At 11:25 PM 4/12/99 -0400, Gavin Thomas Nicol wrote:
I should note that both Simon St. Laurent and Chris Lilley have been making
noise about such "catalogs" recently as well.
And then later:
Rather than extending the MIME headers with lots of
parameter, perhaps just the URI of a single resource like that would allow
greater
extensibility.
Yes, I think this is absolutely the right way to go. The only thing I would
add is that packaging itself has many facets: the "catalog" (drlove, XPDL,
etc.)
definition, packaging in MIME multipart/*, resource name aliasing, push vs.
pull,
mixed mode push/pull, etc. that all need to be considered.
XPDL (http://purl.oclc.org/NET/xpdl) is exactly about this packaging (and I
really should talk with Rick about his drlove project to figure out if we
have much in common.) After a number of rounds on XML-Dev complaining
about the imprecision of the methods currently available for identifying
XML document types (including MIME), I figured it was time to go ahead and
write up a proposal for describing such types.
At this point, I think I'm leaning toward leaving the MIME type
application/xml for documents (and creating something more descriptive for
DTDs) and using a different mechanism to connect the descriptions to the
document. (The XPDL draft currently has five possibilities, and I'm
pondering a sixth.)
Rick's point that a single XML document may be a chameleon of MIME types,
capable of being presented as text/x-myType, application/xml, possibly
text/xml and text/plain, and potentially even more, seems very striking.
We've moved beyond "one document - one type" and dealing with that will be
difficult.
I think I'd rather let MIME identify a basic type - application/xml - and
develop another mechanism for detail beyond that. XML documents share some
key features as far as transfer over networks, which to me at least is what
MIME is best for helping with.
Simon St.Laurent
XML: A Primer
Sharing Bandwidth / Cookies
http://www.simonstl.com