions from using it. Do no evil, do no good?

      Nope. They use Microsoft and IE6.

    • by Shikaku (1129753)

      You know that thing called the GPL? If you distribute binaries outside you must distribute the source code?

      Google doesn't distribute outside. So while they ARE involved in Linux development, they keep things internal and send what code they want to send.

    • by pnewhook (788591)

      Seriously, how much kernel development needs to be done? The OS kernel is the foundation for everything else. If it needs major development, then the project is not very stable by definition.

      Given that Linux has been around for many years, it should be stable and hence need little if any changes. The majority of development should be on the application side.

  • 25% non-corporate? (Score:5, Interesting)

    by highways (1382025) on Thursday January 21, 2010 @07:25PM (#30853830)

    As part of my job, I port Linux to our embedded boards and occasionally hack a driver or two.

    However, in order not to scream out to our competitors "Hey! We're making a new product!", the small amounts of code I send pack at patches (it's a pain in are done so though a nondescript gmail account.

    I suspect this practice is fairly widespread. Therefore, I'd say that 75% is an under-estimate.

  • Oh, man... (Score:3, Interesting)

    by Anachragnome (1008495) on Thursday January 21, 2010 @07:31PM (#30853938)

    Oh, man. To be a fly on the wall when Ballmer reads THIS little line...

    "Within that field, Red Hat topped that chart with 12%, followed by Inte with 8%, IBM and Novell with 6% each, and Oracle 3%. Despite the clear commercial rivalry between those players, central kernel development worked well, Corbet noted."

    And everyone thinks the Faraday Cage around his office was to keep his signals safe. The boys in Security know it is really to keep the chairs in his office...well, in his office.

  • There are definitely plenty of paid coders on the kernel. But are they counting the kernel hackers that companies have chosen to sponsor as paid or as volunteer? Does a grass roots volunteer kernel hacker stop counting once a company sponsors him to be able to contribute full time?

  • by ClosedSource (238333) on Thursday January 21, 2010 @07:44PM (#30854088)

    What percentage of these paid developers work for a company that derives its revenue primarily from software development?

    • by Spit (23158)

      That's a good point and illustrates the advantage of free software from a user perspective. Users are spending money to get the features they need and lose nothing by those features being available to others gratis, especially when other's input is reciprocal.

      • I think you misunderstood my point, but in any case, I don't think users will be happy to pay for features only to find out that others are getting it free. Reciprocation is a nice theory, but too abstract and indirect for most customers to buy into.

    • by zzatz (965857) on Thursday January 21, 2010 @09:12PM (#30854996)

      That's the point - bypass the middleman's sales overhead and profit.

      On one hand, Company A buys software from Company B, indirectly funding the development of the software. If Company A wants changes or new features, they can beg and plead for them, and they might get them. Company A will indirectly pay for development at Company B whether or not they get the changes they want. Company B will then sell the software, possibly incorporating Company A's ideas and improvements, to all of Company A's competitors. Company B's customers pay the cost of the development, plus the cost of sales (marketing, commissions, etc.), plus a markup.

      On the other hand, Company A hires developers to improve software that others have made freely available. They get exactly the changes they want. Company A's competitors also get those changes, but the reverse is true: Company A gets Company C's improvements. Both companies find this agreeable because neither can gain an advantage through the software, and both have reduced the cost of developing it. Company A has cut out the middlemen, avoiding the cost of sales and profits extracted by Company B.

      You can't gain an advantage over your competition by buying your software from a third party, because your competitor can buy it, too. You can't gain an advantage over your competition by hiring developers to write open source software, because your competitor can dowload it, too. There's no difference between open source software and third party commercial closed source software as far as advantage over a competitor. The only way to use software as a competitive differentiator is to develop it internally, keep it closed, don't sell it, and pay the high cost of developing for a single customer - yourself.

      In economic terms, software is a complementary good. Intel sells processors, which are not useful without software. But every dollar spent on software is a dollar that isn't spent on processors. Red Hat is in a similar situation; they sell support, not software, and giving away software makes money available for support.

      The economics are simple. Any software that has a large enough base to support sales in binary form has a large enough base to support shared development under open source licenses with a lower overhead. Selling binaries is a temporary aberration caused by network effects during the initial growth of the market. As the market matures, sales of mass market software will decline.

    • by selven (1556643) on Thursday January 21, 2010 @10:14PM (#30855438)

      Red Hat 11.2%
      Novell 8.9%
      Linux Foundation 2.6%
      Oracle 1.3%

      (among others)

      Source: https://www.linuxfoundation.org/publications/linuxkerneldevelopment.php [linuxfoundation.org]

  • by SwordsmanLuke (1083699) on Thursday January 21, 2010 @07:51PM (#30854178)
    So who decided that the Open Source movement was about *not* making money? I thought it was about enlightened self-interest. If we make the source of today's apps available to the coders of tomorrow, everyone wins. Up-and-comers get a chance to see real-world (and sometimes, cutting-edge) code - and the community (of software developers) gets new devs who show up already knowing some of the things *we* had to figure out the hard way.

    The new guys get the benefit of our experience and in ten years, we get to hire better new guys.
  • ... businesses are learning that contributing to a shared resource has value. And that efforts made to monopolize resources (like patents do) aren't as valuable as was assumed in the past.

  • by LarrySDonald (1172757) on Thursday January 21, 2010 @09:58PM (#30855350)
    I use linux professionally. So does most of the web. We're "forced" to GPL any improvements we have to make in the process of getting the job done. "Forced" is in quotes because fair is fair - so did everyone else including those bat**** crazy people following Linus and Sallman who wrote the seeds that grew into this and frankly I feel I'm getting more then I could ever give (at best correcting the occasional bug). GPL is there so it's clear to the managers that if you have a problem with that, feel free to pay quite handsomely. It's cheaper to improve linux (and/or the rest of GNU) then it is to not use it. Epic score - that was the whole point all along, right?
  • by bug1 (96678) on Thursday January 21, 2010 @10:15PM (#30855440)

    It's not clear from the article why anyone should perceive a contradiction between having high ideals and getting paid to do something you enjoy.

    One day a situation will arise when you will be expected to do something you dont enjoy.

    You will choose between love and money, you will begin to discover how much your high ideals are worth.

"Ahead warp factor 1" - Captain Kirk

Working...