Request for Discussion on relation bewteen OData and W3C's R2*

[Stefan Drees <stefan@drees.name>]

Dear Giovanni,

based on the result from discussion during meeting#3 of OData TC (below 
under details the snippet from the minutes) and the
fact that there was no subsequent discussion in this regard on the 
mailing list
I hereby - as a co-member of the tc - kindly suggest, that we do not 
discuss this in the current state during meeting#4 and I further 
propose, that you either

A)	file a JIRA-Issue against the CSDL component (I guess)
	or maybe OData (in general?) to foster further discussion

or
B)	that we let the issue ripen further before we discuss it again
	upon request by a member	

What do you think?

@All: What do the other members think?

This topic is at least alive. There has been a fresh blog entry 
advertising from the Semantic Web Group at W3C: (Call for Review: 
'R2RML: RDB to RDF Mapping Language' and 'A Direct Mapping of Relational 
Data to RDF' Proposed Recommendations 
Published)[http://www.w3.org/News/2012#entry-9535] (from 2012-AUG-14)

Additional Idea:
Maybe feedback from OData TC members in direction to the W3C-Initiative 
would be helpful? I understood, that OData's storage-agnostic 
"Orchestration"-Approach makes it a potential superset of 
storage-dependant interfaces.



Details from notes of meeting#3:

Discussion:

Giovanni
> I think that both the OData TC and W3C RDB2RDF are trying to address 
the same issue. RDB2RDF is an attempt to raise the level of abstraction 
from relational data to RDF (a model based on graphs), similarly OData 
does but using an entity relationship model. What they have in common is 
that they use HTTP and URI refs to access information and put it on the 
Web, the former using a RDF model and the latter using an ER model.

Pablo comments on Giovanni's question:
> Most applications of OData at the moment do not build their service 
interface directly on top of a database, but mostly onto some additional 
(federating) business logic like catalogs. He recommends to understand 
OData **not** as exposing databases, but exposing a more abstract 
business logic.

Mike
> He sees it mostly like Pablo, and adds that 1. the vocabulary used is 
well separated from concrete storage models. So we expose a model, not a 
database. 2. ODatas data model is an entity relationship model. It is an 
abstract model, well separated from the more physical aspects of the 
realization of the server side storag. Might be a relational store, a 
web server, a triplet store and so on.

Ashok Malhotra (Oracle)
> The W3C Proposed Recommendations are at: [R2RML: RDB to RDF Mapping 
Language](http://www.w3.org/TR/2012/PR-r2rml-20120814/) and [A Direct 
Mapping of Relational Data to 
RDF](http://www.w3.org/TR/2012/PR-rdb-direct-mapping-20120814/)

Giovanni
> States that his personal view is that app developers would benefit 
from having a  uniform way to access this information and the way the 
URI is built is an important aspect of the matter.

Chair calls in since allocated time is over. Suggests to continue and 
perhaps talk in next meeting again on this issue raised by Giovanni.


All the best,
Stefan.

Attachment: smime.p7s
Description: S/MIME Kryptografische Unterschrift