What's in a name? - "spam" / "not spam"

[Michael Janssen]

On Wed, 30 Jun 2004, Mike Brudenell wrote:

> For spam this heading looks something like this:
>
>     X-Blah-MailScanner-SpamCheck: not spam, SpamAssassin (score=-4.8,
>             required 8, BAYES_00 -4.90, BIZ_TLD 0.10)
>
> and for non-spam something like this:
>
>     X-Blah-MailScanner-SpamCheck: spam, SpamAssassin (score=22.781,
>             required 8, autolearn=spam, BAYES_99 5.40,
>             DATE_IN_PAST_12_24 0.75, ... )
>
> The problem arises when trying to use client-side filtering with an IMAP
> mail program.  Such can be set to query the IMAP server to check the text
> of a particular message header.  However the IMAP specification stipulates
> that this match is as a case-insensitive substring.
>
> Thus setting up a search to check the "X-Blah-MailScanner-SpamCheck" header
> for the word "spam" matches both spam and not-spam messages.  (The converse
> -- checking for non-spam messages by looking for "not spam" is fine.)

X-Blah-MailScanner-SpamScore solves this problem. Config Setting
"Spam Score = yes" turns them on. Config items "Spam Score Header" and
"Spam Score Character" manage the details. The Header looks like:

X-Blah-MailScanner-SpamScore: ssssssssssss

Such a score of 12 is found eg by a filter with "ssssssss" (8).  With the
SpamScore header, user can choose their treshold themself, which is a
advantage in some circumstanzes/ for some user.  (Minor) disadvantages are
that some might try to "improve" the filter with to low tresholds and some
might misunderstand the concept and increase the treshold to stop more
spam where they should have reduce the treshold.  Anyway the SpamScore
Header is nice for filtering.


Michael

-------------------------- MailScanner list ----------------------
To leave, send    leave mailscanner    to jiscmail at jiscmail.ac.uk
Before posting, please see the Most Asked Questions at
http://www.mailscanner.biz/maq/     and the archives at
http://www.jiscmail.ac.uk/lists/mailscanner.html