Slashdot Mirror
Slackware Likely To Drop GNOME Support
An anonymous reader writes "After Hewlett Packard, who jumped off of supporting GNOME, Red Hat has followed by splitting their Desktop Linux out to Fedora which is community driven, and now distributions like Slackware have started to drop GNOME entirely in favor of KDE. Read more about their decision here. It looks like companies as well as distributions start focusing towards one solution." Patrick Volderking's quoted message doesn't announce a final decision to drop GNOME from Slackware, however -- and as the followups in that thread note, it could be interpreted as an endorsement of the good job done by Dropline in packaging GNOME for Slack.
Its about time one of them, I don't kare which one, got the upper hand and snowballed.
Choice is good, but if we're going to have a million different distros, then we don't need every single one to have all million software packages too.
This is probably a good idea, for every old joe-schmoe who installs linux, there can be more or less, a unified 'look'
;)), then it probably makes life much easier for everybody.
Being more partial to KDE than GNOME, I don't really see a problem with it, but packaging it is the way to go. If it's a package, that can be 'apt-got' (just for example
Error 407 - No creative sig found
Less bloat for the install. Now maybe we can get Slackware back down to one CD for installation!
I've used KDE and Gnome before, even somewhat recently, but just can't stand the overhead. They both look great, but I'm much happier in Fluxbox. All I do is work in xterms all day anyways.
From what I've heard, Dropline Gnome really is an excellent package. Makes sense for Slackware to drop Gnome support, if there's already an excellent source for a Gnome package for Slackware.
Kudos to both Patrick V. and the Dropline Gnome maintainers! This is how open source should work.
its not that pat wants one DE its that gnome is taking too much effort for so little when dropline is good enough.
Most Slackware Gnome addicts use Dropline anyway since it's a very lively, well maintained project.
;-)
If Todd of Dropline and Patrick work together this could be pretty good for both projects. Of course there is PAM integration in Dropline that Patrick dislikes and therefore he won't include it in the "official" CD set. Slack with Dropline is in fact the best Desktop-Linux Experience I ever had.
Let's hope Todds servers can handle all the load following a slashdotting.
Look out BSD! Gnome is coming your way! /karma to burn
I'm begining to face facts. I still think that GNOME looks better, and is, in many ways, easier to use. But KDE has even made huge progress in these areas in the last year (especially with Konqueror and new skins that finally *don't* look horrible, at least to my eye).
GNOME still has nominally better applications in certain key areas compared to KDE, for example, Ximian Evolution. However, again, KDE has made enourmous progress in this area, all in the last year. It boggles my mind to see how quickly this gap has dissapeared in one area - compare Instant Messaging in KDE and GNOME two years ago (nothing vs Gaim) to now, Kopete has developed so quickly it's just amazing.
One thing I did miss in KDE was Mozilla. But now, we can even use Gecko as a rendering engine in Konqueror, so even, like me, if you considered that KHTML was inferior to Gecko, this "advantage" for GNOME has now dissapeared (also thanks to Apple and Safari).
I still think KDE needs some work, especially in the ease-of-use department (too many settings presented to the user, some intelligent hiding would be appreciated) - but this is improving. And, even as a GNOME user, I have to admit that C++ as a basis is a much superior choice to C, especially considering the kludge that seems to underly GNOME, separate libraries for GTK and GNOME applications with surprisingly few applications taking advantage of the GNOME-only libraries.
If you look at the distributions on the shelves, SuSE is KDE, Mandrake is KDE, Linsipre is KDE (with modifications). You can't buy Fedora at PC World. Any new user getting interested in Linux would probably go here first, and by consequence they're going to get KDE.
So whilst I will keep GNOME around for a while yet, and I think the "race" is far from over (who says there has to be a winner anyway? The whole concept of a "war" is just completely silly), if KDE goes on to become the defacto Linux desktop, then I won't shed that many tears. Of course, GNOME, I'm sure, will be around for a long while yet.
HP and Redhats actions are completely different. HP sponsored SCO's roadshow, so we know how relevant their opinion is. And Redhat's Fedora uses GNOME by default!
Sure, slackware is considering dropping gnome support, but this isn't some kind of mass migration away from GNOME, look at what Novell & Sun base their linux desktops on.
Kudos to the submitter for successfully trolling the editors
Cheers Koz
I thought this rejoicing had something suspicious to it...
More seriously, this whole thing sounds sensationalist to me... RedHat adopting a community model with Fedora, and one fed-up maintainer for a redundant Slackware package do not a mass defection maketh. The HP bit might be worrisome, but.
