Web Accessibility from a Legal Perspective

Speech given to The High Level Group of Representatives of the Member States on Disability

Date: 24/10/2001
Venue: Brussels

Before I begin my presentation proper, I wish to make four remarks:

First, although I am one of the longest serving members from Europe of the Web Accessibility Initiative (WAI), of the World Wide Web Consortium. I am not speaking on its behalf but simply on my own behalf.

Secondly, however, I do not think it would be controversial for me to welcome the Communication from the Commission: Europe 2002: Accessibility of Public Web Sites and their Content (1).

Thirdly, I want to thank this High Level Group for inviting me, balancing that with the observation that it is only in the last paragraph of the Communication (Conclusions 10) that participation by disabled people in the process is actually mentioned and, even then, this concerns "feedback" not framing proposals. Disability representation throughout the process from People First (2), via the Lisbon Conference (3) to the establishment of this Group has been less than satisfactory.

Finally, I want to urge the European Union to continue to provide financial support to the Web Accessibility Initiative so that it continues to operate from, at the very least, a Trans-Atlantic perspective.

What I am going to try to do in this presentation is look at four ideas and how they relate: Law, Standards, Universal Service and Guidelines.

1. Law

Except in the sense of 'Natural law' and of some very basic rules of conduct and organisation, the rules that societies make for themselves depend critically upon the establishment of priorities. In our highly developed societies this prioritisation operates in two ways: firstly, the establishment and enjoyment of different rights often leads to a conflict of more or less equal forces; secondly, the legislative process can only handle a certain load and, again, account has to be taken of priority setting.

For disabled people the priority setting of the European Union in the context of the Information Society has, until this latest Communication, been depressing. I remember how much excitement the People First (2) paragraphs (96-113) on disability access generated, but as proposals have become more concrete, they have proportionately become less generous. Disabled people may be very important in theory but in a forum of competing interests they are marginalised. The evidence of this at a highly practical level is the recent decision to make access to information under the Copyright Directive (4) a matter of subsidiarity, not a right of citizenship under the protection of the Union. At the e-Europe level the highly specific proposals, which were first published in December 1999 (5), were steadily watered down during Ministerial negotiations. The Communication is, to put it bluntly, what we've been left with.

In a now infamous background briefing (6), disabled people were said not to require specific copyright provisions because they were not an economically significant enough group to warrant the attention. Apart from the fact that this statement reduced the worth of citizens to their economic power, it was wholly inaccurate. The demographics and economics are extremely important and I will turn to them in a moment. What I want to say first, however, is that it is my belief that the EU has no stomach for legislation to extend, entrench or in any other way guarantee the rights of disabled people in the Information Society any further than has been done already. If our rights to information in the information society are subject to subsidiarity and if the means of receiving information are to be steadily de-regulated, there is no hope in that direction.

Having got these negative observations out of the way, I want to spend the rest of my presentation being constructive, suggesting solutions for the Union, which are effective and not unduly expensive. What I want to suggest above all else is that optimal accessibility based on market share is a much better strategy within an organisation that is primarily economic not social, rather than trying to work in the context of a rights-based approach which is much more akin to the way these matters are discussed in the United States.

A market approach, however, fundamentally depends on demographics and it is to that which I will now turn.

2. Economics and Demographics

The obvious first problem with the legal aspect of disability is definition. Such definitions as there are apply for largely administrative reasons, to justify special expenditure, protection or privileges. Disability, in other words, is an administrative concept. This state of affairs has been carefully protected by disabled people themselves, particularly the activist leadership, because it protects special budgets for a small number of people. In the area of visual impairment, for example, it is difficult to get activists to agree to expenditure on enlarged print even though that is badly needed; the activists want to spend the money on Braille, which is only used by a tiny minority of the sector. The next problem, which leads from this, is understanding the size of the actual sector. Most people who are legally classifiable as "Disabled" would not self classify in that way; and most of the people who need some kind of customisation function to standard digital products would not be classifiable as disabled. The largest single element of the EU population which cannot access digital information is that proportion who are functionally illiterate. The next group in size is almost certainly made up of those for whom the first language of the place they live is not their own first language. Then there is a massive variety of cognitive barriers to access from mild dyslexia to problems with systems navigation. Next are those with physical difficulties, particularly problems with their hands, who find the manipulation of small devices like the mouse, very difficult; but you would not classify somebody with chronic arthritis as disabled. Next is the group which finds hearing difficult, an increasingly difficult problem in an age of ubiquitous sound and multimedia; finally, the smallest identifiable group, or cluster, are those who have vision problems but, even then, there are far more people who have difficulties reading a standard computer screen than those who would be classified as visually impaired.

In some offices the number of people who cannot read Times New Roman 10-point, the default setting for many computers, is more than 15%.

humanITy's (www.humanity.org.uk) estimate is that if you gave everyone in the United Kingdom a free Pentium and an unlimited training voucher, half the population would not be able to use cyberspace, as currently constituted, effectively. This figure will be lower in some EU countries with much better literacy rates and a little higher in others.

