<SNIP>
for my money, the best time spent is following the *pattern*
based filters and
working on ways to share that information amongst others of
like interest. a good
start would be a site dedicated to the sharing of procmail
recipes, beysian formulas,
etc.
<SNIP>

Yes, this sounds like a good starting point.
No need to debate the value of open-relays; there is a need,
but not in the public circuit.

but to bring this back to James, we are looking at
implementing mailets to process SPAM.
So far, the process is to consult black-hole lists and to
feed mail through a Bayesian filters and what not.

One problem I see with this is that black-hole lists are
arbitrary and pattern matches are too inclusive.

Spam filtering needs to be based on user preference as each
instance will be different; to quote

"
Bayesian spam filters are content-based filters that

- are specifically trained to recognize the individual email
user's spam and good mail, making them highly effective and
difficult to adapt to for spammers. 
- can continually and without much effort or manual analysis
adapt to the spammers' latest tricks. 
- take the individual user's good mail into account and have
a very low rate of false positives. Unfortunately, if this
causes blind trust in Bayesian anti-spam filters, it renders
the occasional mistake even more serious. 
" (http://email.about.com/library/weekly/aa100702a.htm)

The effective SPAM blocker system will be bayesian based
user specific systems.

James currently touts a Bayesian mailet, but employs only an
overall data source and is not concerned with individual
preference; to be an effective SPAM blocker, a relationship
needs to be established between a specific user and her
Bayesian lists.

and of course the necessary functionality to maintain the
lists ...

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to