Re: [netmod] consensus all: timezone-location and draft-ietf-netmod-iana-timezones-03

[Andy Bierman <[e-mail]>]

Hi,

This does not address the political problems that seem to be
associated with the mere existence of a YANG snapshot
of the TZ database, so it is OK if we just use a string.
Since Lada is OK with a string, and that is already seems
to be the consensus, I will withdraw this union typedef idea
so we can finish this draft.


Andy



On Sat, Feb 1, 2014 at 12:54 PM, Andy Bierman <[e-mail]> wrote:

> On Fri, Jan 31, 2014 at 7:20 AM, Ladislav Lhotka <[e-mail]> wrote:
>
>>
>> On 31 Jan 2014, at 16:02, Thomas Nadeau <[e-mail]> wrote:
>>
>> >
>> >       I would like to call consensus on this matter. I've received %100
>> supporting comments for the proposal from the WG with no one not liking the
>> proposal. The only comments received were from Martin and
>>
>> You must have missed my comments, so let me say I don't like the
>> proposal. I think we are throwing out the baby with the bath water.
>>
>>
> I think I have a compromise solution that saves the baby :-)
>
> typedef timezone-enum {
>   // TZ snapshot we have now
> }
>
> typedef timezone-name {
>    type union {
>       type timezone-enum;
>       type string;
>    }
> }
>
>
> 1) The NETMOD WG would maintain this module, not IANA.
> 2) The snapshot would be updated at long-term intervals, such as every 2
> years
>
> This forces the server to support a set of values the client can predict
> and also supports extra strings not in the snapshot.  This prevents false
> negatives
> wrt/ YANG validation, but not false positives (i.e., values have been
> removed
> from the snapshot).
>
> Do timezone names get removed from the database? IMO, this seems rare
> enough not to worry about until the next YANG snapshot RFC is published
> with the now obsolete enums.
>
>
>
>
>> Lada
>>
>>
> Andy
>
>