On Wed, 14 Jan 2004 13:48:00 +0100 Kristian Köhntopp <[EMAIL PROTECTED]> wrote:

> Please have a look at the article at http://www.jerf.org/writings/
> bayesReport.html from Jeremy Bowers. Jeremy has written a tool shown at 
> http://www.jerf.org/images/spam.writer.full.png that assists in writing spam 
> that does not trigger pure bayesian filters.

... based on the corpus available to the sender.

Therein lies the problem with his approach; he can only tune his message
to get past his own filters. Assuming most people have similar mail
corpora (as in the case for site-wide Bayes setups), his approach may
reduce Bayes' effectiveness some. I'm not worried about it.

I quesstion some of his motives and logic (paraphrasing: "let's reduce
filter effectiveness in order to shift focus to other ways of fighting
spam like e-postage and web-of-trust".) The way you stop spam is by
denying spammers access to the internet and by securing resources so
they cannot be abused to send or receive spam. Discovering the social,
legal, economic, and technical means to do that are left as an exercise
for the student.

It's well known that filtering is *only* useful for keeping one's inbox
uncluttered; it does nothing to interdict the flow of crap from
upstream. You want to put a serious dent in spam? IDP broadband
providers that give their customers direct access to port 25 on remote
systems by default. Spam from AOL dropped to almost nothing once they
did that. It's not trivial to filter outbound port 25 traffic but if you
want to provide large-scale broadband access, it's mandatory to curb
abuse. It takes money, expertise, and most importantly, the corporate
will to be a good net neighbor and take responsibility for their users'
actions. Most broadband providers have all but the latter; an IDP gives
them incentive to get it or lose the rest.

-- Bob


-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to