Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
1997
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 1997

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

Search the Archive

Re: NSolve

  • To: mathgroup at smc.vnet.net
  • Subject: [mg8631] Re: NSolve
  • From: Daniel Lichtblau <danl>
  • Date: Fri, 12 Sep 1997 04:10:45 -0400
  • Organization: Wolfram Research, Inc.
  • Sender: owner-wri-mathgroup at wolfram.com

(Fred Lang) wrote:
> 
> Dear Mathgroupers,
> 
> I've found something strange with NSolve in Mma 3.0 (PC and MAC).
> 
> Try the following input:
> 
> In[1] =  NSolve[ x^2 + x(z+1) == 0 , x]
> 
> Out[1] = {{x -> -1.},{x -> 0.}}
> 
> In[2] =  NSolve[x^2 + x z + x == 0 , x]
> 
> Out[2] = {{x -> 0.},{x -> -1. - 1.z }}
> 
> The first output is wrong. NSolve has choosen for me to put z = 0.
> 
> OK! NSolve is a numerical command and z has no numerical value, but
> 
> why don't NSolve give me something like "No numerical value for z, I put z
> = 0" ?
> 
> In Mma 2.2, NSolve gives the correct answer.
> 
> Thank you for your remarks.
> 
> Cheers.
> 
> Fred Lang


Small glitch, fixed in version 3.0.1. A work-around in version 3.0 would
be N[Solve[...]]. This is reasonable because, as you note, nothing
usefyul can be done with inexact methods because we have a non-numerical
parameter anyway.

Daniel Lichtblau
Wolfram Research
danl at wolfram.com


  • Prev by Date: matrix
  • Next by Date: Re: programing: reduce list to cycle
  • Previous by thread: NSolve
  • Next by thread: conics and graphics