KDE 4, distributors, and bleeding-edge software

Posted Jan 28, 2009 19:46 UTC (Wed) by fandom (subscriber, #4028)
Parent article: KDE 4, distributors, and bleeding-edge software

I tried 4.0 in Mandriva for about 15 minutes, before I decided to use 3.5
which was also included in the distro. After all, I thought, they did say
to wait for 4.1 if you wanted to actually use the system.

And indeed, 4.0 was unusable as desktop system, I don't see why they
needed lots of testers to reach that conclusion, they wanted the same bugs
reported a few thousand times before fixing them?

4.1 is much better, it is not a feature rich as 3.5 but it works,
unfortunately they got a weird bug that requires them to make programs
unusable in the name of usability, so I am still with 3.5

Haven't tried 4.2 yet, let's hope they got some sense back.

KDE 4, distributors, and bleeding-edge software

Posted Jan 28, 2009 20:09 UTC (Wed) by roblucid (guest, #48964) [Link]

Having tried a pre-release version, I'd agree, it was very clear it was not going to be ready for a long time. There was a lot of info about nifty features and eye candy around, yet the difficulties of daily usage seemed to be ignored in reviews. Those reviews stoked the desire, but the inability of the DE to "put out" lead to much pent up frustration, which got vented out in blogs and forums.

Aaron did comment at the time, and it seems they really needed to get back to release discipline, to focus minds, and avoid prolonged blue sky development. It's worth using the Google feature "KDE4.0 site:lwn.net" to see some of the comments from January 2007.

At the end of the day, the version numbering system simply didn't give them a way to release early and often, KDE 4. openSUSE included changes to KDE 3.5 to make occasional running KDE 4 and then returning simpler. That should have been in the plan. Make 3.5 & 4 coexist, shipping broken 4.0.x, and then bump the version to signify "stable" at the 4.2 or 4.3 stage.

KDE 4, distributors, and bleeding-edge software

Posted Jan 28, 2009 20:52 UTC (Wed) by dkite (guest, #4577) [Link] (8 responses)

>I don't see why they needed lots of testers to reach that conclusion

It wasn't end users that was required. It was the necessity of putting
together a release to coalesce the development efforts of the project.
Users and testers in the sense of application developers. The libraries and
services were in a state of flux, many developers were waiting for things
to settle down before doing the changes in their applications. A release
was necessary to get all that happening, to test the interfaces, get it all
working together.

I'm not certain how else that could have been done. People have limited
resources, and following constantly changing api's isn't the best use of
them. Some way of nailing things down was required, but not something like
a long feature freeze that would either be ignored or halt the necessary
development.

I think the real problem was people and distros believed what they were
told. It was like ads for medicine that one sees in magazines. Wonderful
pictures followed by two pages of fine print telling you it can kill you.

My suggestion for a marketing campaign was "Software Sucks. Ours Sucks
Less".

Derek

KDE 4, distributors, and bleeding-edge software

Posted Jan 29, 2009 13:02 UTC (Thu) by mrshiny (guest, #4266) [Link] (7 responses)

If you need to release something half-finished in order to get the second half started, then more attention needs to be paid to the transition period. If KDE4 had been compatible with KDE3 in parallel then there wouldn't be a major problem; users could have had stuff replaced on an ongoing basis as the new stuff arrived and matured. Instead Fedora shipped two releases with a very bad KDE because they didn't have much choice about it and they were told the next release will fix everything.

KDE 4, distributors, and bleeding-edge software

Posted Jan 29, 2009 14:20 UTC (Thu) by halla (subscriber, #14185) [Link] (6 responses)

KDE3 and KDE4 _can_ be installed in parallel, in fact, there's been a lot
of effort expended on making it possible. And, if you look at OpenSUSE you
can see it in action. If Fedora, for whatever reason, refuses to make it
possible, it's their problem, not the KDE developers' problem.

KDE 4, distributors, and bleeding-edge software

Posted Jan 29, 2009 17:17 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]

openSUSE and some other distributions have patched KDE 4 heavily to make this possible and it is not supported by upstream developers. Fedora tries to not deviate much from upstream much, especially KDE since folks complained loudly when it happened.

As I pointed out earlier,

http://fedoraproject.org/wiki/SIGs/KDE/KDE4FAQ

KDE 4, distributors, and bleeding-edge software

Posted Feb 1, 2009 15:49 UTC (Sun) by mrshiny (guest, #4266) [Link] (4 responses)

I guess I don't just mean "in parallel" as in "you can install both", but rather a hybrid KDE made up of the old pieces of KDE3 that are working and the new pieces of KDE4 that replace the old. This way, as new pieces become usable and stable they can supplant their older versions.

