A Sneak Preview of KDE 4

An anonymous reader writes "In recent times, a lot of discussion has been generated about the state of KDE version 4.0 and as Linux users we are ever inquisitive about what the final user experience is going to be. This article throws light on some of the features that we can look forward to when KDE 4.0 is finally released some time this year. The article indicates that the most exciting fact about KDE 4.0 is going to be that it is developed using the Qt 4.0 library. This is significant because Qt 4.0 is released under a GPL license even for non-Unix platforms. So this clears the ideological path for KDE 4.0 to be ported to Windows and other non-Unix/X11 platforms."

Comments

Memory

[Anonymous Coward]

"For instance, Qt 4 is designed to save lots of memory and will perform faster."

They need to work more on that cause thats the reason why I'm not using KDE. I like the UI but KDE is just to bloated so I use Gnome instead, even though I hate most of Gnome's UI.

Re:Memory

[Frekko]

This http://ktown.kde.org/~seli/memory/desktop_benchmar k.html article from 2006 shows you how much memory Gnome/KDE use. Even though it is written by a KDE member I can't see why he should have messed with the numbers. As you can see KDE actually uses a bit less (not much though) memory than Gnome.

Re:Memory

[jcupitt65]

Here's an update written by a GNOME person:

http://spooky-possum.org/cgi-bin/pyblosxom.cgi/kde vsgnome.html

tldr: they have (essentially) the same memory requirements.

Re:Memory

[swillden]

And one of the reasons to use C instead of C++ (besides the "easier to make bindings" reason) wasn't that C++ was more "heavyweight"?

Only among those who don't know what they're talking about. There's nothing in C++ that's inherently "heavyweight". Object-Oriented code often tends to be a little heavier because the tools make it so easy to generalize, and OO developers often end up writing code that may be needed someday, but isn't now. C programmers are less prone to that particular error.

However, the flip side of that focus on generality and reuse is that if it's used well it can actually reduce the code size and memory footprint, as well as making development faster and easier. You can, of course, obtain the same level of code reuse in C, but it's more effort to write and much harder to debug, so it doesn't tend to happen.

My impression of the KDE and GNOME situation is that, at present, KDE provides a richer set of libraries with more tools to make application developers' lives easier than GNOME does, and KDE does this with about the same, and perhaps a bit less memory usage. And this is *because* of C++, not in spite of it.

C++ is a bitch to learn, but once you do, it's a very pleasant language to work with, in every respect: expressiveness, power, efficiency, compactness. IMO, C++'s only practical failing is that C++ programmers are too hard to find/create (back to that learning curve). Note that I'm not saying the language can't be improved upon -- I'm comparing it to the other widely-used options, not to some theoretical ideal.

Re:Memory

[NekoXP]

Not only are they rewriting stuff to reduce footprint they're using more and more system components that everyone has (DBUS) rather than KDE-specific things (DCOP for example). Qt4 definitely uses a lot less memory as a whole than Qt3 but it's not backwards compatible.

What you will see is KDE4 by default using a huge glob less of memory, but if you run an old KDE3 or Qt3 app, suddenly memory usage will kind of go up when the compatibility libraries load.. disk usage will go up too because of them. But in most systems, 90% of the time the CPU is fairly idle and memory usage is the most important performance factor; not just memory-limited systems, on huge multi-GB desktops too.

What I really want to see is KDE4 running on Qt4 directly on the Linux framebuffer; get rid of X. Then something like MythTV running on top of it; bringing requirements down by removing some of the extraneous cruft (X no longer has magic mouse and keyboard drivers since the USB HID system does most of the work, would be one example) is a good goal too and KDE4 is also doing some of that.

I'm not sure what direction GNOME is taking, but at least there is a lot less ability to do so with GTK; they pride compatibility without compatibility libraries, and new functionality comes with new applications and rewrites of applications which never made the grade (Ubuntu Edgy had a bunch of them) - it seems to be a more pronounced, feature-rich development cycle with less chances to sit down and optimize something old. Both environments seem to be focussing on simply PROVIDING user experience than optimizing it. However KDE has a lot more baggage; components like the browser, office suite are all part of the KDE offering, which GNOME doesn't have an encumberance on. Optimizing KDE gives more results for less work. Optimizing GNOME seems harder to justify considering very few things will benefit but the toolkit and desktop itself. Maybe I'm wrong though... :D

great

[mrsev]

Ooooh... Kreversi, KMajhong.... both essential components of my desktop experience. The article is a little thin to say the least.

