[I've not yet found in the archives something which addresses this
question; if I missed it I apologise]
I am trying to set up a system whereby I can create lists of
addresses/servers of varying degrees of spamminess. For example, I
might have three lists:
From server constantcontact.com: +2
From server roving.com: +4
To user nos...@example.com: +10
I find myself in the situation in which most, but not all, of the mail
from certain entities is spam. Simply blacklisting them would generate
false positives, and I need to avoid that. However, absent other
considerations (like a low Bayes score) I want the message to be tagged
if not rejected outright (I'm using spamass-milter) based upon its
score. Being able to set up categories such as above would help in this.
Likewise, there are certain users on my system that get mostly spam, but
occasionally legitimate messages. For example, I hope my 8 year old son
doesn't need Viagra, but he does want to read his email from Grandma.
Increasing the score of messages destined to his mailbox, while not
tripping the spam flag in and of itself, would contribute to such a
designation when combined with other indicators.
How would I go about creating such rules? If this has been covered
before or elsewhere, a pointer would be welcome. Thank you!
-Don
