Archive for January, 2011

Will Computer Science Meet accessibility in 2011?

January 18, 2011

I’m a legally blind retired computer scientist. As I gained proficiency with assistive technology for reading, writing, and communicating, I faced similar costs, barriers, grievances, and coping challenges as thousands of other computer adept late career people. However, I also take a keen interest in effectiveness and usability of my access tools and the media they work upon as a total system for processing information in our marvelously plastic brains. And, as former educator, researcher, and manager, I look upon my profession as contributors to both sides of the problem and solution arenas acting under broader social forces from government, demographics, and mainstream technology industries.

May I share my unique experience with you? Here’s my take on the current state of computer science (CompSci) related to Persons with disabilities (PwD)in general and the specific opportunities for visually impaired persons. Assistive technology refers to software like screen readers that use text to speech and keyboard focus interactions with operating systems, applications, and web pages. Accessibility is a matter of degree to which the applications, OS, and web sites support assistive technology. to achieve the same performance and satisfaction as all other users.

responsibilities, accountability, openness, and Opportunities for CompSci

are educational institutions now, in 2011, ready to embrace disability civil rights? Is the academic computing field prepared to integrate advances from the separated assistive technology industry and the generation of students raised with strong but different skill sets? Can CompSci meet its aspirations of providing the 4th R of education for everybody? Will there be movement to re-mediate decades of deficient designs of web information management systems and individual documents? where does CompSci and information technology fit into this solution, or problem, space?

basic accountability as an academic discipline

Like all educational fields that use web resources to assist education, the CompSci and IT fields are clearly responsible for adhering to standards that mitigate barriers for people with disabilities using available assistive technology. Especially where costs of access technology and special skills have been attained through rehabilitation resources or even individual investments, this is immediately a matter of jobs for PwD. Moreover, there are ripple effects for all intermittently or eventually disabled persons or caretakers, or tax payers, and that is everybody several times over.

Have our fields done well so far? No, as shown by flaws revealed traversing the 2010 Computer Science Education week and partner websites (see data below). These are rife with stumbling blocks, and generally exhibiting indifference to established design and usability practices. Barriers are unnecessarily erected, and unfortunate messages of ignorance and indifference indicate a field not so much up with trends in user oriented communication. or even acknowledging sensory differences in users.

domain responsibility of the CompSci field

CompSci and IT bear the additional responsibility of producing the tools, languages, and patterns; the programmers, designers, and testers; the processes, quality assessments, and design strategies; interfaces, interaction models, and transactions; the books, published articles, and motivations; and so on, that underlay the capabilities for educational institutions to meet their basic accountability.

Further CompSci responsibilities are the development of cultures where people with disabilities exhibit their skills and tools to demonstrate how well they can produce software and hardware products and artifacts. Beyond Cultural integration is the need for domain knowledge, e.g. how screen readers and caption systems work and how artifacts must be designed for smooth operation by persons using assistive technology.

CompSci has often promoted pedagogical tools like Alice and scratch that explicitly bar people with certain disabilities getting equal footholds in and excitement about computing. Nevertheless, many people have not only become high functioning but also innovative regarding access technology, including the very products I’m using to write this article. A community of computing oriented professionals have banded together to produce the aforementioned standards, tools, processes, and businesses that await adoption by CompSci and IT.

Computational thinking opportunities await CompSci

In fact, the above strengths and weaknesses of the social motivation for overcoming limits for PwD are truly, really, beautifully illustrative of computational thinking. The widely used WCAG standards are a fledgling “science of accessibility” with tested hypotheses, guidelines,, terminology, and a blogging trail of intellectual progress. Good web pages are all about semantics: markup, logical structure, sound relationships (in a database sense), and progressive enhancement design to transform semantics with syntactic elements like color and graphics. The essence of accessibility is support for multiple representations where access tech supplements or replaces sensory limits. Abstraction, semantics, representations, implementations, relationships, … are the sound principles for achieving the technical aspects of basic accountability and additional responsibilities of computing fields.

Hey, take the challenge! What should CompSci and IT do?

