On Sat, Jan 31, 2004 at 04:43:17PM +0200, Shaul Karl wrote:
>   Excluding Nvidia and ATI, for which I believe I know the answer, what
> manufacturers I am likely to see on ebay that:
> 
>     1) Usually fully and freely publish the specifications of their AGP
>          hardware.
>     2) Got themselves an X driver?

In case you are lumping NVIDIA and ATI into the same category, you have
made a mistake. (In case not, ignore this post.)

ATI has been quite friendly with NDA docs for all but their latest
hardware, AFAIK. A Mach64 driver was even recently written based on
their help.  We are lacking r300 documentation but maybe when their next
gen hardware shows up, they will be more open with it.

NVIDIA, on the other hand, released some obfuscated source code back in
1999 that the Utah GLX driver was based upon.  After that point, they
eventually expressed dissatisfaction with the DRI architecture and have
never released an ounce of documentation or driver code since.  Their
claims are that not only would their object-oriented hardware model be
too hard for open source developers to understand and write high
quality driver software for, there is also too much proprietary and
licensed information embodied in their hardware for them to risk
releasing.  To their credit, they have done an excellent job of
supporting their binary releases.  The unfortunate truth is that
eventually the stream of support for the current cards will end one day
when it is no longer economically feasible for NVIDIA, and we will be
left with more useless hardware that we are incapable of supporting as a
community.  Maybe at this later point they will consider relinquishing
their support monopoly for their hardware, but something tells me the
situation down the road will be worse, not better.

Your request for free publication is undeniably idealistic.  I think it
is a perfectly reasonable compromise to provide specs under NDA to
developers who have shown themselves to be productive and trustworthy in
the past, e.g. by contributing to XFree86 or producing and supporting an
own 3rd-party driver like Tungsten Graphics.  It is a much less risky
investment for the chip manufacturer than freely publishing documentation
for all.  The manufacturer will rarely reach any individuals who would
not have qualified for a NDA anyway, and will most likely end up giving
their competitors ideas they may not have had otherwise.

The downside to the NDA approach is that the driver is less supportable
than without an NDA - once the NDA developer moves on to other things,
abnormal aspects of the hardware may be difficult or impossible to
understand.  Furthermore, he may have left things unimplemented or
undocumented in the code that will make people scratch their heads later
on.

This is all of little importance compared to what we gain through open
source NDA development - a driver that can run on any hardware
platform, be ported to other system software, maintained aside from the
hardware interface, reused in other projects talking to that hardware,
etc.  We would have none of this if graphics companies were not gracious
enough to release their proprietary information under NDA.  If they want
to go the extra mile and provide their documentation freely, for example
like Matrox did in the glory days of DRI, that is all the better for us;
but I do not think we should begrudge NDA development at all, because it
strikes what is IMO a fair balance between the interests of the hardware
manufacture and the interests of the open source community and users.

-- 
Ryan Underwood, <[EMAIL PROTECTED]>

Attachment: signature.asc
Description: Digital signature

Reply via email to