Re: less-restrictive range and domain terms

Peter F. Patel-Schneider writes:

> From: David Menendez <zednenem@psualum.com>
> > However, for the purposes of UI support, which (I think) was Phil's
> > original goal, there's no need to make any implications at the RDF
> > level.
> 
> Why not?  Oh, you mean that the UI doesn't need to make inferences? 
> Maybe, but then the UI won't warn you if you are making
> contradictions (which may be possible in this extension of RDF).

No, I think a decent UI would want to infer things. What I meant was
that any conclusions the UI draws from its use should stay at the level
of human interaction, rather than be applied to the data.

Right now, RDF editors can use vocabulary definitions to detect errors
(contradictions). Annotation properties like "usualRange" and others
would give editors enough information about intended usage to raise
warnings when they find something unusual but not necessarily wrong or
impossible.



(rambling speculation follows)

You could also describe the warnings in RDF. Like,

    _:fido p:owner [ a p:Gorilla ].

could generate a graph describing what's wrong.

    [ a eg:ObjectOutsideUsualRangeWarning;
    ; eg:for
      [ a rdf:Statement
      ; rdf:subject _:fido
      ; rdf:predicate p:owner
      ; rdf:object _:x
      ]
    ; eg:expectedClass p:HumanAgent
    ; eg:declaredClass p:Gorilla
    ; rdf:value
        "Object (class 'Gorilla') out of usual range ('Human Agent')"
    ].
-- 
David Menendez <zednenem@psualum.com> <http://www.eyrie.org/~zednenem/>

Received on Tuesday, 4 May 2004 20:29:10 UTC