Re: [Cbor] I-D Action: draft-ietf-cbor-time-tag-04.txt

[Emile Cormier <[e-mail]>]

On Tue, Jan 17, 2023 at 3:04 PM Vadim Goncharov <[e-mail]>
wrote:

> Then I understand you, but instead of seven new tags I think reference
> to Section 4 here and in 3.5.5 should be prohibited. It allows for too
> large values if a generic #6.1002 duration would be allowed. In fact,
> as most clocks support 1-second resolution, it should give part of
> second (no more than 1 second), like Precision field in NTP.
>

I think the rationale for allowing Tag 1002 for the clock quality
Uncertainty and Guarantee attributes is to allow flexibility in their
resolution, as well as numeric format (integer, float, bignum, etc). If
there was universal agreement on the resolution and numeric format of the
Uncertainty and Guarantee attributes, then we could indeed remove Tag 1002
as a possibility for those attributes. That would eliminate the possibility
of recursion.

I'm not an expert on time synchronization, so I cannot propose a resolution
and numeric format of the Uncertainty and Guarantee attributes. One
possibility would be:

[quantity, exponent]

where 'quantity' is an integer or float, and 'exponent' is an integer for a
10^(exponent) multiplier for the quantity. But this is not consistent with
how Tag 1002 durations work, and I can see the "ugliness" of this
inconsistency.

The Uncertainty and Guarantee clock quality attributes might be used for
scientific applications, so I in that case, don't think they should be
constrained to numeric formats and resolutions used in network time
synchronization.