Games doesent work in Fedora test 3

[Mike A. Harris]

On Tue, 21 Oct 2003, Joe wrote:

>>Why not, if you're playing all of them on an Nvidia card.  What you don't 
>>realize, is that once you install the Nvidia drivers, you nullify any kind of 
>>practical support you could get for any 3d application.  If you've got a 
>>problem with it, bring it up with Nvidia, since they're the only ones that 
>>can help.
>
>I think his point is that all fc t3 users will have problems with these 
>games. Dismissing him with a wave of the hand because he's using an 
>nvidia card may work for awhile, but eventually there may be some brave 
>souls who venture to attempt games with non-nvidia cards, and when they 
>see the same problems and worse, the problems will get the attention 
>they really should be getting now - you create a sort of vicious cycle 
>when nvidia cards are really the only practical choice for a linux 
>gamer, but users of nvidia cards are shouted down when reporting bugs.

And you misunderstand that wether or not someone here would like 
to solve an Nvidia user's problem, that we cant fix bugs in 
Nvidia's drivers, and we can't easily even debug them to find out 
where the problem could be.  It's just not something we can do.  
There are hundreds of capable engineers out there and 
programmers.  Is there anyone reading this list who thinks they 
can debug a 3D acceleration problem with a proprietary video 
driver and proprietary kernel module, without the source code, 
and be able to do ANYTHING for a user having this problem?  I 
challenge anyone out there to do so.

Now if users have problems with open source drivers it is a 
completely different story.  At least the source code is there, 
and some of the documentation is available to some people.  
However from a Red Hat perspective, I have to make a decision 
with prioritizing work.  Hmm, should I spend a week and a half 
debugging a 3D lockup in DRI on Radeon that happens with Unreal 
Tournament 2003 (after spending $70-80 of my own money to 
purchase a legal copy of the game I don't have) or should I spend 
that week and a half fixing 5-10 bugs for commercial enterprise 
customers spending tens of thousands or more dollars on our 
products?

On a completely personal level, I'd much rather have fun 
debugging the video game issue.  ;o)  On a responsibilities to my 
employer and our paying customers level however, I can't justify 
debugging video games as a high priority.  The only time I can 
really justify spending time debugging video game related DRI 
issues is on weekends or in my own personal time.

Theoretically, if I spent all my time making video games work
with DRI, I might make Red Hat's DRI the most rock solid out
there.  But then how long would I have a job here?  ;o)

>I hope the development community finds a better response than
>dumping on nvidia, since nvidia might just decide it's too damn
>much trouble to support linux - and then we'll really be out of
>luck.

Nvidia doesn't provide video drivers for Linux in order for video 
gamers to have working drivers.  Nvidia wrote their video drivers 
for high end 3D users including the movie making industry, 
medical industry, geological research, and other scientific 
applications.  They provide their drivers freely for download to 
the community for other people to use simply as a kind public 
service.

If it became a problem to them to provide these drivers, they
could just as easily not allow public download, and instead
provide them only direct to their high end corporate customers in
the fields that I mention above.  No video vendor out there
provides 3D drivers specifically for people to play video games -
they all provide drivers for one purpose - to make money of sales
of high end video hardware for high end business usage.  The
community is just lucky that they have mostly all unified their
drivers nowadays to support all of their hardware from a single
source.  ;o)

So you can very much think of all of the available proprietary 
drivers out there as a charity to the Linux community, as that is 
what they very much are.




-- 
Mike A. Harris     ftp://people.redhat.com/mharris
OS Systems Engineer - XFree86 maintainer - Red Hat