Re: subgraph/entailment

Enrico Franconi wrote:
> On 8 Sep 2005, at 10:37, Seaborne, Andy wrote:
> 
>>For SPARQL to be useable, it must be possible to make queries  
>>against the exact triples in the base data.  Other modes shoudl  
>>also be possible determined by what service/graph/repository is  
>>being queried.
>>
>>It must be possible to access the base data for SPARQL to be  
>>useable for any application that is involved in building a graph  
>>while at the same time accessing it.  I coined the term "zero- 
>>entailment" as I couldn't find extsing terminology.
>>
>>I understand this to be part of the "Local Query" requirement. I  
>>had in mind that the query language be usable in applications that  
>>require access to the base data so they can add new triples, and  
>>they themselves deliver inference services to other applications.
> 
> 
> The point I am raising is in the case you want to be compliant with  
> RDF MT, which seems to me necessary for SPARQL: we don't want SPARQL  
> to be unable to correctly answer queries under the official standard  
> RDF-MT semantics. 

I am unclear in your argument whether you are proposing this to be the only 
mode of operation.

As
  http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Jun/0039
shows, there are cases where the application writer wish to have a syntactic 
view of the graph.  This has always been a problem for semantic web 
applications when a network transfer is involved (example; try to implement a 
distributed RDF graph using standard serializations).

(We decided that we did not sanction explicit handling of bNodes by label but 
at the same time we didn't put any obstacles in the way.)

I would like to see a SPARQL that recognizes that it is important to have RDF 
semantics (and RDFS, and OWL and ...) for query for the case when accessing 
some KB but also that change and modification of the base graph, and using 
query as the data access mechanism to teh synatctic form is also important and 
"SPARQL compliant".

This impacts your example:

<http://example.org/book/book1> dc:title "SPARQL" .
_:b dc:title "SPARQL" .

It would be limiting to not permit a usage that (1) knows there are 2 triples 
in the base graph and (2) in some way get back "_:b" so that the triple

  _:b dc:creator "DAWG" .

is added, thereby changing the answer to

{ ?x dc:title "SPARQL" . }

This itself can't be the only way to use SPARQL but then nor can a 
specification that precludes this.

	Andy



> The RDF-MT semantics introduces two entailments:  
> the simple one, where basically the graph is un-altered, and  
> therefore it corresponds to what above you refer to as "exact  
> triples" or "zero entailment". Then, there is the RDF-entailment. If  
> you take seriously the official standard RDF-entailment as defined in  
> the official standard RDF-MT semantics, 

I do take it seriously.

> I say that you have to have  
> the behaviour that Bijan and me were arguing about.
> 
> cheers
> --e.

Received on Thursday, 8 September 2005 14:45:11 UTC