Re: JSON-LD terminology

On Aug 29, 2012, at 7:05 AM, Andy Seaborne <andy.seaborne@epimorphics.com> wrote:

> The proposal for a reworded Linked Data definition is better.
> 
>> Some of the data model differences require further discussion and
>> need to be publicly aired, as they impinge on long-held resolutions
>> in JSON-LD.
> 
> Good to hear -
> 
> 
> One specific point:
> 
>> [[ 1. Linked Data is a set of documents, each containing a
>> representation
> of a linked data graph.
> ...
>> 8. IRIs used within a linked data graph SHOULD be dereferenceable to
>> a Linked Data document describing the resource denoted by that IRI.
>> ]]]
> 
> 
> Test case: is foaf:name <http://xmlns.com/foaf/0.1/name> an IRI
> dereferenceable to a Linked Data document?

I would say yes. However, not all vocabularies will conform with this; hopefully, that will change over time. IMO, resources should use content-negotiation to return formats appropriate for the consumer. An HTML representation should be marked up with RDFa to describe its contents. An application using JSON-LD would likely generate an HTTP Accept header that would include application/ld+json to get back a JSON-LD representation if available, otherwise RDF/XML, Turtle or RDFa. This is similar to the constraints we had over RDFa profiles, before they were removed.

> Linked Data is being defined here is mostly silent on format limitations
> but by context the text leads towards "no" because the context is the
> JSON-LD document.  But it is used in JSON-LD examples  :-)

I would interpret silence as not making any normative statements about data formats. In this context, a Liked Data document is a document which adheres to the principles set forth for LInked Data, which is an abstract concept, not a concrete syntax requirement. This is also probably a good time to make the claim that the precepts of Linked Data are themselves based on (or are a superset of) RDF Concepts, which would then say that any RDF document is also a LInked Data document.

Gregg

> 	Andy
>

Received on Wednesday, 29 August 2012 17:12:00 UTC