W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2012

Re: Graph store protocol editor's draft updated

From: Gregory Williams <greg@evilfunhouse.com>
Date: Fri, 17 Feb 2012 13:11:24 -0500
Cc: SPARQL Working Group <public-rdf-dawg@w3.org>
Message-Id: <E6999324-8CCC-4C7E-AEC0-72126CA78B01@evilfunhouse.com>
To: Chimezie Ogbuji <chimezie@gmail.com>
On Feb 17, 2012, at 10:11 AM, Chimezie Ogbuji wrote:

> Seems like we are splitting hairs here.  See my response below.
> On Fri, Feb 17, 2012 at 9:57 AM, Gregory Williams <greg@evilfunhouse.com> wrote:
>>> The GSP document has the following explicit text (in section 4.2):
>>> "In a similar manner, a query component comprised of the string
>>> default can be used to indicate that the operation indirectly
>>> identifies the default graph in the Graph Store."
>> My concern here is twofold. First, this is halfway through a document which is entirely focused on interacting with graph stores, and this is essentially the first time default graphs are discussed.
> This document doesn't define what a graph store is, only how to
> interact with it over HTTP.  Defining (precisely) what a default graph
> is (or a Graph Store for that matter) is out of scope for this
> document.

My concern is that I don't think it's clear that the actual definition is in the Update doc, and that what is given in the GSP Terminology section is meant only as a summary. The GSP says:

"The following terminology is used in this document:  Graph Store - A mutable repository of RDF graphs managed by one or more services [SPARQL-UPDATE]."

From this, I would expect that when I click on "[SPARQL-UPDATE]", I would find this same definition. But while that sentence is essentially in the Update definition, the important bits from the Update document aren't included:

Similar to an RDF Dataset operated on by theSPARQL 1.1 Query Language, a Graph Store contains one (unnamed) slot holding a default graph and zero or more named slots holdingnamed graphs.

I don't think it's made clear that the reader should be clicking through to the Update doc for the definition before continuing reading of the GSP document.

>> Second, I didn't read this as "explicit text".
> I don't think text any more 'explicit' than this is necessary or
> constructive for what this document is specifying - for the reasons I
> have stated above.
>> Certainly not in the way Update deals with defining a graph store ("a Graph Store contains one (unnamed) slot holding a default graph and zero or more named slots holdingnamed graphs").
> That is my point exactly.  *That* specification deals with the
> definition of what a default graph is, *this* document builds on that
> definition and provides a mechanism for interacting with them.

That's fine, but why not say that, then? "This document relies on the definition of a Graph Store in [SPARQL-UPDATE]."

>>> Were you looking for something else beyond this?
>> Yes. I was expecting/hoping for a an actual definition of what the graphstore is.
> This is in the Update specification (which this document has a
> normative dependency on).

I think this this dependency should be made more explicit. I don't want to have to read through entire document all the way to the references section before finding out that the Update document is a normative reference. As it is, there are only two mentions of SPARQL Update that occur before the terminology section's "definition":

Abstract: "This interface is an alternative to the SPARQL 1.1 Update protocol."

Introduction: "Where the meaning of the operations are described, a SPARQL Update equivalent syntax is shown for clarity."

Neither of these give me the impression that I have to be familiar with the Update spec and its definition of a graph store in order to understand the GSP. For me, I think the opposite is true: I read these mentions as suggesting that I can read and fully understand the GSP document independently of the Update document.

>> If the GSP is going to rely on the definition in Update, I think it should have some text to that affect, not just a citation link next to the (different) definition.
> I'm not sure what you mean by 'different' definition.

As I explain above, it looks like a different definition because the GSP terminology section says a graph store is "a mutable repository of RDF graphs managed by one or more services," but this doesn't actually include the important specifics from the Update doc that a graph store "contains one (unnamed) slot holding a default graph and zero or more named slots holding named graphs."

>  Are you saying
> the summary of the  external definition (clearly indicated with a
> normative reference) is a poor summary?
> If so, can you suggest text
> for a better summary?

I'd suggest either including the text from Update about a graph store including one unnamed and any number of named graphs, or adding text that explicitly states that the GSP uses the definition of "Graph Store" from the Update document.

Received on Friday, 17 February 2012 18:11:52 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:01:06 UTC