I recently got a chance to interview George Staikos, the Official Representative for the KDE Project in North America. He addresses some questions on the current status of the KDE project, and about the problems they have faced.
It has been eight years since the start of the KDE Project. At the very end of the project announcement, Matthais said that he is a “dreamer”. Has his dream become a reality or is KDE still someway off?
I think it has. KDE is used by an incredible number of people worldwide. It has matured into a stable platform and even sets the standards in many areas. It’s truly something to be proud of.
Read More @ Linuxtimes.net.
No mention of Huame INterface Guidelines or cleaning up the clutter. KDE may be technically excellent but in terms of eseability it still agls behind.
Wow, that was not even remotely close to half-assed English. Considering you’re Australia I’d have expected better.
Frankly KDE is neck-in-neck with Gnome in usability. The majority of popular distros are built around KDE, for example, SuSE.
Just because KDE doesn’t have a formal HIG they still policies which, with not the same level of success as Gnome, that cover the same thing.
There is no need for trolling; both KDE and Gnome are fine desktops. Personally, I find KDE cross-integration amongst apps (KParts etc) to be the killer, and until I can browse the net and my files with the same program I won’t come near Gnome.
Anyone have an ETA for KDE 4?
Windows has had COM+/OLE etc and it’s being replaced with Linux.
<<Frankly KDE is neck-in-neck with Gnome in usability.>>
No, not even close.
<<Just because KDE doesn’t have a formal HIG >>
http://usability.kde.org/hig/
I wonder if it will be GStreamer?
Here we go again with the KDE vs GNOME thing.
Sorry about my typing, Captain Pinko. I am very tired after editing LaTeX files for the last forty-eight hours. Let me try again:
[There is] no mention of Human Interface Guidelines or cleaning up the clutter [in the toolbars]. KDE may be technically excellent but in terms of useability it still lags behind [the GNOME desktop environment].
I hope that meets your standards.
Why do objective articles result in subjective bickering and personal assaults? If you like Gnome fine, use it. As a neophyte Linux user, I love KDE and the power it affords. Compared to Windows, which offers pathetic control over system look and feel, configuration, access to components, etc., I think KDE is at first tough, but then easy and rational in its design. Kind of like playing a guitar. Or learning OOP versus procedural language. The elegance of design is not immediately obvious.
Linux is complex and powerful. If KDE is hard to use for some folks, maybe Gnome is a better fit. I am not trying to stir sh*t here!!!! HONEST!
P.S. Suse 9.2 and KDE 3.3 rocks. Bottom line.
Comparing KDE’s “HIG” and Gnomes HIG is like apple and oranges, Gnome’s HIG push is a lot more advanced, and more indepth and covers alot more applications (even some GTK only applications). To semi-unify the desktops, the best scenerio would be to entice KDE follow Gnome HIG, or for Gnome HIG (or even HIG-lite) to submitted to freedesktop.
<<To semi-unify the desktops, the best scenerio would be to entice KDE follow Gnome HIG, or for Gnome HIG (or even HIG-lite) to submitted to freedesktop.>>
Never going to happen, the Gnome/KDE devs are way too stubborn. The KDE folks won’t sacrfice functionality[0] and the Gnomers will never taint the holy HIG.
[0] Heres something amusing, a 100+ comment discussion about “Delete vs Move to Trash”. LMAO those KDE-devs crack me up.
http://lists.kde.org/?l=kde-usability&m=110492635002578&w=2
Personally, I use Gentoo & KDE 3.3.2…and it’s probably the finest desktop I’ve ever used.
Can’t wait for the next release (and the subsequent 24 hr. compilation on Gentoo
Why can’t a KDE article for once have only technical or general discussions about KDE, why do all those that seem to have issues with KDE have to come and start flaming? The same goes for GNOME news, but it’s not as bad with those.
Are these anti-KDE folks really trying to keep on flaming until they have forced the developers to give up developing KDE? Now, let us see some more maturity from the non-KDE folks, ok?
