Breaking the barriers: digital accessibility

To make digital content accessible to everyone, a combined strategy is needed: recommendations, legislation, control, and technical guidance should be used in a coordinated way. The recommendations must be reviewed critically, to prevent the means from becoming goals. This process requires that people with special needs participate in it and take an active role in making digital technology easier to everyone.

This document accompanies the author’s presentation in the First Kuwait International Conference on the Role of People with Special Needs in Building the Information Society “Access to Information” (Kuwait, May 1 to 3, 2006). The presentation in PowerPoint form is available. This page is a more detailed description and contains references and links. The page has been converted from the original Microsoft Word form of this document.

The need for accessibility

The information society means that activities of all kinds depend more and more on computer technology, including networks. Therefore, the basic skill of using computers, or “computer literacy,” is becoming a necessity of life. The information society opens huge possibilities, but it also creates new barriers.

Techniques used on web pages, and in digital media, have often been designed and used without due consideration of people with special needs. Using the most modern technique may exclude large number of people, creating unnecessary barriers and obstacles.

To illustrate the problems, we can visit typical web pages using a text-only browser. This gives a rough idea of the way blind people experience the pages when they use a speech-based browser (which converts the textual content to synthesized speech) or a Braille device (which converts the textual context to dot patterns that can be sensed with fingertips). The obstacles created by using images without text equivalents vary from minor inconvenience to total blockade.

For example, the English version of main page of the United Nations (http://www.un.org/english/) starts as follows:

   People from around the world People from around the world People from
   around the world People from around the world "We the peoples" - Link
   to the United Nations Charter People from around the world People from
    around the world People from around the world People from around the
                                   world

     Daily Briefing by the Spokesman press releases Radio, TV and Photo
     Official Documents, Maps Publications, Stamps and Databases The UN
                                   Works
      Peace and Security Economic and Social Development Human Rights
                  Humanitarian Affairs International Law


                       Welcome to the United Nations

                                UN 60 logo

This nonsense has been caused by writing textual alternatives to images without thinking why. The first images are mere decoration and would need empty alternatives. In the present design, the alternatives just cause frustration and confusion.

The graphic presentation of the UN page is the following, with some easily identifiable problems to people with reduced eyesight or other special needs:

(Screenshot of UN main page)
Problems on the page include:

It is true that the UN site contains links to “text versions.” However, text versions are a problem rather than a solution. First, links to text versions do not work in many browsing situations, due to technical mistakes in their design. Second, “text versions” carry the connotation of secondary access that is intended for secondary people: disabled people, please use the back door. Third, since “text versions” are so often less content-rich and less up-to-date than normal versions, people avoid them or remain uncertain of whether they have full access to the contents. For example, there was once a national broadcasting company with a web site with daily news so that the “text version” contained only the headlines of news, with no information on this restriction. “Text version” has become the symbol of secondary quality.

Aspects of accessibility

Accessibility of digital content is much more than accessibility to the blind and visually impaired. However, the problems of the blind are easy to demonstrate and understand, and there are several simple ways of alleviating the problems. It is also socially safer to refer to such problems than to problems experienced by many other groups of people, such as mentally retarded people.

Quite often, the overall complexity of the structure and layout of web pages makes them difficult to understand and use to average people. The complexity can be a serious obstacle to people with special needs. This applies particularly to public sites, which typically have the challenge of disseminating complex information. The simple explanation to this is that authors of web sites are generally well above the average in intelligence, literacy, knowledge, perception, motorics, and cognitive abilities. By instinct, everyone authors for her or his peers or perhaps even for those that are above the author on some scale. Only through conscious efforts can we consider a wider audience.

One of the key problems on public web sites is difficult language, which is often formal and literary, with words that are uncommon in spoken language and with long sentences. It is especially challenging, if not impossible, to people with below-average cognitive skills and to people who need to learn texts in a language that is not their native language. Yet, much of the work of public organizations is, or should be, carried out to serve people with special needs.

Many people have motoric or other problems that prevent efficient use of a computer mouse. Yet, many applications have been designed exclusively for use with a mouse. In web design, simple principles make it possible to use a page without a mouse, typically using the keyboard for all controls. This mostly means simply not using techniques that only work with a mouse.

Why focus on web accessibility?

Activities aimed at breaking the barriers and enabling all people to live in the information society have largely been focused on digital accessibility. Specifically this means accessibility of World Wide Web and intranet pages to all people, including people with disabilities and elderly people. . In this area, the recommendations of the Web Accessibility Initiative (WAI, http://www.w3.org/WAI/) by the WWW Consortium have widely been cited as the generally accepted principles.

