Marza has asked for the wisdom of the Perl Monks concerning the following question:

Greetings

For spam defense, we have TrendMicros InterScan. Since we purchased it(manager made the decession on his own), I can't change it and thus have become the spammunky!

As the filtering has gotten more intense, more and more legal mail is getting quartined. Rather then searching the endless files of spam, I would like to write a reviewer for the mail. Obviously, it can handle the From: To: and Subject:, However, I would like to see the actual message when the header info is not helpful(ie Subject: Hi)

Does anybody have a module to suggest? I have been looking at some of the mail modules on cpan: Mail-abuse, Mail-action, etc.

When it qurantines a mail, it creates two files.

A header file which is not always helpful:

F<JoePowell@mx48.Blkmagicti.biz> T<user@company.com> OWorld Operating<JoePowell@mx48.Blkmagicti.biz> SJoe Bob, City Residents / Compare Rates Online....Instantly! DJoe Bob<user@company.com>

And a mail file which will have the message with varying levels of stuff. Sometimes text, sometimes HTML, Sometimes both.

Thank you!

20040327 Edit by BazB: Changed title from 'Quartine Mail Review'

Replies are listed 'Best First'.
Re: Reviewing quarantined email
by tilly (Archbishop) on Mar 27, 2004 at 07:11 UTC
    I'd suggest sending the quarantined queue through Spamassassin to separate it into things that you want to look for and things that you don't. First of all that will manage to get most of the legal documents extracted from the mess. Secondly you can improve the review of the rest by adding rules that specifically locate legal-type emails to improve your accuracy.

    Of course after a little time with this exercise, you might wonder what value TrendMicros InterScan is adding. But don't ask those questions too loudly. Because your manager wouldn't look good if you did that...

Re: Reviewing quarantined email
by chanio (Priest) on Mar 27, 2004 at 18:22 UTC
    PopFile is a good Open Source written in Perl but installs at any windows alone. Many people is using it as a structure to add some other useful things. Since it is all based in a Bayesan filter, this system keeps on learning what we, humans, consider spam and what not. It also learns how to classify normal emails. As everything it needs some time to learn or someone to pass you his corpus (what it has learned).

    Popfile works the same in any other platform.

    There is an insteresting article about another alternative for LINUX: http://www.stonehenge.com/merlyn/PerlJournal/col07.html .