Is KDE dying? This question, or variants of it, have been asked with increasing frequency in the two weeks since Jonathan Riddell announced that, after the next release, Canonical would no longer pay him for his work on Kubuntu, the KDE version of Ubuntu.
But is the question valid? Or simply unsupported panic?
The issue was raised as soon as Riddell’s news reached Slashdot. One thread in the comments about the announcement was entitled, “Beginning of the end for KDE?” The first post in the thread suggested that KDE will “continue to be developed for years to come but without major backing it’ll probably fade away like a lot of projects do. It’s a shame, I feel KDE had much more to offer than Gnome but long term there could be only one winner and all the major players picked Gnome.”
More recently, the submitter of a Slashdot link about the design plans for GNOME 3’s core applications commented that “for now, I’m sticking to the sinking ship of KDE in the Ubuntu ocean.”
The issue was raised in even greater detail by Jack Wallen, a columnist at TechRepublic. The news, Wallen wrote, “hits the KDE desktop where it counts. . . . This marks the loss of the last major distribution to ship with a KDE desktop.”
Wallen went on to say that he thought KDE would survive, if only as an alternative. However, his concluding comments that “KDE does not deserve such a fate” and “deserves to be made available through some official channel” continues to leave the impression that the news leaves KDE on life support.
Apparently his readers agree. In a poll accompanying Wallen’s article, 12% thought the news meant KDE’s “death knell,” while 52% thought that KDE should release its own distribution.
Clearly, the idea that KDE is in trouble is spreading. The only trouble is, the idea is based on a mis-interpretation of the news, and is unsupported by the evidence.
Leaping to Conclusions
Anyone who believes that KDE is dying would do well to look at history. A Google search for “is KDE dead” returns 38,100 results, most of them in the last five years, but a few over a decade old. By contrast, “is GNOME dead” returns only 6 results. Apparently, the KDE death-watch has a long tradition, even though there has never been much to see.
As for the recent news about Kubuntu, whether it affects KDE is doubtful. Although over the years Canonical and Ubuntu founder Mark Shuttleworth has expressed the occasional enthusiasm for KDE, it has never received the same level of attention as GNOME in Ubuntu.
According to KDE’s Aaron Seigo, Riddell was never paid full-time to work on Kubuntu. In fact, “Jonathan Riddell was working on [the version control system Bazaar] during his work hours for most of the last Kubuntu release cycle and had little time outside of work for Kubuntu. Yet, Kubuntu still made a good release on-time, driven by the community around Kubuntu. So we know it doesn’t die if Jonathan isn’t paid to work on it.”
In fact, a look at the Kubuntu mail forum indicates that an active community supports Kubuntu, and seems determined to keep it alive, despite the discouraging news.
In addition, like Lubuntu, Xubuntu, and other Ubuntu variations, Kubuntu will continue to have access to Launchpad, Ubuntu’s development site, and other Ubuntu resources.
Yet even if Kubuntu were to stop development, the effect on KDE would be relatively small. The truth is, Kubuntu has never been the premier KDE distribution (that would probably be openSUSE).
Mention Kubuntu, and inevitably its past reputation for having quality assurance problems is raised. True, in the last few releases, its quality has improved immensely. Yet, as Riddell admits in the announcement, in seven years Kubuntu “has not taken over the world commercially and shows no immediate signs of doing so.”
In fact, contrary to Wallen’s characterization, Kubuntu has never ranked as a major distribution. The page views on Distrowatch tell the story: in the last four years, while Ubuntu has consistently ranked first or second among distributions, Kubuntu has never been higher than fifteenth, and is currently at twenty-fifth, behind Lubuntu and just ahead of Xubuntu. In these same years, Kubuntu has consistently had less than fifth of the page views as Ubuntu.
Under these circumstances, Seigo is not just putting a spin on bad news when he says, “We have not in the past relied exclusively or even primarily on Canonical’s involvement so, while we regret they aren’t participating as much anymore, the impact is not expected to be significant.”
KDE in the Surveys
An even stronger indication of KDE’s continued popularity is how consistent the preference for it remains in various surveys over the years. These figures show no reason to think KDE is in serious trouble — although possibly it has suffered a small decline in popularity due to the fragmentation of the desktop in the past year. Even the great transition to the KDE 4 series in 2008, which sparked a user revolt, seems to have only temporarily affected the figures.
In 2007, on DesktopLinux.com’s survey, KDE was favored by 35% of the 38,500 who responded to the questions, compared to 45% who favored GNOME.
The Linux Journal’s Readers’ Choice Awards did not give figures before 2008, but, in 2003-2005, KDE was chosen over GNOME each year. The Editor’s Choices for 2006 also chose KDE. In 2008, GNOME won with 45% to KDE’s 42.5%, and in 2009, with 53% to KDE’s 30%. The two environments tied in 2010, and in 2011, GNOME won again, although figures were not given.
