[music-dsp] [OT] Patent: Efficient convolution withoutinput/outputdelay

[Mike Berry]

Tóth László wrote:
> 
> On Thu, 15 Mar 2001, Mike Berry wrote:
> 
> >       The standard for patents includes "non-obvious."  If you can show that
> > something was obvious at the time of invention, then the claim is
> > invalid.
> >
> But what is the way to show that something was "obvious"??
> If it is documented, than it proves that it was "known".
> Put the opposite does no hold, since nobody documents "obvious" things, so
> the lack of documentation doesn't prove that it wasn't known.

	The standard method is to have experts testify, once it gets to trial. 
Prior to trial, basic textbooks are often used.  If something appears in
a basic textbook to teach a subject to someone who does not know it,
then it would be obvious to someone knowledgeable in the subject.  If
something is so obvious that it does not appear in basic textbooks, then
it should be obvious to all involved.
	Failing these, you can also look to papers or patents in the field. 
Lets say that process x is patented and in question.  If you find prior
art which does not describe x, but assumes x (such as "when doing y, do
x 4 times"), then this might be a case for obviousness.  If you can find
this repeatedly, your case is stronger.  The thing to watch out for here
is that the authors might require x without having any knowledge of a
method to perform x.  This means that the method for x is non-obvious,
though the need may be obvious.  Also, the patent might be a novel
method of performing x, even if other methods exist (which means that
the scope of the claims would be limited to the novel method, not other
methods).
	I'm not a patent attourney, so don't use my thoughts here in
litigation. :)

-- 
Mike Berry

dupswapdrop -- the music-dsp mailing list and website: subscription info,
FAQ, source code archive, list archive, book reviews, dsp links
http://shoko.calarts.edu/musicdsp/