[LEAPSECS] suppressing TAI

[Dennis Ferguson]

On 14 Nov, 2011, at 14:21 , Zefram wrote:

> Happy birthday TAI, 1.7 Gs old at 2011-11-14 22:12:46 UTC.

> 

> There's a component of the proposal to abolish^Wredefine UTC that

> refers to "suppressing TAI".  This sounds awfully silly, and more

> than a little reminiscent of the worries one list member has about UT1.

> ("Put the maser down and come out with your hands up.")  So I'm wondering

> what it's actually intended to convey.  What do list members seriously

> imagine would be entailed by the intention to "suppress TAI"?


My guess would be that BIPM Circular T would stop reporting TAI-UTC and
TAI-TA(k), and would recast the other TAI-based measurements they report
into terms of the new UTC.  Oh, and timescales with a TAI basis (via TT),
like TDB and TCG, would be redefined in terms of their relationship to the
new UTC (without actually changing them).  You could still compute TAI if you
wanted to, but it wouldn't appear as an independent variable on the right hand
side of the "official" equations for anything.

I think their picture of how the world should work is that there should be
one precisely maintained timescale which is distributed by all services and
which all other timescales are computed from and/or measured against.  In
this picture TAI is part of the problem, but that UTC-as-it-is-now is not
the solution either is demonstrated by the fact that the three new precision
time distribution services designed in the past decade, Galileo, Beidou and
IEEE 1588, have all chosen something different than UTC (and different from
each other, too!) for their system timescales, while those which have attempted
to use UTC for this (GLONASS certainly, NTP arguably) seem to have suffered for
that decision.  That expanding plethora of timescales, all doing the same thing
but all slightly different, is what they seek to cure, and in this view TAI is
something to be cured.

Dennis Ferguson