[LEAPSECS] Bloomberg announced its smear
martin.burnicki at burnicki.net
Wed Sep 28 03:59:34 EDT 2016
Michael Shields via LEAPSECS wrote:
> On Tue, Sep 27, 2016 at 2:02 AM, Tony Finch <dot at dotat.at> wrote:
>> So, if everyone who is smearing is doing so piecewise-linear, it seems
>> there is a great opportunity for them to get together and choose a
>> consensus set of smear parameters.
> I think that would be a great idea. There isn't time to get agreement
> for leap 37, but it would be very nice if smeared clocks agreed in the
> future. After all, the point of having clocks, as opposed to interval
> counters, is for them to agree.
> The 24-hour smear from noon to noon UTC seems to be a good choice. It
> is only 11.5 ppm frequency change, within the error of most clocks
> anyway, and it is simple to reason about.
Hm, IMO the advantage of the initial smear approach (24 hours before the
leap second) is that smearing is finished as soon as the leap second has
occurred, so the beginning of the next UTC day /hour / minute is accurate.
I *know* the new 12/12 approach only has a maximum error of 0.5s while
the old 24/0 approach has 1 s maximum error, but the questions are:
- If a temporary 0.5 s time error against UTC can be tolerated, should
it matter if the maximum error is 1s?
- With the new 12/12 approach the offset to real UTC steps from +0.5 s
to -0.5 s (or vice versa) over the leap second. Does this affect any
application which normally expects to work with real UTC time?
More information about the LEAPSECS