Re:great

[CastrTroy]

I agree with you about Tetris. The best tetris I've been able to find for the PC is the Nintendo tetris, run on a emulator. It's kind of sad that nobody makes a good tetris game for the PC. It's such a simple game, you'd think it would be easy to program it well.

Performance

[bcmm]

It should also be pointed out that the port to QT is expected to very noticeably improve performance.

When was the last time a new version of Microsoft Windows came out with a faster user interface?

Re:Performance

[eclectro]

When was the last time a new version of Microsoft Windows came out with a faster user interface?

When you bought the new computer?

Re:Performance

[urbanradar]

It should also be pointed out that the port to QT is expected to very noticeably improve performance.

Maybe it's a typo, but just to clarify: KDE is already based on QT. It's just that KDE 4 will be using QT 4, whereas the current KDE uses QT 3.

Re:Performance

[zlogic]

Uhm, Vista runs the GUI faster that WinXP if you have a decent GPU. Just like QT4's Arthur, it uses hardware to do rendering, which is great because even when the system is under heavy load you can still do basic tasks like moving and minimizing windows. The system feels much more responsive.

Re:Performance

[diegocgteleline.es]

It's not about "using the GPU", QT4 is just much faster and eats less resources.

"When Qt designer was ported to Qt 4.0 - only the neccesary changes to make it compile - the libqt size decreased by 5%, Designer num relocs went down by 30%, mallocs use by 51%, and memory use by 15%. The measured Designer startup time went down by 18%"

Now try to imagine the savings for the whole KDE desktop

Re:Performance

[leuffi]

If you upgrade your computer to Core 2 Extreme QX6700, you can enter the Olympics! w

Re:Performance

[swillden]

Software evolves to take advantage of it.

Software that takes advantage of it is fine. Software that wastes it to no advantage is not.

From dot.kde.org

[strider44]

I monitor dot.kde.org pretty closely and there's a few things notable here. Firstly if you look at KDE at the moment it doesn't look much different to KDE 3.x. This is because the frameworks are currently being finished and polished - the interface will be the *last* thing to be finalised - remember guys tip of the iceberg - there's a whole lot more code that you don't see than you do see.

Also, with this article specifically, a few of the graphics are temporary, most notably the background that's pretty obvious in ksysguard. Yes it's horrible for that app, no it won't be there in the finished version. It's a temporary background being used in several apps at the moment for a placeholder.

Also, the start menu isn't finalised yet from anything I've heard, that's the start menu designed specifically for Suse - it's been on Slashdot before.

KDE looks like it will be coming together quite quickly and quite soon. Several major components are pretty much complete and are being polished now. Looks like pretty fun stuff - don't believe anyone who says it's vapourware.

Re:From dot.kde.org

[strider44]

Also, a similar KDE article is at http://dot.kde.org/1167723426/

Read the comments there as well for some interesting info.

Re:From dot.kde.org

[Wanderer2]

Also, the start menu isn't finalised yet from anything I've heard, that's the start menu designed specifically for Suse - it's been on Slashdot before.

Glad you mentioned that before I posted! I was about to rant about how much I hate the look of that start menu. It looks too similar to the Windows XP, expanding-to-fill-the-screen-with-icons-all-over-t he-place one which drives me mad. That said, I do occasionally have trouble finding seldom-used stuff within my KDE start menu (is $APP under Settings, Utilities or System?), so I know my current set-up isn't brilliant either.

Less of the kitchen sink would make KDE better

[DrXym]

KDE is a very slick desktop, but it doesn't seem to know when "less is more". The control center is probably the worst, most confusing configuration application of any desktop I've ever seen simply because the options that 99% need to get at regularly are mixed in with options that only 1% / nobody ever needs to touch. Then you have various K apps such as Konq or KMail where you might have up to SIX different preferences menu items to choose from to configure the app.

