Building a document archive that still opens in twenty years

By Michael, founder of DocAttic · September 2026 · 8 min read

Scanning a document is the easy part. The real question is whether the file you create today will still open on whatever computer exists in 2046 — whether the app that made it is still around, whether the format it's saved in still means anything, and whether the drive it's sitting on has survived. An archive that only works for a few years isn't an archive, it's a postponement. Here's how to build one that actually lasts.

The threat isn't loss, it's rot

People picture archive failure as a dramatic event — a fire, a stolen laptop, a deleted folder. Those happen, but the quieter failure is more common and easier to miss: format rot. A file survives perfectly intact on disk, every bit exactly as it was written, and it still becomes unreadable — because the program that made it stopped being sold, the file format it used was proprietary and undocumented, or the one company that understood it shut down. The bytes didn't go anywhere. The ability to make sense of them did.

This has already happened once to most people's digital lives — old word processor files, a tax program's proprietary save format, a scanning app's own container format that nothing else reads. The fix isn't a better app. It's choosing formats that don't depend on any one app existing.

Choose formats built to outlive the software that made them

Some formats are designed for longevity; others are designed to sell you the next version of a program. The difference matters more than almost anything else in an archive:

• PDF, ideally PDF/A — for scanned documents and anything text-based. PDF is an open, published standard, not a company's private format, and dozens of independent programs read it. PDF/A is the archival subset built specifically for long-term storage: it embeds its own fonts and avoids anything that depends on an external resource, so a PDF/A file made today is designed to render identically decades from now.
• Plain TIFF, PNG or JPEG — for raw scans and photos of documents. These are among the oldest, most thoroughly documented image formats in existence, and that's exactly why they'll still be readable long after anything fancier.
• Plain text or CSV — for any data you'd otherwise keep in a spreadsheet or database. A spreadsheet's native format can become a puzzle in twenty years; a plain text file is never a puzzle.

What to avoid: any format that only one paid program can open — a scanning suite's own proprietary container, an accounting tool's private save format, anything that locks the content behind a license that might not exist later. If a format needs a specific company to stay in business for you to read your own document, it's a liability, not a filing choice.

File-level durability: names and folders a human can read

The most future-proof filing system needs no software at all — just descriptive filenames inside a plain folder structure, browsable by anyone who can open a file explorer. A year-based structure works well: a folder per year, with a short, clear filename inside it, like 2019-employment-contract.pdf rather than Scan_0043.pdf.

Resist the temptation to bury everything inside one app's internal database or its own proprietary library format. A database file that only one program understands is exactly the kind of single point of failure format rot exploits — if that program disappears, so does your ability to browse what's inside. Plain files in plain folders can be opened, copied and read by anything, forever.

Bit-level durability: the 3-2-1 rule

Open formats solve the software half of the problem. The hardware half is just as real: every drive eventually fails, whether from age, a bad sector, or a simple accident, and a single drive is a single point of failure no matter how good the files on it are. The standard, boring, effective answer is the 3-2-1 rule:

• Three copies of anything that matters — the working copy plus two backups.
• Two different kinds of media — for example an internal drive and an external one, so a single bad batch or a single failure mode can't take out both.
• One copy off-site — at another location, so a theft, fire or flood at one place doesn't take out every copy at once.

Two habits make this actually work over decades rather than just once: verify your backups occasionally by actually opening a few files from them, not just trusting the copy completed, and migrate to new media every several years. Drives, discs and even cloud backup services age out; moving the archive forward periodically is part of keeping it alive, not a one-time setup task.

Consolidate before you preserve

Durability only pays off if what you're preserving is worth preserving. An archive that's really ten overlapping copies of the same tax return, scattered across old drives and Downloads folders, wastes backup effort on redundancy instead of on the one document that matters. Clean up first:

• Read how to find duplicate files and PDFs to identify what's genuinely repeated before you back anything up.
• Read how to organize scanned documents, PDFs and records for a plain year-based filing structure to preserve.
• Read how to organize files from old computers and drives if the source material is still spread across multiple old machines.

A clean, de-duplicated archive is smaller, easier to verify, and cheaper to keep three copies of than a messy one — consolidation isn't a separate chore from preservation, it's the first step of it.

Keep it local and independent of any one company

An archive you hold as plain files on drives you own will outlast an archive trapped inside a cloud service's account system — a service can raise its price, change its terms, get acquired, or shut down, and a locked-in archive goes down with it. Plain, open-format files on media you control answer to nobody's business decisions but your own. This independence is part of what the wider Attic family of tools is built around: local archives you own outright, not accounts you rent.

Frequently asked

What is the best format for long-term document storage?
PDF, ideally PDF/A, for documents; plain TIFF, PNG or JPEG for scans and photos; plain text or CSV for data. The common thread is openness — a format several independent programs can read, not one tied to a single paid app.

What is PDF/A?
The archival subset of the PDF standard, built for long-term preservation. It embeds its own fonts and avoids anything that depends on an outside resource, so a PDF/A file made today is designed to look the same decades from now.

How do I store documents so they last?
Save in open formats, use descriptive filenames in a simple folder structure, and follow 3-2-1 backup — three copies, two kinds of media, one off-site — with occasional verification and a move to fresh media every few years.

How long do files last on a hard drive?
There's no fixed number; drives can develop errors after five to ten years from age alone, and any single drive can fail without warning. The fix is redundancy across multiple copies and media, not a better drive.