The Unwritten Rules of Web Preservation
We talk a lot about the technical mechanics of saving the web: the crawlers, the WARC files, the storage solutions. But behind these protocols and commands lies a more fundamental, and much thornier, question. What, exactly, are we trying to preserve?
It’s a question with a deceptively simple answer. We are preserving web *pages*, of course. We capture a URL, its HTML, its images, its scripts, and we store that bundle away. The Internet Archive’s Wayback Machine is the monumental testament to this approach. Yet, anyone who has ever used it knows the peculiar feeling of encountering a “saved” page that feels less like a preserved artifact and more like a ghost. The layout is broken. The images are missing. The interactive element you desperately wanted to click is frozen, a static picture of a button that will never again respond. We successfully preserved the page’s ingredients, but we lost the meal.
This points to a deeper, often unwritten rule of web archiving: the unit of preservation is not the page, but the experience. A webpage is not a document in the way a PDF is a document. It is a set of instructions to a browser, a performance that unfolds in real-time. A news article from 2006 might have had a live-updating stock ticker, a comments section buzzing with real-time reactions, and an embedded video player that used a now-defunct plugin. To truly preserve that article, we would need to preserve the capacity for it to *perform* as it once did. We would need to bottle not just the data, but the context.
The Context is the Content
This is the archivist’s impossible challenge. The modern web is a sprawling ecosystem of dependencies. It relies on external fonts, third-party APIs, content delivery networks, and JavaScript frameworks hosted on other servers. A single webpage can be a fragile assembly of components scattered across the globe. When we archive a page, we try to capture as many of these dependencies as possible, but it’s like trying to capture a conversation by only recording one participant. The context—the other speakers, the room’s acoustics, the social norms—is lost.
This leads to a phenomenon I call “archival drift.” The preserved page slowly drifts away from its original meaning as its context evaporates. A link that once pointed to a vibrant forum discussion now points to a 404 error page within the archive itself. A piece of JavaScript that pulled in the latest weather report is now frozen, displaying a forecast from the day it was crawled. The page is stable, but it is no longer truthful to the lived experience of browsing it when it was “live.” It has become a museum diorama: meticulously arranged, but fundamentally still.
So, what is to be done? The answer isn’t to abandon the effort. It’s to change our expectations and be more honest about the limitations. Perhaps the goal is not perfect fidelity, but meaningful preservation. We can document the context—note the dependencies, describe the interactive elements that were lost, explain what the page felt like. We can prioritize saving the core content—the text, the primary images—while acknowledging the loss of the ephemeral performance. The unwritten rule, then, is that web archiving is as much an act of interpretation and storytelling as it is a technical process. We are not just saving data; we are writing the footnote for a performance that can never be fully restaged.
Notes & further reading
A few pages I came back to while writing this: