[ntp:hackers] ITU and Leap second elimination

Magnus Danielson magnus at rubidium.dyndns.org
Sat Oct 5 14:40:59 UTC 2013


On 10/05/2013 02:52 PM, juergen perlinger wrote:
> On 10/05/2013 01:16 PM, Magnus Danielson wrote:
>> <snip>.
>> Smearing of leap-seconds might fit some applications but not others.
>>
>> time_t alongside a utctai coeff can handle leapsecs nicely.
> Of course. Once you have an API where you can get both values atomically
> in one sweep. But strictly speaking, this is just TAI in disguise.
Agreed. You need to preserve the TAI/UTC duality. The attempt to forge
one time-scale to rule them all is not possible when the requirements is
mutual exclusive. Keeping two, with a simple relationship is the most
practical we can do, then we pick the one best adapted to the technical
situation of a particular application.

Cheers,
Magnus


More information about the hackers mailing list