The IRAF math library provides a collection of general purpose numerical
routines for use in scientific applications programs.  This library is
continually growing as we collect numerical packages from around the country.

At present the library is just a collection of untested routines from various
sources.  As time goes on, these routines will be edited to conform to a
standard which has not yet been developed.  The standard is expected to provide
for different classes of routines, according to how well the routines have
been integrated into the system and tested.  The standard for the math library
will affect programming style, naming conventions, documentation, and the
subset of Fortran which portable library routines are expected to comply with.

A fundamental part of the standard is that math library routines should not do
i/o.  Some of the otherwise numerical routines in these directories print
error messages when an error occurs, using Fortran i/o.  This is bad because
Fortran i/o is not used elsewhere in the IRAF system, and because it is
preferable for a numerical routine to return an error code in the event of an
error.  Before these codes are used in IRAF applications programs, this will
have to be fixed.