[LEAPSECS] The POSIX Time Rationale - in the Working Group's own words

[Steve Allen]

On Fri 2016-12-30T20:20:57 +0000, Poul-Henning Kamp hath writ:
> >It may prove useful to know why the POSIX Working Group (WG) excluded
> >leap seconds, in their own words.
>
> POSIX didn't "exclude leap seconds", they chose not to "add leap seconds".

In the 1988 standard it is clear that they chose not to add the
century leap-year rule.  They preferred simplicity for now rather than
correctness over an interval longer than time_t could then encode.

But the archives of the tz project show that it was well underway at
the time of the 1986 draft and that Bradley White had created the leap
second handling code in tz before the voting on the 1988 standard.
Landon Curt Noll is listed in the 1988 standard as one of the members
and he reported that POSIX knew that yet chose to exclude leap seconds.
https://www.mail-archive.com/leapsecs@rom.usno.navy.mil/msg00109.html

POSIX chose not to attempt precise time.  Anyone trying to do precise
time with POSIX is implementing an extension to the standard.

There is much more record of the consternation over the handling of
time zones in the 1988 standard.  The end result of time zones in the
1988 standard was that POSIX did not require full handling of the
complexity of civil time, but POSIX did allow extensions that do
correctly handle civil time.

In contrast, the end result for leap seconds is that POSIX did not
require handling of leap seconds, and that it explicitly prohibited
their correct handling unless an implementation violates some aspect
of POSIX -- pick which aspect, and everybody chose a different one.

--
Steve Allen                    <sla at ucolick.org>              WGS-84 (GPS)
UCO/Lick Observatory--ISB 260  Natural Sciences II, Room 165  Lat  +36.99855
1156 High Street               Voice: +1 831 459 3046         Lng -122.06015
Santa Cruz, CA 95064           http://www.ucolick.org/~sla/   Hgt +250 m