Is the workforce split between Gnome and KDE hurting Linux?

[Castar]

Ok, no flames intended :D !!! It's just a thought. Even though having a choice is great and having two different desktops is fun and interesting, don't you think that if the two communities combined forces a stronger desktop with better applications could have emerged faster?

I was reading this post (http://www.ubuntuforums.org/showthread.php?t=159879) and kept wondering that if amarok is at the point it is now, why should people spend time to create the same program in GTK (Listen, Exaile) since amarok is developed at a great level? The same of course goes for other application pairs like Evolution/Kontact and so on. I mean, if all the people concentrated on one player, wouldn't the linux experience be better out of the brainstorming? Wouldn't new features be implemented faster than, for example, Media Player giving the Linux community the lead? I know that opendesktop.org tries to define some common standards and all. But if we had one desktop instead of two branches maybe things for the average linux user would be better?

ps: Please let's forget about the architecture differences and so on. Let's suppose that the two communities agreed on using either GTK or Qt or something else.

ps2: Choice is great and I recognise this. Please focus on the effectivenes :)

[bruce89]

Possibly, mabye that is what the Portland Project (http://portland.freedesktop.org/wiki/) is for?

[mscman]

A large problem with this theory is the different mindset between Gnome and KDE users (at least for a good portion of the development community). The point of Gnome is to be simple and not have tons of options, where KDE is supposed to be "fully customizable". Herein lies the problem: the applications are built to reflect the rest of the DE. If an application is built for Gnome, chances are it has less options and just works. An application built for KDE, on the other hand, is going to (generally) have a lot of tweakable options.

By having one program for everything, the choice goes away, and developers instead spend time arguing rather than making the changes to develop better software.

It's kind of like proposing that there be only ONE IM client, or only ONE web browser; people have completely different opinions on how a computer can and should work.

Feel free to disagree with me, this is just how I see it.

[eentonig]

The problem is that the Linux focus isn't on effectiveness.

1. Both Gnome and KDe are so far apart from a development point of view. That it would be highly inefficient to rewrite the code from one integrated desktop to another.

2. Linux is not about creating The single best Desktop experience. It's partly about creating the best desktop experience from your point of view. Hence the massive possibilities of modding and tuning.

Both of these points do have their benefit in giving us cutting edge technology on seperate fields od expertise, while at the same time inflicting a drawback on the overall development progress. But even that 's arguable. Linux is not one operating system for me. It's rather a large pool of Microsoft alternatives. You shouldn't compare Microsoft, Apple, Linux, ... You should compare Microsoft, Apple, Ubuntu, Debian, Suse, Mandrake, ...

And actually, don't compare at all. Choose the one that suits you best.

[Stormy Eyes]

Ok, no flames intended :D !!! It's just a thought. Even though having a choice is great and having two different desktops is fun and interesting, don't you think that if the two communities combined forces a stronger desktop with better applications could have emerged faster?

Don't try to herd cats. You'll just waste your time and annoy the cats.

[G Morgan]

I think it would be of benefit if both used the same toolkit and apps then essentially the choice of DE would just drop to a position of setting defaults. At the moment we have a strange position where KDE is/was probably the best DE but gtk is undoubtedly the standard for Linux apps. Of course the fact that gtk is the standard has meant that GNOME has caught up to make the variations pretty irrelevant.

So if KDE moved to gtk prehaps that would end all the problems in one. In reality though they are moving to qt4 so no progress there really.

[bruce89]

Xfce tries to be lighter than Gnome still. At the moment, however, there is no way to write a program for both Gnome and KDE, without writing the UI twice. That is what the Portland Project is supposed to do. I am not sure how they would get round Trolltech's licencing though.

[aysiu]

Any gain in efficiency you gain from combining efforts will immediately be lost by the time it takes all those involved to actually agree on something.

Imagine all the chefs in the world trying to combine efforts to create the perfect pizza recipe, and they all have to agree on that? It would take forever, and eating pizza would be boring--variety is half the fun.

