ietf-822
[Top] [All Lists]

Re: on the end-to-endness of SMTP

1991-06-10 22:31:21
One of the really beautiful things about RFC822 and RFC821 is that
aside from a gross, evil hack (route-addrs), there is no mention of
routing. Routing in the IP universe is handled by IP and routing
protocols, two layers or more below SMTP. This is clean, elegant
design - it keeps the routing mess out of the hair of the mailer
writer (and the telnet writer, and the FTP writer, and so on).

RFC1123 is the first step toward getting rid of the gross ugly hack -
by deprecating route-addrs. If you read the description carefully you
will note that "support" means that an SMTP MTA must *parse* this
monstrosity, but is in no way obligated to follow it. MTAs are
encouraged to simply try the last hostname in the chain, and bounce the
letter back if that fails. Kudos to the Host Requirements WG.

Percent hack is another matter - that's in the local-part, which you're
not allowed to peek at, unless that domain name on the right hand side
is you. For all you know, that percent sign might be a part of
someone's mailbox. How the local-part is parsed (within the rules of
RFC822) is none of your business. RFC1123 mentions it simply to
document a very commonly used convention.

        Erik E. Fair    apple!fair      fair(_at_)apple(_dot_)com

<Prev in Thread] Current Thread [Next in Thread>