The Fallacy of the Universal Format: How Our Obsession with 'Future-Proof' Files Betrays the Past

In the world of digital preservation, there’s a comforting piece of received wisdom: to save something for the future, we must migrate it into a universal, open, non-proprietary format. Convert those WordPerfect files to ODT. Render those CAD drawings as STEP. Flatten those layered Photoshop documents into uncompressed TIFFs. The logic is sound on the surface—proprietary formats are locked boxes, while open standards are master keys. But in our rush to future-proof, we often perform a quiet act of historical violence. We preserve the data, but we sever it from its native environment, stripping away the very context that makes it a record and not just raw information.

Consider a municipal budget from 1997, originally crafted in an early spreadsheet program. Our preservation instinct says: export it to CSV. The numbers are saved, perfectly readable by any text editor until the end of time. But what have we lost? The carefully set print areas that show what the council members actually saw on their handouts. The cell comments where a clerk explained an odd outlier. The specific font choices and borders that signaled sections of importance. The formulas, which are the logic of the document itself. The CSV is a ghost—a list of values without a soul. We have traded fidelity for fungibility.

The Native Environment as Artifact

This obsession with the universal format assumes that data is separate from its software, that content can be neatly extracted from its container. But for anyone who has used technology, the experience is holistic. The clunky menu, the particular way a cursor blinks, the default palette of colors—these aren’t bugs; they are the aesthetic and functional reality of the era. A flyer designed in 1995 using Microsoft Publisher is a cultural artifact of that software’s limitations and affordances. Preserving only a PDF of it is like preserving a photograph of a sculpture. You get the shape, but not the texture, the scale, or the way the light falls on it in its intended space.

Our drive for the universal often stems from a fear of emulation—the practice of preserving old software environments to run old files. It’s seen as complex, fragile, and niche. A plain text file feels safe. But this ‘safety’ is an illusion of simplicity. It ignores the fact that understanding a record often requires understanding the tools that created it. The undo history of a document, the revision tracking, the unused layers in an image—these are the ‘metadata’ that live natively in the proprietary format, and they are often the first things destroyed in a conversion to something ‘open’.

The true lesson might be that there is no universal format, only different kinds of loss. The open standard may guard against the loss of access due to corporate obsolescence. But the native format, preserved alongside the software and system context needed to render it, guards against the loss of meaning and experience. A complete preservation strategy must embrace this uncomfortable duality. It must value the ‘messy’ original—the file in its natural habitat—just as much as the clean, extracted derivative. Because sometimes, to keep a thing alive, you don’t just preserve the specimen in a jar; you preserve the entire ecosystem it came from. The future may not need just our data, but a way to understand how we worked, thought, and saw the world through our own machines.

Notes & further reading

A few pages I came back to while writing this: