Summary of Lecture for June 22, 1999

	Topics:  when to use inheritance and when to use composition,
		uses of protected and final

	Composition and inheritance are two techniques for reusing 
code.  Each has advantgaes and disadvantages when compared to the other.
	Advantages of Composition over Inheritance:
1.  The new class does not have to support the interface of the object
being reused by being a data field of this class's objects.  The data
field can be private, protected, or package visibility and thus you can
restrict where that object can be accessed. You could make the field 
public, if you wish any client of your new class to be able to access the
composed object's interface (its methods).
2.  The composed data field is an object variable and can thus be 
assigned new objects during the execution of the new class's objects.
Using the upcasting discussed later today, the new objects can be of 
different derived classes of the base class that is the data field's 
declared type.  If this is done carefully, the data field can thus be 
used to support different semantics ( take on different roles) at 
different times during execution.

	Advantages of Inheritance Over Composition:
1.  Objects of the new class can act as if they were objects of the base
class using the upcasting to be described later today.
2.  Methods in the derived class can be automatically called when the user
calls a method of the base class using object-oriented polymorphism as 
discussed tomorrow.
3.  Inheritance supports incremental development.

	These advantages indicate which technique should be used when:
1.  Composition should be used when you want to hide your use of an 
object of a given class from clients of your class;
2.  Composition should be used when you want to take on different roles 
in the same object ( the parent object that contains the composed object
variable) at different times during a single execution without changing 
your code or using conditional statements.
3.  Inheritance should be used when you want to take advantage of either
upcasting or polymorphism.  The most common such usage is when you want
the different items in an array or a collection ( to be covered July 1st)
to come from different derived types of a common base class.
4.  Inheritance should you be used when you want objects of the new class
to implement the same interface as the base class.

	 When should private, protected, public and package visibility
modifiers be used?  The general guidelines are:
1.  object data items should almost never be public.
2.  class data items should almost never be public.
3.  object methods should be public, if they provide a service to the 
clients of your object.  Otherwise, usually they should be protected.
4.  class methods should be protected or package in most circumstances.
5.  class data items should usually be package visibility and sometimes
protected.
6.  object data items should usually be protected visibility and 
sometimes package.  
7.  For final classes, protected above can be replaced by private.

	Upcasting is one of the two foundations of object-oriented 
programming.  The other foundation, object-oriented polymorphism will be 
discussed tomorrow.  Upcasting means that objects of a derived class can
serve as if they were objects of their base class.  This allows one to
declare a base class object variable and assign to it an object of a 
derived class.  Further, whenever a parameter or return value is supposed
to be an object or object variable of a base class, you can use an object
or object variable of the derived class instead.
	Note that objects of the derived class are not actually objects
of the base class.  In particular, derived class objects cannot access
the private data and methods of base class objects.

	We discussed various uses of the final modifier.  final generally
means the modified item cannot change its value once a value has been 
assigned.  static generally means there will be one copy of the modified
item that will have storage associated with it from the moment the 
container ( class or object) is created until that container is no longer
accessible.
	A variable of a primitive type can have the modifier final.  This
use of final indicates that variable is actually a constant.  It must be
assigned a value at the point of definition. If final and static are 
used as modifiers, the item has only one copy and that copy cannot be 
changed.
	If final is used on an object variable, the handle associated 
with that value cannot change, but the actual object referenced by that
handle can have its data fields change value.  Arrays work the same way.
An array acts just like an object except it has no class.
	If final is used on a method, it indicates that method cannot be
overridden in any inheriting class.  This use of final may result in more
efficient code.
	If final is used on a parameter in the called method's header, 
that parameter cannot be changed by the called method.
	If final is applied to a class, that class cannot serve as a 
base class.