Re: [Json] Allow any JSON value at the top level

[Paul Hoffman <[e-mail]>]

On Jun 10, 2013, at 7:38 PM, R S <[e-mail]> wrote:

> On Mon, Jun 10, 2013 at 7:04 PM, Paul Hoffman <[e-mail]> wrote:
> 
> Good catch. When I said "as-is", I meant "don't change the current text". We need to add a point to the differences section for 4627-ECMAScript, and would need to duplicate that in the section for the current-doc-ECMAScript section.
> 
> Wouldn't that be a section that amounts to "here's what you need to do to interoperate with browsers and most other popular implementations"?

Oh gopod no. I certainly hope the WG find that defining "browser" and "most" and "popular" is out of scope.

> Many endpoints only accept specific root elements. For example, json-patch only accepts arrays. In other words, no one will notice if your JSON parser only accepts objects and arrays. 
> 
> What is the benefit of preserving RFC 4627 requirements? I don't understand the rationale accompanying your opinion. The charter mandates that the WG document top-level primitive parsing, no matter what.

Fully agree on the last part. But "document" does not mean "change the requirements". We can say "the requirement is X" and "X differs from 4627 and/or ECMAScript", at least in my reading of the charter.

--Paul Hoffman