  1. clean up our websites, a good goal for Cs education week 2011. Read the standards, use guidelines and tools to re-mediate and assess quality, then do the work. With remediation of technical zits will come a better understanding of the computational thinking issues that should lead to improved designs.
  2. Take responsibility for explaining disabilities and accessibility to educational colleagues. Incorporate local disability service professionals and
    enlist the fear and concerns of university management to assure resources.

  3. audit all pedagogical tools and artifacts and label each for sensory and disability limitations. Then progress toward the better products available while applying computational thinking for more universal representations.
  4. Use the competitive, exciting advances of tablets, smart phones, text to speech, and accessible apps to motivate and explain both how accessibility works and why it matters in our economy. Just open up the hood under the accessibility options and check out the high performing speech interfaces.
  5. Learn to talk with persons with disabilities about their
    needs, high functioning skills, innovative tools, and culture.

  6. Do not feel bad about lack of experience or past mistakes. We are all overdue with a dose of karma, such as this writer who cannot use or maintain security education applets I developed five years ago. Ouch!


Overall, let’s open up a new field of computing, pull publications out of the ACM pay wall, and lead the way through computational thinking.

why not?

Issues, evidence, and epiphanies

are the feds really coming after universities for inaccessibility?

The Obama administration departments of Justice and department of education Office of civil rights have certainly shown signs of action backed up by White House ceremonies and initiatives:

On the positive side, California state University system is often praised for its improvements. Sadly, a funded study of analysis of university web accessibility is hidden in an obscure journal.

If all this comes to fruition right under the noses of congress, regulatory and advocacy will open many doors for computing professionals with a bent toward social entrepreneurship and intriguing technology advances. By the way, the professional accessibility virtual water cooler spreads daily updates on Twitter .

What will happen if universities are forcefully or voluntarily driven into accessibility? We may know by 2012.

why hasn’t accessibility and assistive technology taken hold in computing research and education, ?

As a former educator, I’ll take the all purpose route of blaming the textbooks? One form of blame is the presentation of content as in printed tomes, derived from WORD documents, spruced up by publishers, and embellished with instructor power points all performed without consideration for readability by print disabled students. This forces, I’m not kidding, hundreds of pages to be scanned into electronic forms where most original MS-WORD structure is lost, i.e. hours of labor in an error prone incomplete reverse engineering process.

How dumb is that?well, nationally, this problem is being rectified by bookshare under a department of education contract to adapt, just once in an industrialized manner, many college and K-12 textbooks. However, there isn’t a similar well known cooperative effort specializing in computing texts, or efforts by publishers except for Oreilly Media contributions of its electronic versions directly to bookshare.

Now, consider textbook content itself. Are there any, like more than 0, standard computing texts that contain chapters and exercises on assistive technology and accessibility as recommended in standards and produced by specialized branches of software and publishing industries? Please comment any examples.

the root of all evil in textbooks goes back to curricula accreditation. Omitted there, and frozen into practice, accessibility principles are instead forced into industry workshops, such as Knowbility Access U and Open Web Education Alliance. This further differentiates career paths with web development considered a craft, combining touchy feebly communication, advertising fodder, turnkey content management systems, and a steady flow of freelance or in house jobs open to lesser educated mortals.

The irony is that web accessibility is one of the best exemplars of “computational thinking” that has driven some higher echelons of CompSci leaders. See my 2009 post on many ways accessibility and assistive tech put computational thinking in action for pedagogical practices.

really? is the W3C nurturing a “science of accessibility”?

Read the W3C Web Content Accessibility Guide 2.0 and “Universal Design for Web Applications” by Wendy Chisholm and Matt May for lively explanations and motivation for the WCAG standards.

There’s an amazing amount of thought hammered into shape and utility in these guidelines and scenarios on the w3C web site. Rather than tons of funded research projects to identify hypotheses and perform experiments and build prototypes, the standards bodies combine experiences from developers, authors, consultants, and gadflies who really care about their subject. social and technical consequences. Fights and personalities drive discussions toward articulation and analysis that don’t come out looking like ACM portal abstracts. Nevertheless, pick any recommended practice, e.g. headings and logical structure in web pages, and you’ll find rationale, practical hedges for difficulties, and the basis for better controlled and more academically rigorous investigations.