So, if we are going to aim for a society of universal access, the approach of basing measures on the special needs of a tiny minority is not going to work. You must also bear in mind that this picture is going to be further intensified by the ageing of the population until at least 2030.

So much for a sketch of the demographics. The economics are also not so straightforward as they are often portrayed. Many people who are classified as "disabled" and many millions more who are not but who have accessibility difficulties are not poor; you only have to consult the growing number of people with solid, occupational pensions. You should also look at figures for disposable income and the fall in real terms of the price of ICT access.

So, in summary, the paradigm of disability we are supposed, particularly by fund-raisers, to consider is paediatric, traumatic, severe and impoverished whereas the real paradigm is geriatric, gradual, mild and with some disposable income.

This does not mean that there are not severely disabled people who require special, niche provision but it does mean that the more provision can be covered through general means the more budget there will be left for severe cases. On that basis, you may wish to contemplate special measures for people who are legally Registered as Disabled People for the purposes of special provisions or benefits but this will not be relevant to the vast majority of people who have ICT access problems.

3. Market Boost or Brake?

Looking back for a moment to that briefing note, the extent to which access needs to be legislated should depend on the extent to which the market cannot handle a consumer need; in other words, the smaller the niche the more likely it is to require legislative intervention. This is true in pure theory but in fact one of our problems is that ICT is still terribly subject to producer 'push' rather than consumer 'pull', so engineers control product development rather than marketing departments. This means that many large companies, ignorant of the demographic pattern I have just described, think that disability access to digital products puts a brake on their development. That may be true if you measure the success of your product according to the time between conception and going to market but, as it has been pointed out (7), the critical time lag is that between conception and profit. If this latter criterion is used then accessibility may well be a factor in boosting market share. In ICT the model most developers have is that of the PC industry and the periodic upgrade whereas the era of ICT we are now entering is much more closely related to consumer electronics and 100% market penetration.

Put simply, the incorrect demographic and economic paradigm adversely affect the ICT industries' understanding of the potential market. This may be short term but it is an acute problem for disabled people. There is no prospect that legal measures of any kind will be taken to deal with this but there is no evidence either that general awareness raising is the answer. It seems that the two measures that can be taken are: first, the encouragement by the EU of Member States to pass legislation about disability discrimination which is as specific as it reasonably can be about the right to information. The UK Disability Discrimination Act has been cited in some EU documents but one of its key weaknesses is its failure to be specific about rights to information, particularly, for example, where this comes up against the conflicting right of the author to withhold copyright permission for Braille transcription. Secondly, against the background of industry ignorance or indifference, the rise of SMEs without extensive human resources or legal departments, and general de-regulation, we need to look carefully at universal standards of service for major suppliers, technical standards incorporated into authoring tools rather than compliance manuals for Web designers and guidance on the implementation of these two sets of standards which become documents of primary reference for the settlement of what meets and fails to meet a standard. Before turning to these topics, however, I want to look very briefly at why accessibility is important and I want to point out that we are not simply looking at the rather rigid, traditional, PC-based approach to access which automatically assumes that the key environments are concerned with education and employment. The demographics of disability, the unemployment rate amongst disabled people and the use of domestic and mobile ICT systems all point to the need for a far more general access framework than that for education and employment.

4. Universal Standard of Service

We have to recognise that the kind of exclusion that people suffer when they do not have access to ICTs is changing. When ICT access was largely PC-based, with participation by 25% of the market, the other 75% were excluded but this was not critical. When the figures swing the other way round and 75% have access and 25% do not, then this critical mass of access creates a norm in social and economic expectations. If, for example, social arrangements are based on the spontaneity of total connectivity, those who are not connected will become more isolated as fixed social functions become less popular. The combination of fewer banks but inaccessible e-banking will become an additional burden for disabled people. At the point where digital transactions replace traditional methods it is time to think about general guarantees.

It follows from all that I have said that the basic building block for any guarantee of access is the establishment of a universal standard of service by major organisations such as telephone companies, broadcasters, banks and retailers. Here are four basic reasons for this approach:

First, the potential market for products is such that a universal standard is not forcing suppliers to work in a market segment that is fundamentally unnatural or distorting
Secondly, those who require niche provision can receive it in the context of general provision instead of being segregated, though this will require public/private partnership in supplying services
Thirdly, a universal standard of service does not discriminate between providers
Finally, a standard can largely be established through licensing conditions so it can be achieved without cost to the public sector.

The last point is critical, as is the observation I made about SMEs. My target is not the student Web designer, nor is it the small company marketing a new piece of home-made software; but I am naturally looking at major banks, retailers and broadcasters who obtain their permission to provide service through various kinds of permits and licences.

This is not so easy as it looks. There are already cases where information accessibility has been legislated but where either the communications channel cannot carry the extra content or where the receiver hardware cannot deliver it. This is why any standard has to refer to the information system not to particular parts of it. There is no point insisting that a bank makes its information accessible in various ways if the Telecoms company it uses cannot or will not deliver the content. In other words, the service requirement must be viewed holistically or you might otherwise be in the unfortunate position, watching television five years from now, where you can access the programmes or the commercials but not both.

