Re: Relevant documents on SPIN

On 10/26/2014 11:32, Peter F. Patel-Schneider wrote:
> OWL is not just a syntax, it's a state of mind, and it seems that OWL 
> and SPIN don't share this state of mind.

Correct. SPIN treats OWL as one vocabulary among others, without buying 
into the whole OWL (DL) semantics.

> As mentioned below, rdfs:subClassOf axioms are the only axioms that 
> are considered.
>
> So it seems that SPIN and RDFS don't even share a state of mind then.

SPIN relies on rdfs:subClassOf and rdf:type, but things like rdfs:domain 
and rdfs:range have no meaning for SPIN by default. If people want to 
use these terms of RDFS, then they need to either activate graph-level 
inferencing (or assert the inferences) before running constraint checking.

This WG is trying to come up with a mechanism for closed-world 
constraint checking. This is not a topic that is natively supported by 
RDFS or OWL. Therefore whether OWL and RDFS have a role to play in the 
solution is something that we need to decide on. It is quite possible to 
simply state that these two universes are largely separated, because OWL 
was originally designed with different goals. Then OWL would be left to 
its current use cases (mostly classification problems) while a shapes 
language covers other use cases. SPIN suggests that the meeting point 
between those worlds are the rdf:type and rdfs:subClassOf triples (which 
are arguably by far the most important ontological properties).

Holger

Received on Sunday, 26 October 2014 03:06:03 UTC