- From: Dave Reynolds <der@hplb.hpl.hp.com>
- Date: Mon, 11 Dec 2006 16:42:43 +0000
- To: "Boley, Harold" <Harold.Boley@nrc-cnrc.gc.ca>
- CC: RIF WG <public-rif-wg@w3.org>
Boley, Harold wrote: > Omitting the 'type' attribute can be regarded as specifying the > appropriate 'root' type: a "primitive", arrow, or boolean sort > (http://www.w3.org/2005/rules/wg/wiki/CORE/Conditions/Positive, > section "Multisorted Extensions of the RIF Core"). > > The "primitive" sort is the root as used in order-sorted logics > (often called Any) Ah, that makes sense. > and description logics (in OWL called Thing). Yes but ... > So, the untyped > > <Con>6</Con> > > can be regarded as specifying the owl:Thing root type > (cf. http://www.w3.org/TR/owl-semantics/rdfs.html#5.2) > > <Con type="http://...owl#Thing">6</Con> > > This default type can also be assumed for IRI'zed individuals. > > So, the untyped, webized > > <Con iri="http://en.wikipedia.org/wiki/Number_6"/> > > can be regarded as specifying > > <Con type="http://...owl#Thing" > iri="http://en.wikipedia.org/wiki/Number_6"/> That particular case might be legal but in general I'm not sure the root sort can be OWL:Thing. We want to use IRIs at least for datatypes, property names and class names as well as individuals. In OWL/DL then only individuals are instances of owl:Thing - classes, properties and individuals are mutually disjoint. Perhaps the root sort should be rdfs:Resource. In an OWL/full interpretation rdfs:Resource and owl:Thing do have the same extension but I assume we want to at least leave open the possibility of OWL/DL compatibility. Dave
Received on Monday, 11 December 2006 16:43:12 UTC