From: Peter F. Patel-Schneider (pfps@research.bell-labs.com)
Date: 10/02/01
From: Dan Connolly <connolly@w3.org> Subject: Re: XML schema and RDF datatypes [was: comments...] Date: Tue, 02 Oct 2001 14:42:00 -0500 > "Peter F. Patel-Schneider" wrote: > > > > > [Dan Connolly] > > > >It's also possible to design a language where the type of > > > >a literal may *depend* on a declaration from an XML schema: > > > > > > > > <kr:KRLang xmlns:rdf="http://...new-kr-lang..." > > > > xmlns:ex="http://example/vocab"> > > > > <ex:Person> > > > > <ex:name>John Doe</ex:name> > > > > <ex:shoeSize>10</ex:shoeSize> > > > > </ex:Person> > > > > > > > >so that the "10" above is not a logical constant at all; > > > >not until you find a/the schema for http://example/vocab > > > >do you know how to parse/interpret "10"... i.e. the > > > >meaning of that chunk of XML is dependent on all the > > > >trust issues around following links from one document > > > >to another (not to mention a complete implementation > > > >of XML Schema, an effort several orders of magnitude > > > >larger than an RDF 1.0 parser). > > > > > > > >This sort of language is not a candidate for a future > > > >version of RDF: it fails to meet > > > >one of the basic requirements of RDF: that an RDF document > > > >stands on its own as a logical formula. > > > > I fail to see how this follows. > > I'm not sure I can explain it any better. Maybe Pat H. can; > he seems to have understood. My point is that it is possible to create a semantics for RDF/RDFS/DAML+OIL in which literals are untyped but still functions quite well as a logical formalism in which each document (or any other collection of statements) has a perfectly well-defined meaning. For example the datatype extension to DAML+OIL works this way. To me, this forms a counterexample to the conclusion of the rule. > > I would also appreciate a pointer that provides some measure of support for > > the premise of this inference. (In fact, I would find it very instructive > > to see a concise, authoritative enumeration of the ``basic requirements of > > RDF''.) > > Er... I'm not sure what sort of authority you're after. > The RDF Model & Syntax WG didn't make requirements part > of the 1.0 spec. Yes, and I view this as a serious problem. > This excerpt is from a NOTE that has no formal standing, > but perhaps you'd find it convincing regardless of the source... > > [[[ > Lack of ambiguity > > Some programming languages allow one to introduce identifiers > from new name spaces in such a way that it is not possible to > know which namespace a local identifier belongs to without > accessing both the module interface specifications and checking > which one has with the highest priority, or most recently in the > document, redefined a given local identifier. > > This may have some uses in a programming language such as > Java[Java], but it has a serious flaw in that when one module > changes (without the knowledge of the designers of the other > module), it can unwittingly redefine a local identifier used by the > second module, completely changing the meaning of a previously > written document. Clearly, in the Web world in which modules > evolve but documents must have clearly defined meanings, this is > unacceptable. > ]]] > > -- Web Architecture: Extensible languages > http://www.w3.org/TR/NOTE-webarch-extlang#Ambiguity > W3C Note 10 Feb 1998 > > Perhaps you would find a statement from one of the RDF 1.0 editors > compelling: > > "This is indeed an explicit a design decision ..." > -- Ora to www-rdf-logic Mon, 01 Oct 2001 04:28:11 -0400 Yes, I agree that a comment from Ora should have great weight. However, Ora's comment was to the effect that RDF should be context free, I think, and I also think that an approach similar to that of the one in DAML+OIL would pass muster. > As to concise, I condense the design principles of RDF > as follows when I present it: > > [[[ > Semantic Web > Principles > > terms grounded in URI space > > simple XML usage for use with XSLT etc. > > explicit translation to statements > natural language statements > logical formulas > ]]] > -- http://www.w3.org/2001/Talks/0103daml-kt/slide7-0.html > All fine, and none of which prevent untyped literals with typing provided by RDF(S) mechanisms, at least in my opinion. peter
This archive was generated by hypermail 2.1.4 : 04/02/02 EST