Re: Two Standards ?

It's not that we would only have a shape for any class, it is that we might
have zero or more shapes defined for any subset of the graph of interest,
including classes.

Consider this: I want to assert that all predicates in my RDF graph must be
annotated with rdfs:label. I am not, however, stating that this is
universally true of RDF predicates, and nor is this a class (although it
could be considered a class expression).  We already have RDF(S)/OWL for
descriptions and classification, and I think that shapes are not *only*
special intensions of ontologically-defined classes.

all this being said, I think that a shape is itself a class, and can be
described using RDF/OWL. It has special properties and special powers that
we will imbue it so that intelligent applications correctly apply its WG
defined semantics.

m.

Michel Dumontier, PhD
Associate Professor of Medicine (Biomedical Informatics)
Stanford University
http://dumontierlab.com

On Fri, Feb 13, 2015 at 4:21 PM, Holger Knublauch <holger@topquadrant.com>
wrote:

> On 2/14/2015 8:01, Michel Dumontier wrote:
>
>>   I don't agree that OO people will not understand how to deal with
>> shapes that place restrictions on other object or data patterns - there's
>> nothing complicated or bewildering about it.
>>
>
> But from an RDF triples points of view, you end up with two RDF structures
> that basically do the same thing. When to use which, how to correlate one
> to another, why have a "Shape" for every "Class". These are all questions
> that will come up again and again. They increase the implementation cost.
> They increase the cognitive cost. We may end up with two parallel semantic
> webs. Your queries have to do more work.
>
> Classes are already well-established. If we have to add Shapes *as RDF
> triples* then there need to be good reasons for doing so.
>
> Holger
>
>
>