Re: bang ! in turtle

Le 02/03/2011 10:23, Pierre-Antoine Champin a écrit :
> On 03/01/2011 07:48 PM, Nathan wrote:
>> Ivan Herman wrote:
>>>> ^a [ owl:inverseOf rdf:type ]
>>>
>>> As Sandro said, this is just an explanation, right? Formally, what
>>> you describe is a bnode that can be inferred to be an inverse, etc,
>>> etc
>>
>> I'm happy with whatever words make it happen, one view is that it's
>> sugar, the other is that it saves you declaring inverseOf properties
>> all the time - either one roughly equates the same net result.
>
> No its does not have the same net result, as it does not produce the
> same graph!

Yes.

>
> Consider
>
> :a ^:b :c .
>
> in one case, it produces graph A below (1 triple)
>
> :c :b :a .
>
> in the other case, it produces graph B below (2 triples)
>
> :a _:x :c .
> _:x owl:inverseOf :b .
>
> which entails graph A *only* under the inference regime OWL.

So called "Graph B" is not an RDF document and therefore, not an OWL 
document either. So under no standard semantics this would make sense.

But let us admit for a moment that we have generalised RDF.
In this case, even under the OWL RDF-based semantics, the two graphs are 
not equivalent.

"Graph B" OWL-entails "Graph A" but "Graph A" *does not* OWL-entail 
"Graph B".

>
>  From a strict RDF perspective, this is two completely different things.

Right.

-- 
Antoine Zimmermann
Researcher at:
Laboratoire d'InfoRmatique en Image et Systèmes d'information
Database Group
7 Avenue Jean Capelle
69621 Villeurbanne Cedex
France
Lecturer at:
Institut National des Sciences Appliquées de Lyon
20 Avenue Albert Einstein
69621 Villeurbanne Cedex
France
antoine.zimmermann@insa-lyon.fr
http://zimmer.aprilfoolsreview.com/

Received on Wednesday, 2 March 2011 10:41:30 UTC