The Forgotten History of OOP

The Forgotten History of OOP

Object-oriented programming is a programming paradigm based on the concept of "objects", which can contain data, in the form of fields, and code, in the form of procedures.

Originally published by Eric Elliott at

The functional and imperative programming paradigms we use today were first explored mathematically in the 1930s with lambda calculus and the Turing machine, which are alternative formulations of universal computation (formalized systems which can perform general computation). The Church Turing Thesis showed that lambda calculus and Turing machines are functionally equivalent — that anything that can be computed using a Turing machine can be computed using lambda calculus, and vice versa.

Note: There is a common misconception that Turing machines can compute anything computable. There are classes of problems (e.g., the halting problem) that can be computable for some cases, but are not generally computable for all cases using Turing machines. When I use the word “computable” in this text, I mean “computable by a Turing machine”.

Lambda calculus represents a top-down, function application approach to computation, while the ticker tape/register machine formulation of the Turing machine represents a bottom-up, imperative (step-by-step) approach to computation.

Low level languages like machine code and assembly appeared in the 1940s, and by the end of the 1950s, the first popular high-level languages appeared. Lisp dialects are still in common use today, including Clojure, Scheme, AutoLISP, etc. FORTRAN and COBOL both appeared in the 1950s and are examples of imperative high-level languages still in use today, though C-family languages have replaced both COBOL and FORTRAN for most applications.

Both imperative programming and functional programming have their roots in the mathematics of computation theory, predating digital computers. “Object-Oriented Programming” (OOP) was coined by Alan Kay circa 1966 or 1967 while he was at grad school.

Ivan Sutherland’s seminal Sketchpad application was an early inspiration for OOP. It was created between 1961 and 1962 and published in his Sketchpad Thesis in 1963. The objects were data structures representing graphical images displayed on an oscilloscope screen, and featured inheritance via dynamic delegates, which Ivan Sutherland called “masters” in his thesis. Any object could become a “master”, and additional instances of the objects were called “occurrences”. Sketchpad’s masters share a lot in common with JavaScript’s prototypal inheritance.

Note: The TX-2 at MIT Lincoln Laboratory was one of the early uses of a graphical computer monitor employing direct screen interaction using a light pen. The EDSAC, which was operational between 1948–1958 could display graphics on a screen. The Whirlwind at MIT had a working oscilloscope display in 1949. The project’s motivation was to create a general flight simulator capable of simulating instrument feedback for multiple aircraft. That led to the development of the SAGE computing system. The TX-2 was a test computer for SAGE.

The first programming language widely recognized as “object oriented” was Simula, specified in 1965. Like Sketchpad, Simula featured objects, and eventually introduced classes, class inheritance, subclasses, and virtual methods.

Note: A virtual method is a method defined on a class which is designed to be overridden by subclasses. Virtual methods allow a program to call methods that may not exist at the moment the code is compiled by employing dynamic dispatch to determine what concrete method to invoke at runtime. JavaScript features dynamic types and uses the delegation chain to determine which methods to invoke, so does not need to expose the concept of virtual methods to programmers. Put another way, all methods in JavaScript use runtime method dispatch, so methods in JavaScript don’t need to be declared “virtual” to support the feature.
The Big Idea

“I made up the term ‘object-oriented’, and I can tell you I didn’t have C++ in mind.” ~ Alan Kay, OOPSLA ‘97

Alan Kay coined the term “object oriented programming” at grad school in 1966 or 1967. The big idea was to use encapsulated mini-computers in software which communicated via message passing rather than direct data sharing — to stop breaking down programs into separate “data structures” and “procedures”.

“The basic principal of recursive design is to make the parts have the same power as the whole.” ~ Bob Barton, the main designer of the B5000, a mainframe optimized to run Algol-60.

Smalltalk was developed by Alan Kay, Dan Ingalls, Adele Goldberg, and others at Xerox PARC. Smalltalk was more object-oriented than Simula — everything in Smalltalk is an object, including classes, integers, and blocks (closures). The original Smalltalk-72 did not feature subclassing. That was introduced in Smalltalk-76 by Dan Ingalls.

While Smalltalk supported classes and eventually subclassing, Smalltalk was not about classes or subclassing things. It was a functional language inspired by Lisp as well as Simula. Alan Kay considers the industry’s focus on subclassing to be a distraction from the true benefits of object oriented programming.

“I’m sorry that I long ago coined the term “objects” for this topic because it gets many people to focus on the lesser idea. The big idea is messaging.”
~ Alan Kay

In a 2003 email exchange, Alan Kay clarified what he meant when he called Smalltalk “object-oriented”:

“OOP to me means only messaging, local retention and protection and hiding of state-process, and extreme late-binding of all things.”
~ Alan Kay

In other words, according to Alan Kay, the essential ingredients of OOP are:

  • Message passing
  • Encapsulation
  • Dynamic binding

Notably, inheritance and subclass polymorphism were NOT considered essential ingredients of OOP by Alan Kay, the man who coined the term and brought OOP to the masses.

The Essence of OOP

The combination of message passing and encapsulation serve some important purposes:

  • Avoiding shared mutable state by encapsulating state and isolating other objects from local state changes. The only way to affect another object’s state is to ask (not command) that object to change it by sending a message. State changes are controlled at a local, cellular level rather than exposed to shared access.
  • Decoupling objects from each other — the message sender is only loosely coupled to the message receiver, through the messaging API.
  • Adaptability and resilience to changes at runtime via late binding. Runtime adaptability provides many great benefits that Alan Kay considered essential to OOP.

These ideas were inspired by biological cells and/or individual computers on a network via Alan Kay’s background in biology and influence from the design of Arpanet (an early version of the internet). Even that early on, Alan Kay imagined software running on a giant, distributed computer (the internet), where individual computers acted like biological cells, operating independently on their own isolated state, and communicating via message passing.

“I realized that the cell/whole-computer metaphor would get rid of data[…]”
~ Alan Kay

By “get rid of data”, Alan Kay was surely aware of shared mutable state problems and tight coupling caused by shared data — common themes today.

But in the late 1960s, ARPA programmers were frustrated by the need to choose a data model representation for their programs in advance of building software. Procedures that were too tightly coupled to particular data structures were not resilient to change. They wanted a more homogenous treatment of data.

“[…] the whole point of OOP is not to have to worry about what is inside an object. Objects made on different machines and with different languages should be able to talk to each other […]” ~ Alan Kay

Objects can abstract away and hide data structure implementations. The internal implementation of an object could change without breaking other parts of the software system. In fact, with extreme late binding, an entirely different computer system could take over the responsibilities of an object, and the software could keep working. Objects, meanwhile, could expose a standard interface that works with whatever data structure the object happened to use internally. The same interface could work with a linked list, a tree, a stream, and so on.

Alan Kay also saw objects as algebraic structures, which make certain mathematically provable guarantees about their behaviors:

“My math background made me realize that each object could have several algebras associated with it, and there could be families of these, and that these would be very very useful.”
~ Alan Kay

This has proven to be true, and forms the basis for objects such as promises and lenses, both inspired by category theory.

The algebraic nature of Alan Kay’s vision for objects would allow objects to afford formal verifications, deterministic behavior, and improved testability, because algebras are essentially operations which obey a few rules in the form of equations.

In programmer lingo, algebras are like abstractions made up of functions (operations) accompanied by specific laws enforced by unit tests those functions must pass (axioms/equations).

Those ideas were forgotten for decades in most C-family OO languages, including C++, Java, C#, etc., but they’re beginning to find their way back into recent versions of most widely used OO languages.

You might say the programming world is rediscovering the benefits of functional programming and reasoned thought in the context of OO languages.

Like JavaScript and Smalltalk before it, most modern OO languages are becoming more and more “multi-paradigm languages”. There is no reason to choose between functional programming and OOP. When we look at the historical essence of each, they are not only compatible, but complementary ideas.

Because they share so many features in common, I like to say that JavaScript is Smalltalk’s revenge on the world’s misunderstanding of OOP. Both Smalltalk and JavaScript support:

  • Objects
  • First-class functions and closures
  • Dynamic types
  • Late binding (functions/methods changeable at runtime)
  • OOP without class inheritance

What is essential to OOP (according to Alan Kay)?

  • Encapsulation
  • Message passing
  • Dynamic binding (the ability for the program to evolve/adapt at runtime)

What is non-essential?

  • Classes
  • Class inheritance
  • Special treatment for objects/functions/data
  • The new keyword
  • Polymorphism
  • Static types
  • Recognizing a class as a “type”

If your background is Java or C#, you may be thinking static types and Polymorphism are essential ingredients, but Alan Kay preferred dealing with generic behaviors in algebraic form. For example, from Haskell:

fmap :: (a -> b) -> f a -> f b

This is the functor map signature, which acts generically over unspecified types a and b, applying a function from a to b in the context of a functor of a to produce a functor of b. Functor is math jargon that essentially means “supporting the map operation”. If you're familiar with [].map() in JavaScript, you already know what that means.

Here are two examples in JavaScript:

// isEven = Number => Boolean
 const isEven = n => n % 2 === 0;const nums = [1, 2, 3, 4, 5, 6];// map takes a function `a => b` and an array of `a`s (via `this`)
 // and returns an array of `b`s.
 // in this case, `a` is `Number` and `b` is `Boolean`
 const results =;console.log(results);
 // [false, true, false, true, false, true]

The .map() method is generic in the sense that a and b can be any type, and .map() handles it just fine because arrays are data structures that implement the algebraic functor laws. The types don't matter to .map() because it doesn't try to manipulate them directly, instead applying a function that expects and returns the correct types for the application.

// matches = a => Boolean
 // here, `a` can be any comparable type
 const matches = control => input => input === control;const strings = ['foo', 'bar', 'baz'];const results ='bar'));console.log(results);
 // [false, true, false]

This generic type relationship is difficult to express correctly and thoroughly in a language like TypeScript, but was pretty easy to express in Haskell’s Hindley Milner types with support for higher kinded types (types of types).

Most type systems have been too restrictive to allow for free expression of dynamic and functional ideas, such as function composition, free object composition, runtime object extension, combinators, lenses, etc. In other words, static types frequently make it harder to write composable software.

If your type system is too restrictive (e.g., TypeScript, Java), you’re forced to write more convoluted code to accomplish the same goals. That doesn’t mean static types are a bad idea, or that all static type implementations are equally restrictive. I have encountered far fewer problems with Haskell’s type system.

If you’re a fan of static types and you don’t mind the restrictions, more power to you, but if you find some of the advice in this text difficult because it’s hard to type composed functions and composite algebraic structures, blame the type system, not the ideas. People love the comfort of their SUVs, but nobody complains that they don’t let you fly. For that, you need a vehicle with more degrees of freedom.

If restrictions make your code simpler, great! But if restrictions force you to write more complicated code, perhaps the restrictions are wrong.

What is an Object?

Objects have clearly taken on a lot of connotations over the years. What we call “objects” in JavaScript are simply composite data types, with none of the implications from either class-based programming or Alan Kay’s message-passing.

In JavaScript, those objects can and frequently do support encapsulation, message passing, behavior sharing via methods, even subclass polymorphism (albeit using a delegation chain rather than type-based dispatch). You can assign any function to any property. You can build object behaviors dynamically, and change the meaning of an object at runtime. JavaScript also supports encapsulation using closures for implementation privacy. But all of that is opt-in behavior.

Our current idea of an object is simply a composite data structure, and does not require anything more to be considered an object. But programming using these kinds of objects does not make your code “object-oriented” any more than programming with functions makes your code “functional”.

OOP is not Real OOP Anymore

Because “object” in modern programming languages means much less than it did to Alan Kay, I’m using “component” instead of “object” to describe the rules of real OOP. Many objects are owned and manipulated directly by other code in JavaScript, but components should encapsulate and control their own state.

Real OOP means:

  • Programming with components (Alan Kay’s “object”)
  • Component state must be encapsulated
  • Using message passing for inter-object communication
  • Components can be added/changed/replaced at runtime

Most component behaviors can be specified generically using algebraic data structures. Inheritance is not needed here. Components can reuse behaviors from shared functions and modular imports without sharing their data.

Manipulating objects or using class inheritance in JavaScript does not mean that you’re “doing OOP”. Using components in this way does. But popular usage is how words get defined, so perhaps we should abandon OOP and call this “Message Oriented Programming (MOP)” instead of “Object Oriented Programming (OOP)”?

Is it coincidence that mops are used to clean up messes?

What Good MOP Looks Like

In most modern software, there is some UI responsible for managing user interactions, some code managing application state (user data), and code managing system or network I/O.

Each of those systems may require long-lived processes, such as event listeners, state to keep track of things like the network connection, ui element status, and the application state itself.

Good MOP means that instead of all of these systems reaching out and directly manipulating each other’s state, the system communicates with other components via message dispatch. When the user clicks on a save button, a "SAVE" message might get dispatched, which an application state component might interpret and relay to a state update handler (such as a pure reducer function). Perhaps after the state has been updated, the state component might dispatch a "STATE_UPDATED" message to a UI component, which in turn will interpret the state, reconcile what parts of the UI need to be updated, and relay the updated state to the subcomponents that handle those parts of the UI.

Meanwhile, the network connection component might be monitoring the user’s connection to another machine on the network, listening for messages, and dispatching updated state representations to save data on a remote machine. It’s internally keeping track of a network heartbeat timer, whether the connection is currently online or offline, and so on.

These systems don’t need to know about the details of the other parts of the system. Only about their individual, modular concerns. The system components are decomposable and recomposable. They implement standardized interfaces so that they are able to interoperate. As long as the interface is satisfied, you could substitute replacements which may do the same thing in different ways, or completely different things with the same messages. You may even do so at runtime, and everything would keep working properly.

Components of the same software system may not even need to be located on the same machine. The system could be decentralized. The network storage might shard the data across a decentralized storage system like IPFS, so that the user is not reliant on the health of any particular machine to ensure their data is safely backed up, and safe from hackers who might want to steal it.

OOP was partially inspired by Arpanet, and one of the goals of Arpanet was to build a decentralized network that could be resilient to attacks like atomic bombs. According to director of DARPA during Arpanet development, Stephen J. Lukasik (“Why the Arpanet Was Built”):

“The goal was to exploit new computer technologies to meet the needs of military command and control against nuclear threats, achieve survivable control of US nuclear forces, and improve military tactical and management decision making.”

Note: The primary impetus of Arpanet was convenience rather than nuclear threat, and its obvious defense advantages emerged later. ARPA was using three separate computer terminals to communicate with three separate computer research projects. Bob Taylor wanted a single computer network to connect each project with the others.

A good MOP system might share the internet’s robustness using components that are hot-swappable while the application is running. It could continue to work if the user is on a cell phone and they go offline because they entered a tunnel. It could continue to function if a hurricane knocks out the power to one of the data centers where servers are located.

It’s time for the software world to let go of the failed class inheritance experiment, and embrace the math and science principles that originally defined the spirit of OOP.

It’s time for us to start building more flexible, more resilient, better-composed software, with MOP and functional programming working in harmony.

Note: The MOP acronym is already used to describe “monitoring-oriented programming” and its unlikely OOP is going to go away quietly.
Don’t be upset if MOP doesn’t catch on as programming lingo.
Do MOP up your OOPs.

Thanks for reading

If you liked this post, share it with all of your programming buddies!

Follow us on Facebook | Twitter

Further reading

Goodbye, Object Oriented Programming

JavaScript and Object-Oriented Programming

Python vs Java: Understand Object Oriented Programming

Object-Oriented Programming — The Trillion Dollar Disaster

Object-Oriented Programming is Bad

Hire Dedicated eCommerce Web Developers | Top eCommerce Web Designers

Hire Dedicated eCommerce Web Developers | Top eCommerce Web Designers

Build your eCommerce project by hiring our expert eCommerce Website developers. Our Dedicated Web Designers develop powerful & robust website in a short span of time.

Build your eCommerce project by hiring our expert eCommerce Website developers. Our Dedicated Web Designers develop powerful & robust website in a short span of time.

Hire Now:

Mobile App Development Company India | Ecommerce Web Development Company India

Mobile App Development Company India | Ecommerce Web Development Company India

Best Mobile App Development Company India, WebClues Global is one of the leading web and mobile app development company. Our team offers complete IT solutions including Cross-Platform App Development, CMS & E-Commerce, and UI/UX Design.

We are custom eCommerce Development Company working with all types of industry verticals and providing them end-to-end solutions for their eCommerce store development.

Know more about Top E-Commerce Web Development Company

JavaScript developers should you be using Web Workers?

JavaScript developers should you be using Web Workers?

Do you think JavaScript developers should be making more use of Web Workers to shift execution off of the main thread?

Originally published by David Gilbertson at

So, Web Workers. Those wonderful little critters that allow us to execute JavaScript off the main thread.

Also known as “no, you’re thinking of Service Workers”.

Photo by Caleb Jones on Unsplash

Before I get into the meat of the article, please sit for a lesson in how computers work:

Understood? Good.

For the red/green colourblind, let me explain. While a CPU is doing one thing, it can’t be doing another thing, which means you can’t sort a big array while a user scrolls the screen.

This is bad, if you have a big array and users with fingers.

Enter, Web Workers. These split open the atomic concept of a ‘CPU’ and allow us to think in terms of threads. We can use one thread to handle user-facing work like touch events and rendering the UI, and different threads to carry out all other work.

Check that out, the main thread is green the whole way through, ready to receive and respond to the gentle caress of a user.

You’re excited (I can tell), if we only have UI code on the main thread and all other code can go in a worker, things are going to be amazing (said the way Oprah would say it).

But cool your jets for just a moment, because websites are mostly about the UI — it’s why we have screens. And a lot of a user’s interactions with your site will be tapping on the screen, waiting for a response, reading, tapping, looking, reading, and so on.

So we can’t just say “here’s some JS that takes 20ms to run, chuck it on a thread”, we must think about where that execution time exists in the user’s world of tap, read, look, read, tap…

I like to boil this down to one specific question:

Is the user waiting anyway?

Imagine we have created some sort of git-repository-hosting website that shows all sorts of things about a repository. We have a cool feature called ‘issues’. A user can even click an ‘issues’ tab in our website to see a list of all issues relating to the repository. Groundbreaking!

When our users click this issues tab, the site is going to fetch the issue data, process it in some way — perhaps sort, or format dates, or work out which icon to show — then render the UI.

Inside the user’s computer, that’ll look exactly like this.

Look at that processing stage, locking up the main thread even though it has nothing to do with the UI! That’s terrible, in theory.

But think about what the human is actually doing at this point. They’re waiting for the common trio of network/process/render; just sittin’ around with less to do than the Bolivian Navy.

Because we care about our users, we show a loading indicator to let them know we’ve received their request and are working on it — putting the human in a ‘waiting’ state. Let’s add that to the diagram.

Now that we have a human in the picture, we can mix in a Web Worker and think about the impact it will have on their life:


First thing to note is that we’re not doing anything in parallel. We need the data from the network before we process it, and we need to process the data before we can render the UI. The elapsed time doesn’t change.

(BTW, the time involved in moving data to a Web Worker and back is negligible: 1ms per 100 KB is a decent rule of thumb.)

So we can move work off the main thread and have a page that is responsive during that time, but to what end? If our user is sitting there looking at a spinner for 600ms, have we enriched their experience by having a responsive screen for the middle third?


I’ve fudged these diagrams a little bit to make them the gorgeous specimens of graphic design that they are, but they’re not really to scale.

When responding to a user request, you’ll find that the network and DOM-manipulating part of any given task take much, much longer than the pure-JS data processing part.

I saw an article recently making the case that updating a Redux store was a good candidate for Web Workers because it’s not UI work (and non-UI work doesn’t belong on the main thread).

Chucking the data processing over to a worker thread sounds sensible, but the idea struck me as a little, umm, academic.

First, let’s split instances of ‘updating a store’ into two categories:

  1. Updating a store in response to a user interaction, then updating the UI in response to the data change
  2. Not that first one

If the first scenario, a user taps a button on the screen — perhaps to change the sort order of a list. The store updates, and this results in a re-rendering of the DOM (since that’s the point of a store).

Let me just delete one thing from the previous diagram:

In my experience, it is rare that the store-updating step goes beyond a few dozen milliseconds, and is generally followed by ten times that in DOM updating, layout, and paint. If I’ve got a site that’s taking longer than this, I’d be asking questions about why I have so much data in the browser and so much DOM, rather than on which thread I should do my processing.

So the question we’re faced with is the same one from above: the user tapped something on the screen, we’re going to work on that request for hopefully less than a second, why would we want to make the screen responsive during that time?

OK what about the second scenario, where a store update isn’t in response to a user interaction? Performing an auto-save, for example — there’s nothing more annoying than an app becoming unresponsive doing something you didn’t ask it to do.

Actually there’s heaps of things more annoying than that. Teens, for example.

Anyhoo, if you’re doing an auto-save and taking 100ms to process data client-side before sending it off to a server, then you should absolutely use a Web Worker.

In fact, any ‘background’ task that the user hasn’t asked for, or isn’t waiting for, is a good candidate for moving to a Web Worker.

The matter of value

Complexity is expensive, and implementing Web Workers ain’t cheap.

If you’re using a bundler — and you are — you’ll have a lot of reading to do, and probably npm packages to install. If you’ve got a create-react-app app, prepare to eject (and put aside two days twice a year to update 30 different packages when the next version of Babel/Redux/React/ESLint comes out).

Also, if you want to share anything fancier than plain data between a worker and the main thread you’ve got some more reading to do (comlink is your friend).

What I’m getting at is this: if the benefit is real, but minimal, then you’ve gotta ask if there’s something else you could spend a day or two on with a greater benefit to your users.

This thinking is true of everything, of course, but I’ve found that Web Workers have a particularly poor benefit-to-effort ratio.

Hey David, why you hate Web Workers so bad?

Good question.

This is a doweling jig:

I own a doweling jig. I love my doweling jig. If I need to drill a hole into the end of a piece of wood and ensure that it’s perfectly perpendicular to the surface, I use my doweling jig.

But I don’t use it to eat breakfast. For that I use a spoon.

Four years ago I was working on some fancy animations. They looked slick on a fast device, but janky on a slow one. So I wrote fireball-js, which executes a rudimentary performance benchmark on the user’s device and returns a score, allowing me to run my animations only on devices that would render them smoothly.

Where’s the best spot to run some CPU intensive code that the user didn’t request? On a different thread, of course. A Web Worker was the correct tool for the job.

Fast forward to 2019 and you’ll find me writing a routing algorithm for a mapping application. This requires parsing a big fat GeoJSON map into a collection of nodes and edges, to be used when a user asks for directions. The processing isn’t in response to a user request and the user isn’t waiting on it. And so, a Web Worker is the correct tool for the job.

It was only when doing this that it dawned on me: in the intervening quartet of years, I have seen exactly zero other instances where Web Workers would have improved the user experience.

Contrast this with a recent resurgence in Web Worker wonderment, and combine that contrast with the fact that I couldn’t think of anything else to write about, then concatenate that combined contrast with my contrarian character and you’ve got yourself a blog post telling you that maybe Web Workers are a teeny-tiny bit overhyped.

Thanks for reading

If you liked this post, share it with all of your programming buddies!

Follow us on Facebook | Twitter

Further reading

An Introduction to Web Workers

JavaScript Web Workers: A Beginner’s Guide

Using Web Workers to Real-time Processing

How to use Web Workers in Angular app

Using Web Workers with Angular CLI