As for the actual academic research communities, there’s a strange legacy of publication practices that make it difficult to track the field. Conference papers disappear behind the ACM Digital Library Portal pay wall. Institutional and individual members of ACM have access that people like this retired researcher have to fork over $200 to reach. Even paying the ransom isn’t enough, as I found it exceedingly difficult to negotiate the search interface in the 2008 time frame, and without response to requests for assistance. In other words, the publication pay wall is an inhibitor to the spread of insight on accessibility from perfectly serious and hard working researchers. How silly is that?

The notable exception I track is the work of professor Richard Ladner at U. Washington research and outreach and his prolific junior colleague Jeffrey bigham, now at U. Rochester. WebInsight project publications are available as readable PDF’s organized well by topics and authors that offer the bulk of their funded research.. These publishable fundable research results are intelligible, related to the standards versions of their science, and especially interesting for a user of the technology attracted to computational thinking, i.e. me. But then the papers reference too often into the ACM portal black hole. Wouldn’t the field progress more rapidly if more people could read such publicly funded publications and appreciate the experimental models being applied?

One additional topic I tracked was an award winning paper mentioned in Professor bigham’s blog on web research, namely the collaborative accessibility project at IBM Japan. However, the best I could find was a useful Youtube video on “social accessibility”. Indeed, with additional perspectives from the grass roots operational social accessibility projects webvism community tagging and solana for cracking the evil CAPTCHA barriers facing visually impaired web users. Indeed, find screen reader and accessibility videos on Youtube including Easy Youtube since Youtube itself is marginally accessible.

another interesting area is accessible apps for apple and android mobile products. There are important engineering lessons here regarding accessibility integration into the architecture, with apple doing it well, Google trying to paste on its talkback capability, and Microsoft admitting it blew off accessibility in its win 7 phones. Google Android accessibility is dubbed the “Model T Syndrome” for not applying state of the art engineering techniques, expecting visually impaired consumers to wait years for reasonable functionality and usability.

Finally, for the serious minded computer theory connection, visit the IBM researcher and leading accessibility guru Jim Thatcher articles on practical standards in business as applied to,, and many .gov websites. This wealth of robust reasoning and decades of experience are truly awesome.

What’ is the evidence for bad accessibility practice in the computing field?

Here is a test you can perform yourself.

Start the CSED Week test in Web Aim WAVE analyzer. Yes, click that link and now you’ve been seduced into web page testing! Now, look for the link to Partners, click and see the errors there. Keep going for the partner websites, opening and analyzing each web site. Keep going and you will be amazed at the WAVE complaints as the page structures are revealed in their semantic nakedness.

Lots of errors, right??? Let me explain how the errors affect my reading using an interactive access tech “screen reader”, illustrated in recordings in the 2009 post.

  1. The “missing ALT description” error tells me the web site developers have no clue about accessibility, ignoring the most basic rule. Visually impaired people cannot know what’s in your graphic, why it’s there,if it is decorative or meaningful in context.
  2. At the higher level of page structure are errors in omitted headings, irregular heading levels, and uninformative headings. The basic problem for someone visually impaired is building a reliable map of a page to transform from a linear search by laboriously tabbing from one HTML element to another. The outline tells me quickly what’s on the page, just like the outline of any well written document. Rarely do I find a web page from a CompSci organization with a good outline, often omitting headings entirely. Another indicator is irregular headings, like H4-H1-H3 which usually indicate confusion among semantics of headers and font-style presentation issues better handled by style sheets.
  3. Unlabelled form elements can be a show stopper when leading a person and screen reader through a donation or purchase or registration form. The proper HTML has an explicit corresponded between label and element, call, duh, “Label”. Without labels, the user just hears “edit box” rather than “first name edit box”. Forms are really complex , often associated with transaction timeouts and monumental headaches locating and fixing errors. Again, there are good rules for creating usable forms, which the unlabelled form element error tells me the developer has ignored. Do they want my business?

