[Patents] natural forces

[PILCH Hartmut phm@a2e.de]

Gregory Aharonian wrote:

> >Yet, the distinction between the physical and logical world, between
> >mind and matter, is an eternal one, whose relevance never changed in
> >human history.
> 
> Wrong.  I do not comment on mind and matter, but at least for computing,
> history as written by the great mathematicians of this century, people
> like Turing, Church, Post, Chaitin and others have quite convincingly
> proved that the physical embodiment of an algorithm, hardware or software,
> is irrelevant to understanding the nature of the algorithm - i.e., its
> relevance to other algorithms (to determine novelty/obviousness in the
> patent world), and its general efficiency.

For algorithms that is of course true, I'd say even tautological.
But the patent world is not about algorithms.
 
> What we want to encourage people to develop is new solutions to problems.

I doubt whether that needs encouragement by monopoly.  We have to be more
specific than that.

> Whether or not the solution is implemented in hardware, software, optics,
> DNA computers, etc. is irrevelant to its patenability - in fact, such
> embodiments should not even be mentioned in independent claims, but rather
> dependent claims.  The independent claims should be reserved for 
> expressing the newness of the algorithm invented.

That would mean that once a certain abstract scheme of doing things
(algorithm) is found to work in mathematics, it must also be claimed for
biological, chemical, physical and any other kinds of solutions of quite
unrelated problems.  That is highly dangerous, I'd say.

The traditional approach of the patent system is to monopolise solutions
based on new empirical knowledge, i.e. algorithms containing some new
knowledge about matter that could not have been gathered by mere
abstract-logical reasoning.  It is true that every solution contains an
algorithm, but if it doesn't also contain something else, which is novel
and inventive, then it isn't a patentable solution.
 
> History evolves.  That our laws grew out of a mechanical past does not
> mean that we are forever fixed to just allowing patents on that which
> is mechanical.

All the categories that you and I stated above are eternal and unrelated
to history.  What evolves is something else:  the focus of economic
interest.  The danger that we have to combat is that the patent system
moves together with that focus.  A system that works in one area doesn't
necessarily work in another one.  Where its proper should be defined by 
philosophical categories, not by monetary interest.

If money is everything, why clamor about novelty and inventivity in
Patnews all the time?  After all trivial patents are good for the powers
that have the money.  And novelty/inventivity are only abstract concepts,
just as is "technicity/physical substance/empricial knowledge +
algorithm".  

The only difference is that the abstract concepts which you are advocating
are more widely accepted by patent lawyers and easier to explain.  So you
chose an easier evangelising strategy, we chose a harder one.  In the US,
perhaps you have no other choice.
 
> The relevance of the mechanical should not be forgotten,
> but not treated as gospel, especially since the patents laws as shaped
> by these mechanical and electrical histories were written before the
> great developments in logic and mathematics in the 20th century.  No
> where is the work of Turing reflected in any patent laws.  And if these
> theoretical arguments are insufficient, then one can reflect on the
> implications of hardware/software codesign tools, where designers do
> not even consider hardware or software while designing, leaving the
> embodiment choice to the computer to figure out which should be 
> turned into hardware and which into software, in multiple ways.

As I said many times before:  apparently the conclusion is that the
hardware innovation that comes out of codesign tools is purely algorithmic
and contains no new insights into causality of natural forces.  Maybe
innovations of that type should not be patentable at all.  Maybe some of
their hardware embodiments should be patentable.  Those are open questions
to be decided on the basis of many factors that stand outside this
epistemological debate.  What is clear however is that any patenting of
information works infringes basic freedoms such as the freedom of
expression and therefore needs to be excluded, unless giving up such
freedoms brings us substantial amounts of extra welfare. Even if someone
convincingly argues that in some contexts dictatorship enhances public
welfare, this may not persuade people to opt for serfdom.  In our case, we
even have strong evidence showing that swpat-serfdom leads to less 
welfare.

> So I find it inconsistent with these theories and tools to treat hardware
> and software differently when it comes to patenting, and strongly
> believe that people against software patents are actually against both
> software and hardware patents, and without too much of a leap, against
> all patents in general.  I can respect that view, but not the lesser
> argument of banning just software patents.

You should give up some of your "strong beliefs" and learn to respect that
view.  It corresponds to a basic anxiety in human nature about freedom of
expression.  Fighting against this anxiety is perhaps even more futile
than fighting against the monetary forces which you recommend us to adapt
our strategy to.  I can respect your view that we should not concentrate
on futile fights.  I can even respect your view that fighting for freedom
is futile.  But I can only despise those who lend their hand to the
monetary forces for jointly fighting against freedom.

-phm