Re: Comments on draft-ietf-mimesgml-related-04.txt

[Ed Levinson <[e-mail]>]

Mark,

>From the previous discussion on this list there didn't seem to be
a wide desire to support composite media types.  For me, any discrete
type can be a Multipart/Related "type" but it only makes sense if there
is something to be done for the entire multipart based on that type.
I don't know what a Mul/Rel; type="image/gif" would represent but
HTML and SGML are two arenas where Multipart/Related fits.
The MUA, of course, should not need to understand the "type", only
where to hand the MIME body parts.

The draft is neutral (really silent ;-) on the "type" issue.  What
application (and behaviour) would make use of the types you
suggest?

The paragraph you cite, which is the only change outside of dropping
the change log, was left out in error and was pointed to by the change
log.  It only applies to the start parameter.  In that context do you
see something more to be said?

Best.../Ed

On Thu, 16 Nov 1995 15:10:34 PST Dr. Mark K. Joseph wrote:
> On page 3 near the bottom the following paragraph appears:
> 
>    In the case of a Multipart/Alternative body part containing
>    several entities with identical content-IDs the start entity
>    should be selected using the Multipart/Alternative rules.
> 
> This is the only paragraph discussing the use of compound mime
> types for the type and start parameters.  It is way too terse and
> should be expanded.  For example can any mime type/subtype be used
> for the type parameter (e.g., message/external, image/gif) ?
> 
> An example of the use of type="Multipart/Alternative" at the end
> of the document, as well as, a discussion of which mime types/subtypes
> can and cannot be used for the type parameter will help interoperability
> and my understanding of the draft.