But MIME is a message format. It is not a mail transport. It does not
make any transport related guarantees because that is outside of its
scope.
I agree that we need to keep clear the distinction between message layer
and transport layer, but...
We have an SMTP extension called "8bit-MIMEtransport".
At the risk of stating the obvious, the two documents must agree about
what 8bit-MIME is, so that:
i) any MIME message validly labelled with "C-T-E: 8bit" can be safely
transported by 8bit-MIMEtransport; and
ii) it is no harder to implement 8bit-MIMEtransport correctly than is
necessary (else why not simply mandate full binary transparency?).
I think it would be useful if the advocates of changing the specs could
propose specific text they would like to see to restrict RFC-1521, and
relax RFC-1652.
It would also be useful to hear from any implementors of RFC-1652: do
they, in fact, "preserve all bits in each octet passed using the DATA
command"?
Tim.