Note: this is a long post with webliography in the next article.
There is also a recorded tour of CS web sites as an MP3 download.
Understanding web accessibility through computational Thinking
This post is written for distribution during the first proclaimed National computer science education week, December 7, 2009. My goal is to stimulate awareness within the CSE community of the importance of web and software accessibility to society at large and to the proper development of associated skills within CS curricula. Taking this further, I offer a call to action to renovate our own websites for purposes of (1) improved service, (2) learning and practice, and (3) dissemination of lessons learned to other academic entities, including professional organizations.
recognizing that traditional, accredited CS curricula do not define a role for accessibility, I suggest actions that can be grafted into courses as exercises, readings, debates, and projects. To even more legitimize and improve uptake of accessibility, many of these problems can be cast as computational Thinking in the framework of drivers from society, technology, and science.
Definitions and Caveats
Caveat: I do not represent the blindness communities, standards groups, or any funding agency.
Also, I limit this accessibility context to the USA and visual impairment disability.
here is my personal definition framework:
Definition: disability = inability to independently perform daily living tasks due to physical or mental causes
example: I cannot usually read print in books or news, nor text on a computer screen at size 14
Example: I cannot usually follow a mouse cursor to a button or line of text to edit
Definition: Assistive Technology (AT) = hardware or software that overcomes some limits of a disability
example: A screen magnifier can track a mouse cursor then smooth and enlarge text in the cursor region
Example: A screen reader can announce screen events and read text using synthetic speech
Definition: Accessibility = Quality of hardware and software to (1) enable assistive technology and also (2) support the AT user to the full extent of their skills without unnecessary expenditure of personal energy
example: A web page that enables focus through keyboard events enables a screen reader to assist a user to operate the page with ease, provided hands are working. Same is true for sighted users.
Example A screen magnifier enables reading text and screen objects but at such a low rate that I cannot accomplish much usual work:
Note: I am conflating accessibility with usability here, with usability usually referring beyond disabilities. Informally, to me, “accessibility” means my screen reader is fully operational, not in the way, and there are no reasons I cannot achieve the goal of page success as well as anybody.
Definition: Accommodation = explicit human decisions and actions to accomplish accessibility
Example: Modifying a web page enhances comprehension for a screen reader user, see POSH computational thinking below
Ecxample: Adapting security settings on a PC to permit a job applicant with a screen reader on a pen drive to read instructions and complete tests and forms
Example: A curb cut in a sidewalk enables wheelchairs to moor easily cross streets. Also true for baby strollers, inattentive pedestrians, visually impaired, luggage carts, skateboards, etc.
I base my analysis and recommendations on several domains of knowledge:
Learning and acquisition of skills as a recent vision Loser, becoming “print disabled”, “legally blind”, now at an intermediate skill level
Computer scientist, active for decades in formal methods and testing, highly related to “computational thinking” with broader professional experience in design methods and technology transfer.
Intermittent computer science and software engineering educator at undergraduate and master’s level programs with experience and opinions on accreditation, course contents, student projects, and associated research
Accelerated self-study and survival training from the community of persons with disabilities, the industry and professions serving them, and the means for activism based in social media like twitter, blogs, and podcasts
Lingering awareness of my own failings before my vision loss, including software without accessibility hooks, web pages lacking structural/semantic markup, and , worst of all, omission of accessibility considerations from most courses and projects. My personal glass house lies in slivers around me as I shout “if only I knew then, when I was professionally active, what I know now, as a semi-retiree living with the consequences and continuing failures of my profession.
what is “computational thinking” and what does it have to do with accessibility?
This term was coined by dr. Jeannette wing in a 2006 article, and best expressed in her
Royal society presentation and podcast conversations. for our purposes, CT asks for more precise description of abstractions used in assistive technology, web design, and mainstream browsers, etc. The gold standard of web accessibility for my personal kind of disability, shared with millions of Americans, is the bottom line of reading and interacting with web sites as well as currently normally sighted persons. To an amazing degree, audio and hearing replaces pixels and seeing provided designs do support cooperation of assistive technology at both primitive levels and costs for effort expended. I’ll illustrate some fledgling computational thinking in a later section and by touring CS and other websites, but, sorry, this won’t be a very pleasant experience for either me the performer or listeners.
CSE can benefit from the more rigorous application of CT to meet its societal obligations while opening up new areas of research in science and technology leading to more universal designs for everybody. To emphasize, however, this is not a venture requiring more research before vast improvements can be achieved, but rather a challenge to educators to take ownership and produce more aware computing professionals. …
Driving forces of society, Technology, and science
Here’s a summary of trends and issues worthy of attention within CSE and suggested actions that might be grafted appropriately.
driving forces from society
computer science education has a knowledge gap regarding accessibility
As excellently argued in a course description “Accessibility First”, web design in general, accessibility, and assistive technology are at best service learning or research specialties falling under human computer interface or robotics. where do Cs students gain exposure to human differences, the ethics of producing and managing systems usable by everybody, and the challenges of exploring design spaces with universal intentions.
