classiccmp list (sort of) help requested
brad at heeltoe.com
Wed Dec 13 19:47:59 CST 2006
"Jay West" wrote:
>Maybe spamassassin will find a way to progress in the tit-for-tat war with
I'll tell you what you probably already know. Spam has been morphing
lately on a couple of fronts. People who don't run their own SMTP
servers have probably only noticed one of these fronts - the move to
pictures embedded in mime encoded email. No Bayes classifier can deal
with that (that I know of).
Another front is the rise of spambots. I've seen a huge increase in
"SMTP hangups", where a connection occurs and then is dropped. It's
spambots fishing. Also, tons of email sent to non-existent users at
valid domains. These cause double bouncing and clog up email servers.
I've seen hundreds of them running all over (what appear to be) cable
modem nets in the UK. And other places as well (no slight ment to those
across the pond - just one of many examples). There must a lot of
infected pc's running windows out there.
I've taken to rejecting email at the SMTP level by checking the
recipient - using a qmail variant called magic-smtpd. It's eliminated
the double-bounces caused by email sent from non-existent users to
non-existent users. But, the checking requires you to keep a list of
valid email recipients. This is probably not practical for many domains
but it might work for a list/domain where you know the list of valid
recipients up front. Just a thought.
The sad thing is that email is quickly becoming less effective than fax
and the phone. A high percentage of email is filtered or lost. very
More information about the cctech