--On Friday, November 01, 2002 6:53 AM -0500 Sean Brunnock
<brunnock@server.com> wrote:
> OK. Then explain something to me. How does SpamAssassin differentiate
> between a solicited bulk message and an unsolicited bulk message?
Just to explain one more time, SpamAssassin applies a bunch of (mostly
regular expression) tests against the headers and body of an email, driven
off a list of 'rules' each of which has a certain 'score weight' attached
(which can be plus or minus). At the end of the email it adds up the score
and compares it to a threshhold which you have set in your .cf file, and if
the threshhold is exceeded, it either flags, redirects, or drops the
message, again according to your chosen config file.
Given two otherwise identical messages, both originating from
non-blackholed mail hosts (since one of the optional tests is for
blackholes), one "solicited" and the other "unsolicited," SpamAssassin
would have no way to tell. (There is no such thing as an X-Solicited:
header, and you can bet that if there were, it's the first thing spammers
would include.)
As installed out of the box, SpamAssassin would most likely let BOTH
messages through to your inbox unmolested. If you tune the scoring or the
threshhold downward so that most bulk email is blocked (which some prefer
to do), what you need to do is "whitelist" your chosen list memberships, so
that, for example, list-managers-owner@greatcircle.com as the Sender is a
-10 (or lower) score.
As a practical matter, it might be nice to have "on file" a list of howto's
that allow your managed list to be whitelisted for the major antispam
packages. Then when people ask, you are ready with a recipe.
References:
|
|
