- From: Bob MacGregor <macgregor@isi.edu>
- Date: Fri, 27 Aug 2004 07:46:22 -0700
- To: danny.ayers@gmail.com
- CC: Steve Harris <S.W.Harris@ecs.soton.ac.uk>, "'RDF interesting groupe'" <www-rdf-interest@w3.org>
Danny, Danny Ayers wrote: > Bob MacGregor wrote: > >> So where is the controversy? Possibly is has to do with the problem >> of nested contexts. The >> most general solutions allow contexts to contain other contexts. If >> the named graph solution >> allows one to apply a name to a subgraph of another graph, then it is >> supporting exactly the same kind of >> nesting. The named document solution does not support nesting. > > Sorry, could you clarify that last point - if you can refer to a graph > through the URI of the container document, then surely can't that URI > be used in another graph/document? yes, but I'm getting at a different issue: If you allow for a containment relation between graphs (say Graph C is the union of graphs A and B), then contexts (named containers) can be arranged into a lattice, according to the containment relation. That induces an inheritance relation between A and C, and between B and C (e.g., all metastatements made about C automatically apply to A and B). That implies a new (to RDF) form of inference. Contexts based only on SOURCE don't have containment relations, so they are much simpler to implement. Cheers, Bob
Received on Friday, 27 August 2004 14:47:08 UTC