KDE Frameworks 5.0 In Development

An anonymous reader writes "In addition to bringing up the plans for KDE on Wayland, Aaron Seigo just announced at the 2011 Desktop Summit that the KDE 5.0 Frameworks libraries are being planned for development. This central code will be developed in parallel to future KDE SC 4.x releases until it is ready, as to not cause another KDE 4.0 mistake. When the code is ready, key applications will be ported to the new interfaces." (There's another article at IT World.)

Comments

Feels early

[Windwraith]

Feels actually very very early. After 4.6 being almost identical to 4.5 regarding workflow, bugs left unpatched, and all the little issues KDE4 still has, moving to 5?
Is there a new, breaking release of Qt to catch up with like with KDE4?

Re:Feels early

[KTheorem]

That's it exactly. From what I have read Qt 5 will not have the backwards compatibility for Qt 3 that Qt 4 does. Too many KDE applications still use those compatibility features and so they need to rewrite it so that it no longer does to be able to use Qt 5. Since that will break programs that rely on those compatibility features it is deserving of a version change.

Re:Feels early

[Anonymous Coward]

All brainpower will be on KDE SC 4.

This story is about breaking kde's current libraries into smaller modular pieces, it is not about:
1. KDE SC 5
2. About developing new libraries

Ivan
KDE developer

Re:Feels early

[Enderandrew]

I think the issue is that Trolltech/Nokia is moving past Qt 4 series into Qt 5.

KDE has maintained that kdelibs can't break binary compatibility between major versions. If there is a significant change with Qt, and thusly major changes for kdelibs, then they have major release number.

That doesn't mean a massive rewrite and change necessarily like we saw with KDE 4.

Re:Feels early

[JabberWokky]

You didnt misunderstand, the phoronoix "article" is misleading. ... as often .;.

Seems pretty straightforward to me:

"Application development will not be pausing as we do this: releases every six months of application improvements will continue based on the 4.x codebase. When Frameworks gets to the point where it is ready for serious banging on, then we will start repurposing our highlight applications to the new codebase," Seigo wrote. "We don't want application development to be held up by the library development, and we don't want the library development to create much, if any, need for 'porting' application code. We want 'just recompile and test' to be the common case, with whatever changes do become necessary to be of the simple and even automatable sort.

"If this sounds rather different from how we approached 4.0, that's because it is. The requirements, needs and context for this release are utterly different. We're after evolutionary improvement and broadening our developer ecosystem, and our plans therefore need to, and in our opinion do, reflect that," Seigo added.

KDE 5, then, will not be the paradigm-shifting platform that happened with KDE 4, a move that caused many Linux desktop fans to throw up their hands and complain that KDE 4 should never have been released in its initial state. Criticism of the KDE 4 desktop still exists (this is, after all, the Linux fanbase we're talking about), but it has moved well past the "immature" and "too much change" arguments that once plagued the inboxes of KDE developers.

Re:Feels early

[suy]

Qt 5 will be binary incompatible with Qt 4 because they will reorder libraries and modularize them. KDE will do the same with their Frameworks. However...

Qt 4 will require very few changes in the source code. Yes, they will drop the Qt3 support, but very few KDE apps use the Qt3 classes support (of the apps that I use, only JuK to my knowledge).

KDE 5 and Qt 5 will be mostly a major version bump because of the binary incompatibily that can't happen without a change in the major number. Some applications might not need any source code changes, or very few.

Time for a new API

[Chemisor]

It is clearly time for yet another major API change. People have been writing way too many applications for KDE 4 and this must not be allowed to continue! Having millions of apps is such a waste of effort - we're the Linux Desktop, for heaven's sake, not some lame appstore. Surely everyone can agree that having KDE developers write all the key apps is the way to go. We are the most experienced and the most knowledgeable in using the KDE API, and dammit, WHY WON'T YOU LET US HELP YOU?

I hope they make it like 3.5!

[Lord Lode]

It would be awesome if 5.0 were more like 3.5 again (its behaviour and settings), but with the modern graphics features of 4.0 :)

Re:I hope they make it like 3.5!

[zixxt]

I just don't get the love for the 3.5 series. I liked 3.5.x and used almost everyday, but 4.x series is sooo much better in terms of speed, looks and apps. The KDE 4 series is IMHO right behind the Mac OS X in terms of goodness.

Re:I hope they make it like 3.5!

[marcosdumay]

Well, at 3.5 the plasma environment didn't segfault one or twice when I start my laptop (sometimes locking the X). Also, its applications did have a more sane reaction to keyboard orders (like, if it opens a window, let it have the focus). Also, I could have more than one KDE session without windows appearing clamming that it couldn't lock a file and closing the application I'm using (and if I don't press the "proceed" button, I can keep using the app, no problem, except for the window that stay above it). The possibility of having more than one session open at the same time was the dealbreaker that let me out of Gnome at KDE3/Gnome2 time (before that I didn't give a dam about what DE I was using).

