- From: Chris Bizer <chris@bizer.de>
- Date: Fri, 19 Dec 2003 15:09:43 +0100
- To: "Andrew Newman" <andrew@pisoftware.com>, <www-rdf-interest@w3.org>
Hi Andrew, > > I also think that reification seems the appropriate way to do it. > > I've seen obvious use for a context when applying security to a graph > for example. But a fifth one for identifying statings seems > inefficient. Even at the lowest level the combination of s,p,o would > form a unique identifier. > > What do you do I an application needs to capture provenance information (context) *and* statement specific metadata (like trust ratings of a statement)? See my example in http://www.wiwiss.fu-berlin.de/suhl/bizer/trustcontextjustification/#datamodel 2.2 If you use the fourth element of a quad for the context, you cannot refer to a specific statement. If you use the fourth element as statement identifier you have to use a container or something to group statements what is inefficient and ugly when it comes to queries. Using "s,p,o" as unique identifier is out of the bounds of the RDF datamodel. Chris
Received on Friday, 19 December 2003 09:06:24 UTC