Virtual mailuser with their own bayes db?
cpedaschus at gmx.de
Mon Jun 12 00:41:29 IST 2006
Kai Schaetzl wrote:
>Christian Pedaschus wrote on Sat, 10 Jun 2006 21:19:28 +0200:
>>At least 200 spam/ham, i know, but that's no problem, most of the users
>>have >2000 mails in their box (without all the spam).
>I'm not so much talking of the minimum. There's a difference if you have 1
>Mio Bayes tokens for each single user (site-wide) or only 10.000 for each
That's true, no doubt.
Which brings me to the question: "does anybody know if bayes take the
recipient into his calculation"? If so, it would clear all my concerns
about virtual domains.
I must admit i didn't enter this question in google (yet), just came
home and checked my mails before going to bed ;)
More information about the MailScanner