Re: [dispatch] PING2. Re: JSON Canonicalization Standard

[Anders Rundgren <[e-mail]>]

On 2018-09-06 01:07, Larry Masinter wrote:
> A canonicalization of a space should apply to all valid values, and
> shouldn't map two semantically different values to the same canonical form.

I don't understand what you are referring to here since white space is
insignificant in JSON.


> JSON intentionally allows large integers and floating point numbers outside
> of IEEE range of capabilities (1e400, 99999999999999999999).
>   
> While some implementations will coerce values and round off, I think a good
> canonicalization should leave these values intact, if you're want a JSON
> canonicalization rather than an "ES6 value serialization" canonicalization
> (a strict subset).

The rationale for the IEEE-754 double precision/ES6 restriction used in the I-D
is that this is (AFAIK) the de-facto scheme for current IETF standards defining
JSON objects including such that use big numbers like RSA parameters.

Another reason for taking this route is that using the full JSON Number
notation would require the canonicalizer to work on the text level only
since it is incompatible with most current JSON Parsers (all that do not
use an unlimited BigNumber as underlying JSON Number type).

The JCS scheme is designed to ultimately be integrated in JSON serializers
(only) as an output option.


Although it is considerably easier defining a free-standing canonicalizer
scheme like https://gibson042.github.io/canonicaljson-spec/ it might end-up
getting limited industry support.  The JOSE folks slashed an earlier effort of
mine requiring requiring JSON text preservation. In particular, dealing with
active intermediaries was considered a showstopper.

Anders
I-D: https://tools.ietf.org/html/draft-rundgren-json-canonicalization-scheme-01
JWS+JCS combo: https://github.com/cyberphone/jws-jcs#combining-detached-jws-with-jcs-json-canonicalization-scheme
On-line testing: https://mobilepki.org/jws-jcs/home
> 
> _______________________________________________
> dispatch mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/dispatch
>