- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Wed, 20 Nov 2002 21:12:26 +0000
- To: Dave Beckett <dave.beckett@bristol.ac.uk>
- CC: Brian McBride <bwm@hplb.hpl.hp.com>, Jan Grant <Jan.Grant@bristol.ac.uk>, pat hayes <phayes@ai.uwf.edu>, Jeremy Carroll <jjc@hplb.hpl.hp.com>, RDFCore Working Group <w3c-rdfcore-wg@w3.org>
I think Brian has the wrong end of the stick here ...
While it is true that we have forbidden the language tag from having any
significance for datatyping, it is only in the datatype interpretation that
this prohibition takes place.
Thus
Brian's entailment:
>>>>We know that:
>>>>
>>>> <a> <b> "foo"@@en#<datatype> .
>>>> <c> <d> "foo"@@fr#<datatype> .
>>>>
>>>>entails
>>>>
>>>> <a> <b> _:l .
>>>> <c> <d> _:l .
is a {datatype}-entailment, but not any other sort.
Since the test cases are only dealing with XSD entailments then we don't
have to worry about the general case.
i.e. the prohibition on the lang tag meaning anything is not part of the
logic really, just some arbitrary constraint that we have thrown in.
If we make too much of it then Dave's comment:
> This
> "ignoring the language in datatype interpretation except for
> rdf:XMLLiteral" is seeming increasingly stupid.
will win the day.
General principle: you only unpick typed literals, when the type is one of
those in the set of datatypes you are currently dealing with.
If you stick to that then:
- I suggest we do not even consider the boolean entailment
- I suggest the conclusions should always use the canonical form of a
datatype (Brian's bnode in conclusion pattern is also OK).
These suggestions are intended as concessions to implementors.
(Sorry I have not yet looked at the actual tests)
Jeremy
Received on Wednesday, 20 November 2002 16:12:49 UTC