From: pat hayes (phayes@ai.uwf.edu)
Date: 11/05/02
Guys, I wonder if you might find RDF typed literals useful? These consist of a literal string plus a uriref which is supposed to indicate a datatype, and the convention is that in 'datatyped RDFS' they denote the value of the string under the lexical2value mapping of that datatype, eg "100"^^xsd:number (Ntriples notation) denotes a hundred. (If there is no such datatype, they essentially act like an unknown uriref: ie they denote something but the reasoner don't know what.) The point that we havn't emphasized is that RDF is not itself restricted to any particular collection of datatypes, so you are free to invent your own and to use them in RDF, and the result is legal RDF. For these purposes, what counts as a datatype is something which 1. defines a class of legal lexical forms (strings) 2. defines a mapping from the class of lexical forms to some class of values 2a. in practice, is able to provide an algorithm for computing the pullback of identity under the inverse of the mapping on the lexical space (eg by computing a canonical lexical form though tha s not the only way) 3. has a uriref. So, for example, one way you could use literals to encode, say, urirefs and obey the RDF thought police would be to use xsd:anyURI as a datatype; similarly for XML and rdfS:XMLLiteral; and similarly for any other thing you wanted to encode (invent your own dataype and assert <yourdatatypeURI>. rdf:type rdfs:Datatype . to keep the RDFS parsers happy.) Of course you can define your own rdfs:subClassOf rdfs:Datatype and have rules which apply to it. In fact, if you want to get creative, it could also *be* an rdfs:Datatype.... :-) Best wishes Pat -- --------------------------------------------------------------------- IHMC (850)434 8903 home 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32501 (850)291 0667 cell phayes@ai.uwf.edu http://www.coginst.uwf.edu/~phayes s.pam@ai.uwf.edu for spam
This archive was generated by hypermail 2.1.4 : 11/05/02 EST