W3C home > Mailing lists > Public > public-rdf-star@w3.org > October 2020

Re: weakness of embedded triples

From: Pierre-Antoine Champin <pierre-antoine.champin@ercim.eu>
Date: Tue, 27 Oct 2020 23:17:58 +0100
To: "Peter F. Patel-Schneider" <pfpschneider@gmail.com>, public-rdf-star@w3.org
Message-ID: <2f71590f-3b15-587c-4a4a-475182c10d00@ercim.eu>
Peter,

On 27/10/2020 20:01, Peter F. Patel-Schneider wrote:
> [Back to the original topic of this thread.]
>
>
> I was looking at the semantics of RDF* and I think I have discovered further
> weakness of embedded triples.  In the semantics for RDF* there is a special
> mechanism to allow:
>
> Mary believes << John loves _:a >>
>
> to RDF*-entail
>
> Mary believes << John loves _:b >>
>
>
> But this mechanism is very weak.  As far as I can tell
>
> Mary believes << John loves _:a >>
>
> does not RDF*-entail
>
> Mary believes << John loves _:c >>
>
> Mary believes << John loves _:b >>
>
>
> nor does
>
>
> Mary believes << John loves Susan >>
>
> RDF*-entail
>
> Mary believes << John loves _:a >>
>
>
> Both of these non-entailments seem to me to be significant problems with RDF*.

You are right, these entailments do not hold under the current
semantics, and I agree that this is a serious problem.

Actually, your initial questions in this thread made use realize these
problems -- thanks for that ;-) -- and we have started working on a
revised semantics which is currently in a pull request, and that we
intend to merge soon. In the meantime, the PR is here:

https://github.com/w3c/rdf-star/pull/19

  best

>
>
> peter
>
>
>


Received on Tuesday, 27 October 2020 22:18:15 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 27 October 2020 22:18:16 UTC