Re: Attempted to address blank node issue in R2RML draft

Hi Richard,

Should the first statement be modified to include rr:graphColumn also?
(which btw has an incorrect range in section 5, at least inconsistent with
3.4.1.4, which shows range as string).

When I first read your second sentence, I thought it was implying that maps
generating blank nodes could only specify one graph (but subsequent
sentences clarified that).  I wonder if modifying the latter part of the
first paragraph a bit might help make this clearer up-front?  Here's a
possible re-wording of the first paragraph, use any, all, or none as you see
fit.

Triples generated from an R2RML mapping can be placed into multiple different
graphs using rr:graph, rr:graphColumn, or rr:graphTemplate.  However, blank
nodes should never appear in two different graphs, named or unnamed.
Therefore, R2RML processors generating triples conforming with this
specification *must *create a unique blank node identifier for each graph
into which a blank node is placed.

Bob Scanlon
Revelytix


On Tue, Mar 15, 2011 at 4:05 PM, Richard Cyganiak <richard@cyganiak.de>wrote:

> I've tried to address ISSUE-30 in a new section of the draft:
> http://www.w3.org/2001/sw/rdb2rdf/r2rml/#Blank_Nodes
>
> Its contents are appended at the end.
>
> There was a Usage Notes subsection on blank nodes; I merged its contents
> into this new section as both discuss the scope of blank nodes. Since it was
> the only usage note at this point, I removed the Usage Notes section for the
> time being.
>
> This change might go a *little* bit further than what was indicated by my
> ACTION; if anyone objects to this text, please let me know ASAP and I'll
> revert to something uncontroversial.
>
> Best,
> Richard
>
> -------
>
> Triples generated by an R2RML mapping can be placed into different graphs
> using rr:graph or rr:graphTemplate. Blank nodes in R2RML-generated triples
> are scoped to the single graph they appear in. The same blank node can never
> appear in two different graphs, named or unnamed. A single blank node
> identifier will create one distinct blank node for each graph that receives
> triples involving the blank node identifier.
>
> This implies that triples generated from a single logical table row will
> have different subjects if the subjects are blank nodes and the triples are
> placed into different graphs.
>
> Conformant R2RML processors MAY rename blank nodes when providing access to
> R2RML-mapped graphs. This means that client applications may see actual
> blank node identifiers that differ from those produced by the R2RML mapping.
> Client applications SHOULD NOT rely on the specific text of the blank node
> identifier for any purpose.
>