The extensive webliography below offers the best examples I could find, so please add others as comments. Note that I do not reference digital libraries because (1) the major ACM Portal is accessibility deficient itself and (2) I object to the practice of professional contributions being available only at a charge. The practice of professional society control over publications forces a gulf between academic researchers and a vibrant community of practitioners, including designers, tool builders, accessibility consultants and activists.
Action: Use the above definition framework to describe the characteristics of the following as ordinary or assistive: keyboards, tablets with stylus, onscreen keyboard, mouse, screens, fonts, gestures, etc. How do these interfaces serve (1) product developers and (2) product users? Where is the line between assistive and mainstream technology?
Action: see the proposed expansion of the National computer Science education proclamation in our conclusions. Debate the merits of both the whereas assumptions the therefore call to action. Are these already principles adopted and practiced within CSE?
Disability is so prevalent that accessibility is a uniform product requirement.
Being disabled is common, an estimated 15% of U.S.A. population with serious enough visual impairment to require adjustments from sites designed assuming full capabilities of acuity, contrast, and color. Eyesight changes are inevitable throughout life, even without underlying conditions such as macular degeneration or severe myopia. Visual abilities vary also with ambient conditions such as lighting, glare, and now size and brightness of small screens on mobile devices. considering other impairments, a broken arm, carpal tunnel injury, or muscle weakness give a different appreciation for interaction with a mouse, keyboard, or touch screen. As often said, we will all be disabled some way if we live long enough. Understanding of human differences is essential to production of good software, hardware, and documentation. Luckily, there are increasingly more specimens, like me, willing to expose and explain my differing abilities and a vast library of demonstrations recorded in podcasts and videos.
Action: view You tube videos such as the blind web designer using a screen reader explaining the importance of headings on web pages. Summarize the differences in how he operates from currently sighted web users. How expensive is the use of Headings? See more later in our discussion of CT for Headings.
Action: visit or invite the professionals from your organization’s Disability services, Learning center, or whatever it is called. These specialists can explain disabilities, assistive technology, educational adjustments, and legal requirements.
Action: Is accessibility for everybody, everywhere, all the time a reasonable requirement? What are the ethics and tradeoffs of a decision against accommodation? What are the responsibilities of those requiring accommodations?
The ‘curb cut’ principle suggests how accessibility is better for everyone
Curb cuts for wheelchairs also guide blind persons into street crossings and prevent accidents for baby strollers, bicyclists, skateboarders, and inattentive walkers. The “curb cuts” principle is that removing a barrier for persons with disabilities improves the situation for everybody. This hypothesis suggests erasing the line that labels some technologies as assistive and certain practices as accessibility to maximize the benefits for future users of all computer-enabled devices. This paradigm requires a new theory of design that recognizes accessibility flaws as unexplored areas of the design space, potential harbingers of complexity and quality loss, plus opportunities for innovation in architectures and interfaces. Additionally, web accessibility ennobles our profession and is just good for business.
Action: List physical barriers and adaptations in your vicinity, not only curb cuts, but signage, safety signals, and personal helpers. Identify how these accommodate people with canes, wheelchairs, service animals, etc. And also identify ways these are either helpful or hampering individuals without disabilities. Look at settings of computers and media used by instructors in classrooms. Maybe a scavenger hunt is a good way to collect empirical physical information and heighten awareness.
Action: Identify assistive technology and accessibility techniques that are also useful for reasons different from accessibility? e.g. A keyboard enabled web page or browser tabs support power users.
Persons with disabilities assert their civil rights to improve technology.
while most of us dislike lawsuits and lawyers, laws are continuously tested and updated to deal with conflicts, omissions, and harm. Often these are great educational opportunities on both the challenges of living with disabilities and the engineering modifications, sometimes minor, for accommodations. Commercial websites like amazon, iTunes, the Law School aptitude test, small business administration, and Target are forcefully reminded that customers are driven away by inaccessibility of graphics, menus, forms, and shopping carts. Conversely, recently, I had a quick and easy checkout from a Yahoo small business website, greatly raising my respect and future return likelihood whenever I see the product vendor and website provider.
Devices such as controllers on communication systems, the amazon Kindle, and new software like google WAVE and chrome browser often launch with only accessibility promises, excluding offensively and missing feedback opportunities from persons with disabilities. Over and over, it is shown that the proverbial software rule of increasing cost of fixing missing requirements late is exemplified by accessibility, whether legal or business motivated. While a lawsuit can amazingly accelerate accessibility, companies with vast resources like Microsoft, Oracle, blackboard, and google are now pitted in accessibility races with Yahoo, apple, and others. The bar is rapidly being raised by activism and innovation.
for many The social good of enabling equal access to computing is an attractor to a field renowned for nerds and greed. Social entrepreneurs offer an expansive sense of opening doors to not only education and entertainment but also employment, that now stands around 20% for disabled persons. Many innovative nonprofit organizations take advantage of copyright exemptions building libraries and technology aids for alternatives to print and traditional reading.
