Student Originated Software 1997-1998 Fall Quarter
		A Software Engineering Course at The Evergreen State College

Object Oriented Programming About Double Dispatching Double dispatching is a useful technique for efficiently choosing an algorithm based on the classes of one or more of the arguments of a message as well as the class of the receiver. For example, we might wish to evaluate the following expression: 3 + 4.0 When the SmallInteger 3 receives the message #+, it may have to choose among algorithms designed for Integers, Fractions, and Floats. The fastest way to do this is for it to simply send the argument of the message another message--as shown here: + aNumber <primitive: 1> ^aNumber sumFromInteger: self If we replace variable names with values, this translates to: 4.0 sumFromInteger: 3 Since the float knows that the argument must be an Integer, it knows immediately what must be done to add itself to the integer. Note that the order of the numbers has changed, so that in methods such as Fraction>>quotientFromInteger:, it is the argument that is divided by the receiver, and not the receiver divided by the argument In the system as delivered, there are five operations that are subject to double-dispatching:  + (addition) For addition, the dispatch operator takes the form #sumFrom[a class name]:  - (subtraction) For subtraction, the dispatch operator takes the form #differenceFrom[a class name]:  * (multiplication) For multiplication, the dispatch operator takes the form #productFrom[a class name]:  / (division) For division, the dispatch operator takes the form #quotientFrom[a class name]:  < (comparison--the rest of the comparison operators can be derived) For comparison, the dispatch operator takes the form #lessFrom[a class name]: