From srctran@world.std.com  Tue Dec 12 09:01:32 2000
Date: Tue, 12 Dec 2000 04:01:32 -0500 (EST)
From: Gregory Aharonian srctran@world.std.com
Subject: [Patents] Report on software patents.

Dan,
	You and the others use sorting algorithms as the test case
for software patentability.  Well, I can show you just as trivial
chemical patents.
	Instead, consider the problem of natural language translation.
I have yet to see a really good piece of software to do that.  Why?
Because it is one of many software problems that are not "utterly
trivial".  You solve the natural language translation problem (or
even just a good subset), and you will have companies throwing as
much money at you as they did at L&H before L&H went bankrupt.  And
you will want your many years of R&D into your natural language
technical solution protected.
	And any trivial sorting algorithm more than likely won't
be patentable because of the prior art.
	Every field has trivial things being patented, and non-trivial
things being patented.  Including software.  This is yet one more
criteria that can't be used to separate out software patents.  You
want to have some - go look at the golf ball dimple pattern utility
patents awarded endlessly here in the US.

Greg




