Floating_ Point Numbers.

  1. The possible range of floating_ point numbers is machine dependent. At a minimum, the values will have 5 place accuracy, with a range of approximately 1038.
  2. A floating_ point constant is a sequence of one or more digits, followed by a period, ``

    .

    '', followed by one or more digits. Thus,

    2.0

    is legal, and

    2.

    and

    .5

    are illegal.

    It may be followed by an exponent. An exponent consists of one of the characters ``

    e

    '', ``

    E

    '', ``

    f

    '', ``

    F

    '' followed by a signed or unsigned integer. The value of a floating_ point constant is determined as is usual with scientific notation. As with integers, it is unsigned. Hence, for example,

    0.2

    ,

    2.0e - 1

    ,

    20.0e - 2

    are all equivalent.
  3. Different systems use different printed representations when floating point values are out of the machine's range. For example, when the value is too large, the Mac prints ``+++++'' and the Sun prints "Infinity".