I'm very interested in your opinion on this, and want desperately to
follow it.  You've already put me on the right path of adding to the
spamdb and spamdb.helo instead of starting anew and running rebuild
spam frequently.  That's definitely helped.

Currently we outright reject on DNSBL, URIBL, and Bayesian and score
the rest like suspicious helo.  I'm not set on these ways.

Maybe our corpus isn't good enough, but it always hovers around 1.0,
of course that could be with incorrect data.

When I reject due to bayesian, it goes to discarded and then I'm
manually watching that folder for the time being, adding the confirmed
messages to the spam folder.

I have bayesian set to reject because there are a lot of messages from
hosts that get by greylisting by retrying, come from ip addresses that
aren't blacklisted, and don't contain any bad urls.  Take messages
send from a gmail account (through google's servers), that's just an
ad for something with a url that itsn't in the uribl yet.  I don't
know how I could  block a message like that without bayesian
rejection.  If I'm understanding correctly, the only place it would
get a score is bayesian and that wouldn't be enough to reject on its
own.  I suppose we could have a perfectly crafted bombre, but i don't
know how to keep up with the ever changing messages to do that.

I'm always afraid of being to strict and having false positives.  At
the same time, there are lots of administrators who have a conniption
fit if a spam message slips through to them.

I'm all ears when you h,,ave a moment.

THANKS so much for all of your help.  You're directly helping hundreds
of people here, me especially.  I can't imagine email life without you
and ASSP.


On Sat, Oct 24, 2009 at 10:14 PM, Fritz Borgstedt <[email protected]> wrote:
> ASSP development mailing list <[email protected]>
> schreibt:
>>missed that the new version is out and fixes the issue.
>
>
> By the way: Bayesian blocking and all other scoring is not a
> recommended approach for ASSP.  I dare to say: to the contrary.
> SpamAssassin is working in this way. ASSP should not.
>
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry(R) Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay
> ahead of the curve. Join us from November 9 - 12, 2009. Register now!
> http://p.sf.net/sfu/devconference
> _______________________________________________
> Assp-test mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/assp-test
>

------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Assp-test mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/assp-test

Reply via email to