Ubuntu is a distribution that people have strong feelings about, both pro and con. Last week, those feelings erupted again after former Red Hat employee and Fedora community architect Greg DeKoenigsberg ranted about Ubuntu’s contributions to the GNOME desktop in his blog.
DeKoenigsberg has since apologized, but the issue is still worth a closer look, because it raises several issues about how the free software community works and one of its unspoken expectations.
DeKoenigsberg based his comments on Dave Neary’s analysis of contributions to the GNOME 2.30 release. Speaking at GUADEC, the yearly meeting of GNOME programmers, Neary was interested primarily in patterns in GNOME development.
However, what DeKoenigsberg noticed was the difference in the contributions made by Red Hat versus those by Canonical, Ubuntu’s commercial arm. While Red Hat had made 16.3% of all contributions, topping the list of corporate contributors, Canonical’s contributions comprised only 1.03%.
DeKoenigsberg now says that he wrote the now-infamous blog entry because of his lingering resentment over Canonical founder Mark Shuttleworth’s description in 2007 of Red Hat as a proprietary company. He now thinks that the entry was rash, and that he should not have written it.
However, whatever DeKoenigsberg’s motivations and second thoughts, his comments amounted to an attack on Canonical. Calling Canonical a “marketing organization masquerading as an engineering organization,” DeKoenigsberg described it as being hypocritically adept at claiming credit for promotion of the Linux desktop.
“One of the most irritating things about working at Red Hat was watching Canonical take credit for code that Red Hat engineers wrote,” DeKoenigsberg says. “Of course, Red Hat engineers, being the upstanding sort of chaps that they are, never said a word about it, because they’ve always been too busy carrying the load.” But the truth, DeKoenigsberg alleged, was that, “Canonical has been riding on Red Hat’s coattails for years.”
In less than four days, the blog entry attracted over 300 comments, to say nothing of numerous posts elsewhere agreeing or disagreeing with him — sure signs of how strongly the community cares about the issues behind his comments
The Canonical responses
DeKoenigsberg is far from the first community figure to denounce Canonical and Ubuntu. Most notably, in his keynote at the Linux Plumber’s conference in 2008, kernel developer Greg Kroah-Hartman noted that, over the previous three years, Canonical’s kernel contributions were .1% of the whole. This number placed its level of participation far below those of comparable distributions, such as Red Hat and Novell.
Since 2008, cynicism about Canonical has become commonplace, to the point where Ubuntu Developer Manager Scott James Remnant can talk about “this week’s popular anti-Ubuntu FUD” in a blog entry as though it were routine. The fact that his announcement that Canonical was hiring quickly turned into attacks on its hiring practices in the comments section suggests that his perception is more or less true.
However, DeKoenigsberg’s attack was the first from a major community figure in some time. Moreover, since DeKoenigsberg is no longer a Red Hat employee, or active in free software development, he felt no need to moderate his language. Consequently his comments were much more strongly worded than most criticisms of Canonical.
But, whatever the reason, responses from Canonical were quick in coming. Canonical Chief Operating Officer Matt Asay tweets that Neary’s analysis “tracks *all-time” Gnome contributions. Canonical will never catch up w/ RHT. It’s not helpful data.” Asay is referring to the fact that Red Hat was already founded when GNOME began in 1997, while Canonical did not exist until 2004.
Ubuntu Community Manager Jono Bacon gives a more detailed reply when he blogs that Neary’s original analysis does not include contributions to GNOME that are “(a) not part of official GNOME modules, and (b) hosted and developed elsewhere, such as Launchpad [Canonical’s development site].”
Nor, Bacon adds, does Neary’s analysis include applications developed to run on GNOME that have not been accepted into the GNOME project, such as Ubuntu’s Simple Scan. Furthermore, while Canonical has not contributed heavily to GNOME, Bacon also points out that the company and the Ubuntu community have made “significant upstream investment in other areas such as Upstart, Bazaar, Launchpad, and a full team building Ubuntu.”
However, by far the strangest reply is from Canonical founder Mark Shuttleworth. Instead of addressing DeKoenigsberg’s comments, Shuttleworth attempts to deflect them obliquely by warning about the dangers of “tribalism,” which he defines as “when one group of people start to think people from another group are ‘wrong by default’.”
Tribalism, Shuttleworth explains, is the origin of racism and sexism and “makes you stupid.” By contrast, Shuttleworth describes Canonical and Ubuntu as a place where the prevailing values are to treat everybody with respect. He urges people “Do not be drawn into a tribal argument on Ubuntu’s behalf” and “hold fast to what you know to be true” — that is, to pride in Ubuntu’s leadership and contributions to free software.
