a few open questions

Hi,

I've got a few questions where I feel that some clarifications could
contribute to the discussion.


1) RDF vs. Namespaces

I understand that RDF likes to see absolutized URIs as namespace names so
that it can reliably make statements about namespaces. So what's the impact
of having a relative URI as namespace name? If it's not absolutized, it
might not be unique, and statements made about the namespace might not be
correct. If it is absolutized, it's namespace name will change while the
document moves from location to location, so I again you can't use RDF to
make statements about it.

So from RDF's point of view a namespace name does only make sense if it is
*authored* as absolute URI, correct?


2) xmlns vs. xsi:schemaLocations vs. other mechanisms

It seems that a TBL & followers would like to use (re-use, abuse, overload
whatever...) the namespace name to actually point to a resource which would
contain additional information about the namespace (be it a human readable
document, a schema, RDF...). a) Why would the specifics of this be relevant
to a namespace aware XML processor? b) It has been suggested several times
to use an additional attribute to create this linkage to the intended
semantical information. What's wrong with this approach?


3) Namespace REC vs XPath/XSLT

We have heard that namespace REC and the XPath/XSLT RECs are inconsistent in
the treatment of relative URIs. However, actual implementations of XSLT seem
to stick to the literal comparison. a) It would be interesting to see, *why*
they actually do this and whether the authors plan to change this behaviour.
b) Does that mean that we do NOT have two interoperable implementations of
XPath and XSLT that fulfill the REC as written? In this case XPath and XSLT
might have to be moved back to the status of "candidate REC", if this is
technically possible.

Regards, Julian

Received on Wednesday, 31 May 2000 08:11:38 UTC