AUG
16
2004

NewsForge: Q&A with George Staikos about KDE 3.3

NewsForge has caught up with KDE developer George Staikos to talk with him about the upcoming KDE 3.3 release, KDE's future, Konqueror, KWallet, freedesktop.org, KOffice and more.

Comments

I have a little request for the upcoming version of KDE 4. After a little conversation with some people on IRC I was told that KDE 4 will become a bigger task of change (probably related to QT 4) and therefore I'd like to take the chance to ask for this wish and hope that developers might like this idea.

Read more here: http://bugs.kde.org/show_bug.cgi?id=86986


By Ali Akcaagac at Mon, 2004/08/16 - 5:00am

That sounds like a really good idea, I voted for it! Seems like everyone would win with that change!


By Corbin at Mon, 2004/08/16 - 5:00am

I think this is a very good idea. I voted for it, too


By Michael Thaler at Tue, 2004/08/17 - 5:00am

If you break things in KDE4 then break it completly, not just a little bit to make compromises, and make a clean and efficient design.

have fun hacking.


By chris at Mon, 2004/08/16 - 5:00am

I say if something should be changed but it will break things, then nows the time to do it! Anything that will make KDE faster/smaller/better but may (or will) break old apps is worth it!

Though if you plan on breaking something just for fun that may not be the best idea (unless fun = making KDE better ;-) )


By Corbin at Mon, 2004/08/16 - 5:00am

But KDE alrealdy has a clean and efficient design. There are a few places where changing the design a bit will makes things easier but developers are certainly not going to throw the design that made KDE a success for the last years.


By Philippe Fremy at Tue, 2004/08/17 - 5:00am

hi sorry i know this is offtopic but i have to ask for it somewhere, perhaps it has to do with the feature of kde , so the interview :>

i once had a kde version, when i tried to overwrite an mp3 with antoher mp3
i got a dialog , where i could play both mp3's and after that i could decide which one i overwrite or rename.

i also worked with images , i got presented with 2 image preview's so i could decide not only on the filename.

i dont know if it worked with more things , but i can imagine....

i think it was around kde 3.1 somewhere - then it disappeared , i thouht it was a really really useful feature.

what do you think and know about this ????


By chris at Mon, 2004/08/16 - 5:00am

Install kdeaddons.


By Anonymous at Mon, 2004/08/16 - 5:00am

yes I still have this nice option in 3.2.x


By JC at Mon, 2004/08/16 - 5:00am

I may just of overlooked one, but I wanna learn about 4.0! Someone should do a 4.0 Q&A or article! 4.0 looks like (and sounds) like its going to be a HUGE change, so all the stuff on it will be really intresting!

If I've just not seem them, does anyone have a link to one?


By Corbin at Mon, 2004/08/16 - 5:00am

The trouble is that no one is 100% sure what we're going to do yet! There are a large number of things planned, for example:

- A new multimedia architecture
- An improved notification mechanism
- Better support for MVC in the APIs
- Various API cleanups are already prepared (and marked in the header files)
- Maybe the introduction of a bridge to DBUS

A lot of the changes will depend to some extent on what ends up in Qt 4.0.


By Richard Moore at Mon, 2004/08/16 - 5:00am

This is from the interview:

" KDE4 is just around the corner now -- I predict late 2005 -- "

It is a typo right ? Does he mean late 2004 ?


By MandrakeUser at Mon, 2004/08/16 - 5:00am

What year is this???

Hmm..

Yeah. He probably means late 2004


By Allan S. at Tue, 2004/08/17 - 5:00am

Nope.

KDE 4 will be based on Qt 4.
And Qt 4 will not be ready in 2004, perhaps early 2005.
And then it'll take some time to port programs etc...

See: http://doc.trolltech.com/4.0/tech-preview.html

Quote:
"Depending on the feedback from this Technology Preview, we will provide a second Preview in Q3 2004, enter the beta phase in Q4 2004 and release the final Qt 4 in late Q1 2005."


By tbscope at Tue, 2004/08/17 - 5:00am

"And Qt 4 will not be ready in 2004, perhaps early 2005."

So, wouldn't it make sense to make one more point release,
say 3.4 , before going 4.0 ? 6-8 months from now to Qt4 seems
like too long a wait. And then all the porting you mention, etc.

More than enough time for another point release !


By MandrakeUser at Tue, 2004/08/17 - 5:00am

I don't think they'll do a minor version (e.g. 3.4.0), but will certainly do bugfix point releases (e.g. 3.3.x) in between 3.3.0 and 4.0.0.


By Ian Eure at Tue, 2004/08/17 - 5:00am

Yes, that's my feeling from lurking a bit on the lists. Of course I have no time to develop, so I shouldn't be vocal. I have the feeling though that a minor release (with new features that don't break things) would be ideal. Think that if they go directly for a 4.0 we will either see no new features for about a year, or otherwise the 3.3 branch will end up being more than a bugfix branch, and will absorb new features and hence introduce new bugs (regressions), which would be a mistake IMHO ...


By MandrakeUser at Tue, 2004/08/17 - 5:00am

No, I don't think this is a typo. Qt 4 has to be out first, and there are huge changes to KDE planned (see the other comments).


By cm at Tue, 2004/08/17 - 5:00am

Oops, it seems I haven't hit reload in an hour or so :-P
Nothing new in my comment.


By cm at Tue, 2004/08/17 - 5:00am

KDE 4.0 will probably take more than a year, just as KDE 3.0 did. So late 2005 sounds about right.


By Rayiner Hashem at Tue, 2004/08/17 - 5:00am

Why not to wait until 4.0 is finished ?
No need a release date. Just making a new great release.

And as in the past for a major version it will take longer than a minor version.


By JC at Tue, 2004/08/17 - 5:00am

I do not think that it is a typo.

Late 2004 is just not possible, as Qt4 final will not even be released at that time.

Have a nice day!


By Nicolas Goutte at Wed, 2004/08/18 - 5:00am

A Question, what requeriment for KDE 4? My old PC is AMD DURON 950MHZ, 384MB RAM, GeForce2 32MB, i'm using SuSE 9.3, may my machine run KDE 4?


By Joel Zerpa at Tue, 2005/08/23 - 5:00am

That's funny. You answered to a post about 1 year after the last one.
As to KDE 4 we now know that it's going to be released in late 2006 or early 2007 I think. And I think it will support you old PC but you won't be able to switch all the graphic elements on.
Well, At least I hope it's going to work on old PC since mine is a PIII 900MHz 256 RAM...


By Jonathan Müller at Sun, 2006/02/12 - 6:00am

Its funny you say late 2006 early 2007, I just think it is funny, ... started off saying late 2004 early 2005 not it is 2006 and still no KDE4.

Almost like Steam saying their releases and it is normally like a year or 2 later...

I am not bashing I just happened to come across this thread and thought it was funny.

People should always know to to ask about a release date because 9/10 it is never on time as originally expected, with respect to the developers always finding some bugs they want fixed before release, or new programs they would like to incorperate into it.


By Narcussist at Fri, 2006/02/24 - 6:00am

And... Year later... It is even funnier now! I was googling for kde4 release date and this is what I got. I can not believe that it is still on hold. OK I want it all and I want it NOW. :-)

Well guess I'll have to wait till late 2008/early 2009

PS
Anyone knows will it work on my old Dual Core 5GHz / 4GB RAM / 1GB NVIDIA / 1TB harddrive?


By Encho at Tue, 2007/02/27 - 6:00am

Hm I've read that it's supposed to come out in Fall 2007 and we can expect a preview (i mean a preVIEW not a dev snapshot) sometime in july or so. well that's theoretical of course but still. and as for computer ... i think that might work. try adding some RAM though ;-)


By andrea at Wed, 2007/02/28 - 6:00am

Hm I've read that it's supposed to come out in Fall 2007 and we can expect a preview (i mean a preVIEW not a dev snapshot) sometime in july or so. well that's theoretical of course but still. and as for your computer ... i think that might work. try adding some RAM though ;-)


By andrea at Wed, 2007/02/28 - 6:00am

This thread started on August 2004???
The estimated relase I think is now January 2008...
... maybe, I will beleive it when I see it
nearly 4 years delay so far
I hope this (release) time is for real...
(I wonder what we had said about XP or Vista comming 3 years later that initially scheduled)


By jose marin at Fri, 2007/10/05 - 5:00am

Funny you'd mention it:

From http://en.wikipedia.org/wiki/Windows_Vista :

"On January 30, 2007, it was released worldwide to the general public"
[...]
"Microsoft started work on their plans for Windows Vista ("Longhorn") in 2001, prior to the release of Windows XP. It was originally expected to ship sometime late in 2003 as a minor step between Windows XP (codenamed "Whistler") and "Blackcomb" (now known as Windows 7)."
[...]
"Some previously announced features such as WinFS were dropped or postponed, ..."

Sound familiar?


By cm at Fri, 2007/10/05 - 5:00am

"nearly 4 years delay so far"

What? Read up on your history; October 23rd 2007 is the *only* official release date ever given for KDE4.0. Even going by that article, where a single KDE dev makes a personal prediction of "late 2005", the "delay" is little more than 2 years.

I swear, dot.kde.org users get dumber every day.


By anon at Fri, 2007/10/05 - 5:00am

I can't wait for KDE4 to come out! I just wish that it was going to be ready in time for the next release of Mandriva 2008. But that is cool. I'm getting ready to buy a new machine and I'm exited to use Compiz Fusion. I guess that is on Mandriva 2008. But anyhow KDE is way better than anything else out there.


By MattKingUSA at Tue, 2007/10/09 - 5:00am

IIRC Dirk Müller did quite a lot of work on KHTML but from CVS commits it seems he has mostly stopped working on it. Is there anyone (and who) still working on KHTML?


By MK at Tue, 2004/08/17 - 5:00am

I think it's mostly Leo Savernik, Germain Garand, and Tobias Anton working on various different parts of khtml lately. Stephan Kulow is helping with regressions, Zack Rusin is occasionally doing safari merges, Koos Vriezen is the main guy working on kjs. Waldo is doing occaisional bug fixes, and bj working on the access key stuff.

So yeah, Dirk doesn't seem to be working much on khtml anymore.. not as much as he used to at least.

oh well... Dirk was the main guy behind khtml for a while, but khtml has lost a lot of important developers before. Lars Knoll, the author/founder of the modern incarnations of khtml, became too busy a long time ago, as did Harri Porten, the author/founder of kjs.


By anon at Tue, 2004/08/17 - 5:00am

> Is there anyone (and who) still working on KHTML?

The Safari team is doing major work on khtml, this has "only" to be merged with KDE's "branch".


By Anonymous at Tue, 2004/08/17 - 5:00am

I hope that Dirk is doing IE-extensions to konqueror. Someone at Novell is doing this kind of work.


By Jaana at Wed, 2004/08/18 - 5:00am

> Where do you feel that KDE has to improve to gain more new users?

We must have better quality control; TQM should be our goal. The problems like we had with the last few releases must not continue. Better quality is more important to most users than "cool" features. All regressions are infuriating to users. Therefore, all regressions should be considered for showstopper status no matter how small they might appear -- they should be judged instead based on how many users they will affect.

I noticed that there isn't one word about fixing bugs in the whole interview. A large omission. To gain new users, we need to ALWAYS ship a better quality product than Redmond does.

But, it appears that there is hope. If there will be no 3.4 and 4.0 won't be shipped till late NEXT year then we have a lot of time to fix bugs and improve our quality systems.

To be clear about this: I want to see a better quality KDE product. Is there anyone that disagrees with that goal?

--
JRT


By James Richard Tyrer at Tue, 2004/08/17 - 5:00am

> > Where do you feel that KDE has to improve to gain more new users?
>
> We must have better quality control; TQM should be our goal. The problems
> like we had with the last few releases must not continue. Better quality is
> more important to most users than "cool" features.

Well said. For a start, there needs to be more backporting of bugfixes to previous KDE branches (not just backporting of highly publicized security fixes). e.g. I'm using KDE_3_0_BRANCH and kde-config crashes given dodgy arguments, KMail indexes are frequently corrupted (so if you click on a mail in the search dialog, it lands you in the wrong folder; or message aren't automatically marked as read), Konqueror can't even display http://bugs.kde.org, Konqueror leaks uncontrollably at http://www.kdedevelopers.org, Konqueror javascript doesn't work half the time, starting 2 KDE apps in quick succession outside of KDE crashes with DCOP errors, ark can't open compressed files stored on vfat and brings up a bogus permission error (I'm using an unexpected scheme where I'm in the "vfat" group) etc. etc.

This is not quality. This is not supporting branches about 6 months after the initial release. Even MS gives more support.


By w at Tue, 2004/08/17 - 5:00am

There is no good reason to run KDE 3.0 anymore (and to complain about its bugs today).

> This is not supporting branches about 6 months after the initial release.

KDE 3.0 is over 2 years old!

> Even MS gives more support.

And they request money for their product to do this.


By Anonymous at Tue, 2004/08/17 - 5:00am

> There is no good reason to run KDE 3.0 anymore (and to complain about
> its bugs today).
Upgrading to a new release had the potential to break stuff - risk.
e.g. most businesses are unwilling to upgrade to the next version of
Windows just so that the icons look prettier for example - that want
as little change as possible, just fixing bugs to reduce risk.

For a home user, it's the hassle of upgrading - either compiling from
source or using RPMs that may or may not work depending on how bad their
Linux distributor is. Not to mention the huge download on dialup.

These are pretty legit reasons I think not to upgrade to 3.3 instead of say 3.0.6 (if it ever comes out).

> KDE 3.0 is over 2 years old!
Windows 98 is over 6 years old and still provides at least the same amount
of functionality for me (if not more, due to Office and market share / compatibility).

> > Even MS gives more support.
> And they request money for their product to do this.
Fair enough; KDE developers already create a great desktop when most are volunteers. But this doesn't sit well with the enterprise IMO (
http://desktop.kdenews.org/strategy.html). I'm just not in a position to upgrade because of download and/or compile time.


By w at Tue, 2004/08/17 - 5:00am

> For a home user, it's the hassle of upgrading
> I'm just not in a position to upgrade because of download and/or compile time.

You have the option to pay for a new distribution in a store.


By Anonymous at Tue, 2004/08/17 - 5:00am

> You have the option to pay for a new distribution in a store.
but i'm too cheap to do that since i would have to pay instead of my non-linux parents :)


By w at Tue, 2004/08/17 - 5:00am

But see -- there you've got it. Nobody wants to do the gruntwork for fun. If you want to have the boring, annoying jobs you're just going to have to pay someone to do this. Commercial distributions tend to maintain older KDE branches for much longer -- for major bugs they'll usually release updates until the end of life for that product (normally a couple of years).

But you don't really think people are going to do that in the free time, do you?


By Scott Wheeler at Tue, 2004/08/17 - 5:00am

I do think you misunderstand the issue, open source is not about doing things for free. But sometimes it's more a natural path economy. When you solve a bug for yourself there is no incentive to exclude others. Just by the fact that persons use software they find bugs and submit reports that help others who try to track them down. It's no "social revolutionary" concept for people doing things for free as opposed to capitalism. Wiping out remaining bugs from a mature release is no big deal.


By Gerd Lange at Tue, 2004/08/17 - 5:00am

In the particular culture of KDE, yes, most of the work is done for free. The boring jobs -- such as backporting fixes to other branches of KDE (older than the last stable branch) is usually only done by people that are working to distribute KDE in some sort of "value added" sort of package where bug fixes for older releases are part of the support arrangement.

