PDF Days Europe 2022 — The Hybrid Return After the COVID Gap
PDF Days Europe 2022 took place in Berlin on 16–17 May 2022 — the first European in-person PDF Association conference after the pandemic cancelled the 2020 edition and moved 2021 to virtual. The hybrid format combined in-person attendance with live-streamed remote sessions, a model that has persisted since. Topics included PDF/UA-2 progress, PDF 2.0 tooling maturity, digital signature long-term validation, and structured e-invoicing mandates across EU member states.
Official event page →Key takeaways
- PDF Days Europe 2022 established the hybrid in-person/online format that has persisted through subsequent events, combining the networking value of in-person with the accessibility of remote streaming.
- The draft PDF/UA-2 was introduced publicly for the first time, featuring new structure elements (/Aside, /FENote), WCAG 2.1 alignment, removal of the /B conformance level, and MathML support.
- Five years after publication, PDF 2.0 had solid library support but slow end-user tool and organisational adoption — most PDFs were still being created as PDF 1.7.
- The PDF Association's reference implementations and test suites for PDF 2.0 were highlighted as key resources for implementors ensuring correct conformance.
- Removal of PDF/UA's /B conformance level in PDF/UA-2 was a deliberate signal that accessibility is not a spectrum with a minimum threshold — full conformance is the only standard.
Return to in-person: the hybrid model established
PDF/UA-2 draft: new structures and WCAG 2.1 alignment
/Aside element was defined for tangential or supplementary content (sidebars, pull quotes, advertisements) that is related to but not part of the main content flow. The /FENote element replaced the ambiguous use of generic note elements for footnotes and endnotes, providing a semantically specific type. The /DocumentFragment element addressed the common case where a PDF represents a portion of a larger logical document rather than a complete document.
WCAG 2.1 alignment — PDF/UA-1 had been developed against WCAG 2.0. PDF/UA-2 aligned with WCAG 2.1 (published June 2018), incorporating the additional success criteria around mobile accessibility, cognitive accessibility, and low vision. For PDF specifically, the most relevant WCAG 2.1 additions were 1.4.10 (Reflow), 1.4.11 (Non-text Contrast), and 1.3.4 (Orientation).
Removal of /B conformance level — PDF/UA-1 had two conformance levels: /U (full) and /B (basic). In practice, the /B level had been widely criticised as providing a lower bar that was insufficient for genuine accessibility. PDF/UA-2 removed the /B level entirely, establishing a single conformance requirement. This was a deliberate statement that PDF accessibility was not a spectrum with an acceptable minimum — a document was either fully accessible or it was not.
MathML integration — PDF 2.0 had added support for embedding MathML for mathematical content. PDF/UA-2 made proper MathML embedding normative for mathematical expressions, addressing a longstanding gap where mathematical content in PDFs was essentially inaccessible to screen readers unless alternative text was manually provided.PDF 2.0 implementation: five years in
Frequently asked questions
What new structure elements did PDF/UA-2 introduce?
PDF/UA-2 (ISO 14289-2:2024) introduced several new structure element types compared to those available in PDF/UA-1: /Aside for tangential content like sidebars and pull quotes; /FENote specifically for footnotes and endnotes (replacing ambiguous use of generic note elements); /DocumentFragment for PDFs that represent part of a larger logical document; /Annot for explicit annotation structure; and /Form for form container structure. PDF/UA-2 is based on PDF 2.0 (ISO 32000-2), which itself introduced additional element types. These new elements provide authoring tools and remediation workflows with more precise semantic vocabulary, enabling more accurate accessibility tagging for complex document layouts.
Why was the PDF/UA /B conformance level removed?
The /B (basic) conformance level in PDF/UA-1 (ISO 14289-1:2012) was intended to define a lower baseline of accessibility requirements for situations where full /U compliance was not achievable. In practice, it was widely criticised because the reduced requirements were insufficient to ensure genuine usability for people with disabilities. Organisations could claim PDF/UA-1/B compliance while delivering documents that were not functionally accessible to screen reader users. The PDF/UA-2 working group decided to remove the /B level entirely, making a clear statement: there is one accessibility standard, and either a document meets it or it does not. Documents that cannot meet full PDF/UA-2 requirements should be clearly identified as non-conformant rather than claiming a lower conformance level that might mislead users about actual accessibility.
How does MathML in PDF/UA-2 improve mathematical accessibility?
Before PDF/UA-2, mathematical content in PDFs was represented as graphics or as sequences of characters and operators in the content stream — neither of which was accessible to screen readers without manually provided alternative text. PDF 2.0 (ISO 32000-2:2017) added support for embedding MathML (Mathematical Markup Language) as associated content for mathematical expressions. PDF/UA-2 makes this normative: mathematical expressions in a PDF/UA-2 compliant document must be represented using properly structured MathML, not just visual rendering. Screen readers that support MathML (including JAWS and NVDA with MathPlayer) can then read mathematical expressions correctly — converting notation like "x squared plus two x minus three equals zero" into spoken form appropriate for users who cannot read the visual notation.
Try it in your browser
The PDF tools and standards discussed at this event are available free — no account needed.