HP 50g fails simple SOLVE. Comparing to TI-89, which succeeds
in [HP48]
|
From: Steen Schmidt on 12 Feb 2007 16:21 Marcin Witek wrote: > What guess value for ROOT are you using on HP? Zero. Regards Steen
From: Marcin Witek on 12 Feb 2007 17:16 Steen Schmidt wrote: >> What guess value for ROOT are you using on HP? > Zero. With guess value of 0 my TI-89 Titanium computes this in about 9 seconds, but with guess value of 1 it does it in just 2 seconds. Wit
From: Virgil on 12 Feb 2007 18:18 In article <74ddk2t6tz9y.dlg(a)wit.ummagumma>, Marcin Witek <NOwitSPAM(a)malenstwo.iinf.polsl.gliwice.pl> wrote: > Steen Schmidt wrote: > >> What guess value for ROOT are you using on HP? > > Zero. > > With guess value of 0 my TI-89 Titanium computes this in about 9 seconds, > but with guess value of 1 it does it in just 2 seconds. On my HP49+, the ROOT command finds the real root of the equation in under 1 second from either guess. But MSLV won't work with a guess of 0 because LN(X) is not defined there.
From: John H Meyers on 12 Feb 2007 18:51 On Mon, 12 Feb 2007 17:18:03 -0600: > MSLV won't work with a guess of 0 because LN(X) is not defined there. Didn't my examples work with one starting guess of [(0,0)] ? In the real-valued numeric solver, a single guess of 0 is actually equivalent to the pair of guesses { 0 1 } and single non-zero guesses are adjusted by a very small amount to create a second starting guess. I don't know anything more about how MSLV works, but it did produce both real and complex answers for the original examples, starting from [(0,0)], although that may not have been the best place to start :) [r->] [OFF]
From: Helen on 12 Feb 2007 21:20
On Feb 12, 8:53 am, "JBGM" <Literatron...(a)gmail.com> wrote: > > Besides, the example you give is hardly an example for a "massive > > symbolic computation"... > > But just an example of the device's mathematical capability. The example you gave has next to nothing at all to do with symbolics. Somewhat unexpected if what you are interested in is symbolics. > The world is fairly rich and complex, so be less surprised when you do not > understand causes and reasons; you might be missing the rest of the > iceberg ;-) I won't comment. How you want to do your work is really none of my business. You may disregard my idle comment on this. > The algorithm is simple: if (estimated hours saved * personal labor > hour cost per hour) - (cost of calculator + time invested in training > * personal labor hour cost per hour) > 0 then go calculator. So far, I > go calculator. Maybe your algorithm is a bit too simple... > BTW TI-89 seems to have superior symbolic manipulation > capabilities FOR MY PURPOSE. Look athttp://technicalc.org/tifaq/?tivshp.htm. > HP 50g seems fairly superior in numerics. I am familiar with the differences. I agree that everything depends on your specific purpose. There are cases where the TI may be the better choice, for others the HP is better. Since you can try the caluclators via their emulators, you can find out yourself, and answer your own question much more reliable than any comparison on some website could. Good luck with your work! |