[107416] in Cypherpunks
Babbage's Revenge; Patenting any algorithm reducible to software
daemon@ATHENA.MIT.EDU (Robert Hettinga)
Wed Jan 13 17:49:06 1999
In-Reply-To: <36a00278.43123235@world.std.com>
Date: Wed, 13 Jan 1999 17:04:52 -0500
To: hapgood@pobox.com
From: Robert Hettinga <rah@shipwright.com>
Cc: dcsb@ai.mit.edu, Digital Bearer Settlement List <dbs@philodox.com>,
e$@vmeng.com, cryptography@c2.net, cypherpunks@cyberpass.net
Reply-To: Robert Hettinga <rah@shipwright.com>
At 3:36 PM -0500 on 1/13/99, Fred Hapgood wrote:
> Is it as obvious to everyone else as it is to me that business model
> patents, especially models involving the internet, are absurd from an
> enforcement perspective? I have no idea how US lawyers finesse the
> conflicts with anti-trust in this country, but I can't imagine
> anti-trust regulators in any other country respecting a patent
> that tries to grab for a monopoly this comprehensive. Am I
> wrong about this??
You shoulda been there, Fred. :-).
At the DCSB meeting last week, we found out that this ability, to take
*any* algorithm, including a business process, to reduce it to practice in
a computer, and then to claim patent on it, has been inherent in patent law
all along, but that it took the actual emergence of computer technology
itself to make it manifest. Call it Babbage's revenge. Or maybe Russell's.
It's a logical consequence of the very way that the entire structure of
patent law has emerged, which makes it very hard to deal with. Particularly
since patents were included in the constitution, and there's a lot of
case-law built up around it.
My own personal hack, short of the Shakespearean one, :-), is to claim that
Moore's Law has so accellerated progress that the *length* of patents on
computer-based should be reduced from 20 years after filing, which was
changed from 17 after issue to comply with the WIPO/GATT, whatever, to
somee thing much more reasonable. I'd take the divide-by-seven approach
that Andreeson et. al., concocted at the same time they came up with the
phrase "internet years".
It turns out that 20 years divided by 7 is 2.85, so let's say three years
from filing? :-). The number is a made up one, anyway, and it has
oscillated several times in the history of patents in this country. and
there was something which came out this very morning about how ungodly long
the *copyright* duration is at the moment, how in the matter of a few
decades, it has gone, de facto, from 28 years, to 73 years after the death
of the owner, to 93 years in the case of certain copyrights which should
have expired this year but won't because of some comprimise in the new
international copyright law. In fact, it maybe that this whole armature of
*international* patent law is the problem, as progress happens in different
places in different times, according to the type of economy and business
you're talking about. 20 years from issue makes perfect sense on
agricultural, or even industrial, time-scales, but it certainly doesn't on
a geodesic one.
More practically for patents themselves, which depend on uniqueness and
prior art, an effort is under way, someone at the meeting said, to scan in
all kinds of computer journals, from as far back as possible, and put them
up on the web, free of charge, so that computer patent examiners have some
rather significant evidence of prior art to go by. (My take is that the
USPTO saw itself as a strictly *industrial* enterprise, and that they
didn't think they *had* to do this computer stuff :-), and that they had
ignored algorithms, software, and electronic business practices
accordingly, but that the very structure of the law, with help from some
computer patent lawyers, :-), forced the issue in the last decade or so.)
Finally, somebody else has suggested that we should just march on
Washington and force the patent examiners, at gunpoint, to actually read
Knuth, and then the copyright date, and that that would solve the entire
problem. ;-).
Cheers,
Robert Hettinga
-----------------
Robert A. Hettinga <mailto: rah@philodox.com>
Philodox Financial Technology Evangelism <http://www.philodox.com/>
44 Farquhar Street, Boston, MA 02131 USA
"... however it may deserve respect for its usefulness and antiquity,
[predicting the end of the world] has not been found agreeable to
experience." -- Edward Gibbon, 'Decline and Fall of the Roman Empire'
