Do Interoperability Technologies Help or Hurt KDE Development?

In a recent story at, the author considers the implications of some recent developments announced on the dot -- namely, XParts and QGtkWidget. These projects permit non-KDE programs -- particularly Gtk applications -- to be used within KDE applications. The author opines that these projects will end up hurting KDE development. In particular, the author wonders why a Linux developer would now develop for KDE when they can write for Gtk and have the application work both under KDE and GNOME. What do you think, will these projects encourage Gtk development at the expense of KDE/Qt development? Even if so, does it nevertheless make KDE stronger as a competitor against proprietary desktops? Is it important for KDE developers to create projects that would make KDE apps work inside GNOME apps (such as a KParts container that works with Bonobo) or with Gtk widgets (essentially a GtkQtWidget)?


It will hurt KDE development. This is exactly why Micro$oft does what they do. They know that people will develop for the largest possible audience. As such they make it as easy as possible to port your code from another system. Ie. Porting from Mac has tons of assistants and wizards to help get your code to work on M$ OSes.

The better approach would be to allow KDE apps run in GNOME land but not the other way around. That way people see the best that KDE has to offer and then "lure" them over to KDE.

If you look at (yes I hate to bring them up again but they have gotten where they have because they have been smart) M$. They license their technology very limitedly to Unix to basically encourage folks to come over to M$. Not for M$ developers to go over to Unix. This kind of mentality needs to be brought in for the KDE official approach.

Now the great thing about Open Source and this kind of development environment is that someone else will go and make these amazing products such as XParts & GTK widget containers. However these should not be "officially" supported apps. That way its not something that developers heavily depend on. They then know also that support as KDE is developed is entirely up to that third party and compatibility is not a priority for the KDE core development team.

All in all everything here has been fantastic work. I don't want to discourage it but we have to be prudent about what is done for the greatest good while not impeading anyone else.

Yes, but the big difference between the Windows/M$/etc. world and the Free Software world is this: MS is out to compete with and destroy their competitors in order to further their commercial success. Those are their goals. However I believe (idealistically, I guess) that the goals of free software developers are much closer to "make the world a better place by writing good software for everyone, that is free (speech/beer)".

So in light of what KDE is doing, I think they're doing the right thing. Sure it may or may not have negative effects on their mindshare, but is that really important? If they're improving computing under Linux/BSD/* in any way, then in my opinion it's great. We don't need KDE vs Gnome flamewars. We don't need crazy competition that devolves to the extent of what you see in the commercial software world.

It's actually very refreshing to see the developer who did this, take a step back and realise that free software development is about more than 'my project is better than yours and we don't want anyone to use yours', but to actually look at the bigger picture and work towards something positive for everyone.

by Maarten ter Huurne (not verified)

When Open Source people mention "world domination" as a goal, don't take it too seriously. As I see it, the goal is not to have KDE installed on as many desktops as possible, but to make KDE the best desktop possible both for the users and from a technical point of view. So whether or not interoperability technologies help KDE to dominate the Unix desktop is not important.

Microsoft's tactics may be a commercial success, but it hasn't made them popular, especially among Open Source developers. I wouldn't recommend the way Microsoft treats their competitors as a rolemodel to anyone...

Yes, but on the other hand helix code tries to dominate
the unix desktop by declaring gnome as the
standard desktop. And, because, in my view, kde is technically superior, why should the kde crowd not
try to turn the thing round and use some tactics
to make kde the standard desktop ?

by David B. Harris (not verified)

I'm sorry to say, but you've got a lousy alternative.

From a business perspective, Microsoft's attitudes and actions are pure gold. They're very good at what they do - killing the competition.

However, when you say things like "we should let KDE apps run under GNOME, but not the other way around", what are you doing?!? Everyone who wants to use a full KDE desktop is immediately locked into KDE-only apps. That's absurd. I mean, no offense, I like KDE and all, but KMail just doesn't cut it. It's close, but I use a GTK+-based app(Sylpheed) which has just the features I need, and no more. Now, you'd want me not be able to run this under KDE? How does that help me?

No, to suggest such a thing means that instead of getting ahead by being good, you'll get ahead by making the other guys bad. That's unacceptable, and whenever I see that attitude around me, I make sure to change it. Real Quick(tm).


