Friday, June 13, 2008

Desktops in trouble

The main reason I switched to Linux in 2000 was the availability of a viable desktop, in my case KDE 1.1. The Linux world was very different back then and I was even forced to install proprietary tools in order to have all the functionality I needed. Nowadays we have Open Office, MPlayer, VirtualBox and apart from some Wine emulated stuff there isn't a proprietary program in my entire system. We are inclined to take all those things for granted. The next version of any program is bound to be better than the previous one, so why worry.

Well, there are some disturbing developments and they are happening in the key components of our systems: the desktop. KDE has spawned a new release, KDE 4.x, and although it looked promising at first, KDE is in trouble. People are not only complaining about its instability (which is not a good thing in itself) but also about the direction KDE is taking. It is a change of paradigm. KDE has always been what you wanted it to be. You could install it as is and just use it or tweak it until you were happy with it. Just about everything was configurable and every possible feature was available. That was what the KDE audience liked about KDE.

But the KDE team has taken another path by banning the icons from the desktop, claiming that all that clutter wasn't neat. The folder view was much more powerful, they stated. I won't argue with that. Maybe that one day the KDE audience will learn how to use them properly and won't even blink when that code is taken from the repository. But I don't think this was the proper time to do that.

The change of a major version number has always been a big deal with KDE. I remember that I continued to use my KDE 1.x long before I finally switched to KDE 2.x, simply because it wasn't stable enough for production purposes. The switch itself wasn't a big deal. Ok, theming was much more difficult than it was back in the KDE 1.x days, but as far as basic functionality was concerned it was pretty much the same thing. Although reports are mixed I don't think that KDE 4.x is production quality right now. Most major distros agree with me: KDE 3.5.x is still an option. Since my hardware is slowly starting to fail I'm afraid I will face that decision as well in the near future. And I've already decided: I'll be using KDE 3.5.x.

If you want to lure your users to a new version the best way to do it is to make sure that everything is the same, but better. The KDE developers have failed to do this and I think that may have been the worst development in the entire history of the project. It's also a breach with the previous KDE philosophy: you decide what is good for you; we won't. Instead of letting the user decide which desktop paradigm he prefers, the KDE developers decide for him. That's a philosophy that sounds familiar. The Gnome boys have been saying that for years. You don't want to tell me that in the end they were right, do you?

Gnome is in trouble as well but for different reasons. First of all, there is the Mono controversy. I could dedicate a blog entry itself on that, but I won't. But it is a major problem. It is no secret that Gnome is closely tied to Richard Stallman and his FSF, so when Stallman himself admits Gnome is in trouble and some parts of it might need to be completely rewritten I guess we have a real issue here. But that is not all. Some people think Gnome has become a dead project, because it has ceased to be "exciting and innovative". Ironically, what is their example? Yup, you guessed right. KDE 4.x.
Now, if the revolutionary features in KDE 4 do not improve productivity and ease of use, then I don't know what else will. Plasma is simply changing the way we perceive a desktop, and I think for the better - the folder view is just one things that comes to my mind, but there are others, like the desktop grid and such.

It is a strange cross-over if you come to think of it. KDE developers are starting to treat their users like mindless children just like Gnome has all these years and Gnome developers are looking at the direction KDE is taking. Is this the prelude to a merger like happened on the 3D desktop front with Beryl and Compiz? A flashy, innovative desktop for dummies?

I know there is an end to the KDE 3.5.x branch and when KDE 4.x hasn't proven itself by then or - even worse - slowly slipped into oblivion for the lack of people using it, I'll be forced to choose a new desktop. I have found myself looking at E17 lately, which is the new generation Enlightenment. It is lightweight and fast and looks pretty neat. But secretly I still hope KDE will get their act together and start to provide what users want. A stable, reliable and configurable desktop. "Klickybunti" is nice, but I also want to get some work done.

Update: I've been contacted by core KDE developer Aaron J. Seigo and he assured me that a classical desktop with desktop icons is still possible. So I've asked him to show me how. If he delivers the goods (and he has), I will publish them on this blog, so it will make a nice tutorial for those who want to have an "Old Skool" desktop and at the same time prove or debunk Steven J. Vaughan-Nichols claims that Joe User will be "utterly bewildered by the process". Well, one thing is for sure: they are still listening.

P.S. Comments are always welcome, but can you please refrain from pushing your favorite alternative?


Anonymous said...

You say KDE should not dictate what is best but let the user choose. I'd argue that any desktop environment, no matter how configurable, has most of the important things fixed just in the way the desktop methaphor is designed. E.g. KDE has chosen not to create a Mac-like or NextStep-like desktop, but adopt most of the MS Windows conventions. KDE 4 changes some of the basic designs and concepts that people are used to and that were very similar to Windows in the previous versions. Making KDE 4 more configurable is not necessarily going to make it more familiar, it is a different animal by its nature.

To me the real problem seems to be that a lot of the KDE 4 interface is not very usable and some major things have in fact regressed. This is not just because things differ from the KDE 3.5 / Windows conventions, but because the design just plain is not very good. If the usability was clearly better than in 3.5 for most people, I don't think there would be many complaints about configurability. Hopefully the interface bits get fixed, because the technical stuff under the hood seems to be genuinely good and also a real conceptual step forward.

Anonymous said...

I totally agree with what you are saying about KDE 4.0.x I'm running F9 and KDE 3.5.x was much better.

There is a ton of stuff that doesn't work in KDE 4.0. Kget, for instance. Konqueror has taken 2 steps back as a web browser.

The desktop sucks. The icons have this shading thing that gets highlighted when you click an icon, only it doesn't go off when it should. Its terrible.

Not being able to access stuff on the desktop is terrible too. I am constantly opening Dolphin. What a pain.

Luckily we are talking about an open source manager. Hopefully they will get enough feedback from the community that the project gets back on track - or else someone needs to fork KDE or do a KDE 3.5.x release.

Anonymous said...

"exciting and innovative" - when was Gnome ever this?

Re: kde4 - the removal of icons from the desktop is minor relative to their calling an alpha, 4.0. kde4 has much to do before people will migrate from 3.5.9 but it will happen because when they get it done it will be worth the wait.

Alan Jones said...

Ummm - did you read the article of Aaron's you linked? You can have icons, all he's effectively done is make it modular. Looks to me like the user still has control.

I've been using kde 4 daily since 4.0.1 and while not everything is there, it's definitely usable. 4.0.5 fixed the only bug that was really annoying me. Only other thing I miss is global shortcuts to launch apps.

The talk of distro support seems odd. It's still only months old, I don't see many apps which achieve mass migration within that timeframe.

As someone who's using KDE 4 every day I'd say it's definitely heading in a good direction. If you're looking something to be concerned about with KDE ask Nokia about their desire to "educate" us on allowing them to use DRM with our code.

Anonymous said...

WRT GNOME and Mono, I think there's some misunderstanding. Mono is irrelevant to GNOME no matter what Icaza says. The "cool stuff" you miss out on is inferior and bloated compared to the native GNOME counterparts that are distributed with most GNOME Distros. Granted, SUSE has thoroughly Mono-ized SUSE's version of GNOME, but if you're using SUSE, you've already committed to the Novell-Microsoft scam and even SUSE KDE isn't immune.

WRT to GNOME and "treating users like idiots", I think your preferences are clouding your judgments. I've been using Linux since 1993 and the brief time period I defected to Windows (1997-2001) because of my work, I used CygWin and other Unix ports. I'm not a newbie or an idiot. I've already gone through the pre-Slackware build your own Linux and debug your own drivers stage. I've passed the "oooo shiny" stage. I just want to get work done and would be perfectly happy with even FVWM. All I want is a desktop that stays out of my way and makes the boring things easier so I can focus on the really interesting stuff. GNOME gives me that, and each new refinement shows me that GNOME understands my need and is moving towards them.

KDE (a.k.a. the Emacs window manager) has never been right for me and if GNOME didn't exist, I'd go to XFCE. That being said, I think it's appropriate to defend KDE. KDE 4.0 really was an alpha that was pushed out just to get people to test KDE 4.0 because no-one was and maintaining KDE 3.x much longer was too burdemsone. The KDE leads admitted as much. They made the same mistake that GNOME 2.0 and Mozilla 1.0 made and assuming that such radical (but needed) change didn't have to worry about transitioning from the old to the new. They screwed up.

