Re: error function erf(z) with complex z

[Victor Munoz <vmunoz@macul.ciencias.uchile.cl>]

On Wed, Sep 07, 2005 at 12:03:28PM +0200, Thomas Weber wrote:
> Hello, 
> 
> > >>> from scipy.special import erf
> > >>> erf(2+3j)
> > (-20.829461428577808+8.6873182710560588j) 
> Yes, thanks. Unfortunately, the library used for this is Fortran, so I'm
> still open for other suggestions.
> 
> 

It is not in Debian, but Matpack has an implementation for the Faddeeva
function:

 w(z) = exp(-z^2)*erfc(-i*z),
 
 where erfc is the complementary error function.
 
I extracted the "stand-alone" code, containing that function and adding the 
plasma zeta function Z(z) = i*sqrt{pi} w(z), in which I was really interested.

I made C++ (Matpack is written in C++) and Octave versions, but I don't know
if it's ok to redistribute it, as it is basically the same code, with a few
lines commented to make faddeeva(z) independent of the rest of Matpack code.
>From the copyright notice:

"The software may be modified for your own purposes, but modified versions
may not be distributed without prior consent of the author."

Does extracting a single function from the code constitute a "modified
version"? What about translating an algorithm to another language?

Regards,
					Victor