I can't agree with you in any point. It is not a problem for KDE when there is a better compatibilty with the Gnome Desktop. Perhaps it will bring some people to use gtk instead of QT but where is the problem? The Unix Desktop will never be accepted if there there are two incompatible desktops. But if you can use the same apllications in any desktop system you can decide which to use and don't have any disadvantages because there is one little program that is only availabl for the other desktop system.
This should be the futury: There are two compatible desktop systems and everybody can desides which looks better for him but this decision will not bring any disadvantages! This will beat Micro$oft because the only advantage for beginners is that there is only one deskop system. But I prefere the freedom to choose my system.

by Martin Andersson (not verified)

I don't think any of the c-coders, writing gtk apps, could (or would like to) handle QT without relearning almost everything... This makes it hard to get the gnome/gtk coders to start writing pure kde programs, but with the QGtk classes more programs, that was limited to gnome, will now run as native kde apps and I think that helps kde to be a better and wider desktop..

??? but the qt/kde libraries are much easier
to work with than gtk. So what ?

by Bero (not verified)

I don't think this type of stuff can hurt KDE development.
Looking at the APIs of Qt and gtk, I can't see any developer switching from Qt and gtk by his own choice.

Having GNOME land embedding KDE applications would IMO not be a very good idea for KDE's spreading (though it would be a good idea in general, our enemy is/should be Microsoft, not GNOME) - if people can just run KDE applications natively, where's the reason to switch?

This is nowhere near the M$ tricks because the limitations are missing.

why a Linux developer would now develop for KDE when they can write for Gtk and have the application work both under KDE and GNOME.

because he thinks the kde framework is better?
this is open source development, and the kde team and the gnome team are willing to reach the same goal. No need to be afraid of each other...

by the way, gnome apps already run in kde, but now they can be embedded in kde: that brings more tools to kde users, so this is a Good Thing(tm).

I agree with you :) It shocked me when I first looked at KDE2 source. Amazing - I could *read* it! :)

*THAT'S* why I'll be programming for KDE2. The widget set is all right(nothing amazing), speed is acceptable(not as fast as GTK+ 1.2.x, though), but damnit, I can actually code for it! :)


Not to sound trollish, but people will develop for QT because it is a superior C++ OO toolkit. I don't even want to *know* how ugly GTK apps look embedded in QT programs.


they would indeed look pretty ugly. the GTK+ apps would look beautiful on their own, but if you try and embed them in an ugly QT app, Lord!

Did you ever look into qt ? Doesn't seem so.
gtk uses similar technologies as Xt/Motif, which is from the 80's. I started to look into gtk first, but when i compared it with qt, i didn't bother with gtk any minute longer. Not more than a disturbance to me.

First off,
1.) Asthetically speaking; GTK+ (IMHO) is far more cosmetic then QT.
2.) QT is only a superior C++ OO Toolkit over GTK+ because GTK+ is done with C.

What's with that "iguy" character and his ("The better approach would be to allow KDE apps run in GNOME land but not the other way around. That way people see the best that KDE has to offer and then "lure" them over to KDE." and "yes I hate to bring them (them == MS) up again but they have gotten where they have because they have been smart" END_QUOTES) comments?

Now, not to sound like a prick here or anything, but are all KDE users gnubies now or what? Open Source isn't about capitalism, it's not about strong arming the competition, open source is a collaboration of developers from all over the world, not market share.

That's my 2 cents anyway.
Thanks for reading ;P
[email protected]

are all KDE users gnubies

A large portion of us, yes. KDE makes *nix fun and easy, after all. I think
you're right that people often forget about the community concept of free
software because they're not used to it, but please try to be gentle with
"gnubies". A little time and some experience with these new ideas are required
before they sink in.

Now for my 2 cents on the topic... I think this new ability will be good for
kde. Most free software developers code for enjoyment and those who choose kde
do so because the kde/qt style of coding appeals to them. Also, for those new
to the scene and trying to choose a platform, this could make kde more
attractive because it means that developers who code for kde will have a huge
number of components to reuse, including gtk ones. And of course, gnome will
continue to be a fine choice as well. Those who ultimately choose gnome may
even be encouraged to play with kde, knowing that they can reuse stuff they've
already written. The experience they gain from doing so could bring more cooperation to the environments. So really, I see this as having the potential to benefit a
lot of people without hurting anybody, which is great!