The computing curb cuts principle can motivate professionals, services, and end users to achieve the potential beauty and magic of computing in everyday life, globally, and for everybody who will eventually make the transition into some form of sensory, motor, or mental deficiency. But, first, mainstream computing must open its knowledge and career paths to encompass the visionaries and advances now segregated. All too often persons with disabilities are more advanced, diversified, and skill full in ways that could benefit not yet disabled people.
Action: The ubiquitous bank ATM offers a well documented ten year case study of how mediation led to a great improvement in independent living. for visually impaired people. Take those ear buds out of the MP3 player and try them on a local ATM, asking for service help if needed or ATM is not voice enabled. Using a voice enabled ATM also provides insight into the far more problematic area of electronic voting systems.
the amazon Kindle lawsuit by blind advocates against universities considering, or rejecting, the
device and its textbook market provides a good subject for debate.
Action: On the home front, pedagogical advances claimed for visual programming languages like Alice are not equally available to visually impaired students and teachers. first, is this a true assertion? How does this situation fit the definition of equal or equivalent access to educational opportunities? should the platform and implementation be redone for accessibility? Note: I’ve personally seen a student rapidly learn OO concepts and sat in on Cs1 courses with Alice, but I am totally helpless with only a bright, silent blob on the screen after download. Yes, I’ve spoken to SIGCSE and Alice personnel, suggested accessibility options, but never received a response on what happens to the blind student who signs up for an Alice-based CS course. Please comment if you have relevant experience with accommodations and Alice or other direct manipulation techniques.
The Web has evolved a strong set of standards and community of supporters.
W3c led efforts are now at 2.0 with an evolved suite of standards products, including documents, validator’s, and design tools. standards go a long way enabling accessibility by both their prescriptions and rationales, often drawing on scientific principles, such as color perception. but the essence of web standards is to define the contracts among browsers and related web technologies that enables designers to predict the appearance of and interaction with their designed sites and pages. The theme of WCAG 2.0 sums up as Perceivable, Operable, Understandable, and Robust. we all owe a debt to the Web standards Mafia for their technical contributions, forceful advocacy to vendors, and extensive continuing education.
Web standards are sufficiently mature, socially necessary, and business worthy that open, grassroots motivated curricula are being defined. CSE people who understand CT may well be able to contribute to this effort uniquely. In any case, questions about the relationship of tradition CS education and this independent curriculum movement must be addressed considering the large workforce of web designers, including accessibility specialists. Furthermore, web design inherently requires close designer and client communication, making it difficult to offshore into different culture settings.
Action: Use the #accessibility and #a11y hash tags on twitter to track the latest community discussions, mostly presented in blogs and podcasts. Pick a problem, like data tables, to learn the accessibility issues from these experts. find and create good and bad examples, but note you may need a screen reader software for this. can you characterize the alternatives and tradeoffs in CT terms?
Action: Create or try some web page features in several different browsers. Notice the differences in appearance and operation. Which sections of WCAG apply to noticeable differences or similarities?
Action: What is the career connection of computer science and web design? What are the demographics, salary, portability, and other qualities of web design versus traditional CS and SE jobs?
Transparency and dissemination of federal government data is drawing attention to accessibility
First, a remodeled whitehouse.gov drew accolades and criticisms. New websites like data.gov and recovery.gov appeared to reinforce the Obama administration promises. Disability.gov showed up on my radar screen through its Twitter flow. All these web sources, are now in my RSS feed reading regime. But the websites seem to be still behind on some aspects of accessibility, and under scrutiny by activists, including me. Personally, I’d be satisfied with a common form for requesting data and services, not the elements itself but well evolved interaction patterns through feedback and validation. More importantly, the data sets and analyses are challenging for visually impaired people, suggesting even new scientific research and novel technology to utilize alterative non-visual senses and brain power.
Additionally, innovation in assistive technology and accessibility is recognized at the National Center for Technology Innovation, with emphasis on portability and convergence with mainstream technology. Indeed, apparently, there are stimulus funds available in education and in communication systems.
Action: Visit the various USG cabinet department websites and then write down your main perception of their quality and ability to answer questions.
Action: Find examples of USG website forms users fill out for contacts, download of data sets, mailing lists, etc. How easy is filling out the forms> what mistakes do you make? How long does each take? Which forms are best and worst?
Check out on recovery.gov whether any stimulus funds are being spent on assistive technology. Or perhaps that information is on Deptart of Education sites as plans or solicitations.
Mainstream and assistive technologies are beginning to cross over.
BusinessWeek notes a number of examples:
Clearly mobile devices are driving this change. Embedding VoiceOver in Mac OS, transferred then to products like IPod Touch, has motivated a number of blind “screenless switchers”. Google calls its version on Android “eyes-free”. For those long stuck in the “blindness ghetto” of products costing $1000s with small company support and marketing chains through disability support service purveyors, this is a big deal. Conversely, although limited by terms of amendment under the Chafee agreement, members of Bookshare have enjoyed access to a rapidly growing library of texts, really XML documents, read in synthetic speech by now pocket size devices than cross Kindle and IPod capabilities. There’s never been a better time to lose some vision if one is a technology adopter willing to spend off retirement funds to remain active and well informed. The aging baby boomer generation that drives USA cost concerns will be a vast market in need of keeping up with the government flow of information, electronic documentation, not to mention younger generations.
But, while this Vision Loser is happy with the technology trend, to those disabled around the world working with older or non-existent computing environments this and free, open source trends make truly life changing differences.
Action: What are the job qualifications for working in the areas of assistive technology and accessibility? Is this business are growing, and in what regions of the USA or the world?
social media opens the culture of disability and the assistive markets for all computing professionals to explore.
while the cultures of disability may operate separate systems of societies and websites, in the case of vision impairment, the resources are right there for everybody to learn from, primarily by demos disseminated as podcasts by blind cool Tech, accessible world, and vendors. several annual conferences feature free exhibit halls visited by disability professionals, independent disabled like me, and luminaries like stevie wonder. cSUN is the biggest and a good place to get vendor and product lists. Again, many products can be seen in local disability support services. Local computer societies and CS courses may find well equipped people who can present like my Using things that Talk. This is a vibrant world of marketing closely couple with users, highly professional demos, and innovative developers, often disabled themselves. I personally treasure shaking hands with and thanking the young blind guys behind my Levelstar Icon and NVDA screen readers. Also, mailing lists are to various degrees helpful to the newly disabled, and rarely particular about age and gender. it’s a great technology culture to be forced into.
Action: Whenever you’re in a large enough city, visit their local vision training centers. I think you’ll be welcome, and might leave as a volunteer.
Action: With well over a thousand podcasts, dozens of blogs, and a regular tweet stream, the entry points for learning are abundant. However, the terminology and styles of presenters and presentations vary widely. Consider an example, often used in computer science, like David Harel’s watch, the microwave oven, or elevator controller. How do the state diagrams manifest in speech interfaces? Can you reverse engineer device descriptions using computational thinking? How could this help disabled users or accessibility providers?
Text-to-speech (TTS) is a mature technology with commodity voices.
Screen reader users rely on software implemented speech engines which use data files of word-to-sound mappings, i.e. voices. built into Mac Os, and widely available in windows and Linux, this mature technology supports a marketplace of voices available in open source or purchased with varying degrees of licensing, at a cost of about $25. comparable engines and voices are the main output channel of mobile assistive devices, like now I type on the Levelstar Icon. web pages, books, dialogs, email, … reading is all in our mind through our ears, not our eyes. An amazing and not yet widely appreciated breakthrough of a lineage of speech pioneers dating back to 1939 through DecTalk ATT Natural voices and now interactions with voice
Action: Wikipedia has a great chronology and description of synthetic speech. Track this with Moore’s law and the changes of technology over decades.
Action: Compare synthetic voices, e.g. using samples from vendor nextup.com or the ‘As Your World Changes’ blog samples.
Processor and storage enable more and more talking devices. why not everything?
Alarm clocks, microwave ovens, thermostats, and
many more everyday objects are speech enabled to some degree, see the demos on blind cool Tech and accessible world. I carry my library of 1000+ books everywhere in a candy bar sized screen-less device. but why stop until these devices are wirelessly connected with meaningful contextual networks. Thermostats could relay information about climate and weather trends, power company and power grid situations, and feedback on settings and recommended adjustments. Devices can carry their own manuals and training.
Action: Listen to podcasts on blind cool Tech and accessible world about talking devices and how they are in use by visually impaired people. Reverse engineer the devices into state machines, use cases, and write conversations between devices and users in “natural language”, assuming ease of speech output.
Action: Inventory some devices that might be redesigned for talking, even talkative. Electrical or chemical laboratory instruments, medical devices, home appliances, cars and other moving things, etc. But what would these devices speak? How do they avoid noise pollution? interference? annoyance?
Action: Computer science researchers are great at devising advanced solutions that provide service to relatively few disabled persons. For example, I have no use of GPS because if I’m somewhere I don’t know, I’m in bigger trouble than needing coordinates. This would b different in a city with public transportation, maybe. How do we evaluate technology solutions with the user, not the technology purveyor, as the main beneficiary?
Pivotal technology for visually impaired, the screen reader, is rapidly evolving through open source
A screen reader doesn’t really read pixels but rather the interfaces and objects in the browser and desktop. GUI objects expose their behaviors and properties for the screen reader to read and operate via TTS. Listen to the demos of Cs websites you may be familiar with. Unfortunately the marketplace for screen readers has been priced at over $1000 with steep SMA updates and limits in trials and distribution. Products largely sold to rehab and disability services passed on to users, with limited sales to individuals. This is a killer situation for older adults who find themselves needing assistance but without the social services available to veterans, students, and employee mandated. Worse, product patents are being employed by lawyers and company owners (some non USA) as competitive lawsuits.
however, the world has changed with the development over the past few years of NVDA, Non visual desktop access, originating in Australia with grants from Mozilla, then yahoo and Microsoft. A worldwide user community adapts NVDA for locale and Tts languages, with constant feedback to core developers. gradually, through both modern languages (Python) and browser developer collaborations, NVDA is challenging the market. You can’t beat free, portable, and easily installed if the product works well enough, as NVDA has for me since 2007. It’s fun to watch and support an agile upstart, as the industry is constantly changing with new web technologies like ARIA. The main problem with NVDA is robustness in the competing pools for memory resources and inevitable Windows restarts and unwanted updates.
