[Mimedefang] spamass-milter or mimedefang ?

[Kelson Vibber]

At 09:50 AM 8/21/2003, Marcus Schopen wrote:
>What are the basic differences between spamass-milter or mimedefang?

I haven't looked into much detail on spamass-milter, but I can tell you 
that MIMEDefang is very flexible.  You can set it up to take different 
actions depending on score level or specific rules.  Suppose you wanted to 
reject anything found in both Razor and Pyzor, or convert HTML to text (or 
call Anomy to sanitize it) only if the spam score is higher than a certain 
number, or set up multiple score ranges so that you can label, reject, or 
silently discard a message based on its score.  You can pass through, 
alter, tag, reject, and redirect messages.

MIMEDefang also lets you take action based on non-SpamAssassin 
criteria.  For example, you can have it check to see if the remote server 
is identifying itself using your IP address or hostname.

I have my system set up so that it labels anything above 5 and rejects 
anything above 25.  It also rejects anything found in Razor or Pyzor with a 
score higher than 10, or anything found in more than three DNSBLs (you have 
to pick these carefully, though, and avoid sites that mirror each 
other).  Exceptions are made for Bonded Sender and Habeas.  If the source 
is one which I know does not accept incoming mail, I have it set up to 
discard instead of reject.  Individuals have some limited customization - 
they can bypass SA altogether, or they can choose a different rejection 
threshold (setting it high to never reject mail, or setting it at lower, 
riskier levels).

I don't know if you could set up this sort of configuration in spamass-milter.

Of course, the cost of this flexibility is that you do need to have some 
Perl knowledge in order to make these customizations.  The sample filter 
works very well, though!

>I just need spam scanning. Virusscanning is done by amavisd.

Doesn't amavisd also let you use spamassassin?

Kelson Vibber
SpeedGate Communications <www.speed.net>