- From: Graham Klyne <GK@Dial.pipex.com>
- Date: Mon, 08 Jan 2001 22:18:56 +0000
- To: "Tim Berners-Lee" <timbl@w3.org>
- Cc: <www-rdf-logic@w3.org>
At 03:34 PM 1/8/01 -0500, Tim Berners-Lee wrote: >I have just got hold of Sowa's spec for Conceptual Graph Interchange Format, >and noted down the comparison bwteeen CGs and the semantic web at >http://www.w3.org/DesignIssues/CG >for anyone who is interested, or anyone from the CG community who >would like to verify that it makes any sense. This led me to your 'N3' and 'Anonymous' notes, which had slipped out without me noticing. Considering your comments about expressing quantification. I contemplate an alternative approach, triggered by Pat Hayes' comments about public/proper names. In outline, I suggest a 'properName' property to bind a resource to some label, in addition to 'genid' (a locally generated unique identifier, right?) for the resource. Two different resources that appear in the same context with the same properName value are then presumed to be equivalent in all respects (other than URI). This approach was designed for binding general descriptions into specific instances, but I think there are parallels with quantification. I think this has the advantage that it doesn't in any way modify the semantics associated with RDF resource ids (no special case for genids). It does require an application to be able to track resource equivalence (which I suspect will be needed anyway for applications of any complexity). #g -- Pat Hayes' original comments at: http://lists.w3.org/Archives/Public/www-rdf-logic/2000Oct/0112.html Some later clarification: http://lists.w3.org/Archives/Public/www-rdf-logic/2000Oct/0122.html ------------ Graham Klyne (GK@ACM.ORG)
Received on Monday, 8 January 2001 17:45:50 UTC