Andrew Lentvorski wrote: > James G. Sack (jim) wrote: >> Out of curiosity, how do the current crop of calculators deal with (eg) >> "representing 0.1" exactly? > > Just a side note, Jim. In the last year or two, your questions are > getting really good.
Hey thanks! Though there might still be room for improvement. ;-) Answers, I've already got perfect (although) only a few: - 42 - Oil it, Hit it with a hammer, Buy a new one - It depends > >> Do they use different representations in different ranges? (seems >> unlikely to really work), or do they maybe recognize & distinguish >> rationals? >> >> What magic lurks within? > > Decimal arithmetic. > > Most modern calculators actually store .1 as .1 (my $7 HP seems to, the > TI website indicates that their expensive ones do too) > > They handle everything by using decimal digits and digit serial > algorithms. Given modern processing power and memory, that's not much > of a big deal. IIRC, one of the HP calculators had to be slowed down > because people didn't trust it. Probably the 12C. > > This was a much bigger deal when calculators first came out and memory > was scarce. I guess it's been a long time since I've seen a string of trailing 9s. I just plugged in my hp45, and (after removing the disgusting nicad pack) find it still works! I kinda recall some tricks doing things like 1 10 / .1 - 20 [10^x] * and getting non-zero. ..or maybe my memory is in worse shape than the nicads? Regards, ..jim -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
