This package defines a type for binary
fixed-precision arithmetic. The main differences
between this and Data.Fixed are that this is
binary fixed-point and it's polymorphic in the
underlying representation.
When is this more appropriate than floating
point? You'll mainly want to use this when you
need to be able to represent fractional values
within a bounded range. Fixed-point numbers have
the advantage of uniformity in these cases. On
the downside, you lose precision relative to
floating point numbers as you approach zero, and
you lose the ability to express very large (but
imprecise) values that floating point can
express. On some architectures, fixed-point
arithmetic might be faster than floating-point
arithmetic, but this is probably not the case on
x86.