Skip to main content

Manual check of a PDF/UA document

Bent Kirkegaard
  • Updated

Not all PDF/UA issues can be checked by a machine. Some requirements must be validated by a human. Usually, you do this with a visual check of the logical document structure.

How do you do it?

Useful tools for the visual check are:

These tools offer a simplified structure view so that you can do a visual check very easily and without deep knowledge of PDF structure elements.

What do you do?

Visually check the following topics:

  1. Scroll through the Preview and check if the logical reading order of the document is correct.
  2. Scroll through the Preview and check if all relevant content is tagged (for example, letter headers and footers often contain relevant information such as business hours or bank details, which must be accessible).
  3. Check all figures (and formulas if available) if they have meaningful alternative text (and not just the respective file names or poorly generated automatic text, for example).
  4. Check if appropriate tags (semantics) are used for all the content elements of the document:
    • Titles and subtitles should be <Title> and <Subtitle> tags.
      Note: For PDF 1.7, these tags should be mapped to the <P> role; for PDF 2.0, role mapping is not necessary.
    • Headings should be <H1>...<H6> tags.
    • Text paragraphs should be <P> tags.
    • Lists should be <L> tags (with nested <LI> tags, which in turn usually contain <Lbl> and <LBody> tags); if list continuations are included (i.e., there are several paragraphs that belong to a specific list item, it must be ensured that the list has not been split into several lists).
      Note: Glossaries and lists of abbreviations are also lists (these are referred to as definition lists).
    • Figures should be <Figure> tags and must have alternative text.
    • Tables should be <Table> tags (with nested <TR> tags, which in turn usually contain <TH> tags for the header cells and <TD> tags for the data cells).
      Note: Layout tables are broken down into their respective elements, such as headings and text paragraphs.
    • Captions, e.g., for figures and tables, should be <Caption> tags.
    • Block quotations should be <BlockQuote> tags.
    • Footnotes/endnotes should be <Note> tags.
    • Tables of contents and lists of figures and tables should be <TOC> tags (with nested <TOCI> tags).
    • Internal links and external links such as URLs should be <Link> tags.
    • Formulas should be <Formula> tags and must have alternative text.
    • Form fields should be <Form> tags and must have a clear tooltip.
  5. Check if all characters are displayed correctly.
  6. Check if there are spaces between the words.
  7. Check whether there are other accessibility issues, such as hyphenation or whether text in a language other than the main language is marked accordingly.