I agree with Nathaniel that the inline/attachment distinction should be made
at the body parts, but I think that a separate header line like
Content-Disposition is much better than a parameter on Content-type.
My reason is that every type/subtype would need an optional "disposition"
parameter that speaks of the enclosing scope (usually, an enclosing multipart)
rather than the body. This would make it unlike any other parameter in the
MIME spec, both in its ubiquity and context-sensitivity.
Plus, a separate header line would have less impact on the MIME spec. I
haven't seen any mechanism for "registering new parameters for
already-registered
type/subtypes"; I assume it would have to go in RFC 1341. On the other hand,
new Content-* header lines are explicitly allowed, and could naturally be
defined
by a new RFC.
Per Nathaniel's message I have posted this to ietf-822 and not info-mime. Maybe
a summary of our conclusions should go to info-mime.
Jay