The Data We Couldn't Read: When Public Records Outlive Their Software

We often picture lost data as something deleted, burned in a server fire, or trapped on a corrupted hard drive. But there’s a quieter, more insidious form of loss happening right now in municipal basements and state archives: data that’s physically intact but functionally mute. It’s trapped in proprietary file formats created by software that no longer runs, a digital ghost town where the keys have been thrown away.

I stumbled into this while trying to help a historian friend. She was researching mid-1990s urban planning decisions and had been granted access to a treasure trove—a set of CDs labeled “City Council Assessments, 1994-1996.” The CDs were pristine, the files dutifully copied. But when we tried to open them, our modern machines just blinked back at us. The extensions were unfamiliar: .wpd, .apr, .fm. They were from WordPerfect, AppleWorks, and FileMaker Pro versions so old they’d been discontinued before the iPhone was invented. The data was there, a series of magnetic bumps arranged with precise meaning. But the decoder ring—the specific software that could interpret those bumps into a legible report—was gone.

The Preservation Paradox of the Proprietary

This is the paradox at the heart of so much “open data” and digital preservation work. We focus on saving the bits, the ones and zeros, and that’s crucial. But bits without their context are just noise. The proprietary software that created these files was the context. It defined not just how to draw a paragraph on a screen, but how the data within it was structured—which field was a date, which was a comment, where the margins of a database table were. Without that original software, or its meticulously preserved specifications, we’re left with digital archaeology, trying to guess at meaning from raw binary patterns.

It’s easy to blame this on bureaucratic neglect, but often, it was just standard procedure. A city clerk in 1995 wasn’t thinking about archival formats; they were using the software the department bought, the tool that got the job done. The record was created, the report was printed for the meeting, and a copy was saved to the newfangled CD for “permanent storage.” Job done. The idea that the vendor would cease to exist, or that the file format would become an unreadable relic, wasn’t part of the workflow. The preservation was an accident, and a partial one at that.

So, what do we do with these silent archives? The solution isn’t just technical, though emulation and format migration are vital tools. It’s also cultural. It requires us to see public records not as static documents, but as dynamic systems of information that include their means of interpretation. True open data and digital preservation must advocate for open standards and simple, durable formats (like plain text, CSV, or PDF/A) as matters of public accountability. Because a record you can’t read might as well not exist. The story of our decisions, our governance, and our collective life is at risk of being locked in a vault to which we’ve deliberately, if unintentionally, lost the only key.

Notes & further reading

A few pages I came back to while writing this: