Post by PythonPost by Maciej WozniakThat's how it works in physics: an idiot is
mumbling about pears growing on a willow -
and his obedient doggies are preparing some
new, better definitons of pears and willows.
And then pears growing on a willow are
confirmed; with a little help of new,
better definition of "confirming", of
course.
If you hadn't decided to stay as far as possible to physics you would know
that it is definitely not "how it works".
[for the strayed-in, and perhaps misled kiddies, if any]
Indeed, relativity, or any other theory, has nothing to do with it.
It is pure empirism, about realising the best posible time standard.
Averaging the many atomic clocks to obtain TAI just takes
the observed clock rates into account.
The why of some clocks being systematically slow or fast,
with respect to the average of them all, doesn't come into it.
As a matter of fact BIPM begins the clock comparisons
with the computation of EAL, the Echelle Atomique Libre,
which is just an average that doesn't involve any corrections.
Next they compute TAI from that.
An empirical correlation of clock rate with altitude,
or more precisely with the Newtonian potential,
would work just as well for the purpose,
Jan