[ntpwg] Further to the timestamping issue

[Steve Allen]

On Thu 2008-06-19T10:28:42 -0700, Rob Seaman hath writ:
> Nothing would stop Gambis from
> announcing a leap second for December 2011 at the same time as a leap
> second for June 2009.

Except for the very strong "sic semper faciebamus" which lingers from
the totally haphazard way that actions by the BIH, IAU, and CCIR in
the 1960s and 1970s led to the current practices.  With two of those
agencies no longer in existence any change is hard to effect.

The world would be better off if the broadcast precision time scale
(and "broadcast" now includes internet, POSIX time_t as well as radio)
did not have leap seconds.  The 2003 colloquium on UTC in Torino
admitted as much.  Leap seconds are very suited for humans, and not
suited for machines.

What the colloquium did not point out was that the POSIX specs for
zoneinfo already require enough complexity to handle the leap seconds,
and the mechanisms for updating the zoneinfo files are well tested.
All that's needed is to rename the time scale used in broadcasts,
NTP, and POSIX time_t.

NTP already contains the necessary complexity, if not the desired
features.  All of the missing bits would be much easier to solve
by putting leap seconds in zoneinfo than by messing with NTP.

This discussion should go to LEAPSECS and leave NTPWG.

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