Re: RDF WG Resolutions

On 15/05/13 18:31, William Waites wrote:
> Since blank nodes act
> as variables in the SPARQL I don't think there is a serious
> inconsistency.

In terms of handling blank nodes, they do behave a bit like variables 
but in fact they are handled differently.

SPARQL Query evaluation handles blank nodes in Basic Graph patterns 
specially and is not part of the algebra evaluation. This is needed for 
extension to more powerful levels of inference.  Like many standards, it 
is a consensus/compromise.

SPARQL syntax does not allow a blank node in the GRAPH position.  The 
current evaluation design can't be extended to bNodes in that position.

It is true that IF the bNodes are in all the current correct places THEN 
a simple entailment evaluation can treat bNodes like named variables 
with the one catch that SELECT * does not see them.


SPARQL BGP matching:
http://www.w3.org/TR/rdf-sparql-query/#BGPsparql
http://www.w3.org/TR/sparql11-query/#BGPsparql
   (PatH wrote that text)

GRAPH evaluation:
http://www.w3.org/TR/rdf-sparql-query/#defn_evalGraph
http://www.w3.org/TR/sparql11-query/#defn_evalGraph

Discussion of SPARQL Update in recent email to this working group:
http://lists.w3.org/Archives/Public/public-rdf-wg/2013May/0128.html

	Andy

Received on Thursday, 16 May 2013 09:59:07 UTC