- From: Andy Seaborne <andy@apache.org>
- Date: Fri, 11 Oct 2013 14:39:56 +0100
- To: public-rdf-dawg-comments@w3.org
(personal response) On 07/10/13 20:11, james anderson wrote: > good evening; > > the text in section 5.1 of the "SPARQL 1.1 Graph Store HTTP Protocol" stipulates that > > "If the RDF graph content identified in the request does not exist in the server, and the operation requires that it does, a 404 Not Found response code MUST be provided in the response" > > how is one intended to interpret the passage, "the operation requires that it does"? can it be correct, in particular, for a service to respond to a GET with a zero-length 200 or, would a null graph require a 404? if a 200 is correct, under which circumstances? among the archived discussion about the language in this document, there is one message which identifies the issue to be resolved[1]. it notes, > > > Then "if the specified graph does not exist" must mean that the name is > registered but it names the null graph. But there is another possibility, > which is that the name is not registered in the store. > > which points out, that there are two situations in which "graph content ... does not exist". unfortunately, the message stands without response and the recommendation document remains unclear on the topic. > > if the recommendation were to specify the status code for each case, it would aid interoperability. The target of operations is a Graph Store which is a collection of named slots and an unnamed slot, each slot holds a graph value. The definition is in SPARQL 1.1 Update. GET and DELETE require a web resource to exist otherwise they return 404. The Graph Store HTTP Protocol defines the naming from protocol to graph store slot. A GET, which ends up at a slot with the empty graph, would return 200 and zero triples (which is not zero bytes in RDF/XML, but can be other formats). A GET on a slot that does not exist would be 404. > best regards, from berlin, > --- > [1] : http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Jan/0015.html That raises some rather different issues regarding naming and graphs-as-values. It is not about the target of a SPARQL Graph Store opertation. Section 4 is there to explain this - that talks about "RDF graph content" to try to distinguish graph-as-value and graph in a graph store. Hope that helps, Andy > --- > james anderson | james@dydra.com | http://dydra.com > > > > > >
Received on Friday, 11 October 2013 13:40:31 UTC