From what I understand, spamassassin can only give a static score, there is no way to have score between 0 and max-score. and if we need a variable score we have to create several rules like BAYES_00,BAYES_01,...,BAYES_99 , and let all these rules executed. Although only the first rule will scan the message and save the score in a static variable, and other bayes rule will use this variable if it is executed, but I think this is not really efficient.
If you think it's not efficient.. prove it first. Do some profiling.
. It would be nice if we have only 1 bayes rule, for example BAYES_CHECK, and this rule will give you spam probability between 0.0 and 1.0 (and mapped to the score between -5.0 and 5.0).
Erm, yeah, but how are you going to GA the score? The GA isn't going to output coefficients for a quadratic to use as a scoring formula. Clearly it's not a simple linear function, and not GAing the rule tends to defeat the purpose of SA's design.
------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01 _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
