[NTLUG:Discuss] Postfix Spam Blocking

[Victor Brilon]

Brian wrote:
> 
> That's seems to be the stock answer to anyone who claims SA isn't
> doing what they think it should do :)  
> 
That's generally because it's true to be honest. Seriously, read the SA 
mailing list for 5 minutes and you'll see what I mean.

> Version 2.61, no network checks, Bayesian training up and running,
> with autolearn.  Solid base of spam and ham, with daily feedings of
> both in small quantities.   

There's a *HUGE* part of your problem. SA 3.x is much more clever about 
the new types of spam that are out there. You should really consider 
upgrading if you're able to.

> 
> Don't get me wrong:  The SA rules work great.  The Bayesian training
> sucks (at least on my specific configruation).  After many hours of
> RTFM activities, I've just about decided the learning curve necessary
> to get SA to do what it's supposed to do is simply not worth the time
> invested.

This might not be the place for it, but if you can point out specific 
issues you're having with SA Bayesian, chances are someone has already 
solved the same problems and documented it. I'd be happy to help if I can.

Not trying to turn this into an SA advocacy thread btw ;) It's been the 
best tool I've been able to find for getting rid of spam in my life and 
I'm glad to pass on any knowledge I've gained to help others.

Victor