That's interesting. It's conceivable that for any given safety factor, such as using 8 digits, there might be some narrow special case fraction inputs such that the rounded results are still different. (Just speculation on my part.) As for checking of my script's results, I just used a hand canculator, e.g.,
(5/9)* 2.54 = 1.41111111...
and observed it seemed to round to 4 digits properly in the few cases I checked.
|