I am gentle with gnubies. :)
Everyone is a gnubie at some point in time, right? But come on now, anyone who states that one open source project should somehow have an advantage over another, at the other projects expense; should know better.. I mean, if I wanted to run software developed by bigots, I sure as hell wouldn't have started using the open sourced *nix's in the first place.

I meant no harm with the gnubie comment, I was just pointing out the obvious, and I hope iguy reads this and realizes that that's not what "open source" stands for, and if he does read it, and he understands it, then my job is done. ;]

Cheers all! Happy New Millenia

Fair enough. Happy new millenium to you too!

I don't really think this is the issue at hand between Gnome and KDE. Seeing as IBM, Solaris, Easle, Helix Code, are all companies that are supporting Gnome. The issue that Star Office will be ported as a Gnome office tool. This is really a shame because KDE is hands down the better looking of the two.

Whatever the case, if someone decides to code for GTK it should run in KDE, and vice versa. If a user decides to use KDE or Gnome, the user should be able to install a base RPM or DEB that would allow them to run the other desktop's applications. KDE apps should run in Gnome, and Gnome in KDE.

It's unfortunate, but I'm pretty sure that Star Office won't be ported to a native Linux toolkit(like Enlightenment's, or GTK+/GNOME, or QT/KDE). Apparently, from what I've heard, they'll be porting it to Mozilla's XUL.

Yup, you heard right. A slow Office app, with a slow toolkit, in a slow graphical environment(sorry, Xfree guys! I appreciate all the hard work you do, but it's not as fast as some of the alternative - but really, thank you so much! :).

Yucky, eh?


Actually, almost an equal number of companies have joined the KDE League, some are even in both! Company backing probably won't affect it very much, these companies are using this as a way to get involved in Open Source, so that in 5 years they can say "Hey, eveyone look, we were part of developing this!". Besides, Solaris (er, I assume you mean Sun Microsystems) has a nasty habit of making software that is good, but rather clunky. Besides, KDE has theKompany. Also, KDE apps already run in GNOME, and vice versa. The problem is that GNOME apps dont use DCOP, and KDE apps dont use CORBA, thus rather limiting the interoperability of GNOME and KDE apps.

KDE interoperability technologies just continue the tradition of showing how inferior other technologies are compared to KDE. Qt has shown this with Motif and Windows, GTK+/GNOME is logically the next.

Gnome/GTK+ isn't that inferior.
In fact, the current releases of Gnome/GTK+ are in some ways still superior (although not all).
And the development version can really match KDE/QT.

Nope, it can not. gtk tries to solve an object-oriented problem area (ui programming) with a procedural language (C) instead of an object oriented language (C++). This may work for small programs, but the experiences with Motif showed pretty well that it doesn't work well for any bigger project. Believe, i've programmed Motif for 8 years. And C++ Wrappers around wanna-be object-oriented C-based frameworks showed to be quite crappy.
That's why qt/kde programs will remain faster and more reliably than gtk programs at all times.

by Jacob R. Kreutzfeld (not verified)

Nice said, but what is it what makes QT/KDE so supperior? I always read how supperior KDE is to GNOME, but I never see any evidence. What killerapplication can you offer. What technologie do you have I would not like to miss. Could you please explain your stupid rant with some facts? Otherwise people will easily realize that you are just talking nonsense.

Jacob R. Kreutzfeld

Are you a developer? Have you looked at the source for Qt/KDE or GTK+/GNOME apps? I am, and I have. Qt/KDE is a lot more advanced/modern from a developers perspective. GTK is messy, and hard to understand, as they are typically written in C using odd OO models. Qt is the nicest GUI dev kit I've ever used. Plus, KParts is way cleaner, leaner, and just easier to understand than Bonobo/CORBA or OLE. What about KIOslaves? Konqueror is the anything browser, and can literally browse any service, protocol, or device if someone takes the time to write an IO slave.

