The Wrong Kind of Redundant: How Over-Preservation Can Erase Context
In the world of digital preservation, redundancy is a sacred cow. The logic is sound, almost self-evident: to fight the entropy of bit rot and link rot, you make copies. Lots of copies. You store them in different geographic locations, on different media types, with different institutions. The mantra "Lots of Copies Keep Stuff Safe" (LOCKSS) is a foundational principle. But what if, in our zeal to save everything, we are accidentally building archives that are perfectly preserved yet contextually hollow? What if our redundancy is the wrong kind?
Consider a single PDF document, a city council meeting agenda from 2003. A diligent archivist, following best practices, ensures it is saved in three separate digital repositories, its file format normalized for future readability, and its metadata meticulously recorded. By all technical accounts, this document is preserved. Yet, it exists in these archives as a lonely satellite. The sprawling, ramshackle GeoCities website that hosted the original PDF, complete with visitor counters and animated under-construction GIFs, is gone. The local news blog that hyperlinked to it, providing critical commentary, has evaporated. The web forum where citizens discussed its contents is a ghost town whose whispers are lost. We saved the star, but we let the entire constellation blink out.
This is the paradox of over-preservation. We focus so intensely on the object—the file, the record—that we sacrifice the ecosystem that gave it meaning and weight. A digitally preserved artifact without its native habitat is like a single piece of driftwood on a vast, empty beach. You can analyze the wood, you can ensure it doesn’t decay, but you’ll never understand the forest it came from or the storm that washed it ashore. The context is the story, and our current fixation on individual object redundancy can systematically erase it.
The Aesthetics of Loss
There’s an uncomfortable truth here: loss is not always the enemy. Historians have long understood that the selective pressures of decay shape our understanding of the past. The documents that survive from the medieval period tell a story, but so do the documents that were lost to fire, flood, or neglect. In the digital realm, where we theoretically have the power to save everything, we risk creating a suffocating, undifferentiated mass of data. The aesthetic of a 1990s webpage—the garish colors, the clunky layouts—was part of its communicative power. When we strip a document of its original presentation, saving only the raw text, we are performing a kind of normalization that flattens history.
The challenge, then, is not merely to preserve more, but to preserve smarter. This means shifting resources toward capturing contextual relationships. Archiving a single webpage is a start; archiving the network of pages it linked to and that linked to it is the real goal. It means valuing the capture of an entire interactive web application, with its functional quirks, over the static screenshot of its homepage. It’s a messier, more computationally expensive task, but it’s the only way to preserve meaning alongside data.
Perhaps we need a new principle to sit alongside LOCKSS: CUKSS, or Contextual Understanding Keeps Stories Sound. The goal should not be an immortal, isolated record, but a preserved moment in time, with all its tangled, fragile connections intact. Because saving a single star is a technical achievement, but saving the constellation is what allows us to navigate the dark.
Notes & further reading
A few pages I came back to while writing this: