- From: Andy Seaborne <andy.seaborne@epimorphics.com>
- Date: Mon, 30 Apr 2012 09:02:43 +0100
- To: public-rdf-wg@w3.org
On 30/04/12 08:38, Ivan Herman wrote: > > On Apr 30, 2012, at 09:24 , Pat Hayes wrote: [snip] > >> >>> >>> >>> I must admit I am not sure what >>> http://www.w3.org/2011/rdf-wg/wiki/Graphs_Design_6.1#Blank_Nodes >>> (ie, that blank nodes have a file scope) mean eg, in terms of >>> semantics. If I look at the more abstract level >>> >>> (D, (<u>,G), (<v>,H)) >>> >>> with G and H being different graphs, what does it mean that they >>> share a blank node? >> >> Exactly what it says. There is nothing in the 2004 RDF specs that >> prevents two different and distinct graphs from sharing a blank >> node. (As to whether there SHOULD have been something preventing >> this, maybe so: but in fact, there isn't.) >> > > I am surprised but I of course believe you:-) > > However... Are we sure that existing systems (RDFLib, Jena, > you-name-it) are prepared for this? Many of those have some sort of a > named graph/quad store implemented already, possibly with TriG input, > and it would be good to know whether this would force them to > re-engineer their blank node processing workflow... > > Ivan Jena has no problem with bNodes shared between graphs. It happens - a inference graph and it's base graph share bNodes in the base graph. (For RDFS, the base graph is a subgraph of the inference graph.) In Jena, blank nodes have a system-wide internal id, and are not identified relative to the graph they are in. The identifier is global, it is not a IRI (two different spaces of names). The identifier is not related to the label used in the syntax file. Andy
Received on Monday, 30 April 2012 08:03:12 UTC