At 08:30 PM 6/21/03 -0400, Gordon Cormack wrote:Auto-learn and auto-whitelist use different scoring criteria from those used in spamassassin's spam filtering.
The bayes auto-learning does not use "it's own" scoring mechanism, it uses scoreset 0. This is the score the email would get by the main SA engine if bayes and network checks were off.
Is that really true that auto-learning doesn't take any notice of network checks ? :(
Certianly you do NOT want the bayes scores to feed back into bayes learning, If you can't see why, think "feedback amplifier with positive gain".
Absolutely... I've tried to explain this to people on the list before, but I don't think they believed me. However, I know full well what an uncontrolled positive feedback loop does :)
As for disabling the network checks for auto-learning, that makes sense to me as well, since the bayes code learns from text tokens, not IPs.
And there I disagree - yes, bayes learns from tokens, however what autolearning has to decide is *which* messages to learn, not what should be learnt about them.
It's obvious why whitelists and blacklists are ignored for autolearning scoring, and its also obvious why the bayes score is ignored for autolearning (positive feedback) but its not obvious to me why the RBL checks and/or razor should be left out ?
Surely if a message has several positive RBL checks, a positive Razor hit, and some normal checks, and that pushes it over 15, it should be autolearnt ?
Just IMHO anyway :)
Regards, Simon
------------------------------------------------------- This SF.Net email is sponsored by: INetU Attention Web Developers & Consultants: Become An INetU Hosting Partner. Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission! INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
