Subject: Re: "incentive void" (was Re: A different patent covenant...)
From: <stephen@xemacs.org>
Date: Thu, 5 Oct 2006 21:11:04 +0900

Norbert Bollow writes:

 > No, Ben simply pointed out that (a) was not involved in the
 > scenario and question that I was responding to, which was in
 > reference to a specific example of an innovation.

(a) is necessarily involved, because (a) is a source of benefit from
the patent system.  If you exclude that from consideration, of course
you're far more likely to get the answer that the patent system
sucks.  This is no different from the practice of patent advocates who
ignore the search costs of knowing whether there's a relevant patent
or not, and then come to the conclusion that the system is great.

There are other ways to analyze, for example, show that demand-driven
and demand-inducing innovations can be distinguished in court and *be
known in advance of the patent application*, and thus could
practically be subject to different rules.  But nobody here has even
tried to do that.  (The emphasized phrase is a sine qua non for the
incentives; if it can't be determined ex ante, then the two types
can't be separated for analysis of incentives.)

 > I have argued that that specific example was a demand-driven
 > innovation, and that this appears to be a typical property of the
 > software patents that I have observed "in the wild" during the past
 > decade or so.

Which means nothing without justifying your extrapolation to the whole
set.  You don't have to do a full statistical study of the hundreds of
thousands of existing software patents, but you do have to justify
projecting your necessarily very limited experience to the population.

Nor am I at all convinced by your "argument" that the example was
demand-driven.  For starters, you didn't refer explicitly to even
*one* claim in the patent application to show a correlation between
current demand and the patent.  Furthermore, a correlation does not
create causation; given that this is basically academic work, it seems
reasonably likely that it came from out-of-the-blue based on
considerations of possible applications of a technology Rob Cameron
knows well, rather than any plan to satisfy particular market demands.
(You are referring to the Cameron technology, right?  Not that it
matters, I don't recall you ever distinguishing patents from claims.)

How can you say, except by asking Rob Cameron?  How do you prove
otherwise, if he says "Steve has the right of it"?

 > Software patents are particularly problematic for demand-driven
 > innovations during the period of time during which demand for that
 > kind of innovation emerges.  Any attempt to achieve a meaningful
 > reform should provide convincing arguments that the proposed reform
 > resolves the problems in that area.

Nonsense.  If you insist on that, it's very likely that any reform
will proceed without *any* consideration of what you want.  A full
resolution of any of the problems of the patent system is unlikely,
for one thing.  For another, a proposal needs only to provide strong
evidence that it will resolve *some* problems to be a meaningful reform.

 > Do you believe that it is possible, in any manner which is more easily
 > achievable than total abolition of software patents, to reform the
 > software patents system

Yes.  Legislating the International Characters covenant (or some
descendant) into the U.S. software patent system will not be easy, but
it will be much easier than abolition.  And it will be a major
improvement from the point of view of my own usage of open source
software.

 > so that in the reformed system, at least some kinds of software
 > innovations are patentable, but it is not possible for a company
 > like Microsoft to impose essentially-unlimited patent-related legal
 > and/or other costs on an emerging Free Software based competitor?

If by "competitor" you mean "business", I don't care.  Creating
special privileges for free software *businesses* is not a goal that I
support as a matter of principle.  Microsoft is a threat to *all*
businesses (not just those in software!), and should be treated that
way.  FSBs are a red herring here.

In fact, as an economist, I believe that the encouragement of true
competition *requires* the frequent demise of competitors.  If the
deceased are disproportionately FSBs, I'll be very sad, of course.

What I am willing to insist is in the public interest is encouraging
software distribution with source and broad permission to modify and
redistribute.  If this is also applied to commercial redistributions,
so much the better.  However, free commercial redistribution is not an
absolute requirement for me at this point in time.  Sorry.

 > How would one go about trying to achieve such a quantitative
 > measurement?

Dunno yet; I'm a theorist by nature.  I'm trying to encourage interest
from some empirical folks and local funding sources.  It'll go faster
if you send me money. ;-)