Re: Multigraphs example proposal in RDF 1.1 Primer

On 12/13/13 12:43 PM, Kingsley Idehen wrote:
> On 12/13/13 11:59 AM, Yves Raimond wrote:
>> 1) After the TriG listing, adding something like "RDF 1.1 doesn't
>> prescribe any specific semantics for datasets. Possible semantics are
>> described in [DATASET-SEMANTICS]. In this example dataset we assume
>> graph names represent the source of the RDF data held within the
>> corresponding graphs, i.e. by retrieving<http://example.org/bob> we
>> would get access to the four triples in that graph. In the default graph
>> we associate provenance information with the<http://example.org/bob>
>> graph."
>
> How about something like this:
>
> "RDF 1.1 doesn't prescribe any specific semantics for datasets. That 
> said, there are concrete syntaxes associated with RDF (e.g., NQuads 
> and TriG) that enable the creation of RDF documents comprised of 
> statements (triples) qualified (or partitioned) by an IRI that denotes 
> the location of the aforementioned document. The example that follows 
> demonstrates the how such a document would be constructed using NQuads 
> or TriG.
>
>   In this example dataset we assume graph names represent the source 
> of the RDF data held within the corresponding graphs, i.e. by 
> retrieving <http://example.org/bob> we would get access to the four 
> triples in that graph. In the default graph we associate provenance 
> information with the <http://example.org/bob> graph."
>
> My rough sketch example, for which appropriate substitution can apply 
> re. final product:
>
> TriG:
> <http://example.org/bob>
> {<#s> <#p> <#o>} .
>
> NQuads:
>
> <http://example.org/bob> <http://example.org/bob#s> 
> <http://example.org/bob#p> <http://example.org/bob#o> .
>
> Please note:
>
> We shouldn't conflate the needs of a SPARQL implementation with the 
> semantics of RDF based structured data representation and document 
> content construction practices and patterns. This kind of conflation 
> only leads to problems.
>
> [1] http://bit.ly/Jfs12P -- illustrating the world-view (re. RDF and a 
> Sementic Web) behind the example above.
>
> Hope this helps. 
Little correction, due to an errant cut and paste:

How about something like this:

"RDF 1.1 doesn't prescribe any specific semantics for datasets. That 
said, there are concrete syntaxes associated with RDF (e.g., NQuads and 
TriG) that enable the creation of RDF documents comprised of statements 
(triples) qualified (or partitioned) by an IRI that denotes the location 
of the aforementioned document.

Here is an example that demonstrates how such a document would be 
constructed using NQuads or TriG:

My rough sketch example, for which appropriate substitution can apply 
re. final product:

TriG:
<http://example.org/bob>
{<#s> <#p> <#o>} .

NQuads:

<http://example.org/bob> <http://example.org/bob#s> 
<http://example.org/bob#p> <http://example.org/bob#o> .

Please note:

We shouldn't conflate the needs of a SPARQL implementation with the 
semantics of RDF based structured data representation and document 
content construction practices and patterns. This kind of conflation 
only leads to problems.

[1] http://bit.ly/Jfs12P -- illustrating the world-view (re. RDF and a 
Sementic Web) behind the example above.

-- 

Regards,

Kingsley Idehen 
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen