Only some applications break configuration files. Application Data is usually fine. Plasma usually is the worst affected ( Plugin renames, new functionality requiring file restructure, etc )

bcooksley wrote:Only some applications break configuration files. Application Data is usually fine.

Right.
Actually applications usually don't break configuration files either.

However, reports indicate that different version's of KDE are being run with different local prefix, thus breaking the built-in configuration "import/upgrade" of applications.

It seems that the additionally required upgrade tools are not run reliably thus creating the impression that applications would not deal with their old config correctly.

Unfortunately this makes it almost impossible to detect bugs in the application's upgrade paths should there be any.

Cheers,
_

ivan wrote:at the time I entered the world of free software (late nineties), Red Hat x.0 was usually beta quality, x.1 was final and x.2 was the polished release (just like KDE 4.x).

I remember that too. I also remember that Redhat had a **** reputation because of it. If not for their marketing department targeting corporate CTOs, it would have ended up just another distro. We don't have anywhere near the marketing skills that Redhat does.

A x.0 release means that the developers consider it to be finished. There will of course be bugs, just because it hasn't had wide usage to shake them out, but the numbering indicates that there are no outstanding critical bugs, the software is feature complete, and the development community feels that it is stable. By the time of x.0.2, the software should be rock solid.

I can not agree that it had a **** reputation. Or to be more precise, it did, but it was still one of the strongest distros at the time (with Debian and Slack, and later Mandrake). We all knew that you just don't install x.0 unless you want to have the latest and the greatest. but with a fair share of bugs.

The point-oh release can not be what you said in the second paragraph ever. At least not when the software is concerned.
- feature completeness - what is that? What software is feature complete? What would be the point of making (x + 1).0, if x.0 was feature complete?
- bugs - show me a non-trivial program without bugs, and I'll show you a real, live and kicking Santa Claus.

x.0 is just a mile-stone in every software's life. And it should usually mean that it is ready for public. I agree that it is debatable whether KDE 4.0.0 was ready for public, but I, for one, have had fully switched to it long before the final version was released, so it *was* ready, at least for some part of the public.

To be honest, I think we shouldn't give Linus' words too much weight. Agreed, he wrote a terminal emulator, and this has grown into a full and respectable OS. Also agreed, he is still very active in Linux kernel development.

But, first of all, Linux has grown much larger than Linus Torvalds, and if we gave him all the credit, we would not be fair to thousands of developers all over the world. Linux would not have become Linux if it was for Linus alone...

Second, as far as I know, Linus Torvalds has never developed anything for any DE or WM. So, as far as OS GUIs are concerned, Linus Torvalds is just a regular user, like most of us.

And third, KDE is not a Linux GUI. KDE is a desktop environment which runs on many OSs, one of which is Linux.

So, the connections between Linus Torvalds and KDE are

• Linux is one of the operating systems KDE runs on, and KDE is one of the desktop environments that run on Linux, and
  
• both KDE and Linux are very successful open source projects

No more and no less.

Some people need gods, idols or icons; they are free to follow Linus Torvalds, or anyone else, blindly.

But, fortunately for KDE, new Linux users nor *BSD users nor users looking for an alternative for Microsoft nor users willing to try some other UI for Mac nor Solaris users looking for a decent GUI, see Linus Torvalds as a god, idol or icon. They look for quality, without caring about the opinion of someone who was someone in a different field and/or a different era.

So, if people want to switch to a different DE because Linus Torvalds does, let them. They're punnishing theirselves more than they're punnishing us. Maybe they'll be back, maybe they won't. We'll see.

KDE (as a whole: developers, promo people, forum admins, etc.) should focus on the future, not on the past.

So I say: thank you, Linus, for your opinion; it is appreciated and taken into account, but now, please go back to doing your thing, while we do our's.

ivan wrote:- feature completeness - what is that? What software is feature complete? What would be the point of making (x + 1).0, if x.0 was feature complete?

"Feature complete" is well known concept in software engineering. Without it you might as well just have a cron job tarball sources and stamp versions on them every three months. Feature complete means that it contains the intended functionality for that release. There is typically a release plan that indicates the features for the release. For small projects you can get away without it, but for large projects you can't.

- bugs - show me a non-trivial program without bugs, and I'll show you a real, live and kicking Santa Claus.

I didn't say without bugs. I said without known critical bugs. Big difference. You don't ship software if there is a known bug that corrupts the users's data, for example. It's also expected that there will be a limited number of known moderate to serious bugs. There will always be *unknown* bugs, but the known bugs need to be squashed.

There's too much flame and troll already. I do the same with FC9 and KUbuntu 8.10 Linus does because KDE4 at present state is just not for me. And I am waiting for stable and usabe KDE4. Sure I'll be back as soon as I see important things working.

Why all this disaster with KDE4 happened? Because of terribly wrong marketing and PR. KDE developers are great and doing wonderful work. But KDE PR people and KDE release managers have made and still doing terrible mistakes describing KDE4.x as fantastic new product ready for any distro.
Only after massive criticism they started to publish "Is KDE4.x for you?" statements.

Fedora 9 was terribly wrong providing 4.0 as only one KDE desktop. Even KUbuntu 8.10 was wrong providing KDE4.1 as only one choice. Why they rushed to 4.x? They followed wrong marketing and PR **** of KDE guys.

