Personally I found this a pity because I had implemented it in JUMBO1 and
it worked fine. I had things like:
<?namespace ns="something-i-forget" prefix="CML" src="jumbo.cml.schame.xml"?>
The schema was very similar to the current XSchema, extended for each
element with things like:
- help
- java class mapping
- additional semantic constraints
worked nicely - had to tear it all up (yes, I know the warnings are in the
drafts - :-)
The current namespace spec *deliberately* provides no semantic resolution.
I argued against this because it seemed a recipe for chaos. So far I
haven't been proved right or wrong - we are still in the inaction phase.
The current jumbo has a placeholder:
<?jumbo:namespace ns="http://www.xml-cml.org" java="jumbo.cml.*Node"?>
and similarly for other namespaces.
Not pretty, but it works.
There have been warnings about the complexity of registries - I assume that
the xml-ea1 props file is a registry? Personally I don't yet care. I'd much
rather see us try to make this work than simple wait around. I'm game for
whatever URI<->functionality mapping we choose (as long as it's simple) so
that we can at least get some experience.
David's done it, I've done it - I assume there are others. It's part of
XObjects and needs addressing soonest :-)
P.
Peter Murray-Rust, Director Virtual School of Molecular Sciences, domestic
net connection
VSMS http://www.nottingham.ac.uk/vsms, Virtual Hyperglossary
http://www.venus.co.uk/vhg