Re: [Json] Possible next work for the WG

On Tue, Oct 15, 2013 at 7:51 PM, Paul Hoffman <paul.hoffman@vpnc.org> wrote:

> - I-JSON (a profile of JSON with some interoperability issues nailed down)
>

This, frankly, is the only real reason I’ve invested work in this group and
agreed to edit the -bis.  I want something that people who really rely on
getting clean JSON can point to and say “do this” and not have to write any
extra rules.

- Best practices document for JSON implementers and folks who use JSON in
> protocols
>

Nothing wrong with the idea, but I think that if we do the -bis and then
I-JSON, it’d be superfluous.

> - Canonicalization rules so that two JSON texts can be compared for
> equality
>

History would seem to show that this requires substantial effort and has
very modest pay-off. Also, there are lots of success stories around
digitally-signed JSON already in deployment, which seems to weaken one of
the use-cases for canonicalization.  I’d argue for passing on it.

> - Requirements for JSON schema (which will be needed before the WG
> considers working on an actual schema)
>

I would argue against this because designing schema languages is a crushing
amount of work and easy to get wrong.  Also, in the domain I care about
(RESTful protocol payloads) the benefit seems small.  There is a certain
amount of code you could remove, I suppose, of the form

if json['person']['id-num'] == nil
  // reject transaction with an error message about how the id-num is
required
end

but pushing the Sysiphean schema-language-design rock up the hill feels
like hitting a beetle with a boulder.

>
> The chairs request general comments on these or additional topics for the
> WG. If there is interest, we will re-open the topics on separate threads,
> and then put together a proposal for rechartering. If people really want
> this WG to shut down after 4627bis is done without doing additional work,
> that is a valid view as well.
>
> Note that we have also heard requests for non-compatible additions to the
> JSON syntax, particularly for comments and byte-oriented data. If the WG
> wanted to adopt this work, we would probably need to talk with our ADs and
> the IAB because that might be better handled by Ecma in a possible revsion
> to ECMA-404.
>
> --Matt Miller and Paul Hoffman
> _______________________________________________
> json mailing list
> json@ietf.org
> https://www.ietf.org/mailman/listinfo/json
>