From daniel.probst@gmx.ch  Thu Nov 30 11:15:03 2000
Date: Thu, 30 Nov 2000 12:15:03 +0100
From: Daniel Probst daniel.probst@gmx.ch
Subject: [Patents] Hardware/software codesign

----- Original Message -----
From: "Gregory Aharonian" <srctran@world.std.com>
To: "Robert Bielik" <Robert.Bielik@gyrosmicro.com>;
<patents@liberte.aful.org>; "Gregory Aharonian" <srctran@world.std.com>;
"Bernard Lang" <Bernard.Lang@inria.fr>
Sent: Wednesday, November 29, 2000 7:37 PM
Subject: Re: [Patents] Hardware/software codesign


> > that is an unsubstantiated statement, contradicted by economists.
>
> Here you lie.  SOME, and I repeat, SOME economists, have come out
> critical against software patents.  Many other economists haven't,
> partly because it is very difficult to economically measure the
> impact of patents on any economic process, given the complexities
> of today's marketplace.
>
> SO if you are abandoning mathematical logic, and relying on economics
> to argue against software patents, I encourage you to do so (only
> because you are completely wasting your time doing so).
>

Ehem; economics does not abandon mathematical logic. IMHO as a theoretical
economist a sound basis on a formally consistent, mathematical model is a
prerequisite.

However, I must side with Greg. In the economic theory concerning patents,
software is nowhere seen to be conceptually different from other issues of
intellectual property (biotechnology, textile machines, business practices
etc.). The main concerns are to trade off welfare losses due to patents,
versus welfare losses due to insufficient incentives to innovate. Important
factors influencing this tradeoff are market structure, licensing
inefficiencies, competition effects due to licensing etc.

Nowhere does one make a clearcut conceptual differentiation between
software/hardware (no, not even in the theoretical part of Bessen and
Maskin). Actually economics is lacking a good model of knowledge acquisition
(I 'm racking my brains trying to find soemthing sensible). All models I
know of circumvent the critical issues by giving products a quality
dimension (say the real numbers). Effort (real number) then gives rise to a
probability  of a quality jump. One can then talk of infringement when the
resulting quality is too close to the patented quality. Needless to say
that, while these models do give valueable insights, they are very very far
from reality. Bessen and Maskin's theoretical work makes the point that
complementarities are important in the context of cumulative innovation
(while complementarities are a feature of software, they are not confined to
software).

There is however a difference with software, in that it can/is protected by
both copyright and patents (at least in the US). Reverse engineering code is
quite costly (maybe even proportional to developement costs?). Therefore
copyright  seems sufficient to allow  firms to recoup their developement
costs (nobody in Europe reverse engineers Windows 2000, rewrites the code in
Pascal with French comments and sells the compiled product for half the
price).

Casual empiricism (Bessen Maskin empirical part) suggests that copyright
protection seems to have been sufficiently strong to ensure innovation, and
that the problems of current patent systems seem to outweigh the benefits.

However nothing in economic theory disputes Greg's point that a more "ideal"
patent system with stringent novelty rules might do better than copyright. I
am personally somewhat sceptical to the short/medium term feasability of
such better patent systems, due to the vested interests of the involved
parties. Even glaringly obvious problems like patent offices obtaining fees
for patents (which distorts patent offices incentives) seem to be
politically difficult to tackle.

I have yet to understand the Technizitts concept put forward in this group.
I am yet undecided as to whether I should 1) accept it at a shallow level as
a rule of thumb for patentability, or whether I 2) completely disagree with
its consistency at a more deep (but presumably empirically less important)
philosophical level.

Even if I were to stick with 1), I see no economic arguement why this should
be the basis of a general, optimal patent system. At best it might be second
best given political, social constraints which have yet to be precisely
specified.

I am currently rereading Kolle a few times in the hope that the process
converges to some understanding.

-dan

----------
daniel.probst@vwl.uni-mannheim.de
Dr. Daniel Probst
Economic Theory
Mannheim University
Germany

> Greg Aharonian
>
> _______________________________________________
> Patents maillist  -  Patents@liberte.aful.org
> http://liberte.aful.org/mailman/listinfo/patents
>





