Re: RDF Data Shapes WG agenda for 15 October 2015

On 10/15/2015 4:55, Peter F. Patel-Schneider wrote:
> ISSUE-95
>
> See also
> https://lists.w3.org/Archives/Public/public-data-shapes-wg/2015Oct/0059.html
>
> My view is that two significant problems with SHACL are that there are
> abstract classes and that templates use rdfs:subClassOf for inheritance.
> Simplifications that do not fully overcome these two problems are not worthwhile.

It would be helpful if you could make a counter proposal then.

> ISSUE-96 (same as last week)
>
> I feel that SHACL validation results already contain adequate information to
> identify the construct in question.  Adding more information only complicates
> an already-complex system.  I vote -1 for such additions.

The current spec does not make any property such as sh:sourceTemplate 
mandatory, neither is there an official link from properties such as 
sh:minInclusive to the template or function that has produced the 
violation in the spec. If we want to make validation results 
interchangeable so that tools can switch from one engine to another, 
then we need to come up with something here.

I actually ran into a specific use case yesterday for our Reference Data 
Management tool: We want to distinguish the kinds of violations based on 
their results so that we can display *incomplete* data (sh:minCount) in 
a different dashboard chart than *incorrect* data. We also want a pie 
chart displaying the different kinds of violations. Yes, TopBraid can 
invent its own mechanism to do that, but this means that TopBraid would 
not be able to switch to other (database) vendor's native SHACL engines. 
All we would get back is unstructured results data with random text in it.

Even if you don't think *you* need this feature, please consider 
changing your vote to -0.9 to at least not block this for those who have 
proven needs for it.

Thanks,
Holger

Received on Thursday, 15 October 2015 04:10:34 UTC