KDE 4
There was a funny exchange on the Gamers with Jobs podcast a few months back, where the guys were grilling BioShock designer Ken Levine about some of the terrible games that he's listed in the credits for on his MobyGames page:
Ken: Master of Orion III is on my MobyGames page
Certis: That's right
Ken: Now I need to make a public statement about my MobyGames credit for Master of Orion III
Certis: Please do
Ken: I had a friend who produced it, and my involvement with Master of Orion III was I told my friend don't ship it
That sums up my thoughs on KDE 4.
Basically, KDE 4 isn't finished yet. I've been playing with the quite up-to-date KDE 4 pre-release packages for openSUSE, should be very close to the release version we're being told to expect at the end of the week, but there are bugs and feature regressions everywhere. A lot of this is down to Plasma, the widget engine that's meant to redefine how we interact with the desktop or some such nonsense, which is very clearly incomplete, and currently limited to pulling off bad Kicker and SuperKaramba impersionations.
That's right: KDE's panel, which has long been annoyingly inflexible compared to the GNOME panel, despite rants about how unconfigurable GNOME is in general, is even less flexible now. The new Kickoff main menu, which is based on the modified main menu used in openSUSE's KDE 3.5, is also really annoying, making it much harder to browse through your available applications and essentially forcing you to use the (admittedly quite handy) search box they've stuck in there to find anything.
Of course, this release isn't about the desktop at all -- it's about the platform, and I understand that they really need to get developers working on these new libraries, and getting users to test against them too. In a lot of ways it's like Apache 2.0, which was also barely ready for public consumption, and which was also torn up internally in the name of portability. The new threading models did great things for Apache's performance on NT, but they did nothing for Linux, where most Apache users are running. Likewise, it's great that KDE's new platform-independent underpinnings will help extend it to Windows and OS X, but for the 99% of KDE users running Linux/UNIX, there's a lot less to gain.
Like Apache 2.0, I'm pretty sure that KDE 4 will eventually be something genuinely usable, and perhaps even impressive, but that's 12 months away at least. In the meantime, it's going to put a lot of distro maintainers in a very awkward position.
