Games tend to push certain parts of the hardware to its limits. This
can result in new developments which are very important for scientific
applications: e.g., real time simulations, good memory management, etc.
I recall a lecture by Bill Joy in the early 80's in which he said, more
or less, that, as a new graduate student, he noticed that the computer
(perhaps processor) use in the video arcades seemed to be better in many
ways than what he saw in the faculty offices at Berkeley. He began to
think about better use of terminals, graphical interfaces, etc. We know
many of the things that came out of that.
-sen
Linus Torvalds writes:
> > On Tue, 4 Jun 1996, root wrote:
> > > least make it a module option. The fact remains, if linux is to be
> > > supported by game developers it needs to have the basic video support
> > > required.
> >
> > Ok, here's the fundamental point I was trying to make: Do we want, as
> > a design goal, for Linux to be an OS for game developers? Is "supported
> > by game developers" really a goal for Linux/UNIX?
>
> Actually, I'd say YES here..
>
> I consider games pretty important - we don't want to get snotty and tell people
> about "real work" - we want to have the best all-round OS available, and that
> definitely does mean games too..
>
> In short, I'd _love_ to have lots of good games, and I was _really_ happy when
> DOOM was made available for linux.
>
> > It seems to me that this is not the direction Linux should take, or IS
> > taking. Is Linux supposed to be an alternative to Windows 95 and the
> > Mac or is it supposed to be an alternative to Windows NT, SCO Unix,
> > Solaris, etc...?
>
> Oh, I'd like to be an alternative to Win95 any day.
>
> The SCO Unix and Solaris kind of OS we can already pretty well replace: the
> issues there are not really technical. WinNT is a joke.
>
> We _do_ want exactly the kinds of applications Win95 has, and that certainly
> includes games, even though I consider the "real" applications like Applixware
> even more important.
>
> > Now, if we can make Linux a great game platform WITHOUT compromising
> > these goals, then fine. However, at least Linus and some others don't
> > seem to think that's possible.
>
> Here you're touching on one of the problems. It's pretty hard to support games
> in a _good_ manner without just giving up the whole machine. And I have to
> admit that I tend to use the fast machines where a game under X is perfectly
> reasonable. I'm not talking "solitaire" here, I'm talking raw action, 3D stuff,
> "a better DOOM than DOOM" too.
>
> However, we do have one thing going for us: technology. Machines aren't getting
> slower, and I doubt that the _real_ problem with games under Linux is the fact
> that you should use X to write them. The _real_ problem is that most of the
> developers are under DOS - THAT is why there are more and better games for DOS,
> not because Linux lacks anything in this area.
>
> In short, I'm definitely NOT arguing against games. But we need to be very
> careful about what we do about graphics, and that's why I've been calling for X
> rather than any other solution. Because X is a viable option even in the long
> run, while I don't see any other "standards" emerge (note that "standards" are
> not just pieces of paper: the BEST standards are the de-facto standards that
> never were paper in the first place, but somebody just went out and did it, and
> it was so good that it _became_ a standard).
>
> The GGI stuff might well become a standard that is worth supporting. It just
> isn't that yet..
>
> Linus
>
>