----- Original Message ----- From: "Alton Danks" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, March 18, 2004 3:16 PM Subject: Anyone not using Bayes?
> Hello, > > Is there anyone who is not using Bayes? > Why not? > What are you doing above and beyond the default SA config? > > We are not and I'm wondering just how alone we are. > > Al > I'm the opposite of the scenario you are looking for. We use Bayes only except for the default rules that come with SA. We are a small office with 6 active mail users. We have a Bayes database that is site-wide. It currently has been trained with 2300+ spam emails and 13000+ ham emails. We use SA in conjunction with MailScanner and ClamAV. All of these are configured "stock out of the box" ie no major changes in the configurations. We use RBLs and block about 4500 spam email attempts a week before SA even sees them. SA picks up any others that comes through. These get placed into a Spam folder via procmail. I and my office collegues may receive 1 un-marked email of spam a day. If we receive anything we save it to an IMAP Spam folder. If we get a FP, we save it to an IMAP Ham folder. False positives are very infrequent. They usually occur when joining some newsletter. Once it is fed through as ham we don't see any more FPs. Every night the system runs sa-learn on the appropriate folder. We never see that type of spam again. I know we are a small shop so I'm sure other's milage may vary. However, for us, just using bayes with the stock rules and RBLs has been a huge success with minimal time on my part administering the mail system. Hope this helps, Mike
