Bug in gamma-inc implementation and documentation

[williams]

Noel Welsh send the following:

The gamma-inc function in the science collection has an error:

In gamma-inc-impl.ss it reads:

(define (gamma-inc-F-CF a x)

(define nmax 5000)
(define small (exp double-epsilon 3))
...

That last function call should be expt, not exp.

The documentation for gamma-inc states

"Computes the normalized incomplete gamma function"

but the equation below that line corresponds to the *un*normalised
function. I compared the corrected function against Matlab and believe
it is correctly implementing the unnormalised function.

[williams]

Fixed and released to PLaneT in version 3.6.