[...re-sending...]
-----Original Message-----
From: Mark Baker [mailto:distobj(_at_)acm(_dot_)org]
[...]
I agree that media types aren't suitable for multi-namespaced
documents.
That's why I'm all for making the transition from media types to
namespaces with either vanilla "application/xml", or adorned with the
root namespace.
Does a namespace alone indicate the semantic meaning of an XML document?
Is it 'enough' of a hint to do useful things?
Or do you need the name of the root element as well?
Would schema+rootElement be a better alternative?
Would that sufficiently identify the 'model' of the message content?
Could XML documents without schema use namespace in place of schema
reference?
Is there a more general idea of 'content-model' that might apply to multiple
content-type representations?
Example:
Content-Type: application/xml
Content-Model: http://www.w3.org/2001/12/soap-envelope#Envelope
and
Content-Type: application/x-java-serialized-object
Content-Model: java:org.apache.sample.user.profile
Mike