[LEAPSECS] Crunching Bulletin B numbers (POSIX time)

[Rob Seaman]

On Feb 21, 2011, at 11:32 AM, Poul-Henning Kamp wrote:


> Would abolishing leap seconds make timekeeping less or more complex ?


See http://en.wikipedia.org/wiki/Complex_question

You are assuming that leap seconds are a free parameter that can be abolished (or redoubled for that matter) at will.  Rather, leap seconds are a means to an end, and that end addresses requirements in the problem space.  Those requirements won't vanish even if leap seconds do.

If the ITU wants a timescale without leap seconds, and if that timescale must be distinct from TAI and GPS, then they are free to define a new timescale.  The pertinent requirement that was identified in Torino in 2003 was that such a timescale shall be called something other than "UTC".  Ignoring this requirement doesn't mean it goes away.  Presumably the leapless timescale would have to be called something other than the "volt" or the "ampere" or "kelvin" or "Avogadro's Number" or "pi", too.  A timescale without leap seconds is none of those things, just as it would not be UTC.

Similarly, "Coordinated Universal Time" is subclassed from "Universal Time".  Universal Time in its various forms provides a realization of Greenwich Mean Time.  Greenwich Mean Time is layered on the "synodic day".  By kicking the leg out from under UTC, all the contingent complexity would slide into a tangled mess on the floor.  Yes, this would make timekeeping as a whole significantly more complex.

...but then, the ITU doesn't care about timekeeping as a whole, do they?

Rob