Re: TIB: Re: TI-86 math error?

To: ti-basic@lists.ticalc.org
Subject: Re: TIB: Re: TI-86 math error?
From: Randy Overbeek <ro@mindless.com>
Date: Mon, 5 Oct 1998 23:25:37 -0500 (CDT)
Delivered-To: ti-basic-outgoing@towerguard.unix.edu.sollentuna.se
Reply-To: ti-basic@lists.ticalc.org

At 05:10 PM 10/5/98 -0500, you wrote:
>I would call this a bug in the TI-86.  As near as I can tell, it has to do
>with the way the calculator chooses to store information, which I am not an
>expert on.  I had a lot of trouble with this when doing complex equations
>involving imaginary numbers, hence needing the calculator.  If you do a
>simple equation such as 5+2-->x, the calculator stores the answer of the
>equation (7) to the variable.  Hence x-7=0.  I can't explain it, but if you
>do something like 2(5^(1/2)) --> x  (Sorry about the fractional exponent,
>couldn't figure out a radical symbol) It does not store the answer
>(4.472135955).  I don't know exactly what it does, but x-4.472135955 does
>NOT equal 0 as it should, but rather equals -4E-13.  Just to be sure, I
>manually subtracted 4.4.. - 4.4... and got 0.  I don't know how it works,
>but I thought I'd put in my experience.>

I know you already figured out that -4E-13 is almost zero but I thought I'd
add my thoughts.  The reason you don't get exactly zero is your error in
subtracting the decimal value.  The TI-86 actually figures 2(5^(1\2)) out to
a few more decimal places than it displays.  When you enter x-4.472135955
you are actually saying 2(5^(1\2))-4.472135955.  To the processor there is a
very small difference; -4E-13 to be exact.  This common problem can
occasionally cause problems if you're not careful.

                Randy W. Overbeek  ro@mindless.com