Null Sender Question

[Kai Schaetzl]

Scott B. Anderson wrote on Mon, 6 Oct 2008 15:39:45 -0500:

> Yes, but many, MANY of us Violate the old RFC standard, and the new
> ones recognize the need to reject email that otherwise passes the
> old RFC standard as valid email.

You misunderstand 5321 completely, if you read it that way.

> I would suspect that in turn, a large number of MAIL FROM:<>
> would also fall under this section of RFC 5321.

No, not at all. See, there is a difference between outright rejecting mail 
with <> senders (because of a <> sender) and rejecting mail that is 
unwanted and that bears "accidentally" a <> sender. It was *never* meant 
to require you to accept each and every mail from a <> sender even if your 
system thinks it's spam or harmful.

> 
> This list has debated this point before and I know at least one of
> us simply rejects email with a null sender via sendmail.

Well, that's his problem. If he doesn't have many customers, that might be 
ok for him. My customers would ask me why they don't get any bounce 
notices anymore when they send to wrongly typed addresses etc. Accepting 
<> senders doesn't add anything to spam influx if you have a good anti-
spamn system in place.

In the past
> we agreed it was a direct violation of RFC1123, now it appears there
> is some discretion under applicable RFCs.

Not at all. 5322 clearly requires the use of <> senders for DSNs and 1123 
still applies.

Kai

-- 
Kai Schätzl, Berlin, Germany
Get your web at Conactive Internet Services: http://www.conactive.com