| |||||||||||||
The critics are wrongThe critics are wrongPosted Jul 3, 2008 0:20 UTC (Thu) by Sutoka (guest, #43890)In reply to: The critics are wrong by qg6te2 Parent article: The critics are wrong: KDE 4 doesn't need a fork (ars technica)
Thats not revising history, thats what the KDE developers were saying themselves before the release. "4.0" is just a number after all, and x.0 releases are generally very buggy and having many missing features (hardly 'mostly tested'). As has also been said a lot, the .0 release was more about the libraries than what sat above them. KDElibs was ready for the initial release and for application developers to start working with it, and by releasing 4.0 it became much easier to get your hands on a copy of kdelibs (without spending quite a while manually compiling it from SVN every couple days), as well as it being easier for users of applications that used the kdelibs 4 (it certainly is FAR easier for me to use the normal packages for my distro and just develop my browser against that rather than having to spend *hours* compiling everything needed so I can just get annoyed and by missing dependencies and give up before I even have 'hello world' working). But really, if thats your biggest complaint about KDE 4 (notice the no '.0' for the specific release), then you must not have that much to complain about. People also seem to have forgotten about the 'release early, release often' mantra that many OSS projects follow. Also don't forget they're not charging anything for 4.0, and the day 4.0 was released they didn't kill the 3.5 branch.
(Log in to post comments)
The critics are wrong Posted Jul 3, 2008 0:43 UTC (Thu) by qg6te2 (guest, #52587) [Link] x.0 releases are generally very buggy and having many missing features I'm sorry, but I do not buy the above reasoning. Only unfinished and untested software is buggy and has many missing features. If we all ascribe x.0 software as being a waste of time, nobody would use it (by that reasoning, a fine piece of work, like RHEL 5.0, would be labelled as "junk"). Should we all skip x.0 and go directly to x.1 ? In time, all x.1 software would be considered as not worthy, so should we release everything new as x.2 ? This slippery slope has to stop somewhere. Let's call a spade a spade, and call an alpha release an alpha release. If the new KDE had to be called 4.0, at the very least it should have been tested internally much more -- on the day the 4.0 release, significant stability bugs made themselves known within 2 minutes of usage.
The critics are wrong Posted Jul 3, 2008 0:59 UTC (Thu) by Sutoka (guest, #43890) [Link] >Only unfinished and untested software is buggy and has many missing features. Any change as big as 3.5->4.0 is bound to have lots of missing features and bugs. How about a similar large change... KDE 2.0 was quite similar, and people were saying the same thing about it! >If we all ascribe x.0 software as being a waste of time, nobody would use it A lot of people don't. x.0 releases DO have a reputation of being of lower quality than normal because: >at the very least it should have been tested internally much more It's very hard to do lots of internal testing with a small team (you can often over look a bug staring you right in the face if it's been there for a while but been low priority due to fixing other things... or just your usage patterns, or any number of things). And again, 4.0 was supposed to be more about the libraries which were rather good at that point (there were already a good amount of users of them).
The critics are wrong Posted Jul 3, 2008 4:09 UTC (Thu) by pynm0001 (guest, #18379) [Link] > > x.0 releases are generally very buggy and having many missing features > I'm sorry, but I do not buy the above reasoning. Only unfinished and > untested software is buggy and has many missing features. If we all > ascribe x.0 software as being a waste of time, nobody would use it (by > that reasoning, a fine piece of work, like RHEL 5.0, would be labelled as > "junk"). Should we all skip x.0 and go directly to x.1 ? In time, all x.1 > software would be considered as not worthy, so should we release > everything new as x.2 ? This slippery slope has to stop somewhere. Let's > call a spade a spade, and call an alpha release an alpha release. If you had been around kde-devel or kde-core-devel before the 4.0 release you would know that we wrestled with this exact same issue ourselves. The problem is *exactly* what you said; no one tests alpha releases and beta releases and RC releases. KDE is not the only project to deal with this: http://lkml.org/lkml/2005/3/2/383 is a glimpse into the reasoning behind the change in Linux 2.6 versioning that occurred a couple of years ago. "I want to have people test things out, but it doesn't matter how many -rc kernels I'd do, it just won't happen. It's not a 'real release' " 4.0.0 was actually delayed (http://www.phoronix.com/scan.php?page=news_item&px=Nj...) to get more bugfixing, we didn't just shit out what we had and call it a day. And how in the crap did you end up using it then anyways? :) It was not included by default with any distribution (OpenSUSE 11 is the first to make KDE 4 default IIRC, and that with a much more polished KDE 4 than the 4.0.0 release). So presumably you built it yourself to test it, or you used a LiveCD, or somehow else went out of your way to test it out and gave us no feedback? :( Either way others have been doing a better job at responding point by point so I won't duplicate the effort. But someday real users have to use the software for it to get the testing it needs (unless you'd like to help with KDE Q/A, we could use volunteers. :)
|
|||||||||||||