(1) The specification that causes the floating-point value to be stored internally in the short format. See also double precision.
(2) The use of one computer word to represent a number, in accordance with the required precision.