Most of all, I fail to see how one environment 'getting the upper hand' can possibly be construed as a Good Thing. Nobody serious clamors for less operating systems, less trouser styles, or less pencils. And GNOME is definitely the more professional and efficiently designed, from a purely UI perspective, of the large Free desktop environments.
Until Pat weighs in on this publically I'm not certain about the validity of this claim.
Gnome has long ago lost focus on its goals. It used to be geared towards linux users. It was meant to be a fast and customizable linux DE. Somewhere between 1.4 and 2.0 Gnome development changed. It lost sight of those goals and became geared towards newbies and end-users.
Frankly, it never was as good as KDE at that. Being "user friendly" meant changing the reasons so many of us used and liked Gnome, alienating their base. Gnome became difficult to compile and even more difficult to package. Why can't Gnome install nicely using "make install DESTDIR=~/pkg"?
Pat mentioned in that e-mail that about a third of his time is spent trying to support Gnome, which given the entire size of Slackware is apalling. Spending a third of your time supporting what is around a twelth of the system's size will wear out anyone.
My personal hope is that the Gnome developers will wake up, get their asses in gear, and realize that they're not going to beat KDE on usability for newbies. They need to return to being the fast, custimizable linux DE. I suspect that most of Gnome's old users are now using a plain window manager or Xfce (good stuff).
Slackware, what else when it must be secure, stable, and easy?
All in all, this is not a final decision, it's just a rumor . As long as Patrick Volkerding has not removed Gnome and annouced it either on the Slackware website or in the ChangeLog, I won't believe it...
And this was typed on a Slackware 10 machine running XFCE... Which, IMHO, is so much better than Gnome...
The right to offend is far more important than the right not to be offended. (Rowan Atkinson)
It's not compulsory.
Deleted
This was Patricks' argument for dropping GNOME. Instead of dropping GNOME support, why not communicate with the GNOME community to resolve the issues? This is really a minor technial issue, and I'm sure things can easily be done to make including GNOME as easy as KDE.
Anyway, I'm sure Slackware will never drop GNOME support. People will stop using the distribution in a second!
This is probably why having a single "dicator" maintaining a distribution is a bad idea: He has very little contact with the community. It's not possible for other's to get involved with the development process either. It would be a trivial task to make someone else maintain the GNOME sources in Slackware.
I like Slackware, running slack 10 now, but this makes me change my mind.
KDE can be anything you want it to be. You might have to work at it, but unlike Gnome recently, KDE still gives you all of the configuration options you could want to make the system your own. Chances are that the default is 'Windows like' because since almost everyone has used that, its a good starting point and middle ground.
"I use a Mac because I'm just better than you are."
Latin phrase meaning, "It does not follow." The characteristic feature of arguments that fail to provide adequate support for their conclusions, especially those that commit one of the fallacies of relevance.
"After Hewlett Packard, who jumped off of supporting GNOME, Red Hat has followed by splitting their Desktop Linux out to Fedora which is community driven, and now distributions like Slackware have started to drop GNOME entirely in favor of KDE."
What's up with the quality of trolling on Slashdot these days? Even the article summary trolls are poorly written and transparent these days.
Fedora and Redhat Workstation default to using GNOME for the desktop. Novell hasn't cancelled Ximian's GNOME efforts, and is in fact working on improving GNOME in SuSE. Solaris and Sun JDS both use GNOME.
Not that KDE isn't doing very well for itself as well, with SuSE being a very nice KDE oriented distro, not to mention Mandrake, and many others.
Both are doing just fine - the prospect of some distros focussing on one is not surprising, but I'd hardly call it significant. The whole DE flamewar is mostly rather silly. FreeDesktop.org is doing a good job and increasing cooperation and shared functionality between, not just KDE and GNOME, but XFCE, WindowMaker/GNUStep, and even, to some extent whatever new DE Enlightenment eventually turns out. There are different desktop needs, and different DEs pursue very different goals. As long as FreeDesktop.org manages to continue its efforts to define some good shared base standards things will work just fine.
Jedidiah.
Craft Beer Programming T-shirts
I'll let the zealots take over from here... ;~)
"An infinite number of monkeys typing into GNU emacs would never make a good program."
Why do we have to follow the conceptual desktop UI that MS has laid out? Linux should follow the path to what makes using it easier. A single button under which everything is nested seems unnecessary - there have to be better ideas out there.
