The Archivist's Palate: How Wine Vintages Teach Us About Digital Provenance

In the cellars of Burgundy, a winemaker doesn’t just produce a bottle of wine; they curate a story. The story is one of a specific patch of earth, a particular year’s weather, and the meticulous process of transformation. This story, known as provenance, is everything. It’s the difference between a generic red blend and a Grand Cru that speaks of its origin. This ancient practice of documenting a product’s complete history, from ground to glass, holds a profound and unexpected lesson for those of us trying to preserve the digital world.

We often think of archiving a webpage or a dataset as a single act of capture—a snapshot frozen in time. But a wine’s vintage teaches us that true preservation is about capturing a continuum. A 2015 Romanée-Conti is not just about the grapes harvested in 2015; its character is understood in relation to the wetter 2014 vintage that preceded it and the sun-drenched 2016 that followed. Similarly, a dataset is not an island. Its full meaning is derived from its versions, its corrections, the context of its creation, and the datasets it superseded. To archive it in isolation is to lose its vintage character.

Beyond the Snapshot: The Terroir of Data

Winemakers have a word for this context: terroir. It’s the complete natural environment in which the wine is produced. Our digital artifacts have a terroir, too. It’s the now-defunct API it was pulled from, the specific version of the software that generated it, the political climate the day it was published, and the hyperlinks that once connected it to a wider web of information. Standard metadata fields like ‘date created’ and ‘creator’ are the equivalent of listing only the grape variety. They tell you something, but not the full story. To truly preserve something, we must strive to document its digital terroir—the ecosystem that gave it meaning.

This vinicultural analogy also reframes the concept of preservation failure. A corked bottle of wine isn’t just ‘broken’; it’s a loss of that specific story from that specific place and time. You cannot recreate it. The same is true for a corrupted database or a website crawled without its stylesheets. The loss isn't merely of ones and zeros, but of a unique historical moment’s context and character. The data may be recovered, but its vintage quality—its authenticity within a precise digital environment—is gone forever.

Perhaps the most valuable lesson is one of patience and perspective. Great wine is made not just for immediate consumption but for future understanding. Archivists and data stewards are, in a sense, vintners for the future. We are bottling up the digital harvest of our era, not knowing precisely when it will be opened or what questions it will need to answer. Our job is to ensure that when that future researcher uncorks our work, they don't just get raw information. They get a taste of the time itself, with all its complexity, nuance, and origin intact. They get the full provenance.

Notes & further reading

A few pages I came back to while writing this: