A page that comes out the same for everyone

The promise of an archival document is that the page read today is the same page decades from now. It holds only where the file settles everything the page shows, leaving nothing to the reading software.

Summary

A PDF is allowed to leave decisions to the reading software. A layer can show on screen and stay off the printer. A picture can ask to be smoothed over, and reading software smooths it differently. A clickable area can exist without being visible. A page can trigger an action as soon as somebody arrives at it.

The format allows all of that; the archiving standard was written to take that choice away from the reading software. An archival document settles all of it inside the file, so the page a reader gets is the page that was meant.

What this costs in practice is small. Layers are still allowed — what goes is the automatic state that shows or hides them by itself. Clickable links are still allowed — they simply have to be visible and printable. A button that triggers an action when pressed belongs to a document in use, not to a document being archived.

Technically

Optional content is permitted; automatic state is not. A layer whose visibility follows printing, the screen, the magnification or the reading software's language is refused: the engine does not write the document. The layer itself, and the default configuration naming it, are written unchanged.

An image is drawn from the samples it carries: interpolation refuses the document. Every annotation is printed and shown: none is invisible, hidden, kept off the screen, or turned off the screen when it is selected. Form widgets are written printed by construction, so the rule bites on shape annotations and links.

Actions are judged on their type, never on where they lead. Part 3 permits going to a page of this document, to another document, to a document carried inside this one, to a reading thread, to an address on the web or to a place this document registers under a name, and submitting a form. Part 4 permits those and three more: running a script, playing a media item the file carries, and going to one of the parts a document is divided into. Beyond the type: a field that acts when it is used refuses the document under every part; a field that acts when something happens to it refuses it under part 3 and is allowed under part 4; a page that acts on arrival or departure refuses it under both.

A claimed document is written with at most 8 388 607 objects, and is refused above that, which is the standard's own ceiling rather than the format's.

Clause by clause

Three things the standard settles are left to a checker. A claimed document that also asks for a password is written as asked, and a checker declines it — an archival document is written without one. Transparency, soft masks and blend modes are written as asked; under parts 3 and 4 the engine does not check them. And the engine does not read the additional actions of the document itself, only those of a page.

What the engine does on its own

The engine does all of this as soon as a document claims the standard, whatever else the calling program asks for.

  • Names the default configuration of the layers, which the standard asks for.
  • Writes every form widget as printable.
  • Refuses the document over a layer with automatic state, a smoothed image, a hidden or unprintable shape or link, a forbidden kind of action, a field or page that acts, and a document past the object ceiling — naming, in each case, what it was refused for.

What you supply

The engine states what it was given and invents nothing else: without those values it refuses the document rather than choosing one for you.

  • Layers without automatic state: name them, draw on them, and leave their visibility to the reader.
  • Images asked for without smoothing.
  • Links and shapes left visible and printable.
  • A document without a password, and without transparency where the receiving archive declines it.

The published texts behind these rules

Each published text, and what it asks for
Published text What it asks for
ISO 19005-3 6.9 The configuration of the layers is written out, and no state is left automatic.
ISO 19005-4 6.10 The same, for the part built on PDF 2.0.
ISO 19005-3 6.3.2 Every annotation is printed and shown, and none is invisible, hidden, kept off the screen, or turned off the screen when it is selected.
ISO 19005-4 6.3.2 The same, for the part built on PDF 2.0.
ISO 19005-3 6.2.8 An image is drawn from the samples the file carries.
ISO 19005-4 6.2.7.1 The same, for the part built on PDF 2.0.
ISO 19005-3 6.5.1 Only the listed kinds of action are carried.
ISO 19005-4 6.6.1 The same, with a longer list.
ISO 19005-3 6.4.1 A form widget neither acts when used nor acts when acted on.
ISO 19005-4 6.4.1 A form widget does not act when used.
ISO 19005-3 6.5.2 A page does nothing on being arrived at or left.
ISO 19005-4 6.6.3 The same, for the part built on PDF 2.0.

The other families of requirements

Which part of the archiving standard to claim Every font the document uses is embedded in the file Colours that mean the same thing on every device The document says what its marks stand for A document that carries other files inside it What the file says about itself Checking the claim, before and after

Back to what the archiving standard asks for

See the prices See the examples