Sometimes, being right is no fun. Three years ago, I suggested that the Linux desktop was headed for a future dominated by KDE, and that GNOME would be at a disadvantage. Looking back, I conclude that I was right, if only approximately.
What I did not foresee was that GNOME 3 would not only lag behind KDE for code maturity and innovation, but fail catastrophically with users, resulting in alternative interfaces, ranging from Ubuntu's Unity to Linux Mint's re-creations of GNOME 2 in Cinnamon and Mate.
The collapse is so thorough that GNOME is reportedly now talking about obtaining a twenty percent share of the Linux desktop by 2020, where a few years ago its share was well over forty percent.
I know of no figures for traditional desktop usage in 2012, but LinuxQuestion's 2011 survey showed KDE in front, followed by Xfce. Cinnamon was too new to make the survey at all, and Mate registered only a few percent, but, like Unity, both are almost certain to do better this year.
Some, or even all three of these desktops are likely to do better than the 19% that GNOME 3 managed in 2011. GNOME 3 itself will probably show even further decline. As for actual numbers of users, all traditional desktops are likely to lose ground to mobile devices.
Today, the Cold War of the giants, of KDE vs. GNOME, is over. We are in a new era of diversity (or fragmentation, if you think having more choices is a bad thing). So which, if any desktops are likely to dominate in the next few years?
Which will be the source of major innovations? Which will fail to emerge from the pack? Which are likely to be in the running?
None of these questions are as easy to answer as they were three years ago.
The Linux Also-Rans
Immediately, I would remove Xfce, Cinnamon, and Mate. I regularly use all three, and all of them are likely to increase in popularity, but their development goals are too modest for any of them to dominate the desktop.
Xfce succeeds in its goal of being both "fast and requiring low system resources, while still being visually appealing and user friendly," and a percentage of users will always want this combination. But it has never been a center of innovation, and a massive change of direction would be needed for it to become one.
Similarly, Cinnamon and Mate are dedicated to recreating the GNOME 2 experience. So far, neither has managed more than minor innovation beyond that modest goal, although both are new enough that judging them on their releases to date would probably be rash. They may have been too busy providing the basic GNOME 2 experience.
However, the whole idea of Cinnamon and Mate is to offer what users want -- and that means giving users GNOME 2. In ten years of active development, GNOME 2 was more known for incremental changes rather than radical redesigns, so that suggests we should expect the same from Cinnamon and Mate.
Just possibly, development of one might be discontinued in a few years in favor of the other, but that is probably the largest change to be expected. While GNOME 2 clones are popular now, they are hardly likely to be the long-term future of the desktop. Their popularity may be a temporary reaction to GNOME 3 and could plateau after this year.
Is GNOME Still a Player?
So what about GNOME itself? At the recent GUADEC, GNOME's annual conference, Xan Lopez and Juan Jose Sanchez delivered a presentation entitled "A bright future for GNOME." Contrary to what I expected, the presentation was not ironically titled, but a call to revitalize GNOME that is scheduled to result in the release of GNOME 4 in March 2014.
Lopez and Sanchez identify the problems of GNOME as due to a number of trends, including the shift to mobile devices, the fragmentation of the desktop market, a lack of corporate support, a brain drain, and user disenchantment. In the place of these discouraging developments, they suggested ambitious possibilities for rebuilding that included developing GOME for mobile devices and a GNOME distro.
Strangely, their suggestions did not include greater user involvement in development, which would be a natural solution to GNOME's present state. All the same, coming at GNOME's annual conference, their presentation could be called the first official recognition in the GNOME project that GNOME 3 was a disaster. After eighteen months, such a recognition was long overdue.
However, from its slides, the presentation appears to have been long on strategy and short on tactics -- specifically on interface and software changes. Perhaps the follow-up discussions on the conference corrected this oversight. But, for those with a sense of history, the emphasis sounds uncomfortably close to that of the discussions that led up to GNOME 3. The project is aware that it needs to take some bold steps, but has no definition of just what those might consist.
In addition, the presentation may have an optimistic view of the situation. Shortly before GUADEC, developer Benjamin Otte announced that he was skipping the conference this year, calling it a "self-congratulating echo chamber" and adding "I don’t feel like I would be productive in the current state of things."
Otte followed up on these comments in a longer blog entitled "Staring into the Abyss." In it, he paints an even bleaker picture than Lopez and Sanchez do.
According to Otte, GNOME is losing lead developers, is understaffed, and directionless, and the majority of developers are from Red Hat, which makes it unrepresentative of the general community. The accuracy of his comments are harder for outsiders to assess, but, in 228 comments, only a few people quibbled over details and none, so far as I can see, questioned his general perspective. The closest to an opposing view I can find is the statement that GNOME has faced similar situations before -- which it has, if never such serious ones.