From: Ian Horrocks (horrocks@cs.man.ac.uk)
Date: 01/08/02
On January 8, Mike Dean writes: > This is a draft response from the Joint Committee to the RDF > Core Working Group regarding datatyping. We can discuss > this during the Jan 8 telecon. > > One of the dimensions by which one can categorize datatyping > proposals is by whether individual values are explicitly or > implicitly typed, e.g. whether each occurrence needs to > specify xsd:integer (explicit) or whether xsd:integer is > specified as the rdfs:range of the property (implicit). > > We believe that RDF should allow users to choose either > approach, and have adopted this approach in DAML+OIL. > Allowing the user to specify rdfs:range (with or without an > explicit value type) allows for validation; allowing the > user to specify both allows for consistency checking; I would NOT say this as it encourages a common misconception - that range and domaind constraints are/can be used for "validation" (in the sense of checking for "incorrect" data). Like any other part of the DAML+OIL language, range and domain constraints are simply treated as axioms, and lead to "inference" rather than validation. I.e., if you set the range of P to be integer and add a triple <x P y>, then you can infer that y is an integer. Of course this inference may or may not lead to an inconsistency (e.g., depending on whether or not you can also infer a conflicting type/value for y). Regards, Ian > not > requiring an explicit value type promotes consistency with > current use of XML DTDs and XML Schema. We encourage the > RDF Core Working Group to also allow both explicit and/or > implicit typing, and believe this approach to be compatible > with several of the current datatyping proposals. > > Mike
This archive was generated by hypermail 2.1.4 : 04/02/02 EST