Time has never been a strong suit for me. I tossed out my last wristwatch nearly 30 years ago. There has always be a clock somewhere around...and if there wasn't it must not be that important.

However, to do celestial navigation time is important, or so I read in Dava Sobel's book Longitude.

Navigatrix syncs System Time with the GPS to within milliseconds. The Task Panel Clock time, although reflecting System Time, can be off by many seconds and is not reliable for accurate time reference.

I read somewhere that TDT (Terrestial Dynamical Time), was used from 1984-2000 as a time-scale of ephemerides from the Earth's surface. Though the name changed to TT (and brought into compliance with the General Theory of Relativity) in 2001 it's still TDT = TAI (Atomic Time) + 32.184 seconds.

The whole thing comes out like:

Is GPS time sufficient, or is another time standard more appropriate?

If I remember this right, the GPS' internal system time is indeed different from UTC and was designed that way. However, the time that is reported by the GPS units we have on board is UTC. I think to remember that the satellites are kept up to date with the current difference between GPS time and UTC and that this delta is transmitted as part of the satellite signal so that GPS units can display correct UTC. At least on our old Garmin 128 unit, the UTC displayed is in synch with the time signal sent on WWV as UTC.

So using the GPS-provided UTC in Navigatrix should be fine (accurate to far less than a second).

UTC seems to be a bit of strange breed though. On the one hand it ticks in synch with TAI atomic time on the other hand it is not allowed to stray far from mean solar time and the two parents typically go their own separate ways. The challenge for electronic ephemeris calculators then becomes to predict the leap seconds - which are occasionally inserted into UTC to keep it on speaking terms with both parents - into the future. Both, AA and PyEphem extrapolate this based on historical data (PyEphem actually uses AA code for this) but it's a bit like forecasting hurricane tracks - the error cone gets wider the further you go into the future.

When I last looked the PyEphem code used historical data up to 2011. For the 2016 release of NX we may have to update that code...

Take this all with a grain of salt...the above may be a dangerous mix of a patchy memory and a vague understanding to begin with... I'll happily be corrected.

Let us not forget Capt Slocum's "tin clock"

"The want of a chronometer for the voyage was all that now worried me. In our newfangled notions of navigation it is supposed that a mariner cannot find his way without one; and I had myself drifted into this way of thinking. My old chronometer, a good one, had been long in disuse. It would cost fifteen dollars to clean and rate it. Fifteen dollars! For sufficient reasons I left that timepiece at home, where the Dutchman left his anchor."
"At Yarmouth, too, I got my famous tin clock, the only timepiece I carried on the whole voyage. The price of it was a dollar and a half, but on account of the face being smashed the merchant let me have it for a dollar."

The slop in my shoots more than offsets any questions of the accuracy of gps time tics.

Very true...

if I remember correctly, the "tin clock" even lost its minute hand somewhere along Slocum's journey and he still made it around with his "lunars"...