Re: Keeping R2RML free of Direct Mapping dependency (ISSUE-25)

On Apr 26, 2011, at 17:39 , Juan Sequeda wrote:

> +1 
> 
> I've been wanting to state this for a while: have the Direct Mapping expressed has R2RML. I guess we can talk today about how we should proceed on this.
> 
> IMO, this will be the starting point for users:
> 
> Scenario 1:
> 1) Generate Direct Mapping. The output will be a R2RML file
> 2) Customize the R2RML file 
> 
> Scenario 2:
> 1) Create views in the database of the data that I want to export (assuming I can create views)
> 2) Generate a (customized) direct mapping. user can choose which tables/views want to be directly mapped
> 3) Use RIF/SPARQL Construct
> 
> I'm betting on Scenario 2. Assuming the user has create view privileges, IMO, it will be easier to create view and direct map, instead of learning R2RML. But this is my opinion (and what I'm betting on). Additionally, this "customized direct mapping" is not part of the spec. But I just wanted to get this use-case out there.

I am not even sure what this customization means to be honest...

However: scenario 2 presupposes that the user uses some additional tool at the end (RIF/SPARQL). If so, then I do not really see the value of that intermediate view. I have the impression that I can express everything I would do for a view via, say, RIF, SPARQL, or other tools that massage RDF data...

Ivan


> 
> Juan Sequeda
> +1-575-SEQ-UEDA
> www.juansequeda.com
> 
> 
> On Tue, Apr 26, 2011 at 10:17 AM, Ivan Herman <ivan@w3.org> wrote:
> 
> On Apr 26, 2011, at 17:11 , Alexandre Bertails wrote:
> 
> > On Tue, 2011-04-26 at 09:53 -0500, David McNeil wrote:
> >>
> >>>
> >>>> If the user wants a hybrid of these two models then they
> >>        can generate the Direct Mapping for an RDB and then replace
> >>        parts of it with a hand-crafted R2RML mapping.
> >>>>
> >>>
> >>> I am not sure I understand that one. You mean generate an
> >>        R2RML that would correspond to a Direct Mapping?
> >>>
> >>>
> >>> What I mean is to use a Direct Mapping tool to produce an
> >>        R2RML mapping file for an RDB.
> >>
> >>
> >>        Yes, that is what I meant. An R2RML representation of the DM
> >>        results for that particular RDB.
> >>
> >>        If we go down that route, it would be worthwhile having an
> >>        appendix in either the r2rml or the dm document that gives a
> >>        precise mapping of the dm to r2rml. This should not be left to
> >>        implementers to be figured out separately.
> >>
> >> Agreed. I was thinking the Direct Mapping was expressed in terms of
> >> R2RML, but I see now that is not the case.
> >
> > The Direct Mapping was never intended to be "expressed in terms of
> > R2RML". But I agree that there could a section about it in the spec.
> > Just remember this will never be the normative definition of the Direct
> > Mapping. And the mapping itself will be parameterized by an instance of
> > RDB.
> 
> Agreed and agreed:-) It should be in an informative appendix of one of the two specs.
> 
> Ivan
> 
> >
> > Alexandre.
> >
> >>
> >> -David
> >>
> >
> >
> 
> 
> ----
> Ivan Herman, W3C Semantic Web Activity Lead
> Home: http://www.w3.org/People/Ivan/
> mobile: +31-641044153
> PGP Key: http://www.ivan-herman.net/pgpkey.html
> FOAF: http://www.ivan-herman.net/foaf.rdf
> 
> 
> 
> 
> 
> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf

Received on Tuesday, 26 April 2011 15:45:42 UTC