Everyone agrees that desktop Linux has become more diverse in the last couple of years. But how diverse? And how are the dethroned dominant environments responding?
Those are questions that nobody is asking -- although they probably should.
So far, 2012 has yet to see one of those magazine polls that are our main -- although imperfect -- indications of what desktop environments users prefer. However, with a little digging, a few indicators can be unearthed.
Several indicators are available from Distrowatch, the site that attempts to track distributions.
To start with, as I write, Distrowatch's last ten updates (from September 27 to October 8) include two distributions that offer a choice of default desktops, one that offers GNOME or KDE, and one apiece that uses IceWM, LXDE, and Xfce. Four use GNOME, but none use precisely the same version: one uses Cinnamon, Linux Mint's recreation of GNOME 2: one ships Mate, Linux Mint's set of extensions that converts GNOME 3 into GNOME 2, one offers GNOME 3 with a large number of extensions, and the fourth ships GNOME with its own modifications. Neither KDE nor Unity are among the default desktops of the distributions mentioned.
Any other group of updates from the last six months shows an equal diversity of results. To take a random sample from June 20-29, one has no desktop, one offers the choice of GNOME, KDE, or Xfce, one a choice of GNOME or KDE, one of Unity, and one of KDE. Four ship with GNOME, and one with a privately customized version of GNOME.
The page hits tell a similar story about what those interested enough in Linux to know about Distrowatch are interested in. Cinnamon / Mate, Unity, and Xfce are each represented by one of the top ten distros, KDE by two, and GNOME 3 by two. The rest have no single default, or else use lightweight window managers such as IceWM or Enlightenment.
By contrast, the top ten page hits for 2010 include 5 for GNOME and 3 for KDE.
The only user survey for 2012 that I could locate was a study of Czech users completed two months ago. It showed 24% were using KDE, 17% some version of GNOME 2 (probably Mate or Cinnamon), 20% GNOME 3. Unity accounted for 14%, and Xfce for 12%.
The same study, parsed for Czech Fedora users, showed 46% using GNOME 3, and 27% KDE. Some 12% used Xfce, and 8% GNOME 2. In comparison, 30% of Ubuntu users were -- unsurprisingly -- using Unity, 21% GNOME 3, 18% GNOME 2, 15% KDE, and 10% Xfce.
Assuming that any of these indicators reflect all desktop Linux usage to within more than a few percent would be rash. However, taken together, they clearly show the current diversity of interfaces (or fragmentation, if you believe this situation undesirable).
They also show no strong preference. Although KDE and various flavors of GNOME remain strong, none dominate as decisively as GNOME and KDE once did.
As for Xfce, which for years was the third most popular desktop, either it has slipped from its popularity in 2011 LinuxQuestions.org Members Choice Awards, or else those results were atypical. If Xfce has slipped, it is probably due to the maturation of Cinnamon and Mate -- perhaps users are thinking there's no point to using a desktop something like GNOME when a GNOME variant is available.
Another trend these statistics seem to indicate is that the split in GNOME appears to be continuing. A couple of years ago, Unity added a second interface to GNOME. Today, Linux Mint's recreations of GNOME 2 are popular, and GNOME 3 is shipped with modifications from official GNOME extensions, or else from the distributions themselves.
Moreover, if the Czech survey can be scaled, which GNOME variant users prefer apparently depends largely on what default a distribution uses. However, since Fedora's and Ubuntu's default desktops have each been available to the other distribution for only a few months, that may not be a valid conclusion.
Responding to the New Reality
So far as I can tell, few of the current desktop alternatives have officially responded to the new reality. KDE and Xfce continue much as they have always done, concentrating on development with relatively little attention paid to marketing. KDE's marketing concentrates largely on producing KDE.News (also called The Dot), while Xfce's advocacy list has not had a posting for a year, and its foundation list for six months.
Although KDE is attempting to release its own tablet, that project has so far not succeeded. The project can release the KDE Manifesto, a re-statement of free software principles, as though it were ten years ago, and diversity on the desktop was not a current challenge. Having suffered less than GNOME because of user revolts, KDE apparently has less incentive to change its group habits.
Similarly, while Linux Mint's developers talk on the distribution's blog sometimes about giving users what they want, and maintain mail forums with moderate traffic, online they give no other indication of being overly concerned about the popularity of their work.
Unity developers may be concerned about market share. But given that Canonical would market Unity as part of Ubuntu anyway, their attitude is hard to deduce.
By contrast, the project that has reacted most strongly to the diversification of the Linux desktop is GNOME. But, considering that mainstream GNOME has lost more user share than any other alternative, that is hardly surprising.
Besides, GNOME has focused on marketing for some years now. Its marketing list includes regular postings by such GNOME leaders as Allen Day, Dave Neary, Stormy Peters, and Karen Sandler, and its news releases show some awareness of business standards.
For example, the announcement of GNOME 3.6 is careful to include ready-to-use quotes from several prominent free software community members, while the announcement of KDE 4.9 is basically a friendlier version of the release notes.
However, even by GNOME's previous standards, public relations and strategy has become a larger concern in the last six months.
Officially, the project admits to no problems, emphasizing in its plans the continued development of GNOME 3 along the current lines. Unofficially, though, many members seem resolved to restore the project's popularity and influence.
Some of this concern seems naive. After I suggested that the "conventional wisdom is that GNOME 3 has failed," a thread on the GNOME marketing list considered ways "to push back on negative articles like this.
The responses considered complaints to the site that published the article, positive press releases of what was happening in GNOME, a blog campaign, and suggestions of media training to counter future articles that the group considered negative. In the end, more experienced heads prevailed, which was just as well.
More serious efforts to promote GNOME seem to date from "A Bright Future for GNOME," a talk delivered by Xan Lopez and Juan Jose Sanchez at the 2012 GUADEC.
Lopez and Sanchez identified a number of problems with GNOME, starting with a lack of direction and vision and a loss of developers and users, and suggested a number of redirections. Instead of remaining a traditional desktop, they suggested, GNOME should explore offering cloud services and an app store, and expanding to new hardware platforms.
This talk was followed at GUADEC by a “birds of a feather” session for planning. GNOME emerged from the talk and the planning session with the concept of GNOMEOS -- not quite a separate operating system or a distribution, according to Allan Day, but an installation with tools like a Software Developer's Kit and support for new types of devices.
The goals of GNOMEOS are to make development of both software and interfaces more easy, and to make displaying the project's efforts to interested outsiders easier. These efforts continue largely under the radar, although GNOMEOS was one of the topics scheduled for discussion at the recent GNOME Boston Summit.
This year, too, GNOME has extended its outreach into the community. GNOME observers attended the User Observation Hackfest to observe first hand how ordinary users reacted to GNOME. In addition, the Friends of GNOME campaign started to raise funds and emphasize accessibility, while the GBeers campaign recently begun focusing on grassroots meetings.
Whether such efforts will be enough to restore the popularity of mainstream GNOME is anyone's guess at this point. Some might wonder if any of these efforts, no matter how ambitious or well meaning, will be enough if GNOME doesn't learn to listen to users more. But, for now, what matters is that the diversity of desktops is leading at least one major player to efforts that it might not have otherwise considered.
GNOME's reaction to its change of status seems proof that competition can be creatively healthy. So long as GNOME was one of two dominant desktops, the project had little incentive to innovate. It could simply go on for years introducing small changes without altering much of anything. Now, confronting either a loss of influence or a major shift in roles, it has the motivation to be more adventuresome.
Of course, as the user revolts of the last few years of shown, innovation for its own sake is not enough. GNOME may still be dragged down by its insistence on staying the course with GNOME 3, despite all its efforts to expand in new directions. But what matters is that, with the diversification of desktop Linux, it has reason for trying. For that reason alone, diversification seems worth having.
Somewhere in the next few years, diversification on the desktop may slow down. It might even diminish, with two or three distributions becoming dominant, although given the individualism of Linux users,that seems unlikely.
Meanwhile, I only hope that diversification encourages other interface developers to follow GNOME's example. After the conservatism that the user revolts have caused, desktop Linux needs something to encourage innovation again.