Re: heading toward datatyping telecon

At 11:21 AM 11/2/01 +0000, Brian McBride wrote:

>Graham Klyne wrote:
>>I really don't see an issue here.  Allowing literals as subjects in the 
>>abstract graph, and in N-triples, doesn't change the RDF/XML syntax as 
>>defined.  Furthermore, if we use Pat's MT work in conjunction with RDFS 
>>closures, we can explain how to apply XML data types with the existing 
>>RDF/XML syntax (even if we can't express the graphs that result from RDFS 
>>closure in the existing RDF/XML syntax).
>Oops, am I missing something.  Graham, how should I write:
>   "subject" <rdf:type> <rdf:Literal> .
>in RDF/XML?

You don't.  I'm not sure why you would *need* to do so.

But if one says:

     my:shoe shoe:size "10" .


     shoe:size rdfs:range xsd:integer .

(both of which *are* expressable in RDF/XML)

then (as I understand Pat's proposal), using RDFS closures, one can infer 
the graph-statement that corresponds to:

     "10" rdf:type xsd:integer .

(but relating this to the particular node above that is labeled with "10".)

i.e. the resulting *graph* closure may have literal-nodes in the subject 
position, even though the RDF/XML syntax cannot express them directly.


Graham Klyne                    MIMEsweeper Group
Strategic Research              <>

Received on Friday, 2 November 2001 09:34:35 UTC