The Sourdough Starter of the Internet: Lessons for Preservation from Live Cultures

We talk about digital preservation as if it’s a task for cold storage. We want to freeze a webpage, a dataset, a moment in the network’s flow, and lock it away in a perfect, unchanging state. The dream is a digital fly in amber. But what if the things we’re trying to save are less like static specimens and more like living cultures? This isn’t just a metaphor; looking at the ancient, ongoing practice of maintaining sourdough starter offers some surprisingly practical lessons for keeping our digital commons alive.

First, a starter isn’t a single, finite thing you create and possess. It’s a dynamic colony of wild yeast and bacteria, maintained through regular feeding—flour and water—and use. Neglect it, and it starves; hoard it in the back of the fridge forever without attention, and it eventually becomes inert. Our web archives and public datasets face a similar fate. A dataset dumped once and forgotten, without a plan for updates, corrections, or format migration, isn’t preserved. It’s just data in a coma. Preservation, like maintaining a starter, requires an ongoing ritual of care.

Second, a healthy starter thrives on use and sharing. Bakers often give a portion of their starter to a friend, creating a new, resilient branch of the same culture. This decentralizes the risk. If one jar is forgotten or contaminated, the culture survives elsewhere. Contrast this with our reliance on monolithic, institutional web archives. While invaluable, they represent a single jar. The practice of ‘distributed preservation’—where libraries, communities, and individuals keep their own copies of critical datasets or websites—creates a network of resilience. It’s the digital equivalent of sharing the starter, ensuring no single point of failure can erase our shared record.

Finally, and most importantly, a starter is defined by its activity, not just its ingredients. You can’t fully document a starter by listing its chemical components. Its true essence—its rise, its flavor—emerges from the complex, living process. Similarly, archiving a modern web application isn’t just about saving the HTML, CSS, and images. It’s about preserving some semblance of the interactivity, the APIs it called, the way it felt to use. We’re learning that to save a dynamic digital object, we sometimes need to preserve its ability to ‘bubble,’ even if in a limited, emulated environment. The goal shifts from a perfect snapshot to preserving a functional echo of the living process.

The lesson from the baker’s counter is that preserving living things requires embracing continuity, sharing, and aliveness over static perfection. For those of us tending the digital larder, it means our work is never truly done. It’s a craft of regular feeding, of passing on copies, and of learning how to keep the dormant sparks of interaction alive. We’re not just filing clerks; we’re cultivators.

Notes & further reading

A few pages I came back to while writing this: