From: Greg A. Woods (no email)
Date: Tue Sep 03 2002 - 16:52:06 EDT
[ On Tuesday, September 3, 2002 at 09:54:05 (-1000), Clifton Royston wrote: ]
> Subject: Re: Using blacklists and RBL's with Postfix
>
> In the real world, doing this sort of computation (and dynamically
> reweighting the filters for each user with each incoming message) has
> serious costs. If just running a predefined set of regexes on incoming
> mail can be too expensive, how expensive is recalculating your
> statistical base for each incoming mail going to be?
Running your mail through an arbitrary list of regular expressions,
especially when within the framework of some massively over-bloated
interpreter such as Perl or Python, is likely going to be one hell of a
lot more expensive than a decent implementation of a simple word sorter,
so long as you're not doing your lexical analysis with naive regular
expressions. Even PCRE can't make using REs really quite so cheap.
People have been implementing, or at least talking about using, Bayes'
Theorem for computer analysis of text since the mid 1960's or so (eg. as
noted in "Automated Language Processing", by Harold Borko, Wiley & Sons,
1967).
I haven't yet done an analysis of my own rather large corpus of personal
e-mail, but I expect, as ESR says of bogofilter, that the biggest
problem will be finding an efficient way to store and access the word
lists. However even /usr/share/dict on my machine contains just over
4MB of uncompressed word lists. I'd gladly keep 8MB of RAM in my mail
client dedicated to the cause against spam! ;-)
> This also presumes (like every other anti-spam silver bullet) that
> spammers are completely incapable of adapting, which has been
> repeatedly shown to be wrong.
Do you have a scientific criticism of Graham's assertion about how
robust a Bayesian filter should be even when the "attacker" knows
exactly what algorithm is being used?
-- Greg A. Woods +1 416 218-0098; <>; <> Planix, Inc. <>; VE3TCP; Secrets of the Weird <> - To unsubscribe, send mail to with content (not subject): unsubscribe postfix-users
|
|
|