Re: Validation and cool URIs

On 08/06/2014 03:30 PM, Holger Knublauch wrote:
>
> On 8/6/14, 11:00 PM, Arthur Ryman wrote:
>> The above use cases are orthogonal to use cases where an RDF class defines
>> constraints that must be universally obeyed, independent of the context,
>> i.e. a bicycle must have two wheels.
> Yes, that's what I thought. We have seen many more examples of the latter than
> the former, and I believe the structure of existing ontology languages
> suggests that it should be possible to "attach" constraints to classes. I
> personally believe your use case is the exception, not the rule, but of course
> I would like to see a generic solution that covers all use cases so that the
> standardization process succeeds.
>
> I believe the most pragmatic way forward will be to generalize SPIN so that
> constraint definitions can be hosted by a "Shape" object, and that no rdf:type
> triple is required to trigger evaluation. I am confident that we can come up
> with a reasonable design for that use case, and the syntax will probably be
> almost identical. I am however still hoping to get away with the minimum
> necessary language for the job, and keep it simple and intuitive. But we can
> defer this to the WG.
>
> Holger

Well if you only consider constraints for objects of properties, as per the 
charter, then it may turn out that constraints associated with an RDF class 
are in the minority.  Of course, I strongly feel that constraints for objects 
of properties is not a central aspect of RDF validation, but that is what the 
WG is limited to according to the scope section in the charter.

I would much prefer to see this section generalized to something like:

This work group will address these design goals:

RDF graph validation:  constraints on an RDF graph so that it contains 
sufficient information of the right sort to be usable in particular contexts.

Conformance: ...

Extensibility: ...

The first order of business: ...


peter

Received on Wednesday, 6 August 2014 22:57:37 UTC