Hi Ronan, I myself have been looking at this for the company I work for, and may be able to provide some insight.
DSPAM is an entirely statistical filter, so is similar in behaviour to the BAYES part of SA. It implements a number of more sophisticated algorithms and strategies for recognising spam. It also has some nice touches for feedback (because it is virtually essential that you tell DSPAM when it gets it wrong) SA relies primarily on its regexp matching (with lookups etc). So, each spam detector works in a very different way. One of the options we are considering here is to use both SA and DSPAM together. There are a few HOWTO's around on how to do this, (I've seen a good one using exim, but can't find the link atm) or you could probably write a plugin to SA 3.0 to call DSPAM. Basically, you need statistical filters because spammers are tricksy, but you also need pattern matching, because when you spot something odd, you want to be able to start catching it immediately. Using a two-pronged attack also allows you to use SA as a detector for training the DSPAM filter until it gets good. Hope this helps, Richard -----Original Message----- From: R McGlue [mailto:[EMAIL PROTECTED] Sent: 18 February 2005 09:58 To: [EMAIL PROTECTED] Subject: alternatives I am aware that this is a SA list.. and my posting my generate some flames... but can anyone perhaps highlight some of the main plus points of SA over .eg. Dspam?? One of my objectives here in work is to (even though we are enjoying the best anti spam experience in recent history since i upgraded and twaeaked SA) my boss' are looking for more... not in itself a bad thing. The see the high cpu and memory overhead of SA as a problem so I have unearthed Dspam and am looking into it now. They give a short desciprition on their website as to how it compares ( albeit favourably ) to SA, perhaps this might be useful for SA?? thanks ronan --------------------------------------------------- This email from dns has been validated by dnsMSS Managed Email Security and is free from all known viruses. For further information contact [EMAIL PROTECTED]
