However, this is the second time I've seen this reasoning as argument
for a kernel driver and I think it's incorrect.
Technical issues aside, this is a question of what, fundamentally,
Linux IS and what UNIX IS. You have to look at the direction that Linux
is moving in. Companies like red-hat and caldera are working very hard
to build acceptance of Linux as a reliable, low cost alternative to
Unix(tm). More and more companies are beginning to look at Linux as a
viable place to exapand their market with a vertical application.
Let's look at the requirements of a game. A game needs an environment
where it can get as much performance as possible out of the hardware.
Some of the best games are still written in DOS becuase DOS won't put a
barrier up between the game and the hardware. A game expects to have the
resources of the hardware available at all times.
Now, Microsoft is trying to provide a way to accomplish the same level
of performance while still keeping a reasonable amount of multitasking,
etc... However, Windows 95 is not as stable. It's not built to handle
multiple users. etc...
There is a fundamental, philisophical if you will, difference between
Linux and DOS/Windows. Linux is meant to be a multi user system, capable
of handling all kinds of different services. It is not MEANT to be a
systemm, used by a single person, sitting in front of the console.
Granted, many people, including myself, use it exactly that way.
However, Linus needs (we need)? to look at this issue from an eye toward
commercial
acceptence if he (we?) wish to shake the idea that Linux is a "Hacker's OS."
Two points: 1) There's nothing wrong with Linux as a hacker's OS.
It's just that I think it would do the world a great deal of good to be
able to hold up Linux and make commercial Unix vendors charging thousands
of dollars for their product shake in their boots. I think it would be a
triumph of "grass roots" organization against coorporate greed. Ok, I'm
showing my colors I know..
2) Notice, I'm not arguing AGAINST a kernel level driver. I'm not
arguing for it either. Like I said, I don't know enough of the technical
details.
All I'm saying is that to say, "we want this so we can attract people
who want to play games," is sort of missing the point. Games are fun and
I love a good game, but an OS is a tool. In a perfect world, we would
have one SuperTool that would do everything (it slices, it dices, it
chops, and hops...), but this isn't a perfect world.
It seems to me, honestly, that playing games on Linux is kind of like
buying a Porsche to drive down the street to the store. Yes, it will do
it and it will do it VERY well, but it has so much more to offer.
So, I think a convincing argument to make would involve not games, but
high end graphics applications. Imagine if Linux became THE platform for
digital video/effects? Granted, it's a dream, SGI has a big head start,
and PCs don't really have the hardware (yet?), but it's a much better
reason to build a good graphics engine than games.
Oh, and don't forget another goal is to NOT make linux a proprietary
systemn. Do the xfree people really want to have to create one S3 driver
for Linux and another S3 driver for everything else?
Of course, nothing I've said really argues one way or the other in
terms of kernel vs. user space driver (accept maybe the last point...).
I just think we ought not to think of Linux with such a narrow vision..
Sorry this is so long winded, that's what comes from having a
communication degree. :)
____________________________________________________________________
Mark E. Levitt
Department of Speech Communication, Syracuse University
E-mail: melevitt@syr.edu
Home Page: http://web.syr.edu/~melevitt
PGP fingerprint = B8 A3 AA A6 0F 83 9A BE F2 7A 19 F9 15 79 FE A4
Public key available from http://web.syr.edu/~melevitt/pgpkey.html
____________________________________________________________________