From phm@aliuC2.oas.a2e.de  Fri Jan  5 09:27:46 2001
Date: Fri, 5 Jan 2001 10:27:46 +0100 (CET)
From: phm@aliuC2.oas.a2e.de phm@aliuC2.oas.a2e.de
Subject: [Patents] triviality of innovation steps in logical buildings

> On Mon, Jan 01, 2001 at 05:35:34PM -0500, Gregory Aharonian wrote:
> > And for those patents, software or not, that seem trivial and
> > obvious, there is often enough prior art to invalidate the patent,
> > if given to the patent examiner.
> 
> It is a matter of luck if there is prior art to a trivial patent.
> Your criterion would invalidate only a fraction of trivial patents,
> not all of them. And there is no good reason why a trivial patent
> should have been granted.

We can of course another round of "same for any other industry" vs "no,
the swpat case is specially painful."

But that is a side issue.  The important point is that the novelty test
will eliminate only a fraction of the trivial patents.  And when looking
at the swpat horror gallery, you will find that most of these patents
pertain to trivial solutions of new problems.  Since the problem is new,
and new problems appear every day, the novelty test will actually tend to
eliminate very few patents.  

Again, concerning the side issue of whether there is anything
software-specific in this, I'd say yes.  As soon as you have natural
forces involved, you cannot construct sequentially layer by layer.  Even
lego brick constructions will often fail to work the way you constructed
them in your mind, not due to reasoning errors, but due to deviations
resulting from natural forces that you were unable to reckon with.

When systems become complex and innovation proceeds sequentially, new
problems appear very quickly all the time and each single solution step is
quite trivial.  What is claimed in a swpat main claim usually fits into a
higher-level programming language procedure of about five statements.  
Everything below that is necessarily considered "obvious to the man
skilled in the art", when in fact porting such an algorithm from C++ to
Java may be less trivial than inventing the algorithm itself.  This is
highly arbitrary.  In the end, which layer is claimed depends mainly not
on

	 which layer represents my ingenuity?

but

	which layer prevents my competitor from solving the problem?

which again leads to a model of "functional claims" or claims to
"technical effects", which in fact are problem descriptions. 

-phm 





