Ubuntu is probably the most popular Linux distribution ever. According to the 2007 Desktop Linux Survey, over 30% of Linux users run Ubuntu, and, if anything, the numbers have increased since then.
In June 2009, Ubuntu's users were estimated as 13 million, and the distro was described as "growing faster than any other distribution." When Dell began to preload laptops with Linux, it chose Ubuntu, and Ubuntu is the first and only distribution to have a print magazine dedicated to it. By any measure, Ubuntu is a major player in free software.
Yet this popularity has a flip side. If Ubuntu is the most popular distribution, it is also the most hated. Last year, a survey on the Linux Hater's blog listed Ubuntu as "the most hated community distro." Search Google for "why I hate Ubuntu," and 9260 hits are returned -- compared to 376 for the equivalent phrase for Debian, and 11 for Fedora.
The explanation seems to be that Ubuntu is attacked from several perspectives. Partly, the attacks may be simply a consequence of business as usual. But other grievances also seem to be aired when Ubuntu is attacked. They range from a perception that Ubuntu is stealing from Debian or is an upstart in free software in general to accusations that the distro is not a member in good standing of the community. But perhaps the largest reason is that Ubuntu is a victim of its own success, and has created expectations that it has yet to fill completely.
Complaints represent more eyeballs
Asked to explain the phenomenon, Jono Bacon, Ubuntu's community manager, opines that the negative view of Ubuntu is due to the way that free software is developed.
Quoting Eric Raymond, Bacon comments, "This harks back to one of the cornerstones of open source: given enough eyeballs, all bugs are shallow. When we kicked out a new release of arguably the world's most popular desktop distribution, there were more eyeballs, more hardware, more networks, more devices, more configurations, more expectations and therefore more opportunities for things to go awry inside these attributes, Bacon says. If we then combine this with the natural inclination for human beings to communicate complaints as opposed to share praise, it is not surprising to see reports of things not working."
In particular, Bacon comments on some of the reactions to the recently released Karmic Koala release, which has been heavily criticized for numerous problems.
Calling the Karmic release "adventurous," Bacon says, "All software ships with bugs; that is just the nature of how software works. Linux distributions face additional risk in this area as we pull in thousands of independent upstream projects, and we inherit their bugs, too."
In other words, more people using Ubuntu means that it is being used in more situations, so more problems come to light, especially when the distribution is adding so many innovations. Implicit in Bacon's comments is the idea that, in the long run, Ubuntu will be stronger for the complaints, especially since the distribution's members are working to improve its quality assurance and, presumably to improve their responses to such feedback.
The ire of the old timers
To a large extent, Bacon is right about the causes of the complaints. If you follow the links, many lead to complaints about bugs or features. But many of the attacks on Ubuntu have less to do with technology than with the politics of the free software community.
Of these political complaints, the oldest is that Ubuntu borrows from Debian without giving enough back. Ubuntu not only uses packages from Debian's testing repository for its releases, but Canonical, Ubuntu's commercial arm, has hired many Debian developers. Although these developers often continue to work with Debian packages, over the years Debian mailing-lists have been full of accusations that they give priority to Ubuntu's packages while neglecting Debian ones.
Some of these accusations might be dismissed as the jealousy of those who have not been hired. Moreover, while free software licenses permit -- even encourage -- code borrowing, you can easily imagine Debian members being upset to see Ubuntu becoming successful while their own efforts go unrecognized. As Kevin Bowling describes the situation, "Ubuntu is stable because they are standing on the shoulders of giants," perhaps without mentioning the fact as often as it might.
Yet the largest reason for complaints from the Debian distribution seems to be a difference in culture. Where Debian is democratic to a fault, Ubuntu is ultimately led by Mark Shuttleworth, who half-jokingly describes himself as dictator for life. Just as seriously, while Debian prides itself on non-commercialism, Ubuntu's goal is commercial success.
Consequently, when Shuttleworth offered to help Debian meet a proposed feature freeze, the offer was seen in some quarters as less a generous offer than an attempt to control and manipulate Debian -- especially when the decision to meet the freeze was not debated and voted upon in Debian's usual fashion.
A similar, more general resentment also exists in the free software community in general. For instance, while acknowledging Ubuntu's many improvements in usability, Blackbelt Jones accuses it "of a lack of empathy for me, the experienced Linux user. In re-engineering Debian, Ubuntu creates detours in the familiar highway of GNU/Linux administration, and then doesn't bother to put up a sign. And so, before I know it, I'm off the road."
Jones cites the simple example of labeling Ubuntu's KDE packages "Kubuntu desktop," which emphasizes Ubuntu's sub-distro while confounding the expectations of experts searching for KDE. But other examples abound, such as adding changes to notifications in GNOME before they are accepted by GNOME and replacing the init daemon with the appropriately named Upstart.
Others complain that the new users who are attracted to Ubuntu know nothing about Linux in general. "They seem to think Ubuntu is responsible for all that is good in the [free software] world," Bowling complains. As often as not, changes that are credited to Ubuntu would be more properly credited to the GNOME desktop or another piece of software, and alleged innovations by Ubuntu can also be found in distributions like Fedora or openSUSE.
However, a far more serious charge is that Ubuntu manipulates the free software ethos for its own advantage, making changes within its distribution rather than in specific projects where others can easily share them.
This charge has been whispered off and on for years. But it found its strongest form in the keynote delivered by kernel developer Greg Kroah-Hartman at the 2008 Linux Plumber's Conference. Throughout his presentation, Kroah-Hartman used Ubuntu and Canonical as a example of an organization that was not contributing to the community as it should.
According to Kroah-Hartman, Ubuntu contributed only 100 of 99,324 kernel patches in the last three years -- substantially below the 11,846 from Red Hat or the 7,222 from Novell. He also found similar statistics in contributions to GCC, X.Org, ALSA, and other core aspects of GNU/Linux. His conclusion? "Canonical does not contribute to Linux plumbing."
Further, Kroah-Hartman stated flatly, "Developers who are not allowed to contribute to Linux should change jobs," in effect urging Canonical developers to quit -- a plea that seems to have been almost entirely ignored.
Like other complaints from the community, Kroah-Hartman's keynote can be seen as partly personal -- the fact that he is employed by Novell could indicate that he is attacking a rival of his employer. But, at the same time, the trend he indicates is too obvious to ignore. The numbers are not even close.
The fact is that Ubuntu is a relative newcomer that, in six years, has pushed itself to the top of the free software pack. Following its ambitions, it has not always considered existing factions in the community as well as it might, either in its practices or its design.
But that is not to say that the ire of the old-timers is completely justified, either. The same groups that complain can also show an inconsistent pride in Ubuntu's successes at making free software more mainstream.
The three ages of success
American fantasist Harlan Ellison once said that the successful go through three periods of public reception. When they first burst on to the scene, the successful are feted and acclaimed. In the second, they are criticized harshly and often unfairly. But, if they persevere, they are accepted as an institution, and the criticism subsides.
Collectively, I suspect, Ubuntu is now in the second stage. The honeymoon, in which the majority saw Ubuntu as free software's best hope, is long over. Instead, Ubuntu is at a point where, in some circles, it can do no right -- and the third stage is nowhere in sight.
On his blog, Bacon argues that "criticism is a sign of success." He has a point, but I think it would be more accurate to say that, in Ubuntu's case, criticism is a sign of partial success.
From the beginning, Ubuntu set out to make desktop Linux succeed. It has done much to reach that goal, improving usability, adding features, and shaking older projects out of their complacency.
The trouble is, having created high expectations, Ubuntu has not finished meeting them. The initial excitement created by Ubuntu's bold goals has died away, and world domination -- the half- serious goal of free software -- is only somewhat closer. Disillusion, or at least resigned realism, has set in. Many are disappointed, and in a mood to magnify faults and present half-truths as completely accurate. And these reactions are exaggerated still further by journalists looking for controversial headlines.
Ubuntu should manage to weather such criticism. However, commercial success and improved testing might not be enough. Considering the passion found in the free software community, Ubuntu might also need to find a middle way between ignoring criticisms completely and accepting them as problems that need immediate action.
Until it can find that middle way, Ubuntu is likely to continue the way it is now, fair game for anyone with a grievance, regardless of whether that grievance is completely justified or not.