Re: My first programable HP Message #12 Posted by Mike Morrow on 14 Mar 2008, 1:32 p.m., in response to message #11 by Ken Shaw
That's an excellent example that shows the difference in precision *and* accuracy between results from the 15C and the 42S.
"Precision" tells us how many significant figures an answer is given, while "accuracy" tells us how close the answer is to the true value. A result from a calculation that should be Pi could be shown on machine A as 3.141599999, and on machine B as 3.141592654. Machine A and machine B have the same precision (10 significant digits), but machine B has much better accuracy than machine A.
AFAIK, all Saturn (real or emulated) based scientific calculators (such as the 42S, 32S, 32SII, 48/49/50 series, and even the lowly 20S) performing the same calculation will generate answers of identical (and excellent) accuracy and precision that are much better than those obtained from preSaturn calcs like the 15C.
Not having one, I don't know if the recent Kinpo units such as the 33S and 35S will generate the same numbers as the Saturnbased units.
The 30E/Cseries (Spice) and 10Cseries (Voyager) have the same *precision* as the earlier Classic and Woodstock units. But the firmware algorithms in Spice and Voyager models produce much better *accuracy* than those of most of the earlier models. (The 67 and the 19C/29C have the better accuracy algorithms that appeared in later models.) For example, the same calculation performed on a 34C and a 15C will produce identical results. Likewise, the same calculation performed on a 65 and a 25C will produce identical results, but they will be of lower accuracy than those from the 34C and 15C.
FWIW, a 2499iteration loop of the old classic "Savage" benchmark will execute *significantly* faster on a 25C compared to the later 34C, but the accuracy of the 34C will be better. Apparently the better accuracy of the 34C is obtained with a tradeoff in speed.
Thus, the old Classic and Woodstock machines are lowest in precision and accuracy ratings. The Spice and Voyager machines have the same precision as the earlier models, but better accuracy in general. The Pioneers and other Saturnbased units have both better precision and better accuracy than any earlier models.
These differences will likely be of significance only in long or repetitive or iterative calculations. As one who made the transition from slide rules to calculators in my professional life 36 years ago, it still seems a little unnatural to be worrying about all those digits way out on the right side of the decimal point.
Mike
Edited: 14 Mar 2008, 2:28 p.m. after one or more responses were posted
