NavList:

Frank,

Quite happy for you that you could solve this tricky situation.

Fully agree with you that "  But it was problematic because, as it turns out, the bug is intermittent. And intermittent bugs are the worst! "

I has started to mull over your predicament and had ruled out :

- Incorrect ephemeris computation ? Probably not since it should have yielded incorrect cleared lunar distance.

- Incorrect refraction computation ? probably not either since it should have yielded offset of about 3 arc minutes, which is not the case. Actual offset is 60 times smaller.

- Maybe angles transcription from Decimal Degrees into Degrees Minutes seconds / Degrees minutes, tenths of minute format ??? This could yield the same magnitude as the observed error and does seem to behave somewhat erratically ... In other words that could be some kind of "interface problem" ... That's always where I start looking in case of erratic / intermittent bugs.

- Whats else ?? I could not figure out other tracks to investigate.

But from thousands of hours spent on programming, albeit RPN programming, I have become convinced that nothing - if not nobody - is or can be more stupid than a computer or a calculator.

If it were possible for you to indicate more of that kind intermittent / erratic bug, that could be of interest to the ones of us interested in programming.

Meanwhile, thanks for this excellent news as regards your on-line Calculator.

And happy for you again.

Antoine