Re: relevant paper on using OWL (description logics) for RDF validation

On 10/24/2014 09:02 PM, Holger Knublauch wrote:
> Hi Peter,

[...]

> In the implementation section you briefly mention:
>
> "(Recent work at Mannheim implements OWL descriptions as constraints using a
> similar approach.)"
>
> I would like to hear your thoughts about this approach. I assume you are
> referring to the work announced by Thomas Bosch to the previous mailing list
>
> http://lists.w3.org/Archives/Public/public-rdf-shapes/2014Jul/0250.html

Yes, this is the work I alluded to.

> This is a SPIN constraint library that scans the input graph at usages of the
> OWL namespace and reports violations against the closed world interpretation,
> very similar to what you describe in your paper and to what other OWL
> closed-world implementations such as Stardog ICV do.

Perhaps.  It certainly is a large document that mentions various OWL 
constructs.  To determine that it says what you state above would require 
quite a bit of work.

> All that is needed to
> activate this type of constraint checking for a given graph is to mark that
> graph, e.g. with a spin:imports (or ic:imports) triple.

I'm not sure how this would work.  Are you saying that taking an arbitrary RDF 
graph containing some OWL axioms and adding the right imports triple would be 
all that is needed to pass the graph into a SPIN processor and get the 
behaviour you state above?  This doesn't seem possible.

> Basically what this is outlining is that if something like SPIN would be
> standardized, then a SPIN constraints library can be published that knows how
> to interpret OWL restrictions etc. However, people would have the freedom to
> mix and match OWL with other vocabularies, without being limited by the design
> choices of a language that was driven by the computational theory of
> Description Logics. If someone wants to have their graph interpreted with OWL
> closed-world semantics, they would only need to add a triple
>
>      my:Ontology spin:imports <http://w3.org/something/owl>
>
> and a generic SPIN engine can figure out the rest from there, without
> requiring any hard-coded knowledge about OWL. This would in principle allow
> users to reuse existing OWL ontology definitions.

Again, I don't see how this is possible for an arbitrary OWL ontology.

In any case, this doesn't seem to be the way that constraint validation should 
work.  You need at least three kinds of information - data, axioms, and 
constraints - and this appears to provide only two of them.

> Another question I have is how you would address the issue of redefining the
> semantics of the OWL vocabulary that already has numerous books and papers
> written about it, all explaining the details of a rather different open-world
> semantics. That issue could suggest that a fresh start with a new vocabulary
> (such as Shapes) might be a better option than confusing the market with a
> semantically overloaded vocabulary.

I don't see this as a problem at all.

> Thanks,
> Holger

peter


> On 10/23/2014 16:10, Peter F. Patel-Schneider wrote:
>> I suggested that anyone who is going to present technical stuff at the WF
>> F2F meeting next week send out something about the stuff beforehand.
>>
>> I am here covering my marker.  (In a certain sense - this is an academic
>> paper, not a manual, although this stuff is actually implemented in Stardog
>> ICV, more or less.)
>>
>> Questions welcome.
>>
>> peter
>>
>>
>>
>
>

Received on Saturday, 25 October 2014 22:52:54 UTC