Converts this value to Double.
Finite Float values and signed zeros are represented exactly; infinities and NaN are retained.
the floating-point representation.