I wish they'd follow GNOME or Firefox and realise that overloading the senses with tabs, buttons and checkboxes does not make for a pleasant desktop experience. I'd be happy if all KDE 4 consisted of was a rationalisation of the menus and prefs to slash out most of the crap, or at least move it into an advanced mode where only masochists could see it.

Re:Less of the kitchen sink would make KDE better

[strider44]

It's funny that that's the reason why I detest Gnome - for some reason they got the idea that removing all the options that only 5% of users use is a good idea. Of course all the other options are used by a different 5% each time so in the end you've got the majority of users upset because the option they want has been removed. Note: Put to the side with an unknown and unguessable key combination counts as removed.

But that's OK because Gnome isn't for me.

Please, Gnome is a slim pick up and go desktop for new users, KDE is a customisable and flexible desktop for power, business or techie users. I like it this way, it gives everyone a desktop that they are comfortable with. As a techie, I want KDE to stay the way it is, please don't try to change it to something it is not.

Re:Less of the kitchen sink would make KDE better

[DrXym]

Please, Gnome is a slim pick up and go desktop for new users, KDE is a customisable and flexible desktop for power, business or techie users. I like it this way, it gives everyone a desktop that they are comfortable with. As a techie, I want KDE to stay the way it is, please don't try to change it to something it is not.

No it doesn't give everyone a desktop they're comfortable with. If you put twice as many options in a user's face than they would reasonably expect or ever require they are going to get confused. They'll be scared of even opening prefs or menus for fear of trashing their app or desktop. As I said, if somebody in the minority absolutely needs all those options, create some tool for them to get at them or implement an advanced mode that shows them. For example, I can type about:config into Firefox and fiddle with all the options, without inflicting them on regular users. And Windows has TweakUI and regedit.

There is a very good reason that the likes of Apple, and to a lesser extent Windows & GNOME hide options - it makes the desktop more accessible, more productive and easier to use for everybody. Expecting the vast majority to wade through crap makes for a terrible user experience. Regardless of what advantages KDE might offer over GNOME in other ways I know which side my money would be on in a usability lab.

Re:Less of the kitchen sink would make KDE better

[Godji]

I understand your concern, but on the other hand, KDE is right now the only desktop which allows me to customize EVERYTHING I could ever want to customize WITHOUT going into config text files or the source code itself. This is exactly what I like most about it!

Have you ever had the feeling that "this program is awesome, but there's this really annoying tiny thing I wish I could easily change"? I had a couple of these with GNOME last time I tried it, and I've never had this with KDE 3.x.

GNOME is already a good, clean desktop. KDE is the good, customizable desktop. It's better to have one of each, rather than two of the same, don't you think?

Re:Less of the kitchen sink would make KDE better

[Octorian]

Windows doesn't hide all the advanced options. Many are still easily accessible from the GUI. MacOSX, however, hides them so deep that you need to drop to the command like to tweak them. (Heck, Apple's complete hiding/elimination of options is why I ditch Apple's packaged apps whenever I find a suitable alternative, even if I used MacOSX every day at work)

Likewise, GNOME hides the options so deep as well, that only a poweruser spending the day on Google is ever going to even figure out how to get to them.

At least KDE (and Windows) put the options where you can find them using just the normal flow of the GUI.

This whole "assume the user is a drooling moron or an ubergeek, with *nothing* in-between" really puts off a lot of "competent" Windows users.

Re:Less of the kitchen sink would make KDE better

[Chris_Keene]

"Please, Gnome is a slim pick up and go desktop for new users, KDE is a customisable and flexible desktop for power, business or techie users."

Disagree.

I use Gnome because I have a million and one things to do and so long as the interface isn't annoying, looks ok and doesn't get in the way, then it's good for me.

Don't get me wrong, I'm a power, business and techie user. When KDE 1 came out I spent loads of happy minutes changing every setting just to how i liked it on my home PC. Partly because I could and partly because I found the default kde setup annoying.

I now use Ubuntu (at work) and have never felt the urge to change a single option. Now, the techie in me wants to do cool things at a PC, not change how the taskbar looks.

Re:Less of the kitchen sink would make KDE better

[Mjlner]

"I wish they'd follow GNOME or Firefox and realise that overloading the senses with tabs, buttons and checkboxes does not make for a pleasant desktop experience."

Hear hear!

You're so right! I wish the KDE team would realise that a pleasant desktop experience involves editing .gtkrc-2.0 by hand and adding stuff like

binding "gaim-bindings" {
bind "Return" { "insert-at-cursor" ("\n") }
bind "<ctrl>Return" { "message_send" () }
}
widget "*gaim_gtkconv_entry" binding "gaim-bindings"

This is Clearly preferable and more easily understandable compared to having to click a check box, as you had to do in the bad old days of Gaim.

Re:Less of the kitchen sink would make KDE better

[hamfactorial]

If you're sore about the CTRL+Enter to send option being removed from the GAIM preferences, perhaps you should send your grievances to the GAIM developers. They're in a much better position to consider your issue than the GNOME developers and the general unwashed slashdot masses.

Re:Still not there yet..

[rucs_hack]

Funny how everyone says 'SUSE' these days to rhyme with 'traitorous scumbags' :-)

I'm a new convert to KDE, after years of predominantly fluxbox usage, with the odd dabble into Gnome. This is mainly because I principally used Linux over VNC or ssh, so KDE was out of the question, too slow over the network.

Now I have the novelty of a fast local Linux box, and decided to try out these fancy Graphical Desktops a bit more. The new Gnome is good, but I must say I am becoming more and more impressed with KDE as the days go on. I still like my fluxbox though, simplicity does have it's appeal sometimes. Can KDE ever be that fast though, I doubt it. Not that I care much about load times on KDE, 99% of my computer usage is text editors and the console. Those are two things that run fast on any system.

KDE on windows? Sounds interesting. Windows is just a games environment or dumb terminal into my linux cluster for me normally, I'd love to have KDE on XP. A fast KDE frontend for Vista might actually make me consider buying that heap.

KDE on MacOSX

[hritcu]

I can hardly wait to be able to run KDE apps on MacOSX without having to start an X server, and have proper copy-paste support, correct window stacking and native look an feel. There are KDE applications like Krusader for which there is hardly any alternative on the Mac.

KDE vs. Gnome

[Psychotria]

At the risk of being labelled a troll, I have a few obversations to make. I yearn to return to Gnome (I made the switch from Gnome--which I'd been using for 3 years--to KDE about a year ago. I'm not sure if it's a "feature" of Gnome, but when Gnome apps (at least on my systems) fail, they don't even give a reasonable error message. This may be a design feature, to make it "easier", but, in fact, makes things stupidly difficult. If something fails, then I want to know WHY (at least give me the option of more detailed error messages). KDE is consistent. Gnome isn't (yet). 3 years ago, I would laugh at KDE users, because I knew that "Gnome was best". These days I take a more pragmatic view. Ideoligally, Gnome may be better. In practice, KDE takes the cake.

Re:KDE vs. Gnome

[kfg]

This may be a design feature, to make it "easier", but, in fact, makes things stupidly difficult.

Troll or not I think you have just pegged the perfect Gnome slogan:

"So easy it's stupidly difficult."

KFG

Re:KDE vs. Gnome

[ebichete]

You know what would be really cool? When an application crashes:

- the system collects a stacktrace of the process and the libraries (including version numbers) it was using
- compiles this with extra possible relevant info about your state at time of crash
- (after asking your permission) connects to an online database of crashes and compares your
stacktrace to the database, looking up the debug symbols automatically
- if your stacktrace matches a previously known crash it shows you the bug entry for that crash
- if your crash is unique, it allows you to file a new bug for the developers to look at

That would really be a step forward in improving the stability of free desktop environments.

Maybe someday someone will do this

Re:Kstill Ktoo Kmany Koptions

[kfg]

Is it really so hard to strike some middle ground between no options and so damned many you can't find the one you're looking for?

Yes.

KFG

Re:You know that...

