History of Computing: From the Abacus to the First Computer

Say «history of computing» and most people picture a black-and-white photograph: a room full of cabinets and magnetic tape. Nice image. It is also the last chapter. The story starts much earlier, with sticks, toothed wheels and one stubborn idea that took centuries to find its machine: calculating without a person doing the math.

Computing began long before the first computer. It began the first time somebody asked whether a machine could work through repeated steps on its own. Transistors, microchips, the phone in your pocket. All of it is downstream of that single question. Once you see the chain, a list of dates stops being trivia and becomes a map of why your own tools look the way they do.

What we mean by the history of computing

The history of computing is the story of how we went from counting on our fingers to handing the math to machines, and then from those machines to a worldwide network of computers. It is the history of an accumulation, where every advance stands on the one before it and the leaps only look small until you line them up.

Keep a few words straight, because people use them as if they were synonyms and they are not. Computing is the act of running calculations or instructions. Computer science is the wider study of how information gets processed automatically. Hardware is the physical machine. Software is what tells it what to do. A network links machines to each other, and the internet is the network of networks that joins them across the planet. Mix them up and you end up arguing about the wrong thing, which happens more often than you would expect.

This story needs all of them: machines, programs, protocols and the people who built them. It is not a museum tour either. It is the reason your laptop, your phone and the servers behind every website you visit are arranged the way they are. Those choices go back further than most people think, and they were made for practical reasons you can still trace today. You do not need the dates to use a computer. You do need them to understand why it behaves the way it does.

The origins: from the abacus to the Turing machine

For centuries, computing was human work, and the word still carries the proof: a computer was a person who did calculations. The abacus let a trader keep a running total without trusting memory. Pascal’s machine worked with wheels and gears, and each turn of a wheel carried into the next digit, much like the mechanical counters that survived into the twentieth century. Blaise Pascal’s Pascaline (1642) could add and subtract, and he built it to spare his father, a tax official, from long columns of sums. Charles Babbage’s Analytical Engine (1837) went further than anything before it. The thread that reaches the modern computer, though, starts with a loom.

The Jacquard loom and the first coded instruction (1801)

The Jacquard loom (1801) proved that a machine could follow instructions stored outside itself. Joseph-Marie Jacquard, a French weaver, built a loom that read punched cards to decide the pattern of the cloth. The holes were the program. The fabric was the output. Change the cards and the loom wove something else, with no engineer touching the mechanism. It calculated nothing, yet it fixed the idea that would carry all of computing afterwards: the plan is encoded on a medium you swap without touching the machine. A century later, the first computers would read their instructions exactly that way.

Babbage’s Analytical Engine (1837)

The Analytical Engine was the first design for a general-purpose computer, even though it never got built. Charles Babbage gave it a memory, a processing unit and conditional branching, so it could run any operation it was programmed to perform. He had already drawn a Difference Engine to calculate mathematical tables, and the Analytical Engine went further by making the steps themselves programmable. His own era could not cut gears accurately enough, and the machine stayed on paper. The blueprint was right. It was a computer in everything but metal.

Ada Lovelace and the first algorithm (1843)

Ada Lovelace wrote, in her 1843 notes on Babbage’s machine, the first algorithm intended to be run by a machine: a method for computing Bernoulli numbers. She had translated an Italian account of the engine and buried her own thinking in the notes, which is where that method lives. Her contribution reaches past arithmetic. Lovelace saw that these machines could handle any symbol, music and letters and logic included, and she wrote that the engine «weaves algebraic patterns just as the Jacquard loom weaves flowers and leaves». She became the first programmer, and the first to imagine a computer that was something more than a calculator.

Boole’s logic and Shannon’s bridge (1854)

Boolean algebra (1854) is the reason digital circuits work at all. In An Investigation of the Laws of Thought, the mathematician George Boole reduced reasoning to a two-value system of true and false, plus a handful of operations: AND, OR, NOT. At the time it was pure mathematics with no practical use, and decades passed with nobody building anything from it. Eighty-three years later, in 1937, Claude Shannon showed in his master’s thesis that this algebra described exactly how electrical switching circuits behave. That is where digital electronics began. The ones and zeros moving through every chip today are, literally, Boole’s algebra.