But ok, that second problem appears on a kind of interaction that simply didn't exist at the 3.5 time.

Re:I hope they make it like 3.5!

[qbast]

Let's see:
1) Speed - open dolphin window, start resizing. Watch the contents trying to keep up and failing, each part of window moving at different speed
2) In dolphin go to home directory. Wait 15s until *anything* shows up then 5s more until dolphin is actually usable
3) Memory usage: in kde3 kmail took about 50-80MB ram. In kde4 it is several hundred MB for kmail itself, next several hundred for mysql, akonadi server and pop3 resource. After that I switched to Thunderbird.
4) Even more memory usage: run kmail with nepomuk enabled. Now you have 2 full database servers running (mysql and virtuoso). OMG.
5) Others mentioned it already: zip/tar.gz integration. In KDE3 it rocked, now you have to use crappy Ark.
6) Try using KDE4 application remotely. Works much slower than stuff from KDE3.

KDE4 was supposed to be revolutionary for user. It introduced several "pillars" that were supposed to make it years ahead of competition. Too bad almost all of them failed miserably.
1) Nepomuk - so much promise, so little delivered. Semantic engine, associations, new ways to combine data from different sources - damn, this actually had potential to be a revolution. And what users actually got from all this? Full text search, tagging and rating - none of them really integrated with UI. And in return we sacrifice performance and memory usage.
2) Akonadi - another middleware/unification effort. Was supposed to rewrite everything PIM-related in clean way, deliver high performance, add compatibility with non-KDE world. And what we got? Memory usage of KMail+Akonadi now goes to gigabyte or more if use are unlucky. Performance is way worse and getting GNOME guys on board was never anything more than pipe dream (come one, gnome devs using piece of software that came from KDE? Never happened, never will happen).
3) Plasma. Well, Aaron Seigo is good when it comes to marketing buzzwords - paradigm shifts, new way of using UI, yadda, yadda. And what user got is way to put widgets on desktop. Woohoo, I could do the same in KDE3 with superkaramba.
4) Solid - this one actually turned out ok. Stays out of the way, does its job.

To be clear - I still use KDE4 as my main desktop. Gnome was, is and will be crap. KDE3 is long unsupported. However I am moving piece by piece to non-KDE technologies - replaced Konqueror (after years of using it as my main browser) with Firefox, KMail with Thunderbird.

Re:I hope they make it like 3.5!

[stilborne]

* you can rubber band in the file manager
* weather works just fine with European cities. i live in one, so i know.
* graphics glitches are usually driver related, but we've also fixed a lot of issues (small and large) in the least couple of years
* creating a launcher -> drag it from the file manager, the launcher, from a search in krunner, ... yeah, not hard.

i realize that reality often does not come into play much when people create and then post again and again these kinds of lists. really it comes down to your last point, doesn't it: you have a highly personal preference for 3.5 (which is fine) and you work backwards from that to justify it. your points, however, are outdated or were never correct.

but that's alright. one can't please everyone all the time, of course. we do have many users who are quite happy, and both sides of that (number of users and average satisfaction) appears to be growing with each release. perhaps not amazingly, it is those people, not those who generate lists for re-posting on every article they can find, who get developer focus.

Re:What KDE 4.0 "mistake"?

[Enderandrew]

It depends on perception.

I read dot.kde.org regularly, and Planet KDE. Every single KDE dev was quite clear that KDE 4.0 wasn't for everyone on day one, and it wouldn't have feature parity with KDE 3.5 on day one.

Yet every single tech blogger says they were lied to in this massive fiasco that KDE 4 would be perfect on day one. Where exactly was that statement? I think the problem is that a few distros were pushing KDE 4 as a default desktop before it was fully ready for primetime, and Kubuntu in particular was shipping really broken packages.

If you got a KDE 4 desktop before you personally wanted it, or if you had a buggy desktop, then KDE 4.0 was a disaster and the devs lied, even if that really isn't the case. So Aaron is justified in saying 4.0 wasn't a disaster from a developer standpoint. They needed to get a base release out there for people to test, and for developers to develop for. That didn't mean every user would be happy with it on day one. But since people did have bad experiences, you're not going to convince any of those users that it wasn't some unmitigated disaster.

Oddly enough, the Gnome devs have sworn that one of their biggest goals of Gnome 3.0 was to avoid the KDE 4.0 disaster, and they wouldn't push a massive change out the door on day one. And yet you can argue that the Gnome 3 shell is a bigger change, and a bigger removal of features than the KDE 4 launch. And with KDE, most of those features returned in time. They just hadn't been ported over yet. Gnome 3's shell removes many basic features as a fundamental design decision.

In the end, users should make informed decisions about what desktop works best for them be it KDE 4, Gnome 3, Unity, XFCE, etc.

Re:KDE can suck it

[halivar]

Gnome's mom always posts AC.

Re:Yeah.. key applications

[obarthelemy]

2011 will be the year of the desktops on linux !