PDF vs PDF/A: what is the difference and when do you need each?

You have been asked to submit a document as PDF/A. You have a PDF. Are these the same thing? Not quite — and the difference matters enough that court systems, universities and records regulations specify one rather than the other.

The short version: PDF/A is a stricter subset of PDF designed so the file will still display correctly decades from now. Everything PDF/A allows is valid PDF, but not every PDF qualifies as PDF/A.

The problem PDF/A solves

An ordinary PDF can quietly depend on things that are not inside it. The most important is fonts. A PDF can either embed a font or simply reference it by name, assuming the reader will have it. That assumption holds today, on your machine. It may not hold in twenty years, on software nobody has written yet.

When a referenced font is missing, the reader substitutes something similar. Line breaks shift, characters may render differently, and a document that was carefully laid out starts to drift. For a memo this is irrelevant. For evidence in a legal case, a signed policy, or a thesis that has to remain citable, it is not.

Fonts are the clearest example, but not the only one. A normal PDF may also contain links to external content that has since vanished, transparency effects that future readers might interpret differently, embedded multimedia requiring plugins that no longer exist, or encryption whose password nobody remembers.

What PDF/A actually changes

• All fonts must be embedded, in full, inside the file.
• External dependencies are forbidden — nothing the document needs may live elsewhere.
• Encryption is not allowed, since a lost password would defeat the entire purpose of archiving.
• Colour information needed to reproduce the page faithfully is carried inside the document.
• JavaScript and executable content are prohibited.
• Audio and video content is not permitted.

The result is a genuinely self-contained file. Everything required to render the page exactly as intended travels with it. That is the whole idea, and it is why the format exists.

The cost: bigger files

Embedding complete font sets is not free. A PDF/A version of a document is often noticeably larger than the original, sometimes by several megabytes if it uses multiple typefaces.

If your PDF/A file is larger than the original, the conversion worked correctly. That extra size is the self-containment you asked for.

This trips people up regularly, because we are trained to think smaller files are better. For archiving, the trade is entirely worth it — you are buying certainty about what the document will look like long after everyone involved has moved on.

The conformance levels

PDF/A is not one thing, which adds to the confusion. Several versions and levels exist, and different institutions ask for different ones.

• PDF/A-1b — the original baseline. Guarantees visual reproduction. The most widely required.
• PDF/A-1a — adds structural tagging for accessibility, which is a higher bar to meet.
• PDF/A-2 — allows JPEG2000 compression and transparency, so files can be smaller.
• PDF/A-3 — permits arbitrary attached files, used for things like embedding source data with an invoice.

In practice, always check what the receiving institution specifies before converting. A court e-filing system that wants PDF/A-1b will reject a file that does not meet it, and the error message is rarely helpful about why.

Who actually needs PDF/A

• Court e-filing systems, which increasingly mandate it for submitted documents
• University thesis and dissertation deposits
• Healthcare and financial records subject to long retention requirements
• Government archives and records management
• Any corporate archive intended to outlast the software that created it

If you are not in one of those situations, a normal PDF is completely fine. There is no advantage to converting your everyday documents, and you will only make them larger for no benefit.

Practical notes before converting

Remove any password first. PDF/A forbids encryption, so an encrypted document cannot be converted until it is decrypted. Unlock it, convert, and store the archive somewhere appropriately secure instead of relying on the file's own password.

Run OCR on scans before converting. A scanned PDF/A is a perfectly valid archive of an image, but nobody can search it. Adding a text layer first preserves the words as well as the picture, which is usually the entire point of archiving a document.

Convert last. Editing a PDF/A file after conversion may break its conformance, so keep your working copy as a normal PDF and convert as the final step before submission or deposit.

And check the result opens and looks right. Conversion should not change appearance — if anything, an embedded font may render more faithfully than the substitute your reader was using before.