Re: Input sought on datatyping tradeoff

   [Thomas Passin]
   >Test A1 uses the same predicate and so could be compared.

   [Brian McBride]
   We considered this, and talked ourselves out of it.

      http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jul/0011.html

I don't think this settles the issue.  The example is 

   [From Brian's rdfcore posting]
   _:a dc:title "4th July" .
   _:b dc:date  "4th July" .

   Now add a common super property for all the dc properties:

      dc:property rdf:type        rdf:Property .
      dc:title rdfs:subPropertyOf dc:property .
      dc:date  rdfs:subPropertyOf dc:property .

   This now entails:

   [and here I edit a little, possibly completely missing Brian's point]

      _:a dc:property "4th July"
      _:b dc:property "4th July"

Now we have the same predicate, and hence "4th July" must denote the
same entity, contrary to our previous conclusion.  

I must indeed be missing something.  We've been assuming that for
every property (here, dc:property in particular) there is a parser
that unambiguously specifies how to interpret the strings that occur
as its values, even if we don't know what that parser is.  Well, even
if we don't know what it is in this case, surely we must be able to
spell out a candidate or two.  But as far as I can see there is no
candidate, because in the case of _:b it would have to transform "4th
July" into one value, and in the case of _:a into another.

Hence the triples above are inconsistent, although in practice few RDF
systems will realize that.  (In general it will be undecidable
whether all the parsers for the subproperties of a property are
consistent, but in almost all realistic cases the answer will be
clear.) 

By the way, this settles the question whether all properties could be
subproperties of some super-duper-property.  Obviously, they cannot, so
long as the silly string-parsing idea continues to hold sway.

                                     -- Drew McDermott

Received on Friday, 12 July 2002 18:32:21 UTC