The Whispering Codex: On the Necessity of Unreadable Data

We speak of digital preservation with a vocabulary of clarity: access, readability, open formats. The goal, we say, is to ensure that a record can be seen, parsed, and understood by anyone, forever. It is a noble pursuit, built on the promise of perfect transparency. But I find myself increasingly drawn to the opposite, to the value of the unreadable, the corrupted, the file that stubbornly refuses to yield its secrets.

Consider the humble floppy disk, not one from a government server, but one found in a box in an attic. Its label, handwritten and faded, reads ‘JAN1994’. You acquire the correct hardware, you spin it up, and instead of a tidy collection of WordPerfect files, you are met with a low-level read error. The disk has degraded. The magnetic particles have lost their cohesion. The data is there, but it is a ghost of its former self—a whisper of ones and zeros that can no longer be assembled into a coherent whole.

We would call this a failure of preservation. And in a practical sense, it is. The letter to a friend, the draft of a novel, the household budget—that specific information is lost. But what remains is a different kind of record altogether. The corrupted sector is itself a document. It is a testament to the physicality of digital objects, a blunt artifact of time’s passage written in the language of magnetic decay. It speaks not of what was written, but of the medium itself, of the fragility we so often forget in our age of cloud storage and endless redundancy.

The Aesthetics of Loss

This is not an argument for negligence. It is, rather, a meditation on the layers of meaning inherent in any preserved object. A perfectly migrated text file tells one story: the content. A corrupted file, or data in an obsolete proprietary format for which the key has been lost, tells another, richer story. It speaks of its own creation, its technological moment, and its journey through time. It becomes a digital potsherd—a fragment whose value lies as much in its breakage as in its original form.

In our rush to make everything open and perfectly readable, we risk subjecting all data to a single, homogenizing standard of value. We risk creating a future record where every byte is pristine and functional, but where the texture of technological history has been smoothed away. The struggle to read a thing, the faint echo of its original operating environment, the very fact of its obsolescence—these are not bugs in the archival process. They are features of a deeper, more honest history.

Perhaps true preservation isn't about stopping time, but about faithfully recording its effects. It is about keeping not just the message, but the whisper of the medium that carried it. The unreadable data stands as a quiet, crucial monument to everything that surrounds information: the context, the decay, and the beautiful, inevitable silence that follows.

Notes & further reading

A few pages I came back to while writing this: