The testability of a method, 
is dependent on the visiblity component of the
method. We define the testability of a method in a class to be:
To make the testability estimate of a method a true indicator of the effort needed to
test the method, we take the constant to be the number of unique objects
that participate in the calculation of the metric. That is, we count the
objects used as explicit input parameters, implicit input parameters,
return objects and exception objects. We do not include the implicit
output objects nor the pass by reference, explicit parameters in the count.
This constant is a reasonable choice because its value indicates the
maximum number of objects whose state may have been changed incorrectly.