The Unspoken Language of the Archive: What Do We Lose When We Only Save the Text?
We talk a lot about saving web pages, but we rarely ask what, precisely, we mean by a ‘page’. In the grand, meticulous work of web archiving and digital preservation, the focus often narrows to the textual content. We capture the words, the primary data, the ‘information’ itself. This is a monumental and crucial task. But in this necessary triage, a subtle, silent language is often left to fade: the language of layout.
Consider a government report from a decade ago, perfectly preserved as a clean, text-only PDF. Every fact and figure is present and accounted for. But the original web page it came from told another story entirely. The urgent press release was crammed into a narrow sidebar, visually shouting its importance. The annual budget, a dense and intimidating table, was tucked away at the very bottom, requiring six scrolls to find it. The ‘Contact Us’ link was a bright, impossible-to-miss button, not a string of underlined text.
This visual rhetoric is not a decorative afterthought; it is a core component of the record. It tells us what the creating institution believed was urgent, what was secondary, and what it perhaps hoped you wouldn’t easily find. The proximity of one piece of information to another, the size of a font, the choice of a bold red over a calm blue—these are all semantic choices. They are the non-verbal cues of bureaucracy and public communication. To lose them is to receive a letter where all the words have been typed in a uniform, monotone font, stripped of its underlines, its emphatic ALL-CAPS, its anxious postscript scribbled in the margin.
The Ghost in the Machine
This presents a profound challenge for archivists. Capturing the full render of a website—its look, feel, and function—is a resource-intensive nightmare. It requires saving not just text, but stylesheets, fonts, JavaScript, and a thousand other dependencies that are notoriously brittle over time. The text-only version is a pragmatic, stable, and searchable solution. It is the preserved skeleton, long after the skin and muscle have decayed.
But we must acknowledge that a skeleton can only tell us so much about the creature it once supported. Future historians studying our digital present will have an immaculate record of what we said, but a deeply fractured understanding of how we meant it to be encountered. They will see the data, but they may miss the narrative embedded in the design. They will know the policy, but not its prominence.
The next frontier in public record preservation isn’t just about saving more data; it’s about finding smarter, more sustainable ways to capture context. It’s about developing practices that can somehow bottle the ghost of user experience—the weight of a click, the path of a scroll, the hierarchy of a homepage. Because the truth of a public record often lives not in the paragraph, but in the pixels surrounding it.
Notes & further reading
A few pages I came back to while writing this: