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

Re: Unexpected non-evaluation problem

• To: mathgroup at smc.vnet.net
• Subject: [mg63180] Re: Unexpected non-evaluation problem
• From: David Bailey <dave at Remove_Thisdbailey.co.uk>
• Date: Fri, 16 Dec 2005 07:22:48 -0500 (EST)
• References: <dnrch3$kie$1@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

Hello,

I don't think this is anything to do with the mathematical properties of 
the fourier transform - I think it is a bug caused by the fact that one 
  transform is nested inside another. If you remove the ReadProtected 
attribute on InverseFourierTransform and spill out all the top-level 
code for the operation, you find that some cases of the function are 
conditional on System`FourierTransformDump`$FourierExptransform - a 
global variable - so maybe the two levels interfere with each other.

The interesting thing is that you end up with an expression that 
Mathematica thinks is fully evaluated when in fact it is not! That is 
normally pretty hard to do.

BTW, I assume that if WRI did not want people to read the code for 
internal functions such as InverseFourierTransform, they would have 
prevented users from clearing the ReadProtected attribute!


David Bailey
http://www.dbaileyconsultancy.co.uk