The Myth of Perfect Metadata: Why We Should Embrace Messy Archives
In the hallowed halls of digital preservation, a sacred mantra is whispered like a prayer: good metadata is everything. We are taught that without pristine, structured, and comprehensive metadata, an archived object is a ghost—present but un-findable, saved but lost. This pursuit of perfection has become an orthodoxy, a gatekeeper that often prevents preservation from happening at all. It’s time we challenged this. Perhaps the most effective way to save our digital heritage is not by meticulously cataloging every last byte, but by embracing the beautiful, chaotic mess of the archive itself.
The common advice is paralyzing. A small historical society, a community activist group, or an individual researcher looking to archive a set of websites is often met with a daunting list of requirements. They are told they need a specific schema, controlled vocabularies, persistent identifiers, and a full suite of preservation metadata before they even begin. Faced with this insurmountable task, many simply don’t begin. The perfect becomes the enemy of the good, and in this case, the enemy of the saved. We let vast swathes of digital culture vanish because we were too busy drafting the perfect label for the box it might one day go into.
The Power of the Pile
Consider the traditional archive, the physical kind. Scholars don’t just rely on the official finding aid prepared by a meticulous archivist. They dig. They find connections in the marginalia of a letter, a forgotten receipt tucked inside a book, or the simple, un-cataloged proximity of one document to another. The mess is not a bug; it’s a feature. It preserves context and accidental relationships that overly rigid categorization would destroy.
Our digital archiving practices, in their quest for machine-readable order, often scrub this context away. We atomize collections into discrete, perfectly described items, severing the hyperlinks, the informal network of pages, and the lived experience of how the web actually functions. A WARC file, in its raw, ‘messy’ state, captures this lived experience—the broken links, the ad scripts, the personal blogrolls—all of which are vital to understanding the digital culture of a moment. By prioritizing perfect metadata, we risk creating a sterile, clinical representation of the past, one that is easily searchable but ultimately devoid of its original spirit and interconnectedness.
This isn’t an argument for no metadata. It’s an argument for ‘good enough’ metadata done at scale. It’s a plea for action over perfection. A URL and a timestamp are metadata. A broad subject tag applied by a volunteer is metadata. The text content itself, full of keywords and names, is a form of metadata. We have powerful full-text search tools that can sift through these messy piles far more effectively than any human could with a card catalog. The goal should be to capture the material first and worry about polishing its description later—or let future researchers and improved technology do that work.
By lowering the barrier to entry, we empower more people to become stewards of their own digital history. The most important metadata is the metadata that exists, not the metadata we wished we had. Let’s build messy, sprawling, chaotic archives now. We can always organize them later. But we can never recreate a website that has already blinked out of existence.
Notes & further reading
A few pages I came back to while writing this: