At 00/03/20 15:57 -0800, ned(_dot_)freed(_at_)INNOSOFT(_dot_)COM wrote:
Let me try and state what I think the problem that have been brought up are:
(1) We have to deal with what happens should another one of these suffixes
be defined at some point. Now, I have pointed out that we can define
things in such a way that this is defered until #2 comes along. But
if we want to tackle it now we certainly can. The simplest option is
to insist on these tags being in alphabetical order in any media type
name.
An alternative would be to state that these suffixes are to be
taken inside-out, or outside-in. Assuming outside-in, and
assuming 'charset', generic encodings, and RDF as strawman
examples, the sequence would be something like
text/example-rdf-xml-shift_jis-gzip
i.e. This stuff is RDF experssed in XML encoded in Shift_JIS
and compressed with gzip, or outside-in: To get at this stuff,
first decompress it with g(un)zip, then interpret it as
a sequence of characters encoded in Shift_JIS, then hand it
to an XML parser, and then to an RDF engine.
Regards, Martin.