Re: [Json] Possible next work for the WG

[Jacob Davies <[e-mail]>]

On Wed, Oct 16, 2013 at 9:54 AM, John Cowan <[e-mail]> wrote:
>> - Canonicalization rules so that two JSON texts can be compared for
equality
>
> I think we should explore this, but it's going to be tricky.

Very tricky if we stick to the idea that JSON is only a text format and
does not describe an abstract data model. Who is to say whether 4.5 and
4.500 were supposed to be equivalent? 0 or -0? Or whether the ordering of
keys in an object matters? (It matters to me that the order is preserved
but not that it is included in equality comparisons, but that's just me.)

> As reinvented wheels are often hexagonal, I think if we are going to
> do this we should adopt draft-zyp-json-schema-04 as a base document
> for further work.  This has now expired, but is still available at
> <
https://raw.github.com/fge/json-schema-validator/master/misc/specs/draft-zyp-json-schema-04.txt
>.

How to phrase this: I would be curious how much actual use is made of JSON
Schema or any other formal schema system for JSON. If there are people here
who've found them useful, I think it would be interesting to discuss what
kinds of uses they made of them.

I don't know the history of JSON Schema but the specific link you point to
is pretty obscure (i.e. provide no actual starting point for using it). An
older draft seems to be a bit clearer:
http://tools.ietf.org/html/draft-zyp-json-schema-03

I find the syntax overly verbose and not very idiomatic - e.g. requiring:

   "name":{"type":"string"},

instead of just "name" : "string", and requiring an intermediate
"properties" element when defining objects, but those could be fixed.