However, for a variety of reasons KDE major releases (2, 3, 4) are not binary compatible with older releases and thus the KDE team makes no attempt to do this sort of staged release. However for a majorly disruptive change such as KDE4, a staged release is the best way to achieve your goals whithout alienating your users. It is significantly harder, however, from a programming standpoint. If KDE 4.2 doesn't resolve the problems KDE4 has I suspect KDE will lose a lot of market share.

KDE 4, distributors, and bleeding-edge software

Posted Feb 1, 2009 18:32 UTC (Sun) by halla (subscriber, #14185) [Link] (3 responses)

"If KDE 4.2 doesn't resolve the problems KDE4 has I suspect KDE will lose a
lot of market share."

You know? KDE4.2 is a damn fine desktop environment offering functionality
that isn't available _anywhere_ else, in an attractive, stable and very
usable package. I merely tell you this because the sentence quoted above
seems to indicate you haven't tried it for yourself yet; otherwise you'd
have known already, of course, and wouldn't have felt the need to play
cassandra.

In any case, in my KDE4.2 environment I can run all the KDE3 applications
that I need, and I never had a problem running the KDE4 application under
KDE3 either. Nor running a pure KDE4 or a pure KDE3 environment on the same
laptop.

KDE 4, distributors, and bleeding-edge software

Posted Feb 2, 2009 5:08 UTC (Mon) by mrshiny (guest, #4266) [Link] (2 responses)

KDE4.2 is a damn fine desktop environment offering functionality that isn't available _anywhere_ else, in an attractive, stable and very usable package. I merely tell you this because the sentence quoted above seems to indicate you haven't tried it for yourself yet; otherwise you'd have known already, of course, and wouldn't have felt the need to play cassandra.

You'll have to excuse me. My KDE 4.1 environment is so close to being totally unusuable that I didn't feel the need to rush out and download 4.2 on release day just to see if they finally fixed the problems they shouldn't have introduced in the first place. Call me cynical, or too pragmatic, or whatever, but I feel let down by the 4.0 and 4.1 releases: even after all this time the 4.1 release is still quite unpolished and is, in many ways, a severe regression from 3.x. Don't get me wrong: there are lots of good ideas and intentions in 4.x. But users expect more and in fact need more than what 4.1 delivered. You say 4.2 is awesome. I hope it is; I know the KDE team can deliver awesome. But after being burned I will wait until the updates come from my distro, because I already waste too much time administering my computer instead of getting stuff done with it. And if 4.2 doesn't deliver, I for one will switch, because I won't be able to take it anymore.

KDE 4, distributors, and bleeding-edge software

Posted Feb 2, 2009 7:41 UTC (Mon) by mgb (guest, #3226) [Link] (1 responses)

You fooled me once (KDE 4.0). And shame on me, you fooled me twice (KDE 4.1). Don't expect me to try KDE 4.2.

I too wasted inordinate amounts of time on KDE 4.x. Our laptops and desktops now run Debian Lenny which comes with KDE 3.5.10 - undoubtedly the best KDE release to date and supported by Debian for some years hence.

Kudos for all the experimental work that the KDE team is doing in 4.x. If they keep the good and throw out the bad then there is reason to hope that the KDE 4.x series will eventually surpass KDE 3.5.10.

But as a useful working environment, rather than as an interesting programming exercise, KDE 4.x still has a lot of catching up to do.

KDE 4, distributors, and bleeding-edge software

Posted Feb 6, 2009 2:03 UTC (Fri) by pyellman (guest, #4997) [Link]

Huh. I don't consider myself to be tightly in any loop, but I got the message pretty loud and clear. I kept myself reasonably abreast of opinions and developments by reading occasional early reviews and feedback. I also tested the final beta, 4.0, and 4.1 using live CDs. All in all, not that much more investment of my time than I would spend tracking and testing a favorite piece of software like Amarok. I've tentatively decided that I might be ready to make the jump to KDE 4.2 -- of course, after I find the time to pop a live CD into a spare computer and take it for a test drive. As such, I don't expected to be surprised, disappointed, or fooled.

If the distribution I mainly use (Debian Testing, and still on KDE 3.5.x) had parachuted 4.0 on me without the option of staying with 3.5 for a while, yes, I would have been upset -- but not at the KDE team.

Peter Yellman