Excellent idea, even if it is already present.
Thanks, but now I'm curious. Could you explain in detail *why* you
think it's an excellent idea? Perhaps your explanation would reveal
some other problems that I haven't heard of (i.e. besides ATTMail and
LISTSERV). Perhaps you're referring to X.400?
Should we add a suggestion to MIME/2 that a single message
that has MIME-version: in its first lines of body, and no MIME header,
may actually be a stripped message/rfc822, and ask it to behave
accordingly? This is icky, but not a major change to MIME....
I like this idea. But I'm going to say something even ickier: The
reason why I suggested the use of a "unique" string (i.e. a string
that is damned unlikely to appear in normal messages) is so that
software can interpret that and do "wonderful" things.
For example, (here comes the icky stuff) in the case of a Unix
environment that uses a variant of the "sendmail" program, that
program could look for the unique string, and if it is found, it could
"merge" the headers back into the top-level headers. If sendmail
takes care of this, then the UAs don't even need to be altered; they
can just continue doing MIME the way they used to. (Yes, it's very
icky -- I *did* warn you.)
A more specific example might clarify what I'm talking about. When
sendmail receives the message, the message/rfc822 header is not there
any more, so it says:
From: suzie(_at_)harvard(_dot_)edu
To: joe(_at_)foo(_dot_)com
Subject: new proposal
Comments: dslfkjsdlkfjsdlkfgjsglkjsdlfkjsddslkjfdsf
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Mime-Version: 1.0
Hi Joe,
How's things? [...]
After "merging" the headers back into the top-level, it says:
From: suzie(_at_)harvard(_dot_)edu
To: joe(_at_)foo(_dot_)com
Subject: new proposal
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Mime-Version: 1.0
Hi Joe,
How's things? [...]
Is this icky, or what? Of course, sendmail wouldn't do this if it was
just acting as an MTA, and it wouldn't do this if the receiving user
hadn't told it to, just in case there are any problems with POP-like
systems. (Am I missing something?)
This spec could be added to MIME without resetting to Proposed, since
it's really an extension and not a change, in my opinion.
Erik