If you believe in KDE, don't give up on them. GNOME 2.x really didn't show it's vision until GNOME 2.8. Mozilla didn't show it's vision until Firefox 1.0. KDE 4.x might be in for a rough ride, but I'd be surprised if a future version of KDE 4.x didn't wow you. In the mean time, stick with KDE 3.x. With all the long term support distros out there, it's not going away any time soon.

Anonymous said...

You might take a look at xfce.

I find both KDE and Gnome to be not what I want, as I don't really care that much for all that eye candy. xfce is lightweight by comparison,pretty solid and relatively straightforward to use.



Anonymous said...

XFCE! It's still new enough to be fun, and it is quite feature packed for such a light weight desktop environment.

Anonymous said...

I think this is much ado about nothing.

You say that KDE 4 is not stable--but nobody said it was. And I have heard no one claim that it should be used in a production environment. Many of the apps are still in alpha stage...

Is there any release other than Fedora that has even used it? Fedora is not meant to be stable--it is cutting edge.

I've been happily using KDE for a couple of years. I just installed KDE 3.5x on top of Linux Mint--it is fast, stable, and a delight to use. Give KDE 4 another year.

Anonymous said...

But the beauty of open source is that KDE 3.5.x will be around for as long as people use it. If enough people have a similar opinion of KDE 4.x, then people will continue to maintain KDE 3.5.x.

Anonymous said...

From Aarons blog:
That last bit is important: it means that you can have an Old Skool(tm) desktop with an icon mess if that's what you really, really want. So don't bother with that flame, nobody has anything to complain about. ;)

You can still have containers that look just like icons.

I have to admit I'm still using the 3-series and flirting with E17 as well, I do not see the "lack" of icons becoming KDE's doom.

Anonymous said...

1. Gnome's infrastructure stunk big time.

2. set out to fix that by providing a flexible generic set of APIs that all desktops could use (but primarily Gnome... since what they had stunk). D-bus was actually a variant of KDE's DCOP... but with some enhancements.

3. KDE 3 is partially D-bus aware, but didn't use the session D-bus at all... preferring to use DCOP (if it ain't broke...).

4. KDE 4 and Qt... well Qt wanted a lot of the pipes moved down to it... so that changed a lot of things... also KDE 4 needed to leverage the "right" desktop paradigm, so KDE's way was dropped in preference of This is the source of most of the "instability" and lack of working applications. KDE got it two fold.... by Qt and by changes.

5. KDE 4's new interface.... The Gnome fanboys HATED KDE because KDE's interface looked a lot like Windows (it also functioned like Windows on steroids providing an integration level unmatched by any OS... even Mac). Unfortunately, people change based on complaints and FUD spread about the interface... thus KDE is attempting a new, radical (and pretty non-functional) direction. Gnome fanboys (who will NEVER switch to KDE), are celebrating the new interface!!

6. The beauty of FOSS is that KDE 3.5.x can live on indefinitely. Also, thanks to the work of, the idea of yet another desktop eventually is VERY likely. I've lost faith in both Gnome and KDE's ability to actually design something that the user wants.

KDE 4 LOOKS GOOD, but arguably so does Gnome... for whatever reason LOOKS seems to be the driving factor of both desktops now. And while I think a configurable desktop is VERY important and I loathe the way that both Gnome and KDE have approached it (by destroying the KISS principle entirely), it's obvious that neither is interested in something that just plain works.... sad but true.

Anonymous said...

The more you try to turn the desktop into something that behaves like a Web 2.0 app (i.e. do everything - just not too well), the more grief you're going to let yourself in for.

Murphy was right: the harder you work to improve something, the sooner you'll break it.

I'll just stick with xfce on my new box, and Fluxbox on the old clunker laptop I take on the road.

Anonymous said...

It seems that you haven't even used KDE 4. This post is just a remix of "things people say around", crying out loud just to have a headline. If you don't have anything new to say, why bother? This "KDE 4 is alpha oh god it sucks" thing is got old a long, long time ago.

Anonymous said...

I switched to E17 as a partial solution to xorg's no longer offering virtual desktop sizing/panning. I am very happy with it. Even so, I appreciate that the KDE people work hard to provide a good product. 4.0 may require more tuning than normal so you may want to use 3.5 until 5.0 comes out.

Anonymous said...

I do not have any contribution to add to your post, I only want to say that I am in total agreement with every bit of it and I share the same feelings. Well done and thanks.
Antonio - Belgium

Anonymous said...

Why does everyone assume because KDE released as 4.x that it's production ready? It's a work in progress and every release seems to be getting better and better. Sure - there is missing functionality - but it's getting fixed with each new release. If you want to help shape KDE4 - stop blogging about the shortfalls and start sending feedback to the developers.

What I shake my head at most is how most people seem to have lost the ability to read. From the KDE4 website:

KDE 4 packages are currently in the experimental branch and are not intended to be installed on a production system.

Unknown said...

I'm very dissapointed to read this FUD about Plasma on your blog here. What you've reported is factually incorrect, and you've simply added to the misinformation campaign that is going on right now against Plasma.

If you'd like to discuss the issue directly with me so I can help clear things up with/for you, please drop me an email (aseigo at kde dot org).

I don't even mind giving you a quick phone call or skyping if you'd like something a bit more "broadband".

devnet said...

Half of the things you link to can be taken from your point of view or the opposing view. Posting links to them does only one thing, weaken your argument.

Sorry my friend, but you couldn't be farther from the truth in this respect. If you aren't riding the cutting edge of KDE 4.1 (or, in my case 4.0.82 r8198) then you just don't know. I also ride the forefront of Gnome development being a Foresight Linux user...I get EVERYTHING before Ubuntu. I get EVERYTHING before Fedora.

See, there is innovation still happening. There are things being done that no one has done before. Pulseaudio is a huge example of this. Plasma is another.

KDE 4.1, when released, will be everything it claims it should be and more...stop running crappy, months old iterations of code and instead check out a daily build.

Anonymous said...

I've used KDE4.x on FC9, openSUSE, and Kubuntu. Bottom line: devs love their new playground, users hate it--users lose again. It's pre-alpha, at best and may never reach the heights that KDE3.5.x has. Too bad, really like KDE3.5.x. --dB

Anonymous said...

I've said it before and I'll say it a thousand times: If KDE Devs want to keep forcing me to go THEIR direction I'll tell them where to stick their lousy "direction" and go somewhere else.

I like the way my desktop works NOW, and I don't care to change the way I do things. If that doesn't set well with any one else, TOUGH! I'm not anyone else, I'm me. My point is simple: Make KDE configurable so that I can make it mine, and make it work the way I want to work, or I'll do what apparently some distros are doing, and switch to something ELSE!

I'm fed up with this crap, and if this direction of forcing users to use ONE paradigm and only ONE continues, I might as well go back to mickey soft!! I want my icons on my taskbar, I want some icons on my desktop, and I want everything to be as configurable in KDE4 and more, as it was in KDE3 not LESS!!

And Yes, I've tried and used KDE 4.0.x and the beta KDE 4.1 and they are HORRIBLE in every way. I don't want a "folder view" I don't care how much more "efficient" it is!

Anonymous said...

the last time I used a GNU/Linux desktop it was "ion" after using KDE, WindowMaker, e16 and sawfish.... I switched to OS X after using Linux and its desktops/WMs for over 10 years.

After working around 1 1/2 year with it, I can say that Linux is dead on the desktop.

Anonymous said...

Oh my... If you cannot figure out how to put icons on your desktop in KDE4.X you are in trouble. Drag an drop what ever you like :)

The direction KDE is taking is just awesome. A desktop that can be tweaked to what ever you like. For PC, for UMPC for TV for presentation boards and what not.

I think that the author fo this blog just don't get it :)

Anonymous said...

Personally the only thing I've seen so far in KDE 4.1 beta significantly less configurable than KDE 3.5 is the desktop panels. I will definetely miss them when I switch to KDE 4.1 but I hope additional functionality will be introduced in future versions.

Everything else is pretty much as configurable as in KDE 3.5 or even more!

KDE 4.0 as already pointed out wasn't meant as an end user version. It lacks many components and applications giving the impression that it is worse than KDE 3.5.

KDE 4 imho changes many basic concepts of the computer desktop as we know it from the last decade, thus it is inevitable to feel a bit unfamiliar and limited at first.