Action: download and install NVDA. Listen to demos to learn its use. You will probably need to upgrade TTS voices from its distributed, also open, Espeak.
Action: learn how to test web pages with NVDA, with tutorials available from Webaim and Firefox. Define testing criteria (see standards) and processes. Note: good area here for new educational material, building on CS and SE testing theories and practices.
Action: develop testing practices, tools, and theories for NVDA itself. since screen readers are abstraction oriented, CT rigor could help.
Action: Modify NVDA to provide complexity and cost information. Is there a Magic Metric that NVDA could apply to determine with, say 80% agreement with visually impaired users, that a page was OK, DoOver, or of questionable quality in some respect?
structured text enables book and news reading in a variety of devices..
DAISY is a specification widely implemented to represent books, newspapers, magazines, manuals, etc. Although few documents fully exploit its structuring capabilities, in principle, a hierarchy of levels with headings allows rapid navigation of large textual objects. for example, the Sunday NY Times, has 20 sections, editorials, automobiles, obituaries, etc. separated into articles. Reading involves arrowing to interesting sections, selecting articles, listening in TTS until end of article or nauseous click to next article. books arrive as folders of size usually less than 1 MB. reader devices and software manage bookmarks, possibly in recorded voice, and last stopping point, causes by user action or sleep timer. In addition to audible and National narrated reading services with DRM, The TTS reading regime offers a rich world from 60,000+ books contributed by volunteers and publishers to bookshare and soon over 1M DAISY formatted public books through bookserver.org.
These are not directly web accessibility capabilities as in browsers but rather do read HTML as text, support RS’s reading of articles on blogs, and include browsers with certain limits, as in no Flash.
Over time, these devices contribute to improved speech synthesis for use everywhere, including replacement of human voice organs. Steven Hawking, blogger heroine ‘left thumbed blogger’ Glenda with cerebral palsy, and others use computer and mobile devices to simply communicate speech.
Action: Listen to podcasts demos of devices like Icon, booksense, Plextalk, Victor stream. What capabilities make reading possible, tolerable, or pleasant? Voice, speed, flexibility, cost, access, …?
Accessibility tools are available, corresponding to static analyzers and style checkers for code.
While not uniformly agreeing, accurate, or helpful, standards groups provide online validator’s to “test” accessibility. For example, WAVE from webaim.org, marks up a page with comments derived from web standards guidelines, like “problematic link”, “unmatched brackets”, java script interactions (if java script disabled), header outline anomalies, missing graphic explanations, small or invisible text. It’s easy to use this checker, just fill in the URL. However, interpreting results takes some skill and knowledge. Just as with a static analyzer, there are false hits, warnings where the real problem is elsewhere, and a tendency to drive developers into details that miss the main flaws. Passing with clean marks is also not sufficient as a page may still be overly complex or incomprehensible.
Action: Below is a list of websites from my recorded tour. Copy the link into WebAim.org WAVE (not the Google one) and track the markup and messages to my complaints or other problems. show how you would redesign the page, if necessary, using this feedback.
Action: redesign the ACM digital library and portal in a shadow website to show how a modern use of structured HTML would help.
Action: consider alternatives to PDF delivery formats. Would articles be more or less usable in DAISY?
Action: design suites of use cases for alternative digital libraries of computer science content. which library or search engine is most cost effective for maintenance and users?
Understanding of brain plasticity suggests new ways of managing disabilities
Brain science should explain the unexpected effectiveness and pleasure of reading without vision.
My personal story. Although I was experimenting with TTS reading of web pages, I had little appreciation, probably induced by denial, of how I could ever read books or long articles in their entirety. since it was
only a few weeks after I gave up on my Newsweek and reading on archetypes until my retina specialist pronounced me beyond the acuity level of legal blindness, I only briefly flirted with magnifiers, the trade of low vision specialists. rather, upon advice of another legally blind professional I met through her book and podcasts interviews, I immediately joined the wonderful nonprofit bookshare.org. A few trials with some very good synthetic voices and clunky PC-based software book readers lead me to the best at that time handheld device, the Bookport from APH, American Printing House for the blind. within weeks, I was scouring bookshare, then around 20,000 volumes, for my favorite authors and, wonders be, best sellers to download to my bookport. At first, I abhorred the synthetic voice, but if that was all that stood between me and regular reading, I could grow to love old precious Paul. going on 4 years, 2 GB of books, and a spare of the discontinued bookport, I still risk strangulation from ear buds at night with bookport beside me. Two book clubs broadened my reading into deeper unfamiliar nonfiction terrain and the Levelstar Icon became my main retriever from bookshare, now up to 60,000 volumes with many teenage series and nationally available school textbooks. I tell this story not only to encourage others losing vision, but also as a testimonial to the fact that I I am totally and continually amazed and appreciative that my brain morphed so easily from visual reading of printed books to TTS renditions in older robotic style voices. I really don’t believe my brain knows the difference about plot, characters, and details with the exception of difficult proper names and tables of data (more later). Neuroscientists and educators write books about the evolution of print but rarely delve into these questions of effectiveness and pleasure of pure reading by TTS. The best exceptional research is Clifford Nass A ‘wire for speech’ on how our brains react to gender, ethnicity, age, emotion, and other factors of synthetic speech. such a fascinating topic!
