Tagging an untagged PDF automatically

Tags are what a screen reader follows through a PDF. Without them, assistive technology has to guess what's a heading, what's a paragraph and what order things come in, and it often guesses wrong. A great many PDFs have no tags at all. PDF by Zaweba can add them automatically, which is usually the first and biggest step in making an old document accessible.

What automatic tagging does

It wraps untagged content in tags. Paragraphs become paragraphs. Headings are found by size, so larger text is tagged as a heading. Figures are tagged as figures, ready for alternate text. Links, form fields and comments each get their proper tags.

It also deals with things that shouldn't be read at all. Running headers, footers and page numbers are marked as artifacts. An artifact is content that's there for the look of the page rather than its meaning. A screen reader skips it, so a listener isn't told the document's title and "page 14" at the top of every single page.

What it leaves to you

Lists and tables are not recognised automatically. You make them in the Tags window. It's an honest limit. Both are easy to get wrong by guessing. A table tagged badly can be harder to follow than one not tagged at all, because the screen reader announces rows and columns that don't match what's on the page. So after automatic tagging, look for the lists and tables in the document and tag them by hand.

Checking the result

Headings found by size are usually right, but not always. A large pull quote may be tagged as a heading. A small heading in a dense document may be tagged as a paragraph. Open the Tags window and look down the tree, or use the reading order tool, which shows numbered boxes on the page in the order they'll be read. Anything out of place can be changed, moved or turned into an artifact.

Figures need alternate text. Tagging finds the pictures; it can't describe them. Set alternate text for each one, or mark it as decorative if it adds nothing to the meaning.

Where it fits in the sequence

A sensible order for making a document accessible looks like this. Run the accessibility check to see where you stand. Tag the document automatically. Make the lists and tables in the Tags window. Fix the reading order where needed. Add alternate text, set the language and the title. Check again and save the report.

Make accessible runs the built-in action, then alternate text and the check, for those who prefer the steps strung together. Make accessible is also one of the built-in Actions, so it can be run on a whole folder of files.

Tips for better results

Tag a copy of the document, so the untagged original is still there if you want to start again.

Look at the first few pages carefully after tagging. Mistakes tend to repeat, so a problem on page two, such as a repeated header that wasn't marked as an artifact, is likely to be on every page, and you'll know what to look for.

Don't tag twice. Automatic tagging is for untagged content. If a document already has tags, the Tags window is where to improve them.

Automatic tagging in PDF by Zaweba does the slow, repetitive part of the job, offline and in moments, and leaves you to apply judgement where it counts.