Re: [Ntp] Timescales

[Magnus Danielson <[e-mail]>]

Philip,

On 2020-12-10 01:33, Philip Prindeville wrote:
>
>> On Dec 9, 2020, at 9:29 AM, Steve Allen <[e-mail]> wrote:
>>
>> On Wed 2020-12-09T14:06:50+0100 Magnus Danielson hath writ:
>>> The most robust solutions is creating a time-scale that is some
>>> well-defined shift from TAI, and then provide the TAI-UTC difference
>>> explicitly along with a pre-warning system. Look at for instance GPS [1]
>>> and PTP [2]. I've chosen this path myself and it has worked very nicely.
>>>
>>> So, this is not a crazy idea, it's robust engineering proven in battle.
>>>
>>> What is problematic is able to support backward compatibility to the
>>> classical NTP time-scale as this needs to be handled in all time-stamps.
>> The extended discussion in the NTPWG now is strikingly similar to the
>> arguments in the late 1960s which resulted in leap seconds.
>> A 1969 report analyzing the options pointed out that leap seconds
>> would cause problems in precise timing systems used for navigation.
>> Backward compatibility concerns forced the international agreement to
>> have leap seconds.  Before the first leap second the technical members
>> of the committees proceeded to change their operational systems to use
>> purely atomic time with no leaps, and that choice was repeated for
>> GPS, Galileo, BeiDou, etc. because it was the only robust way.
>
> Maybe I’m just slow, but I’m not getting any of this discussion.
>
> It seems to me that time advances, only, and at a fairly constant rate (barring relativistic shifts as the earth accelerates as it moves in the same direction as the sun’s trajectory, vs. decelerating when the earth’s orbit takes it in the opposite order as the sun’s trajectory, etc.).
>
> It simply goes “tick tock tick tock”… and that’s all it does.
>
> Anything relating to mapping to dates and times on a calendar is a purely “presentational” thing, like preferring decimal over octal or hex or binary…
>
> It seems to me we’re overthinking things.
>
> Or worse, assuming we’re the only ones that really understand time, and deciding “we” should fix it “here” because we understand it, rather than letting “them” fix it “there” because we don’t trust others to do so adequately.
>
> What am I missing?

Through history we kept inventing ways to describe and annotate the
time. Not all of them have their annotations increase monotonically for
a myriad of reasons. Trying to get some order in this quaggmire tends to
be "interesting" depending on how much each person learned and through
their exposure to various solutions, and also lack of exposure to
various solutions. With that comes both benefits and problems. As time
passes, new things is added, new quick-hacks is added, more labyrinth of
details show up and requirements shift over time.

Keeping a common map can be challenging, especially as one has to be
detailed on details and for what case they apply and not.

Whatever simplification one come up with, it ends up causing things to
be painful somewhere else. Getting the set of good guys here agree takes
a bit of talking.

Cheers,
Magnus