- From: Graham Klyne <Graham.Klyne@MIMEsweeper.com>
- Date: Wed, 12 Jun 2002 13:05:13 +0100
- To: patrick hayes <phayes@ai.uwf.edu>
- Cc: w3c-rdfcore-wg@w3.org
I've quickly read through the Lbase spec and scanned the RDF(S) semantics. My immediate thoughts are: (a) Lbase performs a comparable role for RDF(S) semantics that N-triples does for its syntax. (b) The RDF(S) semantics specified by translation to Lbase seems easier to follow and understand than the current model theoretic document. This is, I think, because the explanatory structures used are more familiar to one who has some exposure to mathematics -- does this mean it's also easier to misunderstand? ... And so to some more specific comments on the Lbase specification: Section 2.3: - is the set ID intended to contain any values other than integers and character strings? - the rules of interpretation don't specify the interpretation of names. (Now I think names are not valid formulae, so maybe that's OK, but then again neither are strings and numerals valid formulae. I think a complete set of rules needs to contain all of these.) - I think the syntax allows an expression to contain free variables, but I don't see any corresponding semantics. - are the given interpretations for 'forall' and 'exists' the wrong way round? Section 2.4: - in the description of axiom schemata, it is not clear how one is expected to distinguish "schematic variables" from non-substitutable values. Section 3.1: - just below diagram, a typo: "avove". #g ------------------- Graham Klyne <GK@NineByNine.org>
Received on Wednesday, 12 June 2002 07:51:44 UTC