What concerns me is that, in developing the open source desktop, developers might adopt certain ideas or ideological concepts to the detriment of end-users. All that matters to many open source developers is the sheer beauty of the idea. Everything else comes second.

None of this means that Gnome Shell (or any other innovative open source technology, on or off the desktop) will be poor quality, or outrageously useless. But I really hope they test technologies like Gnome Shell against members of the public, and maybe run a few usability labs. Developers need to remember that they are far from typical users. Above all, I hope the Gnome guys (and all open source developers) are brave enough to realize when an idea has to be abandoned, regardless of how intrinsically clever it is, or how much it deserves to succeed. (Again, please note that this isn't a veiled criticism of Gnome Shell, which I think looks great and is extremely promising.)

As far as Ubuntu goes, however, I think I have a solution for the immediate problem of where to go next. In my opinion, the KDE4 project is maturing very nicely. The recent release of Kubuntu 9.04 shows off KDE 4.2 very well, and the 9.10 release in October will apparently feature the even better 4.3.

So why not move the main Ubuntu release to KDE4? Effectively, the Kubuntu project has been doing advanced research for such a move almost since the beginning of the entire Ubuntu project. They've developed the system software, for example, such as a GUI package manager.