The Case for the Unreadable File: Why Obsessing Over Open Formats Misses the Point
In the world of digital preservation, there is a sacred mantra, repeated at conferences and in best-practice guides with near-religious fervor: "Preserve in open, non-proprietary formats." We are told to convert .docs to .odt, .xls to .csv, .psds to .tiffs. The logic is sound and seemingly unassailable. Open formats are more likely to be renderable by future software, ensuring long-term accessibility. They are the bedrock upon which we build our digital archives. But I want to propose a heretical thought: this single-minded focus on the "open" end-state can inadvertently erase the very context we seek to preserve.
Consider a simple memo, drafted in Microsoft Word in 1995. The standard preservation advice is to export it as a plain text file or a PDF/A. But in doing so, we strip away the metadata that is its hidden signature: the author's name embedded in the file properties, the revision history that might show a contentious edit, the now-antiquated font choice that screams its era. More importantly, we lose the artifact itself. A plain text file is not what the author created and sent; it is a derivative, a translation. The original .doc file, with all its proprietary, clunky, and potentially unreadable glory, is the primary source. The converted file is an interpretation.
The Artifact Versus the Information
This is the crucial distinction we often blur: the difference between preserving information and preserving an artifact. For a data scientist mining decades of climate records, a folder of perfectly formatted CSV files is a dream. For a future historian studying early 21st-century corporate culture, a folder of pristine PDFs is a sterile wasteland. They need to see the clunky PowerPoint presentations with their garish transitions, the Excel spreadsheets with their hidden columns of personal calculations, the Word documents with tracked changes and comments that reveal the human negotiation behind official policy.
These digital artifacts are the equivalent of finding a centuries-old letter with wax seals, unique handwriting, and ink blots. We wouldn't transcribe that letter into a standard typeface on a clean sheet of paper and discard the original, calling the job done. We preserve the physical object because its form is part of its truth. The proprietary format is the digital equivalent of that paper, ink, and wax. It is the authentic vessel.
This isn't an argument to abandon open formats. It is a plea for a more nuanced approach. Our preservation strategy must be two-pronged. Yes, we should absolutely create accessible, open-format versions for practical use and to guard against format obsolescence. But we must also zealously guard the original bitstream, the unaltered, proprietary file in all its obscure complexity. Our duty is to preserve the thing itself, not just a sanitized version of its content. We must save the digital inkblots along with the transcribed text, for they tell a story all their own.
Notes & further reading
A few pages I came back to while writing this: