KDE 4: Understanding the Buzz

With all the excitement surrounding KDE 4 development at the moment people are starting to ask why they have not seen any updates on what KDE 4 will look like. KDE 4 - Understanding the Buzz answers these increasingly common questions by explaining the current status of KDE 4 development and why the exciting work so far is only visible to developers. "Before any new features can be added to KDE and projects like Plasma can get underway, the porting of KDE to Qt 4 has to be completed."

Dot Categories: 

Comments

by minty (not verified)

QT4 on Linux is at best in beta stage, the performace is so slow as to not even be usable.

Unless you link to the qt3 libraries in qt 4 virtually every line of QT in KDE will have to be rewritten.

Might be best to wait, perhaps even fork QT3 to a KQT and do something over time that doesn't break everything

by Aaron J. Seigo (not verified)

> the performace is so slow as to not even be usable

performance is perfectly acceptable here and i'm not even on a particularly powerful laptop.

> Unless you link to the qt3 libraries in qt 4 virtually every line of QT in
> KDE will have to be rewritten

having ported kde code to qt4, i can say that this is hyperbole

> perhaps even fork QT3 to a KQT and do something over time that doesn't break
> everything

this would make 0 sense. there's no need to fork qt3 if we wished to stick with it (we'd just keep using the qt-copy in the 3.5 branch), but that would prevent us from taking advantage of the benefits of qt4 of which there are more than enough to make that a desirable goal.

realize that kde4 is probably ~18 months away so there will be at least a qt 4.1 release before then if not a 4.2.

by James Richard Tyrer (not verified)

> realize that kde4 is probably ~18 months away

Good to see a more realistic estimate.

But, isn't more than a year a long time to go without anythings but bug fix releases?

by Anonymous (not verified)

If you want we can delay KDE 3.5 release nine months after freezing. Would this please you more?

by James Richard Tyrer (not verified)

Very funny. :-D

While I don't see that we need to be in a hurry to release 3.5 -- waiting perhaps an other month for the feature freeze wouldn't hurt. My point is that we should *consider* that we might want to do a 3.6 release -- especially if 4.0 is delayed.

by Anonymous (not verified)

Rather a higher 3.5.x will have a relaxed feature/message freeze.

> My point is that we should *consider* that we might want to do a 3.6 release -- especially if 4.0 is delayed.

A KDE 3.6 release would delay KDE 4.0 even further. Again, we *don't* have the man power for another KDE 3 release. After KDE 3.5, everybody will concentrate on KDE 4.

by James Richard Tyrer (not verified)

> we *don't* have the man power ...

There are many that would tell you that this statement is questionable. More man power (and woman power :-)) is available than core developers think.

> After KDE 3.5, everybody will concentrate on KDE 4.

Perhaps we can put new people on the bug fix releases for 3.5.x then.

After they get all the bugs fixed, they will have to do something while we wait for 4.0. :^)

by superstoned (not verified)

18 months? shit, that's still a long time... KDE 3.5 september, than still more than a year? sucks... at least i hope it won't be delayed much more!

by KDE User (not verified)

Wow, I would say such a long time period might even be suicide for the KDE project. I have bad memories of the KDE 1 to 2 switch. I don't think doing that a second time is a good thing.

KDE 4 should just be a port of KDE 3 to Qt 4 with some small cleanups. If things like Plasma are going to make it take 18 months, then that should be left for KDE 4.1 or KDE 4.2.

by superstoned (not verified)

well, i feel i must agree... altough i'd love to get my hands on a KDE4 with full tenor integration, reworked conqueror, totally different desktop (plasma) and full hardware acceleration, if this would take 18 months, i'd rather have a KDE4 with just some of these :D

by minty (not verified)

I am confussed by your statement that you have ported KDE to QT4 and yet you say it is going to take 18 months. And 18 months is only an estimate. QT 4.0 missed its window by almost a year and it is still a beta product. Whose to say how long it will take to port all of KDE over.

The API did change a lot, even the most basic way to create a widget changed. As it no longer takes an argument for a name. Openoffice and Mozilla were smart not to build upon QT not because QT is a bad product but Trolltech does not understand its customers and their need for stability.

It makes sense for KDE to fork QT now as Trolltech has really changed the game so much. A KQT can be evolved in a reasonable way over time so that not every pice of code has to be rewritten. The only things KDE looses are Windows and Mac compatability that QT offers, but does KDE really care about this.

