[Patents] Re: European arguing for patenting pure math formulas

[phm@a2e.de (PILCH Hartmut)]

> I am completely against swpat, but Greg has a somewhat valid criticism
> of Hartmut's formula, and we will get further with real arguments than
> with false accusations.

This was not a formula for sorting out software patents but just another
possibility of introducing a limit on patentability in analogy to those
limits that Paul B. Laat proposed as a corollary to allowing claims to
purely mathematical constructs.

> Greg's grain of truth: a mathematical proof is relatively easy when
> one knows what is to be proved.

A mathematical proof based on quantum wave properties apparently isn't all
that easy.

> I rephrase Hartmut's statement: A patent on an invention in the realm
> of applied natural science is valid until the day when someone has
> mathematically deduced it from prior knowledge in a cheap process that
> does not benefit from knowledge of the patented material.

I may not understand the last part "in a cheap process ..."

> But this rule may not exclude all software patents, since proving
> code, while easier than quantum mechanics, is still tricky.  Isn't it
> the case that some encryption algorithms rely on a currently unproven
> proposition, which if false would render them vulnerable to easy
> attack?  One might conclude that the algorithms should be patentable.

Maybe that wouldn't be all that wrong then.
I repeat: my proposition wasn't meant for real-life use but as an answer
to Laat.

> I prefer another formulation: No patent should prevent people from
> doing whatever they want with information, including copying,
> transmitting, manipulating, augmenting, or interpreting it.

Yes, that is one basic goal to be achieved by the delimitation rule.
You restrict subject matter in such a way that you do not get claims
on pure information objects or on information processes that run on
state of the art hardware.
Another way to do this would of course be to take Laat's approach of
patenting pure mathematics and at the same time introduce a rule that
whatever is claimed, it can only be enforced against material products
produced in an industrial fashion.
Laat appeared to me rather shallow in that he did not attempt to define
the scope of objects against which math claims should be enforcable.
He just distinguished between "commercial" and "private"/"research", not
realising that these borders have become very vague in a society where
almost anything, including private life tends to be commercialised and
governed by commercial aspects.  How can a philospher not ask "where
exactly should the commercial sphere end".  I think it ends where
information begins. Information cannot be sold, but only crippled and
connected to a license that then is sold.   While crippling/licensing is
quite ok, it should not be considered to be the default paradigm.
It is also quite appalling to me that Laat and other people imbued with
patent movement thinking have no notion of "public".  Why only
"commercial vs private"?  In software "commercial vs public" is a much
clearer and more important borderline.

> In other words, patents may legally restrict the processing of matter,
> but not the processing of knowledge.

> (To write this into law, one must point out that such effects as the
> motion of electrons in information processing are irrelevant.)

--
Hartmut Pilch
Federation for a Free Information Infrastructure
protecting information innovation against the abuse of the patent system
http://swpat.ffii.org