W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2011

Re: rdf dataset parameters in protocol for sparql update

From: Lee Feigenbaum <lee@thefigtrees.net>
Date: Mon, 05 Sep 2011 22:33:38 -0400
Message-ID: <4E658682.6050503@thefigtrees.net>
To: Andy Seaborne <andy.seaborne@epimorphics.com>
CC: public-rdf-dawg@w3.org
On 8/29/2011 12:58 PM, Andy Seaborne wrote:
> On 16/08/11 03:11, Lee Feigenbaum wrote:
>> 1/ The parameters are called using-graph-uri and using-named-graph-uri
>> 2/ If at least one value for one of the parameters is present, then the
>> protocol parameters fully define the RDF dataset that is used for the
>> query pattern matching (WHERE clause) of all operations in the update
>> request. That is, they fully replace all USING and USING NAMED clauses
>> in the request's operations.
>> A third way of stating this: the semantics of including at least one of
>> using-graph-uri and using-named-graph-uri is the equivalent of doing the
>> following:
>> A) Remove all USING clauses from the request.
>> B) Remove all USING NAMED clauses from the request.
>> C) For each using-graph-uri=u in the protocol request
>> For each INSERT/DELETE/INSERT DELETE operation in the request
>> Add "USING u" to the operation
>> D) For each using-named-graph-uri=u in the protocol request
>> For each INSERT/DELETE/INSERT DELETE operation in the request
>> Add "USING NAMED u" to the operation
>> Protocol says nothing further about the semantics of these parameters --
>> everything else is based on the semantics already given for USING and
>> USING NAMED in SPARQL 1.1 Update. This includes things like whether
>> "USING x" mentioned repeatedly in an update request refers to a mutable
>> graph or a graph as retrieved from the Web; this includes things like
>> the interaction between these parameters and WITH.
>> Our options at this point:
>> + Discuss any questions
>> + Consider any alternative proposals
>> + Consider dropping these parameters altogether
>> + Consider adopting this proposal
> This does not address my concern that an update of several operations
> may have different USING clauses in different operations. Overridding
> any USING, and changing operations that don't mention USING, makes it,
> to me, a different situation to query. The structure of the update
> request is changed.
> Proposal: allow using-graph-uri and using-named-graph-uri in the
> protocol if the operations of the request do not have any mention of
> USING/USING NAMED. (A condition on uniform use of USING/USING NAMED
> seems overly complicated.)

I've thought more about this.

On one hand, I think the above proposal is sort of weird to specify in 
the protocol document. On the other hand, my core use cases involve 
specifying USING/USING NAMED externally, and don't really rely on 
overriding them, so I think I could support this proposal.


> Andy
Received on Tuesday, 6 September 2011 02:34:36 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:01:04 UTC