Action: Listen to some of the samples of synthetic speech on my website, e.g. the blockbuster ‘Lost symbol’ sample. Which voices affect your understanding of the content? How much do you absorb compared with reading the text sample? Extrapolate into reading the whole book using the voices you prefer, or can tolerate, and consider how you might appreciate the book plot, characters, and scenery Do you prefer male or female voices? Why?.
Numerical literacy is an open challenge for visual disability.
I personally encountered this problem trying to discuss a retirement report based around asset allocations expressed in pie charts. Now, I understand charts well, even programmed a chart tool. But I could find no way to replace the fluency of seeing a pie chart by reading the equivalent data in a table. This form of literacy, a form of numeracy, needs more work in the area of Trans-literacy, using multiple forms of perception and mental reasoning. Yes, a pie chart can be rendered in tactile form, like Braille pin devices, but these are still expensive. Sound can convey some properties, but these depend on good hearing and a different part of the brain. Personally, I’d like to experiment with a widget operated by keyboard, primarily arrow keys, that also read numbers with different pitches, voices, volume, or other parameters. The escalating sound of a progress bar is available in my screen reader, for example. Is there a composite survey somewhere of alternative senses and brain training to replace reading charts? Could this be available in the mainstream technology market? How many disabilities or educational deficiencies of education and training might also be addressed in otherwise not disabled people?
Is there an app for that?
Action: Inventory graphical examples where data tables or other structures provide sufficient alternatives to charts? Prototype a keyboard-driven, speech-enabled widget for interaction with chart like representations of data. Thank you for using me as a test subject.
Action: Moving from charts to general diagrams, how can blind students learn equivalent data structures like lists, graphs, state machines, etc.?
Web science needs accessibility criteria and vice versa.
The web is a vast system of artifacts, of varying ages,
HTML generations, human and software generated, important, etc. could current site and page accessibility evaluation scale to billions of pages in a sweep of accessibility improvement?
Surveys currently profile how screen readers are used and the distribution of HTML element usage.
Do a web search, in bing, Yahoo, google, or dogpile, whatever, and you’ll probably find a satisficing page , and a lot you wish not to visit or never visit again. Multiply that effort by , say 10, for every page that’s poorly designed or inaccessible to consider the search experience of the visually impaired. Suppose also that the design flaws that count as accessibility failures also manifest as stumbles or confusion for newer or less experience searchers. Now consider the failure rate of serious flaws of, , say, 90% of all pages. Whew, there’s a lot of barriers and waste in them there web sites.
experienced accessibility analysts , like found on webAxe podcasts and blog, can sort out good, bad, and just problematic features. Automated validation tools can point out many outright problems and hint at deeper design troubles.
Let’s up the level and assume we could triage the whole web, yep, all billions of pages as matched with experimental results of real evaluators, say visually impaired web heads like me and those accessibility experts. This magic metric, MM, has three levels: OK, no show stoppers by human evaluators; at 80% agreement; DO OVER, again with human evaluators 80% agreement of awfulness; and remaining requiring reconciliation of human and metric. Suppose an independent crawler or search engine robot used this MM to tag sites and pages. probably nothing would happen. but if…
Action: declare a week of clean Up the web, where the MM invokes real Acton to perform “do over” or “reconcile”. Now, we’re paying attention to design factors that really matter and instigating serious design thought. All good, all we need is that MM.
Action: which profession produces the most accessible pages, services, and sites? computer scientists seem to be consistently remiss on headings, but are chemists or literary analysts any better? If acm.org is as bad as I claim, are other professional societies more concerned about quality of service to their members? what are they doing the same or differently?
How does the quality of accessibility affect the science of design as applied to web pages, sites, and applications?
Accessibility needs a Science of Design and Vice Versa
Accessibility concerns often lead into productive unexplored design regions.
Accessibility and usability are well defined if underused principles of product quality. The ‘curb cuts’ principle suggests that a defect with respect to these qualities is in a poorly understood or unexplored area of a design. Often a problem that presents only a little trouble for the expected “normal” user is a major hassle or show stopper for those with certain physical or cognitive deficiencies. However, those flaws compound and often invisibly reduce productivity for all users. Increasingly, these deficiencies arise from ambient environmental conditions such as glare, noise, and potential damage to users or devices.
Moreover, these problems may also indicate major flaws related to the integrity of a design and long term maintainability of the product. An example is the omission of Headings on an HTML page that makes it difficult to find content and navigation divisions with a screen reader. This flaw usually reveals an underlying lack of clarity about the purpose and structure of the website and page. Complexity and difficult usability often arise from missing and muddled use cases. Attitudes opposing checklist standards often lead to perpetuating poor practices such as the silly link label “click here”.
The ‘curb cuts’ principle leads toward a theory of design that requires remedy of accessibility problems not as a kindness to users nor to meet a governmental regulation but rather to force exploration through difficult or novel parts of the design terrain. The paradigm of “universal design” demands attention to principles that should influence requirements, choice of technical frameworks, and attention to different aesthetics and other qualities. For example, design principles may address where responsibilities lie for speech information to a user, thus questioning whether alternative architectures should be considered. Applying this principle early and thoroughly potentially removes many warts of the product that now require clumsy and expensive accessibility grafts or do-overs.
Just as the design patterns movement grew from the architectural interests of Christopher Alexander, attention to universal design should help mature the fields for software and hardware. The “curb cuts” principle motivates designers to think beyond the trim looking curb to consider the functionality to really serve and attract ever more populations of end users.
The accessibility call for action, accommodation, translates into a different search space and broader criteria plus a more ethically or economically focused trade-off analysis. now, design is rarely explicitly exploration, criterion’s, or tradeoff-focused. but the qualitative questions of inclusive design often jolt designers into broader consider of design alternatives. web standards such as WCAG 2.0 provide ways to prune alternatives as well as generate generally accepted good alternatives. It’s that simple: stay within the rules, stray only if you understand the rationales for these rules, and temper trade-off analysis with empathy toward excluded users or hard cool acceptance of lost buyer or admirers. well, that’s not really so simple, but expresses why web standards groups are so important and helpful — pruning, generating, and rationalizing is their contribution to web designers professional effectiveness and peace of mind.
Action: Reconstruct a textbook design to identify assumptions about similarities and differences of users. Force the design to explore extremes such as missing or defective mouse and evaluate the robustness of the design.
Action: Find an example of a product that illustrates universal design. How were its design alternatives derived and evaluated?
revving Up our computational Thinking on accessibility
POSH (Plain Old semantic HTML) and headings
POSH focuses our attention on common structural elements of HTML that add
meaning to our content with Headings and Lists as regular features. An enormous
number of web pages are free of headings or careless about their use. The
general rule is to outline the page in a logical manner: h1, H2, h3,…,H6, in
why is this so important for accessibility?
headings. support page abstraction. reaching a page, whether first or return
visit, I, and many other screen reader users, take a ‘heading tour’. Using our ‘h’ key repeatedly to visit headings, gives a rapid-fire reading of the parts of the page and an
introduction to the terminology of the web site and page content. bingo! a good
heading tour and my brain has a mental map and a quick plan for achieving my
purpose for being there. No headings and, argh, I have to learn the same thing
through links and weaker structures like lists. At worst I need to tab along
the focus trail of HTML elements, usually a top-bottom, left-right ordering.
Page abstraction enables better than linear search if I know roughly what I
want. for example, looking for colloquium talks on a Cs website is likely to
succeed by heading toward News and Events, whatever. with likely a few dozen
page parts, linear search is time and energy consuming, although often leading
to interesting distractions.
Page abstraction encourages thinking about cohesion of parts, where to
modularize, how to describe parts, and consistent naming. This becomes
especially important for page maintainers, and eventually page readers, when
new links are added. Just like software design, cohesion and coupling plus
naming help control maintenance. An example of where this goes wrong is the
“bureaucratic guano” on many government web pages, where every administrator
and program manager needs to leave their own links but nobody has the page
structure as their main goal.
while it’s not easy to prove, but plausible, SEO (search engine optimizers)
claim headings play a role in page rankings. This appeals to good sense that
words used in headings are more important so worth higher weights for search
accuracy. It might also mean pages are better designed, but this is just
conventional wisdom of users with accessibility needs.
so, we have abstraction, search, design quality, and metrics applied to the
simple old semantic HTML Heading construct.
Now, this rudimentary semantic use of Headings is the current best practice, supplementing the deprecated Accs Tags that all keyboard users can exploit to reach standard page locations, like search box and navigation. Rather, headings refine and define better supplements for access tags. Going further, the ARIA brand of HTML encourages so-called ‘landmarks’ which can also be toured and help structure complex page patterns such as search results. The NVDA screen reader reports landmarks as illustrated on AccessibleTwitter and Bookshare. Sites without even Headings appear quaint and deliberately unhelpful.
The Readable conference program Problem
I recently attended a conference of 3.5 days with about 7 tracks per session.
The document came as a PDF without markup, apparently derived from a WORD
document with intended use in printed form. Oh, yeah, it was 10MB download with
decorations and all conference info.
I was helpless to read this myself. yes, I could use the screen reader but
could not mentally keep in mind all the times and tracks and speakers and
topics. I couldn’t read like down Tracks or across sessions nor mark talks to
attend. Bummer, I needed a sighted reader and then still had to keep the
program in mind while attending.
A HTML version of the preliminary program was decidedly more usable. Hey, this is what hypertext is all about! Links from talks to tracks and sessions and vice versa, programs by days or half-days subdivided on pages, real HTML data tables with headers that can be interpreted by screen reader, albeit still slowly and painfully.
that’s better, but would be unpopular with sighted people who
wanted a stapled or folded printout.
OK, we know this is highly structured data so how about a database? This would
permit, with some SQL and HTML, wrapping, generation of multiple formats, e.g.
emphasizing tracks or sessions or topics,… But this wouldn’t likely distill
into a suitable printable document. Actually, MS WORD is programmable, so the
original route is still possible but not often considered. Of course, it’s often more difficult to enter data into forms for a database, but isn’t that what student helpers are for? Ditto the HTML generation from the database.
The best compromise might be using appropriate Header styles in WORD and
use an available DAISY export so the program in XML could be navigated in our
This example points the persistent problem that PDF, which prints well and
downloads intact, is a bugger when it loses its logical structure. Sighted
readers see that structure, print disable people get just loads of text. This
is especially ironic when the parts originally had semantic markup lost in
translation to PDF, as occurs with NSF proposals.
so, here I’m trying to point out a number of abstraction problems, very
mundane, but amenable to an accommodation by abstracting to a database type of
model or fully exploiting markup and accessible format in WORD. Are there other
approaches? Does characterizing this problem in terms of trade-offs among abstractions and loss of structural information motivate computer scientists to approach their conference responsibilities different?
More generally, accessibility strongly suggests that HTML be the dominant document type on the web, with PDF, TXT, WORD, etc. As supplementary. Adobe and free lance consultants work very hard to explain how PDF may be made accessible, but that’s just not happening, nor will this replace probably millions of moldering PDFs. Besides negligent accessibility, forcing a user out of a browser into a separate application causes resources allocated and inevitable security updates.
Design by Progressive Enhancement<
So, here’s a challenge for those interested in Science of Design, design patterns, and test methods with many case studies on the web, discussed in blogs and podcasts.
Touring CS websites by screen reader
— download MP3
Are you up for something different? Download
MP3 illustration of POSH Computer Science websites 45 minutes, 20 MB
. This is me talking abot what I find at the following locations, pointing out good and bad accessibility features. You should get a feeling of life using a screen reader and how I stumble around websites. And, please, let me interject that we’re all learning to make websites better, including my own past and present.
Note: I meant POSH=”Plain old semantic HTML” but sometimes said “Plain old simple HTML”. Sorry about the ringing alarm. Experimental metadata: Windows XP, Firefox, NVDA RC 2009, ATT Mike and Neo speech Kate, PlexTalk Pocket recorder.
Web Sites Visited on CSE screen reader tour
U. Texas Austin
Firm accessibility statement;
headings cover all links?;
good to have RSS;
No headings, uses layout tables (deprecated);
good use of ALT describing graphics;
all headings at H1, huh?;
non informative links ‘learn more’;
CS at cmu.edu
non informative graphics and links;
unidentified calendar trap;
definitely not POSH
Computational Thinking Center at CMU
strange term probes:;
non informative links PPT, PDF;
CRA Computing Research Association
interminable links unstructured list;
not so POSH
ACM.org and DL portal
irregular headings on main page;
no headings on DL portal;
noninformative links to volumes;
hard to find category section;
Computer Educators Oral History Project CHEOP
don’t need “looks good” if standard;
NCWIT National Center Women Information Technology
doesn’t conform to accessibility statement;
graphics ALT are not informative;
link ‘more’ lacks context;
do over for POSH
So, what to do with these POSH reports?
Clearly, some sites could use some more work to become world class role models for accessibility. At first glance, my reports and those that would be compiled from validator’s like WebAim WAVE indicate that some HTML tweaking would yield improvements. Maybe, but most websites are under the control of IT or new media or other departments, or maybe outsourced to vendors. Changes would then require negotiation. Another complication is that once a renovation starts, it is all too easy to use the change for a much more extensive overhaul. Sometimes, fixes might not be so easy, as often is indicated by the processes of progressive enhancement. This is classical maintenance process management, as in software engineering.
However, hey, why not use this as a design contest? Which student group can produce a mockup shadow website that is attractive and also meets the WCAG, validator, and even the SLGer tests?
Just saying, here’s a great challenge for CSE to (1) learn more about accessibility and web standards, (2) make websites role models for other institutions, and (3) improve service for prospective students, parents, and benefactors.
conclusion: A Call To Action
To the proclamation, let us informally add
whereas society, including the Cs field itself, requires that all information, computer-based technology be available to all persons with disabilities,
whereas computer science is the closest academic field to the needs and opportunities for universal accessibility,
Disabled individuals are particularly under-represented in computing fields, in disparate proportion to the importance of disability in the economic and social well-being of the nation
computer science educators will adapt their curricula to produce students with professional awareness of the range of human abilities, the resources for responding to needs of persons with disabilities
computer science education will be open and welcoming to all persons with disabilities both helping the person to reach their own employment potential and opportunity to contribute to society and (2) inform educators and other students about their abilities, needs, domain knowledge,
See next post for Webliography
Comments, Corrections, Complaint?
Please add your comments below and I’ll moderate asap.
Yes, I know there are lots of typos but I’m tired of listening to myself, will proof-listen again later.
Longer comments to firstname.lastname@example.org. Join in the Twitter discussion of #accessibility by following me as slger123.
Thanks for listening.