W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > February 2003

response to issue pfps-8

From: pat hayes <phayes@ai.uwf.edu>
Date: Mon, 3 Feb 2003 18:15:05 -0600
Message-Id: <p05111b2cba64b174c328@[10.0.100.86]>
To: w3c-rdfcore-wg@w3.org
Cc: pfps@research.bell-labs.com
I believe that there are errors in the treatment of typed literals in
the RDF model theory.  For example, let I be an interpretation where
rdf:XMLLiteral and ex:bar both denote the domain element.  However,
the typed literal "xx"^^ex:bar does not necessarily denote the same
thing as "xx"^^rdf:XMLLiteral because the rule for rdf:XMLLiteral uses
the URI reference directly, and not its denotation.

-----
This is an interesting case. Basically, it reflects an option between 
two ways to think of an XML literal in RDF. On one view, the datatype 
rdf:XMLLiteral should be thought of as simply as another datatype; on 
the other, XML typed literals should be considered a separate special 
case of RDF syntax with its own special rules. Since the built-in 
datatype is somewhat special and has a special relationship to the 
surface syntax (in its treatment of lang tags) , one can rationally 
hold either view, so it is not quite accurate to speak of this 
situation as an 'error'.

However, there is no doubt that the wording of the semantic 
conditions for XML literals is confusing and muddled, and should be 
re-written for clarity. I therefore propose to state the conditions 
referring explicitly to the uri 'rdf:XMLLiteral' as the 'trigger' for 
the special treatment. The situation referred to here, where a 
different uriref denotes the built-in datatype, will indeed not 
trigger the special syntactic treatment. However, this situation 
cannot arise in RDF or RDFS or any datatyped interpretation, since 
there is no way in these languages to state that any other uriref 
must denote the built-in datatype. In OWL-DL and OWL-Lite such a 
statement is syntactically illegal. The only proposed language in 
which such a situation could arise would be OWL-Full, where one could 
assert that

1.  ex:bar owl:sameIndividualAs rdf:XMLLiteral .

Now consider the two triples

2.  ex:aa ex:pp "sss"^^rdf:XMLLiteral .
3.  ex:aa ex:pp "sss"^^ex:bar .

where sss is some wellformed canonical XML, say, and consider the conclusion

ex:aa ex:pp _:x .
_:x rdf:type rdf:XMLLiteral .

this follows from 2. by RDF entailment, but not from 3. However, one 
can imagine a semantic extension of OWL-RDF in which 2 follows from 3 
and 1 together; and then of course the conclusion could be inferred. 
So the issue is whether or not

ex:bar owl:sameIndividualAs rdf:XMLLiteral .
ex:aa ex:pp "sss"^^ex:bar .

together entail

ex:aa ex:pp "sss"^^rdf:XMLLiteral .

and the answer is  no, in RDF(S) or OWL, at present, but maybe they 
might in some future semantic extension, in which case the relevant 
entailment will be possible.

In any case, the present situation seems to me to be coherent and 
also intuitively satisfactory, so (apart from editorial 
clarification) I propose that no action be taken on this issue.

Pat

-- 
---------------------------------------------------------------------
IHMC					(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola              			(850)202 4440   fax
FL 32501           				(850)291 0667    cell
phayes@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam@ai.uwf.edu   for spam
Received on Monday, 3 February 2003 19:13:25 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:55:47 EDT