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

Re: problems with FindRoot: what worked with 4.2 does not work with 5.0

• To: mathgroup at smc.vnet.net
• Subject: [mg47379] Re: [mg47356] problems with FindRoot: what worked with 4.2 does not work with 5.0
• From: Gerry Flanagan <flanagan at materials-sciences.com>
• Date: Wed, 7 Apr 2004 03:16:39 -0400 (EDT)
• References: <200404061036.GAA09266@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

The usual fix for us is to modify the functions so that they only execute 
for numerical input. This can be done by placing a NumberQ conditional with 
some or all of the input arguments. I'm guessing that without the 
conditional, FindRoot now attempts a symbolic derivative, and fails if you 
have a complex function (like anything containing an If statement).

Gerry Flanagan


At 06:36 AM 4/6/2004, Edgar Dachs wrote:
>Computers & Geosciences, 24/3: 219-235) und worked out a
>new version that I tested with Mathematica version 4.2. With 5.0, I now
>find problems that several of my functions do not work any more,
>obviously do to changes that have been employed between 4.2 and 5.0
>concerning FindRoot.
>Below is a little test-function that illustrates the problem: the
>return-value of the function depends on the value of x, as in the much
>more complex petrological functions that I have written, where
>calculations depend e.g. on the value

• References:
  • problems with FindRoot: what worked with 4.2 does not work with 5.0
    • From: "Edgar Dachs" <edgar.dachs@sbg.ac.at>