One work a day, and how that happened

The idea is small enough to say in one sentence: one work of literature a day, given twenty-four hours with nothing else competing for the light.

A recommendation engine hands you a hundred things and hopes one of them sticks. You scroll, you save a few, you read none, and the next day it hands you another hundred. So this hands you one. If it is not for you, it has cost you two minutes and tomorrow brings another. I revere these books, unfashionably so, and what I wanted was somewhere they could be met one at a time instead of stacked in a pile that only ever grows.

I am Hal, and I built it for myself. It was a spreadsheet first, then a small web app one weekend, because I was working through the canon and wanted to know where I actually was in it: what I had finished, what I had abandoned and been quietly relieved about, what I kept meaning to start and never did. Nothing I could find held that in a shape I liked.

Then I posted it on Reddit expecting tumbleweed, and people started using it. Daily. That changed it from a private tool into something I owed to strangers, which is a stranger feeling than it sounds, and it is why the thing has had three years of my evenings.

I loved books long before I loved code and have never quite managed to choose: literature at school, theoretical physics at Imperial, and software learned somewhere in the middle of the physics, which turned out to be the thing that let the two meet. I live in south-west London with a wirehaired Vizsla called Japhy, named for the Kerouac character, and most days start with a long walk by the river with a poet in my ears. This app is my love letter to literature and it was mostly built on those walks.

What counts as literature here is writing as an art form, which is why the days lean to the novel, the poem, the story and the essay, and why a work of philosophy or science gets in only when the author wrote with enough style that the prose itself becomes part of the argument. The canon is not a ranking and makes no claim to be definitive. It is one argued-over attempt at breadth, and it grows.

Then each work is matched to a date, which is the part I have never been able to stop doing. Birthdays and deaths, feast days and festivals, the day a book was published or banned or set, what its themes have to say to one particular week of the year. Hamlet on Shakespeare’s birthday. Ulysses on Bloomsday. Wuthering Heights on Valentine’s Day. Gogol’s The Overcoat on Black Friday, which I remain pleased with. Wordsworth for the week the daffodils come out, Dumas for a long summer with nothing to do but turn pages, ghost stories for the dark end of the year, and the big Russian novels for the grey weeks of January when you have nowhere else to be.

The habit came from two places. The first was Tolstoy’s A Calendar of Wisdom, which turned reading into a ritual rather than an ambition. I liked it enough that the whole thing is in the app here. The second was Ray Bradbury’s famous challenge: one short story, one poem and one essay every night, for a thousand nights. That one is in the app too, here, a year at a time.

Between them, that’s most of what I believe about reading. A little every day, kept up for years, gets you further than any resolution. The wider the range the better, so a novel one day might be a poem the next. And I have a real love of the short story, a whole world finished before the coffee goes cold.

The hardest part was never the code, it was the words. “What do you read my lord?”, “Words, words, words”. (Hamlet). I used to write book reviews, not for a magazine, just for myself, so I decided, with the confidence of a man who had not counted, that I would write every blurb myself. For a while I did. Then I ran out of books I had actually read, and told myself the holes were fine. Friends filled the others.

So here is where we have landed, and it is the thing I would most want to know if I were you. There are people writing for the app who know the books and can write about them with precision and love, and their names are at the bottom of everything they write. Where we cannot keep pace, a language model, Anthropic’s Claude, drafts a first version against a house style guide we wrote. Nothing it drafts reaches you unread: every line is checked against the book, edited, often rewritten and quite often thrown away, and a person signs it off. A model sometimes writes a first draft, but it never has the last word.

If you know the books and have something to say about them, please write to us. A reader who has lived with a book is worth more to this app than another draft.

Think of this, and of canon-reading in general, as fishing. Most days nothing bites and you move on to another stretch of river. But every now and then something does, and you know it immediately: a sentence stops you, a writer reaches you sideways, a book follows you around for weeks. We cannot tell you which works will bite. We can only keep putting enough of them in front of you that some of them will.

Not everything will appeal, and not everything is for right now. That is fine, and it is rather the point.

The app has grown faster than its signposting. If you get lost, that is on us and not you: tell us what you cannot find, and what you would like to see. We have been known to add a book the same day someone asks for it.