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

Branching on error messages?

• To: mathgroup at smc.vnet.net
• Subject: [mg108946] Branching on error messages?
• From: Jack L Goldberg 1 <jackgold at umich.edu>
• Date: Thu, 8 Apr 2010 07:59:42 -0400 (EDT)

Hi Folks,

In a module I have written I call Reduce and in the most common case,  
Reduce does exactly what I want.  In some rarer cases, however, Reduce  
returns the error message

Reduce::nsmet: ... (basically, Reduce doesn't know how to simplify  
transcendental equations.)

Suppose in Line L6 of my module I have this code

L6 = Reduce[L5, var, Reals];

I want something like this instead,

L6 = If[ Reduce[L5, var, Reals] generates no error message,  
Reduce[L5,var,Reals], Module[ ... ];

So if there is an error message, we go to Module[...] which I hope  
corrects things by introducing numerical methods.

An alternative is this: Define newReduce which acts like Reduce when  
no error message of the type Reduce::nsmet occurs and uses my  
numerical techniques otherwise.

I am unsure how to code L6 or whether to explore defining my own newReduce?
Any guidance is appreciated, as usual.

Jack

• Follow-Ups:
  • Re: Branching on error messages?
    • From: "Carl K. Woll" <carlw@wolfram.com>