The Impact of PDF/UA on Accessible PDF

One of the world’s major events on accessibility in IT, the ICCHP 2016 conference (15th International Conference on Computers Helping People with Special Needs) takes place July 13-15 in Linz, Austria, on the campus of the Joahnnes Kepler University. Olaf Drümmer, CEO at axaio software GmbH, will be chairing a special session on PDF/UA and its impact on accessible PDF.

Introduction

The new PDF/UA standard for accessible PDF documents has been published by ISO as ISO 14289-1 in July 2012. Those who developed and support it envision a much improved landscape for accessibility of PDF documents. This applies to the creation and consumption of accessible PDF documents.

PDF/UA regulates how content – graphics, text, multimedia, annotations, form fields – is to be included in a PDF file in order to be considered accessible. While leaving implementation details – how the components are technically incorporated in the PDF file, or how assistive technology (AT) takes advantage of accessible content in PDF/UA documents – to solution developers, it provides a clearly defined benchmark: a first in the field of accessible PDF documents.

According to early adopters – most notably software developers who already ship or at least announced PDF/UA enabled tools and solutions – the wait shouldn‘t be long until those interested in accessible PDF find a much more workable eco-system. Tools are being democratized, such that not only specialists but in principle every document creator can save their documents to accessible PDF at any time. At least one screen reader developer – NV Access to be precise – has already enhanced support for accessible PDF in their NVDA screen reader on the background of PDF/UA, for example access to complex tables.

At the same time it was clear from the beginning that extensive education will be necessary to nudge document producers and consumers towards adoption of PDF/UA. And last but not least it of course had to be proven that PDF/UA really does work as advertised, and does work well.

This STS aims to present an overview of the progress made for accessible PDF in the context of the PDF/UA standard. The four years between publication of PDF/UA and the ICCHP 2016 conference encompasses a time period long enough that there should be some feasible developments in the various fields discussed above.

Challenges

There are numerous challenges on the way towards increased accessibility in PDF documents. The overall eco-system will only begin to bloom – and hopefully to boom – , once all the necessary components and involved stakeholders begin to play together to achieve the much needed synergies.

• positioning: understanding PDF/UA in the continuum between document authoring and online, web based content, including its relation to Web Content Accessibility Guidelines (WCAG) 2.0, ISO 9241-171 (Ergonomics of human-system interaction – Part 171: Guidance on software accessibility) and recently introduced regulations like the EN 301 549 (Accessibility requirements for public procurement of ICT products and services in Europe) or the upcoming “refresh of section 508” in the USA.
• education: getting document software developers and assistive technology vendors, decision-makers, organizations in the public and private sector, accessibility experts, publishers, authors and last but not least the end-users to understand and adopt PDF/UA
• implementation: implementing PDF/UA in authoring tools – mainstream (Microsoft Office, LibreOffice, Adobe Indesign and others) and specialized solutions determining conformance with the PDF/UA standard
• deployment: getting document creators und document consumers to actually take advantage of the value added in the context of PDF/UA
• quality assurance: a reasonable level of quality and conformance with PDF/UA is required to ensure well functioning of the emerging PDF/UA eco-system; validation tools should be used as much as possible to relieve users from tiresome, unproductive tasks, while guaranteeing a high level of accessibility.

Presentations

Yoshiaki Tani presents a paper – prepared together with Takuya Takaira and Akio Fujiyoshi – on Sound-Embedded PDF aiming to help Elementary School Students with Print Disabilities accessing text books. While this concept might also be feasible and worthwhile independent of tagged PDF or PDF/UA, full integration with PDF/UA regulated syntax provides for a degree of accessibility that goes substantially beyond a simple interactive audio enrichment to PDF files.

For many problems there is a general choice of either removing the problem or finding a solution for the problem. Australia seems to be leaning towards removing the problem of insufficient access to PDF files by recommending or even requiring not to use PDF at all for communication and publication purposes, or at least to provide more accessible alternatives in all cases. Driven by the insight that especially the now ubiquitous adoption and use of mobile devices hasn’t provided a more accessible PDF experience to users with or without disabilities alike, Gian Wild discusses a number of actual or supposed strengths of PDF over other content formats, and aims to help decision makers make up their mind regarding preferable content formats in the light of the degree of accessibility that these formats can provide.

Alireza Darvishy shares recent developments and findings on Automatic Paragraph Detection for Accessible PDF Documents. This paper is rooted in ongoing work at the ZHAW (Zurich University of Applied Sciences) on providing readily available tools for deriving PDF/UA files from existing office documents. The current work of the group around Alireza Darvishy focuses on a new algorithm for the automatic detection and tagging of paragraphs in PDF documents. Using this algorithm, remediation of existing PDF documents can be carried out much more efficiently and with a lower error rate – a much needed effect given the number of existing PDF documents that have not already been created in an accessible fashion.

Reports on the current state of affairs regarding the use of PDF by people with disabilities are often based on anecdotal evidence, rough estimates or even just personal opinion. Karen McCall decided to address this unsatisfactory situation and in early 2016 carried out an extensive survey among people with disabilities, targeting aspects – in addition to the usual demographic data – such as whether PDF is consumed at work, or for education or leisure, user experience when accessing PDF, or quality of support by existing PDF readers and assistive technology.

Going beyond what is available or supported today, several groups of researchers have recently explored possibilities how to make rich graphical content more accessible, leaving behind the limitations of either providing alternative descriptions or alternative presentations for content that otherwise is not generally accessible. Substantial progress in this field has been presented at a workshop early 2016 in Japan – The 3rd International Workshop on “Digitization and E-Inclusion in Mathematics and Science 2016” (DEIMS2016) – by groups of researchers around Volker Sorge, Fumiyasu Sato and Siddhartha Gupta. In summer 2015 a group of students from the Johannes Kepler University in Linz, Austria – Markus Weninger, Gerald Ortner and Tobias Morowinger – studied how graphs – such as pie charts, bar charts, scatter plots and so forth – can be made fully and inherently accessible in an automated fashion without having to revert to alternative descriptions or alternative representations. In their project Accessible Scalable Vector Graphics (ASVG): Use of intention trees for accessible charts the team chose SVG – an XML syntax which is based on the same graphic model as PDF – for ease of coding. Their findings nonetheless are applicable not just to PDF and SVG, but also to HTML and EPUB.