In the meantime, I've dropped Gnome on my FC2 box in favor of Windowmaker. It's much much faster, eats many fewer resources, and completely avoids the whole "taskbar" concept. And on the plus-side, my roommates are no longer able to use my computer to do anything because they don't know how to work windowmaker. It's just a blank screen with some funky icons and a paperclip!
OK, I confess I have seen some bad submissions, but what does HP dropping gnome (not that I have ever seen anything in news about this), Redhat's decision to spin off Fedora, and Patrick's decision that dropline is good enough for him to stop wasting his time with gnome's odd build procedures have in common? Troll usually appear in the comments, not the articles. Although timothy did make an effort to unspin the Slackware news somewhat, it is still crazy that he would post such flamebait.
Just for the record - in case you aren't up on the latest news - Redhat still ships a desktop linux that uses gnome, and the Fedora project is still one of the strongest linux distributions, along with Debian and Suse (Novell), who both still include gnome and have no intentions of dropping it. Additionally, Sun and IBM are still committed to gnome.
Disclaimer: I don't like KDE. I miss my old mac.
Schrodinger's cat is either dead or really pissed off...
The only way for KDE to win is if Novell buys and LGPL's QT. Otherwise it is too expensive for small / midsize shops to buy the licenses need to ship their QT projects.
Try getting your manager to approve such a large purchase these days when GTK is free. It is very difficult.
>but in terms of the majority of the linux community KDE isn't even used.
:)
:)
Check online polls, KDE always comes out as no 1.
Look at awards, KDE usually wins the award for being the best available desktop environment
So in terms of the majority of the Linux community, KDE is de leader
Heck, even Linus likes KDE over gnome
Well, the fileselector is much better than the old gnome-file-selector.
Now with gnome 2.8 and udev+dbus+hal the new fileselector rocks! Navigation is much quicker (due to the "directory buttons"). Try it a while, you will love it if you just forget the Microsoft/KDE training you had.
Move Sig. For great justice.
I'm pretty much a gnome fan, but in reading the thread linked to in the story, I have to agree: every gnome since 1.4 has just felt "off" to me.
And having dealt with the hell of compiling gnome on slack, I can't blame Pat a bit.
Funny thing is, although I still use gnome, I've got one box running XFCE and it feels much more like gnome 1.4 did -- I'll probably migrate there as long as I can count on a few GTK+ apps (mostly gnumeric, gvim, and I'll toy with giving up evolution if needed.)
KDE has just never done it for me. I can't put a finger on it, it just doesn't feel right or "open" (yes, I see the irony here.)
The main things that originally attracted me to gnome were a few well-done apps and the clean simplicity of 1.4 -- if only the gnomesters would go back to this root.
Whatever the case, I'd like to echo sentiments here (and on the forum linked to in the article) -- it'd be great if Pat would include a well-integrated Dropline package with slackware, and if Dropline would consider a second 'standard' slackware i486 distro, as this can be counted to run on practically all platforms (the i686 won't.)
1) Uses less resources than KDE
It seems to be using a lot more resources here.
I'm afraid you have no idea what you're talking about.
KDE and Qt also fully support switching out the widget rendering engine - I should know, as I've been writing style plugins that do this for *years* now.
And this isn't a recent feature - this has been available since KDE 2.0.
-clee
KDE can be anything you want it to be.
No it can't. I want KDE to be simple a simple UI that has all the options I use and nothing more. Unfortunately there's still no options for "only show me the important widgets" or "death to sidebars" or "simplify these menus" or "Just make stuff work, and get out of my way dammit!".
When the KDE developers realize that 80% of the widgets on their screens are utterly worthless, a clock applet doesn't need 5 tabs full of options and a file manager is not the same thing as a web browser, I'll go back. Until then, Gnome does almost all of what I want, with less frustration and fewer wasted pixels.
0 1 - just my two bits
You hate KDE, because it feels like Windows. Well, join the club!
But err, what does it have to do with a discussion about GNOME? GNOME feels like Windows, too, and just because it gets dropped from Slackware doesn't mean you have to use KDE. You can do just fine without either one of them, and you can even get GNOME from outside Slackware if you want to run it.
As far as getting to the masses goes... A wise man once said "Build a system that even a fool can use, and only a fool would want to use it." Would you rather be using a system that is best for you, or best for the masses?
