Document revision date: 19 July 1999 | |

Previous | Contents | Index |

This glossary defines mathematical terms and symbolic names used in this manual.

**complex number:** See F_COMPLEX.

**denormalized number:** A floating-point number with a
value very close to zero.

**domain error:** An exception condition resulting from
passing an argument whose value is outside the range of permissible
values.

**exceptional argument:** Any argument value passed to a
DPML routine that does not return a meaningful result, or an argument
defined differently for different platforms.

**F_COMPLEX:** A complex number identifier. F_COMPLEX
indicates that a given routine returns two different values of the same
floating-point data type. See Table 1-2 for more information.

**F_TYPE:** A floating-point number identifier. F_TYPE is
used when it is not necessary to distinguish between the floating
types. See Table 1-1 for more information.

**floating-point number:** See F_TYPE.

**HUGE_RESULT:** For VAX data types, HUGE_RESULT =
max_float.

For IEEE data types, HUGE_RESULT = infinity.

**INV_RESULT:** For VAX data types, INV_RESULT = 0.

For IEEE data types, INV_RESULT = NaN.

**invalid argument:** See domain error.

**max_float:** The largest finite number representable in
the floating-point data types. See Appendix A for more information
on max_float values.

**min_float:** The smallest positive normalized nonzero
number representable in the floating-point data types. See
Appendix A for more information on min_float values.

**NaN:** A floating-point value that is said to be
"not a number" and contains an indeterminate quantity.

**overflow:** An exception condition caused by passing a
floating-point value that is larger than the highest valid
floating-point value. See max_float for additional information.

**range error:** An exception condition that occurs when a
mathematically valid argument results in a function value that exceeds
the range of representable values for floating-point data types.

**underflow:** An exception condition caused by passing a
floating-point value that is lower than the lowest valid floating-point
value. See min_float for additional information.

Index | Contents |

privacy and legal statement | ||

6118PRO_005.HTML |