It depends on what you mean by faster. I use KDE 3.2-beta2 (on a 2GHz P4 with 640MB of RAM), and I find KDE to be much faster. KDE apps are definately slower to startup, but once they are running they are very fast. Mostly, redraw is a lot faster. For example, even in Windows, you'll often notice "expose lag" (the lag between when you move a window over another window and the window beneath repaints). In KDE, this expose lag is relatively rare (though, its worse in 3.2-beta than it was in 3.1, but 3.2 has a new window manager that is still in the process of maturing). In GNOME, the expose lag is very common. Also, when resizing windows, only the most complex KDE windows exhibit a lag between the window contents and the border. Meanwhile, even simple GNOME windows (like GEdit) will show the window contents lagging quite far behind the frame. GTK+ widgets also feel slower. For example, if you load up a few hundred MP3s in Rhythmbox, and try resizing one of list columns, the resize will lag way behind your mouse. In JuK (the KDE equivilent of Rhythmbox), the column stays right under your mouse as it resizes.
Also, don't use VMWare as a measure of the speed of the desktop. X GUIs are very dependent on the quality of the available drivers, and running with the ATI or NVIDIA binary drivers is a world away from running in the VMWare VGA driver.
Its a big strech to believe that a $1000 (actually, its more like $2000, but that's irrelevent) per-developer fee is going to affect commercial development on the platform. $2000 is peanuts to commercial developers. Even smaller companies happily shell out about that much for a copy of Visual Studio, and thrice that for a Rational Rose license. Given the choice, I don't think a lot of developers are going to forsake Qt and commercial support to save $2000.
If developers really don't want to pay the Qt license fee, they won't, and GTK+ will naturally become the most popular choice. But that natural selection isn't being allowed to happen --- Perens is choosing GTK+ because that's what his backers are leaning towards.
Kiosk is just the name of the system! It lets you lock down the system as much as or as little as you want. And it probably could be implemented in GNOME, but it isn't, is it?
This isn't a replacement, just another implementation of X. Not only that, but most likely it will be much more memory hungry and marginally slower (not taking into account GL acceleration). However, it will *seem* a lot faster, because it will explicitly synchronize drawing between apps and the window manager, just like OS X.
X was never slow. For the most part, KDE is faster than XP on my machine. X apps, however, have traditionally have done dumb things, and there have been problems with synchronization. Its these issues that are being addressed.
What technical reasons? GNOME was chosen at RedHat early-on because a lot of the GNOME developers worked at RedHatf. GNOME was chosen by Sun because the Sun UI people were more comfortable with C (that was their official reason). GNOME is on the verge of being chosen by Novell because they bought Ximian before SuSE, because SuSE held out for more money. And Perens doesn't even pretend that technical reasons were involved in his choice.
They can make closed-source applications, they just have to pay for it. If they're making for-pay software, they should be willing to pay for their tools.
That's all besides the point. Perens isn't even giving developers the freedom to choose between GTK+ and Qt for their apps. Perens is just deciding for them.
Don't know how it is in the GNOME camp, but KDE mostly uses auto-layout, so apps don't really specify how many pixels from the edge buttons are. KDE and GNOME are really a lot more different than people think. For example, a lot of people said that KDE should adopt something like the GNOME HIG. Well, much of the stuff in the GNOME HIG is not applicable to KDE. KDE's GUI framework operates at a higher level than that. For example, the XML-GUI mechanism automatically adds the standard actions ("Open/Close/etc") when appropriate.
When that happens, the FreeQt foundation gets together and decides whether the behavior constitutes a breach of the FreeQt agreements, and decides whether to BSD-license the library.
Also, and I'm not the first to say this, shareware is dead. It has been superceded by OSS software.
Heh. Button placement *isn't* a mundane issue. The GNOME camp decided to reverse the layout, and not use the "Apply" model anymore. I don't think that was a bad idea, but it has some major usability issues without having some sort of "undo" feature desktop-wide.
First, kcalc (on my 2Ghz P4) takes less than 2s to start once its in the VM cache. Second, startup speed is problematically slow on KDE due to problems with the linker. That doesn't affect the speed of other parts of KDE. In other respects, KDE is easily as fast as XP, even faster on a heavily loaded machine.
GConf is *not* easy enough to use to be a normal interface. There is a registry-style duplication of the key hierarchy that makes it difficult to find what you're looking for. Also, there is a lot of stuff in there that the user should not be messing with.
Bitch about politics and flamage all you want, and espouse the virtues of just "getting work done" all you want. That doesn't change the fact that throughout the history of computing, technologies have lived or died based not on technical considerations, but political ones like these. Too often, technically inferior solutions have been adopted because they were politically more viable. Us engineers would like it to be different, but such is the way of the world.
Not all MDI is created equal. Check out KDevelop3's IntelliJ-inspired MDI (which has been factored into the KDE MDI widget, so other apps like Kate can use it too!)
There has also been a shift within the projects. KDE used to be the newbie oriented/more restrictive DE. It used to have problematic license terms and be based on a non-free toolkit. GNOME, on the other hand, was built with a free toolkit and had power-user features.
