OT: Getting bayes info from a seperate box(Spam trap of a sorrt)

[brichter]

I figured there had to be someone doing something similar with
MailScanner/Spam Assassin so I figured I would start here.

Has anyone setup another box with separate domain to help increase bayes
scoring?

What I was thinking of doing was setting up a separate box with Mail
Scanner/Spam Assassin with an unused domain of ours.  Put some email
addresses for this unused domain out there (web sites, Usenet groups, get
the ball rolling so to speak) so the Spammers get a hold of them and then
somehow pipe ALL mail that comes into this spam trap domain as SPAM
automatically into bayes that resides on this separate box.

Then somehow share this bayes database with my 3 real Mail Scanner/Spam
Assassin boxes that handle the valid email traffic for our real domains
that receive mail.  Would there be any issues since the destination domain
on all the learned Spam would be different then the domains on the 3 real
mail servers?

Is this a bad idea, can you do a check against a bayes database but not do
any learning? (I am thinking I would not want to have the 3 real mail
servers updating against the spam trap db)

If I did this, would I then want to just disable the bayes use on the 3
real mail servers and JUST have them compare against the spam trap bayes
database. (I am assuming this would increase the efficiency of the 3 mail
servers that would no longer maintain their own bayes databases, as they
currently do)

Waste of time? Or could this positively help fighting spam?

Thanks for any feedback on this.

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the Wiki (http://wiki.mailscanner.info/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).

Support MailScanner development - buy the book off the website!