Re: Datatyping Summary

B3 log:implies

[[[
DanC is concerened that with TDL:

   <mary> <haircolor> "red" .

and a rule:

   ?x <haircolor> "red" => ?x <rdf:type> <redhead> .

one cannot conclude

   <mary> <rdf:type> <rdfhead> .
]]]

There is of course a reification in here, with a use/mention issue! Let's
have three contentious issues in one example.

I wonder whether the TDL ==> S-P syntactic transform [1] is useful here:
(Again only for explanatory purposes, not an implementation suggestion).

If we had
  <mary> <haircolor> _:a.
  _:a <rdf:value> "red".

and the rule

  ?x <haircolor> ?y .
  ?y <rdf:value> "red" .
     =>
  ?x <rdf:type> <redhead> .

then I can see a reading of the rule so that it fires, without any range
constraint.

As I am saying (msg with same delivery time) about the query (B4) TDL
clairifies that one could be talking about the lexical form, or one could be
talking about the value, and permitting a decision about which seems like a
good choice to expose to the layer aboce RDF.

So, in answer to Brian's question.
Is the range constraint required by TDL?
  No. It depends on the rule semantics chosen.
  The rule implementator is free to choose.
  In Dan's case I think he should consistently use the lexical forms of
lexical nodes in the interpreation of the rules. I think I am saying that in
a form #2 reification in [2], literal nodes should be mentions, i.e. in the
model theory map them to the lexical form and ignore the value. But then
form #2 was dropped.


[1]Jeremy Carroll: Re: Datatyping Summary
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jan/0369.html


[2] Jan Grant. Proposals? Re: use/mention and reification
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jan/0236.html

Received on Wednesday, 30 January 2002 04:38:50 UTC