Re: New type suggested: Multipart/Related

Well, I for one did not think the Content-disposition approach cleaner.
(obviously....)

On the DAG issue, I seem to be somewhat clueless:
How is it possible to have a containment hierarchy that is not a DAG?
In hypertext, you obviously want to be able to go in circles, but I did
not think that all possible relationships between body parts were
covered by either Content-disposition or Mainpart=.

Looking forward to your draft (since I do not understand the
Content-disposition).

                      Harald A