[LEAPSECS] Civil timekeeping before 1 January 1972

[Joseph Gwinn]

On Sun, 8 Mar 2015 17:09:59 +0000, Zefram wrote:
> Brooks Harris wrote:
>> It seems to me NTP and POSIX as well as other timescales concerned
>> with "civil time", are essentially disconnected from "reality",
>> expressing "idealized" measurement scales.
> 
> That's very much what they're not.  TT is idealised, and TAI less so.
> 
>>                                                                  I
>> think none of the "civil" timescales are counting in UT - they are
>> measured in SI Seconds, even when prolpetic to 1972.
> 
> NTP doesn't deal with pre-1972 time at all.  (No, the nominal epoch
> doesn't count.)
> 
> POSIX time_t notionally can represent pre-1972 times, but in practice no
> Unix system of that era was synchronised to UTC.  Any use of time_t for
> precise pre-1972 time is heavily retrospective, and the interpretation
> is more governed by the application than by the POSIX standard.  Wild
> pre-1972 Unix time_t values heavily predate the POSIX standard, and their
> interpretation has little to do with UTC.  They are understood to be vague
> UT with usually very poor synchronisation (via the operator's wristwatch).

The actual timescale in POSIX is a found to Seconds Since the Epoch, 
and fractions thereof.  Negative values are not currently defined.

Joe Gwinn