Solution Pros Cons
(a) application/ No change to Not neat to require
foobar-xml MIME specs. mailer to parse the
subtype string.
other con: potentially complicates content-negotiation schemes that
are based on the content-type name. and it's not just mailers that
are affected, since other things use the same content-type namespace.
I think most folks have settled on (a). I still prefer (b) but
my main goal in raising a stink was to get MIME people to think
about the implications of the decision before accepting it.
so I'll stop loudly objecting to (a).
I view the tradeoff between (a) and (b) as one of short-term
acceptance in the case of (a) vs. better long-term sanity in
the case of (b). But MIME may be obsolete for other reasons
before we really have much trouble resulting from these frobs.
caveats: (1) it should be up to the folks defining the type to
decide whether to use the -xml frob or not. (2) the interaction
between the treatment of this frob and the rules about default
handling of mime objects for user agents presenting those objects
need to be clarified. (3) we need to understand how multiple frobs
in the same content-type name work in case we ever need them. *
(4) since "-" is already used in some content-types it might be
worth considering a different separator.
Keith
* I suspect we will need them eventually, though it might be a few
more years. I don't think that XML is the last word in presentation
syntaxes, and some day there might even be generalized semantic
conventions for tagged objects.