...
Conclusions: No show stoppers at present. The issues whereby the DTC inferred to be both depository and non depository, this needs to be fixed (see above). This revolves around insurance. Q: Is the insurance definitional of the kind of thing, or is it regarded as a clue that it is a given kind of thing? The insurance is used as evidence for a thing being a given class, not definitional of it being that class of thing. So we are using facts about insurance as a kind of proxy for the actual facts of the matter. If you are a given kind of thing, it then follows that you have to be insured to be that kind of thing. Note that pellet draws one set of inferences, and TrOWL does not draw that inference. The restrictions in use for the above use cardinality, and Elisa says that means they are not defining a necessary condition. If we had someValuesFrom then both Pellet and TrOWL would regard this as a necessary condition and would draw the same inference. TrOWL does not regard the cardinality restriction as being a necessary condition even though Pellet does. TrOWL recognizes a distinction between cardinality restrictions and the allvalues and somevalues types of restrictions. Up to now we all thought that whether something is a necessary condition was simply a matter of whether it is an equivalent class or a sub class of the restriction, regardless of the restriction type. so TrOWL teaches us a new nuance about how these restrictions work. We need to understand this, and raise a bug on whichever of these reasoners turns out to be incorrect on this. Mike Uschold and Dean Allemang interpret the meaning of a restriction based on what Pellet says. Dean and Mike U prove logically, regardless of the reasoners, what was being shown, and that in this case the two were identical. Elisa notes there are reasons why you would use one kind of restriction over the other depending what you are trying to say, but has not been able to get clarity from the experts (PatelSchneider and McGuinesss), but this is a DL nuance we need to find out the industry consensus one. Elisa will also ask Evren Sirin and Jeff Pan what is happening here. www.cs.vu.nl/~guus/public/owl-restrictions/
Action items
- Elisa Kendall Review old JIRA issues around LEI and some BE reconciliation actions and the list of regulators with David - some of them are in the OWL but somehow not in the specification dcument
.Jira Legacy showSummary false server jira.edmcouncil.org JIRA serverId 6465cde0-d6c9-34a0-a506-4d4ac3461fe5 key FBC-62