Now, the roles have been reversed. KDE is fully-free, and has the power-user features. GNOME (in 2.x) has done a huge over-simplification and has alienated a lot of existing users (though, it got tons of new users attracted to the simplicity). Now, commercial companies appear to be using the LGPL-ness of GTK+ to push it as a proprietory-friendly standard desktop.
If anything, Slashdot has been remarkably consistent in its outlook.
Don't forget that Apple also has two different-looking L&Fs (Metal and Aqua) as well as legacy support for MacOS Classic. Also, Windows has at least three unavoidable toolkits (Explorer, MS Office,.NET).
The KDE and GNOME libs are both LGPL or BSD. KDE and GNOME applications are mostly GPL. GTK+ is LGPL, while Qt is GPL. That means that you can make closed-source GTK+ apps, but to make closed source Qt apps, you have to get a special commercial license from Trolltech.
If they win technically then they'll win in Linux. Don't forget that geeks are the Linux motor. >>>>>>>>> How long will that be true? Its true now, where KDE enjoys a slightly larger user-base because it is slightly better overall (better technology, worse interface), but with Sun pumping out hundreds of thousands of GNOME desktops, how long will geeks have any influence?
Don't we want competion rather than monopoly? >>>>>>>>>>> I don't want either to win. I think that the competition is a great thing for both desktops. But I fear that the commercial industry will force the situation. If KDE ceases to be a viable competitor to GNOME, because of the commercial factor, competition in desktops will be a thing of the past.
What's going on with GNOME appears to be a process of picking the standard. And true to fashion, the commercial software industry has decided to pick the lesser technology *again*, because its more convenient.
The basic problem is that Linux doesn't have to be any better than Windows to be sucessful in the commercial market, because of cost factors, so a "good-enough" solution is just fine.
Windows-style MDI is bad, but KDE doesn't do only Windows-style MDI. The newest version implements a mode called IDEAL (I think its derived from IntelliJ). Its a huge leap over traditional MDI.
I don't really understand the first part of what you're trying to say. My point is that half of all marriages in the US end in failure. Divorce rates among gays can hardly be much worse that that.
The technologies exist, but adoption is poor. When I see a KDE app, I know it'll support KIO and KParts. Even standard apps like GEdit don't support the corresponding technologies in GNOME!
Slashdot Mirror
It depends on what you mean by faster. I use KDE 3.2-beta2 (on a 2GHz P4 with 640MB of RAM), and I find KDE to be much faster. KDE apps are definately slower to startup, but once they are running they are very fast. Mostly, redraw is a lot faster. For example, even in Windows, you'll often notice "expose lag" (the lag between when you move a window over another window and the window beneath repaints). In KDE, this expose lag is relatively rare (though, its worse in 3.2-beta than it was in 3.1, but 3.2 has a new window manager that is still in the process of maturing). In GNOME, the expose lag is very common. Also, when resizing windows, only the most complex KDE windows exhibit a lag between the window contents and the border. Meanwhile, even simple GNOME windows (like GEdit) will show the window contents lagging quite far behind the frame. GTK+ widgets also feel slower. For example, if you load up a few hundred MP3s in Rhythmbox, and try resizing one of list columns, the resize will lag way behind your mouse. In JuK (the KDE equivilent of Rhythmbox), the column stays right under your mouse as it resizes.
Also, don't use VMWare as a measure of the speed of the desktop. X GUIs are very dependent on the quality of the available drivers, and running with the ATI or NVIDIA binary drivers is a world away from running in the VMWare VGA driver.
Its a big strech to believe that a $1000 (actually, its more like $2000, but that's irrelevent) per-developer fee is going to affect commercial development on the platform. $2000 is peanuts to commercial developers. Even smaller companies happily shell out about that much for a copy of Visual Studio, and thrice that for a Rational Rose license. Given the choice, I don't think a lot of developers are going to forsake Qt and commercial support to save $2000.
If developers really don't want to pay the Qt license fee, they won't, and GTK+ will naturally become the most popular choice. But that natural selection isn't being allowed to happen --- Perens is choosing GTK+ because that's what his backers are leaning towards.
Kiosk is just the name of the system! It lets you lock down the system as much as or as little as you want. And it probably could be implemented in GNOME, but it isn't, is it?
Your $2000 also buys you professional commercial support and a more well-documented and feature-filled toolkit.
This isn't a replacement, just another implementation of X. Not only that, but most likely it will be much more memory hungry and marginally slower (not taking into account GL acceleration). However, it will *seem* a lot faster, because it will explicitly synchronize drawing between apps and the window manager, just like OS X.
X was never slow. For the most part, KDE is faster than XP on my machine. X apps, however, have traditionally have done dumb things, and there have been problems with synchronization. Its these issues that are being addressed.
What technical reasons? GNOME was chosen at RedHat early-on because a lot of the GNOME developers worked at RedHatf. GNOME was chosen by Sun because the Sun UI people were more comfortable with C (that was their official reason). GNOME is on the verge of being chosen by Novell because they bought Ximian before SuSE, because SuSE held out for more money. And Perens doesn't even pretend that technical reasons were involved in his choice.
