1. Object Oriented Programming

By Bernd Klein. Last modified: 01 Feb 2022.

Though Python is an object-oriented language without fuss or quibble, we have so far intentionally avoided the treatment of object-oriented programming (OOP) in the previous chapters of our Python tutorial. We skipped OOP, because we are convinced that it is easier and more fun to start learning Python without having to know about all the details of object-oriented programming.

Even though we have avoided OOP, it has nevertheless always been present in the exercises and examples of our course. We used objects and methods from classes without properly explaining their OOP background. In this chapter, we will catch up on what has been missing so far. We will provide an introduction into the principles of object oriented programming in general and into the specifics of the OOP approach of Python. OOP is one of the most powerful tools of Python, but nevertheless you don't have to use it, i.e. you can write powerful and efficient programs without it as well.

Though many computer scientists and programmers consider OOP to be a modern programming paradigm, the roots go back to 1960s. The first programming language to use objects was Simula 67. As the name implies, Simula 67 was introduced in the year 1967. A major breakthrough for object-oriented programming came with the programming language Smalltalk in the 1970s.

You will learn to know the four major principles of object-orientation and the way Python deals with them in the next section of this tutorial on object-oriented programming:

• Encapsulation
• Data Abstraction
• Polymorphism
• Inheritance

Before we start the section about the way OOP is used in Python, we want to give you a general idea about object-oriented programming. For this purpose, we would like to draw your attention to a public library. Let's think about a huge one, like the "British Library" in London or the "New York Public Library" in New York. If it helps, you can imagine the libraries in Paris, Berlin, Ottawa or Toronto* as well. Each of these contain an organized collection of books, periodicals, newspapers, audiobooks, films and so on.

Generally, there are two opposed ways of keeping the stock in a library. You can use a "closed access" method that is the stock is not displayed on open shelves. In this system, trained staff brings the books and other publications to the users on demand. Another way of running a library is open-access shelving, also known as "open shelves". "Open" means open to all users of the library not only specially trained staff. In this case the books are openly displayed. Imperative languages like C could be seen as open-access shelving libraries. The user can do everything. It's up to the user to find the books and to put them back at the right shelf. Even though this is great for the user, it might lead to serious problems in the long run. For example some books will be misplaced, so it's hard to find them again. As you may have guessed already, "closed access" can be compared to object oriented programming. The analogy can be seen like this: The books and other publications, which a library offers, are like the data in an object-oriented program. Access to the books is restricted like access to the data is restricted in OOP. Getting or returning a book is only possible via the staff. The staff functions like the methods in OOP, which control the access to the data. So, the data, - often called attributes, - in such a program can be seen as being hidden and protected by a shell, and it can only be accessed by special functions, usually called methods in the OOP context. Putting the data behind a "shell" is called Encapsulation. So a library can be regarded as a class and a book is an instance or an object of this class. Generally speaking, an object is defined by a class. A class is a formal description of how an object is designed, i.e. which attributes and methods it has. These objects are called instances as well. The expressions are in most cases used synonymously. A class should not be confused with an object.

OOP in Python

Even though we haven't talked about classes and object orientation in previous chapters, we have worked with classes all the time. In fact, everything is a class in Python. Guido van Rossum has designed the language according to the principle "first-class everything". He wrote: "One of my goals for Python was to make it so that all objects were "first class." By this, I meant that I wanted all objects that could be named in the language (e.g., integers, strings, functions, classes, modules, methods, and so on) to have equal status. That is, they can be assigned to variables, placed in lists, stored in dictionaries, passed as arguments, and so forth." (Blog, The History of Python, February 27, 2009) In other words, "everything" is treated the same way, everything is a class: functions and methods are values just like lists, integers or floats. Each of these are instances of their corresponding classes.

x = 42
type(x)

int

y = 4.34
type(y)

float

def f(x):
    return x + 1
type(f)

function

import math
type(math)

module

One of the many integrated classes in Python is the list class, which we have quite often used in our exercises and examples. The list class provides a wealth of methods to build lists, to access and change elements, or to remove elements:

x = [3,6,9]
y = [45, "abc"]
print(x[1])

6

x[1] = 99
x.append(42)
last = y.pop()
print(last)

abc

The variables x and y of the previous example denote two instances of the list class. In simplified terms, we have said so far that "x and y are lists". We will use the terms "object" and "instance" synonymously in the following chapters, as it is often done**.

pop and append of the previous example are methods of the list class. pop returns the top (or you might think of it as the "rightest") element of the list and removes this element from the list. We will not explain how Python has implemented lists internally. We don't need this information, because the list class provides us with all the necessary methods to access the data indirectly. That is, the encapsulation details are encapsulated. We will learn about encapsulation later.

Live Python training

Enjoying this page? We offer live Python training courses covering the content of this site.

See our Python training courses

See our Machine Learning with Python training courses

A Minimal Class in Python

We will design and use a robot class in Python as an example to demonstrate the most important terms and ideas of object orientation. We will start with the simplest class in Python.

class Robot:
    pass

We can realize the fundamental syntactical structure of a class in Python: A class consists of two parts: the header and the body. The header usually consists of just one line of code. It begins with the keyword "class" followed by a blank and an arbitrary name for the class. The class name is "Robot" in our case. The class name is followed by a listing of other class names, which are classes from which the defined class inherits. These classes are called superclasses, base classes or sometimes parent classes. If you look at our example, you will see that this listing of superclasses is not obligatory. You don't have to bother about inheritance and superclasses for now. We will introduce them later.

The body of a class consists of an indented block of statements. In our case a single statement, the "pass" statement.

A class object is created, when the definition is left normally, i.e. via the end. This is basically a wrapper around the contents of the namespace created by the class definition.

It's hard to believe, especially for C++ or Java programmers, but we have already defined a complete class with just three words and two lines of code. We are capable of using this class as well:

class Robot:
    pass
if __name__ == "__main__":
    x = Robot()
    y = Robot()
    y2 = y
    print(y == y2)
    print(y == x)

True
False

We have created two different robots x and y in our example. Besides this, we have created a reference y2 to y, i.e. y2 is an alias name for y.

Attributes

Those who have learned already another object-oriented language, must have realized that the terms attributes and properties are usually used synonymously. It may even be used in the definition of an attribute, like Wikipedia does: "In computing, an attribute is a specification that defines a property of an object, element, or file. It may also refer to or set the specific value for a given instance of such."

Even in normal English usage the words "attribute" and "property" can be used in some cases as synonyms. Both can have the meaning "An attribute, feature, quality, or characteristic of something or someone". Usually an "attribute" is used to denote a specific ability or characteristic which something or someone has, like black hair, no hair, or a quick perception, or "her quickness to grasp new tasks". So, think a while about your outstanding attributes. What about your "outstanding properties"? Great, if one of your strong points is your ability to quickly understand and adapt to new situations! Otherwise, you would not be learning Python!

Let's get back to Python: We will learn later that properties and attributes are essentially different things in Python. This subsection of our tutorial is about attributes in Python. So far our robots have no attributes. Not even a name, like it is customary for ordinary robots, isn't it? So, let's implement a name attribute. "type designation", "build year" etc. are easily conceivable as further attributes as well***.

Attributes are created inside a class definition, as we will soon learn. We can dynamically create arbitrary new attributes for existing instances of a class. We do this by joining an arbitrary name to the instance name, separated by a dot ".". In the following example, we demonstrate this by creating an attribute for the name and the year built:

class Robot:
    pass
x = Robot()
y = Robot()
x.name = "Marvin"
x.build_year = "1979"
y.name = "Caliban"
y.build_year = "1993"
print(x.name)

Marvin

print(y.build_year)

1993

As we have said before: This is not the way to properly create instance attributes. We introduced this example, because we think that it may help to make the following explanations easier to understand.

If you want to know, what's happening internally: The instances possess dictionaries __dict__, which they use to store their attributes and their corresponding values:

x.__dict__

{'name': 'Marvin', 'build_year': '1979'}

y.__dict__

{'name': 'Caliban', 'build_year': '1993'}

Attributes can be bound to class names as well. In this case, each instance will possess this name as well. Watch out, what happens, if you assign the same name to an instance:

class Robot(object):
    pass