Nice little thing that would be real cool to have as a watch.
Minor bug: There's a year 0 JAN 0 but year 0 JAN 1 should jump to year -1 DEC 31
Ah, good point. I collected a whole bunch of Julian-day-number-to-date conversion functions off the internet and then ended up having to write my own from scratch because they all relied on 32-bit integer arithmetic. I imagine there are probably other problems with my converter as well. For instance, the Gregorian calendar started in the 1500s; before that it was the Julian calendar, which differed by a couple of days at the 1AD mark. Probably better to use the Gregorian to go back in time, though, because it matches the solar year better.
Also the planet positions are unlikely to be accurate more than a few hundred years out. Mercury, in particular, has a large first derivative on its orbital elements. It would be fairly easy to add the first derivatives of the orbital elements in; JPL provides those.