nate
> I am curious if the smtpfront pattern matching “solves” the problem of the false positives ???
By that I assume you are referring to:
"Too many false positives from HTML formatted mail......I end up picking false positives out of the junk mail folder far too often"
Pattern matching will by default reject email messages that have attachments with executable content (you can select the basic types of executable content to block).
If you want to block other file type content then you generate the file magic as per the howto and add that to the pattern database (as per howto).
If you exclude all html content then you will probably be in big trouble with your users.
Note that you can copy the patterns database from smev7.0 (which has been updated) and use that in v6.x for more effective blocking.
See a forun post a few months ago about this
So the way to manage all this is to use layers.
1) RBL rejection (configured in jespers Spam Filter or manually as per my howto)
2) Pattern matching rejection (as per howto or default in sme 6.5 & 7)
3) Spamassassin with a senstitive score setting like 4 (use jespers Spam Filter panel)
4) Antivirus filtering - clamav (use jespers antivirus panel)
5) other tweaks to reject spam type messages
eg Reject messages with bad HELO/EHLO entries
Control doublebounce messages
Control mail to invalid addresses & configure mail blocking rules
all as per this howto
http://mirror.contribs.org/smeserver/contribs//rmitchell/smeserver/howto/Mail%20system%20tweaks%20HOWTO%20for%20sme%20server.htmI get virtually no junk mail into my Inbox as it all gets blocked or filtered to the junkmail folder.
I don't recall having a false positive in the junkmail folder for many months now as the current version of spammassassin is quite good.
You can use whitelist or blacklists in the Spam filter panel to help reduce any false positives or negatives particular to your system & your users.
Make sure you keep spamassassin Spam Filter & the antivirus panel updated whenever jesper issues a new release or update.