[Mesa-dev] S2TC - yet another attempt to solve the "S3TC issue"

Joakim Sindholt bacn at zhasha.com
Tue Aug 9 12:23:39 PDT 2011

On Tue, 2011-08-09 at 19:49 +0200, Rudolf Polzer wrote:
> On Tue, Aug 09, 2011 at 10:46:12AM -0700, Corbin Simpson wrote:
> > I should point out something not immediately obvious about S3TC: It's
> > believed that the patents cover any complete pipeline which
> > decompresses S3TC textures according to the S3TC algorithm. It's
> > stupidly broad that way.
> Sure is, which is why S2TC is not able to decode full S3TC, but has a weird
> "fallback" handling that does not make use of inferred colors.
> If only it weren't for the glaring similarity to RPZA's 0xC0 opcode... which is
> VERY similar to S3TC's encoding scheme, and was invented 8 years earlier. That
> can get very interesting, as I assume that Apple tries to defend against the
> S3TC patent exactly by pointing out the similarity to RPZA (the Quicktime 1
> codec).
> Best regards,
> Rudolf Polzer

According to people who actually know about software patent cases, prior
art is near impossible to pull off. The problem is simple: unless you
had the exact same thing bit for bit prior to the patent, it won't work.
Just shy of changing the name, ANY change will invalidate the case for
prior art.
The best (and probably only) way to help mesa would be to obtain a
license for both encoding and decoding of S3TC for it. Alternatively you
could wait till 2017 when the patent expires.

Sorry to disappoint.

More information about the mesa-dev mailing list