> Or using pi = 2 * asin(1.0). Perhaps likelier to give an accurate value.
Or why not
pi = 3.141592653589793238
(or more decimals if you care)
The Haskell standard guarantees that you'll get the floating point
numbers that is closest to the number you wrote. That's why literals
with decimal point are represented with rational numbers and then
carefully converted.
(If you don't do it that way, you're not implementing Haskell. :-)
-- Lennart
