Fuzzy equality comparison                                      [CRC test suite]
-------------------------

	^asserteq^ ^tol^erance { # | ^default^ }
	^asserteq^ { ^noi^sily | ^qui^etly }
	^asserteq^ <exp1> ^==^ <exp2> [^noi^sily|^qui^etly]
	^asserteq^ <exp1> ^/=^ <exp2> [^noi^sily|^qui^etly]
	^asserteq^ <exp1> ^//^ <exp2> [^noi^sily|^qui^etly]

	^vserteq^  <exp1> ^==^ <exp2>

Description
-----------

^asserteq tolerance^ sets the tolerance.  The tolerance is ^default^ at the
beginning of a test or after ^macro drop _all^.

^asserteq^ {^noisily^|^quietly^} sets whether the observed difference is shown.  The
default is quietly at the beginning of a test or ^macro drop _all^.

^asserteq^ <exp1> ^==^ <exp2> asserts

		|<exp1> - <exp2>|  <  tolerance

^asserteq^ <exp1> ^/=^ <exp2> asserts

		|<exp1> - <exp2>|
		-----------------  <  tolerance
		max( 1, |<exp2>| )

^asserteq^ <exp1> ^//^ <exp2> asserts

		|<exp1> - <exp2>|
		-----------------  <  tolerance
		     |<exp2>|

The ^noisily^ and ^quietly^ "options" temporarily override the the ^noisily^ and
^quietly^ settings.


^vserteq^ <exp1> ^==^ <exp2> is equivalent to ^asserteq^ <exp1> ^==^ <exp2> except that
it reports the maximum absolute difference.  ^asserteq^ would report the dif-
ferences for the first observation.  ^vserteq^ is always noisy.  There is no
reason to use the computationally more expensive ^vserteq^ if you do not want
to see the report.

Features
--------

^asserteq^ is implemented in terms of ^assert^; therefore, <exp1> and <exp2> may
be scalar or vector expressions.  ^noisily^, however, is useless in the vector
case because it prints the the results for the first observation only.


Bugs
----

You must put blanks around ^==^, ^/=^, and ^//^.  If <exp1> or <exp2> contains
blanks, you must bind using double quotes.

^vserteq^ sorts the data.  It returns the data in the same order, but with the
sort markers cleared.  (This could be fixed for Stata 3.1.)


Also see
--------

On-line:  ^help^ for ^testproc^