- From: Thomas B. Passin <tpassin@comcast.net>
- Date: Tue, 11 Jun 2002 08:05:14 -0400
- To: RDF Interest <www-rdf-interest@w3.org>
[Patrick Stickler] > > On 2002-06-11 9:14, "ext Joshua Allen" <joshuaa@microsoft.com> wrote: > > >> Awhile back I was thinking about this, and what came to my mind, > >> was this same idea of using UUIDs as a standard identifier for bnodes > > > > Which do you mean: > > A) urn:uuid:blah > > B) _:urn:uuid:blah > > > > I don't think "A" would work. But I think that "B" is completely > > consistent with the RDF Model Theory. > > I meant "B", though not bothering with the prefixes. I.e. > > _:8e4ce93e-7d10-11d6-b93d-0003931df47c > > Thus, we only have three types of label in an RDF graph, URIrefs, > literals, and UUIDs. > Yes, if it doesn't have a leading scheme identifier, then you know it's a bnode. Cheers, Tom P
Received on Tuesday, 11 June 2002 08:05:16 UTC