KDE 3.x (and Linux because of it) didn't won me instantly the moment I first logged in but it took it's time. I guess same goes for KDE 4.x.

Anonymous said...

I accept and appreciate your view which I am sure is meant to be constructive but I have to say that it's quite human to resist change.
I remember when XP came out in 2001, people were hating it (though not as much as Vista). What happened 7 years later is that they even came up with a "Save XP" website, resisting another change. This time Vista.

What will happen after KDE 4? Well as usual people will resist KDE 5 and protest against it saying that KDE 4 was the best KDE ever :)

I have been watching the development of KDE 4 for a long time now and I have to say that it has come a long way and it continues to evolve at a very fast pace. There are still a few uncomfortable areas in KDE4.1, which is a beta but they are also being ironed out and I am confident that it will be what everybody hopes and even more in the very near future.

Concerning stability, I am presently using openSuse 11 RC1 with KDE 4.1 Beta1 - rock solid. I haven't had a single crash until now. There are a few rough edges yes, but they haven't crashed my desktop as yet.

I would advise you to wait for the final release and give 4.1 a shot. There are a lot of wonderful improvements over 3.5x. The panel still lags behind Kicker in some ways (even though one can change it's size easier than Kicker's and it has true transparency) but apart from that one can see improvements everywhere.

Plasma is a true innovation and a brilliant technology, it's really something new,. A lot of us fail to see the vision behind Plasma but I am sure that we will all see the light very soon and I can already hear the wow!
I wish you a lot of fun with 4.1 :)

Anonymous said...

> But the KDE team has taken another path by banning the icons from the desktop, claiming that all that clutter wasn't neat.

Are you trolling or just slow in the head? The DEFAULT is to keep the icons on the desktop as a regular folder view. What's your beef?

Anonymous said...

I think that KDE should be focusing on a dual interface -
1) A simplied ultramobile low overhead laptop interface with support for low resolution/small screens and touch screen, and mostly fixed function
2) A full desktop which is fully customisable with three default configurations that can be defaulted to: one that mimics Windows, one that mimics OSX, and one that mimics Gnome. To combine ease of use with customisability it should be possible to revert to the default of any of these three with the click of a button.

Having said this though, what is really important for KDE is not new features, but a reduction of resource consumption, increase in speed, continuing improvement in stability, rather than an unstable, new version with feature bloat.

Anonymous said...

Honestly - i think the majority of users don't give a damn about this desktop thing at all, and would be perfectly happy with good old win95 style. What they need is a solid platform to run their apps, and that's where the free desktops still feel fragmented and premature (although some progress has been made).

Anonymous said...

KDE is no longer for computers that don't have new CPUs or large chunks of ram. When tic-tac-toe can sit you at 90% CPU usage with at 1.8GH chip, you know something is amiss.

XFCE is the current most configurable option for linux - although it is missing a few useful features, such as SMB and SSH browsing. FUSE is supposed to make that a non-issue, but it doesn't seem to be well supported by major distros when using XFCE,

Currently the option is to keep nautilus --no-desktop as a launcher icon; or konqueror.

XFCE is lightweight, modular and fast. With a little bit of trickery, you can even use Gnome components in the taskbar, or as a screensaver.

Gnome is quite slow, and KDE 4.x positively churns on older machines.

Anonymous said...

This is probably going to be more of a ramble than a cogent post but this topic is evocative. One thing I've always wondered about with opensource desktops is what if they weren't trying so hard to be Windows? The more an opensource desktop chases after Windows the more destined it is to be come Windows. And with that comes a dumbing down for the masses. This isn't an indictment. One can argue that the genius of Windows is making a complicated system like a computer available to anyone. If you read blogs or how-to questions related to Windows it is truly amazing how little users have to know and yet still use the Windows OS. I mean some of the questions indicate the truly clueless. But yet, ever since I started using Linux (about 2001) there is the annual question "is Linux ready for the desktop" or more accurately is there an opensource desktop environment ready for the masses. What if a desktop came along that was interruptive? Revolutionary rather than evolutionary. I haven't had a chance to see it in action but what I read about Sugar (OLPC's desktop) seemed to me to be a new way to think about a desktop. It looked to me like it was network centric. Like I said, I haven't seen it in action so I could be very mistaken about it. My question is, what -given the power of opensource development- could a user interface look like if "we" gave up trying to out Windows Windows?

