[theora] Steve Jobs about theora

[Greg Ewing]

Shayne Wissler wrote:
> If/when it turns out to be true, don't say I didn't warn you.
> 
> There is no reliable defense against patents, particularly given how
> it they have been applied to software.
> 
> 
> Shayne

There is a good defense. When someone actually gives you the patent 
numbers of the patents that they claim you infringe. You can often code 
around it. I have read a lot of "codec" patents. Many can be avoided by 
simply doing things differently-but mathematically "equivalent" as far 
as the format is concerned.

Example: One patent makes some claim that "lifting" a established DCT 
transform matrix is an invention thats patentable. But the patent claims 
are restricted to integer arithmetic with bit shifts. So the same 
lifting scheme using floats would avoid *all* the claims in this patent.

There is also quite a few patents hitting the expired list. So you can 
use their claims to the letter and have a prior art defense that may in 
fact work. But code around seems more prudent without a big legal fund.

This is one reason that "patent" FUD campaigner's never give specifics 
against any OS project. By the time it mattered we wouldn't be infringing.

Cheers
Greg