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

Re: Re: Globally limiting precision or accuracy

• To: mathgroup at smc.vnet.net
• Subject: [mg61058] Re: [mg61032] Re: [mg61010] Globally limiting precision or accuracy
• From: Andrzej Kozlowski <akoz at mimuw.edu.pl>
• Date: Sun, 9 Oct 2005 01:35:07 -0400 (EDT)
• References: <200510070737.DAA03251@smc.vnet.net> <200510080648.CAA20910@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

On 8 Oct 2005, at 15:48, Andrzej Kozlowski wrote:

> You certainly cannot be sure that any machine precision
> computation is accurate even to three digits.

Of course I meant "you can't be sure unless you can show that it is  
by hand"; in other words the fact that you got a certain number of  
digits of precision in the input of a machine precision computation  
alone does not guarantee anything about the precision of the output.

Andrzej Kozlowski

• References:
  • Globally limiting precision or accuracy
    • From: Lee Newman <leenewm@umich.edu>
  • Re: Globally limiting precision or accuracy
    • From: Andrzej Kozlowski <akoz@mimuw.edu.pl>