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

Re: question on PATCH (SPARQL 1.1 protocol and SPARQL 1.1 RDF Dataset HTTP Protocol )

From: Axel Polleres <axel.polleres@deri.org>
Date: Thu, 10 Mar 2011 08:26:29 +0000
Cc: "SPARQL Working Group" <public-rdf-dawg@w3.org>
Message-Id: <D7DC8F4B-62EA-47C8-9D8E-4E1439A25030@deri.org>
To: Chimezie Ogbuji <chimezie@gmail.com>
Hi Chime,

thanks for the clarification, makes sense. 

What I don't really get is the "directly or indirectly" in 
"requests that manage the graph associated with the RDF graph content identified (directly or indirectly)"

Is it, you say that PATCH for an update request without GRAPH in the modify-template, applies to the graph in the graphstore given with the graph paramenter? Wouldn't it then make sense here to simply forbid GRAPH entirely in the modify_template for such requests and say that 

 PATCH graph <g1>
  payload:
    DELETE/INSERT { triples_block }
    ...
 amounts to 

    DELETE/INSERT { GRAPH <g1> {triples_block }}

 in SPARQL protocol.

As it stands, it seems that also 
that you could have <g1> both as a parameter for ?graph and in the request, ie.
 PATCH graph <g1>
  payload:
    DELETE/INSERT { GRAPH <g1> {triples_block }}
would be allowed, or do I misunderstand something?

cheers,
Axel

 
On 9 Mar 2011, at 17:24, Chimezie Ogbuji wrote:

> Hello Axel.
> 
> On Wed, Mar 9, 2011 at 12:01 PM, Axel Polleres <axel.polleres@deri.org> wrote:
> > Hi all,
> > I think the PATCH section in the current SPARQL 1.1 RDF Dataset HTTP Protocol doc [1] looks a bit alien.
> > It seems to just say that PATCH can be used for UPDATE requests, which has nothing directly to do with the Dataset protocol, it seems.
> 
> There was a previous thread that resulted in the current language.  In
> addition, we had several discussions in one or more teleconferences on
> this topic.
> 
> See: http://lists.w3.org/Archives/Public/public-rdf-dawg/2010OctDec/0458.html
> 
> That (informative) part of the document is saying PATCH can be used
> for a restricted form of UPDATE requests that are RESTful (i.e., the
> requested UPDATE message only manipulates the graph content being
> addressed via PATCH, for instance) and recommends appropriate HTTP
> behavior and error conditions in that scenario.  This does have to do
> with the Dataset protocol since it is due to the fact that it is a
> RESTful protocol that has constraints on the interface that the
> restrictions and guidelines are given in that section.
> 
> > That indicates probably, that this rather belongs to our "regular" protocol document?
> 
> I don't think so.  The regular protocol document does not have to
> adhere to any (RESTful) constraints that the dataset protocol does,
> which is why we have the latter in the first place.
> 
> > It's probably just not yet in the 1.1 protocol document because no one has been working on the 1.1 protocol document for a while,
> > but I think the whole Informative section on PATCH in could be dropped in the HTTP Dataset protocol doc or just be shortened to a reference to the Protocol Doc.
> 
> > Opinions?
> 
> I disagree for the reasons given above, there has already been comment
> threads on the topic and some indication for the desire to support a
> RESTful mechanism for using the UPDATE language, the current text is
> informative, and there is currently nothing equivalent to reference in
> replacing it.
> 
> -- Chime
> 
Received on Thursday, 10 March 2011 08:28:33 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:45 GMT