Re: [Json-canon] [EXT] Re: Support for a WG

Thanks for the followup Allan.

In my opinion the application layer schemas for defining what the JSON can
look like (optional values different values meaning the same thing etc),
should be verified by application layer. I would like to keep the
canonicalization (and potential signatures) at the level where they say the
JSON has not changed.

When it comes to empty empty values I think those are values too, they
might convey some meaning to the application layer.and with empty I mean
{"empty": ""}. And since I do not want the canonicalization to be involved
with application level details optional attributes are not a concept.

Does that make sense?

Cheers
//Samuel

On Tue, Jan 22, 2019 at 5:47 PM Allan Thomson <
athomson@lookingglasscyber.com> wrote:

> I’m not sure how you can create a consistent hash value without
> considering how empty/optional fields in a JSON structure are treated
> consistently across implementations.
>
>
>
> So if the same optional/empty fields exist from 2 different
> vendors/instances then do they end up with the same hash?
>
>
>
> Allan
>
>
>
> *From: *Samuel Erdtman <samuel@erdtman.se>
> *Date: *Tuesday, January 22, 2019 at 8:37 AM
> *To: *Allan Thomson <athomson@lookingglasscyber.com>
> *Cc: *"Struse, Richard J." <rjs@mitre.org>, Carsten Bormann <cabo@tzi.org>,
> "json-canon@ietf.org" <json-canon@ietf.org>
> *Subject: *Re: [Json-canon] [EXT] Re: Support for a WG
>
>
>
> I do not think we should go into schemas, that is one of the things that
> makes XMLDigSig inconvenient. But instead keep to simple JSON stringifying
> with predictable result.
>
> On Tue, 22 Jan 2019 at 16:19, Allan Thomson <
> athomson@lookingglasscyber.com> wrote:
>
> Particularly important in what Rich says is also is to ensure the same
> hash is created where those JSON objects have optional or empty/missing
> properties that are defined in the schema of the object.
>
> Regards
>
> Allan
>
> On 1/22/19, 4:31 AM, "json-canon on behalf of Struse, Richard J." <
> json-canon-bounces@ietf.org on behalf of rjs@mitre.org> wrote:
>
>     The use case is to enable other standards that use a JSON
> serialization to be able to count on two objects, each with the same
> contents, having the same hash value based on their representation in JSON.
>
>     Does that help?
>
>     On 1/18/19, 1:47 AM, "Carsten Bormann" <cabo@tzi.org> wrote:
>
>         On Jan 18, 2019, at 03:50, Struse, Richard J. <rjs@mitre.org>
> wrote:
>         >
>         > A standardized deterministic canonicalization of JSON data
> streams is essential
>
>         For what?
>
>         I think we would all benefit if you could explain your use case.
>
>         Grüße, Carsten
>
>
>
>     --
>     json-canon mailing list
>     json-canon@ietf.org
>     https://www.ietf.org/mailman/listinfo/json-canon
>
>
> --
> json-canon mailing list
> json-canon@ietf.org
> https://www.ietf.org/mailman/listinfo/json-canon
>
>