Eugene van der Pijll wrote:
> it's the difference
>between GPS and TAI(GPS) that is fixed, right?
Forgot to reply to this earlier.
Yes, strictly speaking the equation is
TAI(GPS) = GPST + 19 s
I've been using the term "GPS time" slightly loosely, to refer to what
we might call GPST(TAI), defined as
GPST(TAI) = TAI - 19 s
TAI(GPS) stays within about 10 ns of TAI. In purchron the difference
between GPST and GPST(TAI) is swamped by the uncertainty with which TAI
is known, so I think I'm justified in dropping the "(TAI)" there.
-zefram
