Re: [Json] I-JSON Tpic #2: Top-Level

[Nico Williams <[e-mail]>]

On Wed, Apr 30, 2014 at 6:42 PM, Paul Hoffman <[e-mail]> wrote:
> On Apr 29, 2014, at 8:40 PM, Tim Bray <[e-mail]> wrote:
>
>> I’m sorry, but the central idea of I-JSON is to explicitly rule out all the interoperability problems identified in RFC7159: How to produce maximally-interoperable JSON.
>
> This is a good point, and one that I forgot when I said "this isn't needed". There are probably some parsers out there that expect the original JSON restriction; given that, repeating the restriction here seems fine.

Conceding that for the time being, the restriction to repeat was that
JSON texts must be objects or arrays at the top-level, NOT just
objects.

Now, as to whether we should repeat that restriction, given that every
parser I know of at least has an option to permit non-object/array
types at the top-level, I find the interop argument wanting.  If I
know some application uses I-JSON, and if I know I-JSON permits any
type at the top-level (supposing it were to), why would I have any
trouble implementing that with existing parsers?

I haven't carried out an exhausting analysis of existing parsers, and
since I can't -no one can-, I might concede that point even if no one
confirms the existence of any JSON parsers that don't have an option
to permit any type at the top-level.  For now I'm just not buying the
interop argument.

Nico
--