incorrect descriptions of examples of property path queries

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

incorrect descriptions of examples of property path queries

Patel-Schneider, Peter
Several examples in the property paths section of SPARQL 1.1 Query Language do
not do what they claim to do.

*************
Some forms of limited inference are possible as well. For example, for RDFS,
all types and supertypes of a resource:

   { <http://example/thing> rdf:type/rdfs:subClassOf* ?type }
*************

This ignores any subproperties of rdf:type or rdfs:subClassOf, which need to
be consulted when determining types and supertypes of a resource.

*************
All resources and all their inferred types:

   { ?x rdf:type/rdfs:subClassOf* ?type }
*************

As above.

*************
Subproperty:

   { ?x ?p ?v . ?p rdfs:subPropertyOf* :property }
*************

This ignores any subproperties of rdfs:subPropertyOf, which need to be
consulted when determining subproperties of a property.

*************
Elements in an RDF collection:

   { :list rdf:rest*/rdf:first ?element }
*************

This ignores subproperties of rdf:rest and rdfs:first, which need to be
consulted when examining RDF collections in RDFS.


I think that all these need to be called out.


peter

Reply | Threaded
Open this post in threaded view
|

Re: incorrect descriptions of examples of property path queries

Axel Polleres-4
Dear Peter, 

Indeed, I consider this *is* called out by spaeking about "forms of limited inference". No claim is made that either RDFS or any higher entailment regimes would be covered completely here.

Best regards,
Axel

--
Prof. Axel Polleres, WU
url: http://www.polleres.net/  twitter: @AxelPolleres


On 31.10.2014, at 15:41, Peter F. Patel-Schneider <[hidden email]> wrote:

Several examples in the property paths section of SPARQL 1.1 Query Language do not do what they claim to do.

*************
Some forms of limited inference are possible as well. For example, for RDFS, all types and supertypes of a resource:

 { <http://example/thing> rdf:type/rdfs:subClassOf* ?type }
*************

This ignores any subproperties of rdf:type or rdfs:subClassOf, which need to be consulted when determining types and supertypes of a resource.

*************
All resources and all their inferred types:

 { ?x rdf:type/rdfs:subClassOf* ?type }
*************

As above.

*************
Subproperty:

 { ?x ?p ?v . ?p rdfs:subPropertyOf* :property }
*************

This ignores any subproperties of rdfs:subPropertyOf, which need to be consulted when determining subproperties of a property.

*************
Elements in an RDF collection:

 { :list rdf:rest*/rdf:first ?element }
*************

This ignores subproperties of rdf:rest and rdfs:first, which need to be consulted when examining RDF collections in RDFS.


I think that all these need to be called out.


peter

Reply | Threaded
Open this post in threaded view
|

Re: incorrect descriptions of examples of property path queries

Patel-Schneider, Peter
I disagree.  The qualification states that certain kinds of limited inference, e.g., RDFS subClassOf determination, are supported.  This is emphasized in "for RDFS, *all* types and supertypes of a resource" which directly states that all RDFS types and supertypes will be computed.  To compute this requires also looking for subproperties of rdf:type and rdfs:subClassOf.  The example SPARQL query does not do this, and thus does not match the wording.

peter



On Oct 31, 2014, at 4:40 PM, Axel Polleres <[hidden email]>
 wrote:

Dear Peter, 

Indeed, I consider this *is* called out by spaeking about "forms of limited inference". No claim is made that either RDFS or any higher entailment regimes would be covered completely here.

Best regards,
Axel

--
Prof. Axel Polleres, WU
url: http://www.polleres.net/  twitter: @AxelPolleres


On 31.10.2014, at 15:41, Peter F. Patel-Schneider <[hidden email]> wrote:

Several examples in the property paths section of SPARQL 1.1 Query Language do not do what they claim to do.

*************
Some forms of limited inference are possible as well. For example, for RDFS, all types and supertypes of a resource:

 { <http://example/thing> rdf:type/rdfs:subClassOf* ?type }
*************

This ignores any subproperties of rdf:type or rdfs:subClassOf, which need to be consulted when determining types and supertypes of a resource.

*************
All resources and all their inferred types:

 { ?x rdf:type/rdfs:subClassOf* ?type }
*************

As above.

*************
Subproperty:

 { ?x ?p ?v . ?p rdfs:subPropertyOf* :property }
*************

This ignores any subproperties of rdfs:subPropertyOf, which need to be consulted when determining subproperties of a property.

*************
Elements in an RDF collection:

 { :list rdf:rest*/rdf:first ?element }
*************

This ignores subproperties of rdf:rest and rdfs:first, which need to be consulted when examining RDF collections in RDFS.


I think that all these need to be called out.


peter