Hi Tom,

> On 21 Jul 2017, at 18:34, tom p. <[email protected]> wrote:
> 
> Leif
> 
> K/V
> 
> Not something that I would expect to implement but rather my views are
> based on tracking the work of NETCONF and RESTCONF, the latter being
> JSON.  I see the difficulties encounterd in the definition of RESTCONF
> and am very clear that JSON as a language is incomplete and somewhat
> loosely defined, sloppy even.

Without taking a position on K/V versa JSON: I thought 
draft-ietf-jsonbis-rfc7159bis-04 was fairly complete and can be implemented 
without relying on any other documents. Yes, there are some interesting corner 
cases, but I doubt they will affect MTA STS use.

>  I do not see the availability of
> libraries as affecting this - if the underlying technology is incomplete
> and loosely defined, then what a library that implements just that
> technology has to offer will perforce be incomplete and loosely
> defiined.
> 
> K/V I find precise and unambiguous.
> 
> Tom Petch
> 
> 
> ----- Original Message -----
> From: "Leif Johansson" <[email protected]>
> To: <[email protected]>
> Sent: Friday, July 21, 2017 12:10 PM
>> 
>> We had a (less than spirited) discussion in Praha about the K/V vs
> JSON.
>> 
>> There was no clear consensus for switching from JSON to K/V based on
>> the proposal from the spec authors sent to the list last week.
>> 
>> Victor expressed his views clearly and I'll not ask him to repeat
>> himself. I will attempt to summarize the discussion - feel free
>> to correct me!
>> 
>> The situation is that of the active participants that are also
>> implementors one (postfix - Victor) has said that he does not believe
>> that postfix will ship with a standard JSON-based implementation but
>> that this doesn't preclude third-party plugins.
>> 
>> The other major implementor who is also an active participant (google)
>> has stated that they have a JSON implementation but that they are
>> also eager to close this issue. One more implementor (Chris) has
> stated
>> no preference but also said that his (native) JSON parser is somewhat
>> smaller than his K/V one.
>> 
>> Based on this and falling back to the IETF WG process as chair I would
>> conclude that there isn't enough consensus to make a change and that
>> we are staying with JSON.
>> 
>> I would prefer to have more opinions (implementors in particular) but
>> we are also going to get past this. In particular I would like to hear
>> from implementors who (like Victor) have a problem living with either
>> choice.
>> 
>> Please provide your feedback by EOB 1st of August.
>> 
>> Cheers Leif
>> 
>> _______________________________________________
>> Uta mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/uta
> 
> _______________________________________________
> Uta mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/uta
_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta

Reply via email to