🚀 go-pugleaf

>Maciej Stachowiak wrote:
> > I think most people can agree that mathematical knowledge should not
> > be patentable. But:
> > 
> > * The means of performing a computation is mathematical knowledge.
> > 
> > * Any possible scope of a "software patent" is a means of performing a
> > computation. All computations that can be performed by a computer are
> > mathematical, whether or not you express them that way. Inside, the
> > computer is just pushing bits.
> > 
> > * Therefore software is mathematical knowledge.
> > 
> > * Therefore software patents ought not be permitted.
> > 
>
> I used to be opposed to software patents and this was one of my main
> arguments.  However, if you really think about it, the argument is less
> persuasive.
>
> The problem with the argument is that if you try hard enough, just about
> anything can be dismissed as mathematical knowledge. 

No; the laws of nature in particular cannot be deduced by pure reason
(I'm sure Klaus Shilling will object to this, being a Platonic
Idealist, but whatever). Thus I am totally OK with any invention that
depends on natural law, i.e. anything physical. (Note: software
patents have no dependence on physical law; an abstract Turing
computability is independent of physical law).

> In fact, the
> logical extension of your argument is that anything that you can do on a
> computer cannot be patentable since it is just an extension of
> mathematical knowledge.

> Many physical inventions are first fully simulated by computer prior to
> being built.  In what way does a simulation differ from a computer
> program?  Are these inventions to be denied patents?

I think the invention should be patentable, but not the program for
simulating it.

> You can claim that physical inventions are patentable because they apply
> mathematical knowledge to do actual work, but making such a claim while
> insisting that software should not be patentable is implying that
> software doesn't do actual work.  I contend that software does do actual
> work, and thus should be patentable.

No, the difference is physical law and it's applications, which are
discovered inductively, vs. mathematical knowledge which is discovered
by pure reason.

> Intellectual arguments about the meaning of the word "math" aside, my
> real reason for supporting software patents remains that the granting of
> patents encourages innovation by providing a legal means by which a
> company or organization can profit from developing new software
> algorithms.

Comapanies have enough incentive to do this without software
patents. It certainly went OK before software patents existed. In
fact, the point of patents is to allow disclosure of such discoveries
instead of having them remain trade secrets, but since many software
patents are independently reinvented this is clealry not too
important.

 - Maciej Stachowiak

-- 
guile -c '(define c call-with-current-continuation)(define(f s)(lambda(o)(let
((n(c(lambda(x)(o x)x))))(let p((n n)(l(string->list s)))(cond((not (null? l)
)(display(car l))(p(c n)(cdr l))))))))((f"Js nte ul akr")(f"utAohrGieHce\n"))'