An obvious reason to the focus on WWW accessibility is the impact and fast development of the WWW. Another reason is that in that environment, there are so many problems that can be recognized and solved, at least partially. In fact, the work has largely been related to the needs of the blind and other visually impaired people, because there are specific techniques that can be used to help them. It has been said that scientists do not see problems before they have some idea of what the solution might be; this may apply to people in general. For example, you might never regard accessibility to the blind as a problem, if you had no idea of how it could be solved, i.e. how blind people could use web pages and other digital content.

In the information society, information is often available in digital format only. This opens new possibilities, since digital formats can be used to render the content in a manner that is suitable to each person, using different fonts, colors, etc., or perhaps as synthesized speech or accessed via search functions. This is the technical basis for the importance of WWW accessibility. However, the same basic problems and possibilities arise in other forms of digital publishing. Reference books, learning material, and other types of information are increasingly published on CD-ROM, DVD, memory sticks, etc. At present, the format is usually rigid, often using small fonts, insufficient contrast of colors, etc. What is needed for accessibility is not larger fonts but open data formats and software that supports them, so that the representation of information can be tailored according to individual users’ needs. This conflicts with vendors’ attempts to protect their material against illegal copying and distribution.

The four challenges

There are four major challenges in digital accessibility. We have insufficient information about the different special needs that people have. We know some ways to deal with the needs, but we also know that they solve part of the problems only. Therefore, we need new techniques and methods. The third challenge is to define, at different levels (international, national, organizational), the accessibility goals and policies to be applied, with due consideration of potential conflicts with other goals. Finally, we need to find ways to ensure the actual implementation and to check its effectiveness, with feedback to the other phases. The implementation phase is greatly dependent on awareness raising among users and authors, but especially managers.

Laws or advice?

The WAI recommendations mentioned above have had a relatively small impact on reality. Most web pages are more or less inaccessible to large groups of people with special needs. Obviously, effective actions are needed to make accessibility common.

Should accessibility be made obligatory by law and regulations, or is it better to just give recommendations and guidance? In Europe, the latter approach has been taken, and it has largely been unsuccessful. Recommendations, even those made by the European Commission, the European Parliament, and national governments, are ignored by most web authors. They have had some impact on public web sites and intranets. However, there is a considerable credibility problem: even the organizations that issue recommendations usually fail to comply with them themselves.

Moreover, the European recommendations mostly just echo the WAI recommendations, instead of specifying more concrete rules. It is difficult to see what the added value is. Sometimes it might be negative, because web authors and designers have to deal with two or more sets of recommendations instead of one. They get confused and frustrated.

Competitions have also been suggested as a method of promoting accessibility. In Finland, detailed quality criteria have been developed and used. However, evaluating sites with them seems to have little practical impact. Measuring does not make sites better. Honorary rewards do not affect page authors and designers much.

In the United States, part of the worldwide accessibility recommendations was turned into law, especially in the so-called Section 508 legislation. It was mostly derived from the WAI guidelines, but it contains only rules that can be described in a sufficiently exact manner. This means that rules must be defined so rigorously that it can be objectively, or even “mechanically,” decided whether they have been followed. This tends to move the focus from accessibility as a whole to those aspects that are enforceable by law, and they are often not the most important aspects. There is a side-by-side presentation of WAI guidelines and Section 508 rules at http://jimthatcher.com/sidebyside.htm.

Moreover, the legislation mainly applies to web pages and intranets of organizations funded by the federal government. It also contains the exemption that some rules may be ignored if fulfilling them would require an “undue burden.” This is both realistic

The US approach has the benefit of making some work on accessibility obligatory in some contexts. One of the problems of the voluntary approach is that since accessibility has a cost—though often much smaller than expected—it is usually not interesting to private companies, or even to public organizations, which live under many pressures. To put it crudely, free competition as well as efficiency requirement works against accessibility. Quite often, accessibility means costs more than tangible gains, in the short range, and the range is so often just a quarter of a year. When accessibility helps large user communities, such as elderly people, it may in fact imply competitive gains, but companies still don’t see this, as a rule. They classify accessibility as serving small minorities only, such as the blind, and they think they just cannot afford it.

Both legislation and recommendations easily become a way to push the responsibility to others. The politicians and officials may think that by issuing laws or by giving recommendations they have done their part of accessibility. It is then natural that they may even forget to follow the recommendations (or the spirit of the law) in their own activities. As a perhaps extreme symptom of this, when the European Parliament had approved a resolution on accessibility of public web sites, explicitly stating that the PDF document format shall not be used as the only format for any information, the resolution itself was available only in PDF for quite some time.

When guidelines, techniques, and tools become the goal

Another problem with both laws and recommendations is that they easily become the goal rather than means. In particular, the WAI guidelines, though extremely valuable when used properly, have become part of the problem, too. They are often applied without understanding the principles behind them and, most importantly, the needs of disabled people. This may result in striving for “accessibility” as a goal that is regarded as identical to the WAI guidelines, rather than actually making sites more accessible to people.

