Mike Milinkovich on Open Source Consortia @mmilinkov

Mike Milinkovich’s talk at OSS 2015 on “How the Eclipse Community Works”

TRENDS

  • Trend #1: Software über alles (ref: Software is eating the world)
  • Software is becoming the value-creating differentiator (impl: hardware is getting commoditized faster than software)
  • Trend #2: Quoting Immelt (GE): Every industrial company will become a software company
  • Milinkovich: Then, every company is becoming an open source company; also, software avg unit cost approaches zero
  • Many companies will not be able to pivot fast enough from license sales to new business models; they will fail

  • Trend #3: Top 3 brands in the world are Apple, Microsft, and Google — they are all software brands
  • Microsoft Windows as of now is more open than Android (because of open development process, not just license)
  • Of course, Java and Linux are much more open, both in terms of license and process
  • Need to distinguish openness of platform and ecosystem (two different variables)
  • Trend #4: Foundationitis (too many foundations now)
  • There used to be three foundations: FSF, Apache, Mozilla
  • Along came Eclipse, the first one to marry open source with corporations
  • Now it is getting crowded, an explosion of open source foundations, all of them consortia (trade associations)
  • Linux Foundation is spawning all kinds of foundations, taking a 15% sales cut, plus service charges
  • Managing a platform through a consortium makes the platform more trustworthy
  • [Also, the platform becomes non-differentiating. DR]
  • Trend #5: Developers developers developers!
  • Quoting O’Grady: The new kingmakers: How developers conquered the world
  • Developers decide on trends, sales. Example: SOAP was corporate but REST won by devs voting with their feet
  • Thus, developer recruitment is strategic

WHAT DOES ECLIPSE DO

  • Foundations = GOVERNANCE + infrastructure + IP management + projects & processes + ecosystem development
  • “Pay to play” (i.e. paying fee + investing resources, leading a project) is a winning strategy, creates skin in the game
  • Board setup of Eclipse created broad diversity (different types of members), beneficial to projects
  • Eclipse eventually gave way to Github (git at least, now struggeling over bug tracker)
  • Foundations used to compete on forges; now with Github and their money, this is a lost cause
  • IP management at Eclipse is “insane” (four full-time people ensuring “IP cleanliness”)
  • The Eclipse Foundation audits possible suppliers, may reject them; analyses full-depth bill-of-materials
  • Eclipse reliably has its release trains leave the station

ECLIPSE PRINCIPLES AND OPERATIONS

  • Eclipse board stays out of the operations of projects
  • Principle of meritocracy
  • Principle of transparency
  • Principle of openness
  • [Cf our principles of open collaboration. DR]
  • Eclipse is vendor neutral (implies that all infrastructure is open source software)
  • Vendor neutrality can be very annoying i.e. having to use Bugzilla over Jira
  • Freedom of action
  • Eclipse owns all trademarks (abuse of trademarks is #1 way of how companies try to screw communities)
  • To ensure independence, the ED can be fired by the board, but also needs to be able to withstand pressure
  • Corporate transparency (who contributes what, how much, needs to be very clear, easily accessible)
  • As a platform, Eclipse needs to be highly predictable (for follow-on work, products)

CORPORATE SHENANIGANS

  • How corporations are trying to play foundations…
  • Trick #1: Astroturfing (pretending to be open, etc. while you are not)
  • Trick #2: “Founder seats” i.e. special rights like founder seats appointed by specific companies
  • Trick #3: Flat fees for board seats (excludes small companies from leading, influencing)
  • Trick #4: Extra votes (e.g. Pivotal gets 5 votes), accumulation of board seats
  • Trick #5: Affiliate Gerrymandering
  • Trick #6: Trademark ownership (e.g. NodeJS trademark)

OPEN SOURCE CAPTURE

  • How the game is played: Use of more IP to capture the code (trademarks, patents, specificatios, certification, …)
  • Example: Java and Sun
  • [Cf: The Java IP Story https://dirkriehle.com/2011/06/30/the-java-ip-story/. DR]
  • In conclusion: It is possible to have a happy marriage but lots of things can go wrong…

One Reply to “Mike Milinkovich on Open Source Consortia @mmilinkov”

  1. Interesting summary to what must’ve been an interesting and insightful talk.
    Two claims that seem provocative that I’d love to read more about:
    * impl: hardware is getting commoditized faster than software
    Seems like there ought to be data to support this or not, perhaps profitability of hardware and software firms? Strikes me that at least there’s a great deal of variation within hardware and software. Other claim “software avg unit cost approaches zero” would indicate software commodification. How to characterize Apple probably matters. Other claim that it is a software brand — is it? Apple has always been differentiated on hardware as well.
    * Microsoft Windows as of now is more open than Android (because of open development process, not just license)
    I’m not a Microsoft watcher, but this is pretty surprising, I’d like to read about the specific form Windows open development processes now take.
    Another claim seems incomplete:
    * “Now it is getting crowded, an explosion of open source foundations, all of them consortia (trade associations)”
    I guess this must refer to new foundations, as pre-Eclipse ones mentioned, as well as SPI, OSI, Conservancy, and many non-generalist and non-US ones are charities. New open source charities are still being established, otherwise the status of US IRS 501(c)3 determinations for new open source charities would not be a concern to anyone. The trend is probably correct though, off the top of my head I can’t think of a really major new open source charity in the past few years other than The Document Foundation. I imagine part of this has to do with non-vendor-driven projects being on a slower growth path overall, and on a slower path towards a governance crisis and resulting attempt at resolution (even if that is astroturfing). Disclaimer: I’m on the board of Software Freedom Conservancy so say that is the best place for non-vendor-captured projects to call home. But probably the most interesting bullet points concern how Eclipse has attempted to be a non-captured trade association; I’d be curious to see a list of tradeoffs between that sort of trade association and charity foundations, from the perspective of the former (having heard the latter perspective many times).
    I would also enjoy reading more on how “abuse of trademarks is #1 way of how companies try to screw communities”, in part because trademarks have often put even charity open source foundations at odds with user communities.

Leave a Reply