Nine Centuries on Parchment: The Data Retention Lesson Britain Keeps Failing to Learn

A Dataset That Refused to Die

In 1086, William the Conqueror's commissioners recorded the taxable wealth of most of England — landholdings, livestock, mills, plough teams — in what became the Domesday Book. Nearly a millennium later, historians still read it. The parchment is intact, the ink legible, the semantics (with scholarly effort) recoverable.

Now consider the typical modern counterpart. Corporate datasets routinely degrade in usefulness within ten to fifteen years, not because the bits decay but because the context does: formats fall out of support, applications that interpret the data are decommissioned, and nobody documented the schema. We have built the least durable information systems in British history and convinced ourselves this is normal.

The 1986 Embarrassment

The BBC's Domesday Project marked the original's nine-hundredth anniversary with a laser-disc and microcomputer-based multimedia survey, compiled from contributions by over a million people, largely schoolchildren. It was a genuine technical triumph — and within sixteen years the custom hardware was obsolete. By the early 2000s the discs were effectively unreadable. A modern digital edition of a 900-year-old record nearly perished in a single generation.

Rescue efforts eventually recovered the content, but the lesson stands: preservation is not a property of media, it is a property of ongoing stewardship. The parchment survived through continuous institutional care. The laser disc failed the moment institutional attention lapsed.

Why UK Organisations Should Care

The obligations are real and various. Companies must retain accounting records for at least six years. NHS records follow retention schedules stretching across decades. Financial services, under FCA rules, face long-tail record-keeping duties. Local authorities hold planning and environmental data with century-scale relevance. The gap between statutory obligation and technical practice is where audit failures and data-loss incidents breed.

A Practical Longevity Checklist

• Prefer boring formats. CSV, plain text, and well-documented XML will outlive every proprietary binary you are tempted by. If a format needs a licence to read, it is a liability.
• Separate data from application. If your records can only be interpreted by one version of one application, you have stored a program, not data.
• Document the semantics. A schema file is not enough. Write down what each field means, in prose, dated and versioned.
• Schedule migration, don't improvise it. Review stored formats on a fixed cycle — every three to five years — and migrate proactively while the source format still has tooling.
• Test restoration, not just backup. A backup nobody has ever restored is a hope, not a control. Attempt full recovery annually.
• Assign ownership across decades. Data outlives teams, projects and platforms. Retention duties need a named owner in the organisation, not in a repository.

The Stewardship Mindset

The Domesday Book's durability was never an accident of materials alone. Monastic custodians copied, stored and protected it through centuries of institutional upheaval. The 1986 project failed not because its technology was poor but because nobody planned for the day the technology would be ordinary.

Longevity, in short, is an architectural and organisational commitment. Parchment simply made the commitment visible.