PDF/A Explained: The Format Built to Outlive Your Software

Will a PDF you save today still open correctly in the year 2050? For most files the honest answer is 'probably' — and for a national archive, a court record, or a pharmaceutical trial dossier, 'probably' is not good enough. That gap between probably and certainly is exactly what PDF/A exists to close. The 'A' stands for archiving, and the format is an ISO standard (ISO 19005) that defines a stricter, self-contained subset of ordinary PDF.

The core idea is simple: a PDF/A file must carry everything needed to reproduce its pages, forever, inside itself — and must not contain anything whose meaning could change over time. In practice that means fonts must be embedded (no 'use whatever Helvetica the reader has'), colors must be defined with explicit color profiles, and all content must be present in the file rather than fetched from elsewhere.

Equally important is what PDF/A forbids. No JavaScript, no audio or video, no encryption, no external references, no transparency in the oldest profile. Each banned feature is a bet the archivists refused to make: encryption ties readability to key management surviving for decades; scripts tie rendering to engines that will not exist; external links rot. If a feature could make the document render differently in fifty years, it is out. This is also why a PDF/A file is often slightly larger than the equivalent ordinary PDF — full font embedding is the price of independence.

When do you actually need it? Only when a receiving institution says so — but that is increasingly often: electronic court filing systems, government procurement portals, tax archives, and regulated industries commonly require PDF/A conformance and validate submissions automatically. For personal long-term archiving it is a sensible choice too; for everyday sharing, ordinary PDF remains perfectly fine.

Practical notes from the trenches: converting to PDF/A is easiest when the source document is clean — fonts available, no broken structures, sensible images. A damaged file should be repaired first, and a scan-based document benefits from OCR beforehand, since an archival file you cannot search is only half an archive. And because PDF/A forbids encryption, protect the transport channel instead of the file when you must submit sensitive material: an encrypted PDF will simply fail validation.