If you read the history of KDE and Gnome, you'll realize how much they both need each other.

[Castar]

A large problem with this theory is the different mindset between Gnome and KDE users (at least for a good portion of the development community). The point of Gnome is to be simple and not have tons of options, where KDE is supposed to be "fully customizable". Herein lies the problem: the applications are built to reflect the rest of the DE. If an application is built for Gnome, chances are it has less options and just works. An application built for KDE, on the other hand, is going to (generally) have a lot of tweakable options.

By having one program for everything, the choice goes away, and developers instead spend time arguing rather than making the changes to develop better software.

It's kind of like proposing that there be only ONE IM client, or only ONE web browser; people have completely different opinions on how a computer can and should work.

Feel free to disagree with me, this is just how I see it.

I totally agree. In the amarok example, what is the point to replicate Amarok with both Listen and Exaile? I mean it's not even one but two GTK amarok clones :D ! Maybe choice is hurting the effectiveness after all, even just inside the GTK community? (Again, amarok is just the example)

[Castar]

I think it would be of benefit if both used the same toolkit and apps then essentially the choice of DE would just drop to a position of setting defaults. At the moment we have a strange position where KDE is/was probably the best DE but gtk is undoubtedly the standard for Linux apps. Of course the fact that gtk is the standard has meant that GNOME has caught up to make the variations pretty irrelevant.

So if KDE moved to gtk prehaps that would end all the problems in one. In reality though they are moving to qt4 so no progress there really.

Fortunately or unfortunately Gtk has become the standard since it enjoys most of the commercial support (not that it matters, still). I agree that the philosophy between the two is poles apart but in my opinion, one has to use the best tools he can get his hands on. In this case my personal opinion is that qt4 fits the bill. But that's only me :) and it mustn't be the automatic choice necessarily.

[bruce89]

My first Linux experience was an old Mandrake 7.2 CD that my brother left lying around, and both KDE and GNOME looked the same, and they looked similar to Windows. Now though, both are completely different, it's amazing how they have progressed in these 6 years.
(Nicked from http://www.ubuntuforums.org/showpost.php?p=1157466&postcount=20)

[bruce89]

Fortunately or unfortunately Gtk has become the standard since it enjoys most of the commercial support (not that it matters, still). I agree that the philosophy between the two is poles apart but in my opinion, one has to use the best tools he can get his hands on. In this case my personal opinion is that qt4 fits the bill. But that's only me :) and it mustn't be the automatic choice necessarily.
It has the commercial appeal because they don't have to pay licence fees for GTK+, but they would have to for QT.

[Castar]

It has the commercial appeal because they don't have to pay licence fees for GTK+, but they would have to for QT.

But since the "new" desktop would be open-source, why pay for royalties? The Qt is free for open-source, right? In any case, let's not stick to the mainframe problem. I'm completely talking about labour division.

[G Morgan]

Any gain in efficiency you gain from combining efforts will immediately be lost by the time it takes all those involved to actually agree on something.

Imagine all the chefs in the world trying to combine efforts to create the perfect pizza recipe, and they all have to agree on that? It would take forever, and eating pizza would be boring--variety is half the fun.

If you read the history of KDE and Gnome, you'll realize how much they both need each other.

While that is most certainly true you can use 1 toolkit and maintain choice. If KDE moved to gtk and ported all the apps you'd still have all the KDE mentality throughout and GNOME would still have its mentality in its apps. The benefit is that you only need load gtk so those you wish to fully customise by pulling in some GNOME apps and some KDE apps can do so without breaking the style and more efficiently.

So essentially you'd still want GNOME, KDE, XFCE and all the various WM's out there but all using the same basic tools and providing their philosophy within a more standard framework.

I'm not saying KDE should drop everything and start porting but it is something that could be encouraged over the long term (i.e. write new KDE apps for gtk and slowly port things over time, most people run 2 toolkits now so theres little loss in efficiency to do this in the time in between).

