A couple of the issues you raise are good ones, but not really grist for
the ietf-822 mill. In particular, the identification of incoming FAXes (I'm
talking here about "real" FAX, as opposed to "network" FAX) is an enormous
problem, and one which the NETFAX working group is looking at. You might
want to encourage them (hint) to look at this problem a little harder if
you really need a solution of some kind.
The issue of file protections as a hindrance to message/external-body is
also an important one, but somewhat outside of this group's focus. As
external references of this sort and others become commonplace on the
network (and MIME is by no means the first or only effort that is using
such mechanisms) we will need to see corresponding mechansisms appear in the
protocols that are used to resolve external references (e.g. FTP, NFS, AFS,
etc.). It was my understanding that AFS is actually rather well-off in this
regard, but then again you may not be using it for all I know. Specifically,
it would be a Good Thing if a UA had a way of assigning a protection to
one or more files that allow the external reference to be resolved without
compromising anything else. A finite lifetime on such things would also be
a good idea, since otherwise links of this kind will clog up everything
(as you said).
Anyway, thanks for the food for thought, and I look forwarding to any
other comments you might make on MIME.