Re: [OK] Re: [OK?] Re: comments on "SPARQL Query Language for RDF"

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
> > The scoping graph is always isomorphic to the original dataset - the  
> > only things that may change are the names of the bnodes. This has no  
> > effect on the answer set, apart from having possibly different bnode  
> > names.
> 
> Sorry, I meant scoping *set*.

Ah. Well, for (this version of) SPARQL the scoping set contains exactly the URIs, the Literals, and the bnodes from the scoping graph. It is a theorem that unde
r these restrictions the definitions given in 2.5 uniquely identify the answer set - modulo renaming of bnodes - such that the answer set is exactly the outcome
 of the homomorphisms between the query and the scoping graph.

cheers
--e. 

Received on Tuesday, 7 March 2006 15:47:15 UTC