MathGroup Archive 2002

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Off by 0.00000001, Why?


Under Mathematica 4.2, I obtain the same result as you -- off by 1 in 
the last decimal place displayed from what is shown in Maeder's book.

That was on a Pentium 4 PC.  Could it be that Maeder was using a 
different hardware platform, and that that alone would account for the 
difference?  Actually, I'd be surprised if that were the reason, because 
I applied NumberForm[%, 15] to the result and obtained:

    0.00157908413803032


Steven T. Hatton wrote:
> I'm going through Dr. Maeder's book, _Compute Science with Mathematica_, 
> entering the examples into Mathematica and evaluating them.  Here is my 
> current notebook:
> http://baldur.globalsymmetry.com/proprietary/com/wri/notebooks/csm-examples.nb
> 
> I've Noticed that in a few instances my results differ slightly from his.  I'm 
> wondering why this is happeneing.  One would expect that the same algorithm 
> would produce identical results regardless of the system on which it was run.  
> I'm running on 4.2, and it's certain Dr Mäder was using an earlier version.  
> Could that be the cause of the descrepency?  
> 
> One example is the very last evaluation in section 1.1.6.  His book shows 
> 0.00157909, whereas I get 0.00157908.  This may not seem like a big deal, but 
> I heard of one company immediately losing a banking contract for accumulated 
> errors of this magnitude in their software.  It sounds like a good way to 
> lose a space probe as well.
> 
> Any thoughts on this?
> 

-- 
Murray Eisenberg                     murray at math.umass.edu
Mathematics & Statistics Dept.
Lederle Graduate Research Tower      phone 413 549-1020 (H)
University of Massachusetts                413 545-2859 (W)
710 North Pleasant Street
Amherst, MA 01375



  • Prev by Date: Re: Different EPS exported file from frontend and kernel
  • Next by Date: MathLink and Notebooks
  • Previous by thread: Off by 0.00000001, Why?
  • Next by thread: Re: Off by 0.00000001, Why?