Anonymous said...

Lately I had been finding KDE with too
much bloat. I starting to look at
distros with a mini version of KDE
(e.g., minime, linut mint - miniKDE, etc)
or XFCE.

Just my 2 cents.


Anonymous said...

Sorry, but what you are saying about KDE4 is just not right. If you actually tried KDE4.1 and learn ed about what is planned for 4.2, you'd know that KDE4 provides (or rather will provide) everything:
- new Concepts for those who want something new
- the oldschool desktop
- configurability to the maximum (KDE)
- but basically easy (gnomish ;) )


Janne said...

Honest question: You said you are a consultant? Do you push this level of ignorance on your clients as well? If you do, then I feel sorry for them. If you had actually spent few minutes researching and thinking about this issue, this blog-post would not exist.

I find it REALLY annoying that people whine and complain about things that they are utterly ignorant about. So this blog-post was about how "KDE-developers removed icons from the desktop!", when in reality nothing of the sort happened. Yes, Aaron used eye-catching words in his title, but if people actually bothered to read (and understand) the words beyond the ones in the title, this "controversy" would not exist.

Yes, the old system of having icons on the desktop was removed. But it was removed with something that offers the user more power and flexibility, while giving the user a system that is for all intents and purposes 1:1 identical to the old system. Yes, you can still have your precious icons on your desktop.

It's really weird when users are being offered more power and flexibility, people like the owner of this blog start to whine how "KDE-developers are trying to force us in to doing things their way!". If this change is "forcing", then why isn't the old system not forcing as well? I mean, with that old system you had exactly ONE way of doing things. If you didn't like it, your only option was to not to use it at all! Now users are being offered power and flexibility, and people start to whine how they are being "forced" to do something! Really, this whining makes exactly ZERO sense to me!

I don't know which surprises me more: the level of ignorance in the world, or the fact that people openly flaunt their ignorance in blogs such as this! And these comments are coming from someone who is supposedly an "IT-consultant"....

And before you ask: How is this new system better? Well, you can display files 'n stuff from several locations, as opposed to just one location (like with the old system). You can filter what is being shown (so it will only display your music-files for example), and in the future you can have smart views that only display stuff that meet certain criteria (all attachments sent by certain person for example), Also, this new system could be automated, so that it only displays files that are relevant to the task you are currently doing. And those are all things that were not possible with the old system. Yet you are trying to tell now how the old system is somehow "better"? How exactly is it better? It's inflexible. It's featureless. It's nothing but a dumbing-ground for files and icons.

And still, when users are being offered something new and better, that replaces a old and busted idea, they start to whine. It just boggles the mind!

Anonymous said...

I generally agree about loving KDE's customizability (and frankly, my configuration is a long way from MS-Windows-like).

But there are a few things that aren't customizable that annoy me: (1) The window-list menu has a hard-coded "rearrange-windows" item that can destroy a carefully laid-out window arrangement, and that can't be fixed without hacking the source; and (2) "klipper" is insistent on its mis-interpretation of X-Windows selection, in a way that messes up, e.g., Motif apps coming from AIX (and some of us don't have a choice of which supercomputer we have to use ;-( ).

Anonymous said...

I too have been looking at E17. I think this will be the next desktop, but they still have a long way to go, and everything is native to E17 yet.

I am looking forward to this though. I always like a good come back story, and maybe E17 will bring Enlightenment back into the spotlight once again.

Jasim Basheer said...

I could not see the video posted here. (flash plugin not installed on my Grand Paradiso on ArchLinux - I'd rather live without Flash). But I've been looking on a way to create shortcut for a few hours.. And here is what I learned..

First change the desktop from the default one to 'Folder View' (Right click somewhere on the Desktop, Click Appearance Settings, and change the 'Type' to Folder View).

From then, it is almost same as any conventional desktop. Right click, Create New etc. etc.

From my very personal perspective, KDE would have done better to keep 'Folder View' as the default, and include the Add Widget functionality here also. Seems the Widgets in the 'Default' desktop works here also. It is simply a .desktop file.