Archiving, accessibility and metadata

Documents an archive will still open in twenty years, that a screen reader can read aloud, and that an accounting system books by itself.

Some documents have to outlive the software that made them. An archival PDF carries everything it needs inside it — the lettering, what its colours mean, a description of itself — so that nothing it depends on can go missing later. Public bodies and long-term archives ask for exactly that, and there are programs that check it.

The same care is what makes a document readable by somebody who cannot see it. The file can carry its own structure — this is a heading, this is a table, this is the wording that stands for a picture — so a screen reader announces the page in a sensible order instead of reading out a jumble.

What it saves you

An archive accepts the document
Every archival sample shown here is put through an independent validator against the published rules before it is published, so the claim on the page is a checked claim.
An accounting system books the invoice without a person
An electronic invoice carries, sealed inside the very page a person reads, the same invoice in the form a machine books from. Business-to-business invoicing in France is moving to exactly this.
A reader who cannot see it can still read it
A document that carries its own structure is announced in order, with its pictures described, instead of arriving as one undifferentiated block of text.

See it done

Every one of these is a real PDF this library wrote, with the source that wrote it.

A sheet that measures what it says

Three rules — centimetres, inches and points — on a page that states it must print at exactly the size it was drawn, and that the reader is to offer nothing else. That second request only exists from PDF 2.0 on, so the file declares 2.0 rather than changing version without you knowing.

A register whose every part says where it came from

Two sittings bound into one file. Each page carries a record of its own — what it is, whose hand wrote it, when — the photograph on the second page carries one too, and each transcription is tied to the page it transcribes rather than to the file at large. Tying an attached file to anything but the whole document only exists from PDF 2.0 on, so the file declares 2.0.

An archival sheet carrying fields of its own

A records office's reference, the department it came from, the date until which the document must be kept and its classification, in a PDF/A-3 file. No standard knows what a retention date is, so each field arrives with a description of itself — its type, who it is written for, what it means — which is what makes a checker accept information you invented. The page prints the same list, so what a person reads and what a machine reads cannot drift apart.

A tagged, accessible document

A document that says what each run of ink stands for, and which runs are only trimmings. Its text and its table describe themselves, a paragraph and a cell at a time. Validates as PDF/A-3a and PDF/UA-1.

A tagged list whose column is a picture

A parts list whose first column is a picture rather than a word. Each picture is marked as one and carries the words a reader hears in its place, so nothing in the list is lost on somebody who cannot see it. Validates as PDF/A-3a and PDF/UA-1.