RE: [xml-dev] Meta-somethingorother (was the semantic web mega-permathread thing)

["Joshua Allen" <joshuaa@microsoft.com>]

> The question at hand is when one is publishing
> data should it be published in XML+Namespaces or
> in RDF.  You're allowed to use any reasonably

Hold on; if you publish in RDF, then you *are* publishing in
XML+Namespaces.  

I thought the question was, "when modeling data (presumably for
interchange), should you model it using XML data model or RDF data
model?"

Getting RDF *syntax* out of the picture, just think of modeling *any*
data in XML.

Is it ever a best practice to model your XML data in such a way that:
a) it represents a collection of real-world "items" with "properties"
b) all nodes which represent property names are clearly distinguished
(no implicit property names, conventions are clear, etc.) from actual
items or property values
c) optionally, all "items" have IDs which can be referenced in property
values

IOW, even if you do not use RDF, do you ever find that it's a best
practice to model your data in the way that RDF would?

I think the answer is "sometimes".  I suspect that Elliotte has built
real-world schemas that are practically equivalent data model to RDF.  I
wouldn't challenge him to defend *why* he did it, because sometimes it
just makes sense.  Sometimes it doesn't.  I don't see a controversy
here.