  4. Standalone link names are important for, like headings, a link abstraction allows rapidly skimming for general context and specific refinements.”Click here”, “here”, “read more”, and “learn more” require the screen reader user to search around for context. See post “I don’t want to click here” for a humorous take on this annoying practice.

Webaxe guide to introductions to accessibility and its demo podcasts is a good place to start and also entertaining. WebAim Web Accessibility in Mind also offers an annual empirical analysis of screen reader use and many checklists and guidelines. One caveat is that WAVE, although free and easy, is susceptible to flaws of any static analyzer with false hits, cascading errors, and interpretation of results. However, our tests show that it readily exposes often embarrassing mistakes just waiting for correction. My favorite was a major CompSci blog with hidden text offering Viagra remedies.

While many of these complaints relate primarily to technical communication, there are true design problems related to search tasks, as on the ACM Digital Library, and on large multi-organizational websites like universities. Beyond accessibility, as in supporting technology, are issues of bandwidth limitations, small screen mo vile devices, and user choices on browser script security. While not formalized as in “structured programming” or “object-oriented design”, the recommended engineering practice is “progressive enhancement”, starting from a purely semantic page that covers the basic content and separates presentation layers which a browser can strip away to assure the content is preserved in many contexts. It cannot be emphasized too much: the person using a screen reader is working directly with the semantic content provided by the developer. Designer focus on color, fonts, graphics, and interactivity are truly only “in the eyes of the sighted reader” and may add to but should not obscure the essential page content. and use cases. In other words, the analyses provided by tools like WebAim WAVE and even more important, the mental model in the person using a screen reader provide a favor to page designers by pointing out flaws.

And, is there any good news?

Definitely,when cultural divisions are bypassed, are growing assemblage of tools that enable someone losing vision to maintain their computer skills, provided they can access the training and guides to re-build their own environment. Admittedly, regaining capabilities after vision loss requires months of hard work, willingness to learn new approaches, and acceptance of major life changes.

  • AThe free, powerful, open source screen reader NVDA (NonVisual desktop access) competes with established $1000 pricey products on Windows platforms. I truly enjoy, and donate to, the mailing list of international users who daily test and share advice on this Australian generated project. Its developers are blind, primarily using python. These guys deserve a major computing award for their global contributions and professionalism in their twenty-something age ranges.
  • The miracle of Text to Speech that activates the hearing sense into an alternative channel into our brains where reading actually takes place. While older people may take more time to rewire their brains after vision loss,it’s truly remarkable that vision can be so minimalist in computer usage, provided accessibility is engineered into our software and information sources. Now, we’re poised to take on the challenge of “information visualization without vision”, seriously a cognitive and technological adventure in literacy and openness.
  • Bookshare and NFB News Line downloadable a alternative for print disabled services that brings literally 1000s of great books and daily newspapers to our fingertips in wireless seconds. Never did I imagine I could have such a great store of information to support my retirement book club, lifelong learning, and social entrepreneurship activities period. Materials are read by synthetic speech from DAISY, an XML based, international standard for audio and text content.
  • Levelstar Icon Mobile Manager and Docking Station, designed and distributed by a blind engineer, that streamlines my access to Bookshare, NewsLine, Twitter, email, and RSS. Most sighted, and now blind, people will enjoy an immense number of accessible iPad apps, a direction I’ll soon be taking myself.But the Icon sets a high bar of throughput I don’t expect to find on any other device by avoiding screens, using spoken menus and text reading. Another award worthy young technologist for CompSci to learn from.The implementation software for this handheld LINUX box is python and sqlite.
  • The #a11y Twitter community of accessibility gurus, blindness advocates,normal blind working folks, and inspiring authors lifts me up every day with humor and an unbelievable syllabus of linked readings. I never expected to find such a “School of Twitter” in social media that could fill my local personal and professional void. I especially value AccessibleTwitter website and demonstration for its common sense, ease of use, and challenge to the big clunky Twitter, which is, of course, the data source and API.

