Gardeners and Builders: The Architectural Divide in Digital Preservation

I recently spent an afternoon trying to recover an old project from a stack of decaying CD-Rs. The struggle wasn’t just with the physical media, but with the software it depended on—a specific version of a program that no longer exists. It was a stark reminder that preserving a digital object is one thing; preserving the world it lived in is another. This challenge highlights a fundamental schism in the digital preservation community, a divide between two philosophies I’ve come to think of as the Gardeners and the Builders.

The Gardeners take an ecological view. Their primary goal is to keep the original artifact alive and accessible, much like a gardener tends to a rare plant. Their tools are emulators and virtual machines—entire digital ecosystems meticulously recreated to host the original software. When you use an emulator to play a 1990s video game, you are experiencing the Gardener’s work. The data on the disk is the authentic artifact, and the emulator is the lovingly maintained soil and climate that allows it to run. The virtue of this approach is its fidelity; you see the software, bugs and all, exactly as it was. The drawback is the immense, ongoing effort to cultivate these digital greenhouses, which must be constantly updated as the underlying hardware and operating systems they run on themselves become obsolete.

The Builders, in contrast, are architects of the future. They are less concerned with preserving the original runtime environment and more focused on ensuring the data's core functionality and meaning can survive. Their tool of choice is migration—translating a file from an obsolete format into a modern, open one. Think of converting a WordPerfect document to a plain text file, or a proprietary CAD drawing into an open standard. The Builder’s work prioritizes accessibility and longevity over perfect authenticity. You lose the specific look and feel of the original, but you gain a version that is far more likely to be readable by future systems. The risk, of course, is that something is always lost in translation—a specific font, a hidden macro, the precise layout that gave the document its original character.

This isn’t a battle with a right answer. Each approach serves a different purpose. The Gardener’s method is essential for historical accuracy, for art, for complex software where the experience is the artifact. We need Gardeners to show us what the first web pages really looked in Netscape Navigator. The Builder’s method is pragmatic and scalable, crucial for preserving the informational essence of public records, scientific datasets, and legal documents where the content, not the container, is paramount.

The tension between these two camps isn't a weakness; it's the sign of a healthy, thoughtful field grappling with an impossibly complex problem. The real danger lies in relying on one approach to the exclusion of the other. A world of only Gardeners would be a museum of beautiful but inaccessible curiosities. A world of only Builders would be a library of potentially altered, flattened texts. Our best hope for a durable digital memory is to support both—to fund the builders constructing the robust libraries of the future, while cherishing the gardeners who keep the original digital flora alive.

Notes & further reading

A few pages I came back to while writing this: