Avoid Common Object Oriented Programming Pitfalls

What will we cover in this tutorial?

I did them myself. When you first learn about object oriented programming, you get exited about it and want to turn everything into objects.

Very often it fails to make the code easier to maintain and understand. Yes, that is one core goal of creating object oriented programming, to make it easier to maintain and understand.

Often we just turn the traditional way of programming into objects. But that is not the point. The point is to model it in an object oriented way.

Before that, let’s just look at why we use object oriented programming.

Step 1: Why use object oriented programming?

When introducing object oriented programming to beginners it often introduces too many concepts at once.

Simply because object oriented programming is awesome and can do so many things.

Let’s just start simple with the core of the object oriented programming idea.

We want to make your program easier to understand and maintain.

That is the goal for any programmer. Also, when it comes to object oriented programming.

Object oriented programming tries to make the link between the real world and the programming world as close as possible. We humans understand the objects in the real world better than we understand how a computer pushes bits and bytes around in the memory and CPU.

Luckily, we do not need to understand all that. We just need to understand how to program the computer through a programming language.

Object oriented programming tries to make that easier with modeling the programs with objects, which are related to the way we humans understand things.

Let’s try with a simple example of a Stack.

A Stack has a top and bottom, it has operations push and pop.

How would you model the above without using object oriented programming.

stack_size = 8
stack = [None]*stack_size
top = -1 # empty stack

def pop():
    global top, stack
    if top == -1:
        return None
    else:
        top -= 1
        return stack[top + 1]

def push(element):
    global top, stack, stack_size
    if top + 1 >= stack_size:
        stack += [None]*stack_size
        stack_size *= 2
    top += 1
    stack[top] = element

def print_stack():
    global top, stack, stack_size
    for i in range(top + 1):
        print(stack[i], '', end='')
    print(" (size: " + str(stack_size) + ")")


print_stack()
for i in range(10):
    push(i)
print_stack()
for i in range(8):
    pop()
print_stack()

That is confusing code, right?

First of all, we use global variables. That makes the function calls hard to understand, as they have side effects.

This is made more confusing than necessary. It does use Python lists like an Array. That is not needed, but it is just to exemplify how difficult it is to make intuitive code if you model the world like a computer works.

Step 2: Using object oriented programming to solve the above (first step – the less bad, but still not good solution)

First time you are asked to create stack using an object oriented approach, you will probably do it in a non intuitive way.

Say, you think of a Stack like a object. The stack is the full object.

What does a stack consists of?

Well items which are piled on top of each other, and you can take the top off.

How could that be modeled?

Often the straight forward way from a classical thinking way into the a class.

class Stack:
    def __init__(self):
        self.stack_size = 8
        self.stack = [None]*self.stack_size
        self.top = -1 # empty stack

    def pop(self):
        if self.top == -1:
            return None
        else:
            self.top -= 1
            return self.stack[self.top + 1]

    def push(self, element):
        if self.top + 1 >= self.stack_size:
            self.stack += [None]*self.stack_size
            self.stack_size *= 2
        self.top += 1
        self.stack[self.top] = element

    def print_stack(self):
        for i in range(self.top + 1):
            print(self.stack[i], '', end='')
        print(" (size: " + str(self.stack_size) + ")")


s = Stack()
s.print_stack()
for i in range(10):
    s.push(i)
s.print_stack()
for i in range(8):
    s.pop()
s.print_stack()

If you inspect the code, it is actually the same code. Which in some ways has improved it.

  1. We do not have global variables anymore. They are tied to the Stack class.
  2. The function calls push and pop are also tied to the Stack class.
  3. Also, when we use the Stack, it is clear from the context, as it is tied to the variable s, in this case.

So what is wrong?

Well, it is still not simple to understand what happens in the code. It takes time to understand, even with this simple code.

The functions use variables like top and assigns it to -1 if the stack is empty. It requires you to investigate pop and the constructor to understand that.

The key is to keep it simple.

So how to do that?

Step 3: A simple way to model it

Let’s take a look at the drawing again.

A stack.

A stack actually consists of object on it. Hence, the stack is the abstraction that keeps objects in a specific order.

That said, we need to model that closer, as it will be easier to understand for the reader of the code.

The objects on the stack we will call Nodes.

What characterizes a Node? It lies either on the bottom or on top of another Node.

What characterizes a Stack? It knows the top of the stack and can push and pop Nodes.

How can that be turned into code?

class Node:
    def __init__(self, element=None, next_node=None):
        self.element = element
        self.next_node = next_node


class Stack:
    def __init__(self):
        self.top = None

    def push(self, element):
        self.top = Node(element, self.top)

    def pop(self):
        element = self.top.element
        self.top = self.top.next_node
        return element


s = Stack()
for i in range(20):
    s.push(i)
for i in range(10):
    print(s.pop(), '', end='')
print()

How is that code? Easier to understand?

Of course, normally a Stack would as a minimum have a helper function is_empty() to return if stack is empty. That can be added easily. You see how?

class Node:
    def __init__(self, element=None, next_node=None):
        self.element = element
        self.next_node = next_node


class Stack:
    def __init__(self):
        self.top = None

    def push(self, element):
        self.top = Node(element, self.top)

    def pop(self):
        element = self.top.element
        self.top = self.top.next_node
        return element

    def is_empty(self):
        return self.top == None


s = Stack()
for i in range(20):
    s.push(i)
while not s.is_empty():
    print(s.pop(), '', end='')
print()

Do you get the sense of it now?

Model the code as closely to the reality we humans understand. It makes the code easy to read and understand. Hence, easy to maintain.

Leave a Reply