Re: Proposal to resolve ISSUE-86

On 10/16/2015 6:08, Dimitris Kontokostas wrote:
>
>
> On Sat, Oct 10, 2015 at 3:33 PM, Dimitris Kontokostas 
> <kontokostas@informatik.uni-leipzig.de 
> <mailto:kontokostas@informatik.uni-leipzig.de>> wrote:
>
>
>
>     My proposed resolution supports this case. I repeat and extend my
>     proposal to cover Arnaud's and Peter's commets
>
>
>     For ontology / vocabulary (schema) designers who want to associate
>     shapes with their schema definitions the spec will suggest the
>     following 3 means of association:
>     1) the shapes will be written in the same document with the schema
>     2) the shapes graph will be imported using owl:imports
>     3) the shapes graph will be linked from the schema IRI with the
>     property sh:schemaShapes. SHACL engines MAY use this link to load
>     additional shapes in a validation
>
>     It is a common pattern to keep schemas along with the data in the
>     data graph to facilitate data querying.
>     In these cases, options (1) and (2) will result in having the
>     shapes in the data graph. If the user wants to perform a
>     validation with a separate shapes graph, the shapes and the schema
>     will have to be duplicated into the shapes graph.
>     Option (3) keeps the shapes and the schema definition separate and
>     gives the option to the user to load them on demand.
>
>
> Another take to simplify the proposal and avoid objections:
>
> To associate an ontology or vocabulary (schema) to a set of shapes 
> definitions the schema designer MAY link the schema IRI to the shapes 
> graph IRI with the property sh:chemaShapes. SHACL engines MAY use this 
> link to load additional shapes in a validation setting.

Could we call this property sh:defaultShapesGraph? A scenario would then be

     SKOS-Core sh:defaultShapesGraph SKOS-Shapes .
     SKOS-Shapes owl:imports SKOS-Core .
     MyConcepts owl:imports SKOS-Core .

The system can then "infer"

     MyConcepts sh:shapesGraph SKOS-Shapes .

as a default recommendation for engines. No changes to existing SKOS 
models would be required - just the schema would have this extra triple.

Holger