W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2002

Re: Datatyping literals: question and test cases

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Wed, 30 Oct 2002 10:25:58 +0300
To: "RDF core WG" <w3c-rdfcore-wg@w3.org>, "ext Graham Klyne" <GK@NineByNine.org>, Patrick Stickler <patrick.stickler@nokia.com>
Message-ID: <fQ5ZJvNhrqln.KjlXVNct@mail.nokia.com>

_____________Original message ____________
Subject:	Datatyping literals:  question and test cases
Sender:	ext Graham Klyne <GK@NineByNine.org>
Date:		Wed, 30 Oct 2002 09:38:43 +0300

This arose when trying to adjust CC/PP to take account of the datatyping 

For starters, under the current datatyping decisions, I understand that:

   _:x ex:prop "10" .
   ex:prop rdfs:range xsd:integer .

is not satisfiable under under any interpretation that conforms to the 
xsd:integer datatype constraints.

I also understand that, following DanC's requests,

   _:x ex:prop "foo" .
   ex:prop rdfs:range xsd:string .

will be satisfiable in conformance with xsd:string datatypeconstraints.

	No, it is *not* satisfyable. (and since when does one member's
	request result in a major change to a WG decision?)

But, what about this:

   _:x ex:prop "http://example.org/" .
   ex:prop rdfs:range xsd:anyURI .

	This also is not satisfyable.

	The ramifications of string-based semantics is that
	there is *no* implicit datatyping of any kind and one must
	explicitly type each and every occurence of every value
	and rdfs:range assertions are only compatable with
	explicitly typed literals.

	It is like having to cast every lexically represented value
	in a programming language rather than basing the 
	interpretation on the type of the variable. E.g.

	one must always explicitly say

		int x = (int) 1 + (int) 2 + (int) 3;


		int x = 1 + 2 + 3;

	as the latter is interpreted as

		int x = "1" + "2" + "3";

	and thus just as a compiler is going to treat the above
	as a type conflict, so also will inlined literals conflict
	with any datatype range assertion.

	(I wonder if all the WG members realize just how divergent
	the RDF string-based semantics are from every programming
	language on the planet and just how *#& EUR 'd up it is)



PS:  in my subject line, I say "test cases", which these are strictly 
not.  How does one express a satisfiability question as a test case?  I 
suppose that is something is not satisfiable then it entails anything you 
care to choose:

    ex:a ex:notSatisfiable ex:b .


    ex:anySubjectYouLike ex:anyPropertyYouLike ex:anyObjectYouLike .


Graham Klyne
Received on Wednesday, 30 October 2002 03:28:58 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:16 UTC