Are you even using KDE 2? If so, you should notice some huge improvements over GNOME. For one thing, it's fast. It starts up quicker than KDE 1.1 or any version of GNOME. It also has a much smaller memory footprint. An how about Konqueror for a killer app? It's by far the best browser for the Linux platform. Once they improve Java/JavaScript support, I will remove Netscape and never look back. How about KOffice? KWord, although a bit young and lacking a few features, loads up in little more time than it takes me to load a terminal. It also does an admirable job reading those accursed Word documents. What does GNOME have? StarOffice? Give me a break! SO takes about 2+ minutes to load on my 600mhz PIII. It also likes to fill the whole desktop, acting like a window manager of its own.

I strongly disagree with you.
I have written tons of GTK+ and Gnome programs, and GTK+/Gnome definitely isn't a big mess.
It's a very good environment to program in, and certainly not difficult to understand.

I have to agree with Aaron since I did the same.
While Gnome was my primary desktop (I did not like the look of KDE1).
I had a look at KDE /Gnome and I came to the conclusion that KDE/Qt
is from a programmers point of view a lot easier. (And yes, I had a
look at gtk-- which is incomplete and not very well documented).
So while GTK+ might be easy for you, you should at least have a look
at KDE/Qt (assuming you know some C++) and probably you will find
out that KDE/Qt is easier to program :-)

Happy new year, Marco

I have looked a few times at KDE/QT source code.
They look a little cleaner than C code, but that doesn't make GTK+/Gnome a complete mess.
Have you ever looked at gnome-libs' source code?
One of the cleanest code I've ever seen!

I bet you have never written a line of GTK+/Gnome code and never seen any Bonobo/ORBit code.

So what? Nautilus can do the same.
KIO and kioslaves? I bet you've never heard of gnome-vfs and it's modules.

KDE 2 is not faster than KDE 1. It's even slower.
Gnome and KDE 1 are waaay faster than KDE 2.
Smaller memory footprint? Yeah right, in your dreams.
KWord? Ha! Pratically unusable.
You might lauch at StarOffice/OpenOffice, but you can't deny the fact that it's better and going to be *much* better!
SO takes about 45 seconds to load on my P166.
KOffice takes about 1 minute.
And OpenOffice is going to be *much* faster!

KDE might be better in some cases, but definitely not far superior!
If you compare KDE's cvs version with Gnome's cvs version, you'll see that Gnome is a pretty good match.

> And OpenOffice is going to be *much* faster!

Don't want to disturb your dreams, but concerning
OpenOffice there was an article in c't
(an excellent German computer magazine) saying that

  • a lot of code is not very well documented
  • most of the comments are written in german
  • there is a general lack of interest
  • the whole package has about 10000 of files and 9 Million LOC

And even if you are not a programmer you should know that
millions of line of code with almost no (readable) documentation
is no fun to code. So please wait for your OpenOffice but
don't be disappointed if they need more than 2-3 years to
break it in smaller pieces. In the same time KOffice as well as
the (old?) office package are more stable and faster than
the star office package. IMHO the gnome developers should
forget the OpenOffice stuff and continue coding the good old
office package which looks more promising than the Sun stuff.

Happy new year, Marco

by Jacob R. Kreutzfeld (not verified)

Well, I've written programs for GTK+/GNOME and it was a lot of fun. The API is neither messy nor hard to understand. It would be nice if there were full C++ bindings for GNOME, but that does not imply technological inferiority.
Also could you tell me, what you did to make KDE2 start faster that GNOME1.2? On my machine GNOME still starts faster than KDE2. By the way a fully bonoboized GNOME app needs about 5 secs on my P233 but KOffice apps need about 20 secs. Both are tested without any GNOME or KDE libraries loaded before. So how do you see, that KDE2 starts faster than GNOME. Still I don't see, what speed's got to do with superiority.
Do you seriously want to compare Gnumeric with KSpread, Abiword with KWord, Evolution with KMail or Dia with ? (GIMP). You would terribly loose. I haven't seen any KApplication witch appeared to be so usefull, that I needed it. However I use lots of GNOME apps every day and wouldn't like to miss them.

I think you must be dreaming that KOffice is a usefull tool, KDE is faster than GNOME etc.

