Re: [Json] Topic #4: Software Behavior

My mental model is that some working group, for example over in JOSE, says
“messages MUST be I-JSON” and then people who are implementing the code
that receives such messages tells the JSON library to go into i-json mode.

On Mon, Apr 28, 2014 at 1:52 PM, Jacob Davies <jacob@well.com> wrote:

> It seems to me that all the requirements of the I-JSON profile (which
> I think is a good idea) are to be satisfied at the emitting end. The
> reading end can be I-JSON compliant, but the whole point is that any
> old JSON parser is very likely to be able to read I-JSON, isn't it?
>
> With no separate media type and no in-band marker, how can an parser
> possibly know it is looking at an I-JSON document? Therefore how can
> the compulsory error-handling behavior be specified? Or is the idea
> that an I-JSON parser is this strict about *all* JSON documents? In
> which case, this is essentially an effort to get the common
> understanding of "normal JSON" to be identical with I-JSON since
> I-JSON parsers will reject some legal JSON documents (which is true of
> any parser of course). Again I think that's probably a good idea as
> far as numbers & duplicate keys & illegal escapes are concerned.
>
> _______________________________________________
> json mailing list
> json@ietf.org
> https://www.ietf.org/mailman/listinfo/json
>