two comments on XML Schema, RDF, and OWL datatypes (ISSUE-29)

I was doing some more looking into the datatype matter.


I found that RDF datatype are *not* compatible with XML Schema
datatypes, contrary to the claim in RDF Concepts at
http://www.w3.org/TR/rdf-concepts/#section-Datatypes

	5. Datatypes (Normative)

	The datatype abstraction used in RDF is compatible with the abstraction
	used in XML Schema Part 2: Datatypes [XML-SCHEMA2].

because there is no requirement in RDF that each value in a datatype
have a lexical form (from
http://www.w3.org/TR/rdf-concepts/#section-Datatypes)

	Each member of the value space may be paired with any number
	(including zero) of members of the lexical space (lexical
	representations for that value).

but there is in XML Schema at
http://www.w3.org/TR/xmlschema-2/#value-space

	2.2 Value space

	[Definition:] A value space is the set of values for a given
	datatype. Each value in the value space of a datatype is denoted
	by one or more literals in its lexical space.

I'm not sure what, if anything, should be done to fix this problem.


I also believe that Jeremy's analysis in
http://lists.w3.org/Archives/Public/public-owl-wg/2007Nov/0505.html 
is incorrect.  Jeremy says

	My analysis is that the value space of any rdfs:Datatype is an
	owl:DataRange, because it satisfies CEXTI(e) subset LVI, but not
	conversely, because of the lack of a lexical-to-value mapping.

The only semantic requirement on owl:DataRange is that its instances
have all of their instances in LV.  This does *not* imply that instances
of owl:DataRange are instances of rdfs:Datatype (or vice versa).

It is possible to construct (trivially) an RDF-valid lexical-to-value
mapping for every OWL 1.0 datarange, namely the empty mapping (it is
also possible to do much better), so it might be possible to eliminate
owl:Datarange in favour of rdfs:Datatype, except that OWL 1.0 dataranges
don't need to have a URI, but RDF datatypes do need to have a URI.


peter

Received on Wednesday, 2 January 2008 15:05:24 UTC