With autumn approaching, many people start wondering whether KDE4 will be released on time and if it will be usable. Although I didn’t follow the development that closely, I heard enough to get an idea about it, that the current “gamma proposal” simply seem to confirm: even if KDE4 will be released, it will not be in an usable state.
This is not to say that the KDE developers aren’t doing a very hard job. They are doing an heck of a job, and I’m sure that on the long term it will produce a fantastic desktop environment. But the current development method is unlikely to easily produce results on short terms.
What I think is still a big problem is the size of the base packages: having to sync all the release cycle for the packages with the rest of KDE makes difficult the incremental evolution of the software. It’s quite easy to see that a lot of the software that brought users to KDE 3 came from extragear or outside, rather than from the base packages: Amarok, K3B, Kile, TorK, Konversation, … You don’t see people saying they love Kopete (we all had our hopes up after the 0.12 series was split out, but then it was static again); SuperKaramba and Akregator also suffered from bitrot and featurerot since they got into the base packages.
The other problem is that a lot of features got added to the list for KDE4; while this is good, they take time to develop, test and polish. This is why people at KDE are thinking about releasing a ‘gamma’, that is neither a beta nor a release.. Well, of course even if the code is released as 4.0 the quality of the code won’t change magically with the name. Whatever you call it, before next year, KDE 4 won’t be ready for prime time. Even KDE 3 was unusable for me till 3.2.
Another problem seems to be the scarse consideration for distributions; well, for non-SUSE, non-KUbuntu distributions, or probably for no-money distributions. Beside CMake being far from trivial for many distributions, there seems to be little idea about usage of shared libraries: Amarok will use an inline modified copy of libplasma. Let’s add to that some debatale dependency, like packages that were altready removed from Gentoo.
Bottomline, I’m sure that KDE4 will be, in the future, a terrific desktop environment, but I’ll probably continue using KDE3 for a long time.
Unfortunately, I really agree with you.I started reusing KDE with KDE 3.4 because it found the previous ones too slow.Furthermore… I don’t want to switch to Kubuntu only for KDE.And event if I’m sure KDE 4.x (x >= 2) can be terrific, I’m pretty happy with my KDE 3.5 on my Gentoo.BTW, why not releasing 3.7 stable?
Actually the purpose of gamma is to have a firm release date for KDE 4.0. Its hard to have a big release party and media event if a bug pushes back the release date a week. It was not a technical decision at all.I do agree with you entirely about the need to shrink the kde* packages. Most distros like Kubuntu just ignore what KDE considers to be its core apps and package a combination of core and extragear – with mostly extragear since thats where quality is. (The kde* packages probably are used more by Gentoo users, since Gentoo doesn’t give much guidance on which software is any good).
Actually thats not true, it was a technical decision too. .0.0 releases are notoriously bad (for any software), part of the idea is to make KDE 4.0.0 like KDE 4.0.1.
Same with microsoft products. ‘never use before sp1’…
A couple things:* Amarok won’t be using inline libplasma…changes have been made to plasma to support use cases of programs outside kde*. The discussion about Amarok using an inline version wasn’t a lack of awareness about shared libraries — it was a lack of ability for libplasma’s API to handle Amarok’s needs, and these days it’s hard to get KDE 4.0 API changes through.* KDE developers are not generally thinking that distros will pick up 4.0. Many parts of it will not be fully functional/ported/written before 4.1. AFAICT 4.0 will really be a technology showcase — here’s the new and important stuff we’ve been working on, it mostly works and does some amazing things…now give us some time to bugfix it and get it all cleaned up. Plus this will give time for distribution maintainers to wrap their heads around it.* Kopete can be quite nice for some things — hell, they can handle v4l webcams, which is something gaim/pidgin has been trying to get for two years. But I do agree that since it went back into kdebase improvements and work that have been made have been less visible, and I think it’s hampered by the KDE release cycle…release early release often often benefits apps like that.