[tribaal]

Hum, diversity is a strenght, not a weakness.
That's my humble opinion of course, and some might call me a darwinist... but still :)

- trib'

[aysiu]

Good point, G_Morgan. One toolkit would certainly make a lot of people happy. One desktop environment wouldn't be such a hot idea, though.

[Castar]

While that is most certainly true you can use 1 toolkit and maintain choice. If KDE moved to gtk and ported all the apps you'd still have all the KDE mentality throughout and GNOME would still have its mentality in its apps. The benefit is that you only need load gtk so those you wish to fully customise by pulling in some GNOME apps and some KDE apps can do so without breaking the style and more efficiently.

So essentially you'd still want GNOME, KDE, XFCE and all the various WM's out there but all using the same basic tools and providing their philosophy within a more standard framework.

I'm not saying KDE should drop everything and start porting but it is something that could be encouraged over the long term (i.e. write new KDE apps for gtk and slowly port things over time, most people run 2 toolkits now so theres little loss in efficiency to do this in the time in between).

That could be an effective scenario in the long term! :cool:

[fuscia]

an actress once said to george bernard shaw "with my looks and your brains, think of the children we could have." shaw responded "but what if they end up with my looks and your brains?"

[bruce89]

QT for free software is under the GPL, but companies would have to buy a licence from Trolltech if they wanted a commercial applcation to use it. Most companies don't bother, and they use GTK+ instead, except from Google Earth. I think that one toolkit would be a good idea, especially since GTK+ is "more free". You are allowed to object to my last statement.

[Castar]

QT for free software is under the GPL, but companies would have to buy a licence from Trolltech if they wanted a commercial applcation to use it. Most companies don't bother, and they use GTK+ instead, except from Google Earth. I think that one toolkit would be a good idea, especially since GTK+ is "more free". You are allowed to object to my last statement.

No reason to do so. It has proved its effectiveness on many occasions :D

[G Morgan]

QT for free software is under the GPL, but companies would have to buy a licence from Trolltech if they wanted a commercial applcation to use it. Most companies don't bother, and they use GTK+ instead, except from Google Earth. I think that one toolkit would be a good idea, especially since GTK+ is "more free". You are allowed to object to my last statement.

It's certainly better than QT's free for one, expensive for the other plan. Personally I have no problem with LGPL for libraries, long term there will always be some proprietry apps and its better to make it so they can work with and support an open standard for windowing than to say hands off to them.

Of course apps are something else altogether.

[mscman]

I totally agree. In the amarok example, what is the point to replicate Amarok with both Listen and Exaile? I mean it's not even one but two GTK amarok clones :D ! Maybe choice is hurting the effectiveness after all, even just inside the GTK community? (Again, amarok is just the example)

I don't think you quite understood my point; I'm saying that having duplicate programs isn't necessarily a bad thing.

To put it in a way you can understand, look at cars. Some people drive SUVs because they like to sit higher on the road and feel safer. Others drive sports cars because they think they're cooler to drive (or maybe they like the wind through their hair :D ). And still others like to drive the little compact cars because they have better gas mileage and are more economical.

This is the idea behind multiple programs that have the same purpose. You can even expand this idea to operating systems. One system/program/car will not suit everyone's needs. Individuality is something that has to be accounted for in programming, as in everything in life. If all of the car makers decided on ONE style of car, made it ONE color, and said it would cost ONE price... I think you get the idea.

Now we can decide on the same tools to use; all of the car makers put an engine and seats in the car (at least I hope so...). What changes is the way the parts interact.

Just thought I'd clarify my position... :D

P.S. Great thread! I have to say that you are one of the first people I've seen that can post an idea in the forums and not take offense when someone disagrees. Kudos to yourself (and everyone else who's posted).

[Castar]

P.S. Great thread! I have to say that you are one of the first people I've seen that can post an idea in the forums and not take offense when someone disagrees. Kudos to yourself (and everyone else who's posted).

