Re: [Graphs] Proposal for Named Graph Semantics

On 8 Apr 2011, at 15:03, Alex Hall wrote:
> The use cases that I saw this proposal as most useful for addressing were:
...
> 3. Having a coherent story to tell when a newcomer to RDF and SPARQL asks, "what does that IRI in this FROM clause mean?"

For the simple abstract syntax proposal, my answer would be this:

"The FROM clause is part of a query that you're sending to a SPARQL service. The service allows you to query a particular dataset. Within that dataset, there are multiple graphs. The IRI in the FROM clause selects one of these graphs. Now what's in this graph? That depends on the dataset, and the answer would be different for every SPARQL service. In the case of this SPARQL service here, the graph for any IRI is the RDF that you get by resolving the IRI on the Web and sending it through an RDF parser. This is a quite common pattern. But in another case, the graph for an IRI might be the result of running entity extraction over the HTML that you get back when looking up the IRI. In yet another case, the graph for an IRI could contain the web server access log for the web page with that IRI. In yet another case, the dataset may only have three graphs named with three particular IRIs. You have to refer to the documentation for the particular SPARQL service to know what to expect in any given graph."

Best,
Richard

Received on Friday, 8 April 2011 15:43:18 UTC