Re: graph terminology

On 8/17/12 11:04 AM, Thomas Baker wrote:
> On Fri, Aug 17, 2012 at 10:20:23AM -0400, Sandro Hawke wrote:
>>> You have RDF data sources (datasets or triple collections)
>>> associated with RDF data source names (which can take the form of
>>> an IRI). How about that?
>> +1    (I'm happy with those terms, and I take your point, in an
>> earlier email, about different naming conventions working for
>> different communities, even if I'm not sure which communities are
>> more or less ready for RDF.)
> +0.5 - the terms seem okay as long as there's a good story for the difference
> between a "data source" and a "dataset", and between a "dataset" and a "triple
> collection".
>
> Tom
>
Tom,

Yes.

Data Source -- where you get the data from
Dataset -- an actual collection of triples.

Thus, you can have a single Data Source Name associated with one or more 
datasets; which can also be associated with one or more triples.

I believe this can also be mapped to what's expressed in many a VoiD 
graph out in the wild. It also works for SPARQL, even if a Named Graph 
-- in some cases --  maps (ultimately) as a specific Dataset i.e., a 
Dataset with a Name (which could be an IRI).

In a sense, we are still doing denotative indirection albeit not being 
driven by actual IRI/URI de-reference, using HTTP. Instead, this happens 
as part of the SPARQL solution construction.

-- 

Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen