On Mon, 8 Mar 1993 10:24:34 -0500 (EST), Nathaniel Borenstein
<nsb(_at_)thumper(_dot_)bellcore(_dot_)com> said:
Nathaniel> Does anyone object strongly to model #2? If not,
Nathaniel> then all we need to do is choose a syntax (I'd vote
Nathaniel> for a content-type parameter) and a default (I'd
Nathaniel> argue that the default disposition be "inline",
Nathaniel> because I believe that model was implicit in RFC
Nathaniel> 1341. If you don't believe that, look at the
Nathaniel> multipart examples in RFC 1341, many of which
Nathaniel> simply don't make as much sense with an
Nathaniel> "attachment" semantics.
This is exactly in line with what I have written so far. I add an
additional disposition type of "hidden" which is for use with compound
document types; this keeps information from being displayed twice
(once by the MUA during MIME parsing, once by the reader for the
compound document type.)
I lean toward the separate header, since it keeps disposition info
separate from type info, which to me is an important conceptual
separation. It would be nice to put the filename as a parameter to the
content-disposition as well; it belongs there, as being
non-type-related, and might address the clash with
message/external-body NAME semantics which Crispin pointed out.
This will probably not be feasible, though, given that 1341 is so far
along.
-Rens