At 8:21 AM -0600 4/21/05, Jonathan Duncan said something about:
Ok, I found the "bayes_path" option, thanks!Here are two options I do not understand. If I am going off the score that spamassassin gives a message why would I want it in my bayes database? If SA already knows it is spam or not spam based on my score level how would these options be helpful?----------------------------------------------------- bayes_auto_learn_threshold_nonspam n.nn (default: 0.1)The score threshold below which a mail has to score, to be fed into SpamAssassin's learning systems automatically as a non-spam message.bayes_auto_learn_threshold_spam n.nn (default: 12.0)The score threshold above which a mail has to score, to be fed into SpamAssassin's learning systems automatically as a spam message.-----------------------------------------------------
Becuase SA can sometimes be wrong *gasp*, so you don't want to feed spam into the bayes DB if its 'kinda-spammy', only if its 'VERY spammy!'
i.e. don't feed items with a score of 6 into the DB, only scores of 12 and above.
Ditto for making things very 'hammy' - you don't want to feed an item of 4.5 into the ham database, since it will have lots of spam elements. But an item of .1 (or even a negative number!) is definitely clean of spam-elements and should be fed to the DB.
In this way, the spam/ham dbs are built on their own as people get more and more spam. yes, re-feeding spam in thru sa-learn is the best way to do it becuase it then catches spam that was originally not caught.
Matt -- :-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+:-:+ Matthew I. Cohen http://www.iwbyte.com/ It Won't Byte Web Design & Hosting. Ob. Quote: "Reality is for those email: lists@xxxxxxxxxx people who can't handle Fantasy" ====================================================================== This is <vps-mail@xxxxxxxxxxxx> <http://www.perlcode.org/lists/> Before posting a question, please search the archives (see above URL).