W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2006

Re: Editorial thread for BGP matching

From: Enrico Franconi <franconi@inf.unibz.it>
Date: Wed, 25 Jan 2006 20:40:17 +0100
Message-Id: <A380E3A9-4E6F-4FE8-BA1C-7247DB1EDDBC@inf.unibz.it>
Cc: RDF Data Access Working Group <public-rdf-dawg@w3.org>
To: Pat Hayes <phayes@ihmc.us>


On 25 Jan 2006, at 20:37, Pat Hayes wrote:
>>>>> I thought the point was that it is always all IRIs (oops s/URI/ 
>>>>> IRI) and some set of blank nodes (but not all of them). That  
>>>>> then ties to "The identifiers introduced by S all occur in B."  
>>>>> in the BGP matching defn.
>>>>
>>>> Well, yes, you are right.
>>>
>>> More precisely: in sparql we have all the B includes only the  
>>> bnodes in G'; then, we have to leave room for future extensions.  
>>> For example, for OWL data queries, B does not include any bnode;  
>>> and there is an understanding of RDF entailment where you may  
>>> want to have all bnodes in B.
>>
>> OK, I have to read what I write :-)
>>
>> s/all the B/that B/
>
> It must be more nuanced than this for the general case. I don't  
> think we can stipulate that B must contain *all* IRIs, in general.  
> For example, OWL-DL would want to restrict the IRIs allowed in  
> answer bindings to exclude the excluded vocabulary. (?Enrico?)

Ah, yes, you're right!

> Also, someone might offer a specialized query service that focused  
> on giving particular classes of answer and ignores others, for  
> example. I think if we use B then we should just keep it as general  
> as possible in the general case, and require future specs to  
> specify it appropriately in each case.

yep, it makes sense.
So we could go back to my proposal of just saying that it should be  
an arbitrary subset of the RDF terms.

--e.
Received on Wednesday, 25 January 2006 19:40:36 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:25 GMT