Determining the sign of a number

[Samuel A. Figueroa]

I am wondering why the IEEE Standard for Floating-Point Arithmetic does not
recommend that a function be provided to determine the sign of a given value.
Is it perhaps because a function that copies the sign from one number to
another is more general, so that if this function is provided as recommended
in the appendix, a more specific function that only determines the sign of
a given value isn't needed?  If the answer is yes, is there enough reason to
also provide this more specific function?  (One would hope that this more
specific function would incur less overhead than the more general one.)
Sam Figueroa
figueroaacs.nyu.edu