Subject: Re: [Freesw] priorart.org
From: "Frank BENNETT (フランク ベネット )" <bennett@nomolog.nagoya-u.ac.jp>
Date: Mon, 14 May 2001 09:49:22 +0900

On Sun, May 13, 2001 at 02:38:46PM -0700, Greg Broiles wrote:
> At 07:44 PM 5/12/2001 -0400, kragen@pobox.com wrote:
> >Prior art the patent examiner considered during the patent application
> >process cannot be used to invalidate a patent, because the examiner
> >already knew about it when they decided the patent was novel.
> 
> Unless the examiner made a mistake.

Hmm.  I posted a note that pointed at the leading case on this subject.  If
that hasn't reached the list, let me know and I'll send it in again; the case
is useful, and the note offers some other comments relevant to this
discussion.

The law provides for reexamination proceedings inside the PTO, as well as
judicial appeals.  The law, in essence, is that the PTO _cannot_ reverse the
initial grant of a patent in reexamination proceedings, if the sole evidence
for doing so is prior art (documented to have been) referred to by the initial
examiner.  A court of law _could_ reverse, but the court's firm rule is to
defer to the specialist judgement of the initial examiner.

So from a free software perspective, there are two ways to play defense:
create a clearly indexed easy-reference database of prior art, in the faith
that the initial examiner will then be more likely to deny software patent
applications; or do nothing in that line, thus fostering a lower incidence of
prior art citations, and leaving more patents open to challenge at the PTO
reexamination and judicial hearing stages.

> >Making prior art easier to search makes the life of patent applicants
> >easier.
> 
> This is true if you assume that an applicant would prefer no patent or a 
> limited but defensible patent over a broad but vulnerable patent. I'm not 
> sure this is a good assumption, given the scope and defensibility (and 
> reasonableness) of issued patents, especially for software.
> 
> There is no obligation on the part of applicants to search for prior art - 
> only an obligation to disclose the prior art they're aware of. The less 
> they're aware of, the less they disclose, and the less the Patent Office 
> sees prior to issuance.

If you favor maintaining high search costs as the defensive strategy, the
procedural handling of patent challenges becomes important.  If the cost of
proceedings is high (it is), a repeat player with deep pockets will have an
advantage.  If there is provision for preliminary injunctive relief or
interim damages (I don't actually know if this is the case), the party
attacking the patent would need enough economic staying power to cover that as
well.  On the numbers, I think that the free software community stands to
suffer more pain from high search costs than its counterpart.

The numbers are not the only factor in the mix, of course.  Concentrating
efforts on a few, high-profile instances of litigation, argued from hard
principle, can heighten public awareness of the issues.  But it's risky to
abandon the strategy of "win little, win often" -- unless you are convinced
that it has nothing to offer (possible reasons being that one believes that
patent examiners are hopelessly corrupt, or that they are incapable of
reason).