Saturday, January 22, 2011

Ubuntu: Where Did the Love Go?

When Ubuntu first appeared, the free and open source software (FOSS) community was delighted. Suddenly, here was a distribution with the definite goal of usability, headed by a former space tourist who not only understood computer programming but had the money to throw at problems.

The only objections were that Ubuntu was ripping off Debian, the source of most of its packages. For everyone else, Ubuntu and its parent company Canonical seemed everything FOSS had been waiting for.

Now, in 2011, that honeymoon is long past. Although Ubuntu remains the dominant distro, criticisms of its relationship with the rest of FOSS seem to be coming every other month.

What happened? Ubuntu supporters sometimes dismiss the change as jealousy of Ubuntu's success.

But, although that may be an element, the change in attitude is probably due chiefly to the gap between the expectations created by Ubuntu and Canonical in their early days and their increasing tendency to focus on commercial concerns.

Instead of being the model corporate member of the community that it first appeared, today Ubuntu/ Canonical increasingly seems concerned with its own interests rather than those of FOSS as a whole. No doubt there are sound business reasons for the change, but many interpret it as proof of hypocrisy. Added to the suspicion towards the corporate world that lingers in many parts of the FOSS community, the change looks damning, especially when it is so clearly documented in Canonical's corporate history.
A Brief History of Canonical and Ubuntu

After Ubuntu's first release in October 2004, Ubuntu/Canonical seemed in many ways a model FOSS entity. Nor was there much reason to doubt that initial sincerity. Shuttleworth, in particular, who was then the main speaker for both Ubuntu and Canonical, made considerable efforts to express support for other aspects of FOSS.

For example, Shuttleworth emphasized that "we all win, when Red Hat has a win." He made a special point of attending DebConf, Debian's annual conference, and of insisting that "Every Debian developer is also an Ubuntu developer" at a time when relations between Debian and Ubuntu were strained.

However, even in the first years there were signs of isolationism. Ubuntu/

Canonical insisted on using the proprietaryLaunchpad for development rather than existing free tools. Launchpad components did not begin to be released under free licenses until 2007, and the entire code was only released under the Affero GNU General Public License in 2009.

Similarly, in November 2006, Shuttleworth himself created controversy when he invited openSUSE developers to join Ubuntu. Although Shuttleworth later claimed that the offer was a response to Microsoft and Novell's cooperative agreements (Novell being openSUSE's corporate sponsor), it was widely condemned as an effort at corporate raiding unprecedented in the FOSS world, and Shuttleworth apologized a few days later.

However, the real turning point in Ubuntu/ Canonical policy appears to have been Shuttleworth's failure to convince other FOSS projects to coordinate their release cycles.

Shuttleworth first made the case in December 2006 that "it would be nice at the beginning of an Ubuntu release cycle to have a really confident picture of which projects will produce stable releases during those few months when we can incorporate new upstream versions. It would be even better if, during the release cycle, we knew immediately if there was a *change* in what was going to be released."

Over the next few years, Shuttleworth continued to stress the advantages of coordinated releases, arguing that it would allow centralized bug tracking, and suggesting that the cooperation might extend to common training materials.

The FOSS response, though, showed a distinct lack of interest. Many, including KDE's Aaron Seigo, saw the suggestion as squeezing projects into a uniformity that might not fit their needs.

Faced with this unenthusiastic response, Shuttleworth used a keynote at the O'Reilly Open Source Convention in July 2008 to urge a different approach to cooperation, challenging the community to rival and surpass Apple in usability within the next two years.

Given Ubuntu's emphasis on usability and Shuttleworth's own interest in interface design, this challenge was not unexpected. It fit, too, into the growing interest in usability at the time.

However, the FOSS community saw no reason to focus on usability under Shuttleworth's leadership, or within his schedule.

When changes proposed by Ubuntu were slow to be accepted in GNOME -- some say out of hostility -- Shuttleworth began making interface changes to GNOME within Ubuntu. They were accompanied by the announcement of an elaborate new look for Ubuntu that included complicated color codes and a new font.

