- From: Andreas Kuckartz <a.kuckartz@ping.de>
- Date: 10 Jun 2014 10:04:02 +0200
- To: "Harry Halpin" <hhalpin@w3.org>
- Cc: "public-fedsocweb@w3.org" <public-fedsocweb@w3.org>, "internal-socbizcg@w3.org" <internal-socbizcg@w3.org>
Harry Halpin: > We are sticking to JSON (and JSON-LD) for interop requirements, > although people are free to additionally serialize however they want > (RDF/XML for those who want it, or Turtle, or even custom > bytestrings). > > In detail, people can serialize however they want, put forcing specs > to test on a myriad of often-not-so-well developed or adopted data > exchanges would make the CR process of creating a unified test-suite > for the Social Web rather cumbersome. Since most of the Web uses JSON > we'll use that for REC track, we're sticking to it, but folks can > publish alternatives as Working Group Notes, Community Group reports, > or IETF Internet Drafts. That mostly seems to be answering different questions. Maybe my subject was misleading? I was only suggesting a specific requirement for the specification ("that JSON-LD is used in way which makes it an RDF-serialization of the data") - not to replace JSON-LD or add another serialization. Those interested in more technical details are referred to this section of the JSON-LD specification: http://www.w3.org/TR/json-ld/#relationship-to-rdf But the reply from James Snell indicates that such a requirement would be in scope, and I am very interested to see what he has suggested. >> (I would also like to see informal Turtle examples in the future >> specification documents, not only JSON-LD ;-) Please note the "informal". That is a practice which has already been used in W3C Recommendations multiple times. Examples generally are declared "non-normative" anyway. The JSON-LD specification, for example, does this: http://www.w3.org/TR/json-ld/#conformance Cheers, Andreas
Received on Tuesday, 10 June 2014 08:04:45 UTC