So is there no good way possible to automate learning spam and/or ham from
users? It sounds like the only way possible to do this without losing
headers and/or encoding formats is if I intercept everybody's email before
they pop it off and go through it manually to separate the ham from the
spam. You can't just stop learning because as the spams change the
filters will become useless over time... but going through everybody's email
manually would be an impossible task. There must be a better way to handle
the spams that get through to the users....
-Jeff
----- Original Message -----
From: "Loren Wilton" <[EMAIL PROTECTED]>
To: <users@spamassassin.apache.org>; "'Theo Van Dinter'"
<[EMAIL PROTECTED]>
Cc: <users@spamassassin.apache.org>
Sent: Tuesday, February 28, 2006 1:09 AM
Subject: Re: question on training spamassassin
Thanks. I read the wiki.
This is unfortunate because many clients are still using this client:
"Microsoft Outlook Express: It does not appear to have a redirect option"
True statement, but not necessarily important.
I'm running OE, and I have spam and ham mb's set up on the Linux box as
imap
folders, and those appear as shared folders in OE. When I want to train a
message as ham or spam, I right-click on the message (or group selection)
and select "copy to folder", then select the ham or spam mailbox. For
spam
you could just as easily do "move to folder" and it wouold disappear from
the user's machine.
There is a cron job on the linux box that scrapes the folders once a day
and
does training.
With slightly more effort you could have users forward the spams as an
attachment to ham and spam mail accounts. They would have to start a new
message, set the To address to "spam" or some such, and drag/drop the spam
into the body of the new mail, then send it. They can then delete the
spam,
or read the ham. The "more effort" is because you will now have to unwrap
the extra layer of inclusion on the received mails. Its also possible
that
this could lose some of the things like extraneous base64 encoding that
tend
to be good spam signs.
Loren
