NetNews Usenet Archive 1992 #18/NN_1992_18.iso/spool/comp/sys/hp48/4362

Internet Message Format | 1992-08-21 | 3.6 KB

Path: sparky!uunet!dtix!darwin.sura.net!mips!sdd.hp.com!uakari.primate.wisc.edu!usenet.coe.montana.edu!news.u.washington.edu!uw-beaver!ubc-cs!unixg.ubc.ca!ochealth
From: [e-mail] (ochealth)
Newsgroups: comp.sys.hp48
Subject: Re: Hp48 and TI85 compared
Message-ID: <[e-mail]>
Date: 22 Aug 92 04:10:33 GMT
References: <[e-mail]> <[e-mail]> <[e-mail]>
Sender: [e-mail] (Usenet News Maintenance)
Organization: University of British Columbia, Vancouver, B.C., Canada
Lines: 68
Nntp-Posting-Host: unixg.ubc.ca

In article <[e-mail]> [e-mail] (George chow) writes:

>I've always wondered why the HP48 is so slow at plotting. My Casio fx-8000G
>also beats my 48sx at plotting. The only calculator that's slower than my
>HP is the Sharp Super-Scientific.

I think the plotter app. might have a little too much RPL, instead of ML.
The HP hardware aint slow! Some of the graphics routines can put a line
on screen pretty quickly. Too bad they didn't optimize a bit more.
>
>>Simultaneous Linear equations? The HP48 can solve these, and more that 30 too.
>>(assuming memory permits)
>
>If you mean that you can formulate the system as Ax = b where A is a matrix
>and x and b are vectors and just invert A, then no, that's not a feature.
>Matrix inversion isn't a stable operation (just crack open your linear algebra
>text).

IF you enter b A / the HP48 does NOT invert A, but it does the proper elimation
with pivoting etc to return a good appproximation. The Hp48 can also do a
a residual calculation, for even better accuracy. Matrix inversion is not
the best way to solve a linear system, and the HP doesn't do that. Actually,
if a matrix is ill conditioned, you're not going to have any more luck with
elimination than with inversion.
>
>>>The TI85 performs some fudging (rounding, et al) on calculations.
>>>
>>If I remember TI, they don't really round the answer, they just display
>>part of the answer (ie the first 12 digits, and the 13 and 14 are kept
>>internally). Sort of like the HP with 10 FIX on all the time, but trailing
>>zeros aren't displayed.
>
>That sounds just like my TI-66.

Yup! and most other TI scientifics.
>
>>>The TI85 also performs numerical integration and differentiation in the
>>>graphical environment.
>>
>>How fast is it? Does it return an error estimate? I had a TI56 that did
>>Simpson's rule, and it was SLOW. When it finished you had NO CLUE whether
>>the answer is meaningful. The HP uses an adaptive algorithm which will
>>usually return usefull results to an integral, and give you a good idea
>>whether the numerical result is close, or way off.
>
>Hmmm... I must have missed something from reading my HP's manual. How do you
>get an error bound from the HP on your integral? (What do you mean by 'and
>gives you a good idea whether the numerical result is close'?)

By setting the FIXed digits, the HP will give a result accurate to the displayed
digits (more or less). It will leave an error bound in the varaible IERR after
a numerical integration. I can't remember how IERR is calculated, but it gives
a bound for the absolute value of the error or something.
>
>
>>>Rick Homard
>>>
>>                                                      Insert VapourFeature ^^^
>
>George
>


--
______________________________________________________________________________
jpm: [e-mail]
                            Happily using OS/2 2.0 because MS Windows isNT ___
                                                      Insert VapourFeature ^^^