Thanks. I don't see the point in being offended for others' opinions, so why do it? :D

I understand that people need different things out of either cars/media players etc. but I'm starting to feel (and not only in linux) that too much choice can hurt :) . In this sense, I admire the Gnome philosophy because it tries to keep it simple.

[Sushi]

gtk is undoubtedly the standard for Linux apps.

It is? I must have missed that memo.

No, "split workforce" does not harm Linux. KDE-developers are Qt/C++-developers, while GNOME-developers are GTK/C-developers. GNOME-developers wouldn't be very good KDE-developers, and KDE-developers wouldn't be very good GNOME-developers. But as things are right now, both desktops benefiot the other, when they can look at how the other DE does things, and learn from it.

And even if the developers could develope the "other" desktop, there's no guarantee that end-result would be better. Apple has fraction of the programmers and resources than Microsoft has, but Windows is still crap, whereas OS X is miles ahead.

Would it be better if we pooled all the resources for "one desktop to rule them all"? What if the desktop they created weren't any good? We wouldn't have any alternatives, since we just pooled all our eggs in to one basket. And related to this: merging the two wouldn't work either. Toolkits are different, languages are different, philosophies are different. You simply CAN'T merge them. If you did, end-result would be a watered down version of the two. It would not have the simplicity of GNOME (so current GNOME-lovers would hate it), and it wouldn't have the configurability and features KDE has (so KDE-users would hate it as well).

[Sushi]

QT for free software is under the GPL, but companies would have to buy a licence from Trolltech if they wanted a commercial applcation to use it. Most companies don't bother, and they use GTK+ instead, except from Google Earth. I think that one toolkit would be a good idea, especially since GTK+ is "more free". You are allowed to object to my last statement.

I don't really see how GTK+ is "more free", just because it allows the user to create prorietary, closed-source apps. This is basically BSD vs. GPL all over again. Proponents of GPL favor GPL because companies can't use it and then close the source. It's similar in Qt vs. GTK. With GTK, companies CAN use it, and not contribute anything back. With Qt, that's impossible.

Over the years I have seen the Qt vs. GTK-argument rage on. And proponents of GTK usually say "but you can create commercial apps with GTK without paying anyone one dime". At that point I always think "when did Free Software movement move from creating great Free software, in to satisfying the whims of pushers of proprietary software?". And the question remains even today.

I think that Qt does have best of both worlds. If we had only GPL'ed toolkits, then we wouldn't have many commercial apps. But if we have toolkits like GTK, then we get commercial apps that do not contribute anything back. With Qt we get those commercial apps, AND the companies making those apps contribute back.

[Sushi]

I totally agree. In the amarok example, what is the point to replicate Amarok with both Listen and Exaile? I mean it's not even one but two GTK amarok clones :D ! Maybe choice is hurting the effectiveness after all, even just inside the GTK community? (Again, amarok is just the example)

Because developers of Listen wanted to create something, and developers of Exaile wanted to create something slightly different?`Maybe Exaile-developers disliked some things in Listen (or vice versa, I dunno which came first), and decided to creat his own piece of software that doesn't have those problems?

It's all about "scratching an itch". Exaile doesn't scratch the itch Listen-developers have, and vice versa.

[nocturn]

the goals of the two projects are too far apart to merge them. They can share ideas though.

[adkap]

Two display managers are better than one, but two display managers are also better than three. Competition spurs creativity, ingenuity and serves as an excellent motivator.

If KDE and Gnome combined, there would be only one major gui for Linux. No real choice if you want a snazzy ui. The dev's wouldnt really have to worry about what you want. They could implement whatever THEY feel is important without considering you. After all, there is no other option, right? It's the same story with Microsofts Windows. Windows has the market share, and Microsoft only cares about preserving that at all costs. Then look at the competition, Gnome, KDE, and OSX. With all three you see dev's listening to what the users say/want (though most of the time you have to speak loudly and repeat yourself alot. lol) and you see *real* genuine creativity. OSX is the perfect example of creative prowess that can result from competition.

[aysiu]

Too many chefs in the kitchen.

Reinventing the wheel.