I think that KDE developers can be blamed for only one thing. They haven't provided separate config/work environment for 3.x and 4.x to have both 4.x and 3.x in one place. That's why people can not use KDE3.5 for everyday work and KDE4 for fun.

@Brandybuck
Well, KDE 4.0 *had* all the features it was supposed to have, and as far as I know, haven't had any corrupt-data bugs.

So, the /feature complete/ by your definition (which I mostly agree with) is not what the users though of. /Feature complete/ in most user's pov was to match 3.5.x feature-by-feature.

Every user lost from KDE is a bad thing, but it's not they will never ever again use it...
Even if I would to switch the DE I would still want to see what happened to my old DE.
I don't think that there is anything like point of no return in computers and their software...
OK I'm taking this from my experience, but still most users (are hopefully similar).
I was running Linux only box, but I did install Windows 7. I am using KDE, but when Gnome3 comes I'll try it somewhere (probably not on the same install as KDE as then the menu gets very cluttered..)
I might try E17 to see what's all the fuss about it... I might change Linux for BSD or Solaris, or (yes) even Windows...
I don't know.
But I know that no matter what I'll still use KDE or will check it out if my primary DE won't be KDE.

Thing go even worse then KDE guys can imagine.

It takes a lot of efforts to move all settings, accounts, network setup, etc to other environment. I use KMail for years but now I'm moving to other mail client because KMail crashes constantly. So will I return? Do not know, too much time to setup things again. They broke Konqueror and many other things, I moved all my passwords for sites to Firefox and I do not want to move it back by hands again.

I just tested 4.2 release and decided to move to Gnome last day-to-day apps and settings. I do not believe KDE will be good in nearest future because they did not it good enough for year after 4.0 release. It is still broken, apps crashing, apps missing, and so on.

Aaron may say mantras that KDE was bold move toward progress and 21st century but for me personally it was move to nowhere. I may expect exactly the same behavior on 5.x series when I'll be pushed to the same idiotic process of re-configuring everything. New "bells and whistles" do not make me so happy I'd like to spend hours and hours restoring my working environment.

KDE guys may say that they work for free but why they can not establish KDE foundation to get paid core developers like FreeBSD project does? Paid workers are obligated to do smooth version transition and other things users expect.

All KDE4 thingy is the stupidest event happened in last 5 years. It shows the biggest drawback of FOSS - dependency on foolish decisions of few project leaders. And what is most enraging that those leaders insist that they are completely right and there were no mistakes.

@alukin: KDE PIM applications will mostly have to be completely rewritten in some cases for the new Akonadi framework. This should fix most of those crashers.

You also failed to define "many other things". What does this mean? Which applications are affected? The developers cannot fix problems they are not aware about. Please file bug reports if they do not already exist, preferrably with a full backtrace and detailed reproduction information.

alukin wrote:
I just tested 4.2 release and decided to move to Gnome last day-to-day apps and settings. I do not believe KDE will be good in nearest future because they did not it good enough for year after 4.0 release. It is still broken

How?

apps crashing

All systems suffer from crashing apps. I haven't seen KDE4 suffer from this problem any more than some other software does. 4.0 yes, 4.1, not really.

apps missing

Such as?

and so on.

It would be nice for a change to receive tangible feedback as opposed to vague stuff like "it's broken and so on".

KDE guys may say that they work for free but why they can not establish KDE foundation to get paid core developers like FreeBSD project does?

Would that make things better? many of the core-developers are already paid to work on KDE.

Paid workers are obligated to do smooth version transition and other things users expect.

Um, no they aren't. If they are, how can you explain the disaster that is XP > Vista-transition? Reemmeber: that was done with paid workers only.

alukin wrote:I think that KDE developers can be blamed for only one thing. They haven't provided separate config/work environment for 3.x and 4.x to have both 4.x and 3.x in one place. That's why people can not use KDE3.5 for everyday work and KDE4 for fun.

Not true!

You can have as many versions of KDE installed side-by-side. It's just that many distros' devs are lazy bums and don't want to maintain both 3.5 and 4.X.

Don't believe me? Have a look here - I had 3 versions of KDE installed at the same time because KDE has a option to install into different places. The Gentoo devs used the option, your devs didn't.

So the blame is not on the KDE devs. PR, maybe. But the biggest blame is on the distros who don't think of their KDE userbase.

apps missing

Such as?

The ability to manage printers from with the system settings applet & also ability to configure network settings (IP address etc) from the network settings applet in the system settings come to mind off the top of my head.

Please don't take this as criticism I love KDE & have been using it for a long time but I feel there are some missing features still.


Cary

caryb wrote:

apps missing

Such as?

The ability to manage printers from with the system settings applet

I do believe that printing has been significantly beefed up in 4.2 Although I don't print with Linux, so I don't have personal experience.

also ability to configure network settings (IP address etc) from the network settings applet in the system settings come to mind off the top of my head.

To my knowledge that functionality has never been provided by KDE itself, but by a third-party (usually the distro). That said, it seems to me that Knetworkmanager works just fine in KDE4 (I use it all the time) and it can manage network-settings....

Please don't take this as criticism I love KDE & have been using it for a long time but I feel there are some missing features still.

It seems to me that the two features you mentioned are included (although I'm not 100% sure about printing, since I don't print).