- From: Ross Horne <ross.horne@gmail.com>
- Date: Tue, 19 Oct 2010 17:24:21 +0100
- To: Pat Hayes <phayes@ihmc.us>
- Cc: "public-rdf-dawg-comments@w3.org Group Comments" <public-rdf-dawg-comments@w3.org>
Pat, Can I emphasise the accuracy of your point. Current recommendations are stuck with RDF graphs and RDF named graphs as sets. As a result the latest specification of SPARQL Update provides operational semantics with respect to an index representing time. This makes things very complicated. This complication can be avoided by moving to some notion such as an `RDF resource' which is not a set, as you outline. In short, from the perspective of operational semantics, I strongly agree with you. Ross On 19 October 2010 15:13, Pat Hayes <phayes@ihmc.us> wrote: > > The key point however is that in order to allow for 'updates' while still retaining identity, there has to be some way for the same URI to produce different RDF graphs at different times. This goes beyond the named graph model , which assumes that names are rigid designators for graphs, so (even if we think of the named graph as an information-resource) means that it would be an error to speak of updating a named graph. Maybe we ought to call them named graph resources or some such. > > Pat
Received on Tuesday, 19 October 2010 16:25:47 UTC