KDE 3.5 Beta 1 "Kanzler" Released

The testing period for the next major KDE release has begun with the release of KDE 3.5 Beta 1, codenamed Kanzler. This will be the last major release in the KDE 3 series so make sure it turns into the best one by downloading and testing today. The 3.5 Beta 1 information page gives the download link as well as an important warning on using Qt 3.3.5. Packages are currently available for Kubuntu or you can use Konstruct to guide you through the compile.


Can't wait for the asyn filters in kmail

By eze at Thu, 2005/09/22 - 5:00am

kmail used to rock more for me. I have a long list of folders for my professional correspondence. It used to be I could press m for move. Then the list of folders came up and when I pressed say p it would take me to the first folder starting with p. A very nice feature that for some strange reason has been removed.

When I get a letter from a new person I make a new folder. That used to be easy, but now the new folder becomes a subfolder of the inbox. I have to leave the letter and go up to the top folder before I make the new folder, and try to remember the name while I do that. I realize the new way may be more logical, but the old way worked a lot better for me.

I also have an intense dislike for the decision to move the mail directory to be a subdirectory of .kde I help some 40-50 people with their kde-desktop and I have done so for several years. Every so often we solve problems by removing the .kde directory to get a fresh start. It is just too complicated to try to look through all the config files. Now we of course have to be extremely careful with that since important data have been moved into the .kde directory. Lots of people combine the usage of kmail with say mutt, logging in to the mail with mutt when they travel, and using kmail when they are at home. That has been made less convenient. So I do not think it is all progress what is happening.

By Erik K. Pedersen at Fri, 2005/09/23 - 5:00am

Well I don't know about the first two paragraphs, but I can't understand moving the folder into .kde either. Like you I have blasted mine out, or recommended to others to delete it when apps are acting up.

You can however change the location. Just open up ~/.kde/share/config/kmailrc and find the [General] section. Add the following line ->


Change the path of course to wherever you want. Then just put your mail folders in there and you're good to go.

By Andrew Rockwell at Sat, 2005/09/24 - 5:00am

Deleting .kde can be simply disastrous even if you don't have all your mail there. There are loads of other important data stored there, such as your calendar and contacts (by default, at least). If you really *have* to delete something, you might consider deleting only ~/.kde/config, but that is generally a bad idea too...

By Henrique Pinto at Sat, 2005/09/24 - 5:00am

Well this is a nice codename ...
--tk from germany :)

By no name at Thu, 2005/09/22 - 5:00am

Sorta political :)

By Sam Weber at Thu, 2005/09/22 - 5:00am

and Beta 2 or the release might be called Kanzlerin. We'll find out...

By burki at Thu, 2005/09/22 - 5:00am

How about Kaos?

By kujhg at Thu, 2005/09/22 - 5:00am

KStoiber anyone?

By Matt T. Proud at Thu, 2005/09/22 - 5:00am

Remember KDE 3.4 beta 1: Krokodile?

The KDE code names -- your best source on current events in Germany.

By Martin at Thu, 2005/09/22 - 5:00am

First things first: THANK YOU to all people involved in this release :-)

A question: has a decision been made on when to target for the release of 3.5 final ? The release schedule still shows an "undecided" status ...

Thanks !
-- MU

By MandrakeUser at Thu, 2005/09/22 - 5:00am

From how I unserstand the recent hype around klik, the following plan should be possible to implement:

1. Compile kdelibs+kdebase+kde[something].
2. Get yourself a fairly recent xorg X server.
3. Modify the standard klik wrapper script so it ...
. . a) ...starts Xnest/(or, maybe nxagent, the resize-able and fullscreen-able Xnest)
. . b) ...starts the KDE desktop inside the Xnest
4. Pack everything into a .cmg file.
5. Call it a deal and tell the world you have now an easy way to demo KDE-3.5 key features to them.

Anybody already working on this?

A coward

By ac at Thu, 2005/09/22 - 5:00am

you dont know how big this thing would be !!!


By chris at Thu, 2005/09/22 - 5:00am

But I know.

I tell you: it is do-able.

And it will be able to run on any hardware where you can run a current KDE, and still have some 100 MByte virtual memory free, outside the swap area.


By Kurt Pfeifle at Thu, 2005/09/22 - 5:00am

Even more easy,

With NX technology desktop sharing will get real good soon. You'll just be able to connect to a remote server running the latest and greatest.

By Bobke at Thu, 2005/09/22 - 5:00am

Ah, good to hear about NX.

Are you suggesting I should give up with the klik-able KDE-3.5 and instead setup an NX server with 3.5 running, and accommodate a user account on it for you?

By Kurt Pfeifle at Thu, 2005/09/22 - 5:00am

### "(...) following plan should be possible (...)"
Yes, it is.

### "Anybody already working on this?"

By Kurt Pfeifle at Thu, 2005/09/22 - 5:00am

Installed without any trouble and Just Works[tm]!
Ubuntu and KDE rock!

By rik at Thu, 2005/09/22 - 5:00am

Damn straight. No problems, just a straight, apt-get update, apt-get upgrade.

By matt at Thu, 2005/09/22 - 5:00am

Having some libc6-issues here holding back some packages...

The following packages have unmet dependencies:
amor: Depends: libc6 (>= 2.3.4-1) but 2.3.2.ds1-20ubuntu14 is to be installed

I know it's OT, but any idea?

By Ineiti at Thu, 2005/09/22 - 5:00am


after I upgraded to KDE 3.4 (from IIRC 3.2) I lost the feature to automatically reload websites periodically (in the "Extras" menu). I use SUSE9.1 and all other menus in the Extras menu seem to be still available.

Has this feature been user-friendlisized (removed by self-proclaimed usability experts) or is it still there (in 3.5)? If yes, in what package?

Thanks a lot for clearing this up.

By Roland at Thu, 2005/09/22 - 5:00am

This is a plugin now. Can be (re-)activated in the settings.

By Joachim Werner at Thu, 2005/09/22 - 5:00am

self-proclaimed usability experts is ok for me , as long as you dont mean it negativly. Kde as a whole is made by so called self proclaimed coders.

If someone is thinking about usability and remove a feature , its ok for me as long as he has thought about it.

If i like it or not is another matter. Dont think your way to use Software ist the best. You figure out anotherway oder just do the things different....

my opinion: I hate it if someone thinks bad of usability people just because he thinks his way of computerusage is the ultimate best way.....
So keep removing Things people use, only do it in a consistent and logical way.


By chris at Thu, 2005/09/22 - 5:00am

Self proclaimed coders can be seen to be coding, if they can't code their patches don't work and are rejected. Self proclaimed usability experts can make it less usable and get away with it.
"If someone is thinking about usability and remove a feature , its ok for me as long as he has thought about it."
I don't think we should ever remove features unless no-one wants them. Having less options doesn't make the software more usable. Let gnome have the simple-is-better people, KDE can have all the features you ever need, that way there's sense to having the two environments with people choosing between them.
"If i like it or not is another matter. Dont think your way to use Software ist the best. You figure out anotherway oder just do the things different...."
Software should serve me, not the other way around. I'll get it doing what I want if I have to patch the source myself.
"my opinion: I hate it if someone thinks bad of usability people just because he thinks his way of computerusage is the ultimate best way....."
I think badly of usability people because all they have done is made my programs less usable.
"So keep removing Things people use, only do it in a consistent and logical way."
No. If I wanted someone who thought they knew better than me what's usable taking away features I use, I'd be using gnome.

By mikeyd at Thu, 2005/09/22 - 5:00am

I see 2 big problems with "usability experts":

1) They talk a lot about the "average user", but they actually mean some (often imaginary) below-average user. They say when the so called "average user" doesn't need it, it's worthless. This way is stopping all innovation.

Most usability-features, that set Unix apart (multiple desktops, a third mouse-button that is acutally useful, Unix-style copy/paste, etc.) all were made before the "usability experts" declared all non-novice features worthless.
And if you look at tabbed browsing: It was "usability experts" who said all the time that MDI is "just wrong" and for a long time they also spoke against tabbed browing.

2) They think "usability studies" that put complete beginners for half an hour before a computer are the only metric that is to be used. However the average user in the real world doesn't use a computer for half an hour, he will use it for YEARS, probably DECADES.

According to usability tests all keyboard-shortcuts are worthless, because in the first half hour of use you won't use them. Keyboard-shortcuts come from a time where there were no usabilty experts, I doubt they would have a chance to get included today.

There are so many OBVIOUS ways to improve usability.

For example most mice have more than just 3 buttons+wheel. Wouldn't it be great to use the 4th mouse button for something useful? For example as another "Alt"-key? That way you could grab a window anywhere by pressing LMB+4thMB or resizing near the edges with RMB+4thMB.

Well, that behaviour isn't in MacOS, it isn't in Windows and beginners don't crave for it in the first half hour of computer use - so forget it, it simply won't happen.

Another example: We all have "Windows"-keys on our keyboards, why not use it for some more than just popping up the K-menu?
Wouldn't it be great to have standardized shortcuts for "maximize", "minimize", "next desktop", etc.?
Using one of the most accessable keys for just one function is a waste in my opinion, so I use Windows+left/right to switch between desktops and Windows+up/down to switch between applications on a desktop. (What is currently Alt+Tab and Alt+Shift+Tab)
Isn't it obvious that Alt+Shift+Tab is bad usability and something better should be introduced? Yes, keep Alt+Tab/Alt+Shift+Tab for compatibility, but there must be a better solution than that. And because applications use Alt-shortcuts and Ctrl-Shortcuts, wouldn't it be OBVIOUS to use Windows-key for everthing about the window-manager?
And what about Alt+F4? Wouldn't be Windows+Esc be more logical and (gasp) intuitive?
Yes, I know I can set all this in kcontrol (and I did), but we all know that a cross-project standard about this would be just great.

But again: It's not in Windows, it's not in MacOS, beginners are unlikely to use it in their first half hour in computer use, so it ain't gonna happen.

That is EXACTLY what usability experts should do: Organize keyboard shortcuts, create standards (preferrably together with GNOME), use hardware that is available and currently unused, etc.
Reorganizing kcontrol - AGAIN - is just a waste of everybodies time.

By Roland at Thu, 2005/09/22 - 5:00am

KControl was a mess, now its fine.

Usability discussions make al lot of sense.

It is not only newbie simplification, it is also about unification. E.g. always use the same keyboard shortcut for the wanted action.

By gerd at Thu, 2005/09/22 - 5:00am

This is actually quite contrary to the way that usability design actually works.

"They talk a lot about the 'average user', but they actually mean some (often imaginary) below-average user."

The standard mechanism for user modeling is called "personas". Personas can represent a non-technical user or a power user. For a given persona you try to map out the tasks that you expect from such a user and then proceed from there. See Tina's blog for more:

"They think 'usability studies' that put complete beginners for half an hour before a computer are the only metric that is to be used."

Again, really wrong. Please look around and read some usability tests. They almost always describe the type of users that they're testing and why. It's usually not (or at least not exclusively) complete beginners.

"According to usability tests all keyboard-shortcuts are worthless [...]"

I bet you can't find a single usability test that suggests that.

Really, you're not impressing people here by ranting about things which you don't even have the most basic clue about.

By Scott Wheeler at Fri, 2005/09/23 - 5:00am

"Personas can represent a non-technical user or a power user."

And that's completely nonsense because people have different levels of experience in different areas of computing.

Example: Me.

I'm a programmer, I have no problems administrating Unix-machines, but I use a word-processor maybe once a month and I have never done anything more than use "bold"/"italic"/"underline" and the font size in a word processor. I have absolutely no idea how to write chain letters, how to add footnotes or anything else in wordprocessing.

So what am I? Am I a power-user or a newbie?

And you know what? I don't care about hundreds over hundreds of settings in a wordprocessor [b]BECAUSE I DON'T SEE THEM ANYWAY[/b]. And it's the same vice-versa: Newbies don't care about "too many settings" because they just use the defaults.

By Roland at Fri, 2005/09/23 - 5:00am

Look, read the links. You'll see that the personas aren't labeled "power user" or "newbie" they're labeled "John Smith" or "Bob Villa" or whatever. They're meant to be constructions of semi-real people. Your argument doesn't make any sense in that context. And the bit about settings just completely indicates not really even understanding what usability is. Hint: it's not an alias for "those guys that want to remove everything".

By Scott Wheeler at Fri, 2005/09/23 - 5:00am

Well, yeah, it's just that i don't see much improvement happening. Usability-improved menu-on-top is a kind of nightmare to handle now (since you can't really configure it without text editor anymore). IIRC kmail had "check mail for this account" in account context menu -- and it's gone. The context menu has 3 items. It's not like it was overcrowded or anything. And it kept me swearing while i was using kmail.

Toolbars in KDE need cleaning up badly for ages now. But still we don't have an usable toolbar setup as default (or do we? i have been setting up custom toolbars for so long now...). We still fiddle with kcontrol though. But for one, i spend virtually no time in kcontrol. I do spend time in konqueror though. And the toolbars and menus are generally just too big to fit into my little brain. And, uh, well, the less trained personae have _much_ more trouble finding anything than i have.

As for usability experts, i am yet to see any significant contribution. Kmail only got worse for me (actually, i just use mutt now). As i said, i don't really spend time in kcontrol, and when i do, it's because someone shuffled things around. I still don't use most of the default toolbars and reconfigure them in nearly every app. The default shortcuts are a bit silly, not using win-key as a modifier. The tradeoff seems to be getting users from windows retrained quickly, but incurring the same crappy scheme on complete newbies. The "kde default for 4 modifiers" makes just so much more sense. Three-level k-menu is on the edge of bearability -- it's constant reshuffling doesn't help either (good i can just use mini-cli...). On the improvements front, i didn't really notice anything worth mentioning.

Yeah, there is a problem with actually getting some feedback from those usability experts. From the projects on openusability, very few actually have some reports. And those that are there are pretty limited in scope.

In short, usability salvation is not coming anytime soon. And it's still mostly coders' job to get the interface as right as possible.

By mornfall at Fri, 2005/09/23 - 5:00am

Most of this I can go along with -- our usability effort as any sort of organized affair is still in its infancy.

One thing that is worth picking apart a little bit is the comment on the big problems. As a programmer you can't start in the KDE project and start refactoring major library components. You usually start with a 5 line patch somewhere that does something you thought was pretty cool at the time, but in the grand scheme of things is pretty irrelevant.

It's no big surprise that our usability folks start in more or less the same way. We're just now starting to have a few people respected enough that they can actually question some of the bigger issues in KDE and have someone listen.

OpenUsability naturally isn't a silver bullet that will solve all of our problems, but when you consider that it's mostly about half a dozen people they've had a significant impact. If we support them they may grow into a larger group that can tackle larger percentages of KDE.

On KMail, well, I have to say that the improvements were subtle, but I feel in the right direction. The folder properties dialogs were simplified, the filter dialog was split into a main and advanced tab and well, I don't remember other things off the top of my head, but well, good things. :-)

In a nutshell I think it's important to be supportive of the stuff going on so that it will ideally grow into something that is on the whole more significant to KDE.

