Jeffrey Weiss <email@example.com> writes
>May I disagree. I'd love to block using sendmail, but users would be
>within their rights to sue if mail addressed
>to them--spam or not--is filtered out without their request.
>It depends on your client base. ISP or employees?
luckily my clients are employees of the university and not prone to suing.
Nevertheless I share the concern of being careful in preventing delivery,
and the filtering system I implemented classifies spam into definite or
probable classes. Messages considered definite spams are not delivered to
the addressed recipient, but this only happens for rules I've added in
response to seeing particular spam messages. Possible spams have an
"X-SPAM" header added and it is left to clients to filter on that basis,
should they trust my rules. At worst I receive messages classified as spam
and can forward any mistakes onto the recipient.
I think I gain more flexibility over sendmail filtering because I use a
hacked version of the TIS fwtk smapd frontend to sendmail which allows
filtering on arbitrary headers (and the envelope & body if I choose to).
Does anyone know whether there's a system in place for disseminating spam
rules? I'm thinking of something like a trusted hierarchy using ssl to
deliver to 'subscribers'.