G’Day
Did somebody mention Australia? No worries mate 
I find KDE more productive and better looking then GNOME but there are some things KDE can learn from GNOME.
s/KDE/Windows
<<[0] Heres something amusing, a 100+ comment discussion about “Delete vs Move to Trash”. LMAO those KDE-devs crack me up.
http://lists.kde.org/?l=kde-usability&m=110492635002578&w=2>>
I remeber a few years ago, there where the same discussions on the gnome-lists.
I remeber I clearly as if it just happened yesterday. A lot of people saying “If you do this and do that, I’ll hate Gnome and go to KDE.”
I wonder if that means, they are comming back now…? *shudder*
Yep, that is true too.
Not meaning to troll here, but I don’t see how anyone can claim that the Gnome HIG are a good thing when they’ve forced the backwards confirmation dialog on everyone. It may be theoretically better, but really, did anyone actually try them on a typical (i.e. non-hacker) computer user? You can say all you want about people who have never seen a computer, but the fact is the main target market should be people migrating from windows. Not to mention simple incompatiability with every non-gnome program. That has been enough to put me off gnome completely.
KDE is completely devoted to immediate functionality in practice rather than the theoretically best solution. Hopefully this will mean KDE switching over to GNOME technologies as they mature and the theoretical advantage starts to show. But integration has to be a two-way street. So far I have not seen GNOME adopt one thing from KDE. They’ve published their own setups as standards and KDE has adopted them or not, but I haven’t seen them change anything to be compatiable. In the same time KDE has announced a switch from DCOP to DBUS, an abandonment of aRts, and included support for using your gnome theme for your kde apps by default. If this keeps going, we will end up with GNOME as it is competing with KDE containing the best bits of GNOME, and it should deservedly die. Feel free to correct me if this isn’t the case, but this is the impression I get of how the integration is going.
>> Not meaning to troll here, but I don’t see how anyone can
>> claim that the Gnome HIG are a good thing when they’ve
>> forced the backwards confirmation dialog on everyone.
1) Don’t whine about the backwards order of buttons. It can be changed with the latest release of Gkt+ now.
2) You are trolling because the HIG is not just about pestering people with non-standard dialogs. Good luck with your bad-ass KDE dialogs and windows. Which are, in my opinion, a place worthy in the hall of shame for UI designs.
I always wonder how true these GNOME v/s KDE fights are.
I have used RedHat 5.2 to Mandrake8-8.2 and now Slackware 10. My home directory is the same since Mandrake 8 which is at least 2 years.
Any time I do a reinstall or switch the productive partition(I usually triple boot, one production, two testing grounds), I restore my home directory and everything is back where it was. I have not tweaked KDE settings in years except for trying plastic theme sometime back.
Initially I did spend around 15 minutes cleaning up konqueror and konsole to suit my needs. But that was over 2 years. I would not consider it too much. In fact, once in a while I create a test user to see the new KDE default.. Otherwise I would just miss it..:-)
In all honesty, I decided to give GNOME a shot and hence change the X login from KDM to GDM. On first login, I could not find any button to shutown or restart the machine.
This is Slackware GNOME, so I believe it should be closest to its source form. My friend swears that the button is there in Fedora GNOME..
Back to KDE faster than I can blink.
And to the poster who think KDE file dialogs are not good, to each his own. I would have opted for KDE even if its UI was uglier than windows 3.1 because of the functionality it offers.
But the UI is not ugly..:-) It is beautiful OTOH.
And don’t forget GNUstep.
KDE’s HIG is the counterpart of the KDE User Interface Guidelines, which have been around for ages. It crosses over with Gnome’s HIG in certain areas, but a lot of what the Gnome HIG requires is in the UIG or is implicit in QT and the core libs. For example, KDE applications don’t need to think about icon spacing on the toolbar, pixel-perfect alignment, etc. That all gets inherited. KDE’s HIG will never be the same as Gnome’s HIG.