I hope I'm full of shit and it never comes to that, but that's what I'm afraid of.
They can make closed-source applications, they just have to pay for it. If they're making for-pay software, they should be willing to pay for their tools.
That's all besides the point. Perens isn't even giving developers the freedom to choose between GTK+ and Qt for their apps. Perens is just deciding for them.
Don't know how it is in the GNOME camp, but KDE mostly uses auto-layout, so apps don't really specify how many pixels from the edge buttons are. KDE and GNOME are really a lot more different than people think. For example, a lot of people said that KDE should adopt something like the GNOME HIG. Well, much of the stuff in the GNOME HIG is not applicable to KDE. KDE's GUI framework operates at a higher level than that. For example, the XML-GUI mechanism automatically adds the standard actions ("Open/Close/etc") when appropriate.
When that happens, the FreeQt foundation gets together and decides whether the behavior constitutes a breach of the FreeQt agreements, and decides whether to BSD-license the library.
Also, and I'm not the first to say this, shareware is dead. It has been superceded by OSS software.
Heh. Button placement *isn't* a mundane issue. The GNOME camp decided to reverse the layout, and not use the "Apply" model anymore. I don't think that was a bad idea, but it has some major usability issues without having some sort of "undo" feature desktop-wide.
First, kcalc (on my 2Ghz P4) takes less than 2s to start once its in the VM cache. Second, startup speed is problematically slow on KDE due to problems with the linker. That doesn't affect the speed of other parts of KDE. In other respects, KDE is easily as fast as XP, even faster on a heavily loaded machine.
GConf is *not* easy enough to use to be a normal interface. There is a registry-style duplication of the key hierarchy that makes it difficult to find what you're looking for. Also, there is a lot of stuff in there that the user should not be messing with.
A *lot* of KDE developers are doing something very similar to that (see the dot.kde.org story).
Bitch about politics and flamage all you want, and espouse the virtues of just "getting work done" all you want. That doesn't change the fact that throughout the history of computing, technologies have lived or died based not on technical considerations, but political ones like these. Too often, technically inferior solutions have been adopted because they were politically more viable. Us engineers would like it to be different, but such is the way of the world.
Not all MDI is created equal. Check out KDevelop3's IntelliJ-inspired MDI (which has been factored into the KDE MDI widget, so other apps like Kate can use it too!)
There has also been a shift within the projects. KDE used to be the newbie oriented/more restrictive DE. It used to have problematic license terms and be based on a non-free toolkit. GNOME, on the other hand, was built with a free toolkit and had power-user features.
Now, the roles have been reversed. KDE is fully-free, and has the power-user features. GNOME (in 2.x) has done a huge over-simplification and has alienated a lot of existing users (though, it got tons of new users attracted to the simplicity). Now, commercial companies appear to be using the LGPL-ness of GTK+ to push it as a proprietory-friendly standard desktop.
If anything, Slashdot has been remarkably consistent in its outlook.
Don't forget that Apple also has two different-looking L&Fs (Metal and Aqua) as well as legacy support for MacOS Classic. Also, Windows has at least three unavoidable toolkits (Explorer, MS Office, .NET).
The KDE and GNOME libs are both LGPL or BSD. KDE and GNOME applications are mostly GPL. GTK+ is LGPL, while Qt is GPL. That means that you can make closed-source GTK+ apps, but to make closed source Qt apps, you have to get a special commercial license from Trolltech.
If they win technically then they'll win in Linux. Don't forget that geeks are the Linux motor.
>>>>>>>>>
How long will that be true? Its true now, where KDE enjoys a slightly larger user-base because it is slightly better overall (better technology, worse interface), but with Sun pumping out hundreds of thousands of GNOME desktops, how long will geeks have any influence?
Don't we want competion rather than monopoly?
>>>>>>>>>>>
I don't want either to win. I think that the competition is a great thing for both desktops. But I fear that the commercial industry will force the situation. If KDE ceases to be a viable competitor to GNOME, because of the commercial factor, competition in desktops will be a thing of the past.
What's going on with GNOME appears to be a process of picking the standard. And true to fashion, the commercial software industry has decided to pick the lesser technology *again*, because its more convenient.
The basic problem is that Linux doesn't have to be any better than Windows to be sucessful in the commercial market, because of cost factors, so a "good-enough" solution is just fine.
Windows-style MDI is bad, but KDE doesn't do only Windows-style MDI. The newest version implements a mode called IDEAL (I think its derived from IntelliJ). Its a huge leap over traditional MDI.
I don't really understand the first part of what you're trying to say. My point is that half of all marriages in the US end in failure. Divorce rates among gays can hardly be much worse that that.
The technologies exist, but adoption is poor. When I see a KDE app, I know it'll support KIO and KParts. Even standard apps like GEdit don't support the corresponding technologies in GNOME!
Heh, not surprising you're a GNOME user.
Funny you should use GNU/linux. Since the whole "using LGPL to gain commercial popularity" is completely against what RMS and GNU espouse.