On the night of 28 August 1758, at the Hôtel de Cluny in Paris, a twenty-eight-year-old clerk named Charles Messier was tracking a comet through the horns of Taurus when he found something in the way.
It was a faint, cloudy patch of light, a little elongated, with no tail and no definite edge. It looked like a comet. Messier had been hired by the astronomer Joseph-Nicolas Delisle for one purpose, to be the first person in France to recover Halley's comet on its predicted return, and he had spent months sweeping the sky for exactly this kind of smudge. So he watched it. He came back the next night, and the night after. It had not moved. Comets move. This thing sat in Taurus like a stain on the glass.
He wrote down its position so he would not be fooled by it again, and went back to looking for comets.
A LIST OF THINGS TO IGNORE
That entry became the first line of a list. Over the next twenty-three years Messier, and later his colleague Pierre Méchain, kept adding to it: every fuzzy, stationary, comet-shaped object that a comet hunter might waste a night on. He published forty-five of them in 1774, and a final version of 103 in 1781. Later historians, going through his notes, brought the count to 110.
The purpose of the list was purely negative. It was a catalogue of false alarms. Messier had no interest in what the objects were, only in where, so he could skip them.
He was very good at his actual job. He found thirteen comets. Louis XV called him the ferret of comets. He was elected to the Académie des Sciences and the Royal Society. By the standards of his own century he was a success, and the success was the comets.
Nobody remembers the comets. Not one of the thirteen has a name anyone uses. They came, they were bright for a season, they left, and the sky closed over them.
Everybody remembers the list. The first entry, the stain in Taurus, is the Crab Nebula, the wreckage of a star that Chinese astronomers watched explode in 1054. M31 is the Andromeda galaxy. M42 is the Orion Nebula. M45 is the Pleiades. M13 is the great cluster in Hercules, hundreds of thousands of stars packed into a ball. M57 is the Ring Nebula, a dying sun's last breath drawn in glowing gas. Every March, amateur astronomers across the northern hemisphere stay up all night attempting a Messier marathon, trying to see all 110 objects between dusk and dawn. The catalogue of distractions became the most beloved catalogue in astronomy.
Messier died in 1817 believing he had spent his life on comets.
THE REPOSITORY WITH FOUR COMMITS
Every developer has a Messier list, and almost nobody looks at it.
It is the repository with four commits and a README that says work in progress. The command-line tool written at eleven at night to fix an annoyance that nobody else had. The parser started on a train. The framework abandoned the moment it made the real problem clear. The game engine that never rendered a second frame. The Advent of Code folder that stops on day nine every year. The thing built specifically to avoid the thing that was due.
We touched on the side project bloom once before, as one kind of gap in a graph. It deserves its own catalogue.
These were the smudges on the way to the comet. The comet was the job: the product, the release, the codebase with a company's name on it. The comet was what you were being paid to find. The side project was the stationary blur that got in the way, noted down and skipped over so you could get back to sweeping.
Look at the graph of anyone who has been writing code for ten years and the comets are easy to spot. They are the dense bands, the sustained months of daily work, the bright clusters around a launch. They are also, in nearly every case, gone. The company was acquired. The product was sunset. The codebase was rewritten by people who never read the original. What was bright for a season has left, and the sky has closed over it.
The smudges are still there. The four-commit repository is still there. Sometimes someone found it. Sometimes it got a star from a stranger in another country who needed exactly that annoyance fixed. Sometimes it grew, in a corner, over years, into the thing the developer is now known for. The library that a whole ecosystem depends on was almost always somebody's Messier object first: noted, ignored, returned to on a slow Tuesday, and never meant to be the point.
CATALOGUING THE SMUDGES
On an atlas, the side projects have a look.
On the Chronicle, they appear as scattered bright points that never consolidated into a band: a flare in one week, silence, another flare two months later, the signature of something worked on when there was time and put down when there was not. On the Helix they show up as a change in texture within a single loop of the spiral, a patch of irregular light against the steady rhythm of the job around it. On the Planisphere they tend to sit in the same months every year, the holiday weeks and the dead fortnight in December, the only times the sweeping stops long enough to look at the stain in Taurus.
The atlas does not know which stars were the comet and which were the smudge. A commit is a commit. A day is a day. The four-commit repository and the ten-thousand-commit monolith are rendered in the same ink, sized by the same rule, and placed in the same sky. We think this is the right way to draw a life, because a life does not know either, at the time. Messier did not know. He thought the list was the mistake.
WHAT SURVIVES THE COMET
There is a version of a developer's story that only counts the comets. It is the CV. Company, title, dates, the product that shipped. It is true and it is thin, and it leaves out the objects that turn out, decades on, to be the ones worth seeing.
The other version keeps the whole list. The things that were built while looking for something else. The false alarms that stayed in the sky after the real thing had gone.
Messier died in 1817 having found thirteen comets. We kept the other list.
Your side projects are on the chart too. Enter your GitHub username at codexstellarum.com and preview your atlas - free to try, prints from £25.