Proposal to close ISSUE-69 adopting current draft

ISSUE-69 [1] is about whether strings with language tags shall also 
count as string valued. Here is an example:

ex:Product
     a sh:ShapeClass ;
     sh:property [
         sh:predicate skos:prefLabel ;
         sh:datatype xsd:string ;
     ] .

In my proposal, both of these are valid:

ex:MyProduct
     a ex:Product ;
     skos:prefLabel "Car" ;    # xsd:string
     skos:prefLabel "Auto"@de .

I believe this matches users' expectations, i.e. if sh:datatype is 
xsd:string then rdf:langString should be included. I believe Peter 
argues that this deviates from how XSD Schema Datatypes work, but I 
believe this argument is weak and we should favor intuitiveness over 
formal details in this case. We would of course document this contract 
well. I have not seen anyone use rdf:langString in rdfs:ranges or OWL 
restrictions.

Also, how else could people specify the common case "xsd:string or 
rdf:langString": I guess it would require some complex OrConstraint that 
will be a pain to parse and implement by all tools.

So my proposal is to close ISSUE-69 by adopting the semantics 
implemented in the current draft:

http://w3c.github.io/data-shapes/shacl-ref/#AbstractDatatypePropertyConstraint
http://w3c.github.io/data-shapes/shacl-ref/#hasDatatype

Holger

[1] http://www.w3.org/2014/data-shapes/track/issues/69

Received on Tuesday, 28 July 2015 00:22:48 UTC