You just need a good balance of those two cliches. I think we already have it.

[B0rsuk]

Thinking that it's possible to have just 1 window manager is very naive. There are multiple window managers precisely because people have different needs.

[kripkenstein]

Personally, I think the arguments about "KDE vs. GNOME" are hurting Linux, not the fact that we have both KDE and GNOME... but not the discussions about those arguments, like this thread, those are fine ;)

More seriously, though, I think that having 2 or more desktop environments is good and healthy for Linux. Sometimes code projects 'get stuck' - you can't tell in advance if a project will simply become unworkable, unimprovable in a way you didn't originally intend. Look at Vista, for example - what a mess. Even with all the money in the world the project foundered for years. Similarly, look at GNU/HURD - great idea, stalled for way too long. People say frightening things about the codebase of OpenOffice - if they are right, and the project can't move forward past some point - we will have AbiWord, KOffice, etc.

My point is that in the future, GNOME or KDE might run into difficulties - who knows what kind, any kind. If one project 'dies', then the Open Source community will have the other.

It isn't good to put all your eggs in one basket, and so forth.

[Castar]

I'm bringing this back from the dead, but I was reading this (http://thelinuxadvocate.blogspot.com/2006/08/kde-vs-gnome-why-are-we-still-fighting.html) and people are sooo divided on the issue.

Just like this thread :P

[mips]

I like Gnome for it's simplicity. KDE on the other seems more configurable and KDE apps also seem better. thats just my opinion.

[VitaminG]

Though I've been using Linux for a while now, I haven't really got into these types of debates until now. What I don't get is why DEs can't just have advanced and beginner modes. As several have said, aside from QT vs. GTK, the primary differences are different levels of detail in control. Many apps have varying levels of detail, some will use fewer, more general settings, while others will choose to set many, specific settings. Some will fall between these categories. We could simply ask for this setting at setup, and have it easily accessible for changes later. Then, base everything on one, extensible codebase, with extensible being the keyword. It would be a challenge, but the OSS movement has taken on bigger challenges and succeeded.

[synic]

I totally agree. In the amarok example, what is the point to replicate Amarok with both Listen and Exaile? I mean it's not even one but two GTK amarok clones :D ! Maybe choice is hurting the effectiveness after all, even just inside the GTK community? (Again, amarok is just the example)

I'd like to just say that Exaile started out mostly as a learning experience. I've found that the best way (for me) to learn a new language/toolkit is just to dive right in. My purpose was to learn python and wxWidgets. As always, though, whenever I do one of these "learning" projects, I end up implementing a feature that I can't live without. Exaile has been fun to write... and I think that choices are what Linux is all about.

Adam

[Qrk]

I think there is a size for open source projects beyond which they don't get much productivity gains per devoloper. If Linux only had one desktop with twice the devolopers as both KDE and Gnome... I think it would get pretty crappy pretty quick. Also, the friendly competition between KDE and gnome got us two amazing desktop environments. I like both KDE and Gnome better than the XP gui.

Also, if anything, XFCE has shown that there is room for a third, full featured, DE.

[DigitalDuality]

d

[Castar]

I'd like to just say that Exaile started out mostly as a learning experience. I've found that the best way (for me) to learn a new language/toolkit is just to dive right in. My purpose was to learn python and wxWidgets. As always, though, whenever I do one of these "learning" projects, I end up implementing a feature that I can't live without. Exaile has been fun to write... and I think that choices are what Linux is all about.

Adam

And it's a great player! I never said the opposite. Ok, this is different since you wrote it to learn to program the specific toolkit.

But, if you wanted a feature really bad why not implement it in Amarok? That's my point.

[argie]

I don't think the workforce split is hurting Linux. I understand that a lot of these people don't get paid for what they do. So forcing them to do something they don't like is sure not a good way to get stuff done.

Also, it's nice this way, by the general opinion, KDE - flashy, super configurable, Gnome - nice for a beginner, simple. There, you have choice, that wondeful word that's probably been used a 100 times in this thread already.