I must have been unclear - A bayesian spam filter is based on the probability that a given mail is BAD or GOOD from an analysis of a sampling of BAD and GOOD mail.
As a user, I would have control over this sampling - described earlier. As a James user, I would like to see this functionality implemented. This differs from your earlier view in that I see a personalized collection as opposed to a larger, centralized collection. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