Please correct me if I got my facts wrong.
this is silly. all of the qt-linux code is GPLd. thus, you may always use it for anything sans fee, and no of course you can't release it under a bsd license any more than you could do the same with the linux kernel.
all dual-licensing means is that you can do things that you wouldnt be able to do under the GPL (bsd, proprietary software) by paying a fee to the owners of the copyright.
the windows licensing is a separate issue. rather than being dual-licensed, this separate codebase is not released under the gpl. the kde-windows people are working on porting the gpl'd qt-nix framework to windows, if Trolltech were enforcing restrictions beyond the gpl they would not be able to do this.
U.S. War Crimes blog. Email for free Mandriva support.
HP cancelled their GNOME on HP-UX port, which should tell you more about HP-UX than GNOME... ie. that HP-UX is not their leading workstation OS anymore, so it doesn't require active graphical desktop development. HP continue to be involved in the GNOME Foundation, to great effect.
Issues that caused me to switch to KDE circa Redhat 8...
:-)
:-)
1) Miguel de Icaza.
I will never forgive him for beginning work on Mono, fracturing the limited number of developers for the GNOME Desktop. Setting it back probably years behind KDE. For What? A Microsoft red herring planted there strategically to insure any Linux Desktop application framework built on Mono could be stopped easily using copyright, DMCA and patent law....SHOULD it become too popular.
2) The Lack of decent or equivalent KDE development tools. KDevelop? KDesigner? KCacheGrind? KDevelop Assistant? The list is endless and the above applications will squash anything the GNOME community has like a grape to develop fine bugfree native Linux applications.
If you do not have a coherent development framework how the hell can you develop anything decent? No wonder the Distro/End User GNOME community is fundamentally stressed out. These sorts of complaints do not exist in the KDE community.
There are different ones.
But they do not involve resorting to talk out in the open about dumping a desktop linux initiative such as GNOME. This is VERY serious.
The last gaffe that happened of this sort was xfree86....which is now relegated to the dust bin of history. But, AT LEAST it was reborn better than ever!
Perhaps, what is required....is a FORK of the GNOME Desktop project? A fork of GNOME may breath new life into addressing some of its ill's...one of which is listed below...
3) The Object Oreintation Thingy. I am really sorry if a lot of the GNOME developers think OOD when it comes to the GUI apps is so passe' I think GObject library is a throw back to the stone age, personally. I mean for Christ sake, if your going to reinvent the Object Oreintation of your GUI framework just because you cannot/do not/will not learn C++, you get the build complexity we keep reading about that is killing the GNOME release cycles.
This is a CLUE: Adopt, understand and learn how to build a OOD/OOP conceptual framework for your interfaces and DUMP GObject. Stop reinventing what C++ already gives you. With that RANT I present Exhibit A:
#include
struct GTypeModule;
struct GTypeModuleClass;
gboolean g_type_module_use (GTypeModule *module);
void g_type_module_unuse (GTypeModule *module);
(ad naseum)
I really FEEL for you if you have to deal with the kind of crap above.
4) Finally, though I am not a GNOME fan by any means, I would hate to see the distro's...drop GNOME. It is too early to decide on a Linux Desktop architecture, per se, because there are not enough mature options out there. If you cut too many options out too early you kill a lot of innovation. That is something I feel will happen if distro's start telling people we are not supporting GNOME, if you want it go somewhere else and get the RPM's....and GOOD LUCK! We need options to fight Microsoft when they start excercising their massive patent portfolio. Which IS going to happen by the way when they start running out of money....which won't be too far off into the future. Most American companies in the software biz can't innovate their way out of a paper bag, so expect Microsoft to radically step up the Patent attacks in early 2006.
Don't ask how I know that year either.
I won't tell.
-hackus
Got Geometrodynamics? Awe, too hard to figure out? Too bad.
I use packages that require GTK and KDE. My single biggest gripe about KDE and Gnome is that for me to function, I need 400megs of crap if I want to make sure I have a good foundation for me work. This is just retarded. Now is the time for all good distributions to merge for the sake of the open-source community. Both packages are excellent. Time to make the community more mean and lean, I don't care if it is knome or gde, just pick a fricken API.
it is also the reason why there is a GTK-QT engine and no QT-KDE engine.
When writing the GTK-Qt engine, I actually found Qt's theming system far more flexible than that of GTK. Your "reasoning" for why there is not a Qt-GTK engine is rubbish. I have yet to see a GTK theme that can beat a Qt theme in terms of rendering speed or appearance.
-- Wibble
The logic, if you had read the article, is that Gnome is a nightmare to package, especially if you happen to be the sole maintainer of an entire distribution.
