To view parent comment, click here.
To read all comments associated with this story, please click here.
The KDE team didn't do that, the various distributions did. The KDE team said they wanted people to try KDE 4.0, in order to get feedback. They did not say they wanted people to use KDE 4.0 as their primary desktop ... it wasn't ready for that.
They deliberately released admittedly beta quality code as a point zero release and then tried to justify it by attempting to redefine the standard expectations of what a point zero release is for the sake of getting wider testing. I'll agree that the distros were partially to blame for pushing it as a replacement for 3.x so early, but they aren't solely to blame by a long shot.
In late 2007, the kde folks were facing a number of chicken and egg problems that were only going to get resolved by making a release.
Internally, there was tension between the parts of KDE4 that were mature and stable (kdegames, the libs, kdebase outside of kwin & plasma) that were itching to get their work out while kwin's compositing and plasma were quite late in becoming usable and there were still other devs going on and on with blue sky work.
Kwin & plasma were needing big improvements & bug fixes in drivers and the rest of the X graphics stack, and those issues weren't going to get resolved until there were at least some users putting pressure on upstream.
So a release was scheduled, and timeline set. The schedule was pushed back twice, and plasma was still in a really, really, rough state, but once it was barely functional (launch apps, show a taskbar & half broken system tray, draw a clock, etc), they released.
Then, in the midst of all this, as a result of poor communication, several distros planning their spring releases felt the need to pick up 4.0 and drop maintenance 3.5, and by the time it was clear that 4.0 was in poor shape, it was too late to change back.
Then prior to the 4.1 release, to take advantage of some important new features in Qt, plasma was heavily reworked in a way that broke most of the widgets, and by the time 4.1 was released in July, it wasn't in much better shape.
Was it a mess? Yes. Was it the end of the world? No. Did it burn some users? Yes. Did it attract a whole bunch of new developers hacking on the code. Yep - and that is partly why KDE is in a much more mature state now.
Edited 2011-04-06 23:22 UTC
I just want to correct one point: KWin was not in an unstable change in 4.0 or needed a release badly. It is true that OpenGL compositing support was new, but disabled by default. KWin did not start to enable compositing by default before 4.2.
So the KWin the user of 4.0 had was the same rock stable, feature rich window manager as used from 3.5.





Member since:
2005-07-06
Which is why the KDE team should've known better than to push a beta quality release into the mainstream before the associated apps had a chance to properly catch up.
The Gnome team went for continuity and as smooth a transition as possible incrementally updating during the entire lifespan of the 2.x series and pruning out old code before finally switching a good portion of the 2.x releases over to the new libraries. You end up in a position where it's entirely feasible to support Gnome Classic in parallel with Gnome-shell for a long time to come, should the market prove to prefer it that way.