[strider44]

Only when it means Amarok on Windows and Macs. That's a good feature of KDE 4.

NOOOOOO.......

[advocate_one]

Plasma will provide the next generation desktop experience in KDE 4.0. It is planned to integrate three separate applications namely the Kicker (Panel), KDesktop and Super Karamba (Widgets) into a single application. And the surprise of all things is that it will be possible to run the beautiful Dashboard widgets of Mac OSX in KDE 4.0.

I like things as they are with separate applications. If Kicker hiccups and falls over I can relaunch Kicker, if Super Karamba falls over, then I can simply restart Super Karamba, if the desktop falls over then I can restart the desktop... if the "all in one app" Plasma falls over, than what??? do I have to restart KDE? I don't want flaky Super Karamba widgets threatening the entire desktop... and I only want to run Super Karamba if I want to, not by default...

Re:NOOOOOO.......

[N7DR]

Super Karamba gets taken out by buggy widgets.

1. Right. It is also true, as you say, that it "can run for days on end"... but even running a simple bug-free widget set, it rarely runs for a week without spontaneously crashing.

2. The threading model in SK is fundamentally flawed. All widgets halt while one is updating. So if for some reason one takes a long time (network connection is down, it performs a heavy calculation, etc.) the entire SK portion of the desktop simply freezes for a length of time.

3. There is a fundamental design flaw in any system that is designed to contain objects, but which can be taken down by a buggy object.

I love the idea of SK. And I run it all the time. Right now I have 11 themes (widgets) on my desktop simply because over the course of the past few years I have found desktop widgets to so useful. But I sure hope that's it's vastly improved in KDE 4. Right now, I get embarrassed when I talk to OSX or Windows users when it comes to the subject of desktop widgets.

Re:Still not there yet..

[Aladrin]

If by KDE you mean 'KDE, all the apps hosted on the KDE site, and all the features of KDE like the FISH protocol' then I am -so- with you. If apps like K3B would run also, nothing could stop me from buying it.

But if you mean just the window manager and such, and not Quanta/Konqueror/Konsole... I'd have to pass. KDE is useful to me. It's not about looks.

Some criticism of gnome mostly past

[dbIII]

I hate gnome due to the cretinous idea of implementing an obfiscated MS Windows style registry on what was at the time a non-portable linux environment - and having one of these registries per user spread over multiple files named after the three stooges and others (not kidding) in a form where you cannot export the settings to another user let alone another computer. This has been fixed to a degree, not everything needs gconf anymore and with Sabayon many of the settings can be exported. However it remains an environment where a user can render their laptop unusable via a screen resolution applet which sets stuff in the weird registry instead of in the X windows configuration files where it belongs. Try fixing that over a bad mobile phone link to a remote area some time. Cretinous behaviour like making things executable without permission also occured at one time, but I believe was fixed rapidly.

Fortunately the people that wanted a version of MS Windows that they wrote themselves running on linux (only) but not understanding the features of the platform have moved on - leaving us with two fairly decent environments with just a few remaining flaws.

Kool!

[glavenoid]

KNow,if Konly Kthey Kwould Kstop Kalling Keverything KSomething or Kother!

Re:Kool!

[Stevyn]

I GNOw what you mean

Re:Kool!

[WindBourne]

geah, gthat gis gsuch ga gpain.
MSPerhaps, MSthat MSshould MSstop MSeverywhere.
.netof .netcourse, .netI .netbe .netwrong.
ior imaybe iyour igrip is inothing.

Re:Hmm , let me guess...

[ErroneousBee]

Stop kidding yourself that KDE is anything other than a windows rip off.

Actually its a CDE ripoff.

CDE predates win95, and was based on the many desktop WIMP environments around in the late 1980s, such as HPs VUE.

A lot of the things you imagine are Windows interface paradigms are actually basic HCI stuff (Fitts law, Roman language left-right convention, and whatnot) that pretty much dictate colour schemes, icon size, icon behaviour, left to right conventions, etc.

