Question

I am trying to create a tree structure (a binary tree) that is capable of holding two different types of class (a sphere and a rectangle).

For obvious reasons my sphere and rectangle will have different methods for getting their size (getSize()) and I also intend to have a constructor (for both classes) thats takes two objects (two spheres OR two rectangles) and combines them to create a larger sphere or rectangle.

How should I approach coding a node so that it can store either a sphere or a rectangle at a node calling the appropriate methods when required?

Would a simple interface accomplish this if i cast objects to the type i need?

Thanks,

DMcB

Answer 1

I would create three classes.
An abstract class Shape that contains all common code for rectangles and spheres.

public abstract class Shape{
   // contains all common code related to shapes
   // such as child elements
   Shape parentNode; // This will help navigate up
   List<Shape> children; // This will help navigate down the tree
   // Define, merge, split methods which are common to all shapes
   // define shape specific methods
}
public class Rectangle : Shape{
   // Implement shape's abstract methods for this class
}

public class Sphere : Shape{
   // Implement shape's abstract methods for this class
}

All shape specific methods should be left abstract such as getSize(), drawShape(), mergeShape().
Also, maybe not relevant but the Composite design pattern might be of good use for this problem