The Turing machine (1936)

The Turing machine (1936) is the theoretical model that defines what can and cannot be computed. Alan Turing described an imaginary machine able to follow instructions and solve any computable problem. The logical blueprint of what a computer would need in order to exist, with no hardware at all. He was answering a formal question about what mathematics can decide, and the machine was the tool he invented to answer it. That paper, «On Computable Numbers», closed that open problem. Turing did not build a machine. He described what it would take to build one, and that order, think first and build second, shapes everything in computing since.

What was the first computer in history?

The first computer in history has no single name. It depends on what you call a computer, and several candidates have a real case.

• Zuse Z3 (1941), in Germany: the first working, program-controlled general-purpose machine, though electromechanical rather than electronic.
• Colossus (1943-44), in the UK: electronic, but built for one specific purpose, decoding encrypted messages. It was engineered by Tommy Flowers at Bletchley Park, and its existence stayed classified for decades, which kept the achievement hidden from the public for thirty years.
• ENIAC (1945), in the United States: the first fully electronic, general-purpose one, and the one that became famous. It filled a room, weighed around 30 tonnes and did roughly 5,000 additions per second.

If by «first computer» you mean the first electronic, programmable, general-purpose machine, the answer most people accept is ENIAC. If you mean the first one that actually ran a program, Germany’s Z3 takes the medal. Every label measures a different thing, which is why the argument never quite settles. History is more fun than a headline.

There is a third answer almost nobody gives, and it is the one that explains what came next. The ENIAC was programmed by re-wiring cables and flipping switches, so moving from one problem to another could take days. The decisive leap was the stored program: keeping instructions in the same memory as the data. John von Neumann laid it out in 1945 in his First Draft of a Report on the EDVAC, and almost every computer built since follows that model, from a microcontroller to a data-centre server.

This chapter also carries an unpaid debt. ENIAC was not programmed by its engineers alone. A team of six women, Kay McNulty, Jean Bartik, Betty Holberton, Marlyn Meltzer, Frances Spence and Ruth Teitelbaum, translated the problems into instructions the machine could run. They were some of the first general-purpose programmers, and for years they stayed out of the official photographs. Their work was rediscovered decades later, when historians went back to the records.

Timeline: the milestones that changed everything

This is the chain of inventions that runs from the punched-card loom to modern language models. Each piece solved one problem and made the next one possible.

The chip era and the personal computer (1947-1981)

The transistor (1947) pulled electronics out of the age of vacuum tubes. At Bell Labs, John Bardeen, Walter Brattain and William Shockley built a component that amplified signals and switched them, much like the tubes in ENIAC, only smaller, cooler and far more reliable. It earned them the 1956 Nobel Prize in Physics. A decade later came the integrated circuit (1958). Jack Kilby at Texas Instruments and Robert Noyce at Fairchild, working separately, packed many transistors onto one piece of silicon, and the size and cost of electronics collapsed. The first working circuit Kilby showed in 1958 held just a handful of parts, and that was enough to prove the idea. Kilby took the 2000 Nobel Prize in Physics for it. In 1965, Gordon Moore made the observation that still carries his name: the number of transistors per chip was roughly doubling every two years. That single trend, kept up for decades, is why the machine on your desk today outperforms the room-sized systems of 1960.

In 1969, the same year a man walked on the Moon, ARPANET linked two remote computers for the first time and sent a first message: «LO». It was meant to be «LOGIN», cut off when the system crashed on the second letter. Email arrived in 1971 from Ray Tomlinson, who picked the @ sign to separate user from machine because it never showed up in people’s names. Also in 1971, Intel introduced the 4004, the first commercial microprocessor: a whole CPU on a single chip. Two years later, Ethernet (1973) standardised wired local networks, and it still underpins most wired networks today (IEEE 802.3).

With the hardware shrunk down, the computer left the lab. The Apple II (1977) was one of the first machines aimed at ordinary buyers, and the IBM PC (1981) standardised the category with an open architecture and Microsoft’s MS-DOS. Other manufacturers could build compatible clones, and the PC became the office standard for decades. That open model did as much for the spread of the computer as any single chip inside it.

From the internet to the web (1983-1998)

On 1 January 1983, ARPANET adopted TCP/IP as its single protocol. That day, the internet’s «flag day», marks the technical birth of the network as we know it: one set of rules that splits, sends and reassembles data no matter which machines sit underneath. Different networks could finally talk to each other without a custom translation layer for every pair. The design kept the intelligence in the machines at the edges, leaving the network in the middle to move data. On top of that network, Tim Berners-Lee proposed the World Wide Web in 1989. By the end of 1990 the first server and browser were running at CERN, and in 1991 the system opened to the public. The first web page explained how to make web pages. The web is a layer of services built on the internet, and that layer is what made the internet easy enough to go massive.

That same year, 1991, Finnish student Linus Torvalds published the first version of the Linux kernel and gave it away freely. Today Linux runs most servers, supercomputers and Android phones, and its open development model changed how large-scale software gets built. Then in 1998, Larry Page and Sergey Brin solved the problem the web had created by growing unchecked: ranking results by relevance using the links between pages. Finding information stopped being an act of faith.

The era of services and AI (2006-today)

The modern cloud started in 2006, when Amazon Web Services began selling storage (S3) and computing power (EC2) on demand. Any company could rent infrastructure instead of buying hardware, and building a digital product got radically cheaper. A small team could rent what it needed and pay only for what it used, instead of committing to a data centre of its own. A year later, in 2007, Apple unveiled the iPhone and reset what people expected from a phone. Smartphones existed before it, yet the iPhone is the one that made mobile computing ordinary.

The last piece arrived with data. In 2012, a neural network called AlexNet won the ImageNet computer-vision competition by a huge margin over every previous method. That result proved that deep neural networks, trained on large datasets with graphics cards, could beat everything that came before, and it opened the investment cycle behind today’s AI. The popular version landed on 30 November 2022, when OpenAI released ChatGPT and millions of people began using language models to write, code and analyse.

Year	Milestone	Why it matters
1801	Jacquard loom	Instructions coded on punched cards
1837	Analytical Engine (Babbage)	The first concept of a programmable computer
1843	First algorithm (Ada Lovelace)	Programming is born
1854	Boolean algebra	The basis of all digital circuits
1936	Turing machine	Theoretical foundation of computing
1943-44	Colossus	One of the first working electronic computers
1945	von Neumann architecture	Program stored in the same memory as data
1945-46	ENIAC	First electronic general-purpose computer
1947	Transistor	Goodbye to vacuum tubes: smaller and more reliable
1958	Integrated circuit	Components fit on a chip
1969	ARPANET	The embryo of the internet
1971	Email	The @ sign separates user and machine
1971	Intel 4004 microprocessor	A whole CPU on a single chip
1973	Ethernet	The standard for wired local networks
1977	Apple II	The personal computer reaches the mainstream
1981	IBM PC	The computer reaches home and office
1983	TCP/IP	The internet as we know it
1989-1991	World Wide Web	The internet becomes easy to use
1991	Linux	Open source dominates servers
1998	Google	Bringing order to the web’s information
2006	Cloud (AWS S3 and EC2)	Infrastructure on demand
2007	iPhone	Computing fits in a pocket
2012	AlexNet	Deep learning takes off
2022	Generative AI (ChatGPT)	Language models reach the mainstream

The generations of computers

We count the generations of computing by the technology that made each one possible. The word generation sounds tidy, and the reality is messier. These are convenient and slightly loose labels, because the generations overlap and no clean line separates one from the next. Still, the pattern holds: each one brings smaller, cheaper machines and more people using them.

• First (1940s-50s): vacuum tubes. Giant, hot, unreliable machines. ENIAC and UNIVAC I (1951), the first commercial computer built in the United States, are the reference points.
• Second (1950s-60s): transistors. Smaller, safer and, for the first time, within reach of companies and universities.
• Third (1960s-70s): integrated circuits. Software as we know it is born and compatible computer families appear, with IBM’s System/360 (1964) as the landmark.
• Fourth (from the 70s): microprocessors. A CPU on a chip makes the machine so cheap that the personal computer enters homes and offices.
• Fifth (today): artificial intelligence, cloud and ubiquitous computing. The power no longer lives in one box. It is spread across devices, servers and services talking to each other over the network.

Read that list from top to bottom and the direction is obvious. Each generation handed the next one a smaller, faster and cheaper platform, and each platform pulled in a new group of users who had never touched a computer before, whether at work, at school or at home.

From mechanical computing to today’s digital infrastructure

Today’s computing is the accumulated result of those milestones, one stacked on the next. The transistor enabled the integrated circuit. The integrated circuit enabled the microprocessor. The microprocessor enabled the personal computer. The connected computer produced the web, the web multiplied the data, and that data made it possible to train artificial intelligence at scale. Nobody planned the whole thing. Each piece solved a problem and left the next one ready.

The infrastructure you use today is that same stack, only taller. When you pick between the different types of databases to store information, when you design a subnet to segment a company network, or when you chase the jitter that ruins a video call, you are working on layers built in that order and for those reasons. Knowing the history pays off. It tells you why today’s technology has the shape it has.

Here is the practical payoff. Every layer you touch at work sits on top of the ones below it. The cloud rents you machines that still obey von Neumann. The database you tune writes to storage that keeps bits the same way Shannon’s circuits did. Fix a problem at the wrong layer and it comes right back, because you never reached the cause.

The pattern repeats at every step: a new capability creates a new dependency, and that dependency creates new problems nobody had to solve before. Computing grows by piling up layers, which is why almost no technical problem today makes sense from the top layer alone.

Frequently asked questions

What was the first computer in history?

It depends on the criteria. The ENIAC (1945) is considered the first electronic, general-purpose computer. The Zuse Z3 (1941) was the first programmable and functional one, though electromechanical. Colossus (1943-44) was electronic, but built for a specific purpose.

Who invented the first computer?

There is no single inventor. Computing is a chain: Turing supplied the theory, Zuse the first running program, Mauchly and Eckert built ENIAC and von Neumann defined the stored program. Before them, Babbage and Ada Lovelace laid the foundations.

Who is considered the father of computing?

Alan Turing is the most cited figure for his theoretical foundations. Charles Babbage is cited for the design of the Analytical Engine, and John von Neumann for the stored-program architecture that almost every computer still follows.

Who was Ada Lovelace?

A nineteenth-century British mathematician who wrote the first algorithm designed to run on a machine, in her notes on Babbage’s Analytical Engine. She is considered the first programmer in history.

When was the internet born?

ARPANET, its direct predecessor, went live in 1969. The adoption of TCP/IP on 1 January 1983 is considered the technical birth of the internet as we know it.

Which came first, the internet or the personal computer?

The internet. ARPANET went live in 1969, long before the PC reached homes in the early 80s.

When was computing born as a discipline?

As a science, with the Turing machine (1936) and the theory of computation. As an industry, in the 1940s and 50s, with the first commercial computers.

What is the difference between computing and computer science?

Computing is the act of carrying out calculations or instructions. Computer science is broader: it covers the study of automatic information processing, including algorithms, data structures, operating systems, networks, databases and artificial intelligence.

The bottom line

The history of computing comes down to one simple idea, delegating calculation, and that idea took centuries to find its form before it accelerated without brakes. From the punched cards of the Jacquard loom to language models writing text in real time there are almost two centuries, and most of that change happened in the last eighty years, from ENIAC to today.

Eighty years ago a computer filled a room. Today the phone in your pocket holds more power than the machines that put a man on the Moon. Every generation since ENIAC has changed what systems can do, who gets to use them and what for, and the next one is already here, built on everything above.

If you want to see where this is heading, read what Web 4.0 is, the chapter being written right now.