By Scott Wheeler at Fri, 2005/09/23 - 5:00am

Sorry but your rant is not very interesting: usability studies usually defines the expertise of the user they are targetting. Not all studies are for beginners.
I bet that Adobe makes usability studies for photoshop, for example.

As for the transition from tabbed browsing, if you happen to stop ranting and look, you'll notice that what people were criticizing is MDI which is different from tabbed browsing: opera offers full MDI (with icons inside the window) but I only used it in "tabbed mode", konqueror and Mozilla only provides tabs and I never heard someone asking for full MDI..

Also IMHO, tabbed browsing is still immature, as for example when one tabs generate a popup on mozilla, it blocks me even though I may be looking at another tab.. It should be less intrusive, for example sound a beep, makes the tab flash two times and then stay red: now I know that something is asking for my attention on this tab, and *I* decide when I'll want to deal with it.

By renox at Sun, 2005/09/25 - 5:00am

> wouldn't it be OBVIOUS to use Windows-key for everthing about the window-manager?
Agreed (and about the mouse button too).
Note that in fact it is already possible to have this in a limitated way: I've configured my RHE3 to iconify all the windows with the win+D setup, unfortunately it only works for the current workspace, which is annoying when you have a window which is displayed on all the workspaces, and I've no idea how to change this behaviour, to iconify all the windows of all the workspaces..

By renox at Sun, 2005/09/25 - 5:00am

It was always a plugin (see, what has changed is that now users have the option to choose which plugins are displayed.

By Richard Moore at Thu, 2005/09/22 - 5:00am

I just checked and at least in my installation I can't find it in the "plugins" section of the settings. I also couldn't find it in other sections and in kcontrol. Where are these settings?

Thanks a lot

By Roland at Thu, 2005/09/22 - 5:00am

Have you been looking at the right place? It's not "Plugins" in "Configure Konqueror...", it's in the Menu "Settings/Configure Extensions". Then choose the tab "Tools" and activate "Auto Refresh Plugin". I've sent you the German instructions via mail, so I'm a bit surprised that didn't help.

@Richard: Yes, technically it was a plugin, but from a usability point of view I'd only call it a plugin if it can be configured.

By Joachim Werenr at Thu, 2005/09/22 - 5:00am

Thanks a lot, it's right in there!

(BTW this is another example why we should have all settings in one place and never ever introduce "advanced settings" or even (shudder) a registry-clone)

By Roland at Thu, 2005/09/22 - 5:00am

Isn't it the windows registry that keeps settings together nicely in one place in consistent and hierachical format? And regedit is not meant to be the primary interface for editing it btw.
I yet have to understand this constant the-registry-sucks bashing. My .kde dir is polluted with leftovers from old programs too.

By uddw at Thu, 2005/09/22 - 5:00am

"Isn't it the windows registry that keeps settings together nicely in one place in consistent and hierachical format?"

Yes it "keeps it together", wether the user wants it or not.

Some of my settings are still from 1997 when I used *Solaris*. I still use some of those.

He, try to reuse anything from a registry from 1997.

Also the registry is very limited (knows only values and keys, unadequate for many, many settings) and undocumented.

"And regedit is not meant to be the primary interface for editing it btw."

Oh really, what's the great advantage about it? I see a lot of trouble and problems with the registry:

- Incompatible between versions
- Hard to extract settings for a single program
- Not documented and documentable
- When messed up very hard or even impossible to repair
- Impossible to do (cp mysettings settings; /bin/program; cp factorydefault settings)
- Needlessly adds another concept (The whole filesystem is a hierarchical tree-like structure, no reason to add another similar structure)

I wouldn't find it so bad if there were some advantages that would outweigth all these problems. But I still haven't found it.

"My .kde dir is polluted with leftovers from old programs too."

Yeah, just do "mv ~/.kde ~/.oldkde", restart, then copy all the settings you really need from .oldkde to .kde

You will have a very hard time trying to do the same with the registry.

