- From: Pierre-Antoine Champin <swlists-040405@champin.net>
- Date: Wed, 31 Jan 2007 17:05:17 +0100
- To: Adam <adamsobieski@hotmail.com>
- CC: semantic-web@w3.org
Adam a écrit : > In reifying statements from other sources or in a document discussing > its own statements, it appears that we have two alternatives. First, we > can use the syntax where the statement is an element and contains > subject, predicate and object elements. This leads to replication, where > an author must repeat the statements in this context. The second > approach is to consider them to be a subgraph literal which again > removes them from their original context or otherwise duplicates them. > > I would like to propose a third alternative to the discussion while the > topics of reification and provenance in RDF are being considered. This > idea uses the relationship between RDF and N3: that any RDF-based > document can be represented uniquely as a sequence of triples. I do not quite agree: any RDF-based document can be represented uniquely as a *set* of triples. The order of the triples is not relevant. Of course, one may define a canonical order (the most obvious being alphabetically sorting the triples), but that would make it costly to compute for any document that is not represented in that canonical form. Furthermore, I expect that some RDF "resources" would not be "documents", where one can query the triple but not get an exhaustive list of them. Pierre-Antoine
Received on Wednesday, 31 January 2007 16:05:31 UTC