The Obsolete File: On Dying Formats and the Ghosts in the Machine
We often praise the virtues of open data in the abstract, celebrating the sheer volume of information made available for public good. We champion the act of preservation itself, building digital fortresses to withstand the erosion of time. But there’s a quieter, more insidious problem that all the terabytes in the world can’t solve if we ignore it: the silent crisis of the executable context. We’ve been sold a comforting idea—that preservation is about the file, the bitstream, the perfect copy. But what happens when the file is perfectly preserved, yet utterly, irrevocably mute?
Consider the .fla file. For over a decade, it was the lifeblood of a certain kind of digital creativity. It was the source file for Adobe Flash, the engine behind a wild, innovative, and often wonderfully bizarre era of the web. Countless animations, interactive art projects, and early web games were born as .fla files. Today, we have meticulously archived millions of these files. They sit safely on servers, checksums verified, pristine and untouched. And yet, they are digital ghosts. The software required to open them, to interpret their specific structure, to give them life, is gone. Adobe Flash Player is deceased, and the authoring environment that created these files is a relic. The data is open, but its meaning is locked in a tomb of obsolescence.
This is the flaw in the received wisdom of simple preservation. We treat files as self-contained artifacts, like books on a shelf. But a book, while requiring a literate reader, makes few other demands. A digital file, especially a complex one, is not a static text; it is a set of instructions. It requires a specific, often proprietary, interpreter to enact those instructions. Without it, the preserved file is not a document but a puzzle without a key—an ornate lock with the combination lost to time. Our archives are filling up with such ghosts: CAD drawings from defunct engineering software, early hypertext documents from forgotten platforms, scientific data sets saved in formats that no current program can parse.
The challenge, then, is not merely to preserve the object but to preserve the means of its execution. This is the true frontier for open data and digital preservation. It’s a far messier and more complex endeavor. It involves archiving the software itself—the operating systems, the runtime environments, the dynamic-link libraries—and the legal and technical frameworks to run them in emulated spaces. It’s about preserving the context, the ‘how’ alongside the ‘what’. Initiatives like software preservation networks and emulation-as-a-service platforms are the new librarians, not just cataloging books but ensuring the lightbulbs in the reading room still work.
Every time a popular software platform sunsets, it leaves behind a digital ghost town. The buildings are all there, the streets perfectly mapped in our open data repositories. But the power is off. The real work of the archivist now is not just to be a keeper of static records, but an electrician for the past, desperately trying to reconnect the wires so that the lights, and the life, can flicker back on. The file is not enough. We must also save the spark.
Notes & further reading
A few pages I came back to while writing this: