When you write code that calls function len(), you don’t care how len() does what it does. You don’t stop to think about whether the code of the function contains two lines or two thousand lines, whether it uses one local variable or a hundred. You just need to know what argument to pass in and how to use the result that’s returned.

 

The same is true of functions that you write, like this function that calculates and returns the average of a list of numbers:

 

_____________________________________________________________

 

def calculateAverage(numbersList):

    total = 0.0

    for number in numbersList:

        total = total + number

    nElements = len(numbersList)

    average = total / nElements

    return average

 

___________________________________________________________

 

Once you’ve tested a function like this and found that it works, you no longer have to worry about the details of the implementation. You only need to know what argument(s) to send into the function and what it returns.

 

However, if one day you find that there is a much simpler or faster algorithm to calculate an average, you could rewrite the function in a new way. As long as the interface (the inputs and outputs) does not change, there is no need to change any calls to the function. This type of modularization makes the code more maintainable.