by Adam Michael Wood
December 12, 2018
December 12, 2018

Object-Oriented Programming: Making Complicated Concepts Simple

Object-oriented (OO) programming is a programming paradigm that includes or relies on the concept of objects, encapsulated data structures that have properties and functions and which interact with other objects.

Objects in a program frequently represent real-world objects — for example, an ecommerce website application might have Customer objects, Shopping_cart objects, and Product objects. Other objects might be loosely related to real-world equivalents — like a Payment_processor or Login_form. Many other objects serve application logic and have no direct real-world parallel — objects that manage authentication, templating, request handling, or any of the other myriad features needed for a working application.

Object-Oriented Programming

History of Object-oriented Programming

If it seems like object-oriented programming is particularly well suited to modeling real world systems, that isn’t a coincidence. The concepts now associated with object-oriented programming, including most of the language for describing it, originated with the Simula programming language, which was used for simulating real world systems. Classes, subclasses, and objects were added to Simula in 1967.

In the 1970s, Alan Kay coined the term “object-oriented” while working on Smalltalk. This was in reference to the extensive use of objects in the language, and their place as the fundamental unit of organization.

The spread of object-oriented programming was relatively slow at first. Common Lisp introduced the Common Lisp Object System, adding objects to Lisp, in the late 1970s. In the 1980s, the Eiffel programming language was invented, which was fully object-oriented. The inventor of Eiffel, Bertrand Meyer, wrote Object-Oriented Software Construction, which was one of the first books to widely disseminate the concepts of OO programming.

In the 1990s and into the 2000s, object-oriented programming really took off. This coincided with a massive explosion in the tech industry generally during that period, and the sudden influx of programmers into large tech companies. One of the selling points of object-oriented programming, as promoted by its advocates, is that separation of concerns and loose coupling (see below) enable teams to work on large systems more effectively.

By the end of the first decade of the 21st century, object-oriented programming had largely become the accepted “one right way” of developing most software. Newer programming languages, especially dynamically-typed interpreted languages like Python, Ruby, and JavaScript, tend to assume, or even enforce, an object-oriented approach. Meanwhile, many older languages, like Fortran and COBOL have added object-oriented features.

Important Object-oriented Concepts

Object-oriented programming is a concept — a way of approaching the problem of designing a computer program. It is not a hard-and-fast set of rules or a specific methodology. Different programming languages approach OOP differently. Some are more or less strict, and some are just different. What follows is a general overview of object-oriented design concepts. See our sister article, Object-Oriented Programming for resources and details that explain how these concepts are implemented in many different languages.

Separation of Concerns

The central guiding principle which informs all the other concepts in object-oriented programming is separation of concerns. This is the idea that logically distinct bundles of data and functionality should be separated into distinct bundles of application code.

This idea is not unique to object-oriented programming. Most programming languages allow for separation of concerns, and most programming paradigms promote it. However, separation of concerns is of special importance to object-oriented programming. Strong separation of concerns is the impetus behind nearly all other OO concepts.

Objects and Classes

It seems, at first, that the fundamental concept in object-oriented programming is objects. In a way, that’s true. But from the developer’s perspective, the fundamental concern is classes.

A class defines an object; an object instantiates a class. For an example, let’s keep thinking about an ecommerce website. There are a lot of customers, but they all do the same thing. So there would be a Customer class, in which the programmer defined what a customer is, and how it works. Then, each individual customer is represented in the system by a customer object.

Since class definitions specify all the implementation details, most of the work of developing an object-oriented program involves working with classes, not objects.

Classical and Non-classical OO Languages

There are a handful of languages which incorporate a large number of object-oriented concepts, but which don’t have a construct specifically called a class. Most notable is JavaScript, which uses prototypes, or generic objects which are then cloned. The terminology, the underlying philosophy, and even the “under the hood” implementation are different — and yet the developer experience is largely the same.

Languages with classes, which represent the majority of OO languages, are sometimes called “classical.” This is both a pun on the word “class,” and an acknowledgement that classes have become the conventional approach to object-oriented programming.

Methods and Properties

An object has variables, which are usually called properties or attributes in most languages. For example, a Customer object might have properties like shipping_address or name.

Objects also have functions, called methods. These are things that the object is capable of doing, and are also defined within the class. For example, the Customer object might have methods for checkout, change_shipping_address, or logout.

One of the challenges of developing good object-oriented programs is determining how to split up functions among several potential classes. For example, should a checkout method belong to the Customer or to the Shopping Cart? Does the Customer own the logout method, or does the Customer call the logout method of some Authentication_manager object?

Another challenge for many new OO programmers is understanding how to design classes that don’t represent real-world things. Understanding that ecommerce needs Products and a project management system needs Projects is pretty easy. But large object-oriented programs have countless classes defining all sorts of abstract bundles of functionality, like object relational mappers, object factories, and controllers (just to name three easy-to-understand examples).

Navigating these concerns takes practice, of course. An object-oriented developer can also be aided by an understanding of design patterns, and by a development process that includes modeling.

Here are a few resources to help you better understand class, properties, and methods:

Message Passing and Dynamic Dispatch

Classes, and a number of other concepts covered here, have become inexorably linked to object-oriented programming. However, according to the computer scientist who coined the term, object-oriented programming is about message passing.

Message passing means that objects communicate with each other by sending messages. This sounds trivial, but it affects how function calls are made. Rather than one object calling another object’s functions directly, an object would pass a message to another object. The message contains a method name and any associated parameters. The called object then has the ability to handle that method call however is needed.

This ability for each object to determine its own response to method calls as needed is called dynamic dispatch. It allows objects to evolve over run time, and respond to messages according to their present state.

Message passing can be a little confusing, especially since most languages in use today only implement one particular kind (the method call). Here are a few resources to help you better understand it:

Encapsulation

Closely associated with message passing and dynamic dispatch is the concept of encapsulation. Encapsulation means that code outside of an object cannot directly access its internal data. If, for example, one object needs to “know” a property of another object (for example, a Shipping_manager might need the shipping_address of a Customer object), then the calling object can’t find that data out directly; it has to pass a message “asking” the Customer object, which then returns the needed data. (This is usually done with a method call.)

Not all object-oriented languages enforce strict encapsulation; some allow an object’s properties to be accessed (get and set) directly. However, many experts in OO development encourage programmers to adhere to encapsulation principles (by not calling properties directly), even in languages that allow it.

There are many benefits to encapsulation. Primarily, it allows any details or side effects to be abstracted away from the point at which the property is called. For example, each time a particular value is read, you might want to check when it was last calculated, and re-calculate it if certain conditions are met. Being able to add, remove, or change that logic in one place at any time, without affecting any of the other places in the code which make the call, is a benefit of encapsulation.

Inheritance and Composition

Inheritance and composition are two ways that various classes (and their instantiated objects) are related to each other.

Inheritance describes “is a” relationships: a User is a Person; a Customer is a User. In this example, there might be a Person class which defines all the properties and methods of Persons in general (name, birthday). Then a User class would extend the Person class, to add things like a username attribute or login and logout methods. A Customer class might then extend User, adding those things that only a customer needs, like order_history or billing_address.

Composition describes a “has a” relationship, when one object “owns” another object. For example, the Customer class might own a Billing_address, which is itself a full-fledged object defined in a class. Like the inheritance hierarchy, the compositional hierarchy can easily be several layers deep. For example: a Customer might have a Shopping_cart, the Shopping_cart has several Products, each Product has a Vendor, and so forth.

Much of the design work in object-oriented development has to do with mapping out these relationships of inheritance and composition. The class diagram, which is a part of the Unified Modeling Language is an invaluable tool for visualizing these relationships.

Polymorphism

Polymorphism (or, more specifically, subtype polymorphism) is a concept that naturally comes out of inheritance. If a parent class (or “Super class”) has a particular method, then all of its children (“subclasses”) will also have the method. Each subclass might implement the method differently, but calling objects do not need to know which of several subtypes they are calling. They can treat all different subtypes of objects as equivalent.

Polymorphism is a big subject, and is not without controversy. Here are a few resources to help wrap your head around it:

Learn More About Object-oriented Programming

For a deep understanding of object-oriented programming, a number of classic and contemporary books on the subject are invaluable, and can hardly be replaced by online tutorials.

For a more practical, and contemporary, understanding of object-oriented programming concepts, try one or more of these online resources:

Important Object-Oriented Languages

Not all programming languages support object-oriented programming. Some languages are designed to do nothing but support OOP. Others allow for a variety of approaches to programming. Still others appear to be object-oriented, but implement the concepts of object-orientation in non-standard, incomplete, or just plain unusual ways.

Below is a list of some of the more popular OO-capabable languages, with notes about their approach to Object Orientation, and some resources to help you get into OOP with that language.

The internet is filled with essays that look at language X vs language Y — most of which just scratch the surface. We suggest you check out this more general Programming Language Comparison. This thoughtful analysis looks at specific details of several popular OO languages, discussing how particular OO concepts are implemented in each.

C-Based Languages

C is not an object-oriented language. However, it is at least possibleto write OO code in it:

None of the techniques described in those books and articles are particularly robust or easy to deal with for non-trivia programs. Rather than trying to stretch C to its limits, there are three direct descendants of C which include object-oriented language tools.

C++

C++ is multi-paradigm, which means that it supports several different programming methods (PDF), including object-orientation. It is based on C, and developed specifically as a way to add support for OOP’s concept of classes.

C#

C# (“C sharp”) is another C derivative, mostly designed as an improvement over C++ for use in Microsoft’s .NET Framework.

Objective-C

Objective-C was developed about the same time as C++, with essentially the same goal — the addition of OOP capabilities to C. Today, Objective-C is really only used in Apple’s Cocoa Development Platform for OS X and iOS, and GNUstep, its Open Source alternative.

Common Lisp

Lisp is not essentially Object-Oriented. However Common Lisp, one of the more popular Lisp dialects, includes the Common Lisp Object System (CLOS), which provides OOP features.

The approach to Object-Orientation implemented in CLOS is radically different than the way OO is handled in other languages. This means that CLOS is not a great place to start if you want to learn OO in general, or apply OO in other types of languages. However, if you are getting into Lisp, CLOS is very important. Additionally, if you find OOP especially intriguing, you’ll enjoy studying CLOS to see OO concepts in a different light.

Erlang

According to some conventional points of view, Erlang is not an object-oriented language. However, there is a fascinating minority opinion to the contrary, and its proponents state that Erlang Is the Most Object-Oriented Language, or even that Erlang is the only true Object-Oriented language.

Much of this also applies to Elixir, as well. On other hand, maybe this is just OO-style, not true OO.

F#

F# (“F sharp”) is a multi-paradigm language. Its core is really functional programming, but it includes support for OO and attempts to reconcile these two different approaches to programming.

Fortran

Fortran is the oldest programming language still in common use. When it was invented in the 1950s, there was no such thing as object orientation. Explicit OOP support was added to Fortran with the Fortran 2003 release of the language. But it was possible to implement some basic OOP concepts in earlier versions of the language.

Go

Go is a relatively new language, having been developed by Google in 2007. There is a lot of discussion about whether or not Go is object-oriented.

Go doesn’t have anything in it called “object” or “class,” but it does have some analogous structures. From a certain point of view, then, you can think of Go as object-oriented.

Many people in fact do think that Go is OO:

Java

Java was built from the ground up to be object-oriented. Here are some great resources to get you going with Java:

JavaScript

A lot of people don’t think that JavaScript is really object-oriented. But a lot of other people think that it definitely is.

We feel our job is to help you find useful resources, not take sides in unwinnable debates. So here are some OOP resources for JavaScript:

PHP

PHP did not support even basic OO concepts until version 4, and didn’t support full object-orientation until PHP5.

Python

Python codes does not need to be object-oriented, but the language fully supports it. Check out these resources:

Ruby

In Ruby everything is an object, even “primitive” data types like strings and integers. You add two integers by calling the addition method on one of them; you find the length of a string by calling the length method on it.

Smalltalk

Smalltalk is probably the most influential object-oriented language, having had a profound effect on the languages that came later. As a result, it is good to have at least some familiarity with it. These resources will get you started:

On the Other Hand…

If you want to be a well-rounded developer, it is important to understand both the good and bad of any concept. Even though object orientation has become the dominant way of understanding programming, there are legitimate criticisms to be made — both of the ideas themselves, and the culture of hype surrounding them.

Here are a handful of resources presenting the counter-argument to OOP:

Object-oriented Programming Today

While object-orientated programming remains an important methodology in computer programming, it no longer holds its place as industry standard for “good software development.” Other paradigms, especially functional programming, are coming to the forefront, along with new languages that support them. Meanwhile criticisms of object-oriented programming, as well as of the design patterns movement, are on the rise.

Still, a large amount of the software written today is object-oriented. If you are working in any of the common object-oriented languages, you need to have a good understanding of basic object-oriented concepts.


More Interesting Stuff

We have more programming guides, tutorials, and infographics related to coding and development:

  • C++ Developer Resources: in addition to information about C++, there is more information about object-oriented programming.
  • Linux Programming Introduction and Resources: although not specifically about object-oriented programming, this discussion of the many levels of Linux programming is bound to energize you.
  • INTERCAL Introduction and Resources: if you find object-oriented programming hard, studying INTERCAL will make it seem easy. This parody (or joke) language is so complicated and horrible that even the deepest discussion of polymorphism will seem pleasant.

What Code Should You Learn?

Confused about what programming language you should learn to code in? Check out our infographic, What Code Should You Learn? It not only discusses different aspects of the languages, it answers important questions such as, “How much money will I make programming Java for a living?”

Text written by Adam Michael Wood. Compiled and edited by Frank Moraes.

About Adam Michael Wood

Adam specializes in developer documentation and tutorials. In addition to his writing here, he has authored engineering guides and other long-form technical manuals. Outside of work, Adam composes and performs liturgical music. He lives with his wife and children in California.