[Patents] are claims OR'ed or AND'ed?

[srctran@world.std.com (Gregory Aharonian)]

Ramon writes:
>The strategy is to draft the first claim in the most broad possible
>way, and the other claims describe the invention with more detail. In
>this way, if a judge finds the first claim obvious, that would not
>automatically invalidty the other claims.

Mostly correct for US patents.  For US patents, there are two kinds of
claims: independent and dependent.  Independent claims are usually
broad, and are "OR"ed together (in the words of the questioner).
Dependent claims are associated with specific independent claims, and
narrow the scope of the independent claim (i.e.  ANDed to the
independent claim).

There is one more narrowing that goes on, at least in the US.  All claims
are interpreted with respect to the patent specification.  For example,
you could see a really broad claim ("A method of finance, comprising
sort a list of income figures"), and think, "That is outrageous as a
claim" - and maybe so, but the claim is limited to the financial sorting
techniques explicitly described in the patent application, which could
be quite narrow in scope.

So in many cases, a really broad claim is actually quite narrow (though
I have been arguing for years that broad independent claims should be
explicitly narrower, promoting some of the narrowness of the dependent
claims and specification into the independent claims).

It is a problem that affects all areas of technology, including software,
and if patent offices could focus on giving their examiners more time,
money and resources to search for prior art and then demand applicants
narrow their claims (if not withdraw their patent application al
together), maybe of the stupid patents in all areas of technology,
including software, would disappear.  As would most of the arguments
against software patents.  It is a solvable quality control problem.

Greg Aharonian