HD with Xvideo on i915GM (i810 driver)

[Alex Deucher]

On 11/6/06, Andrew Barr <andrew.james.barr at gmail.com> wrote:
> On Monday 06 November 2006 14:07, galenz at zinkconsulting.com wrote:
> > On Nov 5, 2006, at 12:06 PM, Torgeir Veimo wrote:
> > > On Sun, 2006-11-05 at 15:00 -0500, Andrew J. Barr wrote:
> > >> There are some issues with XvMC in general, especially in open source
> > >> drivers,
> > >> because of the minefield that is otherwise known as multimedia
> > >> patents. It's
> > >> hard to implement this stuff fully without infringing on at least
> > >> some
> > >> patents, from what I understand.
> > >
> > > _Releasing the docs_ on the MC part wouldn't infringe any patents?
> > >
> > > --
> > > -Tor
> >
> > I don't think you can infringe on a patent simply by writing about it.
>
> Part of the issue, at least in the ATI-nVidia world, is that the hardware
> itself infringes on some patents and releasing docs and/or source code (for
> that part of the chip) would advertise to the patent holder that there was
> infringement. I don't think that this is the case with Intel, however,
> because they release open source drivers that are planned to be fully
> functional (minus some patented texture compression and some other minor
> details, I think) that they make no attempt to obfuscate.

I think this is a convenient excuse for open source developers and HW
manufacturers, but I don't think it holds water.  The register specs
do not reveal the implementation or even whose cell is actually used.
when the register interface is designed it rarely has anything to do
with the actual cell so much as it has to do with relativity to the
rest of the chip's programming interface or similarity to known
interfaces such as the directx or other specs.


Alex

>
> > -Galen
> > _______________________________________________
> > xorg mailing list
> > xorg at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/xorg
>
> --
> Andrew Barr