Delete vs. Move To Trash: There is a very real question of expectation in destruction of documents, and retrieval. If someone accidentally loses their work, that’s the worst sin that a DE can achieve. Making a dialog box take two more seconds to read vs. losing three weeks of work. It’s right to think about it for a while.
Standards: Gnome has adopted a few standards too; they’re working on DBUS as well, along with menu specifications, MIME type recognition and a few other things. The KDE lot are just getting there faster in some ways. DCOP is not being switched to DBUS. DBUS is modelled to DCOP and is a superset to a certain degree. A DCOP->DBUS bridge is being built; DCOP isn’t being stripped out and applications won’t stop working.
ARts: It was good enough when Linux barely had audio and KDE was 1.x; these days, something more powerful is needed. These things take years to write… you can’t just write the perfect sound server in a few weeks.
Gnome Themes: KDE can use KDE or Gnome themes. If you find a random theme online and you love it, it’ll work in KDE. It may work under Gnome; it may not. I know which option is better for me.
KDE Immediate Functionality: The opposite is actually true. KDE developers spend a lot of time debating and trying to find the best solution to problems. KDE releases are also feature-based, rather than Gnome’s time-based releases. The intention is to develop something good, have it mature, and then release when it is ready. Sometimes a developer has a rose-coloured view of the feature’s stability, yeah, but by and large, it works. In terms of best functionality, DCOP and it’s child DBUS beat the snot out of CORBA, a mainstay of Gnome architecture. This is not to say that CORBA isn’t a good thing in certain cases. It just sucks on the desktop.
Re: Yourname: 1) Yay! Seven years of development, and the users are listened to!
2) KDE dialogs are happily standardised. The UIG and inheritance from the core libs and QT cover most of dialog design. KDE mailing lists don’t get posts saying “I was writing a document and when it asked me to save changes, I hit no by accident, because the buttons are the wrong way around” and at least the KDE file selector has a location bar (wtf is with that? Ctrl+T to get one, isn’t it? A fact this is mentioned nowhere visibly). Honestly, I’ve never heard a KDE user complain about the dialog boxes, Gnome users complain about KDE dialogs, but everyone else gets along with their lives. Dialogs are not even on the radar.
Good summary.
Dialogs are not even on the radar.
I’d like to see some cooperation about the look and feel of File Open/Save dialogs via freedesktop.org. It would be great if Open Office, Gnome and KDE all had dialogs which looked the same, even if the code to invoke them might be different.
There is nothing exciting or innovative about dialogs, and ‘pissing contests’ about which toolkit has the best looking ones don’t get you very far.
I would have thought button order on a dialog needs to be configurable anyway. If you prefer your OK button on the left of the dialog for a ‘left to right’ language like English, it follows that you would prefer it on the right with a language like Hebrew. Maybe we could have scrollbars configurable to be either on the right or the left of a window too (I personally prefer them on the left side).
…I don’t see how anyone can claim that the Gnome HIG are a good thing when they’ve forced the backwards confirmation dialog on everyone…did anyone actually try them on a typical (i.e. non-hacker) computer user?…
Please, do us all a favor and do this:
Go to Home Depot, and purchase something with a credit card.
Go to Lowe’s, and do the same.
Which is easier?
Personally, I find Lowe’s to be easier because the final OK button after signing your name on the digital pad is on the RIGHT — in Gnome HIG/Apple Mac order. Why is this easier? Because I write left-to-right, and after signing my name the pen is (obviously) on the right of the screen.
Home Depot uses the “obviously more correct” OK CANCEL order, so if I try to go fast I invariably hit cancel instead of OK.
After seeing that, I started making the same observations at various other checkout lines in supermarkets and other stores. Invariably, for me, the easier to use checkout items have OK/ACCEPT/whatever on the RIGHT and Cancel on the LEFT. Again, this is probably easier because I’m right handed so my hand gravitates toward the right side of the machine.
