Re: [Json] json-pointer as URI fragment syntax for application/json

[R S <[e-mail]>]

On Mon, Sep 23, 2013 at 9:08 PM, Paul C. Bryan <[e-mail]> wrote:

> **
> The Application Area Working group raised and addressed the question of
> ambiguity during review of the Internet-Drafts. It concluded, by consensus,
> that it was not a substantive issue, despite your protestations.
>

I agree that the WG agreed to use it.

It did not get IETF consensus as a general-purpose JSON fragment
identifier, and the RFC contains text saying exactly that: "the fragment
identifier syntax for application/json is not JSON Pointer."

Feel free to argue the actual technical point, btw.

- Rob





> On Mon, 2013-09-23 at 20:57 -0700, R S wrote:
>
> On Mon, Sep 23, 2013 at 8:49 PM, Paul C. Bryan <[e-mail]> wrote:
>
>  Kris Zyp was one of the two leads in the JSON Schema working group, and
> whose contributions are attributable. He is co-author of RFC 6901.
>
>
>
>  Sure, sure.
>
>
>
>  What about the other point?
>
>
>
>  - Rob
>
>
>
>
>
>
>
>
>
> On Mon, 2013-09-23 at 20:49 -0700, R S wrote:
>
> On Mon, Sep 23, 2013 at 8:42 PM, Paul C. Bryan <[e-mail]> wrote:
>
> It was designed by the JSON Schema working group.
>
>
>
>
>
> Oh, I see. That credit doesn't appear in RFC 6901.
>
>
> Still seems buggy as a general-purpose fragment identifier, since it can't
> distinguish between arrays and objects.
>
>
> - Rob
>
>
>
>
>
>
> On Mon, 2013-09-23 at 20:41 -0700, R S wrote:
>
> Who designed it?
>
> Anyway, it is buggy as a general purpose fragment identifier.
>
> - Rob
>
> On Sep 23, 2013 8:34 PM, "Paul C. Bryan" <[e-mail]> wrote:
>
> JSON Pointer was designed for multiple purposes, including as a fragment
> identifier in URIs.
>
> On Mon, 2013-09-23 at 09:11 -0700, R S wrote:
>
> This is a bad idea. JSON pointer was designed for a narrow patch use case,
> and can't disambiguate objects and arrays.
>
> The JSON RFC shouldn't list every RFC that uses JSON.
>
> - Rob
>
> <no hat>
>
> On Sep 20, 2013, at 10:00 PM, "Manger, James H" <
> [e-mail]> wrote:
>
> > RFC 4627 doesn't define how URI fragments work for application/json
> content.
> >
> > It would be helpful for 4627bis to say that a fragment (that begins with
> a "/" or is empty) is a JSON Pointer [RFC 6901]. A JSON Pointer identifies
> a specific value in a JSON text.
>
> Changing the IANA MIME registry entry can be done, but I propose that we
> don't do it as part of the -bis because it is not really a clarification.
> This update can be done with a separate draft that updates the registry,
> but I admit that people tend to just read the RFCs, not the registry and
> therefore this is likely to be missed.
>
> How do others feel about making a change in the -bis that is "things that
> have happened since 4627" that don't change the JSON grammar?
>
> --Paul Hoffman
> _______________________________________________
> json mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/json
>
>
> _______________________________________________
> json mailing [e-mail]://www.ietf.org/mailman/listinfo/json
>
>
>
>
>
>
>
>
>
>
>
>  _______________________________________________
> json mailing [e-mail]://www.ietf.org/mailman/listinfo/json
>
>
>