lirtosiast wrote:
M. I. Wright, that's a rounding error in the 10^(log()) calculation (it's off by, I think, 10^-12), as OldMathTeacher said earlier. Using Ans10^(-int(log(Ans instead eliminates the error entirely.
You're right, I had foreseen that this sort of problem may occur. It's not really the TI-OS as such, but the calculation itself which appears to give 0.2379 [displayed result] but is actually a tiny bit out. The calculator works to 14 s.f. and displays to 10 s.f., so in most contexts those tiny errors are unimportant and undisplayed, whereas here they are all-important, influencing how many digits there are.
Your workaround is nice; manipulating the figure only by an integer power of 10 should, as you say, keep exactly to the decimal digits you want it to.