UKOLN is supported by: Beyond Compliance - A Holistic Approach to Web Accessibility Brian Kelly UK Web Focus UKOLN University of Bath Bath, UK This work is licensed under a Attribution-NonCommercial-ShareAlike 2.0 licence (but note caveat) Acceptable Use Policy Recording/broadcasting of this talk, taking photographs, discussing the content using , instant messaging, blogs, SMS, etc. is permitted providing distractions to others is minimised. Acceptable Use Policy Recording/broadcasting of this talk, taking photographs, discussing the content using , instant messaging, blogs, SMS, etc. is permitted providing distractions to others is minimised. Resources bookmarked using ‘ techshare-2007 ' tag

2 About This Talk Work carried our by accessibility researchers in the UK: Reviews limitations of WAI approach to Web applicability (described at W4A 2005) Describes holistic approach for e-learning accessibility (described at W4A 2006) Applies previous work to new ‘edge case’ of culture on the Web (described at W4A 2007) Introduces a Stakeholder Model to help ensure sustainability of approaches to accessibility Proposes ‘Accessibility 2.0’ as term to describe approach which builds on WAI’s successes Acknowledgements to co-authors: Profs Helen Petrie & Stephen Brown and David Sloan, Jane Seale, Lawrie Phipps, Simon Ball, Patrick Lauke, Andy Heath and Fraser Hamilton

3 W4A 2005: Reprise At W4A 2005 we presented “Forcing Standardization or Accommodating Diversity…”: The practical difficulties of using a “standard” to encapsulate design requirements to accommodate a diverse set of needs under a diverse set of circumstances and contexts The achievements and limitations of WCAG in supporting this The resultant difficulties (and absurdities) from legislation and policy – that makes inappropriate reference to WCAG Using the example of the e-learning sector we pointed the way to a more holistic view of Web accessibility We received many positive comments on the ideas we presented WAI’s Limitations

4 Limitations of the WAI Model WAI model relies on conformant Web sites, conformant authoring tools, conformant user agents …and conformant users! A common complaint of “standardistas” – “the user needs to take responsibility…” There is value in this argument – but there are practical shortcomings And user technophobia/laziness/lethargy is only one obstacle  How many users know they are “disabled”? WAI’s Limitations Also note importance of evidence-based research. Various UK accessibility studies seem to find that lack of evidence of accessibility of Web sites for PWDs and conformance with WCAG guidelines!

5 The SiteMorse Simplicities SiteMorse’s automated tool showed poor WCAG compliance across several well-known UK disability organisations. The responses from the disability organisations belittled the findings of automated tools and give a much greater emphasis of testing by users. TechDis News Item, Jan 2005 WebExact (Bobby) shows problems WAI’s Limitations If disability bodies find compliance with guidelines difficult does this mean: They don’t care about accessibility Accessibility is hard & costly The guidelines have flaws (and note legal ramifications if an inaccessible Web site cites such data in their defence) If disability bodies find compliance with guidelines difficult does this mean: They don’t care about accessibility Accessibility is hard & costly The guidelines have flaws (and note legal ramifications if an inaccessible Web site cites such data in their defence)

6 The Importance of Context We argue Web accessibility is about supporting users achieve real world goals From Beyer & Holzblatt (1998) – the more you know about your target audience the more you can design to support them So the goal of “universal accessibility” has changed to supporting a defined set of users in the best possible way… How can we use WCAG to achieve this? WAI’s Limitations Note in the UK ‘widening participation’ seems to be preferred to ‘universal accessibility’ – with the latter sometimes leading to universal inaccessibility (“we can’t use JavaScript so we’ll not allow anyone to gain benefits it can provide”  )

7 Holistic Approach Follow-up work awarded prize for Best Research Paper at ALT-C 2005 E-learning conference This approach reflects emphasis in UK on blended learning (rather than e-learning) Kelly, Phipps & Swift developed a blended approach to e-learning accessibility This approach: Focusses on the needs of the learner Requires accessible learning outcomes, not necessarily e-learning resources Holistic Approach

8 Application To Culture Accessibility for information / factual resources is easy(ish) Accessibility for edge cases including learning, culture, research, gaming, communications, assertion of identity (teenagers on MySpace), …: More challenging Needed to allow providers of Web-based cultural services to enhance accessibility Generic model will provide broader framework for variety of Web uses Holistic Approach

9 Universal Accessibility? NormalCancer Man against snow, Austrian Tirol 1974, reproduced with permission of the photographer: Professor Paul Hill The Great Masturbator by Salvador Dali (1929) The Duck-Rabbit CRAFT BREWERY Holistic Approach

