Linked by Thom Holwerda on Tue 14th Sep 2010 22:42 UTC
Ubuntu, Kubuntu, Xubuntu If there's one consistent piece of criticism that gets lobbed in Canonical's and Mark Shuttleworth's direction, it's that they do not contribute enough code - or anything else for that matter - to the Free software world. Mark Shuttleworth has apparently had enough, and has written a very, very lengthy blog post detailing how he feels about this criticism.
Thread beginning with comment 441241
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[2]: duh
by Zifre on Thu 16th Sep 2010 00:09 UTC in reply to "RE: duh"
Zifre
Member since:
2009-10-04

If anything, X.org and Mesa are terribly understaffed.

X.org, definitely. But I think Mesa is actually doing pretty well. Progress is happening fast, and it is one of the most active FOSS projects. (Linux and Mozilla are the only projects I know of that might be more active.)

Missing support for GL ES 2.0 and buggy support for GL 2.0-only may result in KDE producing two back-ends for KWin that share next to no code. If Xorg/Mesa supported OpenGL 3.x, 4.x, and ES 2.0 no duplicated KWin development was required.

First of all, Mesa does support GL ES 2.0 (it is used by Wayland). I'm not sure how complete or correct it is. Second, although different graphics APIs might not be compatible, that hardly means that they would share little code. 90% of what works in GL ES 2.0 will also work in GL 2, 3, and 4. GL 3 and 4 would only be used for optional eye-candy or performance enhancing features. The core rendering stuff would be basically the same. Lastly, the renderer is a very small part of a window manager. Drawing rectangles onto the screen is not terribly complicated. The hard part is window management.

Nouveau also could need much help, especially in power management and support for composite window managers.

I agree there. If Canonical could pay a few developers to work on Nouveau, progress would be much faster (and it is already very impressive). It would probably be the most useful thing that they could do with their money.

Reply Parent Score: 2

RE[3]: duh
by Fettarme H-Milch on Thu 16th Sep 2010 13:25 in reply to "RE[2]: duh"
Fettarme H-Milch Member since:
2010-02-16

X.org, definitely. But I think Mesa is actually doing pretty well.

If it was so well, where is OpenGL 3.x support? I'm not even beginning to ask for 4.0.

Btw: While both are technically separate projects, they pretty much go hand in hand for 3D-accelerated GPU drivers.
Xorg's GPU drivers can't support higher versions of OpenGL than Mesa offers. And Mesa is f*cked when drivers incorrectly claim to properly support feature X of OpenGL 2.x.

Reply Parent Score: 3

RE[4]: duh
by Zifre on Thu 16th Sep 2010 22:46 in reply to "RE[3]: duh"
Zifre Member since:
2009-10-04

If it was so well, where is OpenGL 3.x support? I'm not even beginning to ask for 4.0.

First of all, why do you need 3.0? Of course it's nice to have, but certain things are more important. Here is Mesa's progress toward 3.0: http://cgit.freedesktop.org/mesa/mesa/plain/docs/GL3.txt. A large portion is already done. And not all of those features are terribly useful. I might rather see them work on performance or Gallium then specific OpenGL features. And almost nothing uses 3.0 right now anyway, so it doesn't matter much.

Xorg's GPU drivers can't support higher versions of OpenGL than Mesa offers.

That doesn't make sense. X.org drivers implement things like EXA and other 2D stuff. They do not rely on Mesa at all. Mesa does all the 3D, and it has almost no dependencies on X (as demonstrated by Wayland).

Of course both projects would be better off with more developers, but I think Mesa is doing better than X.org right now.

Reply Parent Score: 3