If documents have to be kept for ten or twenty years, an ordinary PDF isn’t enough. A regular PDF may rely on fonts installed on the computer, external elements or even scripts – fifteen years from now these may no longer exist, and the document may not open the way it used to. PDF/A was created precisely to solve this problem.
What is PDF/A?
PDF/A is a variant of PDF standardised for long-term preservation. The core idea is simple: the document must contain everything needed to display it. Fonts are embedded in the file. Colour definitions are embedded in the file. Anything that could change or disappear – scripts, audio and video plug-ins, links to external sources – is not allowed.
The result is that the document will open in exactly the same way today and in twenty years’ time, regardless of the software used.
How the PDF/A versions differ
The standard’s designations include numbers and letters, each with a specific meaning.
- The number (1, 2, 3) indicates the generation of the standard. Later generations allow more: PDF/A-2 adds layers and JPEG 2000 compression, PDF/A-3 adds attachments, i.e. the ability to embed, for example, an XML file in the document.
- The letter indicates the conformance level. b (basic) only guarantees that the document will look the same. a (accessible) additionally requires a text layer and a document structure, so that the content can be understood by software, not just by a person. u is an intermediate option: it requires that the text can be extracted as Unicode.
In practice: if you just need the document „to survive”, PDF/A-2b is enough. If you’ll need to search the archive and the documents must be readable by screen readers, choose a variant with the letter „a”.
PDF/UA is not the same thing
PDF/UA (universal accessibility) is often mentioned alongside PDF/A. It is designed not for preservation but for accessibility: so that the document can be read by a screen reader, text-to-speech or magnification software. For the public sector and educational institutions, this is increasingly a requirement rather than a nice-to-have.
One thing is important to understand: a scanned document without a text layer is inherently inaccessible. To a screen reader it is a blank image. So the path is always the same: text recognition first, and only then saving as PDF/UA.
Common archiving mistakes
Archiving a scan without recognition. Formally the document has been saved; in practice, nothing can be found in it. Five years on, an archive like this is a box of paper – just an electronic one.
Leaving metadata and hidden content in place. A PDF can carry the author’s name, earlier revisions, comments and form data. Before handing a document over to an archive or a third party, it’s worth cleaning this up – FineReader has a dedicated function for removing hidden data, which strips OCR text layers, comments, attachments, bookmarks, metadata, links and scripts.
Confusing redaction with a black rectangle. A black box placed on top is just a drawing over the text – the text underneath remains and can easily be extracted. Proper redaction removes the data irreversibly, including headers and footers, comments and metadata. The difference between the two can sometimes prove very costly.
Archiving images at excessive resolution. A 600 dpi colour scan is usually unnecessary for archiving and takes up a lot of space. MRC compression can reduce the size dramatically while keeping the document readable.
A practical workflow
- Scan at a resolution of at least 300 dpi.
- Correct the image – skew, background, perspective.
- Recognise the text, specifying the correct language.
- Review any uncertain recognition results.
- Remove hidden data if the document is going to be passed on.
- Save as PDF/A – the „a” variant if you need search and accessibility.
If you have a constant flow of documents, these steps can be automated: you specify a watched folder, and files placed in it are processed according to predefined rules. This feature is available in the Corporate edition – you’ll find the differences between editions in the comparison table.
Licences – FineReader PDF Corporate with automated processing, as well as FineReader PDF Standard and FineReader PDF for Mac – are available in our shop.