Re: proposal for ISSUE-47 - no required access to the shape graph

Although I am on Peter's side, what I suggest is a compromise on both sides.

I say we allow the core to be be defined in the current way if people see
this more convenient but not allow access to the shapes graph outside of
core.
This allows full interoperability between SPARQL based constraints and
frees us from "excluding recursion by blackmail" as Holger noted.

This would be my suggested resolution unless there are other use cases we
can consider.

Dimitris


On Fri, Jun 19, 2015 at 1:28 AM, Holger Knublauch <holger@topquadrant.com>
wrote:

>
> On 6/19/15 6:19 AM, Peter F. Patel-Schneider wrote:
>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA256
>>
>> I propose that SHACL not require that the shape graph be accessible when
>> evaluating constraints on the data graph.  I feel that this allows for
>> effective use of SHACL on more common use cases, including those that use
>> remote graphs such as graphs only accessible via SPARQL endpoints.
>>
>
> ?shapesGraph access is used in the current spec. If you want to drop this,
> you need to present a worked-out counter proposal on how this can be done
> without such access.
>
>     SHACL
>> does not require that the shape graph be accessible when evaluating
>> constraints on the data graph because it is possible to write
>> self-contained
>> queries for all approved SHACL requirements.
>>
>
> Not correct. The requirements do not exclude recursion from sh:valueShape,
> therefore recursion should be supported - at least we should make honest
> attempts to resolve it, and not use this as an excuse to push a specific
> implementation strategy.
>
> Holger
>
>
>


-- 
Dimitris Kontokostas
Department of Computer Science, University of Leipzig & DBpedia Association
Projects: http://dbpedia.org, http://http://aligned-project.eu,
http://rdfunit.aksw.org
Homepage:http://aksw.org/DimitrisKontokostas
Research Group: http://aksw.org