Further complications are caused by the fact that the general web accessibility principles themselves are partly out of date. The WAI guidelines are several years old and reflect, in part, technologies that are not relevant any more. The guidelines even contain principles that actually reduce accessibility.

For example, the accesskey principle was aimed at making web sites navigateable using keyboard only. The idea is that you can use a web page with a common keyboard as the only input device. This is important to people who cannot use a mouse or cannot use it accurately enough to point at items on screen. The accesskey concept was introduced into web authoring to allow authors to assign keyboard key combinations for navigation on a page. However, due to poor design of the accesskey technique, such assignments tend to make navigation more difficult, in most experts’ present opinion. For example, a web page might assign the key combination “Alt+F” for some page-specific function, like entering a field; but this may mask out the built-in functionality of “Alt+F,” such as opening a File menu in a browser. Yet, the accesskey principle has been included into the WAI guidelines and approved by different authorities, such as the UK government, which even mandates a “standard set” of accesskeys. Thus, many authors are now required to spend time on adding features that are known to be harmful.

Similarly, checklists can lead to problems. Checklists are meant to help authors to evaluate and improve the accessibility of their pages, using simple criteria. In practice, most checklists that have been composed are essentially as large as the WAI guidelines. Thus, they tend to be too extensive for most authors to grasp and use. Moreover, they may contain requirements that are out-of-date or even misguided from the beginning, with little possibilities for fixing the problems. One reason to this is that checklists have typically been created, by national and other organizations, as projects, i.e. as activities with predefined goals, resources, goals, and deliverables. This implies that usually nobody is interested in the matter after the project has ended. Any errors made in defining the checklist will remain carved in stone, at least until the next revision project. Thus, checklists should be taken care of by continuous activities rather than projects.

Worse still, the software tools for evaluating web site accessibility are wanting, and largely just useless. Most of the accessibility guidelines cannot be appropriately checked by automatic tools. Yet, some of the tools (such as the Bobby checker) have been interpreted as setting the standards, so their misleading or even outright wrong reports have used as a basis of redesigning sites.

Web accessibility is both extremely simple and extremely complicated. On one hand, it is easy to recognize ways to improve the accessibility of almost any site. There are simple steps forward that can be taken. Often this means just removing features, such as fixed font sizes, or refraining from creating them in the first place. On the other hand, accessibility ultimately means being accessible to any human being, and this is an unreachable goal. It is impossible to cover the entire variation in people’s abilities and capabilities of using web content. We need to work with the problems we have recognized, with the techniques we know, but we should humbly confess that we still know and address just part of the problems. It is dangerous to regard accessibility as a fixed set of rules, or even as a fixed set of concrete goals.

Thus, people with special needs should express their dissatisfaction with web sites that are advertised as “accessible,” if the sites in fact pose serious problems to them. They may need special help in this, since accessibility problems often include difficulties in sending feedback, such as forcing people to use a form rather than E-mail or phone.

Understandable texts

As mentioned previously, most public web sites use far too complicated language. They use the language of bureaus. They have complicated sentence structure, lots of difficult words, specialized abbreviations, and technical terms that might be completely unknown outside the organization. The texts are difficult even to people with academic education, very difficult to average users, and often completely inaccessible (not understandable even with great effort) to people with cognitive below the average, or elderly people who do not know modern vocabulary. This is one of the very real problems of accessibility. The problem exists on web sites of private companies, too, though in different forms. Quite often, such sites have young people as the target group, and this might mean that they use very modern language, with words that you can’t find in any normal dictionary.

Yet, there is just one vaguely formulated guideline on the language of texts, and there are no specific measurable criteria of understandability, still less automatic tools for measuring it. Computable readability indexes exist for English as well as some other languages, but there has been a general tendency to avoid them as too “mechanical.” As a result, most evaluations of accessibility completely ignore the readability issue. Sites might still display icons that claim conformance to recommendations that require that the simplest language possible be used!

Further research and development is needed to create tools for analyzing and improving the understandability of texts. When sufficiently adequate language-dependent analyzers are available, specific requirements can be imposed, if desired. At present, we could impose coarse rules. For texts in English, you could use e.g. the Flesch Reading Ease. It is a measure of readability in a 0–100 scale, with 100 being most readable, and 60–70 being reasonably readable to the public. This measure only counts lengths of sentences and words, but it can be calculated easily e.g. using tools in Microsoft Word. Requiring that Flesch Reading Ease be larger than 30 would be a start. (The present document has the figure 34.0.) Later we could impose stronger requirements, especially for texts that are very important to the public in general or to people with special needs.

