Re: [dspam-users] Model email service

* Chris Ryland <cpr@emsoftware.com> [2005-12-30 14:50:24 -0500]:

> No, my point is that DSPAM is *perfect* for us--it did take me quite
> a few hours over many weeks to learn it, set it up (twice, actually),
> and train it, but it's infinitely worth it. We've basically solved
> our spam problem, even with very public and well-known (12 years)
> email addresses.
>
> We've used SpamAssassin on a previous server, and it was effectively
> useless.

I agree with you about DSPAM. At the moment I'm only using it as a
single user on our server, but I'm already looking forward to
implementing a shared group and inoculation for the rest of the user
base.

I mainly got fed up with SA because of its enormous resource hunger.
Detection rates where OK, actually, when using the bayes filter. I
can imagine that it wasn't good enough for you if you've got
quasi-honeypot mail addresses in use.

DSPAM is much leaner with as least as good results as SpamAssassin,
but the lack of really good documentation and some pitfalls are its
major downside in my opinion. This is probably because there is only
one maintainer who doesn't have that much time to polish the docs.

--Frank
Received on Fri Dec 30 15:11:40 2005