  • I’m also grateful for professional opportunities to potentially influence the direction of computing through the CMD-IT Center for Minorities and Disabilities in ITan, its Board of Advisers, and energetic organizer. I’ve written two other posts input to an NSF Task Force on CyberLearning, and hopefully await an insightful report.
  • Close to home, I appreciate the opportunity to connect with a few local disability professionals and volunteer groups. I’ve seen first hand how a broken rehab system requires enormous cooperation and energy to bring to ever more baby boomers losing vision the tools and experience I managed to find for myself. For all the $$$ spent on research, the chain of referrals and services beyond the medical plateau leaves so many of us just hanging on precariously while trying to find our ways through the inevitable grieving and depression cycles. It shouldn’t be this way in a
    wealthy world, requiring not charity but rather planned delivery of existing resources, as related in Jane Brody’s NYTimes articles on vision loss.

The 2011 CompSci Meets Accessibility Manifesto

And that latter point is where my disappointment with the handling of assistive technology and accessibility in computing has lead me to put considerable effort into writing up this critique. We just have to do better in accountability within institutions, domain responsibility for our professionals, and awareness of the depth of effectiveness of our computational thinking methods. Thousands of jobs depend directly on our outcomes for accessibility and quality computing products, plus centuries of better quality of life for everyone sooner or later. Let’s make accessibility meet computer science professionally in 2011.

We’re now at a teachable moment for assistive tech and accessibility in computing education. Everybody has the basic functions in their hands, literally, and for free. Windows users can download capable free open source NVDA screen reader and try testing web pages. Android and IOs users turn on their text to speech and learn credible NonVisual manners of using myriad interesting and useful apps. Come on, anybody can learn to work like a low vision person so the days of descending into the exorbitantly expensive blind ghetto for access tech is over. Anybody from now on who produces inaccessible pedagogical products or sloppy web pages is out of excuses. Your artifacts are testable, the testing tools are available, the engineering practices are wedded with the science of accessibility in standards. and people with sensory limitations like my hazy vision have those access tools at their fingertips, skilled and raring to use products made for mainstream but accessible if properly designed. So, failure to step up to this challenge and do the right thing, which really isn’t so hard and actually is good for business, is a choice of accountability, responsibility, and opportunity.

Beyond Universal Design – Through Multi-Sensory Representations

January 8, 2011

<The following recommendation was offered at the CyberLearning workshop addressed in the previous post on CyberLearning and Lifelong Learning and Accessibility. The post requires background in both accessibility and national funding policies and strategies.

This is NOT an official statement but rather a proposal for discussion. Please comment on the merits.

Motivation: CyberLearning must be Inclusive

To participate fully in CyberLearning, persons with disabilities must be able to apply their basic learning skills using assistive technology in the context of software, hardware, data, documentation,, and web resources. Trends toward increased use of visualizations both present difficulties and open new arenas for innovative applications of computational thinking.

Often, the software, hardware, and artifacts have not been engineered for these users, unforeseen uses, and integration with a changing world of assistive tools. Major losses result: persons with disabilities are excluded or must struggle; cyberlearning experiments do not include data from this population; and insights from the cognitive styles of diverse learners cannot contribute to the growth of understanding of cyberlearning.

Universal Design Goals

Universal design embodies a set of principles and engineering techniques for producing computational tools and real world environments for persons usually far different from the original designers. A broader design space is explored with different trade-offs using results from Science of Design (a previous CISE initiative). Computational thinking emphasizes abstraction to manage representations that lead to the core challenges for users with disabilities and different learning styles. For example, a person with vision loss may use an audio channel of information received by text to speech as opposed to a graphical interface for visual presentation of the same underlying information. The right underlying semantic representation will separate the basic information from its sensory-dependent representations, enabling a wider suite of tools and adaptations for different learners. This approach transcends universal design by tapping back into the learning styles and methods employed effectively by persons with many kinds of disabilities, which may then lead to improved representations for learners with various forms of computational and data literacy…