Another key point that is sadly overlooked is that OK/CANCEL shouldn’t be used according to the Gnome HIG. Nor should Yes, No, Cancel style dialogs. Instead, actual verbs such as Save, Print, and Delete are used. This greatly increases understanding, as it requires less thought to read and interpret the question being asked. Compare “Do you want to delete the file $file? Yes, No” vs. “Do you want to delete the file $file? Cancel, Delete”.
i was wondering which it was going to be. DBUS? HAL? Spatial Nautilus? but it seems we have decided to return to that tried ok-cancel/cancel-ok argument. i actually like this one cause its so simple, and illustrates the differences between kde and gnome quite well.
KDE descisions are based on opinion, and gnome descisions are based on design.
before i get flamed, these are different, not superior/inferior. gnome people wonder why the hell kde would order buttons in a fashion that usability research has shown to be inferior. kde people wonder why the hell gnome would do something that would piss off its users.
see how easy that is? its designers vs the masses. something of a philosophical debate rather then a technical one. some people believe in the adage “the customer is alwas right”. then theres those of us that the customer has no idea what right is, they only know what they want.
what it boils down to is this, if you dont like kde, use gnome. if you dont like gnome, use kde. if you dont like either, give xfce a try. but bickering about a de when there are so many available is rather silly.
oh man, just ran across this poking around the kde site
One of the greatest things about Open Source/Free Software is that there does not have to be one “winner” and some number of “losers”. Multiple projects can work and succeed together. Just because KDE strives to be the best it can be does not mean it must exclude all other projects. The only “battle” KDE fights is the “battle” for self-improvement.
never seen that summed up quite so well.
It has matured into a stable platform and even sets the standards in many areas.
I wonder what kind of standars was he talking about.
I think the enterview lacks of clue questions, questions that regulars KDE users want to hear.
“It may be theoretically better, but really, did anyone actually try them on a typical (i.e. non-hacker) computer user?”
Yes! MacOS X uses the same button order! MacOS X is often praised for its userfriendliness.
IMHO kde 3.4 will start a complete new trend in KDE development. i’m reading many discussions by KDE devs about “feature craze vs usability” and i think plastik being finally the default theme is a hint of the efforts KDE devs are putting into refinement. I guess the next major release, KDE4, will be “the” unix desktop, really.
this also means cooperation with freedesktop.org (gstreamer!), integration with project utopia (although KDE already has DCOP…) and usability tests on native kde/qt killer apps like k3b, amarok, quanta etc etc
PS: please don’t feed the trolls
All the propoganda by Gnome fans and each time I try it, I come back disappointed. KDE gives so much joy to work and play with.
For me it is more natural if the OK button is on the left, because I read from left to right and so the OK button (which is the one I use more often) is the first button I see and the Cancel button is the second button I see.
The whole discussion is totally useless, because honestly I think it does not matter at all where the OK and the Cancel button is. There are many arguments why Cancel/OK is better, but there are as many arguments why OK/Cancel is better.
The important thing is consistency: all applications should have OK/Cancel in the same order. And even if Linux people don’t want to hear that, but there are a lot of people that also (have to) use Windows (I am happily not one of them), so please just make the order the Windows the default one and make it configurable, if people cannot live with the fact that the have the OK/Cancel buttons the same way they are in Windows.
I would also welcome it very much if the file requesters would look the same in KDE and GNOME. I am a KDE user and I use mostly KDE applications, but sometimes I also use some gtk applications (gimp and firefox) and I always get annoyed by the file requester. I don’t say the KDE filerequester is the best, but I am used to that and probably the GNOME people are used to the GNOME filerequester and it would be very nice if GNOME applications could use the KDE filerequester when run with KDE and the other way round.
