| |||||||||||||
KDE 4, distributors, and bleeding-edge softwareKDE 4, distributors, and bleeding-edge softwarePosted Jan 28, 2009 20:29 UTC (Wed) by pynm0001 (guest, #18379)In reply to: KDE 4, distributors, and bleeding-edge software by rahulsundaram Parent article: KDE 4, distributors, and bleeding-edge software Instead of finding out whom to blame, it would be much nicer to solve the communication break down as the article points out. Merely identifying where fault lies is not effective, that is true. But I think keeping these kinds of problems from happening in the future does depend on better work from the downstream distribution vendors. The problem as always is manpower, for development and testing. I just think if a distro couldn't handle such a large change that it should not be integrated, which is a wise lesson to learn no matter what software you're integrating.
(Log in to post comments)
KDE 4, distributors, and bleeding-edge software Posted Jan 28, 2009 21:33 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]
The problem is that, it isn't always predictable how a release would turn out ahead of time. Downstream distributions have to usually incorporate a development snapshot ahead of time and it be nice to have a clear and regular bi-lateral communication on that expectations for that particular release throughout development instead of just at the end where it is difficult to revert back and lose all the QA time and effort.
One of the things that could have been done better is for upstream to make it parallel installable and support that configuration. Some downstream distributions patched things heavily to provide KDE 3 and KDE 4 together and that was repeated duplication of effort without upstream support that could have been avoided.
From Fedora's standpoint:
KDE 4, distributors, and bleeding-edge software Posted Jan 28, 2009 21:41 UTC (Wed) by pynm0001 (guest, #18379) [Link]
I understand. We actually did quite a bit of work on making KDE 3 and KDE
4 able to coexist. For instance most of the executables that could conceivably end up in the same PATH got "4" tacked onto their name (so now kbuildsycoca is kbuildsycoca4 for instance). I only have ever really used the KDE 3 app from a KDE 4 desktop technique where it worked fine, so I apologize if the Fedora guys tried it and were not able to get KDE 3+4 working well enough.
I've noticed it's amazing at the number of issues I've seen over the years
KDE 4, distributors, and bleeding-edge software Posted Jan 29, 2009 0:40 UTC (Thu) by drag (subscriber, #31333) [Link]
> Merely identifying where fault lies is not effective, that is true. But I think keeping these kinds of problems from happening in the future does depend on better work from the downstream distribution vendors.
Sometimes it's nobody's fault.
Everybody operates on limited amount of information and given that set of information then their actions may be 100% correct. Given that they were probably correct with the limited amount of information that is available and bad things end up happening it means that they were wrong, but not wrong in a bad way. Just plain wrong. It's not their fault they were wrong it's just how it is. Could not of happened any other way.
Very very often bad things happen were everybody involved behaved properly and gave good sound decisions. However given their limited abilities to discern reality they were still wrong.
So very often trying to figure out who to blame is not just a waste of time, it's a aggressively bad move and totally counterproductive. Aka "a witch hunt".
People have this odd concept that everything around them is so tightly controlled that if something turns to shit then it _just_has_ to be somebody's fuck-up. Maybe their own, maybe somebody else's. Then there is a notion that if something is not under control and it spins out then it should be put under control. Both notions may or may not be correct given the circumstances.. but most often are just wrong.
---------------------
If it's a problem that is re-occuring then you can take steps to prevent it.
If it's not a problem that is reoccuring and is just random chance.. then quite often it's not really worth your time dwelling on it at all. Just pick up the pieces and move on.
In the case of the KDE 4.0 end-user-relations disaster, it may or may not be a misrepresentation or self-delusion on the parts of the KDE crew, or it may be a simple miscommunication and disconnect between the devs people and their extended user base... but either way it's pretty irrelevant and the action to correct the mistake is the same.
The corrective action is, from what I can tell, "Learn the lesson and otherwise ignore it, and move on. Lets try not to do that again!".
Seeing how there is no plans for KDE 5.0 we can be pretty safe to say that it's not likely to repeat itself any time soon.
-------------------------------
In the defense of 'release soon and release often' lets remember the fate of the 'Enlightenment' project.
As you all probably remember Enlightenment used to be a very high quality Window Manager from around the days of Gnome 1.x. It was nice, high quality, had a number of visual and usability enhancements that made it quite unique.
They took the approach of slow and careful development while doing a rewrite and at the same time shot for the moon. They essentially stagnated for years and years and now exist with a much much reduced user base and dramatically less attention from folks.
|
|||||||||||||