RE: HP48-HP50G : Error functions ERF(z), FRESNEL integrals, PROBIT, etc.

[Gil]

Version 9b

Now inverse functions of the error function ERF and ERFc,
called respectively INV¦ & INV¦c,
admit as argument p a complex number p = a+ib, b≠0.

For the argument p of INV¦(p):
IpI < 1.

For the argument p of INV¦c(p):
I1-pI < 1.

On that purpose, a new subroutine called 'inv¦z' was used.
Better here to keep all the 50 terms of the numerator and of the denominator.
To save memory, however, you can use (approximate) real numbers, instead of the (exact) integers, in the 2 {} lists of the above mentioned subroutine 'inv¦z'.

Regarding PROBIT function
A slight change in the program, but the argument of PROBIT function was kept to remain a real number p in [0 1].