Beyond Universal Design as Research

beyond Universal Design” suggests that striving for universal design opens many research opportunities for understanding intermediate representations, abstraction mechanisms, and how people use these differently. This approach to CyberLearning interbreeds threads of NSF research: Science of design and computational thinking from CISE +human interaction (IRIS)+many programs of research on learning and assessment. +…

Essential Metadata Requirements

A practical first step is a system of meta-data that clearly indicates suitability of research software and associated artifacts for experimental and outreach uses. For example, a pedagogical software package designed to engage K-12 students in programming through informal learning might not be usable by people who cannot drag and drop objects on a screen. Annotations in this case may serve as warnings that could avoid exclusion of such students from group activities by offering other choices or advising advance preparation. Of course, the limitations may be superficial and easily addressed in some cases by better education of cyberlearning tool developers regarding standards and accessibility engineering.

Annotations also delimit the results of experiments using the pedagogical software, e.g. better describing the population of learners.

In the context of social fairness and practical legal remedies as laid out by the Department of Justice regarding the Amazon Kindle and other emerging technology, universities can take appropriate steps in their technology adoption planning and implementation.

Policies and Procedures to Ensure Suitable Software

For NSF, appropriate meta-data labeling then leads to planning and eventual changes in ways it manages its extensive base of software. Proposals may be asked to include meta-data for all software used in or produced by research. Operationally, this will require pro posers to become familiar with the standards and methods for engineering software for users employing adaptive tools. While in the short run, this remedial action may seem limiting, in the long run the advanced knowledge will produce better designed and more usable software. At the very least, unfortunate uses of unsuitable software may be avoided in outreach activities and experiments.
Clearly, NSF must devise a policy for managing unsuitable software, preferably within a 3 year time frame from inception of a meta-data labeling scheme.

Opportunities for Multi-Sensory Representation Research

Rather than viewing Suitable Software as a penalty system, NSF should find many new research programs and solicitation elements. For example, visual and on visual (e.g. using text-to–speech) or mouse version speech input representations can be compared for learning effectiveness. Since many persons with disabilities are high functioning in STEM, better understanding of how they operate may well lead to innovation representations.

Additionally, many representations taken for granted by scientists and engineers may not be as usable by a wider citizenry with varying degrees of technical literacy. For example, a pie chart instantly understandable by a sighted person may not hold much meaning for people who do not understand proportional representations and completely useless for a person without sight, yet be rendered informative by tactile manipulation or a chart explainer module.

Toward a Better, Inclusive Workforce

Workforce implications are multi-fold. First, a population of STEM tool developers better attuned to needs of persons with disabilities can improve cyberlearning for as much as 10% of the general population. Job creation and retention should improve for many of the estimated 70% unemployed and under-employed persons with disabilities, offering both better qualities of life and reduced lifetime costs of social security and other sustenance. There already exists an active corps of technologically adept persons with disabilities with strong domain knowledge and cultural understanding regarding communities of disabilities. The “curb cuts” principle also suggests that A.D.A. adaptations for persons with disabilities offer many unforeseen, but tacitly appreciated, benefits for a much wider population and at reasonable cost. NSF can reach out to take advantage of active developers with disabilities to educate its own as well as the STEM education and development worlds.

Summary of recommendation

  1. NSF adopt a meta-data scheme that labels cyberlearning research products as suitable or different abilities, with emphasis on the current state of assistive technology and adaptive methods employed by persons with disabilities.

  2. NSF engage its communities in learning necessary science and engineering for learning by persons with disabilities, e.g. using web standards and perhaps New cyberlearning tools developed for this purpose.

  3. NSF develop a policy for managing suitability of software, hardware, and associated artifacts in accordance with civil rights directives to universities and general principles of fairness.

  4. NSF establish programs to encourage innovation in addressing problems of unsuitable software and opportunities to create multiple representations using insights derived from limitations as of software as well as studies of high performing learners with disabilities.

  5. NSF work with disability representing organizations to identify explicit job opportunities and scholarships for developers specializing in cyberlearning tools and education of the cyberlearning education and development workforce.

Note: this group may possibly be
National Center on Technology Innovation


Get every new post delivered to your Inbox.