I believe there is an error in the RDF concepts document.
The problem is in the definition of the XMLLiteral datatype ; the
value space is ambiguous: it is defined to be disjoint from the value
space of the XML schema datatypes, but does not say anything about other
datatypes. It is therefore unclear whether there is overlap between the
value space of XMLLiteral and, for example, the new datatype rdf:text.
This is a problem for any language relying on this datatype, in
particular, RIF and OWL 2.
Consider, for example, the statement
"<bla>"^^rdf:XMLLiteral = "123@en"^^rdf:text
it is possible to decide whether this statement should hold or not,
i.e., whether is it derived from the empty rule set/ontology.
I propose to add the following text to the RDF errata document .
The following text, in the second bullet of the subheading "The value
space" of Section 5.1 of the RDF Concepts document:
disjoint from the value space of any XML schema datatype [XML-SCHEMA2]
should be replaced by:
disjoint from the value space of any other present or future
datatype (e.g., XML schema [XML-SCHEMA2]), with the exception of future
sub- or supertypes of this datatype
The reason for the replacement is that the definition was ambiguous: the
relationship between the value space of rdf:XMLLiteral and the value
spaces of datatypes not defined in the 2001 version of XML Schema (e.g.,
rdf:text) is unclear. For example, one cannot infer from the
specifications whether there is overlap between the value spaces.