MathGroup Archive 2007

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

Search the Archive

Re: NIntegrate with change of variables (discovering also a bug in


Hey Dimitris,  That's fantastic!  However I'll run into the same problem regarding the Log function I described in another post here:  Take for example the contour for the Zeta function. On the lower leg of the contour, if I were to use:

ContourNIntegrate[z^{s-1}/(Exp[-z]-1),z->rExp[-pi i],{r,infinity,1}]

Mathematica would interpret z^(s-1) as:

z^(s-1)=Exp[(s-1)(Log[r]+pi i)] and not:


z^{s-1)=Exp[(s-1)(Log[r]-pi i)] as is required.  Any way we could configure Mathematica to make that substitution?


  • Prev by Date: Re: Re-defining Log over it's branch cut
  • Next by Date: Re: Trouble with TraditionalForm in Mathematica 6.0
  • Previous by thread: Re: questions
  • Next by thread: simplification of 0/0 to 1?