- From: Olaf Hartig <olaf.hartig@liu.se>
- Date: Fri, 30 Aug 2019 10:36:10 +0000
- To: Dan Brickley <danbri@google.com>
- CC: "public-rdf-star@w3.org" <public-rdf-star@w3.org>, David Booth <david@dbooth.org>
Dan, I'm not sure I understand. What do you mean by "graph edges"? Is your idea to use an RDF* triple to capture something like, e.g., "the graph denoted by IRI ex:mygraph contains the edge/triple (ex:Bob, foaf:knows, ex:Alice)." ? Olaf -----Original Message----- From: Dan Brickley <danbri@google.com> To: Olaf Hartig <olaf.hartig@liu.se> Cc: "public-rdf-star@w3.org" <public-rdf-star@w3.org>, David Booth <david@dbooth.org> Sent: Fri, 30 Aug 2019 12:05 Subject: Re: combination of RDF* and graph-level metadata (named graphs) Quick thought: we could also potentially use rdf* annotations to attach multiple graph edges to a named graph. That feels slightly less heavyweight than classical named graphs, although maybe the difference is trivial? Dan On Fri, 30 Aug 2019, 08:52 Olaf Hartig, <olaf.hartig@liu.se> wrote: > David, > > On torsdag 29 augusti 2019 kl. 14:14:39 CEST David Booth wrote: > > [...] > > My main concerns: > > > > - It must be easy to make statements about an entire graph -- a set of > > triples -- rather than one triple at a time. At present RDF* does not > > allow this, but my understanding is that it could be extended to do so. > > IMO this is critically important. > > I don't see why this would be necessary; I mean, I don't think we need > another > such approach. The approaches that you mention below (named graphs, as a > data > model feature, and the N3 syntax to talk about graphs) already give us the > means to make statements about an entire graph. Therefore, in contrast to > these approaches, RDF* focuses on making statements about individual > triples > (similar to standard RDF reification and edge properties in Property > Graphs). > > Making statements about entire graphs and making statements about > individual > triples are orthogonal issues. Of course, there are use cases in which we > want > to be able to do both within the same dataset. To this end, the concept of > a > named graph may simply be extended to be a pair consisting of an IRI (the > graph name) and an RDF* graph (rather than an RDF graph). Then, it is > possible > to use the IRI to make statements about the graph as a whole, and within > the > graph you may have (nested) RDF* triples to make statements about some > particular triple. > > > - It should be harmonized with other existing mechanisms, such as > > named graphs and N3's ability to talk about graphs. > > Does the approach outlined above (named RDF* graphs) address this concern? > > Olaf > > > > Thanks! > > David Booth > >
Received on Friday, 30 August 2019 10:36:40 UTC