One of the speakers at Spamcon 2004, talked about the effectiveness of automatically generated white lists. As I recall, his scheme depended upon two sources of info: the mail addresses that typically appeared in your To: From: and Cc: lines in your corpus of ham, during training and automatically collected from similar info. in mail that you send out. This of course assumes that you don't correspond directly with spammers. <g>
I was wondering how this might be integrated with SA. In particular, how does one intercept addresses on the outbound mail route? Would this be done in some sort of milter (like mimedefang, if you're using sendmail as your MTA)? It seemed to me that if one can build a more effective white list, that the number of false positives can be reduced considerably, and thus, the spam cut-off could be lowered, making sure more actual spam makes it over the dam. Separate question (may be a faq): is there a database (SQL, etc) implementation of from/received white lists? This would make updating a whitelist easier, and more efficient than hacking the user_prefs file, for example. ------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk