Real/Expr Package Ver.1.0


There are 3 subdirectories:

1. src

     contains source codes for the Real/Expr package. To compile
these, you need g++ 2.7.2 with lib-g++ 2.7.2.  These have been tested
on Solaris 2.5.1 (SunOs 5.5.1) and SunOS 4.1.3.

     Real is a class for a variety of real number representations
(int, long, double, BigInt, Rational, BigFloat).  The first three
(int, long, double) are built-in machine types, BigInt and Rational
are big number representations based on GNU package.  BigFloat is our
multi-precision floating point package.

     Expr is a class built on top of Real.  It supports the following
operators and the function:

          +, (unary and binary) -, *, /, sqrt(),

the following exact comparisons:

          ==, !=, <, <=, >, >=,

and the function

          approx()

which approximates an instance of Expr to some specified
precision. For a discussion of relative and absolute precision, see
the paper or tutorial.


2. prog

     contains several sample test programs.

     The Makefile here assumes the sources for Real/Expr are in the
     relative path ../source.
	

3. doc

     contains UNIX man pages.


WEB SITE:      http://simulation.nyu.edu/projects/exact/
BUG REPORT:    ouchi@simulation.nyu.edu
LAST UPDATE:   Dec, 31, 1996

