NavList:

   

Reply

I wrote:
> Delta T was *negative* in 1901. That unusual circumstance revealed a bug
> in my program, which in turn was due to a bug in my positional astronomy
> DLL. It's in a feature which isn't used much, and fortunately there's an
> easy workaround. I'll update the Web site with a description of the
> workaround, and issue a fix soon.

Correction: the bug was in the lunar distance program itself. There was
nothing wrong with SofaJpl, the positional astronomy "engine" that helps 
with the computations. It has a function to convert time (duration, not 
epoch) to angle. The result is put into the interval [0, 2π) radians, so 
it's never negative, even if the angle is. This behavior is precisely 
documented. I simply forgot and used the function improperly.

The code at my site has been corrected. In addition, the observation
data for the Arago lunar distance and occultation have been added to the
existing examples in the source code.

http://home.earthlink.net/~s543t-24dst/sofajplNet/LunarDist2.html

(In the Lunar2.vb file.)

-- 

   

Reply