[ILUG] Future of Mesa ?

[John P. Looney]

On Thu, Jan 27, 2000 at 04:45:51AM +0000, Smelly Pooh mentioned:
> The impression I got is that this sample implementation is just that, a
> "sample" implementation.  3D hardware developers are supposed to do their own
> implementation (same way you'd have different openGL drivers for a Voodoo card
> and a Nvidia card on windows).

 True, but given an implementation of OpenGL, there can't be more than 10%
of it you'd have to re-write to add hardware acceleration. I'd hope it's
somewhat modular, with all the low level rendering stuff off into a few C
files that call X functions in a sample implementation.

> If you check www.nvidia.com/News.nsf they also
> have an announcement of cooperating with SGI and VA Linux to bring out the
> first hardware accelerated OpenGL implementation for Linux.  Nobody seems to
> be really sure what to do with Mesa, read the comments on the article on
> slashdot.org
 
 Hmm. They seem to be jumping the boat. Xig have had a hardware
accelerated OpenGL out for a fair while, and MetroX have been shipping an
X server & OpenGL implementation for about five years. Proper, tested &
SGI badged OpenGL at that.

> >    Mesa is still a very good software implementation.  Its great for those
> >    who want to see how things like rendering on the fly is done in software...
> Nobody wants to see any form of 3D done in software anymore, the fact that
> just about every graphics card manufactured in recent years has hardware 3d
> acceleration might indicate that.

 Indeed. I was of the idea that Mesa formed the backend of the OpenGL/
XFree86 4.0 implementation. That's why Brian Paul got the job with
Precision Insight, yeah ? It's all a bit mad. Imagine the grief that could
have been saved had SGI open sourced OpenGL five years ago. Mesa wouldn't
have had to be written. Gamers could have bundled OpenGL libraries with
their games. Direct3D wouldn't have got off the ground.

 Sigh. Like Netscape with the Mozilla project, it's a "Open Source this
software, and Microsoft can never kill it" idea. And it's too late. Had
Mozilla been started 3 years ago, when browsers were still relatively
small, or OpenGL been started before 70% of 3D programmers knew more
DirectX than OpenGL, it would have killed Direct3D as a gaming platform.

Kate

-- 
Microsoft. The best reason in the world to drink beer.
http://www.redbrick.dcu.ie/~valen