Jacob R. Kreutzfeld

Ummm... how can you compare a bonobo app to Koffice? You should state exactly what app you are referring to when you compare it to Koffice. For me (on a P233), Konqueror loads up in around 5 seconds. Loading up a new Konqi window is less than a second. Also, what koffice app took 20 seconds to load, surely you are doing something wrong there? I would say it takes about 10 seconds to get a koffice part up. But like you say, load speed is not amazingly important, although it does make people happy if their apps load in less than a minute.

Apart from Gnumeric and Evolution, the apps you mentioned are barely GNOME apps. I mean, Abiword? It doesn't use bonobo. All the koffice programs allow embedding. It works too. Seamlessly. Just because you don't find Koffice a good tool, doesn't mean others don't. I think Kword is useful (though buggy), Kspread is useful and getting a lot better very quickly, Kmail is not part of koffice but all the same is great mail app (Aethera (Magellan) is the one you should be comparing with Evolution), kivio is an awesome diagramming program, Kpresenter is excellent (best presentation app I have seen that's opensource), Krayon is making excellent progress etc. YOU must be dreaming that koffice is not useful.

by Jacob R. Kreutzfeld (not verified)

Well, I'm using the packages from unstable debian. If I start KDE first, then KOffice apps indead need about 5-10 secs to start. If I don't start KDE then it takes about 20 secs.

Abiword doesn't use bonobo jet but it will very soon. Besides that is Abiword regarded as full GNOME-app if you compile it with the GNOME-flag. The point about bonobo is, that there is no stable release jet. As soon as a stable bonobo is out you will see lots of bonobo support everywhere. Many apps already use it.
Anyway KOffice can hardly be regarded as stable (I mean stable). KOffice apps crash left and right. Most GNOME-office apps are more stable than KOffice apps even though thereis no stable release jet.

Jacob R. Kreutzfeld

I belive that technologies like Kparts and QGtkWidget can hurt comercially, but as it is, neither KDE or GNOME/GTK are comercial. I think it's great to have this sort of functionality, it just allows more things to be available.

I primarily use GNOME, but I always keep up to date my KDE packages, and I reasearch and try out every technology I find interesting, in both KDE,and GNOME.

The KDE vs. GNOME wars should be nonexistant - Having more than one enviornment is a good thing, it allows choice for a user. And interoperability between the two only makes it better.

Both KDE and GNOME copy off each other in some areas, but that's good- if it works far better than anything else, is there a good reason NOT to copy it? I will always support both KDE and GNOME in everything they do, because choice is good. Interoperability is better.

One thing that I would like to see happen between both projects is the component technology. I think having both enviornments share the component technolgy would make things alot better for sharing code, etc. When I heard about the rumor of KDE adopting Bonobo as their component technology, I was very excited to hear that, unfortunately it wasn't true, but I really wish it would have happened. Bonobo makes sense for both envirnments because it is not toolkit dependant (as far as KParts, I have read that it does depend on KDE libraries-- correct me if I'm wrong), and would only be good for both.

Gnome has support for KDE's "task tray", although not perfect, it allows a GNOME user to use a KDE application in a GNOME enviornment without trouble. Why would a GNOME user want to run a KDE application, why not just use the GNOME equivalent? Easy, the KDE program may work better for that specific users needs. In my case before I found out about GNOME's rp3 dialer, I always used kppp, because it worked better than anything GNOME offered. Even today while I use rp3, there are many things I would like to see implemented in it that kppp has always had.

These are my thoughts (and more) on the subject. Please don't turn this into another KDE vs. GNOME war, as this was not my intention. I only wanted to explain why interoperability is a good thing between KDE and GNOME


I strongly disagree, because it is a one-way street. While the developement of XParts and other methods to allow "foreign" embedding is a strength of Kde, there is no evidence that Gnome has any interest whatsoever of providing ways to embed kde components in Gnome. Even if Kde had adopted Corba instead of Kparts, Gnome would make interoperability more difficult than it need be.

