Re: file attachments in MIME

In the new scheme, you're right -- there's no conflict with the use of
name in external-body, and the use of name for octet-stream simply
becomes archaic.  The conflict came about in trying to define a global
parameter that did what the octet-stream name parameter did, but for all
content-types.  Content-disposition could be a very neat solution for
this problem.  -- Nathaniel