The Recipe as Archive
On building alcatra.pt, and what it means to document something that lives in kitchens
There is a beef stew from Terceira Island, in the Azores, that has been made more or less the same way since the 1450s. It is called alcatra. It is cooked in unglazed clay pots, sealed with lard and dough, slow-braised for hours with wine and onion and allspice and bay. The older the pot, the better — each batch seasons the clay a little more, leaving something of itself behind for the next. In Terceira, the pots are heirlooms. People speak of them the way other people speak of land.
I grew up eating this stew. A few months ago I decided to build a website about it.
I want to write about what happened in between those two sentences, because it turned out to be less straightforward than I expected — not technically, but in the way that any project about something you love is never really about the thing. It is about you, and what you think is worth keeping, and how much you trust the internet to hold it.
IThe Problem with Documenting Living Things
The first question I had to answer was not which JavaScript framework to use. It was whether a website was even the right form.
Alcatra is not a recipe in the way most people use that word. It is a practice. It lives in the hands of the people who make it, in the specific clay of specific pots, in the arguments about how long to braise and whether to use chouriço and what kind of wine. It is transmitted the way all the best things are transmitted — by proximity, by repetition, by watching someone do it in a kitchen until you know it in your body.
A website can hold text and images and maps. It cannot hold the smell of the stew after four hours in the oven, or the specific motion of sealing the pot with your thumb, or the way the color of the broth tells you whether you’ve gotten the allspice ratio right. Whatever I built would be, at best, a shadow of the thing.
I built it anyway, because shadows are not nothing. They record the position of what casts them. And there are people in California and New Jersey and Ontario and Brisbane who grew up eating this food and don’t know why their grandparents’ version tastes different from what they find when they visit Terceira, and don’t know that the version they grew up with — the diaspora version, adapted to different clay and different wine and different schedules — is itself a document of the migration that brought their family to wherever they are now. Telling that story felt worth doing even imperfectly.
IIChoosing the Right Weight
Once I decided to build it, I had to decide how.
My first instinct, trained by years of professional software development, was to reach for a framework. React, probably. Maybe Next.js. A headless CMS. A proper data model. All the apparatus of modern web development, applied to the problem of a stew.
I opened a new project, typed npx create-next-app, and then sat there for a while thinking about what I was actually doing.
The site I was imagining was, at its core, a list of recipes and a few maps. The data was not going to change more than a few times a year. The audience was diaspora communities and food historians and people who were curious about Azorean culture — not particularly likely to be running performance-sensitive applications on low-end devices, but also not people who needed anything elaborate. The site needed to be readable, beautiful, informative, and fast. It did not need a React reconciler.
I closed the terminal and started with an empty HTML file.
This is not a manifesto for vanilla JavaScript. Frameworks exist because they solve real problems, and most projects I work on professionally have those problems. But this project did not. It had a different problem: I wanted the code to feel proportionate to the content. Alcatra is made with six ingredients. The website that documented it should be made with a similar kind of restraint. There is a coherence between the subject and the approach that feels important to me, even if I can’t fully articulate why.
The finished site is about 50 kilobytes of assets. It loads instantly on any connection. There is no build step, no dependency tree, nothing to update or secure or worry about. It will still work correctly in ten years. I cannot say that about most of what I build.
IIIColor Drawn from Photography
One thing I did that I am proud of: I derived the entire color palette from photographs of the dish.
Not from a mood board, not from a color theory exercise, not from what I thought would look good. I looked at photographs of alcatra — the deep wine dark of the broth, the terracotta of the clay pot, the green-grey of the island landscape, the cream of linen tablecloths at festival tables — and I built the palette from what I saw. Every color on the site has a referent in the physical world of the thing it documents.
This probably sounds like a small thing. It is not, to me. So much of what we build on the web is visually disconnected from its subject matter — aesthetic choices made in the abstract, borrowed from design trends or brand guidelines or whatever Dribbble featured last month. The result is sites that all look like each other, regardless of what they are about.
I wanted alcatra.pt to look like alcatra. I don’t know if I fully succeeded. But the attempt itself changed how I thought about the rest of the design decisions. Every choice became a question: does this serve the thing? Or is it serving me?
IVDiaspora as Data
The maps were the part I thought most about.
Terceira is a small island in the middle of the Atlantic. But over the last two centuries, its people have spread across the world — to New England, to California, to Ontario, to the South Island of New Zealand, to the coast of Queensland. They brought alcatra with them. They adapted it to local ingredients. They made it for the Holy Spirit Festivals that followed them wherever they went, those feasts that mark the spring and feed the whole community and have been unbroken in some Azorean towns since the thirteenth century.
When you put pins on a map showing where people are making this stew today — Sacramento, New Bedford, Hamilton, Parramatta — you are not showing food. You are showing migration. You are showing where boats landed and families settled and communities formed around the continuity of a practice that connected them to a place they had left.
I did not expect to feel the weight of that until I started entering the coordinates. There is something about the visual representation — all these scattered points, each one a kitchen, each kitchen a thread running back to a small island in the Atlantic — that made concrete something I had only ever felt abstractly. The diaspora is not an abstraction. It is specific people in specific places making a specific stew.
The maps are the best thing on the site.
VWhat a Website Can and Cannot Hold
I have been thinking, since I built this, about the relationship between documentation and the thing documented. About what it means to try to preserve something that is fundamentally alive.
The honest answer is that alcatra.pt does not preserve alcatra. It cannot. What it preserves is a record of alcatra at a particular moment — the recipes that existed, the communities that made them, the geography of the diaspora in 2024 and 2025. In fifty years, if someone finds this site, they will find a snapshot of a living practice, the way you might find a photograph of someone’s grandmother in her kitchen. The photograph is not the grandmother. But it is not nothing.
I think this is the most important thing I learned from building it: documentation is not preservation. Documentation is testimony. It says: this existed, in this form, and these people made it, and it mattered to them. The act of testifying has value independent of whether it succeeds in keeping the thing alive.
The clay pots will keep being used. The stew will keep being made. The festivals will keep happening in Oakland and Ludlow and Cambridge and Whanganui. I did not save any of that by building a website. But I said: I see you. This matters. Here is what it looks like from here.
That feels like enough.
alcatra.pt is live. If you have a family recipe you’d like to contribute, there’s a submission form on the site. I’d like the archive to grow.