Not logged in

A class may inherit another class. This means that the inheriting class (also called subclass or derived class) starts with all the methods and member variables that the inherited class, (also called superclass or base class) has. The subclass can then have its own, additional methods and member variables.

One situation when inheritance can be useful is when you want to create two or more classes that have a common part. Birds and fishes, for example, are different, with different characteristics, but they do have much in common. You can make use of this by creating a class, for example called "animal", with the common properties. Then you define the two subclasses "bird" and "fish", which inherit from "animal".

At other times you already have a class that almost does what you want it to, but you would like to add something to it. For example, a class "connection", which models an Internet connection, may have everything you need except for a time limit on how long you can be connected, You could then create a new class, "restricted_connection", which inherits from the old connection class, but with the time limit added.

In both of these situations, we have what is sometimes called an is-a relationship: a bird is an animal, a restricted_connection is a connection. We recommend that you use inheritance in this way: to model is-a relationships.

You use the keyword inherit to let a class inherit from another class. For example, to create the sub-classes bird and fish, which both inherit from animal, you would write:

class bird
  inherit animal;
  float max_altitude;

  void fly()
    write(name + " flies.\n");

  void eat(string food)
    write(name + " flutters its wings.\n");

class fish
  inherit animal;
  float max_depth;

  void swim()
    write(name + " swims.\n");

A bird like Tweety can do anything an animal can do, and it has all the data that an animal has. But it can also fly (the method fly), and it has a maximum altitude (the member variable max_altitude).

Note that the class bird has its own method called eat. There was one in animal too, but the new one overrides the old one, and will be used in all bird objects. If you have a method in the subclass with the same name as a method in the superclass, the module in the subclass hides or overrides the method in the in superclass.

If you still want to call the method in the superclass, you can prefix the name with two colons (::). That is what is done in the eat method: after fluttering its wings at the sight of the food, the bird will do the actual eating, and that is done with a call to ::eat.

You can now use our two new classes:

bird tweety = bird("Tweety", 0.13);
tweety->max_altitude = 180.0;

fish b = fish("Bubbles", 1.13);
b->eat("fish food");

animal w = fish("Willy", 4000.0);

One thing that needs explaining is the last line in the example above:


The variable w is of type animal, and that class has no method called swim. But that doesn't matter, since Pike always looks at the object that is stored in the variable. In this case, Pike looks at the contents of the variable w, finds that it is a fish, and then calls the method swim in that object. Looking at the actual object like this is called dynamic binding. (The opposite, to just look at the type of the variable and ignore what's actually in it, would be called static binding.)

As we said, we used inheritance to express is-a relationships. But there are other ways of using inheritance, for example to simply get access to some functionality. If you write a program that needs to work with a file on your hard disk, we could inherit the file-handling class Stdio.File, and then use all the methods in that class as if you had written them in your own program:

inherit Stdio.File;
// ...

This works, but we recommend that you create an object of the type Stdio.File instead, and call the methods for that object:

Stdio.File the_file;
// ...