MathGroup Archive: December 1999 [00025]

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

Bug? in PolyLog

To: mathgroup at smc.vnet.net
Subject: [mg20925] Bug? in PolyLog
From: fgrieu at micronet.fr (Francois Grieu)
Date: Wed, 1 Dec 1999 01:49:56 -0500 (EST)
Sender: owner-wri-mathgroup at wolfram.com

While trying to fix my Dickman function, I uncovered what appears
to be a bug with the numerical implementation of PolyLog.
It gives an imaginary part for PolyLog[2,x] for x>1
when PolyLog[2,x] is purely real in this domain.

Here is a session:

Mathematica 4.0 Kernel for Power Macintosh
Copyright 1988-1999 Wolfram Research, Inc.
 -- Terminal graphics initialized -- 

In[1]:= PolyLog[2, 2.0]

Out[1]= 2.4674 - 2.17759 I

In[2]:= Simplify[Im[PolyLog[2, 2]]]

Out[2]= 0


Any clues ? I am interested by what happens for other releases
of Mathematica.

   Francois Grieu

Prev by Date: Can I slove Delay-differential equations with mathematica?

Next by Date: Checking for non-complex numerics

Previous by thread: Re: Can I slove Delay-differential equations with mathematica?

Next by thread: Re: Bug? in PolyLog