Re: ISSUE-139 proposal for today's telco

If we drop the special treatment of inverse property constraints as 
suggested here:

https://www.w3.org/2014/data-shapes/wiki/Proposals#ISSUE-41:_Property_Paths

then the matter of universal applicability is only to distinguish 
between node constraints and property constraints. That removes many of 
the previously meaningless cases that didn't make sense for 
non-literals. If we combine that vote with a variation of proposal 2a at

https://www.w3.org/2014/data-shapes/wiki/Proposals#Topic_2:_Boilerplate_Macro_injection_for_SPARQL

then we have a way forward, because a single, efficient query could 
always be used. This means that engines could always execute each 
context (regardless of whether it makes sense). However, it remains 
important to have a way for tools to select "recommended" constraint 
parameters (e.g. to make sure "primary key" does not show up under "node 
constraints"). I still need to find the time to think through the best 
design pattern for that, e.g. using shape definitions or whether we 
still need something like sh:context.

Holger


On 9/06/2016 16:33, Dimitris Kontokostas wrote:
> As mentioned earlier I would be fine with a hybrid approach for this 
> issue.
> Here's my draft idea to move forward with this issue
>
> As Peter suggested we should have a single default implementation for 
> all supported contexts.
> Users declare which contexts their components supports and provide a 
> single implementation for all those contexts. whenever a component is 
> used in a wrong context, the SHACL engine will return violations for 
> all focus nodes. SHACL engines may additionally use the context 
> information for UI building or suggesting optimizations in the shapes 
> graph.
>
> Users may override the default implementation based on a simple 
> filtering mechanism (as described in this mail 
> https://lists.w3.org/Archives/Public/public-data-shapes-wg/2016Jun/0033.html 
> ).
> A SHACL engine will try to see if any overridden implementation 
> applies for the current instance or use the default as fallback
>
> Best,
> Dimitris
>
> -- 
> Dimitris Kontokostas
> Department of Computer Science, University of Leipzig & DBpedia 
> Association
> Projects: http://dbpedia.org, http://rdfunit.aksw.org, 
> http://aligned-project.eu
> Homepage: http://aksw.org/DimitrisKontokostas
> Research Group: AKSW/KILT http://aksw.org/Groups/KILT
>