Re: JSON-LD terminology

On 8/31/12 7:16 AM, Andy Seaborne wrote:
> On 31/08/12 12:05, Kingsley Idehen wrote:
>> On 8/31/12 5:20 AM, Andy Seaborne wrote:
>>> On 31/08/12 10:02, Markus Lanthaler wrote:
>>>> Would you, and the other members of the RDF WG, be fine with just
>>>> dropping
>>>> the sentence "In particular, any document based on an RDF 
>>>> serialization
>>>> format is a Linked Data document"?
>>> Not really.  I think it's an important, including for JSON-LD which
>>> has the "toRDF" function.  RDF is the data model.
>>> The statement does not say that only RDF is linked data.
>> Andy,
>> RDF and Linked Data are not the same thing.
> The statement
> "In particular, any document based on an RDF serialization
> format is a Linked Data document"
> only claims the one way relationship
>   set of RDF serializations subset of Linked Data documents
> and not about whether a linked data document is RDF.

It claims any RDF document is a Linked Data document. That simply isn't 
accurate. The RDF spec makes no claims about IRIs being:

1. de-referencable
2. http: scheme based.

It simply requires IRIs to be used as denotation mechanism. That's it.

> No owl:sameAs is implied.
> In the context of JSON-LD, which defines a function to go from 
> application/ld+json to RDF, it makes sense to point out that it is 
> part of that system.

Yes, but it can be expressed in a manner that doesn't conflate RDF and 
Linked Data. All conflation is ultimately detrimental when the goal is 
clarity via consistency.

>> There isn't an owl:sameAs
>> relationship between these two entities. To push this overtly or
>> covertly serves no beneficial purpose. We don't need confusion obscuring
>> the virtues of either.
>> In the context of Linked Data It should be made very clear that RDF is:
>> 1. an option
>> 2. an implementation detail.
> I think the proposed text is not at odds with (1).

It shouldn't be at odds with the items 1 and 2 combined. I also don't 
see how the current test (in anyway) infers  that RDF is an option, 
especially to readers who simply skim through specs.

We have two distinct issues here:

1. RDF as the standard basis (or foundation) for Linked Data 
implementation as proposed by the W3C.
2. RDF as the only option (via concept conflation) for Linked Data 
implementation as proposed by the W3C.

Item 1 is what needs to be conveyed. It protects against:

1. Political FUD from R-D-F reflux types
2. W3C mandate (as opposed to recommendation) perception.

Every RDF document isn't a Linked Data document.
>     Andy



Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web:
Personal Weblog:
Twitter/ handle: @kidehen
Google+ Profile:
LinkedIn Profile:

Received on Friday, 31 August 2012 11:29:44 UTC