Yet another sign of the state of GNOME is that Debian, which has defaulted to GNOME almost as long as there was a GNOME, will now use Xfce instead.
At best, GNOME has a hard struggle ahead of it. Its position in the future will depend much on its immediate plans. But, right now, whether mainstream GNOME can recover anything like its former dominance seems unlikely.
A possible contender might be Ubuntu's Unity. Unity is slightly more tolerant of different work flows than GNOME 3, and no one can say in its interface experiments are not attempts at innovation.
The trouble is, exactly how popular Unity might be is nearly impossible to evaluate. Ubuntu's commercial patron Canonical likes to talk about the number of computers that ship with Unity, but those figures are received by most observers with considerable skepticism. Moreover, as with Linux itself, no one knows how many machines it remains upon once the sale is made.
Considering that Ubuntu repositories allow another desktop to be installed in fifteen minutes or less, and that users' attacks on Unity are almost as numerous and impassioned as those on GNOME 3, the desktop's popularity may actually be far less than Canonical insists.
As for being an innovator, Unity is the vision of Ubuntu and Canonical founder Mark Shuttleworth and a small team of designers. Whether its innovations are anything that the majority of users want is questionable, to judge by their reception. You might even say that, by allowing the needs of mobile devices to influence Unity's design, Shuttleworth and his designers have regressed the desktop rather than advanced it.
If Canonical's statements are accurate, then Unity might manage to emerge as the major desktop, if only for commercial reasons. However, I wouldn't be surprised if independent polls suggest a different story than Canonical's.
From the current state of affairs, it seems probable that KDE will continue to be the leading desktop for the next few years. Not only does KDE have a mature code base, but its options accommodate everyone from those who want a basic desktop to those who want to experiment with a variety of desktop. It even has a two-year head start on interfaces for mobile devices.
True, KDE's last major release was received with almost as much hostility as GNOME 3 or Unity. But, KDE appeared to listen to complaints, and subsequent releases satisfied most of them. Although comparing various user polls suggests that KDE may have lost 7-9% of its users after releasing KDE 4.0, that is less than half what GNOME lost, or maybe as little as a third.
KDE does have some challenges. In particular, its attempt to produce a tablet has been delayed by the sorts of problems that can be expected in bringing a new product to market, especially by those with no experience in the project.
However, while GNOME seems in a communal decline, in early 2011, the latest period for which I could find figures, KDE was steadily adding developers.
KDE's largest problem is that, with the just-released 4.9 release, it seems to have settled into incremental releases, with less direction than it had a year or two ago. However, in the aftermath of GNOME 3 and Unity, a cautious approach to changes may be exactly what the average user wants.
If I am right, KDE is going to continue to be the desktop with the strongest combination of user popularity and innovation. This isn't an unusual position for it; if you look through the user polls, KDE repeatedly switched first and second place with GNOME over the last ten years.
However, if KDE does remain dominant, it will do so under terms very different from those of the past. Instead of competing mainly with GNOME, with Xfce a distant third, KDE is likely to have a far less commanding lead, with several desktops not far behind it.
Even more importantly, the criteria for judging desktops will be different than in the past. A decade ago, KDE and GNOME releases were judged by the new features they introduced. For the next few years I suspect that user choice will be the main factor in choosing desktops.
Small changes and redesigns may still be welcome, but, more than anything else, users are likely to value familiarity in a desktop, as well as the ability to work the way they choose. Possibly, too, they will appreciate a desktop that runs as well on their phone or tablet as on their laptop.
As a desktop that juggles innovation with customization, KDE is in the best position to be the main player in such an era. But who will the other major players be? We'll have to wait and see -- an answer that says even more about the future state of the Linux desktop than which one is the leader.