Then, faced with the choice of supporting these changes in an old version of GNOME or transferring them to GNOME 3.0, Ubuntu announced Unity, a shell for GNOME that was originally designed for netbook computers, would be its new desktop.

This growing tendency to develop in-house has been accompanied by other signs of insularity. As early as April 2006, Ubuntu replaced init, the standard bootup program in GNU/Linux with Upstart, largely to reduce start times. In much the same way, in November 2010, it announced the eventual replacement of Xorg, which provides the graphical interface, with the mostly unproven Wayland.

Since both init and Xorg are flexible enough to provide the sorts of improvements that Shuttleworth advocates, the suspicion is that such decisions are not technical, so much as political. That is, what concerns Ubuntu/ Canonical is not the technical merits of the applications, but its ability to dominate the projects that dominate its software stack.

Other decisions that have negatively affected the perception of Ubuntu/ Canonical include the ending of Gobuntu, an Ubuntu variant that included only free software; a restrictive contributor's agreement; and the question of how affiliate fees for its online music store will be distributed.

The result of all these decisions and actions is an increasingly disillusioned perception of Ubuntu/ Canonical among some FOSS veterans. In 2008, kernel developer Greg Kroah-Hartman observed that despite its popularity, Ubuntu developers had contributed less than one percent of the patches to the kernel over the previous three years.

In much the same way, in August 2010, former Fedora community architect Greg DeKoenigsberg noted that only 1.3 percent of the code in GNOME 2.30 was attributable to Ubuntu. Currently, there is even a small group of regular critics who are sure to appear in the comments beneath any policy announcement made by Shuttleworth or his senior management. Despite its continued popularity among new users, at times Ubuntu has looked surprisingly like a pariah in the last few years.
The Reason for it All

This summary is enough to establish the change in Ubuntu's approach to FOSS. Yet what is missing is why it happened.

The idea that the changes are due to Jane Silber replacing Mark Shuttleworth as CEO is a tempting but incomplete explanation.

True, Silber is more business than FOSS oriented, and has emphasized projects like Ubuntu One, the cloud storage service, which is aimed more at corporate customers than individuals. But the change in attitude was happening long before Shuttleworth stepped down in December 2009 to focus on usability issues. Anyway, as founder, Shuttleworth remains active in decision-making. At most, the change in CEO only adds to an existing situation.

One possible pressing reason for the change is that, so far as anyone outside the company can tell, after seven years, Canonical remains unprofitable. Any investors other than Shuttleworth may be understandably concerned for their investment, and this pressure is probably a growing influence on Canonical decision-making.

Moreover, even if there are no impatient investors, a company that is unable to show a profit after seven years risks being identified as a failure.

However, another reason may be Shuttleworth's own frustration over his interactions with the FOSS community. In Shuttleworth's first blog entry, he states that "successful open source projects are usually initiated by someone with a clear vision and also the knowledge to set about turning that vision into reality," and mentions his strong interest in interface design.

Could Shuttleworth's calls for uniform release cycles and a focus on usability be not just practical suggestions for improvement, but also a bid to become the dominant leader in FOSS?

That is not a question you can ask anybody without the interview turning immediately hostile, but its answer might help to explain the change in direction.

If this really was part of Shuttleworth's motivation (and I do him the courtesy of assuming that he might also have been idealistic), then the bid for dominance not only failed, but failed specifically in an area that was personally important to him. For most people, such failures would be more than enough to explain an increasing insularity -- if you can't dominate the greater world, why not carry out your plans at home, and prove that they work there?

We will probably never know all the reasons why Ubuntu/ Canonical changed from the embodiment of FOSS hopes to more of a business enterprise. Given the high expectations that Ubuntu initially raised, perhaps the disillusionment is inescapable -- and discourages me and everyone else from sufficiently acknowledging that, in day to day operations, the spirit of FOSS seems very much alive at Ubuntu / Canonical.

However, at the level of strategic planning, the fact that something has shifted is undeniable. For those of us who remember the initial expectations, FOSS can only seem poorer for the fact.

No comments:

Post a Comment