PDF Accessibility Standards — PDF/UA

In this tutorial, you will learn about PDF Accessibility Standards. We cover key concepts, practical examples, and best practices to help you master this topic.

PDF/UA (Universal Accessibility), formally ISO 14289, is the international standard that specifies the technical requirements for accessible PDF documents — covering tagged structure, reading order, metadata, font encoding, and navigation.

What You'll Learn

You will understand what PDF/UA requires, how its seven key principles apply to real documents, how it relates to WCAG 2.1 and 2.2, and how to audit a PDF against its criteria.

Why It Matters

WCAG covers web content but does not fully address PDF-specific requirements such as tagged structure trees, artifact marking, and font-to-Unicode mapping. PDF/UA fills that gap. Many government and enterprise procurement contracts require PDF/UA Compliance, and courts reference it in accessibility lawsuits.

Real-World Use

A university must publish 10,000 PDF course catalogs annually. WCAG alone does not specify how tables should be tagged in PDFs. By adopting PDF/UA, the university creates a single technical standard that their remediation vendor follows, and the resulting PDFs pass both automated checks and screen reader testing.

PDF/UA Principle Overview

flowchart TD
  A[PDF/UA ISO 14289] --> B[Tagged Content]
  A --> C[Logical Reading Order]
  A --> D[Unicode Character Mapping]
  A --> E[Metadata]
  A --> F[Natural Language]
  A --> G[Interactive Elements]
  A --> H[Security]
  B --> B1["All content must be tagged"]
  C --> C1["Tree order = read order"]
  D --> D1["Fonts mapped to Unicode"]
  E --> E1["Title, language, subject"]
  F --> F1["Language changes marked"]
  G --> G1["Form fields and annotations"]
  H --> H1["No restrictions on assistive tech"]

The Seven PDF/UA Requirements

1. Tagged content. Every visual element in the PDF must be represented in the structure tree with an appropriate tag. Artifacts (decorative elements like page borders) must be marked as artifacts.

2. Logical reading order. The order of elements in the structure tree must match the intended reading order. This is often violated in multi-column layouts and complex tables.

3. Unicode character mapping. Fonts must include a /ToUnicode CMap so that text can be extracted reliably. Without it, screen readers may output garbage characters.

4. Metadata. The document must declare its title, author, subject, and language in the XMP metadata. The title should be human-readable, not a filename.

5. Natural language. The default language must be set in the document catalog. Any section where the language changes must be explicitly marked.

6. Interactive elements. Form fields and annotations must have tooltips, proper tab order, and roles mapped to standard tags.

7. Security. Security settings must not prevent assistive technology from extracting content for accessibility purposes.

# Pseudo-code for checking PDF/UA compliance
import pdfplumber

def check_pdfua_basics(pdf_path):
    issues = []
    with pdfplumber.open(pdf_path) as pdf:
        meta = pdf.metadata
        if not meta.get("title"):
            issues.append("Missing document title")
        if not meta.get("language"):
            issues.append("Missing document language")
        # Check tags presence (simplified)
        if not pdf.docinfo.get("Tagged"):
            issues.append("Document is not tagged")
    return issues

print(check_pdfua_basics("report.pdf"))
# Example output: ["Missing document title", "Document is not tagged"]

PDF/UA vs WCAG

Common Mistakes

1. Treating PDF/UA as optional — Many organizations achieve WCAG AA on their website but ignore PDFs. Lawsuits often cite inaccessible PDFs as separate violations.

2. Assuming automated checks are enough — PAC 2021 and other checkers verify tag presence but cannot confirm semantic correctness. A <H1> that wraps an entire page passes the checker but is wrong.

3. Missing language changes — A single quote in French within an English document must be tagged with a language switch. Most PDFs omit this.

4. Security blocking assistive tech — Setting "Copying content" to disallow can also block screen reader access. PDF/UA requires security settings that do not interfere with accessibility.

5. Ignoring artifact marking — Page numbers, headers, footers, and decorative lines must be marked as artifacts. If tagged, they clutter the structure tree and confuse screen readers.

Practice and Challenge

1. What are the seven requirements of PDF/UA ISO 14289?
2. Why is Unicode mapping important for accessible PDFs?
3. How does PDF/UA requirement 4 (metadata) differ from WCAG 3.1.1?
4. What happens when a PDF has security settings that prevent assistive technology access?
5. Which tool can you use to check PDF/UA compliance?

Challenge: Download a sample PDF from a government website and run it through the PAC 2021 PDF checker. Report which PDF/UA criteria pass and which fail. Write a paragraph explaining the most critical failure.

FAQ

Mini Project

Select a 3-page PDF that contains headings, a table, and at least one image. Use the PAC 2021 PDF checker to audit it against PDF/UA. Document each failure with the specific PDF/UA requirement number, and write a remediation plan with estimated effort for each fix.

What's Next

Proceed to Tagged PDF Structure to understand how tags work internally, then continue with Reading Order in PDFs to learn how to verify and fix content flow.