1. How Spam Filters Work

The recent surge in Spam has prompted many of you to contact us about
our Spam filters; I'd like to take the opportunity to tell you how our
anti-Spam filters work, and about the recent outbreak.

All anti-Spam filters use a method called "heuristics" to compile a
subjective score for how "Spammy" a message is. When a human reads text,
they can quickly and easily reach a conclusion about it, but it is
difficult to quantify how the decision is reached. Our Anti-Spam vendor,
Sophos, constantly updates a list of rules to check e-mail against.
These check how the message was addressed, what text appears in it,
what's attached, etc., etc.

For example, we just can't assume that an a message with "Sex" in the
subject or body is Spam, but we can say that it is a little more
*likely* to be so - so the Sophos engine gives it a score, and if enough
rules are "hits" it tips over into the "gray" category (and gets a new
subject line starting with 'PMX[#') or the "black" category, where it is
held in quarantine (see the end of the message for information about
accessing your Quarantine).

Spammers watch anti-Spam software closely and adjust their messages
accordingly. They use a variety of means to acquire your personal
address. They alter their message text or hide their message in a
graphic. Also, they lower their "Spam score" by including the language
Congress specified for legitimate advertisements on the web in the
CAN-SPAM act at the end of their messages. In this they make specific
claims that you requested to be on their mailing list and so forth. They
are breaking the law (again), but enforcement of CAN-SPAM on a (for
example) fly by night operating from Bulgaria is basically impossible,

The past three weeks we were subjected to a torrent of Spam from one of
these companies, using techniques to slip past our Spam filters. When
this happens, we report the offender to Sophos, and to Spamcop, a
service we use for "Remote Block lists." In addition, when we find a
severe violator like the recent aggressor ("Dawn 2 Dusk Associates") we
blacklist them on our won servers. We have blocked tens of thousands of
e-mails from that one company since we identified the source of the
messages.


2. Accessing your Quarantined E-Mails
We provide an easy web interface to your quarantined message and you may
choose to receive a daily summary of those messages in you e-mail as well.

Sophos PureMessage E-Mail Quarantine
https://mail.eckerd.edu:28080/

* view messages that have been held in quarantine by our anti-spam system
* choose whether to receive that daily "Quarantine Digest"
* choose which senders should never be quarantined
* choose which senders should always be quarantined


3. What else can you do?
Don't give your e-mail address to companies that will sell it, don't
post it to public web spaces, and don't publish it on your web pages.

If you use Netscape or Thunderbird as your mail program, open up Tools
--> Junk Mail Controls and tell it to move Junk messages to your "Junk
folder." This combined with Sophos PureMessage will screen you from
nearly all Spam messages.

If a certain sender keeps Spamming you, you can "Blacklist" them in
Webmail from the links above the message. Typically, though, Spammers
use an endless array of random addresses and subject lines to get around
these blocks.

If you have a serious problem with a Spammer, you can report the problem
to <postmaster@eckerd.edu> -- what we need is the full headers of the
message, and you can pull these up by clicking "Message Source" above
the message in Webmail.