10 Articulating the Approach The "Tangram Metaphor“ (Sloan et al, W4A 2006) developed to avoid checklist / automated approach: W3C model has limitations Jigsaw model implies single solution Tangram model seeks to avoid such problems This approach: Encourages developers to think about a diversity of solutions Focus on 'pleasure' it provides to user This approach: Encourages developers to think about a diversity of solutions Focus on 'pleasure' it provides to user Holistic Approach

11 Tangram Model & Testability Baseline 1 Testable "WCAG 2.0 success criteria are written as testable statements …" (nb. automated & human testing ) Issues: What about WCAG principles that don't have defined success criteria (e.g. "content must be understandable")? What about 'baselines' – context only known locally What about differing models or / definitions of 'accessibility'? Note vendors of accessibility testing services will market WCAG tools e.g. see posting on BSI PAS 78 Tangram model can be used within WCAG Distinguish between testable (ALT tags) and subjective (content understandable) Supports baselines Discussions with developers on prioritisation of automated testing Holistic Approach

12 Tangram Model Note that similar moves to modularity are the norm in many W3C standards Model allows us to: Focuses on end solution rather than individual components Provided solutions tailored for end user Doesn't limit scope (can you do better than WAI AAA?) Use automated checking – but ensures emphasis is on user satisfaction Guidelines/standards for/from: WAI Usability Organisational Dyslexic Learning difficulties Legal Management (resources, …) Interoperability (e.g. HTML validity) Accessibility metadata Mobile Web …

13 Stakeholder Model Common approach: Focus on Web author Sometimes user involved Sometimes led by policy-makers This approach: Often results in lack of sustainability Web accessibility regarded as ‘techie’ Not integrated with wider accessibility issues Not integrated with training, development, … Priorities not negotiated with others There’s a need to integrate approaches to accessibility more closely with (diversity of) service providers

14 Repositories – Case Study Discussion on repositories list: “Why PDFs of research papers? What about accessibility?” “Important battle is open access. Let’s not add extra complexities.” My response: Open access is important (and PDF is easy) but let’s also: Engage with various stakeholders (incl. publishers) Develop (holistic) policies Explore other options to enhance accessibility And I found Scribd – a Web 2.0 services which creates MP3 from MS Word/PDF: Enhanced accessibility from MS Word master & Flash interface

15 The Cathedral & The Bazaar 2.0 Traditional WAI ApproachProposed Approach E-learningBlended learning CentralisedDevolved Single solutionVariety of solutions Slow-movingRapid response Remote testingTesting in context IT solutionBlended solutions Objective testingContext to testing Medical modelSocial model Accessibility as a thingAccessibility as a process Clear destination (AAA)Focus on the journey Accessibility as a cathedralAccessibility as a bazaar

16 Accessibility 2.0 Need to build on WAI’s successes, whilst articulating a more sophisticated approach. Accessibility 2.0: User-focussed: It’s about satisfying user’s needs Rich set of stakeholders: More than the author and the user Always beta: Accessibility is hard, so we’re continually learning Flexibility: There’s not a single solution for all use cases Diversity: There’s also diversity in society’s views on accessibility (e.g. widening participation, not universal accessibility) Blended solutions: Focus on ‘accessibility’ and not just ‘Web accessibility’

17 The Legal Framework This approach is well-suited for the UK legal framework: SENDA/DDA legislation requires "organisations to take reasonable measures to ensure people with disabilities are not discriminated against unfairly" Note that the legislation is: Technologically neutral Backwards and forwards compatible Avoids version control complexities The legislation also covers usability, as well as accessibility Other country’s legislation also talks about ‘reasonable measures’

18 Our Next Steps Accessibility Summit II: Held at JISC TechDis offices, York in Nov invited accessibility researchers, practitioners & policy makers in HE, public sector & disability support organisations Agreement on various concerns of WAI’s approach Recommendation to develop roadmap: more research, evidence-gathering, engagement, … Museums and Web 2007 Professional Forum: 50+ participants at international conference in April Further agreement on need to build richer approaches to accessibility for cultural heritage orgs Accessibility 2.0 term added to Museums Wiki

19 The Good News Our approaches have helped inform WCAG 2.0 developments WCAG 2.0: Technology neutral Based on four key principles (normative) (POUR):  Perceivable  Operable  Understandable  Robust Separate (informative) guidelines for HTML Reflects out holistic approaches The challenges: Providing a context outside of the principles e.g. “Super Cali go ballistic, Celtic are atrocious” Testing in contextCommissioning

20 Conclusions To conclude: WAI has provided a valuable starting point Need to develop a richer underlying model Need for Web accessibility to be placed in wider content There's a need to an evidence-based approach and less ideology Contextual approach & tangram metaphor aim to help inform such developments Accessibility 2.0 term can articulate a renewed approach

21 Questions Questions are welcome