Hi.
Source routes were implicitly deprecated from Internet mail with
RFC 975 at the beginning of 1986 and explicitly deprecated with
RFC 1123 in 1989. The DRUMS work concluded that SMTP senders
should be prohibited from sending them out but that SMTP
receivers should be (i) required to accept the syntax and (ii)
either to follow the source routing rules for any routes they
accepted or to ignore the routing specifications and use the
mailbox address only.
These decisions were reinforced by knowledge of the nightmares
that explicit mail routing had caused in the UUCP environment,
where such routing persisted much longer than it did in the
Internet/ SMTP one.
To accomplish that goal, a slightly modified version of the
source routing rules from RFC 821 was incorporated into Appendix
C of 2821 and section F.2 of Appendix F was introduced to
deprecate the things.
It obviously was not sufficient, since we still have people
quoting text that specifies, e.g., that relays incorporate their
own domain names into reverse-paths, a feature that, as far as I
know, has not been actively used by anyone with SMTP for more
than 20 years (I suspect that there may still be pockets out
there, but they certainly aren't high-frequency).
To prevent further confusion, I propose that we do one of the
following:
(1) Modify Appendix C to clearly prohibit SMTP-senders from
inserting their own domain names into the return-path, thereby
construction source routes in violation of the existing
prohibition on clients sending out any such thing. This is not
a change in functionality, just changing some appendix text that
can be read as contradicting some text in the body of the
specification.
This option requires some small tidying up elsewhere, but the
Appendix C changes are the core of the work.
(2) Decide that source routes are ancient history and remove
Appendix C entirely, moving all discussion of source routes into
the present F.2, and shifting the rules for forward-pointing
source routes from "MAY ignore" (and use the Mailbox information
and MX records only) to "SHOULD ignore".
Reintroducing source routes and requiring support for them
would, of course, constitute adding functionality, so that
option, which would otherwise be the third case above, is a
non-starter.
The current working version of -03 incorporates a version of
option (1) above.
john