Re: Updated JSON-LD spec to more closely align w/ RDF data model

On 07/04/2013 05:02 AM, Markus Lanthaler wrote:
> On Thursday, July 04, 2013 1:34 PM, Peter F. Patel-Schneider wrote:
>> Appendix A defines nothing in terms of anything related to RDF.  Its
>> "normative" summary defines default graph, named graph, graph, edge,
>> properties, nodes, IRI, blank node, blank node identifier, JSON-LD
>> value, typed value, language-tagged string, and list and none of
>> these definitions mention anything about RDF.
> They are definitions (bold & italic) because we use those terms throughout the document and that's the way how you reference them with ReSpec.

Huh?  What in ReSpec prevents pointing at definitions in different documents?  
It appears to me that class="externalDFN" is part of ReSpec precisely to 
permit pointing at external definitions.

> Would you like to see a "as defined in RDF Concepts" everywhere?

That would be somewhat better.  Even better would be to say something like 
"datasets, graphs, triples, nodes, blank nodes, ... are as defined in RDF 
Concepts"  (with appropriate pointers, of course).   Even more better would be 
to use class="externalDFN" and point directly to the base definitions in the 
body of the document.
>
>
>> It is totally unnecessary to have all these redefinitions, even when
>> there were real differences between the JSON-LD data model and
>> definitions from RDF Concepts.
> For someone with an RDF background that's true. For everyone else it isn't. Sure, we could just require that everyone reads and understands RDF Concepts first but we wanted to avoid that.

And here we are back at the beginning of this discussion.

I have never advocated that "everyone reads and understands RDF Concepts 
first".   I don't see where you could ever have gotten the idea that I have 
ever even hinted that this might be the case. However, I strongly hold the 
position that anyone who delves deeply into Linked Data must, of necessity, 
understand the notions underlying Linked Data as it currently exists, and 
these notions include triples, graphs, blank nodes, etc., as they are defined 
in RDF.  Basing JSON-LD on different definitions does both Linked Data and 
JSON-LD a disservice.
>
> Would including a non-normative summary of the data model in the same style be OK? I'm fine with almost everything that isn't just a diff to RDF's data model.
And I'm fine with a diff to RDF's data model.  This is, after all, the 
definition of JSON-LD.  It needs to be firmly grounded on the RDF data model, 
not just because this is a product of the RDF working group, not just because 
this is a product of W3C, but because for JSON-LD to be truly about linked 
data as it currently exists, JSON-LD needs to both be and be seen to be based 
on linked data as it currently exists, and that means being based on the RDF 
data model.
>
>
>> Consider, for example, the way that generalized RDF Datasets are now
>> defined
>> in RDF Concepts.   The JSON-LD data model could have been defined in an
>> analogous manner all along, building on previous definitions instead of
>> redoing them.
> If all the definitions such a definition relies on are in the same document it is indeed trivial.

Huh?  What does being in a different document have to do with anything, 
particularly when the other document is a product of the same working group?   
Is there some fundamental reason that pointing to definitions in different 
documents is a bad idea?

>
>
> --
> Markus Lanthaler
> @markuslanthaler
>

peter

Received on Thursday, 4 July 2013 13:28:39 UTC