On Thu, 26 Feb 2004 21:40, Christopher Sawtell wrote: > On Thu, 26 Feb 2004 21:07, Nick Rout wrote: > > On Thu, 26 Feb 2004 20:27, Christopher Sawtell wrote: > > > I am protected by the Brightmail product installed at Paradise. > > > It catches about 90% or so. Leaving me to deal with about 10 at the > > > most per day. So that means that pre-filtering I get of the order of > > > 100 per day. A fair few of the Nigerians get through, and a few about > > > prescription drugs. Brightmail seem to have got the virus plague pretty > > > well under control. > > > > > > I have kmail filters set up so that color=#ffffff and charset=Big-5 are > > > spam tags. That gets rid if a fair number too. > > > � > > > > well brightmail won't fit in my mail chain as it doesn't pass throuugh my > > isp's mailserver (and they don't filter stuff that they queue as a backup > > mx) > > > > are you using spamassassin on your box? > > No. Just as I was in the middle of "doing something about viruses and spam" > Paradise introduced the Brightmail product and I found that other things > were more interesting or otherwise important. However I did do a fair > amount of research and experimentation. My purely subjective choice was for > the SpamOricle project. I got as far as building it, training it and > testing it out. It seems to go really well. I trained it with thousands of > messages both good and bad. > > I read many well constructed and cogent arguments which suggested that > SpamOricle is superior to SpamAssassin. My tests would support that. > > > I tend to use many different clients on many different boxes so client > > filtering is no good to me. > > I can appreciate that. > > imho SpamOricle would suit you to a 'T'. > http://pauillac.inria.fr/~xleroy/software.html
I may look into it. spamassassin is doing famously now I have the version with bayesian filtering.