Gnome most certainly is a commercial project. Helix, RedHat, and Eazel are entirely commercial, regardless of GPL. Kde has some commercial influence, but has a good mix of commercial and non-commercial participants at this time. In Miguel's latest interview, he proclaimed that "volunteers" can't handle a major project like Gnome's Evolution and that he prefers commercial management to the use of volunteers in Gnome development and promotion.

Still, with GPL, we can use code and concepts from Gnome and Gtk in our own endeavours. Recently I borrowed some really useful code from GPaint, and likewise Gnome and Gtk developers have borrowd from Kde (as in the gtk html component which was mostly copied form kthml version 1.x). But this does not mean that the Gnome project, as a project, has any interest whatsoever in reciprocating by making it easier to use Kde components in Gnome, aside from providing a place for the Kde menu in the Gnome Panel.

XParts is useful for a lot more than embedding Gnome components. Much also remains to be explored with Java and Kde, for example.

After time and time again giving the Gnome the benefit of the doubt, I've conculued that commercial interests (to recover the investment of time and capital in ventures like Helix and RedHat Labs and Eazel) totally drive the Gnome project. Kde is more laid back, and more free to experiment and explore. Kde's primary commercial benefactor, Trolltech, is now well established and probably relies less on venture captial and more on a positive flow of revenue. I really do feel that the Gnome organization regards Kde, not MS, as its primary competitor, and that the Gnome organization regards it as a life or death struggle.

This doesn't mean that efforts like XParts are in vain. They will prove very useful in the long run. But Gnome will cooperate only when it has no other alternative. This seems unlikely so long as Gnome and related commercial endeavours spinning off of Gnome which really direct Gnome's development like the tail wagging the dog remain on life support from major players like Sun, IBM and AOL.

A Gnome equalivant of QGtkWidget can't be done, because Gnome is written in C.
But Gnome's Bonobo architecture is capable of embedding KParts component, and it will happen in the future.

Gnome is *NOT* a commercial project!
While HelixCode and Eazel are commercial, they produce Free/open source software.
And if you still claim that Gnome is a commercial project, how does that make KDE not commercial too?
QT is created by TrollTech and KWord and KDEStudio by theKompany, both are commercial companies.
And who cares if it's commercial or not?
The people working at those companies are humans too, aren't they?

"Kde's primary commercial benefactor, Trolltech, is now well established and probably relies less on venture captial and more on a positive flow of revenue."

And what makes you think HelixCode, Eazel and RedHat aren't?
You are just trying to spread FUD, that's why.

"Gnome organization regards Kde, not MS, as its primary competitor, and that the Gnome organization regards it as a life or death struggle."

Yeah right. You keep beleving that kind of things, I won't stop you.

"But Gnome will cooperate only when it has no other alternative."

What about Mozilla? Gnome can use GtkHTML too you know.

Well, Gnome isn´t a commercial one but HelixCode is ;-) and in opposition there´s no KDE "for the poor" and a KDE "for those who pay", here there´s only one KDE and that is free and not developed by a company.

As far as theKompany goes, KWord ins´t developed by them (AFAIK) and KDEStudio is not an official part of the KDE distribution while KDevelop is.

Now, when it comes to the commercial extend, I do care. But there is a difference for me in a commercial desktop environment providing everything for daily work and a desktop environment that does the same but is completetly independent.

Third party apps that are written to earn money nowadays have to support one or the other desktop to integrate smoother into the user´s envirnoment and *both*, KDE and GNOME allow commercial development of applications and they will arise sooner or later although we try to provide something free for all needs.

Happy new year,


The software stays GPL'ed.
So anyone can grab Nautilus/Evolution/RedCarpet/whatever's source code and develop it further, even when HelixCode/Eazel doesn't support them anymore.

> Well, Gnome isn´t a commercial one but HelixCode > is ;-) and in opposition there´s no KDE "for the
> poor" and a KDE for those who pay", here there´s > only one KDE and that is free and not developed
> by a company.

Dude, what shit are you smoking?
HelixGNOME is free.
If you want to pay you can have it sent to you on a CD. If you don't want to pay, you download it, just like anything else. All developments made to Helix GNOME are put into the main GNOME source tree if they make sense (like it doesn't make sense to put the logos into the source tree.)

Stop talking crap about stuff you don't know.