That said, I compiled KDE CVS and it already looks quite nice and is very stable. khtml feels faster and seems to render more web pages correctly (even though I find hardly any webpage that Konqueror from KDE3.3 does not render correctly). JuKs UI looks a little nicer and shows a small track history, kpdf is completely reworked and it is really amazing now. Kolourpaint is also improved. I really look forward to the release of KDE3.4:-)
> Never going to happen, the Gnome/KDE devs are way too
> stubborn. The KDE folks won’t sacrfice functionality[0]
> and the Gnomers will never taint the holy HIG.
…and thats why I use KDE, where I can do what I want to do.
Re: Buttons: To find the cancel button, I look for the big, gaping, red X that implies “that’s a negative!” But to each his own, I s’pose. I’m a KDE user, but GNOME gets this right, IMO, only in that the button order is configurable. Apparently, it’s really an important issue to some people…
KDE4: Is there a mailing list posting / other resource that describes how Qt4 is going to change KDE, or other info related to what all these “great changes” are going to be in KDE4?
All the propoganda by Gnome fans and each time I try it, I come back disappointed. KDE gives so much joy to work and play with
I have the same problems with KDE.
“It may be theoretically better, but really, did anyone actually try them on a typical (i.e. non-hacker) computer user?”
Yes! MacOS X uses the same button order! MacOS X is often praised for its userfriendliness.
MacOS also has a paradigm of ejecting/unmounting a disc/CD/floppy by dragging it to the bin (which is highly unintuitive and implies that it will be destroyed; if this was consistent, all files dragged to the bin would be printed
).
Just because Apple does something, doesn’t mean it is a good idea. Likewise with Windows.
(Same guy posting from home)
Standards: I’m not denying that GNOME supports standards, I’m just saying I’ve never seen GNOME switch to a KDE standard wheras it’s happened the other way a few times. Making it appear (I’m not saying this is the case, just that it looks that way) that KDE is putting more effort into integration than GNOME is.
ARTS: Exactly, but this shows what I meant afterwards. The GNOME people went with GStreamer as a theoretically superior solution, while KDE went for arts. Arts was working but limited, while gstreamer was buggy for a while. Now that it has matured, it’s clearly the better solution, and I expect KDE to switch to it pretty soon. But in the meantime, KDE has had a more useable (imo) audio system.
GNOME Themes: Exactly. And again it looks like it’s KDE making more effort at integration than GNOME.
KDE immediate functionality: The impression, and that’s all it is, that I got is that GNOME will always choose the theoretically superior solution, perhaps reasoning that it will be better in the end, wheras KDE will kludge when necessary to get something that works now, on the basis that it can be switched later if they need to. Arts is a perfect example. Or look at the Qt/gtkmm arguments – gnome people tend to argue “gtkmm follows stl conventions which will make it more useable for c++ people” wheras kde people tend to argue “developers who’ve used both say they find Qt’s funny c++ extensions make it more useable”.
RE GNOME and KDE: In a default gdm install, you have to click some button, “actions” or some such, then you get a menu with options to shut down or reboot. But I think that might be disabled by default, so you need to login as root and enable it in the gnome control centre.
ARTS: Exactly, but this shows what I meant afterwards. The GNOME people went with GStreamer as a theoretically superior solution, while KDE went for arts.
No they went for ESD, just so they wouldn’t have to adopt arts or even discuss adopting arts. GStreamer wasn’t around then.
Now that it has matured, it’s clearly the better solution, and I expect KDE to switch to it pretty soon.
KDE will probably have a KDE multimedia layer to allow different back-ends and won’t adopt anything. The most superior option is NMM (Network Multimedia – http://www.networkmultimedia.org/). It’s well structured and well written. The GStreamer folks don’t seem to know anything about good object-oriented design in systems like this.
Another problem is that GStreamer was supposed to be a lightweight system-level multimedia framework that anything could plug into. We’ve now got various parts of it depending on GConf, and henceforth dependencies on things like bloody orbit.
