LSTSRV-L Archives

LISTSERV Site Administrators' Forum

LSTSRV-L

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Keith Moore <[log in to unmask]>
Fri, 9 May 1997 11:53:13 -0400
text/plain (43 lines)
        It is a well-established principle that an SMTP server
        may refuse to accept mail for any operational or technical
        reason that makes sense to the site providing the server.

Sure, you can do this.

But when an AOL.COM SMTP server which is listed in the DNS as a mail
exchanger for one of AOL's domains, refuses to accept a
reasonable-sized message with a perfectly valid return address that is
destined for a valid recipient at that domain... then AOL is not
providing the level of mail service that its customers have every
right to expect.

Source-routed addresses are not forbidden.  They're archaic, perhaps
obsolete.  But they're still valid, and some sites still generate
them.  And in my experience, source-routed addresses aren't very good
indicators of spam.  A lot of spam uses such addresses, but so does a
lot of legitimate mail.

In my experience, invalid return addresses (either invalid syntax or a
nonexistent domain, in either the From header or envelope) are a FAR
better indication of spam.  My lists forward any message lacking a
valid return address to the list maintainer.  This catches a great
deal of spam, and very few legitimate messages.

IMHO, you have far better grounds for rejecting a message on the
grounds that its return address is invalid -- especially if the
envelope return address is invalid -- than for rejecting valid
source-routed addresses.

>     Our general approach is one whereby if we can't determine that
> we could bounce a message if we had to (i.e., a message comes in
> for a nonexistant AOL user, or for one whose mailbox is full), then
> we won't accept that message.

That's a reasonable criterion.  But the use of a source route doesn't
mean that you can't bounce it.  Especially given that 1123 clearly
allows you to strip the route portion.  If stripping the route from
the return-path yields an invalid address, *then* it's reasonable to
refuse to accept the message.

Keith

ATOM RSS1 RSS2