For the most part, the spam filter in Entourage has correctly identified the
unsolicited commercial e-mail I've received since upgrading from Outlook
Express. That's been a nice feature. I've always wondered what criteria the
filter used, though, and am even more curious now. Here's why:
Most messages with subject lines written in all caps are tagged as spam,
which seems an easy, if crude, guideline. Likewise, long and complex
originating domains seem more likely to be tagged. But I've seen two kinds
of exceptions to both these criteria that cast doubt on the rationale behind
the filtering.
One is that Entourage tags every message I receive from Northwest Airlines
as spam, even though I've told the filter to remember the NWA domain. This
may be the consequence of a Microsoft programmer disliking the airline, but
somehow I doubt it. This one remains a mystery.
Another is that I received eight variations on the same spam message this
morning, addressed to both my main accounts. (The user name is identical,
but the domains are different.) Entourage didn't catch a single one of them.
All the messages were formatted the way most spam is, with complex HTML and
a convoluted (faked) originating domain. The only thing that was different
about this crop of messages was that the subject line was capitalized
normally, in sentence case as a native speaker might have done it.
The recipient line showed that mine was just one of many guessed-at
addresses. That tells me I'm not on as many targeted lists of confirmed
addresses as I am in the way of the address bots -- but I still get the
spam. These were probably trawler messages, sent out to net live addresses
for more aggressive spamming later on.
Why has the spam filter stopped filtering? How can I retrain it?
-- Clayton Bennett
[EMAIL PROTECTED]
--
To unsubscribe: <mailto:[EMAIL PROTECTED]>
To search the archives:
<http://www.mail-archive.com/entourage-talk%40lists.boingo.com/>
