erf - The error function.
erf computes the error function:
/x 2
2 | -t
y = ------- | e dt
sqrt(pi) /0
deff('y=f(t)','y=exp(-t^2)');
erf(0.5)-2/sqrt(%pi)*intg(0,0.5,f)
erfc , erfcx , calerf , cdfnor ,
W. J. Cody (code from Netlib (specfun))