2. An overview of the Oxmail features
As described above, the Oxmail service scans for both junk mail content and virus payloads in email messages. In addition to this, the current identification of mail from suspicious sources on the Internet (known as RBL Warnings) is still in place.
When the ClamAV software reports an attached virus we can be certain of its existence because the techniques used to identify viruses are precise and virtually foolproof. It is with this reasoning, and also because viruses pose the most destructive threat to the University's IT resources, that all email messages containing viruses are rejected. A message is returned to the sender of the email telling them that it has been destroyed and did not reach its intended recipient(s).
By contrast, the process of identifying junk mail content is quite imprecise, as the software must interpret the meaning of sentences within an email message. Because of this no other emails will be rejected, and instead the Oxmails assign a score to each of your email messages. With a higher numerical score a message is more likely to be junk mail. Conversely the lower the score (it may even be negative), the less likely a message is to be junk mail.
OUCS can add information to emails such as a junk mail score or RBL warnings because there is an additional portion of an email message that you do not normally see in your email client. This section contains what are known as Headers, which are messages or instructions to your email client software. Here is an example of the Headers we might add to a suspicious message:
X-Oxmail-Spam-Status: score=12.0 tests=NIGERIAN_BODY, OX_PTR_MISSING, OX_RBL_SPAMCOP, RISK_FREE, SUBJ_ALL_CAPS X-Oxmail-Spam-Level: ************
Don't worry if you can't understand these messages, your email
client software will do this for you. You can see in the second
of the two messages that there are twelve stars, meaning that
this email was given a junk mail score of
The score is a number between
+100. It's not likely that you will ever see a
score more than around forty, and most legitimate mail should
receive a score of less than five.
We don't bother generating a spam score for
large messages because
they would consume a lot of system resources and are unlikely to be