Similarly, LinuxQuestion’s Members Choice Awards in 2007 gave GNOME 42% and KDE 40%. In 2008, the results reversed themselves, with KDE at 43% and GNOME at 40%.
Another reversal happened in 2009, with GNOME at 43% and KDE at 40%, while in 2010, GNOME won again, although no figures were given.
However, the most interesting figures are LinuxQuestion’s for 2011. These are not only the most recent figures, but the first since GNOME 3 and Ubuntu’s Unity were released. In these results, KDE finished first at 33%, while GNOME 3 (GNOME Shell) was third at 19%, behind Xfce with 28%.
Other major desktop environments? Unity was far back with less than 5%. Linux Mint’s evocations of GNOME 2, Mate and Cinnamon, had several percentage points apiece — an achievement that is all the more remarkable because they were available for only a few months of 2011. Trinity Desktop Environment, the KDE 3 fork, had 15%, while LXDE had 7%.
What these figures show is that, for years, KDE’s and GNOME’s popularity have been close. The most that might be said is that GNOME may have had a small advantage overall in the last five years. Yet even that is uncertain. A margin of error is never given in any of these surveys, but would probably have been several percent.
However, the situation may have changed in the last year. If you give credence to LinuxQuestions’ results, in 2011 the choice of desktop environments became considerably more complicated then in previous surveys. After years of consistently finishing third, Xfce moved into second place. Alternatives that in earlier years barely registered in these surveys have suddenly become more popular — most likely because of dissatisfaction with GNOME 3 and Unity, and, to a lesser extent, with KDE.
Certainly, the popularity of these other alternatives appears to have come at the expense of both GNOME and KDE. But where KDE dropped by as much as 7-9%, depending on how you calculate its previous popularity, GNOME dropped 21-26%, depending on whether you count GNOME 3 and Unity together or separately. Possibly, though, GNOME’s decline is temporary, like the drop in KDE’s popularity in the Linux Journal survey for 2009, the first year after KDE 4.0 was released.
Of course, these surveys have limited validity. A particular problem is that those who reply are self-chosen. In general, survey participants are probably more knowledgeable than many users of free desktop environments.
Still, the figures’ general consistency suggests that they do have a rough validity, especially if you don’t get down to decimal points. They suggest that either KDE remains relatively healthy or else, if it is in trouble, then the trend is too early to reliably detect.
Moreover, any decline may not be significant. KDE has reached lower levels of popularity in the past and recovered. Moreover, figures from KDE system administrators indicate that, at the same time that KDE’s user figures were dropping in 2011, the monthly number of new KDE developers had reached pre-KDE 4 levels. If KDE really is less popular than before, apparently the project is in small danger of becoming less active.
Reasons for the Leap
So why have people been so quick to extrapolate from the change in Kubuntu’s status to KDE’s decline?
Perhaps people are believing what they want to believe. Yet I think the answer needs to go deeper than that.
To start with, Kubuntu’s name recognition apparently deceives onlookers into thinking that it is more important to KDE that it actually is.
For another thing, KDE’s popularity seems greater in Europe than North America. KDE e.V, the project’s governing body, is based in Europe, and, increasingly, it seems, so are many of its larger projects. North Americans may underestimate KDE’s popularity simply because, for them, GNOME has a larger local presence.
But the main reason is probably that, ever since the creation of the GNOME Foundation in 2000, GNOME has pursued ties with corporations much more aggressively than KDE has until very recently. The result of GNOME policy’s is that, as Wallen points out, the major distributions install with GNOME as the default desktop environment.
The trouble is, the default desktop is a poor metric for success. All distributions ship with a choice of desktop via their repositories, and, increasingly, installation disks are available for both GNOME and KDE for distributions such as Linux Mint and openSUSE.
Even on a distribution like Fedora, which defaults to GNOME, the KDE spin is consistently the most popular variation, with over twice as many downloads as the next-popular Xfce spin. In many distributions, the main difference in status is that the default GNOME disk images are available a few days to a few weeks before the KDE images, and that might be due to a shortage of developers as much as the demand for each desktop environment.
But, whatever the reasons, all indications are that KDE, with or without Kubuntu, is much the same as it has ever been. If anything, GNOME might actually be in far worse shape, although a single year’s decline can hardly be called a trend.
By all indications, KDE is within its normal range of popularity over the last five years. Any decline is slight, and, so far, has not affected KDE’s ranking among desktop environments. Next year might be a different story, but, until then, any funeral arrangements or mourning for KDE seem premature.