In Europe, there are special guidelines on writing in “Easy-to-Read Language,” at http://www.inclusion-europe.org/information/eetr.htm. This means language adapted for people with below-average language skills or cognitive skills. Though little known, these guidelines are very helpful when applied accordingly. They also remind us of the fact that writing naturally and fluently, in a manner that makes texts sufficiently simple for the majority, still does not address all the special needs that people have.

A combined strategy

I can see two ways in which the situation can be improved. First, some accessibility requirements should be made compulsory to public organizations and private enterprises. This is comparable to enforcing physical accessibility of buildings by wheelchairs. A legal requirement creates equal burden to all competitors. This approach needs to be limited to requirements that are important to many people, can be enforced effectively, and do not create undue difficulties and costs. For example, creating easy-language versions of text or graphic presentation that acts as an alternative to text means parallel content production, so its cost can be excessive. The same applies to producing content in different languages or for different age groups, or to preparing synchronized subtitles for multimedia content. However desirable such production might be, it would be unwise to force organizations into it. On the other hand, it is usually quite reasonable to require that web pages be completely functional when accessed without images or used without a mouse, although we need to allow a relatively long period of transition.

Second, the existing and future recommendations should be accompanied with publicly funded support to accessibility work. In the European Union, such an approach has been attempted, in the sense of creating a network of “centres of excellence” in accessibility. This network, however, mainly exists on paper and as a bureaucracy; it lacks public visibility and credibility. Cooperation between accessibility specialists in different organizations can be useful, but it is little direct impact. What is needed is practical guidance and help in evaluating sites, improving them, and most importantly incorporating accessibility principles into the design of new sites and reforms of sites. It is generally much more expensive to fix accessibility problems than to avoid them by proper design. This means that accessibility needs to be integrated into the production process, rather than imposed as a separate layer. The forms of guidance and help need to vary, but they should be directed mainly to manager-level staff and to people who create and tailor authoring environments.

The support should also include the development of new technologies, such as text analyzers that help to make content more understandable. This is strongly language-dependent. It is not only a technological problem but also a matter of making such analyzers part of the set of tools that authors use routinely, perhaps as a built-in function of a normal word processor or a web authoring program. Leading software vendors cannot be expected to take care of such things without initiatives, information, and pressure from national authorities and experts.

Participation

People with special needs must be involved in the policy at all stages. Instead of treating them only as people to be helped, we must encourage them to participate actively.

Many principles of accessibility are essential to some minorities with special needs but also useful to much wider groups of people. To take a simple example, if a web site uses a very small fixed font size, removing this obstacle is crucial to some people, but many more—perhaps the majority of users—will find browsing more convenient. We should encourage people to complain loudly. Usually people who find something just a little inconvenient will just live with it. However, one man’s inconvenience can be a serious obstacle to someone else. The disabled, while defending their right to access digital content, also constitute an avantgarde force.

Accessibility techniques, as all technical innovations, may have spinoff effects that exceed the impact of their original use, at least commercially. Devices and software that are necessary to some disabled people can become useful or entertaining to many other people. Keyboard shortcuts in programs are a common example of this. A haptic mouse is a device that gives tactile feedback via pressure, so that it allows blind people to sense graphic shapes and patterns. In addition to such usage, a haptic mouse is nowadays used just for fun—it feels good.

This is a very positive development: it helps the disabled by making the products mainstream products that participate in normal development, marketing, and competition that tend to make the products better, cheaper, and something you can buy in a normal shop.

The needs of the few can be a convenience to the many. People with less than perfect eyesight constitute a very large part of the population, but they seldom raise their voice and demand better software, better web pages, or larger font in printed matter. If producers can be forced to make font size, colors, and other presentational features user-controllable to allow access to visually disabled people who need large fonts, then the majority benefits, too.

People with different needs

I would like to end with a perspective that extends the concept “people with special needs” to cover all people. After all, we are all different, and we have different needs. At one extreme, some differences are so essential in the use of modern information technology that they require special attention. At the other extreme, the differences are matters of personal taste in minor details.

Looking from this perspective, accessibility is part of a larger complex that also covers localization, customization, and personalization. Localization means making information technology adapt to various linguistic and cultural habits and preferences, instead of forcing them to adapt to the technology. This is increasingly important as the WWW becomes truly worldwide and needs to support the use of writing systems and languages that radically differ from English. Customization means making the technology tuned to different types of use and user communities, and personalization takes this to the level of an individual.

Personalization can mean, for example, possibilities for setting the visual appearance of a web site match the individual user’s esthetic preferences. This typically involves a choice between different basic designs and a possibility to change its details. Often regarded as luxury and a method of attracting customers, personalization can be important to accessibility, too. People with special needs, such as people with cognitive and communicative disabilities, may greatly benefit from a personalized user experience. It helps them feel at home, and personalization can be used to remove parts of user interface that might be distracting or even risky.


Jukka K. Korpela, jkorpela@cs.tut.fi