Another problem is the bindings. Creating and especially maintaining bindings for languages and development tools is one thing, but going the other way in creating bindings from GStreamer to Qt and KDE is just totally unmaintainable. We’ve seen that with the C API being used directly for GStreamer because the bindings are out of date. In terms of the KDE people discussing maintainability after arts that’s pretty ludicrous, and imagine what kind of a mess there’ll be in the run up to a release.
Standards: I’m not denying that GNOME supports standards, I’m just saying I’ve never seen GNOME switch to a KDE standard wheras it’s happened the other way a few times.
I’m still waiting as well.
“Just because Apple does something, doesn’t mean it is a good idea. Likewise with Windows.”
Until people massively flame about MacOS X usability (like they do now with Linux usability), copying MacOS X is a good idea because it’s the best way to answer to trolls.
I guess the general agressive GNOME vs KDE lobbying in this thread has put your shorts in a twist which can explain the negative tone of your post, but let me clear up a few things with GStreamer for you in case you really believe them.
a) Which option is clearly the most superior one is of course a mostly subjective call. And since you don’t qualify that statement or you claim that GStreamer developers have no idea about good object-oriented design I can only say I think you are wrong on both accounts. Judgeing by your next statement I guess you are just making random claims based hearsay or pre-judgement rather than personal experience.
b) GStreamer is lightweight. Which is why the GPE handhelds GUI for instance can base their media player on it. The GConf support you where talking about is an optional package to offer some simple gconf integration for developers who wants it. GStreamer and its plugins builds fine without it and many applications do not use the GConf stuff and would work fine if it where not installed on their system. The only reason I can see for you to have gotten the impression that GStreamer have a hard GConf dependency is due to some distributions packaging all the GStreamer stuff they use into one package, like Red Hat. Which means the Red Hat plugin package depends on a lot of things. Other distributions package each plugin by themselves and thus if you are not using any GNOME apps, just AmaroK for instance, you would not get gconf pulled in by your package system when you get the needed GStreamer plugins installed. We could do a long discussion about which packaging policy is the best, but to make it short, it has little to do with GStreamer itself, its a packaging question. That said we are planning to split up our main plugins package into several smaller packages soon. Which might lead distributions to package things a bit differently as they will be building their packages from several smaller tarballs instead of one big one.
I guess the general agressive GNOME vs KDE lobbying in this thread has put your shorts in a twist which can explain the negative tone of your post, but let me clear up a few things with GStreamer for you in case you really believe them.
Hence the tone of your reply I guess.
Hands down, any default multimedia framework has absolutely got to be maintainable from a KDE perspective. That’s the thrust of my post, not the KDE vs Gnome crap (hence your response). The bindings issue is certainly a very huge and real problem there, and I don’t think I’m wrong at all quite frankly.
As for GConf – well, things have been optional on many platforms in the past. It really depends on where the development goes though.
Judgeing by your next statement I guess you are just making random claims based hearsay or pre-judgement rather than personal experience.
The debate of C and C++ will rumble on and on. The fact is, for a system as componentised as a mulimedia framework needs to be you need languages and tools which provide good object-oriented features. In C++ you get that ability now to have good natively compiled efficient code, with the somewhat higher level features you need for development.
Tell me why I’m wrong.
The thrust of your point does not make sense. Just like other components that KDE uses GStreamer already have a large developer community maintaining it, including people who also work on KDE applications. Using GStreamer is no bigger problem for KDE than using X-windows, glibc, the linux kernel, hal and a host of other technologies coded in C.
GStreamer is highly object oriented and heavily componentised. The KDE hackers who actually use GStreamer are not complaining about GStreamer cramping their development in any way. The bindings which you seem to feel is such a huge problem is according to Scott Wheeler not that big of a job to keep working nicely, it has more been a combination of the lack of time on his part and a bit of a chicken and the egg situation where no apps use the bindings so he hasn’t felt them to be top priority and due to the lack of the bindings being up to date the application developers have instead opted to use GStreamer directly.