Subject: | |
From: | |
Reply To: | |
Date: | Mon, 5 Jun 1995 19:22:03 +0200 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
On Mon, 5 Jun 1995 00:36:49 EDT Ben Chi <[log in to unmask]> said:
>We've just installed 1.8b and it's failed its first anti-spam test. (And
>the Subject: line even said "ADVERTISEMENT"!-) Have we failed to turn
>some feature on?
It order to stop a spam, your server must either detect it on its own, or
receive an alert from another server *before* it receives the spam.
Unfortunately, with the current BITNET situation the latter is realistic
only for TCP/IP servers. It isn't unusual for LISTSERV to go offline at
core sites during peak hours, and in fact one of the largest core sites
recently reported that their LISTSERV was offline 5 1/2 hours on the
average weekday. With that kind of delay, there is no hope for alerts to
propagate quickly enough over BITNET to be effective. So, for practical
purposes, and for US sites, the spam has to be detected by the (1.8b)
servers in your region, since crossing the core can take hours. The
detection process always involves letting a few copies through. You can
only suspect a spam once you have seen a certain number of copies,
otherwise it's just an ordinary message. LISTSERV doesn't attempt to
filter messages based on the contents, because this would result in a lot
of false positives.
Eric
|
|
|