The only thing I can think of that is a Windows thing is the position of a main menu button in the bottom left, its easier to mouse to the top of the screen than to the bottom because of the way the muscles in the hand/arm work. In truth the KDE button can be located anywhere, its just the default themes that just happen to position it there, cos that's where most computer users look to find a central control.

What is thin about it?

[WindBourne]

It shows some graphical pics of games that have been converted to SVG (nice to say the least). Then in the article, it talks about the various projects that are working on core libs. Once those are fleshed out, then more apps will come into focus. I would say that this is actually a pretty good preview of very unsettled work. As to the desktop, well, there will be more.

Re:Windows port ?

[mrchaotica]

Perhaps you're unaware of the fact that QT, the API that KDE is based on, has always been cross-platform. The only thing that ever stopped KDE from running on Windows or Mac OS before was licensing -- it wasn't GPL'd on platforms other than unix/X11 until QT 4.

Pardon my ignorance...

[camperdave]

Why are they McNamed (I'm coining a new word. McNamed: Named in such a way as to be associated with an organization, entity, or product) anyways? Aren't applications written to run on any desktop. I seem to be running gnome, because there is an "about gnome" item on my system menu. However I have several "K" applications, such as kpovmodeller, that run just fine. Clearly they don't rely on KDE, so they must be written to some generic desktop standard. So, why the mcnaming? (They can't all be written by the same people who are writing the desktop software, can they?)

GTD Anyone?

[soloport]

Agreed. Hard to understand what people are even talking about when it comes to "bloat". Gnome, KDE, Windows -- with today's hardware it's not like one has to wait for minutes for an application to load any more.

I use KDE for the sheer convenience and ease of use. Windows seems like it's virtually stood still in time for the last four or five years. KDE has far surpassed it, ease-of-use-wise. Gnome is still such a joke. I don't get it. How is it that Firefox, Thunderbird (at least on Linux) and other packages have to emulate Gnome when it comes to: finding files (GIMP and Firefox try to be so Gnome-like -- sucks!), the whole "Would you like to do this? No? Yes?" anti-natural-language (but oh-so geek-orthodox) OK / Cancel thing. Why do so many distros (Red Hat, Ubuntu) have Gnome as the default? Makes no sense.

I'll trade a little "bloat" for "getting things done" any time.
/rant

Re:GTD Anyone?

[soloport]

Guess what I'm saying is, the whole "bloat" argument used to be so pro-Gnome. Today, it's just a tired, out-dated argument. Can we at least get a better one going? ;-) I think the Gnome as distro-default makes sense from a political point of view. It's probably the case that more Gnome developers know Red Hat, GIMP, Firefox developers than do KDE devs. But from an end-user's perspective, I still can't understand Gnome's popularity.

It's like the Motorola 68020 architecture and Assembly instruction set vs Intel's convoluted silicon / Assembly instruction set. Why, oh why did the Intel processor have to become the most popular?! Well, back in the day, when board layout and Assembly code mattered, it mattered, OK? ;-)

It's rather lonely in here!

[LordPhantom]

Am I the only one that wants a WM, not a Desktop? I'm still using a rather heavily-customized version of E16, and anxiously awaiting E17. Sure, I use a bunch of stuff from the KDE project as drivers (arts anyone?), but I would love to see more options built into the "eye candy overhead" of KDE (i.e. turning off components, etc) to increase perfomance.

That said, the main reason I avoid KDE and GNOME is performance - most of the stuff it does is just overhead I don't need (and why does KDE -not- use the xscreensaver interface instead of their rather....useless wrapper?!!), but if they were able to improve the reliability of KDE, and make it possible to lower it's footprint to something that is, say, just -slightly- more than your average WMW I might consider it.

Re:Hmm , let me guess...

[dbIII]

None of those features? It's OK if you have never seen CDE - but pretending to know about it and getting it wrong on all counts just to win an argument by repeating garbage is annoying and is not OK even if a lot of people do it.

It is true there is nothing with "start" acutally written on it even if the menu comes up from the bottom from each panel button - but many of the things you are talking about here were even on the first mac with a few small differences.

Remember - KDE gets criticism for looking different enough from MS Windows to confuse people, and a lot of this behaviour comes from CDE.