[LEAPSECS] A consolidated approach.

[Rob Seaman]

Poul-Henning Kamp wrote:


> Actually, I'm pretty sure time is entirely independent of which way you orient the Earth.


Well, Ernst Mach and Albert Einstein might be among those who quibble :-)

Threads on this mailing list (and the original Navy list) have often made an implicit assumption that "time" is some Platonic ideal.  Rather, whether a timescale is atomic or layered on Earth orientation or some other phenomena, ultimately the clock in question relies on some measurement process.

Our society certainly equates "time" with "time-of-day".  My point being that this list is for discussing the requirements for civil timekeeping, not some esoteric technical timescale.  This is an engineering question, not philosophy.  (Irrelevant digressions about apparent timescales directed to /dev/null.)  We have any number of degrees-of-freedom available for tweaking.  Completely disconnecting civil time from time-of-day is not one of the possibilities.

Leap seconds are a mechanism for synchronizing with time-of-day.  There are other possible mechanisms.  It is not the members of this list (any of us) who have demonstrated an unwillingness to consider all the actual possibilities.


> Shouldn't we discuss what we want from our timescale in the future, rather than which 30 year old computers we will need to replace ?


Of course.  We could have been discussing any number of interesting questions rather than spending 10 years fending off a badly designed proposal that is pursuing an inane hidden agenda.

That said, any world-scale re-engineering project had better include requirements derived from legacy systems.

Rob