MathGroup Archive 2005

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

Search the Archive

Re: Bug in 5.1??

  • To: mathgroup at smc.vnet.net
  • Subject: [mg54386] Re: [mg54349] Bug in 5.1??
  • From: Andrzej Kozlowski <akoz at mimuw.edu.pl>
  • Date: Sun, 20 Feb 2005 00:07:47 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

  These are just a few examples of the strange behaviour of 
Mathematica's so called "significance arithmetic" at very low 
precision. These are not bugs but consequences of the way significance 
arithmetic works. It is not suitable for very low precision 
computations.
This issue has already caused many heated debates on this list in the 
past and I do not wish to enter another. The best way to deal with this 
is not to use low precision arithmetic for such purposes. It is never 
needed.

Andrzej Kozlowski


On 19 Feb 2005, at 08:31, Gehrmann, James E wrote:

> I'm running Mathematica v. 5.1 under WinXP and had the following 
> interaction with Mathematica:
>
> 5>0
>
> True
>
> 5.>0
>
> True
>
> N[5]>0
>
> True
>
> N[5,3]>0
>
> True
>
> N[5,2]>0
>
> False
>
> N[50000,2]>0
>
> False
>
> Are the last two results a bug or am I not understanding the N 
> function?
>
> Thanks for you help.
>
>


  • Prev by Date: Re: How do I get a DSolve'd function
  • Next by Date: Re: Replace in recursive equations
  • Previous by thread: Re: Re: Bug in 5.1??
  • Next by thread: Re: Bug in 5.1??