Re: [apps-discuss] I-D Action: draft-ietf-appsawg-json-patch-01.txt

["Paul C. Bryan" <[e-mail]>]

On Mon, 2012-03-12 at 14:55 -0700, Mike Acar wrote:

> This version of the draft doesn't address the bad interaction with the
> JSON Pointer spec, namely that "add" takes a JSON Pointer to a value
> which does not yet exist, and the Pointer spec says that in that case,
> an implementation MAY abort with an error. Is that going to be
> addressed in -02?


I'm of the opinion that expressing a not-yet-existing target as a JSON
Pointer is the most simple and intuitive method so far. I've had no
problems implementing the specs as a result of this. I agree that the
language in the JSON Pointer spec is still somewhat awkward and could
still stand to be improved.

The suggested alternative of requiring the parent and target member to
be expressed separately is more verbose and complex. Over-verbosity of
JSON Patch is already an objection that has been raised more than once
on this list, and I'm not inclined to exacerbate it further unless there
is consensus from the members of appsawg that it should be adopted.

Paul