MathGroup Archive 2006

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

Search the Archive

Re: Accuracy of Norm[{0``1}] is infinity?

  • To: mathgroup at smc.vnet.net
  • Subject: [mg67800] Re: Accuracy of Norm[{0``1}] is infinity?
  • From: Jens-Peer Kuska <kuska at informatik.uni-leipzig.de>
  • Date: Sat, 8 Jul 2006 04:56:05 -0400 (EDT)
  • References: <e8lfjn$qsu$1@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

Hi,

it is a property of the computer numbers, that 0 has an
infinite precision.

Regards
    Jens

Tammo Jan Dijkema wrote:
> I wonder why the accuracy of Norm[{0``1}] is infinity, while the
> accuracy of Norm[0``1] is 1, as expected. This is only the case for the
> two-norm, the precision of Norm[{0``1},1] is 1.
> 
> Does anybody have a clue why this is the case? I think it would be more
> consistent if the accuracy of the norm of {0``1} would be 1, since
> Accuracy[Sqrt[k.Conjugate[k]]]==Accuracy[Total[Abs[{0``1}^2]]^(1/2)]==1
> 
> Regards,
> 
> Tammo Jan Dijkema
> 


  • Prev by Date: Non Linear fit problem
  • Next by Date: RE: 3D graphics headache
  • Previous by thread: Accuracy of Norm[{0``1}] is infinity?
  • Next by thread: 3D graphics headache