Re: R2RML: publishing the second draft

Hi Alexander,

This looks very good.  Here's a few comments/suggestions:

   - It looks like tableOwner is missing in TripleMapClass -- that should be
   an optional property, per ISSUE-23
   - It looks like inverseExpression is missing in SubjectMapClass and
   ObjectMapClass; suggest inserting above graph properties to keep those
   together at the end or property listings  (it's in PredicateMapClas but
   without a range)
   - RefObjectMapClass.joinCondition range is a string (currently missing)
   - Suggest moving template up under column in SubjectMapClass and
   ObjectMapClass, for aesthetic reasons (it's one of 3 alternatives to specify
   a subject or object, the other 2 are at the top of the property list)


Bob Scanlon
Revelytix



On Tue, Mar 15, 2011 at 6:01 AM, Alexander De Leon <me@alexdeleon.name>wrote:

> Hi
> I created a UML class diagram of the R2RML syntax:
> http://mccarthy.dia.fi.upm.es/rdb2rdf/img/R2RML.png
>
> Any comments/corrections?
>
> Alex
>
> On 15/03/2011, at 00:25, Boris Villazón Terrazas wrote:
>
> > Hi Souri
> >
> > On 14/03/2011 17:21, Souripriya Das wrote:
> >> Latest Editors' draft (v 1.42):
> http://www.w3.org/2001/sw/rdb2rdf/r2rml/
> >>
> >> Worked on the R2RML document to address the following issues to make it
> ready for publishing:
> >>      • Issue-22: added note to Sec 1.3
> >>      • Issue-24: changed "child." to "{child.}" and added a note about
> allowing complex join conditons.
> >>      • Issue-26: added rr:unnamedGraph IRI as a possible value for
> rr:graph property to indicate that the generated triples should be stored in
> the unnamed graph (besides storing them into specified named graphs, if
> any).
> >>      • Issue-31: Changed "blank node label" to "blank node identifier"
> >> Issues to be addressed:
> >>      • Figures need to be consistent with the current text. (Boris,
> could you please upload the latest figures or send us the pointers.)
> > I will upload the figures, once I update them.
> > In addition, Alex is creating a new figure that describes the R2RML
> specification, we will send you this figure.
> >
> >>      • Issue-28: Needs a test case. (Boris, could you please add a test
> case.)
> > Done, in a previous "Test Cases" email.
> >
> > Boris
>
>