The Data Librarian's Lumber Yard: Lessons from Wood Grain in Digital Preservation
We spend a lot of time talking about the bits, the bytes, the formats, and the protocols. It’s easy to think of digital preservation as an act of pure information science, a battle against obsolescence fought in the abstract. But sometimes, the most useful metaphors come from the physical world, from crafts that have dealt with preservation for millennia. Lately, I’ve been thinking about woodworking, and more specifically, about wood grain.
To a carpenter, wood grain isn't just an aesthetic feature; it’s the material’s biography. It tells the story of the tree’s life—the years of drought packed tight, the seasons of plenty, the direction of prevailing winds, the scars of lost branches. It dictates how the material will behave when cut, planed, or stressed. Working against the grain is a recipe for splintering and ruin. Working with it requires reading its history and respecting its inherent structure.
Our digital collections have grain, too. It’s just far less obvious. The ‘grain’ of a born-digital public record set might be the sprawling, inconsistent taxonomy its originating department used for twenty years. It’s the legacy database fields that were never populated but can’t be removed. It’s the specific political context in which a certain dataset was first released, which explains its curious omissions. It’s the idiosyncratic folder structure of a defunct civic website, a structure that hundreds of citizens learned to navigate. This is the internal narrative, the growth rings of the data.
Too often, in our zeal to ‘preserve’ or ‘standardize,’ we plane this grain away. We migrate data into pristine new schemas, flattening its quirks. We strip away the ‘messy’ contextual metadata that seems irrelevant to the core data points. In doing so, we risk creating a perfectly preserved but utterly splintered artifact. We’ve saved the lumber but lost the story of the tree, and in the process, we’ve made the material more brittle, less understandable to anyone trying to trace its origins or understand its original use.
The woodworker’s lesson is to study the grain before making a cut. For us, this means archival practice must prioritize deep contextual documentation—not just of the data’s content, but of its creation, its use, and its anomalies. It means that migration paths should aim to retain, not erase, traces of original structure. It means recognizing that sometimes the ‘flaw’—the odd field name, the duplicate record—is a critical part of the artifact’s authenticity, a knot in the grain that tells you where a branch once was.
By treating our collections with the respect a craftsperson shows to their material, we move beyond mere technical rescue. We commit to preserving not just the data, but its character and its history. We learn to work with its grain, ensuring that future scholars and citizens aren’t left holding a pile of perfectly milled, context-less splinters, wondering what kind of tree it ever came from.
Notes & further reading
A few pages I came back to while writing this: