I think this is particularly poorly named, but here is what I get from the
"help" system in Maple version 7...
Calling Sequence
fnormal(e)
fnormal(e, digits)
fnormal(e, digits, epsilon)
Parameters
e - an algebraic expression, or a list, set, relation, series, or
range of algebraic expressions
digits - (optional) number of digits for floating-point evaluation
(defaults to the value of the global variable Digits)
epsilon - (optional) error tolerance for ``fuzzy zero'' (defaults to
the value Float(1,-digits+2))
Description
The value returned by fnormal is an expression equivalent to e under the
assumption that all numeric values with magnitude less than epsilon may be
considered to be zero.
In addition, all floats in e which remain nonzero are converted to floats
with digits precision.
fnormal preserves numeric type and sign information as much as possible.
Thus, for example, fnormal(1e-20*I, 10) = 0.*I not 0. . This ensures that
branching behaviour is generally not affected by fnormal. Use simplify(expr,
zero) to remove 0 real or imaginary parts of complex floating point numbers.
If e is a list, set, range, series, equation, or relation, then fnormal is
applied recursively to the components of e.