John,
JL> (For the record: I did suggest to Ted that Jabber logs shouldn't be
JL> considered the "public record". He disagreed. I'll take my chances on
my view is that the jabber session is a public working group meeting.
JL> It is _really_ unfortunate that RFC821 MAIL-From has that name.
I agree. The last few months of discussion have been unpleasantly
enlightening to me, in that regard.
JL> It
JL> has led to untold confusion over the years. It is more correctly
JL> called "Envelope-From",
In the postal world, it is called "return address". In fact, that is
its real function. The fact that the return address is often the from
address is an unfortunate correlation, when trying to discuss messaging
semantics.
JL> I'd like to suggest that there's absolutely nothing wrong with
JL> bounces that must use the RFC821 Bounces-To
There is no reference to Bounces-To in RFC821 or RFC2821. Which address
are you referring to?
JL> Every MTA along that path has had access to the full text; so there's
JL> no _new_ privacy issue in relaying the text back.
The question is not one of privacy, but of authority to set the value in
MailFrom.
JL> So, while, rewriting the Bounces-To may _sound_ like forgery when we
JL> call it "Mail From", it really is nothing of the kind.
That is why we need to be clear about the lines of authority for setting
values. As a practical matter, I view MailFrom as being under the
authority of the entity referenced in the RFC2822 Sender field.
d/
--
Dave Crocker <dcrocker-at-brandenburg-dot-com>
Brandenburg InternetWorking <www.brandenburg.com>
Sunnyvale, CA USA <tel:+1.408.246.8253>