The Unstable Foundation: On the Peril of Perfect Preservation
In the world of digital preservation, the highest commandment is often "preserve everything." We are taught that the ultimate goal is a pristine, unaltered copy of a file, a website, or a dataset, frozen in digital amber for all time. This is the gold standard, the mark of a successful archivist. But what if this pursuit of perfect fidelity is not just a Sisyphean task, but a fundamentally flawed one? What if the very act of striving for a perfect copy builds a more fragile, not a more resilient, archive?
The common logic is seductive. A bit-for-bit copy is the truth. Any deviation is corruption, a loss of authenticity. We build systems with checksums and fixity checks to police this boundary, sounding alarms at the slightest hint of change. But this view ignores a crucial, counterintuitive reality: the digital objects we seek to save were never static to begin with. They were born into a state of flux. A website is not a book; it is a process. Its links rot, its embedded media changes, its very rendering depends on a fleeting consensus between browser and server that evaporates the moment it is captured.
By insisting on a single, "perfect" version, we create a brittle monument. We are preserving the shadow and calling it the substance. The true nature of a dynamic digital resource is its behavior, its interactivity, its connection to a living network. A WARC file that perfectly preserves the bytes of a 2005 Flash website is, in a very real sense, a perfect preservation of a corpse. It has all the original parts, but none of the life.
Perhaps a more honest and ultimately more durable approach is to embrace a philosophy of managed reinterpretation. Instead of seeing change as the enemy, we could see migration and emulation not as necessary evils, but as the natural continuation of a digital object's life. The goal shifts from freezing an object in time to preserving its essential function and meaning, even if that requires translating it into a new language—a newer file format, a different rendering engine.
This is a scary proposition. It introduces subjectivity. It means that the archived version of a thing might look or behave differently than it did on a specific Tuesday in 1999. But it acknowledges that context is everything. A spreadsheet from a public database is worthless without the software to calculate its formulas; a 3D model is a meaningless string of numbers without a engine to render it. Preserving the function is, in many cases, a higher fidelity to the object's purpose than preserving its exact digital particles.
This isn't an argument for carelessness. It is an argument for a different kind of rigor—one focused on preserving meaning and utility rather than just bits. It asks us to build archives that are living gardens, not hermetically sealed vaults. It accepts that some loss is inevitable, but argues that by focusing on the spirit of the object rather than the letter of its code, we might just save more of what actually matters in the long run.
Notes & further reading
A few pages I came back to while writing this: