Coding in the 1840s: Who Wrote the First Computer Program?

Everyone’s talking about AI like it is about to take over the world. Fair enough. But if you want to understand how we got here, you have to go back to the first step. In the last episode, a loom taught us that machines can follow instructions. Now we go hunting for a bigger claim: who wrote the first computer program?

There is one awkward detail. In 1842 and 1843 there was no working computer in the modern sense. There was a plan for one: the Analytical Engine. So this becomes a very Victorian puzzle. Can you write a program for a machine that does not exist yet?

The machine and the rules

To write a program, you need rules. You need a machine model that is specific enough that someone could, in theory, take your instructions and implement them in hardware or punch them onto cards.

Babbage’s Analytical Engine was designed around three big ideas:

• The Store (memory) to hold numbers.
• The Mill (the arithmetic unit) to do operations.
• Punched cards to control operations and data movement, inspired by industrial automation.

In modern terms, the translation is almost rude in its simplicity: Store = RAM Mill = CPU Cards = Program + I/O

The important part is not the metaphor. It is the control structure. The card deck can specify an operation, point to memory locations, introduce constants, and it can also support repetition and conditional jumps. That is the difference between a fancy calculator and a general-purpose computer.

Menabrea’s linear equations (1842)

In 1842, Luigi Menabrea published a report describing Babbage’s machine and included a worked method for solving simultaneous linear equations. Think of it like finding where two lines meet, but written in a form that maps neatly onto Store and Mill operations.

This is a serious candidate for “first program in print”, because it goes beyond vague description. It outlines steps and intermediate results in a way that looks very much like a procedure designed for the Engine. It also proves something important: yes, you can write programs without having the machine in the room.

The catch is that it reads more like a walkthrough than a fully developed, reusable algorithm. It hints at repetition, but it does not lean hard into explicit looping and full trace tables. If your standard is “first published program”, Menabrea is strong. If your standard is “first fully worked algorithm you can run step by step without guessing”, we keep digging.

Lovelace’s Note “G” (1843) and the Bernoulli algorithm

In 1843, Ada Lovelace translated Menabrea’s paper into English and added Notes A to G. The notes are the main event. In Note “G”, she outlines a complete procedure for computing Bernoulli numbers.

You do not need to love the maths to appreciate what Note “G” does well. It is machine-oriented. It is precise about what is read from the Store, what the Mill does, and where results go. Most importantly, it includes two ingredients that make programmers relax their shoulders: explicit looping and a trace table that records values after each step.

That combination is why Note “G” is often treated as the first published, fully worked algorithm for a general-purpose computer. Not the first time anyone ever thought about procedures, of course, but a particularly clear “this is what the machine would do” presentation.

Babbage’s drafts and correspondence (1830s to 1843)

Before anything hit print, there were drafts. Babbage filled notebooks with Engine procedures in the late 1830s and early 1840s. If your standard is priority (earliest surviving program, even unpublished), that evidence matters. It suggests Babbage was writing “code on paper” years before the public versions existed.

Then there is 1843 correspondence. Lovelace is shaping a publishable exposition, asking questions, adding structure. Babbage reviews, proposes corrections, and suggests improvements. It looks like co-development in the way early work often is. If you have ever shipped software, this will feel familiar. The first program, like many modern ones, did not emerge from a single genius in a vacuum.

The standard of proof (your verdict)

Here is the part that annoys people who want a simple answer. Different standards give different answers, and all three standards below can be defended.

So, who wrote the first computer program? The honest answer is: it depends on what you mean by “first”. If you value earliest writing, Babbage is hard to ignore. If you value first in print, Menabrea has a strong claim. If you want the first complete, reproducible algorithm in print, Note “G” is the headline act.

Why this matters today

Software began as writing. Not code running on hardware, but careful instructions written for a machine that was still a concept. That is a strange origin story, and it also explains something modern programmers keep rediscovering: the hard part is not the silicon, it is the clarity.

Also, collaboration is not a modern invention. The first “code” was shaped in conversation, drafts, and revisions. So when people argue about credit in tech today, we are not dealing with a new problem. We are replaying a Victorian argument with better Wi-Fi.