[I think there's enough bickering going on about standardising mime-types
to make it worth a Last Call. If the IESG had gone ahead without one,
and decided to either accept or reject on their one there would have
been enough people annoyed be the decision to start another round of
mass executions, and it's a lot harder to get bullets in Stockholm]
Personally, I don't have any problems with having a plethora of top-level
types for cases such as this. Apart from the meta-types 'multipart' and
'application', I always thought that the first part of the type name was
meant to distinguish between major 'media' types - text, image, audio,
etc- in general, it's pretty easy to negotiate and convert between
subtypes of a top level type, but much harder to convert at the higher
level.
This distinction is useful, and necessarily implies a top level type for
each major semantic group capable of multiple concrete representations.
Forcing all possible structured types under a single top level simply to
try and impose a pattern of order over an essentially chaotic world is
the Geneva way.
LET A THOUSAND MIME TYPES BLOOM!
Simon
-----
April 19th == 4/19. 4+19 = 23. Any Questions?