🚀 go-pugleaf

In article <1993Feb23.172436.26281@sco.com>
	staceyc@sco.COM (Stacey Campbell) writes:

>In article <C2wKMC.2zD@gator.rn.com> larry@gator.rn.com (Larry Snyder) writes:
>>>*furthermore*, with the card, there is no mention whatsoever of SCO or
>>>*NIX and there are no drivers for SCO or *NIX ???????
>>
>>Some video boards, yes -- you are correct -- however the higher end
>>"high-res" boards usually include unix drivers [...]

>Our X11R5 product supports over 110 seperate video cards, of the ~20 or
>so cards that run at resolutions higher than 1024x768 I am aware of only
>one vendor that also ships working X11R5 driver support that integrates
>into our X server (for multiheaded servers, etc).

I think that what is happening, at least for the time being, is that X
support for non-standard video cards on Intel-based systems is
coming from neither OS vendors or hardware vendors, but from third parties.

Presently there are four such third-parties that I know of. Three of
them, SGCS, MetroLink and Pittsburgh Powercomputing, offer commercial
add-on servers for R4 systems that support a very broad range of
accelerator cards, at resolutions up to 1600x1200x256col. The fourth is
the committee which has developed the XFree86 package, a robust
implementation of X for most non-accelerated SVGA cards.

Some of these third parties offer value added options such as real-time
video. All of them have a habit of keeping more in step with developments
in X Windows itself. (For example, all of the third parties are well into
X11r5, while many of the servers offered by the OS vendors are still on
X11r4.)

What would be best is if the X server was flexible enough so that the
vendor-independant "kernel" could be kept separate from the
vendor-dependant "drivers", to use the Unix analogy. Then board makers
could include X Window drivers without having to supply entire X servers, 
a move which would no doubt encourage more participation by the makers
of video boards. Then we could go to a system like with serial boards or
host adaptors, where the board maker supplies the driver, but not the
vendor-independant stuff.

Until that happens, there is a choice of either the X server offered
with each UNIX (sometimes as an option), or to spring for the
high-performance third-party servers for a little extra cash (except, of
course, for XFree86, which is certainly ready to take its place with Kermit,
TeX, gcc, Cnews, Linux and GNU Emacs as the most significant freeware products
of all time).

>I would suggest that rather than expecting to get a working driver when
>you shell out big dollars for high-res/high-speed video hardware, you get
>the Hardware Compatibility Guide for your particular flavor of Unix and
>make sure it says in black and white that your hardware is supported.

Amen to that. It has always been my practice that the software to be run
determines the hardware to run it, not vice versa.

ie, the application determines the (choices of) OS which determines
the (choices of) hardware. The third-party X servers (some of which will
run under SCO too) add another dimension of flexibility.

-- 
 Evan Leibovitch, Sound Software Ltd., located in beautiful Brampton, Ontario
         evan@telly.on.ca / uunet!utzoo!telly!evan / (416) 452-0504
       If you eat too much Nutrasweet, do you get artificial diabetes?