Re: ISSUE-95: What should go into which Turtle file?

Holger,

My view is that we are defining the built-in sh:Constraint subclasses
as part of the vocabulary. This involves some standard properties from
RDFS and some new metadata properties that we define in the SHACL
vocabulary that annotate the sh:Constraint subclasses. However, the
definition of the built-in constraints does not require SPARQL
definitions. Nevertheless, we can provide SPARQL definitions for the
benefit of implementers and validation processors. I'd therefore
suggest that we have a second vocabulary file that owl:imports the
base vocabulary file and adds the SPARQL.

The Shapes definitions of course go in a shapes file, not either
vocabulary file.

-- Arthur

On Sat, Jan 30, 2016 at 1:37 AM, Holger Knublauch
<holger@topquadrant.com> wrote:
> Arthur, all,
>
> as I am working on the new metamodel draft (see [1] for references) I am
> wondering about where to draw the lines between the files. My original
> understanding was:
>
> 1) RDFS file defining all terms in RDF Schema only (classes, properties,
> ranges, domains)
> 2) SHACL file extending the schema with SHACL constraints (details of
> properties)
>
> Questions:
>
> a) Where shall the vocabulary of the extension mechanism go (e.g. sh:sparql,
> functions). We could either add them to the RDFS file or limit them to the
> SHACL file. The latter option may reduce the controversies (some people are
> really just interested in the core language anyway).
>
> b) Shall the SHACL file also contain the SPARQL queries? I have no strong
> opinion here, as we could inject the SPARQL queries from a third file
> maintained outside of the WG, as long as we have stable URIs for the
> subjects.
>
> Thanks,
> Holger
>
> [1]
> https://www.w3.org/2014/data-shapes/wiki/ISSUE-95:_Metamodel_simplifications
>

Received on Thursday, 4 February 2016 16:35:34 UTC