$Id: numeric_type.omh 2506 2012-10-24 19:36:49Z bradbell $
/* --------------------------------------------------------------------------
CppAD: C++ Algorithmic Differentiation: Copyright (C) 2003-12 Bradley M. Bell

CppAD is distributed under multiple licenses. This distribution is under
the terms of the 
                    Eclipse Public License Version 1.0.

A copy of this license is included in the COPYING file of this distribution.
Please visit http://www.coin-or.org/CppAD/ for information on other licenses.
-------------------------------------------------------------------------- */

/*
$begin NumericType$$
$spell
	const
	std
	valarray
$$

$index numeric, type$$
$index type, numeric$$

$section Definition of a Numeric Type$$

$head Type Requirements$$
A $icode NumericType$$ is any type
that satisfies the requirements below.
The following is a list of some numeric types:
$code int$$, $code float$$, $code double$$, 
$code AD<double>$$, $code AD< AD<double> >$$.
The routine $cref CheckNumericType$$ can be used to check
that a type satisfies these conditions.

$head Default Constructor$$
$index default, numeric constructor$$
$index constructor, numeric$$
The syntax 
$codei%
	%NumericType% %x%;
%$$
creates a $icode NumericType$$ object
with an unspecified value.

$head Constructor From Integer$$
$index int, numeric constructor$$
$index constructor, numeric$$
If $icode i$$ is an $code int$$,
the syntax 
$codei%
	%NumericType% %x%(%i%);
%$$
creates a $icode NumericType$$ object with a value
equal to $icode i$$ where $icode i$$ can be $code const$$.

$head Copy Constructor$$
$index copy, numeric constructor$$
$index constructor, numeric copy$$
If $icode x$$ is a $icode NumericType$$ object
the syntax 
$codei%
	%NumericType% %y%(%x%);
%$$
creates a $icode NumericType$$ object $icode y$$
with the same value as $icode x$$
where $icode x$$ can be $code const$$.

$head Assignment$$
If $icode x$$ and $icode y$$ are $icode NumericType$$ objects,
the syntax
$codei%
	%x% = %y%
%$$
sets the value of $icode x$$ equal to the value of $icode y$$
where $icode y$$ can be $code const$$.
The expression corresponding to this operation is unspecified; i.e.,
it could be $code void$$ and hence
$codei%
	%x% = %y% = %z%
%$$
may not be legal.

$head Operators$$
Suppose $icode x$$, $icode y$$ and $icode z$$
$icode NumericType$$ objects where
$icode x$$ and $icode y$$ may be $code const$$.
In the result type column,
$icode NumericType$$ can be replaced by any type that can 
be used just like a $icode NumericType$$ object.
$table
$bold Operation$$     $cnext $bold Description$$               $cnext
	$bold Result Type$$  $rnext
$codei%+%x%$$        $cnext unary plus                        $cnext
	$icode NumericType$$    $rnext
$codei%-%x%$$        $cnext unary minus                       $cnext
	$icode NumericType$$    $rnext
$icode%x% +  %y%$$  $cnext binary addition                   $cnext
	$icode NumericType$$    $rnext
$icode%x% -  %y%$$  $cnext binary subtraction                $cnext
	$icode NumericType$$    $rnext
$icode%x% *  %y%$$  $cnext binary multiplication             $cnext
	$icode NumericType$$    $rnext
$icode%x% /  %y%$$  $cnext binary division                   $cnext
	$icode NumericType$$    $rnext
$icode%z% += %y%$$  $cnext computed assignment addition       $cnext
	unspecified          $rnext
$icode%z% -= %y%$$  $cnext computed assignment subtraction    $cnext
	unspecified          $rnext
$icode%z% *= %y%$$  $cnext computed assignment multiplication $cnext
	unspecified          $rnext
$icode%z% /= %y%$$  $cnext computed assignment division       $cnext
	unspecified      
$tend

$head Example$$
$children%
	example/numeric_type.cpp
%$$
The file
$cref numeric_type.cpp$$
contains an example and test of using numeric types.
It returns true if it succeeds and false otherwise.
(It is easy to modify to test additional numeric types.)

$head Exercise$$
$index exercise, numeric type$$

$list number$$
List three operators that are not supported by every
numeric type but that are supported by the numeric types
$code int$$, 
$code float$$, 
$code double$$.

$lnext
Which of the following are numeric types:
$code std::complex<double>$$,
$code std::valarray<double>$$,
$code std::vector<double>$$ ?

$lend

$end