On Fri, Jul 18, 2003 at 06:09:44PM -0700, Mark Vojkovich wrote: 
>    Ironically, the Linux desktop community doesn't target the
> only potential business case there is.  It's often at odds with
> it.  Workstation users like a platform that doesn't change and anything
> that risks damaging OpenGL behavior (like RandR support or alpha
> blended cursors) is generally not well accepted.

I guess this thread is getting offtopic, but fwiw, from what I can see
of customers and other data this battle is already won. Linux is
definitively cheaper and meets the technical need for just about
everyone, where UNIX workstations are hanging on it's by inertia and
due to certain ISVs. There are not a lot of additional features needed
to win these customers, other than continuous performance gains.

For the 3D performance gains everyone is basically 100% dependent on
nvidia and ATI; if you guys make the improvements they're there,
otherwise they aren't. Nobody else has the ability to address this
issue that I know of.

People are looking beyond this market, because the workstation market
simply is not the largest in the scheme of things, and isolated by
itself has little reason other than a body of in-house and ISV apps to
be Linux-based rather than Windows- or Mac-based. The historical
reason it was UNIX-based was the high end hardware; that reason is
gone.

This paper has some good points:

  http://www.osafoundation.org/desktop-linux-overview.pdf

There is substantial interest in the world today in Linux as a
desktop, and enough of that interest is among potentially paying
customers who understand the technical realities. For now, granted,
we're talking primarily about highly managed and locked-down desktops
with a restricted application set.  But for whatever reason,
empirically many people want to use Linux for anything it's capable of
doing, as soon as it's capable of doing it.

If the X-based platform is to be capable enough for the current
interest in managed desktops, and the few-years-out interest in more
general desktops, we have to be making progress intelligently and
continuously. That's how we're getting there on the server, it's how
we got the workstation market, and it's how we'll get future markets.

There isn't a business case here yet for nvidia, or for desktop ISVs.
However, for the organizations looking at Linux desktops today, those
things aren't essential; which gives a possible way out of the
chicken-and-egg/bootstrap problem. There is already a business case
for many operating system vendors and hardware OEMs, judging by their
publicly-announced actions over the last year.

Nobody can predict the future, of course. Many things can happen.

I don't believe there is a fundamental tradeoff between what all these
various customers want to do, from a technical standpoint.  However,
to properly balance all the competing requirements, looking at the
client-side GUI platform as a whole rather than as any single unit
(window system, desktop shell, toolkits, applications) will be
important. Fluidity and adaptation are essential, properly balanced
with regular ABI-stable releases.

Havoc

_______________________________________________
Devel mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/devel

Reply via email to