Re: Yet more XSD/RDF

> I completed my investigation doc.
>
> See:
>
>
http://lists.w3.org/Archives/Public/www-archive/2002Nov/att-0092/02-index.html


that's an impressive work!

> More important points are:
>
> - there are issues with the XSD document: most is clear, a small part is
> clear in the opposite direction.
> - types derived by restriction correspond to rdfs:subClassOf
> - don't use xsd:QName, xsd:ENTITY, xsd:ENTITIES

right

> - even if we stick to the bits that are clear then implementation looks
> hard.
> Particularly any test that relies on finiteness of a datatype, or of an
> intersection between datatypes.

right

>
> Another hard entailment:
>
> _:a rdfs:subClassOf xsd:postiveInteger .
> _:a rdfs:subClassOf xsd:byte .
>
> entails
>
> _:a rdfs:subClassOf xsd:unsignedLong .
>
>
> (I don't think my paper helps much with that one).

well, at least it helped to have some rules such as in
http://www.agfa.com/w3c/euler/xsd-rules
e.g.

  { ?s a xsd:int, xsd:nonNegativeInteger } => { ?s a xsd:unsignedLong } .
  { ?s a xsd:short, xsd:nonNegativeInteger } => { ?s a xsd:unsignedInt } .
  { ?s a xsd:byte, xsd:nonNegativeInteger } => { ?s a xsd:unsignedShort } .

but we can indeed forget about completeness ;-)

on the other hand it is surely the case that
"1.0"^^xsd:decimal and "1.0"^^xsd:float
denote the *same* number (they unify)
and that can be clearly realized
e.g. by checking that they are
numeric and then do a compare
after some type promotion

-- ,
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/

Received on Tuesday, 26 November 2002 12:46:55 UTC