Re: Direct Mapping

On Mon, Sep 6, 2010 at 3:02 PM, Richard Cyganiak <richard@cyganiak.de>wrote:

> On 6 Sep 2010, at 20:33, Juan Sequeda wrote:
>
>> A RDB2RDF system with a R2RML file will always define a RDF graph.
>>
>> So what makes the direct mapping different?
>>
>> They way I see it, is that the R2RML file is generated automatically, and
>> with this file, given a RDB2RDF system, you will have a RDF graph. If
>> other
>> see it different (I can tell that Eric does), then please shout out.
>>
>
>
> The goal of the direct mapping is to have a simple, canonical RDF graph
> representation of every possible database (let's call it the “direct
> graph”).
>
> There are two ways how you could define the direct mapping:
>
> 1. As a mapping from a database to a direct graph
>
> 2. As a mapping from a database to an R2RML file that maps the database to
> the direct graph
>
> I would prefer the first option, because it better supports the
> specification of RDB2RDF mappings using generic RDF-to-RDF transformation
> technologies such as RIF, SPARQL CONSTRUCT, R2R, and so on. If the direct
> mapping is specified as the first option, then RDB2RDF mappings can be
> written using these technologies as a mapping from the direct graph to the
> desired target RDF schema/vocabulary.
>

Both options involve presenting the direct mapping rules for the mapping
from RDB to a direct RDF. Option 2 simply represents this mapping in R2RML.


So you think that a direct mapping shouldn't output the R2RML file? I think
it should because this file is the basis for people to work on and start
customizing it.


> Hence I'm with Eric here.
>
>
>  The automatic mapping file that is generated in D2R is equivalent to the
>> Direct Mapping (right Richard?).
>>
>
> Well I'd say the *graph* produced by an auto-generated D2R mapping file is
> equivalent to the direct mapping.
>

and I'd call the auto-generated D2R mapping file the Direct Mapping file. So
D2R does option 2 then.

>
> Best,
> Richard