Re: Why does RDF* allow triples as objects?

William,

I understand you are dissatisfied with the design of RDF, and believe that RDF would have been better if certain decisions had been made differently. Such sentiments are not uncommon among those who have worked with RDF. I myself am no exception.

Nevertheless, we have to acknowledge that RDF* is an extension of RDF-as-is, and not an extension of some RDF-that-could-have-been.

As a matter of fact, in plain RDF, rules of the form

    { ?a :p1 ?b } => { ?b :p2 ?a }

already produce syntactically invalid triples under certain circumstances. So I don't see why one would expect rules of that form to always produce valid triples in RDF*.

Again, allowing triples only as subjects is sufficient to address the stated use cases for RDF* (annotation of statements, and PG interoperability), and makes it easier to define appropriate syntaxes and user interfaces.

Richard



> On 1 Sep 2019, at 14:33, William Waites <wwaites@inf.ed.ac.uk> wrote:
> 
> (Introduction: Hi everyone, I've worked with RDF off and on for 12 or so years,
> recently to help with describing genetic circuits.)
> 
> Richard, why are triples allowed as objects in RDF*? My guess would be because
> arbitrary asymmetries have a cost. We already have the problem that literals
> aren't allowed as subjects which upsets reasoning with even simple rules.
> 
> Why should
> 
>    <<:moon :consistsOf :greenCheese>> :isBelievedBy :bob.
> 
> be allowed, but not,
> 
>    :bob :believes <<:moon :consistsOf :greenCheese>>.
> 
> especially under the rule that you would expect,
> 
>    { ?a :believes ?b } => { ?b :isBelievedBy ?a }.
> 
> Best wishes,
> 
> William Waites | wwaites@inf.ed.ac.uk
> Institute for Language, Cognition and Computation
> School of Informatics, University of Edinburgh
> 
> -- 
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>

Received on Monday, 2 September 2019 09:53:14 UTC