[Spambayes] Do you need to continue training ham?

[Skip Montanaro]

    Tim> Ideal would be to train on a small random sample of all the email
    Tim> you get, and expire training messages over time too.  That seems
    Tim> hard.

This doesn't help with the Outlook plugin, but the sb_upload.py (nee
proxytee.py) script accepts a '-r prob' flag which says to upload the
message to sb_server.py (nee pop3proxy.py) with probability 'prob'.  It
keeps you from drowning in messages which demand classification, but also
scratches the random sample itch.

FYI.

Skip