Have you ever personally built Gnome 2.x from source tarballs without problems? Have you ever successfully changed the target install directory, so that making a package (tarball, rpm, whatever) is easy? And that's not even counting the new libraries popping up all the time, often with undocumented dependencies. And then there's miserable pages like this, which have the basic list of dependencies, but only provide links for 3 of them.
By comparison, KDE is simple to build. It's just a dozen or so source tarballs, all of which do the "./configure ; make ; make prefix=/temp/package_to_be_tarballed install" thing quite easily, without major dependency issues. X.org or XFree86, QT, and a recent XML2 library are all that's needed, last I checked.
Slackware dropping Gnome has very little to do with how the two desktops compare when being used, and everything to do with how they compare when building from source. If this alleged email from Patrick is true, then it just means that he's sick and tired of Gnome's chaotic, maintenance-intensive mess of libraries. I don't blame the guy.
define "important widgets"
What's important to you might be irrelevant to someone else, and what's useless for you might be used every day by someone else.
Microsoft learnt that the hard way with the idiocy of their hidden menu options in Office 2000.
That doesn't mean that there isn't room to improve things - there definitely is - but just ripping out half the UI doesn't solve anything. One of the main goals behind KDE has always been that there are NO hidden options (as in not exposed somewhere in the GUI). If you ever have to edit a config file - or launch a generic configuration application that is nothing more than a thin wrapper around directly editing a text file, then it's a bug.
Also your comments about konqueror kind of show that you've never really used KDE, or you'll never like it.
You're seeing Konqueror as two different applications crammed into one. But it's not. Tt's a universal browser and viewer via embeddable parts and pluggable protocols - which enables it to handle filesystem browsing and management as well as web browsing as just two of the many things it can do - and all by simply providing a light framework for other parts to do the work.
If you don't agree with that approach, you'll never like KDE because it's fundamental to it.
Advanced users are users too!
I was actually willing to mod you down, but as I didn't feel ignorance as a fair reason (and neither in the modding list)...
BenjyD refers Slack as a one-man distro just because Pat created it and is mostly its only maintainer and official packager. When Slackware was supported by a CD distributor (it was Walnut Creek?) he had a few lieutenants, but I guess he currently does the job alone.
On the other hand, I guess Slackware ALWAYS has multiuser, as Linux by design always was, and only in the very early days the Linux init sequence was just to start bash (and that doesn't means that multiuser capability wasn't quite there).
I didn't see that, but I'm booting Slack since 1995 and I never heard anything such a non-multiuser Linux distro, besides those end-user oriented new distros as Lindows.
Got Pike?
Does anybody else here refuse to use KDE simply because of its retarded naming scheme?
Did you mean "retarded", like:
* gnibbles
* grip
* gaim
* gnome-about
* gnome-bug
* gnome-calculator
* gcalctool
* gnome-character-map
* gnome-desktop-item-edit
* gnome-dictionary
* gnome-dump-metadata
* gnome-font-install
* gnome-gen-mimedb
* gnome-gtkhtml-editor-1.1
* gnome-keyring-daemon
* gnome-moz-remote
* gnome-name-service
* gnome-open
* gnome-panel
* gnome-panel-preferences
* gnome-panel-screenshot
* gnome-print-manager
* gnome-pty-helper
* gnome-search-tool
* gnome_segv
* gnome-stones
* gnomevfs-cat
* gnomevfs-copy
* gnomevfs-info
* gnomevfs-ls
* gnomevfs-mkdir
* gnomine
* gnotski
* gimp
* gimptool
etc., etc.
I love the smell of flaimbait in the morning...
The issue here is that getting Gnome built is a headache that Pat finds onerous given that he is known to prefer KDE, and while Todd is happy to distribute Dropline Gnome, Pat might be excused for not wanting to duplicate the effort.
If the biggest thing you can find to bitch about is whether all the names start with a G(nu) or Gnome vs. K(de), then I'd say both desktops have succeeded beyond their wildest dreams.
Personally I use both, but I use Gnome for my personal account. GTK is cross platform; so is Qt. My guess is Qt might be better for Windows porting, but as far as Linux itself goes I don't really see much difference. In both cases I just configure until it works the way I want.
Programming is another issue, but I haven't done enough with either to say which is truly "better", and it would just be my personal opinion anyhow. After working with 2-3 other GUI toolkits over the years, I realized they all basically work the same, some just have a cleaner programming interface or more default/standard widgets.
