RE: RDFCore WG: Datatyping documents

> John F Schlesinger wrote:
> 
> > Jonathan said: "Please do not bind the "xsd" prefix to the

> > Please let us not forget that the prefix 'xsd' is purely 
> conventional and
> > has no semantics except when related to an xmlns attribute. 


Absolutely. However:


>>It is
> perfectly
> > correct for RDF data types to associate xsd with
> > http://www.w3.org/2001/XMLSchema#


Nope, it is definitely not correct. See [1], the URI you mention is to
be used by XML Schema itself to refer to it's build in datatypes. 
The XML Schema spec explicitly gives a URI for an XML application other
than itself to use it's build in datatypes, that URI is

http://www.w3.org/2001/XMLSchema-datatypes

So, to use the Int type in a language other than XSD one must use
something like the following:

http://www.w3.org/2001/XMLSchema-datatypes#int

I believe the prefix is an irrelevant subject, as long as known prefixes
are not used; using a widely used prefix (although perfectly legal) may
cause confusion.
One thing that worries me is the notation of facets; I don't know how
common this is but I guess it should be accepted by a WG since it's
inside a recommendation.

Finally, there is the subject of Type Libraries [2], that are far more
interesting.

[1] http://www.w3.org/TR/xmlschema-2/#namespaces
[2] http://www.w3.org/TR/xmlschema-0/#Libs


Manos

Received on Tuesday, 5 February 2002 04:10:29 UTC