On Thu, 10 Jun 2004, Markus Hofmann wrote:
The OCP/SMTP profile item seems pretty clear. Any specific things
that need to be considered when phrasing a charter item on that?
I am not an SMTP/IMAP/POP/etc expert, unfortunately. Perhaps due to my
lack of in-depth understanding of all these "e-mail" protocols, I
wonder whether a more general "mail" or even MIME profile should be
considered for OCP _instead of_ SMTP profile.
In OCP/HTTP we tried very hard to keep most HTTP-level protocol
interactions away from OCP. For example, OCP/HTTP does not know about
[persistent] HTTP connections. OCP/HTTP works on individual HTTP
messages, leaving the actual HTTP "state" out of scope.
I suspect we can do a similar trick with SMTP. Then, it may be
possible to use the same profile for other protocols that deliver
e-mail or even any MIME messages. Just like OCP/HTTP focuses on an
HTTP message, OCP/MAIL would focus on an e-mail (MIME?) message and
will be applicable to SMTP/IMAP/POP/etc. protocols. We may have to
treat tracing and bypass specially, but that would still be easy.
Are SMTP/IMAP/POP/etc. protocols similar enough from mail adaptation
point of view? Does even more general MIME adaptation make sense?
Note that I am not looking for more work for the WG. I am just trying
do understand whether we can cover more protocols with essentially the
same amount of work if we extend scope from SMTP to other mail
protocols or even to MIME in general.
P.S. HTTP messages are almost MIME messages, but there are a few
important differences that make it difficult to handle an HTTP
message as just a MIME message. Is that true for SMTP and other
e-mail protocols? Do they "modify" MIME, each in its own