The whining about package dependencies is just that -- whining. Go ahead and try and install something that requires IE components under Windows and see how far you get if you manage to remove IE. The same goes for Gnome's "Bonobo" CORBA support or Qt under KDE. If the package was built with particular software in mind it will need to have it installed.
Or is everyone going to start crying about all the HTML display components that require Mozilla as well? Perhaps you'd like to get rid of glibc because you like another ANSI C library?
Wah.
Wah. Wah. Wah.
I do not fail; I succeed at finding out what does not work.
Actually I'd argue that we do.
It doesn't need to be unified but it does need to be standard, that way we're all on the same page, which cuts a lot of redundancy out of writing consumer level books and tutorials. That will help Linux move into the desktop. When someone says it should look like this it should, rather than the author having to give 10 examples of how it might look and finishing with "Check your documentation" at that point novice users put it down and go buy Windows.
I also think that by having a grossly popular desktop more gifted developers can focus on more than one project, rather than having to worry about being a GTK or QT expert they can just learn whatever everyone is using and there by make software easier, that's the number one reason Windows even took off in the first place. This would mean when someone makes a good calculator we can call it calculator and not gtkalc or Kalculator or something. I'm not saying variety doesn't have it's merit but standardization has huge merits aswell
I might get mod down for this... but here it goes.
My company recently made the switch to Linux, replacing most of our Windows desktops with Linux (servers are all already *NIX).
I was invovled with the project since the planning stage, and everyone seemed to agree that GNOME was the best choice because at the time (and it might still is), GNOME was the default desktop for most commercial distros. We thought to ourselves: "Oh well, these guys must know something that we don't." Most of us ran KDE, we gave GNOME a small test drive, decided that it looked easy enough and voted for it.
Big mistake.
First of all, GNOME lacked documentation on how to customize it. For gconfd, the GNOME web site only provided 2 links, one of which is dead, and the other was last updated in the year 2000. I asked around on IRC, posted on forums and newsgroups, emailed the GNOME developers, but I did not get any responses. I ended up taking apart all the %gconf.xml files myself, and saving a profile and writing an ugly script to convert it for every user. I am sure there is a better way, but either no one has done it, or nobody cared to share.
What's worse, are the bugs. There are minor bugs that really put a dent on the overall Linux experience, especially for those users that we just switched over. Some of them have already heard about how great Linux is, and how "stable" it is. This only makes them angrier when their Nautilus window craps out and leaves them a core dump (shows up as a little bomb). I looked up some of the bugs, most were already filed, but none fixed. Just a little while ago, there was an email on the nautilus list asking people to help fix bugs, so I think some of the developers agree with me that there are way too many outstanding bugs. When I asked some of the GNOME developers, the response I got then was to "upgrade to 2.6, it is much better than 2.4!". Sounds familiar? Yup, Microsoft told me the same thing.
The similarity doesn't end there. I installed 2.6 and tested it. In my opinion, it was worse. Yes, the spatial view is kind of cool, but you know what it reminded me of? Windows 95. And there is no easy way to turn it off (I would have expected to have it as an option in the drop-down menus). It was not more stable either, but I WAS running an early build of it. I, again, complained to some people about how 2.6 did not quite live up to my expectations, and the answer? "Wait for 2.8, it's GREAT!"
All of this is not helping the Linux desktop movement, especially in my company, where the management was already not really happy about switching over to an "inferior" OS. This just gives them more "evidence" to talk about: "We were right. My WindowsXP box crashed much less often. Linux IS a piece of crap!" But in reality, it was only Nautilus that was crapping out when connecting to a WebDAV mounted drive, not the underlying OS... but they won't understand that, would they?
What we need is a grand unified desktop API. One where I can call "createIcon()" or "queryIcon()" or "deleteIcon()", etc., to add, query, delete, or otherwise manipulate the user's desktop(s). Trying to support KDE 2, KDE 3, Gnome, and any other potential desktops is impossible. We have a "create icons" tool for our (commercial) product, and of those who have owned the tool, one was fired, two were laid off, and the latest just quit, all in the span of 2 years. That's actually two independant statements, completely unrelated, but it is an interesting fact to me :-)
In short, a common desktop API would be incredibly useful. From a purely commercial standpoint, it would be just as useful to have only one Linux desktop. Personally, I'd love to see the opensource competition that drives each project to become better, but there does need to be some co-operation, just like OOo and KOffice and others are standardising on common XML document formats, making it easier for not only document interchange, but for others to write to the spec. We need that programmability for the desktops, too.