Re: on compliance testing for SHACL

Yes, Sandro reminded us of this during the weekly meeting. We will 
suggest a format for implementation reports and plan to look more into 
the test suite next week - it is a few versions behind and needs to be 
updated. I believe the usual process is that anyone can submit test 
cases and the WG will evaluate them. In cases where such submissions 
were "rejected" as you state, we didn't even have the infrastructure in 
place to record them.

Holger


On 23/02/2017 23:47, Peter F. Patel-Schneider wrote:
> The working group needs to provide a description of how compliance testing
> for SHACL will work.
>
> As the result of SHACL validation can be an RDF graph the testing
> methodology needs to be abe to determine whether this graph is consistent
> with the requirements placed on it by SHACL.  Otherwise it will not be
> possible to either determine whether a SHACL implementation conforms to the
> SHACL specification or whether SHACL implementations are interoperable.
>
> As the mandated behaviour of SHACL implementations includes behaviour that
> might not show up in final results the testing methodology needs to be
> sensitive to the internal behaviour of implementations.  Otherwise it will
> not be possible to determine whether a SHACL implementation conforms to the
> SHACL specification.
>
>
> The working group needs to provide criteria for acceptance of tests.
> Several reasonable tests have been proposed and have been rejected by
> members of the working group.  Therefore the criteria for acceptance of
> tests need to be made clear before advancement to Candidate Recommendation
> status.
>
>
> Peter F. Patel-Schneider
> Nuance Communications
>

Received on Thursday, 23 February 2017 23:27:26 UTC