Eventually KDE will fork a version of QT, as I doubt that Trolltech will be in business as we know it today. I mean here they are in their hey day and they still can't make money, or so I surmize by their need of second round funding.

by Andy Parkins (not verified)

> The API did change a lot, even the most basic way to create a widget changed. As it no longer takes an argument for a name.

The API did change but if you think that is "the most basic way to create a widget" I think you are overstating.

w = new QWidget( parent, "MyWidget" );

Has become:

w = new QWidget( parent );
w->setName( "MyWidget" );

It's not that fundamental a change really is it? It could be done with a sed script. It seems like a perfectly acceptable change to me. Why? Because the code is more readable. Putting settings in a constructor is never nice.

It sounds to me like your complaint is that Trolltech broke the API between qt3 and qt4. Well of course they did - it would have been Qt3.6 (or whatever) if the API were the same. The Qt3 API was stable for a good long time; as I'm sure the Qt4 API will be.

by Anonymous (not verified)

> I am confussed by your statement that you have ported KDE to QT4

He wrote "KDE code", not "KDE" as in complete desktop. Learn to read!

> and yet you say it is going to take 18 months.

Because KDE 4 is more than just porting to Qt 4.

> QT 4.0 missed its window by almost a year and it is still a beta product.

IIRC it was more half a year and it's released already.

> Trolltech does not understand its customers and their need for stability.

You're trolling, you're not even Qt spelling correctly. How long was Qt 3 line the current and how many years will Qt 3 be continued to be maintained? There is not even set an end of maintenance for Qt 3 yet.

> It makes sense for KDE to fork QT now as Trolltech has really changed the game so much.

Changed game? Hu? Made Qt/Win GPL maybe?

> Eventually KDE will fork a version of QT

No.

> I mean here they are in their hey day and they still can't make money, or so I surmize by their need of second round funding.

You don't have a clue about how such stuff works...

by Aaron J. Seigo (not verified)

thanks for saving me the typing =)

by minty (not verified)

The fact that the creation of QWidget changed and no longer supports a name argument is just an example of how much the code did change in rather maininless ways. Setting colors, setting window captions all had cosmetic changes , which as someone stated can be modified with a sed script. Well perhaps it can, I eman lok at the tool QT provices qt3to4. The last time I ran this script which was 4 months ago, over 50 percent of QT3 example applications failed to be converted. So I am not a big believer in Script toools that will magically bring your code up to date.

All I am saying is that Trolltech needs to think of end users in the commercial space. Who unlike academic and government labs do not have time nor money to contantly keep changing their codes to reflect Trolltech fashion of the day.

Although I do not work for Trolltech like some here, I am biased, as I convinced this company I work for to purchase QT. We have some real business concerns about using QT. Why because QT3 is being put out to pasture, and QT 4 from our codingin experiences is still a beta product.

Lets just see how long it takes for KDE, KOFFICE and all the KDE apps to be migrated to QT4. And I hope that when it doees happen QT4 is not made obsolete with QT5.

by ac (not verified)

Please spell check your posts in the future, thanks.

by Dolio (not verified)

Sometimes making an improvement requires breaking backward compatibility.

If you don't like that, then perhaps you should consider getting out of the computer industry (or almost any industry, for that matter).

Would you care to rant about how your ISA SCSI controller won't fit in a PCI-Express slot? How about how nobody makes parts for '85 Caprice Classics anymore?

by Mark Hannessen (not verified)

trolltech is a nice company that writes a superiour toolkit, I love the way they deal with things in general and qt4 is nothing more then a needed evolution of there product. and what they did makes perfect sense.

by Michael O'Keefe (not verified)

>> Trolltech does not understand its customers and their need for stability.

>You're trolling, you're not even Qt spelling correctly.

No doubt. And if you check >mintys< e-mail address, aol user. Traditional evidence of a clueless internet user, probably singed up via Windows ME.

As far as stability, I've been impressed with the ease of porting code between different versions of Qt. It is quite clear Trolltech pays much attention to their customers, especially when compared with tools like Microsoft VB (end of the road for their users, and they are not happy), and the numerous, annoying times my Java code has broken with new releases of the JDK. Ask anyone about trials and tribulations with XML.

-Mike

by mhn (not verified)

This buzz explanation is itself generating lots of buzz (finally, as someone said). Keep it up! :) What good would be a great DE without people fining out about it?

by hannes hauswedell (not verified)

i really hope they did something to improve speed of listviews in qt4...
searching, filtering oder otherwise working on listviews is soooo slow!
how come the qtable widget is advertised with being able to scale a square milllion cells without problems, but filtering a listview with >10k items freezes the app for a second?