INTERFACE Real;
IMPORT Word;

This package defines some basic properties of the built-in float type REAL.

Index: REAL; floating-point; generics

TYPE T = REAL;

CONST
  Base: INTEGER = 2;
  (* The radix of the floating-point representation for T *)

  Precision: INTEGER = 24;
  (* The number of digits of precision in the given Base for T. *)

  MaxFinite: T = 3.40282347E+38;
  (* The maximum finite value in T.  For non-IEEE implementations,
     this is the same as LAST(T). *)

  MinPos: T = 1.40239846E-45;
  (* The minimum positive value in T. *)

  MinPosNormal: T = 1.17549435E-38;
  (* The minimum positive "normal" value in T; differs from MinPos
     only for implementations with denormalized numbers. *)

CONST
  MaxExpDigits = 2;
  MaxSignifDigits = 9;

MaxExpDigits is the smallest integer with the property that every finite number of type T can be written in base-10 scientific notation using an exponent with at most MaxExpDigits. MaxSignifDigits is the smallest integer with the property that floating-decimal numbers with MaxSignifDigits are more closely spaced, all along the number line, than are numbers of type T. Typically,

      MaxExpDigits    = ceiling(log_10(log_10(MaxFinite)))
      MaxSignifDigits = ceiling(log_10(Base^Precision)) + 1.

CONST Brand = "Real";

PROCEDURE Equal(a, b: T): BOOLEAN;

Return a = b. The result is undefined if either a or b is an NaN (not a number) value.

PROCEDURE Hash(a: T): Word.T;

Return a hash value derived from a. The result is undefined if either a or b is an NaN (not a number) value.

PROCEDURE Compare(a, b: T): [-1..1];

Return -1 if a < b, 0 if a = b, or +1 if a > b. The result is undefined if either a or b is an NaN (not a number) value.

END Real.