“The Gregs are entitled to their opinions,” he writes (his only direct reference to DeKoenigsberg), but, so far as he is concerned, people like Bacon have already answered the criticism and moved beyond it, and so should everyone else.
In a subsequent blog, Shuttleworth accepts DeKoenigsberg’s apology, adding that “We should start every discussion in free software with a mutual reminder of the fact that we have far more in common than we have differences.”
More than an argument
It is easy to pick out the flaws on both sides of this issue. On the one hand, DeKoenigsberg responds as though Neary’s analysis and one or two other indicators were enough to condemn Ubuntu, when clearly they are not. It is this incompleteness that Jeffrey Stedfast is apparently responding to when he mocks DeKoenigsberg’s comments.
On the other hand, the responses from Ubuntu are just as incomplete. Asay may have a point that Canonical started contributing later than Red Hat, but the contributions are still disproportionate when allowances are made for the fact. Similarly, Bacon ignores the fact that one reason why Canonical contributions are not being accepted into GNOME is that the company has chosen to proceed on its own rather than at the slower pace at which the GNOME project moves.
As for Shuttleworth’s comments, his condemnation of DeKoenigsberg is justified, but his own attempt to assert superiority is as good an example of tribalism as anything that he is responding to.
In the end, neither position seems solid, and both DeKoenigsberg and Shuttleworth should be congratulated for making sure that the argument did not get further out of hand (although I have no doubt that on some level, others are continuing it somewhere on the Internet).
Still, the argument is worth noting for what it implies about the community. On the most basic level, it demonstrates — if any proof is needed — the passionate feelings that Canonical and Ubuntu provoke. After six years, Canonical and its successes are apparently still viewed by many as upstarts. While supporters see Canonical as taking free software to new levels of acceptance, others see it as taking advantage of the community’s success, even though — as Stedfast points out — nothing in free software’s official rules should make that unacceptable.
Perhaps the commercialization of free software is partly to blame. The argument, after all, centers on two companies that work with free software. According to Neary, volunteers made 5% more contributions to GNOME than Red Hat, so you could argue by the same logic that was used in the argument that Red Hat is therefore taking advantage of the community. Yet, so far as I have seen, nobody is making that argument. Not that flame wars don’t happen frequently within the community, but could the tribalism that Shuttleworth decries be so intense because profit is involved?
More importantly, the two sides in this argument represent two very different views about what constitutes good citizenship among free software developers.
Partly, as Dylan McCall suggests, GNOME “can’t, as a community, decide whether they like the idea of external projects building new environments on the Gnome platform. . . . I think there’s one camp that thinks Gnome should be a user-facing product, with its own special branding and its own distinctive look that everything ships in pristine condition . . . . Then there’s another camp that sees Gnome as a starting point with lots of handy tools (and common modules) for distributions to build operating systems. . . . That first camp sees Gnome as a monolithic project; only internal work is worthy. The latter camp sees Gnome as something akin to GNU.”
However, the issue goes far beyond GNOME, and into the unspoken traditions of free software in general. Traditionally, the place for distributions to make changes is upstream — that is, in the component projects themselves. I remember, for instance, when I was working on Progeny Debian ten years ago that when the question of whether the company had the right to make changes to GNOME within the distribution (instead of contributing those changes to the GNOME project) the issued seemed important enough that we spent an afternoon discussing it.
Good free software citizens, the argument went, should make their modifications available for everyone to share. Progeny only went ahead with the changes after a survey revealed that what we were doing was already considered acceptable within the community.
But that is exactly what Canonical and Ubuntu have not done in recent years. In the rush to increase usability and to move towards profitability, Canonical has not been waiting to contribute its changes in notifications or title bar buttons or anything else back to GNOME.
Instead, Canonical has pushed ahead with its changes while keeping them within the Ubuntu distribution. It has every right to do so, of course. And perhaps its efforts will, in the end, redefine the degree of unilateral changes that are acceptable in a distribution. Meanwhile, though, the response that such efforts provoke should not be surprising, especially from those like Red Hat that have generally been more careful about contributing changes upstream.
Behind DeKoenigsberg’s anger and the responses from Canonical lies, if not a cultural clash, then a clash of changing values. The argument itself hardly matters, but the values on each side and — more importantly — which prevails could wind up mattering far more than who was right or wrong.