W3C home > Mailing lists > Public > public-rdb2rdf-wg@w3.org > September 2011

Re: Cardinality of predicate maps in predicate-object maps

From: Souripriya Das <souripriya.das@oracle.com>
Date: Tue, 20 Sep 2011 14:27:40 -0400
Message-ID: <4E78DB1C.7010903@oracle.com>
To: "public-rdb2rdf-wg@w3.org RDB2RDF" <public-rdb2rdf-wg@w3.org>
[notation: POmap => PredicateObjectMap, Pmap => PredicateMap, Omap => 

Since this "small" change over the 2nd Working Draft was not formally 
raised as an ISSUE, we never really were able to consider it formally as 
a group. So, this small change (effectively a "shortcut") was reverted 
back to the original scheme presented in the 2nd Working Draft that 
required a POmap to consist of exactly one Pmap and exactly one Omap.

Also, unless there is a pressing need for such a shortcut, asked for by 
the user community at large, requiring implementations to allow a POmap 
to have >1 Pmaps and one Omap would unnecessarily burden the 
implementer. If needed, multiple POmaps, each with one Pmap and one 
Omap, can always be used.

- Souri.

Richard Cyganiak wrote:
> This is a comment on the Last Call draft of R2RML [1].
> Section 6.3 [1] defines predicate-object maps as allowing only exactly one predicate map.
> I originally proposed to allow multiple predicate maps in [2] along with some rationale. EricP explicitly +1'd this specific change in the ensuing thread [3]. This specific proposal drew no further comments in the change, and it was implemented in this revision, in early July:
> $Id: Overview.html,v 1.94 2011/07/07 12:51:32 rcygania2 Exp $
> It remained like this for the following two months. However, it was reverted immediately before LC here:
> $Id: Overview.html,v 1.151 2011/09/13 18:37:42 sdas2 Exp $
> I note that Souri is on the record before Last Call as saying: “we had those four points only: …” [4] and this change is not among the four listed. I acknowledged the four points in [5] and didn't receive any response that indicated any additional planned changes.
> So I think this change should not have been made according to the process and that it should be reverted.
> Best,
> Richard
> [1] http://www.w3.org/TR/2011/WD-r2rml-20110920/#predicate-object-map
> [2] http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2011Jul/0009.html
> [3] http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2011Jul/0012.html
> [4] http://www.w3.org/2011/08/30-rdb2rdf-minutes.html
> [5] http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2011Aug/0184.html
Received on Tuesday, 20 September 2011 18:29:11 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:00:26 UTC