Re: questions about rdfs:Datatype [Was: RE: Seeking normative definition of datatyping]

>Thanks for the explanations ...
>
>
>In RDF Semantics there is Section 4.3 (informative) on datatype 
>entailments: rdfD1 and rdfD2 -
>
>What is supposed to happen for an input like
>
>
>xsd:decimal rdf:type rdfs:Datatype
>aaa ppp "120"^^xsd:decimal
>
>in view of  http://www.w3.org/TR/xmlschema-2   Section 3.2.3   in 
>particular the "Note" and 3.2.3.1 ?
>
>How many  triples are supposed to result from applying rdfD1 and 
>rdfD2 - in case the RDF  processor understands xsd:decimal.

The rules should not be thought of as computational rules. They are 
valid inference rules, but a processor is not obliged to use them as 
written in forward-inference mode. I will add a sentence to clarify 
this point.

The direct answer to your question is that they would, or could, 
generate potentially infinitely many conclusions, formed by adding 
'.0' at the end and any number of leading zeros at the beginning, of 
the lexical string. Obviously it would be more practical to restrict 
the use of the rules to a mode where the number of conclusions 
generated were restricted. For example, it would be fine, in 
practice, to always use them so as to replace any lexical form 
representing a decimal by its canonical form with all irrelevant 
zeros suppressed. Whenever there is such a canonical lexical form, 
restricting the rules so as to generate only that canonical form will 
always find identities in at most two steps.

As you probably know, dealing with equations when performing 
inferences is always somewhat computationally fraught, since true 
equations can often generate infinitely many correct but silly 
conclusions. These equations - which is what these rules are, in 
effect - are no exception.

Pat Hayes
-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   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 Thursday, 5 December 2002 12:05:35 UTC