- From: Pierre-Antoine Champin <pierre-antoine@w3.org>
- Date: Fri, 16 Dec 2022 14:26:26 +0100
- To: "Peter F. Patel-Schneider" <pfpschneider@gmail.com>, public-rdf-star-wg@w3.org
- Message-ID: <e764182a-3f8d-844c-1d1e-c4ca4e3f6c08@w3.org>
On 13/12/2022 16:45, Peter F. Patel-Schneider wrote: > As far as my proposal goes, when determining entailment all that > counts is the final result so there is again no need to use the > mapping for determining entailment. That is not entirely true. Since this graph << a b c >> :p :o. and that graph _:x1 rdf:subject a . _:x1 rdf:stated-subject "a"^^xsd:string . _:x1 rdf:predicate b . _:x1 rdf:stated-predicate "b"^^xsd:string . _:x1 rdf:object c . _:x1 rdf:stated-object "c"^^xsd:string . _:x1 :p :o. are semantically equivalent, any complying system will have to deal with the mapping one way or another. Of course, they are free to internally store quoted triples as a new kind of term (which, as far as I know, is the case of most if not all existing RDF-star implementations [1]). But then they will be faced with a number of issues. First, when they receive the example above, they will need to "compile" the reification triples into their internal representation of a quoted triple. But if they received _:x1 rdf:subject a . _:x1 rdf:stated-subject "a"^^xsd:string . _:x1 rdf:predicate b . _:x1 rdf:stated-predicate "b"^^xsd:string . _:x1 :p :o1. they would have to store it as is, because there is no quoted triple (yet) being completely described. But as soon as the missing information is added, they would have to "compile" it to a quoted triple. Then what should they do on receiving _:x1 rdf:subject a . _:x1 rdf:stated-subject "a"^^xsd:string . _:x1 rdf:predicate b . _:x1 rdf:stated-predicate "b"^^xsd:string . _:x1 rdf:object c . _:x1 rdf:stated-object "c"^^xsd:string . _:x1 rdf:object d . _:x1 rdf:stated-object "d"^^xsd:string . _:x1 :p :o. ? The same goes for queries. Such system would have to treat SELECT ?x ?y { ?x rdf:subject ?y } as a special case: even if the predicate rdf:subject is never used explicitly in their internal representation, they would have to "generate" it on the fly for each quoted triples in their graph. All in all, such implementation can not just ignore the mapping and deal with the semantics in their own way. The mapping leaks in too many places. pa [1] https://w3c.github.io/rdf-star/reports/
Attachments
- application/pgp-keys attachment: OpenPGP public key
Received on Friday, 16 December 2022 13:26:29 UTC