Re: [Json] On flat vs nested JSON encoding style

[Anders Rundgren <[e-mail]>]

On 2016-02-05 02:09, Phillip Hallam-Baker wrote:
> On Thu, Feb 4, 2016 at 7:17 PM, John Cowan <[e-mail]> wrote:
>> Anders Rundgren scripsit:
>>
>>> This ecosystem is rapidly diminishing since ECMAScript nowadays mandates
>>> strict insertion/declaration order for serialization.
>>
>> That affects only ECMAscript implementations, though, so there are no such
>> guarantees on the server side, or server-to-server, unless Node.js or
>> something similar is being used.
>
> +1
>
> We are doing RFC7159 JSON. Order is not guaranteed. A sender can emit
> the request in any order they like. Code that depends on the order of
> elements within an object is broken.

Well, the W3C Web Payment IG/WG folks who claim they are working with
a new (and presumable secure) system for the Web Payments have not yet
begun dealing with how they are going to sign their JSON messages.

This will be a litmus test, not for RFC7159, but for RFC7515:
https://github.com/golang/go/issues/14135#issuecomment-177265555

In the meantime, I'm helping JSON tool-vendors providing support for
"Predictable Serialization" a la ES6/V8.  With the sole exception of
"Number", this is not rocket science and that particular rocket has already
been built by Google and the blueprints are readily available as well.

Maybe we can come back to this issue in 18 months and see where it went?

Anders