And you're wrong about it being no big deal; code is a moving target and often the fix that goes into the current unstable development version is completely different from what would have been needed to fix a release from 2 years ago. It's not a trivial process and it's certainly not much fun.


By Scott Wheeler at Wed, 2004/08/18 - 5:00am

Wait, so you want extended support for obsolete versions, but don't want to pay a distribution that'll give you that support??? You can't have it both ways!


By Rayiner Hashem at Tue, 2004/08/17 - 5:00am

> Well said. For a start, there needs to be more backporting of bugfixes to
> previous KDE branches (not just backporting of highly publicized security
> fixes). e.g. I'm using KDE_3_0_BRANCH and kde-config [...]

> This is not quality. This is not supporting branches about 6 months after
> the initial release. Even MS gives more support.

So I take it you're volunteering to provide better support for older versions?
Great. When can you start?

No, seriously, who should do it? It would need manpower that the KDE team doesn't have. Besides KDE is not a company with paid employees so it would be hard to find someone willing to do that. It's not a very attractive task to backport *large quantities* of fixes to versions with a very low number of users.

And the point would be? Making 3_0_BRANCH into a bad version of 3.3?

What's your reason for sticking with 3.0? It must be a very good reason because KDE has seen performance improvements and bugfixes since then, besides the usual feature enhancements.

If you really need a version that old you might consider paying someone to do the backports for you.

---

OTOH quality control for the released branch and the new release is must, no doubt about that.


By cm at Tue, 2004/08/17 - 5:00am

> And the point would be? Making 3_0_BRANCH into a bad version of 3.3?
Why are there still Linux 2.0 releases? Why doesn't everyone use 2.6?

To minimise change, higher stability (if things were backported).

> What's your reason for sticking with 3.0?
Apart from that, my personal reason is the download and/or compile time is scary.

> It must be a very good reason because KDE has seen performance
> improvements
For sure no. Time a 3.0 startup against 3.3.

> and bugfixes since then,
Which should have been backported :)

> besides the usual feature enhancements.
Agreed.

> If you really need a version that old you might consider paying someone to
> do the backports for you.
That won't work - you need the actual developers doing it IMO.
Else other fixes get swept underneath the carpet (e.g. a structural redesign to fix some bugs can't be backported directly - must hack around it in older branches).

But I admit that most of us are volunteers so I can't do much more than my share. But if KDE is targeting the enterprise, this doesn't sit well.


By w at Tue, 2004/08/17 - 5:00am

>> It must be a very good reason because KDE has seen performance improvements
> For sure no. Time a 3.0 startup against 3.3.

How can you tell "for sure" if you're not knowing/running it? You're just trolling.


By Anonymous at Tue, 2004/08/17 - 5:00am

> > > It must be a very good reason because KDE has seen performance
> > > improvements
> > For sure no. Time a 3.0 startup against 3.3.
>
> > How can you tell "for sure" if you're not knowing/running it?

*sigh*. i'm a developer. i compile and run head (qt-copy, arts, kdelibs & kdebase only) under a different prefix under an ordinary user. i am not in a position to compile 10 other modules i need everyday (kdenetwork, kdegraphics, kdeaddons, kdegames etc.) _and_ under a different prefix (since that's a waste of time recompiling qt-copy, arts, kdelibs & kdebase). a 1998 computer is just too slow for that.

if you're going to claim that i'm not compiling with optimization options as strong as my distro's binary 3.0, i have been developing for kde since 3.0 with the same build options as now with 3.3. so i can tell the difference.

you can't tell me that kde requirements haven't gotten stepper. remember around the time of kde 3.0, ppl suggested moving up from 64mb of ram to 128mb for reasonable performance.

> > You're just trolling.
And you're just grandstanding? And/or just to conclusions too quickly.


By w at Tue, 2004/08/17 - 5:00am

Pages