Re: A multiple bodypart multiple content type message

[Steve Kille <S.Kille@xxxxxxxxxxxx>]

 >From:  "Ned Freed, Postmaster" <NED@edu.claremont.hmcvax>
 >To:    ttl@fi.sti
 >Subject: Re: A multiple bodypart multiple content type message
 >Date:  Fri, 15 Feb 1991 18:22 PST

 >I agree; the combination of RFC934 and RFC1049 seems to provide a
 >clear, concise way of handling multipart messages that does not suffer
 >from line count fragility or the ugliness of more complex encapsulation
 >schemes. The RFC934 part of this stuff is already called for by the
 >X.400 to RFC822 gateway RFCs 987 and 1148.

 > <deleted>

 >This is an excellent proposal and I for one support it fully. It may be
 >necessary to > cull some additional information from headers specified
 >in additional existing RFCs, plus perhaps some yet-to-be-written RFCs to
 >specify additional information not previously considered.

 >				Ned

A few comments.   First, as anyone who as tried to code it will find,
RFC 934 (with or without 1049) does not give you enough for the 1148
mappings.  It specifies a format for forwarding a list of messages (defined
recursively) with a leading comment.  It does not properly deal with a
message with a sequence of body parts.   In our implementation, we have
been able to twist it to work using separators of the form:
  --------------- Start of Forwarded Message #6

  --------------- End of Body Part #2

and so on.  This really needs clarifying.   


The reason that I cc'd this group on my (plan for) proposed extensions to
1148 was that I am definintely not going to let that document stray into the
area of extending RFC 822 for support of structured messages.  I think that
this extension should be done, and I'd like to see it done in a manner which
can be easily referernced from 1148.  Clearly this group is the right place
for such work.

In addition to the problem mentioned above, it would be useful for such a
standard to tackle:
  - Body Parts Identified by object identifiers
  - The X.400 separation of body parts into separately identifiable 
  	content and parameter parts

I believe that it is important to define a means of specifying body parts,
which can be used both in X.400(88) and in this extended 822.   This will
allow new MMM UAs to easily use either the exteneded 822 or X.400 and to
minimise gatewaying problems.


Steve