PDF/A and PDF versions explained: what an archive-ready PDF needs

A short history of the format

PDF began as an Adobe specification. The first version was PDF 1.0, and revisions followed up to PDF 1.7, released in November 2006. In January 2007 Adobe submitted PDF 1.7 for standardisation, and it was published as the international standard ISO 32000-1 in July 2008.

The next edition, PDF 2.0, is ISO 32000-2, first published in 2017. ISO 32000-1 is written so that it can also be used to interpret files made to any of the earlier versions, from PDF 1.0 to 1.7.

What the version number means

The version number says which edition of the core PDF rules a file was written to. It sits at the very start of the file: the characters %PDF- followed by the number, such as %PDF-1.7.

That header isn't always the full story. From PDF 1.4 on, a file may state a different version inside its catalogue, deeper in the file, and that value overrides the header. Some identification tools read only the start of a file and can get it wrong.

What PDF/A is and what it forbids

PDF/A is a family of ISO standards, ISO 19005, for constrained forms of PDF meant for long-term preservation. The idea is simple: anything the file needs in order to display must be inside the file, described in a way that doesn't depend on a particular device or program. The main rules for PDF/A-1 to PDF/A-3:

• All fonts must be embedded, and they must be legally embeddable.
• Encryption is not allowed, so no open password and no permission restrictions.
• No external dependencies: nothing the page needs may be fetched from elsewhere.
• Audio, video and 3D content are forbidden, and so is JavaScript.
• Colours must be specified in a device-independent way.
• Metadata must be stored in XMP, including a block that says which PDF/A part and level the file follows.

PDF/A-1, -2, -3 and -4 in brief

PDF/A-1 (2005) is based on PDF 1.4. PDF/A-2 (2011) moves to PDF 1.7 and adds JPEG 2000 compression, transparency and better tagging for accessibility, and it allows attachments only if they are PDF/A themselves. PDF/A-3 (2012) is identical to PDF/A-2 except that files in any format may be embedded. PDF/A-4 (2020) is based on PDF 2.0.

Parts 1 to 3 also have conformance levels. Level b (basic) guarantees the visual appearance and is typical for scanned pages. Level u adds the requirement that all text has Unicode equivalents. Level a (accessible) meets every requirement in the standard, including structure.

Why courts and portals ask for it

An organisation that has to keep documents for years wants files that don't rely on fonts installed on one computer or on links that may break. Some courts write this into their filing rules. For example, the formatting requirements for documents filed electronically with New York's Appellate Division say they shall comply with text searchable PDF archival format (PDF/A).

If a portal asks for PDF/A, read its exact wording. It may name a part or level, or require searchable text, which a plain image scan does not have.

How to tell whether a file claims PDF/A

A PDF/A file identifies itself in its XMP metadata, with entries for the part (for example 1) and the conformance level (for example B). The file name still ends in .pdf.

Adobe Acrobat has a preference, View Documents In PDF/A Mode, set to Never or Only For PDF/A Documents. When it is on, a PDF/A document opens in Read mode to prevent changes, and a message appears in the document message bar. If you turn the mode off, you can edit the document.

Where KeepLocalHub fits

None of our tools creates PDF/A. Treat whatever comes out of Merge PDF, Split PDF, Organize PDF pages or Compress PDF as an ordinary PDF, even if what went in was PDF/A. The sensible order is to assemble and arrange your pages first, then convert to PDF/A as the very last step in software that can create and validate it, and send that file without editing it again.