Re: Requirements update

On Aug 9, 2011, at 10:51 AM, Dan Brickley wrote:

> On 9 August 2011 18:24, Gregg Kellogg <gregg@kellogg-assoc.com> wrote:
>> We discussed the requirements [1] on the call today. I made a couple of
>> small changes, but I mainly wanted to get feedback from the list on the
>> specific JSON-LD markup requirements:
>> 
>> A JSON-LD document must be able to express a linked data graph.
> 
> Good stuff. Nitpic time.
> 
> Q: What is the definition of a "linked data graph"?

You may have just been looking at the points in the email, which are taken from the requirements document which does define a "linked data graph".

> e.g. If my datamodel consists of ordered n-tuples, such that:
> 
> a) these 4-tuples:
> 
> *) uri-1 author uri-2 "John Smith"
> *) uri-1 author uri-3 "Alice Example"
> 
> vs
> 
> b)
> 
> *) uri-1 author uri-3 "Alice Example"
> *) uri-1 author uri-2 "John Smith"
> 
> ...capture a meaningful difference, does the requirement for a "linked
> data graph" include respecting such aspects of my data model? Or the
> fact I'm using 4-tuple representations rather than 3-tuple?

Without some concept of quoted graphs, or a similar context model, we don't have a way to express 4-tuples.

Regarding ordering, point 15) indicates that array order is not significant, and that the data model is that of an unordered graph. Point 16), however, says that there SHOULD be a way to express such order. In fact, the latest version of the spec defines {"@list"", ["John Smith", "Alice Example"]} to have a defined order, and the RDF conversion rules using rdf:List semantics to do this, but this is not required in a pure-JSON implementation.

> The draft reads like an attempt to describe RDF's abstract data model
> without naming it, or perhaps, without hooking it into associated
> perceived complexities. That's fine, but if so the doc needs to say "a
> linked data graph is ..." ... and then complete the definition either
> inline or by reference.
> 
> The version I see in http://json-ld.org/requirements/latest/ does have
> such language, "A linked data graph is a labeled directed graph, where
> nodes are subjects or objects, and edges are properties."; is the WWW
> version older or newer than the copy mailed here?

The versions should be in sync, as I committed and pushed that change before I sent the email. Changes to the GitHub version are automatically synced to the site.

> cheers,
> 
> Dan

Thanks for the feedback, and please offer any other suggestions you might think are useful. As you note, we are running parallel with RDF, but not explicitly relying on it in order to be relevant to the JSON and non-RDF Linked Data communities, but we should not be inconsistent with the goals of RDF, and there should be a way to get from JSON-LD to RDF and back; which is what I implement in my Ruby version.

Gregg

Received on Tuesday, 9 August 2011 19:19:44 UTC