Algorithm 680: Evaluation of the Complex Error Function
Poppe G. P. M., and Wijers C. M. J.
ACM Transaction on Mathematical Software, vol 16, N. 1, March 1990, p 47
_______________________________________________________________________________

This file has been added in order to compute numerically erf(z) and
erfc(z), for any complex argument z.

The current routine computes actually w(z), defined as :

                exp(-z^2) * erfc( -i*z)

We can deduce easily:

                w(i*z) = exp(z^2) * erfc(z)

so:

                erfc(z) = w(i*z) * exp(-z^2)

and:

                erf(z) = 1 - erfc(z)

The routine name has been changed, from WOFZ to DWOFZ as it is defined
for double precision.

É. Canot, 10 jan. 2012

