Hi folks,
The Open Bioinformatics Foudation's recent subscription to the RBL+ list (http://mail-abuse.org/rbl+/) has done a great job at seriously cutting down the amount of spam that leaks onto our mailing lists. It does not however, protect us from virus-laden email messages as the members of biojava-l have found out on multiple ocasions. We are eventually going to deploy antivirus scanning on all of our inbound and outbound email but until that happens we need an interim solution. Typically the way that most large mailing lists handle this is to employ a "no attachments" policy. All attachments are either stripped at the MTA level or converted into plaintext by external helper applications. The side effect of this is that it also removes HTML-email which 90% of the time is spam anyway. The feedback from people I asked about doing this on our server was that it could be "too drastic". Instead of stripping anything MIME-encoded I've made some experimental changes to 2 of our largest lists: biojava-l and bioperl-l. What I've done is configured the lists to "hold" messages that contain suspect "content-type:" fields. What this mean is that messages won't be "stripped" but they will be blocked and held for moderator attention. Anything that is spam or suspicious will get blown away by an OBF mailteam volunteer. Messages that look OK will get passed on to the list. This is the best compromise I can come up with between "stripping everything" and converting the lists to 100% moderated forums. One side effect is that our "hold" patterns are going to block HTML messages wich is probably a good thing. Another side effect is that innocent messages may get held up or delayed as they wait for moderation. This is mostly unavoidable. For those that care, here are the patterns we are trying to use to hold suspect message: Content-Type: .*multipart Content-Type: .*mixed Content-Type: .*rich As a general rule to avoid having your emails held for approval people may wish to keep the following in mind: (1) Be polite to text-only email readers; don't send HTML messages to the list. (2) Don't send file attachments; post URLs or links within your message Feedback directly to me or to [EMAIL PROTECTED] is welcome. I'll let people know how this experiment goes - most likely in our next newsletter scheduled for mid-March. Regards, Chris -- Chris Dagdigian, <[EMAIL PROTECTED]> Life Science IT & Research Computing Freelancer Office: 617-666-6454, Mobile: 617-877-5498, Fax: 425-699-0193 PGP KeyID: 83D4310E Yahoo IM: craffi _______________________________________________ Biojava-l mailing list - [EMAIL PROTECTED] http://biojava.org/mailman/listinfo/biojava-l
