Re: Fw: RDF/JSON

Sandro Hawke <sandro@w3.org> wrote on 04/26/2013 08:17:34 PM:

> On the substance of the comment -- I find it rather compelling and 
> am not sure how to proceed.   The bit about "Databases like to query
> on predicates" was particularly interesting....   An odd contrast to
> direct access from JS.
> 
> Arnaud/Martin, how ideally would you like to proceed on this?   Just
> make RDF/JSON a note?   Or something else?    Should the two be 
> aligned in those "details" where they differ?  On the point about 
> documentation, would it help if there were a simple, mostly self-
> contained section of the JSON-LD spec you could point to?
> 

Hi Sandro,
as you know I'm not thrilled about the proliferation of serialization 
formats for RDF so I can certainly sympathize with Markus's reluctance to 
promote RDF/JSON at the same time as JSON-LD. This being said, as Martin 
explained we have a need JSON-LD does not meet. To be able to use RDF/JSON 
in our products we'd like to have at least a stable reference. Of course, 
a Rec would be best but given the time constraints the WG is under we will 
settle on simply having it as a Note.

It is tempting to want to merge the two into one spec but beside the 
impact this would have on the JSON-LD timeline we think this isn't a 
desirable option. JSON-LD already has two forms and adding a third one 
would lead to a spec that is even more complicated and nobody is happy 
with.

Best regards.
--
Arnaud  Le Hors - Software Standards Architect - IBM Software Group