KDE SC 4.5 Just Got Hit With A One-Week Delay

[pvtcupcakes]

Not again! Switching a few applications to Akonadi and leaving the rest as is was the second big mistake after the 4.0 disaster. It may work for people who only have simple e-mail accounts. Groupware functionality and IMAP storage support however is awfully bugged.

So if Akonadi support doesn't make it into 4.5 it seems that 4.6 will be the first release where the full KDE framework is finally implemented. I really hope that they did a good job in designing this framework so it can be used for a long time. The last few years were a disaster!

From what I understand, Akonadi, Kmail and the rest of KDE PIM are being updated in 4.5.1. So it's only a one month delay.

[ChrisXY]

Stop using Kubuntu or Arch and get a working distribution, like openSUSE.

Since Archlinux mostly delivers vanilla Packages do you tell me that KDE needs heavy patching in order to work?

[pingufunkybeat]

Vanilla KDE on Gentoo (minimal patching, if any) works just fine.

[ChrisXY]

Plasma hasn't segfaulted for years here. Then again, I don't do fancy stuff, just a couple of widgets and hiding panels.

"All the time" sounds extreme. Have you reported a bug?

I have tried many versions from 4.0 to 4.4.0 and every single version segfaulted often. Of course I always deleted ~/.kde4 before.

Until 4.3 it was easily reproducable i.e. when changing the panel theme twice in a short time.
When 4.4.0 came out I of course tried it and had 3 Plasma segfaults in one hour without doing any fancy stuff either. For example I just had placed the qalculate plasmoid on the desktop and clicked on it to demonstrate it to someone other because it is really cool - but then plasma segfaulted hard. Before it had worked. It's just random.
I'm not even speaking of plasma not threading right (maybe it's better now, I didn't test it since 4.4.0): Try the qalculate plasmoid and just enter a calculation that will take a long time and your whole plasma desktop will be stalled.

[yotambien]

It traditionally was.

It changed when they stopped the desktop version and rolled out Fedora, and changed again when Ubuntu showed up.

OK. I didn't suspect this. Not that it means that Ubuntu was fed solely from other distributions, of course.

This stuff is hard to measure. The web-based estimated have been hovering around 1% for Linux since as far back as I can remember.

I think that the impression that Ubuntu doubled or tripled the Linux installed base, as assumed by some, is quite wrong.

It's impossible to count the Linux machines and for that the web-based estimations vary a lot.
As I said before, in the greek Ubuntu forums we have 5000 noobs. Where did they come from? They didn't use Linux before. Most of them of course are dual booters. That counts them as windows or Linux machines?

Yes, it's hard, but that's what makes of it an interesting problem, I guess. I really think web statistics are the way to go. For instance, they automatically take into account difficult cases such as the one you mention, i.e. dual booting users. In theory, if you are just interested in what people use in the desktop--as opposed to the number of Linux installations--those users should contribute a fractional amount proportional to the time they spent with each OS. Assuming they browse the web with both systems you'll have that fractional contribution at once.

There are some problems, of course. First, extrapolating results from one website to the general population is quite dubious. The most I think you can reasonably do is to extrapolate to the geographical region where the website offering the data is mainly accessed from. Still in this restricted case, the website in consideration has to have an as wide as possible audience for the extrapolation to be meaningful. Google would, of course, be ideal. Sites such as The New York Times or The Times would not be the worst out there (for the US, of course), although they would very likely understimate young users by a long shot. I thought of Wikipedia as being perfect, but probably in this case there is an education bias at play.

The other obvious problem I see is the margin of error of these statistics combined with the low numbers of Linux. With results in the order of 0.5 to 3% it's difficult to trust any changes. For instance, the Wikimedia OS statistics go back to April 2009 (as far as I know), with the following results:

Win 89.5
Mac 6.05
Lin  1.49
iPh  0.93
BB  0.15
Sym 0.09

The last one, from June 2010:

Win 86.32
Mac 6.63
Lin 1.88
Iph 1.78
BB  0.41
Sym 0.21
iPad 0.19

From those numbers, Windows appears to be less used to access the web. I can even believe that it is somehow less used in general, given that a) its decreased 'market share' is not only explained by the increased presence of mobile platforms; b) Mac and Linux experienced a modest increase during the same period. However, where would it be licit to extrapolate these numbers to is a difficult question, and the low Linux percentages would be quite affected by any uncertainties. Even more if we wanted to break them down by distribution. Still, better than nothing. With statistics from the right website, spanning back to when Ubuntu was released this should be doable.

[Prescience500]

I love KDE, but the only problem I have with it is it's focus on apps (not that GNOME is terribly different.) The first thing I do when I install a KDE distro is to rip out the KDE apps like Konqueror, Kmail, Kopete, and what not and replace them with what my own software. I imagine that many people do the same. Besides, distros just go and package whatever they want anyway.

[pingufunkybeat]

The apps (actually, the infrastructure) are the main point. It helps develop powerful apps easily. Stuff like K3b and Amarok.

Apps like Firefox and OpenOffice implement half of KDE themselves just to get the needed functionality. Both of these have historic reasons for doing this, and it's near impossible to rip out all the app-specific code.

Using just KWin and Plasma is a bit of a waste, IMHO. FWIW, I've often run a different window manager, but mostly KDE apps.

[Prescience500]

By apps, I meant non-infrastructure apps like office, web browser, calendaring, address book, email, instant messaging, media player, etc. Things like dolphin, phonon, kpackagekit, almost everything in system settings, and what not are good.

[pingufunkybeat]

I think it's good to provide everything you need for everyday life. If people then choose a different or more powerful program (such as Gimp, OpenOffice or Firefox), then that's also OK.

The good news is that OpenOffice has KDE4 integration now, and Konqueror can use Webkit, making it a real competitor again (KHTML is starting to fall behind due to the lack of momentum).

[korpenkraxar]

Since Archlinux mostly delivers vanilla Packages do you tell me that KDE needs heavy patching in order to work?

Heh if only you knew how much my OpenSuse-running friend and I have been over this

Anyways, running the latest RC on Arch. Seems to be doing ok except for some weird coloring of the task panel every now and then. Could be the theme I am using though.