- To: mathgroup at yoda.physics.unc.edu
- Subject: ComplexExpand
- From: Elliot Weinberg <weinberg at symcom.math.uiuc.edu>
- Date: Thu, 10 Feb 1994 09:08:15 -0600
The following peculiar phenomena seems to be special to certain installations of Mathematica. I can reproduce it on the logins of some of my colleagues, but not on others. Nor does this happen on my personal copy for a MacIntosh. The Wolfram people have not been able to locate the difficulty; it does not arise on any of their versions. I write to ask if anyone else can reproduce the problem or suggest the solution. ComplexExpand behaves strangely on my Sparc station unless it is 'initialized' by first evaluating it in a specific case. Here is an example. I ask to expand Exp[I x] and get the wrong answer: mira:::math Mathematica 2.2 for SPARC Copyright 1988-93 Wolfram Research, Inc. In:= ComplexExpand[Exp[I x]] Out= 1 In:= Exit Starting a brand new session I can get the correct result by first asking to evaluate an expression which yields a nonsymbolic result. mira:::math Mathematica 2.2 for SPARC Copyright 1988-93 Wolfram Research, Inc. In:= ComplexExpand[Exp[Pi I]] Out= -1 In:= ComplexExpand[Exp[I x]] Out= Cos[x] + I Sin[x] I stumbled across this phenomena when reviewing a program I use in a symbolic algebra course. It ran fine last summer under 2.1, but blows up for me in 2.2. ComplexExpand was used in an incidental way, so it took a while for me to figure out the source of my difficulty. The examples above use ComplexExpand in a way different from that used in the program, but clearly illustrate that something strange is happening.