See http://www.infoworld.com/article/04/06/01/HNnaantispam_1.html for a
report on the absurdity of the US patent system - a system which the US 
is trying to ram down the throat of the rest of the world (eg the
European patent law "reforms", and the mooted changes to Australian
patent law under the proposed US-Australia Free Trade Agreement).

Some of the claims mentioned for this particular patent are doubly
absurd, particularly the use of Bayes rule for email classification,
because such use is obvious from the literature - there are papers which
discuss the use of Bayesian methods to classify email and Usenet
messages going back many, many years, although they don't mention "spam"
because the word meant processed ham when the papers were written, not
unsolicited email. Furthermore, Paul Graham posted an influential and
very widely read paper discussing the use of Bayesian techniques for
spam filtering in August 2002 (see http://www.paulgraham.com/spam.html
), several months before the application for the patent in question was
filed. In all likelihood this patent can be invalidated, but it now
means that time and effort has to be invested in doing so, and in the
meantime, researchers and implementors will be wary of using Bayesian
techniques in spam filters. Thus useful progress is yet again seriously
impeded by software and algorithmic patents - in part due to sheer
stupidity by (and perhaps deliberate underfunding of) the US patent
office (and US patent laws) and greediness on the part of the of the
patent claimant, who must have known that there was prior art for their
claims, but who filed them nonetheless. 'Twas ever the way of the world,
I suppose.
-- 

Tim C

PGP/GnuPG Key 1024D/EAF993D0 available from keyservers everywhere
or at http://members.optushome.com.au/tchur/pubkey.asc
Key fingerprint = 8C22 BF76 33BA B3B5 1D5B  EB37 7891 46A9 EAF9 93D0



Reply via email to