By Roland at Thu, 2005/09/22 - 5:00am

IIRC KConfigXT only knows about keys and values, grouped by single-hierarchy sections. How is that not limited?
The key documentation is nice, but they are not included in the config file itself, so it's not really helpful without special tools.

You can dump registry subtrees to human-readable files and import them later if you feel the need to do that.
You can attach custom ACLs to every single item.

I agree that it's easier to move files around that fiddeling with all kinds of special purpose tools. I guess that's just the windows philosophy, apart from the obvious performance reasons.
Reminds me of the reiser4 pseudofiles vs. extended attributes discussion btw.

I really have no reason to defend the windows registry, but KDE's configuration system is not so far superior that the Windows Registry is pure crap in comparison.

By uddw at Thu, 2005/09/22 - 5:00am

The first part is already underway, I'm afraid.

By ac at Thu, 2005/09/22 - 5:00am

Fair enough - I regard it as a plugin because it is optional if you choose install it.

By Richard Moore at Thu, 2005/09/22 - 5:00am

Hello people,

Please excuse me for the stupid question, but how do I compile the beta through Gentoo ebuilds with all the debugging information enabled? I tried the 'debug' USE flag, but the tracebacks were still missing symbol names, so I must be missing something. What DOES the debug flag do, exactly (or, for non-Gentooists: just what does the --enable-debug=full configure option do)? How does the 'nostrip' Portage feature differ, and can I enable it only for certain packages?

Thanks, and sorry for the mildly offtopic post. I just really wish to do what I can to help make KDE 3.5 the smoothest KDE 3 ever. :)

By Anonymous Coward at Thu, 2005/09/22 - 5:00am

Have a look at your C(XX)FLAGS some can make debugging deficult. I think it is stated in the manpages for GCC, which ones it is.

By Morten Sjøgren at Thu, 2005/09/22 - 5:00am

Thank you! I read the manpage, which was insightful, so thanks for the useful pointer. My CFLAGS are fairly conservative, though, and shouldn't be a problem.

Would you happen to also know 1) what --enable-debug=full does exactly, and 2) if I can turn on 'nostrip' only for compilation of certain packages?


By Anonymous Coward at Thu, 2005/09/22 - 5:00am

You might want to look at the kde .ebuild's so you can see just what is happening at the ./configure stage. --enable-debug=full *should* be all you need, but it might be getting cancelled out somewhere.

By John at Fri, 2005/09/23 - 5:00am

Oh, I did! I just don't know what --enable-debug=full does exactly, nor how it relates with the 'nostrip' feature. Apparently the two are unrelated -- 'nostrip' is a purely compilation-time thing, and instructs the package management system not to strip the executables from its symbols, and --enable-debug is a purely KDE thing that does... other stuff -- and both probably need to be turned on for full debug support. So I did.
I was just hoping that USE="debug" would instruct Portage to also turn the 'nostrip' feature on and clear out compilation flags such as -fomit-frame-pointers, but apparently, it doesn't.

By Anonymous Coward at Mon, 2005/09/26 - 5:00am

Check the flags don't have something silly like -fomit-frame-pointers in them.

"Conservative" still usually means more than your average distro with gentoo..

By Alistair John S... at Fri, 2005/09/23 - 5:00am

Thank you kindly. :)
I'm at -O2 optimization level, which turns off -fomit-frame-pointers on the architectures where it prevents debugging. I had -fomit-frame-pointers explicitely on before and turned it off before compiling. Thank you for the advice!

This being said, I'm having loads of issues with KDE 3.5, and I don't know how many are bugs and how many might be local configuration side effects. Do you know where I could discuss those without uselessly spamming The Dot, some forum somewhere? Thanks. :)

By Anonymous Coward at Mon, 2005/09/26 - 5:00am

Is planned a new Klax release?

By Anonymous at Thu, 2005/09/22 - 5:00am