This post is meant to be read in the context of this post.
I do not find fault with Canonical/Ubuntu does to invest in KDE/Kubuntu, but rather with what they (don’t) do in neglect of KDE.
So, Kubuntu and Ubuntu are treated equally, right? Unfortunately, the neglect is routine. The most recent case of which happened today. The recent upload of bluez 4.x breaks bluetooth in Kubuntu entirely. You can read about it for yourself here (Scroll down to near 15:00), though I will highlight especially important parts here.
15:01 pitti ScottK: does superm1 know about the KDE regression?
15:01 ScottK pitti: He does.
15:01 pitti he worked a lot with the packages recently
15:01 ScottK My impression was he pretty well shrugged.
Edit3: superm1 has been very nice and fixed kdebluetooth, and is looking at hal shortly. This is great, the below comments however, are really the problem. Bugs do usually eventually get fixed, but it’s the attitude that coutns.
15:02 pitti hmm
15:02 seb128 I don’t think anything went especially wrong there
So an untested upgrade breaks supposedly-equal Kubuntu, and nothing is wrong? Hmm…
15:02 ScottK The Kubuntu dev who’s mostly looked into bluetooth (Tonio) is mostly MIA at present.
15:02 seb128 it’s just a lack of manpower issue
15:02 persia It’s more about needing upstream to do something. None of the few people hacking the bluetooth stack now understand kdebluetooth
15:02 pitti http://launchpadlibrarian.net/18310466/Traceback.txt looks like an API change
15:02 MootBot LINK received: http://launchpadlibrarian.net/18310466/Traceback.txt looks like an API change
15:02 seb128 we can’t block ubuntu in a broken state because nobody tests other desktop variants
15:02 ScottK seb128: There’s nothing especially wrong with breaking Kubuntu just before release?
15:03 Riddell persia: it would be better not to do major foundation changes after beta
seb128 having ubuntu broken just in case a derivative would break is not a solution eithjer
I thought that Ubuntu and Kubuntu were equal… The main problem here is that API changes were made without testing with KDE and uploaded anyway. The resounding response we get here is that Kubuntu is merely a derivative, and uploading untested API changes is OK for that reason.
Continuing on, we arrive at bug 259436. Basically, wiki login was broken for every browser except Firefox, another lack of testing/act of neglect. Yes, it did get fixed today (thankfully), but for almost 2 months this bug was a severe impediment to Kubuntu development. Main Inclusion Reports couldn’t be written. Agenda points for meetings couldn’t be set. Release announcement wiki pages were writable only by those who used Firefox. (Such as myself) This bug has almost affected one third of the entire development cycle. (Short by a week or so)
The response here was basically “use Firefox”. So, great! The default browser of Kubuntu is rendered useless for all things ubuntu-wiki for a third of the development because of a lack of testing, the only alternative being to install Firefox. This brings me on to my third and final point, also involving Firefox and the neglect of Kubuntu on that front.
Say somebody wanted to use the wiki bad enough the last two months that they were willing to install Firefox in Kubuntu. The first thing that they would notice is that it would pull in half the Gnome software stack. Very not nice for Kubuntu users. But what has been done? Nothing. Apparently Kubuntu isn’t equal with Ubuntu to the Ubuntu Mozilla team either. Others have noted this too, along with similar issues with installing the java plugin for use with Kubuntu. No consideration is made for the KDE environment and craptons of unneeded packages are introduced and forced upon any Kubuntu user who wants to use Firefox.
Edit: I’ve been notified that this is actually fixed. I can give another example in it’s place, though.
Up until shortly Alpha 5 KNetworkManager was broken by NetworkManager 0.7, once again due to an API change. At one point it looked like Kubuntu was simply going to be overlooked and left with an non-working networkmanager. Luckily this was fixed in svn, and now we are currently running an svn snapshot of knetworkmanager, which isn’t ideal but at least it works…
I agree with Aaron Seigo. Ubuntu routinely neglects Kubuntu, not because of what it does to support it but because of what it does to neglect it. The Kubuntu Team tries hard to deliver the best KDE possible, but that becomes hard to do when the maintainers of other packages routinely neglect the needs of Kubuntu, the Blue-headed stepchild. If Canoncial/Ubuntu want to become serious about Kubuntu/Ubuntu equality major mindset changes will need to be made across the board. Otherwise the stereotype of Ubuntu neglecting KDE will continue, because so far the stereotype has proven to be true.