I would like to clarify. theKompany did not write KWord and doesn't sell KWord, we simply made a committment to help support 2 developers working on KWord for 2 years as part of our committment to KDE.

KDEStudio is a great tool, we are have been responsible for it for almost a year now, but it is not a commercial product, it is strictly a free open source project.

And yes, we are human :)

by Anonymous Coward (not verified)

While it is true that the various GNOME companies do not have an interest in playing nicely with KDE, there is nothing stopping J. Random KDE or GNOME Developer from writing a compatability module. That's the whole point of free software, remember?

Who would want to use GNOME with KDE2 out? A retard?

by Maarten ter Huurne (not verified)

If the number of supported systems for an embeddable component is the only concern, a Java applet would be the ideal solution. Not only KDE and GNOME support it, but also plain X11, Windows, Mac etc. Ofcourse there are also downsides like the overhead of the virtual machine (especially startup time and memory consumption) and a widget toolkit that is either limited (AWT) or not responsive enough (Swing).

My point is that the reasons for developing embeddable components for KDE are the same reasons for developing for KDE in general. Many developers like the way KDE and Qt work and will prefer to write KDE components even if it means their components won't run under GNOME (yet?) without porting.

My point was never to say that we should stop this kind of development. However as a suggestion as an intelligent use of resources was my primary intent. (Really need to get more sleep)

The real issue, in my mind, as others have stated is the fact that this is Open Source product. The other developments such as XParts & QGtkWidget are fantastic. However until GNOME and KDE decide willingly to "get along", exterting energy by the core KDE team to work on this cooperation is a major waste of time, resources and skills in my book.

(Because I mention GNOME here this is not a flame fest against GNOME. I use both GNOME & KDE on a regular basis.)
The other point is everyone kept mentioning that this is Open Source, not commerical. Actually it is now. With the GNOME Foundation, including Sun and 1/2 a dozen other commerical companies, GNOME does have significant amount of commercial intrests at heart now. Granted KDE has commercial intrests also. But from an Engineering & Political standpoint I see it very difficult to justify making something such as these two tools part of the core development.

For this kind of effort to succeed there needs to be a political agreement between the two core development teams. A standard of some kind has to be agreed upon, be that official standard or verbal standard. Without this, the development done can be "contaiminated" and "corrupted" by either core team. What I mean by this is on version KDE you can embed Gtk Widgets. However Gtk v 1.2.3 comes out two minutes later. Now KDE which claims to be able to embed and use GTK widgets suddenly can't with all new versions that use the latest greatest GTK widget. I as dumb user comes along and tries to use program that claims KDE compatibility but it uses the latest GTK widget. I dump KDE (not knowing any better) and go to GNOME. However in my case GNOME is too complicated and I decide to dump Linux all together. This is not what anyone wants.

Okay.. off my little horse. I've dealt with this issue on other developments in my history. Everytime its been the same thing. If there isn't a political agreement between the two core teams doing that kind of cross development is doomed to various forms of failure. However third parties can succeed. That is my point. Let the third parties do what they want. Just don't make it something that the core development team needs to concern themselves with right now.

Oh no, not more Gnome Foundation FUD...


And if you don't believe me, read the Gnome Foundation FAQ:

Gnome trying to break KDE's interoperability?
That'll never happen!

It depends on how you look at it. In my mind, you shouldn't be asking whether or not it hurts KDE, but rather, whether or not it hurts the KDE user. Sure, you could easily "help" KDE by making KDE apps work in GNOME, but not vice versa, although I think that would only hurt KDE. By doing that, people would be more obligated to use GNOME because it runs everything (most everything), while KDE is still only holding 50%. I think what KDE needs to do is worry more about trying to support other stuff, like GTK, than trying to become more popular. With advanced GTK support built-in to KDE, you could easily swing the tides. The goal should not be to bring developers to KDE, but to bring KDE to developers.

I believe that this is a good thing. how could giving your project more functionality hurt it? on a personal dirty end user level I feel more inclined to use KDE because they are trying to give the users all the funtionality that they can pack into the desktop. KDE is an impressive Desktop and doing this will just expand it's appeal to a broader audience, which in turn will cause developers to program with the K desktop in mind.