Interested in building 'community' Mr Corporation? Please swipe here
Having a foundation doesn’t necessarily mean a great deal
Feature Money isn’t new to open source. Capitalist pigs have always been quite at home amidst the unwashed developer masses, spinning up companies to capitalise on the success of JBoss, MySQL, and more.
What’s different today is the lengths to which corporate interests go to feign “community".
In the early days of open source, for example, it was enough to trot out a .org developer site to match the .com corporate site. But today you need a foundation, which is why it seems we mint a new open source foundation every other week.
But do these newfangled foundations actually deliver on their promise of creating community from a morass of corporate, competing interests? It’s not clear.
Who’s your sugar daddy?
Those who advocate foundations insist on their value. Sam Ramji, executive director of the Cloud Foundry Foundation, took the stage at OSCON to lay out the reasons open source foundations have become so fecund of late.
The problem, he maintains, is that open source has become too popular for its own good. In response to the hypothetical question: “Why are there so many foundations,” he answers: “It’s economics. Open source has won. Popular projects are now economically attractive. This breeds competition and distrust.”
More specifically, Ramji points to pesky venture capitalists as fueling this distrust, arguing that as VCs get involved, wrapping heavily funded companies around popular open source projects, developer trust in those projects diminishes.
The tonic to all this carpetbagging cash, he suggests, is the open source foundation: “The move to multi-vendor open source enables competing corporations to collaborate.” No one wants to assign copyright to a competitor, so pooling IP interests in a foundation seems like the way to go.
Who made whom?
Implicit in this argument is the assumption that vendors offer the most compelling code; that their involvement is critical to open source success. I’m not sure this holds up. At least, not consistently.
For example, while IBM and others aggressively hired top Linux committers, they were only able to identify those top committers ... after the commits had been made. While the developers no doubt appreciated getting paid to contribute thereafter, the reality is they were contributing long before these corporate interests slapped an employee badge on them.
Regardless, even if we don’t accept that open source depends upon corporations getting along, need we accept that foundations are the proper way to achieve this?
My former colleague and long-time open sourceror Henrik Ingo believes so. Ingo holds that the biggest open source projects (Linux, KDE, Apache, Eclipse, Perl+CPAN, Mozilla+Addons, Gnome, Drupal and GNU) “are developed as collaborative community projects governed by non-profit foundations". In fact, he continues: “No single vendor project has so far been even close to reaching their magnitude.”
He then concludes, “There appears to be a glass ceiling limiting the growth of the large single-vendor projects,” projects such as MySQL, MongoDB, JBoss and so on. You know, the projects that dominate their respective categories.
It turns out that measuring the size of the developer community (Ingo’s approach) is an imperfect measure for accessing adoption and impact of the underlying project. MySQL, for example, is orders of magnitude more popular than its next nearest foundation-developed project.
Or take OpenStack, the number-one open source cloud, according to a 2014 Zenoss survey, and a model of corporate interests run amok in foundation land. As Joyent chief technology officer Bryan Cantrill posits: “One of the biggest problems with OpenStack is there is so little that companies agree on there ends up being so many decisions left up to the end user.”
Foundation? Check! Zillions of developers? Check! Fewer than 1,000 deployments? Check! (Meanwhile, single-vendor, non-open source “project” Amazon Web Services is growing at such a torrid pace that industry onlookers call its growth “bonkers.”)
Do you want to build a foundation?
Which is not to say that foundations are useless. It depends on the foundation in question.
The longer I’m in open source, the more I appreciate the Apache Software Foundation, which provides just enough help to projects like Cassandra and Storm to help them thrive, without burdening them with the pomp and circumstance of many other corporate foundations.
Even without an ideal foundation like Apache, foundations can help. As Twitter’s open source guru Chris Aniszczyk told me: “It's better in the long term [to have a foundation] than just having one company control the code; foundations set expectations.”
But we should resist the idea that slapping a foundation on a project somehow makes it community friendly. Too often a foundation’s “community” translates to “the moneyed interests the single-source open source vendor has on speed dial.” I know this. I’ve done this.
But that doesn’t make it right.
As Karen Sandler, executive director of the Software Freedom Conservancy, puts it: “Foundation models are [different]” and “may not be actually neutral". She’s right, and may be describing many (most?) of recent additions to the foundation outbreak.
All of which means that developers need to carefully evaluate the projects to which they’re inclined to contribute. Just because it has a foundation doesn’t mean it has a community. ®