ACTION-226: Find out if existential restriction on hassubsystem casuses poor reasoning performance
Find out if existential restriction on hassubsystem casuses poor reasoning performance
- State:
- closed
- Person:
- Krzysztof Janowicz
- Due on:
- November 29, 2016
- Created on:
- November 22, 2016
- Associated Issue:
- Subsystem
- Related emails:
- [Minutes SSN] 2016 11 22 (from phila@w3.org on 2016-11-23)
Related notes:
In the current version of the SSN, hasSubsystem (http://purl.oclc.org/NET/ssnx/ssn#hasSubSystem) is a subproperty of DUL:hasPart
<rdfs:subPropertyOf rdf:resource="http://www.loa-cnr.it/ontologies/DUL.owl#hasPart"/>. I am unsure why we did not use DUL:hasComponent. This is in contrast to properties such as expands in DUL that are explicitly described as proper parts (http://www.ontologydesignpatterns.org/ont/dul/DUL.owl#expands).
Proper Parthood
PPxy \leftrightarrow (Pxy \and \lnot Pyx)
(Assuming that Pxy is not a primitive), one could use overlap instead)
Pxy \leftrightarrow \forall z[Ozx \rightarrow Ozy]
Hence, there is no clear reason why a system should not be its own (and only) part. This means that our current definition of System should not cause any trouble, and this is not even mentioning the OWA which also ensures that we do not run into issues (but works on a completely different level). That said, as System does not use any owl:equivalentClass axioms we should still be fine after removing the existential quantification.
see minutes to https://www.w3.org/2015/spatial/wiki/Meetings:SSN-Telecon20161129
Kerry Taylor, 29 Nov 2016, 22:07:21Display change log.