Re: [Cbor] CBOR tag for RFC 3339 full-date values

[Carsten Bormann <[e-mail]>]

Hi Joerg, Mike,

we discussed the date tag proposals at the CBOR interim today.
The meeting consensus was that we should merge the date tag part of Joerg’s specification (i.e., both textual and numeric date references) with Mike’s draft, and spend a 1+1 tag (which would indeed conveniently be tag 100 = ‘d’).

I would like to work with the two of you to generate a -01 of Mike’s draft with text from Joerg, for discussion on May 25th (or whatever replacement we find for the physical IETF107 meeting.)

(We would look at time-of-day and timezone right after registering the date tag from -01; maybe in a -02.  My personal view would be that the time-of-day tag from Joerg would go right in, but we might want to go for a 1+2 tag for timezone [29818 for ’tz’ maybe?].  But let’s discuss this when we are there.)

Is that a good way forward?

Grüße, Carsten


> On 2020-03-11, at 16:25, Carsten Bormann <[e-mail]> wrote:
> 
> Hi Joerg,
> 
> I’m sorry we didn’t handle your requests properly.
> Let’s try to fix this now.
> 
>> On 2020-03-04, at 08:42, Richter, Jörg <[e-mail]> wrote:
>> 
>> Hi,
>> 
>>> The need for a date (not time) tag has come up before; I just can’t remember where.
>> 
>> One year ago I sent a proposal für a date tag [1].
>> So far it has not made it into the overview of the assigned CBOR-tags [2].
>> 
>> I would very much appreciate if the suggested tag number is used for the date-tag, since we already use it internally.
> 
> What I like about Mike’s draft is that he references RFC 3339, which already is an established specification for how to write dates textually.  For that half of the proposal, I don’t think there is a technical difference with your proposal (except that Mike suggested a different tag number).  Beyond Mike’s draft, we would need to decide that we like to use the same tag with an epoch-based numeric date.  I think this is OK; it just requires users of the tag to specify which of the internal structures (or both) they would accept.  Note that we are running into the same issue with tag 1 (as used in EAT in the RATS WG), so that may help accepting that this is a normal thing to so.
> 
>> What is missing for the suggested tag numbers to be assigned?
> 
> I think we just have to make the decision to spend the 1+1 (short) tag.
> That becomes somewhat easier with including the numeric date, because that is indeed quite compact and merits a short tag.
> 
>> Btw. I have another proposal for currencies and money amounts at [3] whose tags were not assigned either.
> 
> Let’s get to that, too, right after we are done with the date tag.
> 
> Grüße, Carsten
> 
> 
>> - Jörg
>> 
>> [1] https://j-richter.github.io/CBOR/date.html
>> [2] https://www.iana.org/assignments/cbor-tags/cbor-tags.xhtml
>> [3] https://j-richter.github.io/CBOR/currency.html
>> 
>