Re: AW: [wot-ig/wg] minutes - 26 April 2017

On 5/2/17 3:00 PM, Kovatsch, Matthias wrote:
> Hi Peter
> 
> How does your work relate to
> https://tools.ietf.org/html/draft-ietf-ace-cbor-web-token-04? (Since you
> mentioned JWT is your main use case.)

Well, opinions differ about COSE. I'll leave it at that for now.

> For the more general case, I am a bit confused about giving JSCN the
> status of a new format. Isn't it more like best practices how to use
> CBOR correctly?

Your intuitions are correct: JSCN does consist of best practices for
using CBOR in certain contexts, but it is also a format in the sense
that lossless round tripping requires rules that need to be followed for
the sake of interoperability.

Note that JSCN is still young - it's being proved out in code, and the
specification needs more work. Although I'm happy to discuss it here,
Jeremie (who's not on this list AFAIK) is the primary author and would
no doubt welcome contributions at https://github.com/quartzjer/JSCN so
we can get it to a more stable state.

Thanks for listening. :-)

Peter

> 
> Kind regards
> Matthias
> 
> 
> Sent from my phone, limitations might apply.
> 
> -----Original Message-----
> *From:* Peter Saint-Andre - Filament [peter@filament.com]
> *Received:* Tuesday, 02 May 2017, 22:26
> *To:* Peintner, Daniel (ext) (CT RDA NEC EMB-DE)
> [daniel.peintner.ext@siemens.com]; Kazuyuki Ashimura [ashimura@w3.org];
> Public Web of Things IG [public-wot-ig@w3.org]; public-wot-wg@w3.org
> [public-wot-wg@w3.org]
> *Subject:* Re: AW: [wot-ig/wg] minutes - 26 April 2017
> 
> Yes, there are many factors and multiple tradeoffs. A major focus for
> JSCN is the ability to properly handle JWTs in a lossless way. For
> non-security use cases, CBOR itself is close to ideal.
> 
> Peter
> 
> On 5/2/17 6:29 AM, Peintner, Daniel wrote:
>> Hi Peter,
>> 
>> Thank for your pointer.
>> 
>> JSCN (ore respectively CBOR) is definitely one candidate.
>> 
>> Having said that, there are other formats we might want to look at
>> (Smile, EXI4JSON, ...).
>> 
>> I looked at results/examples referenced in JSCN [1] which show JSON (318
>> bytes) to JSCN (187 bytes). I checked EXI4JSON which gets down to 139
>> bytes (see demo at [2]).
>> 
>> So size is one aspect but there are many more aspects we should take
>> into account.
>> 
>> -- Daniel
>> 
>> [1]
>> https://quartzjer.github.io/JSCN/draft-miller-json-constrained-notation-00.html#rfc.section.6
>> [2] http://exificient.github.io/javascript/demo/processJSON.html
>> 
>>  
>> 
>>  
>> ------------------------------------------------------------------------
>> *Von:* Peter Saint-Andre - Filament [peter@filament.com]
>> *Gesendet:* Donnerstag, 27. April 2017 23:16
>> *An:* Kazuyuki Ashimura; Public Web of Things IG; public-wot-wg@w3.org
>> *Betreff:* Re: [wot-ig/wg] minutes - 26 April 2017
>> 
>> On 4/27/17 1:40 AM, Kazuyuki Ashimura wrote:
>> 
>> <snip/>
>> 
>>>    should start activity to look at concise descriptions for TD
>> 
>> Regarding concise descriptions, you might want to look at some work my
>> colleague Jeremie Miller is doing on JSON Constrained Notation:
>> 
>> https://github.com/quartzjer/JSCN
>> 
>> We're intending to begin standardization of this soon.
>> 
>> Peter
>>

Received on Tuesday, 2 May 2017 22:09:14 UTC