I've been using SA for a little while now but am still not up to speed on all it can do. We are mostly a Windows shop here so we have very few techs who know Linux that well. We're using SPAMD to filter email and use a Windows-based email proxy to send/receive info to/from the SPAMD host.
It doesn't appear to exist but I was wondering if there was anyway to 'learn' or 'forget' emails for bayes classification through the SPAMD protocol? If not, I would respectfully suggest that as a future feature enhancement to the protocol. As it is now, we have no way of learning or forgetting false positives (or true negatives). Thanks.
<<attachment: winmail.dat>>
