Thinks for your feedback. Here's what I want to do.
There will be only one user. An exim filter will save the message to a
file and execute a script that will pipe the message into a program I
wrote to modify the measage and pipe it into dspam. The output of
dspam, which I hope will be a fraction like spamprobe between 0 and 1,
will be poped into a perl script that will return words that will
indicate how spammy or happy the result is. This text result will then
be put into a header and Spamassassin with score it.
So - this should be fairly easy I hope. What do you think?
Timothy Spear wrote:
>Marc,
> I like it and I dislike it. :-)
> It is very fast, and learns very well. However, the primary
>disadvantage is that it is a post SMTP session based solution. At least how
>I have implemented it.
> The other disadvantage is that each user is required to once in a
>while go in and scan or review email which has been marked as spam (again,
>as I have configured it).
> Overall, I like it as the final layer of defense against spam which
>the users have interaction with. All other layers/tools are at the system
>level, and are controlled by the system admin.
>
>Tim
>
>-----Original Message-----
>From: exim-users-bounces@??? [mailto:exim-users-bounces@exim.org] On
>Behalf Of Marc Perkel
>Sent: Tuesday, November 15, 2005 9:32 PM
>To: exim-users@???
>Subject: [exim] Who likes DSPAM?
>
>Just looking for a quick opinion - who likes DSPAM?
>
>Been using spamassassin and spamprobe with it as a second bayesian
>filter. The second filter is doing bayes on modified messages. I enhance
>the headers and strip most of the body and feed the results back into SA
>for scoring. Been working really well except I'm having problems with
>spamprobe crashing and corrupting it's database.
>
>And - I'm not using MySQL with SA and I like it - so - looking at DSPAM.
>
>So - who likes/dislikes dspam and why?
>
>
>
--
Marc Perkel - marc@???
Spam Filter: http://www.junkemailfilter.com
My Blog: http://marc.perkel.com
