Let's remind you about the concept of pointers as we will be dealing with them in this chapter. To begin with, imagine that you have a house that you want to sell. Lacking time, you contact an agent to find interested buyers. So, you pick up your house and take it over to the agent, who will, in turn, carry the house to anybody who may want to buy it. Ludicrous, you say? Now imagine that you have a few Python functions that work with images. So, you pass high-resolution image data between your functions.

Of course, you don't carry your house around. What you do is write the address of the house down on a piece of scrap paper and hand it over to the agent. The house remains where it is, but the note containing the directions to the house is passed around. You might even write it down on several pieces of paper. Each one is small enough to fit in your wallet, but they all point to the same house.

As it turns out, things are not very different in Python land. Those large image files remain in one single place in memory.

What you do is create variables that hold the locations of those images in memory. These variables are small and can easily be passed around between different functions.

That is the big benefit of pointers—they allow you to point to a potentially large segment of memory with just a simple memory address.

Support for pointers exists in your computer's hardware, where it is known as indirect addressing.

In Python, you don't manipulate pointers directly, unlike in some other languages, such as C or Pascal. This has led some people to think that pointers aren't used in Python. Nothing could be further from the truth. Consider this assignment in the Python interactive shell:

>>> s = set()

We would normally say that s is a variable of the set type. That is, s is a set. However, this is not strictly true; the variable s is rather a reference (a safe pointer) to a set. The set constructor creates a set somewhere in memory and returns the memory location where that set starts. This is what gets stored in s. Python hides this complexity from us. We can safely assume that s is a set and that everything works fine.