Re: shapes-ISSUE-93 (hsolbrig): SHACL engine vs. SHACL instance requirements [SHACL Spec]

I've thought about this one a bit and I think that it's actually a mistake 
to talk about "SHACL engines". I think we've been using this term to refer 
to validation engines but as we've said before there are other possible 
uses of SHACL so validation engines aren't the only possible types of 
"SHACL engines".

With that in mind I think we should try to limit the text of the spec to 
the definition of the language without referring to any sort of "SHACL 
engines". This would be defining what it takes for a SHACL 
document/schema/shapes graph to be conformant with the spec and the 
associated semantics.

Then we could define what it takes for an implementation to be a 
conformant validation engine, validating an instance graph or node against 
a shapes graph. But this should be done in a separate section if not 
document.
--
Arnaud  Le Hors - Senior Technical Staff Member, Open Web Technologies - 
IBM Software Group


"RDF Data Shapes Working Group Issue Tracker" <sysbot+tracker@w3.org> 
wrote on 09/24/2015 01:47:58 PM:

> From: "RDF Data Shapes Working Group Issue Tracker" 
<sysbot+tracker@w3.org>
> To: public-data-shapes-wg@w3.org
> Date: 09/24/2015 01:48 PM
> Subject: shapes-ISSUE-93 (hsolbrig): SHACL engine vs. SHACL instance
> requirements [SHACL Spec]
> 
> shapes-ISSUE-93 (hsolbrig): SHACL engine vs. SHACL instance 
> requirements [SHACL Spec]
> 
> http://www.w3.org/2014/data-shapes/track/issues/93
> 
> Raised by: Harold Solbrig
> On product: SHACL Spec
> 
> Portions of the spec describes what it means to be a compliant SHACL
> "engine".  As an example, Section 3 states "Compliant SHACL engines 
> MUST support all these constraints".  Other compliance points, 
> however, appear to contain recommendations about what would 
> constitute a good SHACL schema.  As an example, section 3.1 on 
> Property constraints states that a sh:property reference SHOULD have
> an rdf:type triple.  From the SHACL engine perspective, there is 
> nothing we can do with this assertion, because SHOULD is only a 
> recommendation, so an engine will need to work correctly whether or 
> not the rdf:type is actually present.
> 
> Similarly, the document recommends the use of rdfs:comments and 
> rdfs:labels, but there doesn't appear to be any assertions about how
> this would change the behavior of compliant SHACL engines.
> 
> I would propose that we create a new document style with a different
> format that will allow us to include all of these these requirements
> and suggestions but would differentiate SHACL requirements from 
> "good coding style" recommendations.
> 
> 
>