5. Web Accessibility and Standard Setting

Having made these rather general observations, let me turn now to a discussion of Web Accessibility, what we can learn from the process so far and how we need to proceed in the future.

If you are going to impose some kind of theoretical standard like, for example, accessibility, there needs to be a set of benchmarks or technical specifications to give meaning to the abstract idea. In the concept of the World Wide Web and Accessibility, this has been achieved through an open, consultative, consensual framework largely drawing on the ICT industries' own expertise. It could be argued that this is rather cumbersome compared with economic and technical developments, that it is too 'light touch' and too often overlooked but of all the possible approaches it is the best for the purpose because only industry-wide solutions will be acceptable and then operable.

Because we are talking about the World Wide Web the standards also need to be uniform. So far the WAI has primarily borne in mind the setting of standards for PC access but with the growth of Web-enabled mobile phones and Web-based television and radio, there will be an enormous temptation for national initiatives and the consequent breakdown of cross border inter-operability. This is a case where subsidiarity is definitely a bad idea.

There was a very critical time when some Member states wanted to establish a quite separate EU standard from the WAI; and even now that has been avoided, there are still some Member States who think they should write WAI variants.

These may in some respects be improvements on the WAI Guidelines themselves but the way to achieve improvement is through the global process not unilateral tampering. As interoperability becomes ever more important, slight tampering will simply make data non interoperable. In other words, any nationalist variants may give a few people a short term advantage with better or clearer guarantees but the more prescriptive these are the more likely they are to produce medium and long-term disadvantages. The EU cannot continually call for pooled sovereignty from Member States where it really matters and then be relaxed about subsidiarity or even about an EU standard diverging from WAI. I have to say again here that this is my own opinion and that of my organisation, humanITy, I am not speaking for the World Wide Web Consortium (www.w3.org).

The story so far, however, points to a severe dilemma. What people really want is a set of technical standards, which tell engineers what they must implement. To a very limited extent the WAI can do this but it does not solve fundamental problems. WAI has, therefore, quite rightly established a set of general principles with illustrative examples. This may make for rather untidy law in some ways but we will have to accept that what we are trying to achieve with the Web, as with other ICT access, is a set of principles which begins something like:

Every citizen has a right to access information on and services from the World Wide Web and other communications networks where such information and/or services have been provided by the public sector and commercial organisations such as utilities, broadcasters and major goods and service suppliers such as banks and retailers except where the relevant authority has specifically reserved an accessibility responsibility (e.g. the supply of specialist access peripheral hardware and software) or where an exemption has been granted because of the size or stage of development of the supplier.

In other words, this makes a general statement about a principle and grants an exemption.

Having got this far, however, it has to be recognised that the terms of the access will have to be agreed on a general basis rather than on a set of technical, engineering specifications. I will go on to explain why this is so but I just want to repeat that the second building block, along with the idea of a minimum, universal standard of service is the establishment of global principles and, where possible, global good practice and applications for Web (and other information systems such as WAP) accessibility. We should, in that context, ensure that Europe:

has a much clearer collective voice
is much more proactive in WAI
ensures that the balance of disability discussions better reflects demographics.

6. Interoperability and Alphabet Soup

Here is a brief list of devices, which can receive digital information, just to give you an idea of the complexity of the task we face:

Devices/Appliances/Receivers: Brief list of information devices to bear in mind when creating information:

Apple Mackintosh
Browsers - desktop, pocket and text-based; Palm, Symbian / epoc; off-line Games consoles
I-Mode phones
Screen readers (for telephone access or visually impaired people)
WAP phones
Web-pads (pen operated)
Web TV
Windows PC.

I want to illustrate the problem with technical specifications versus general principles with a simple example. At the moment we recommend that designers do not make images more than 150 pixels wide for hand-held devices but there is every chance that screens will get smaller and that 150 ixel images may have to be re-designed for, say 135 pixels. In that context there isn't much point in making a hard and fast guideline about pixel numbers. The rule would have to be something like: "Ensure that the pixel specifications for your graphic fit the screen size of the output device". You might then want to go on to refer to a more general principle which says something like: "Because many people cannot see tiny images and because some devices receiving your information do not have a screen, ensure that the text version is complete and does not depend on the graphic".

So far, this principle is very general. You may then wish to go on to specify that the best possible way of achieving your goal is to use XML (extensible Mark-up Language) with a style sheet for each output (receiving) device). You may then go on to say that people may wish to use mark up languages other than XML and variants of it but their output must meet the standard they would achieve if they were to use XML with Style Sheets. So, from a general principle you have gone into some technical detail about how you achieve the output and you have given a technical benchmark, XML and style sheets, against which any other information design technologies have to be judged.

So what we have here is the need for some general principles, which have to deal with the whole alphabet soup of new technologies. There is no point making a set of rules which specifies an output device like a satellite television channel or a games console because you will have to re-write the legislation when the next set of devices comes along; that was the big mistake with copyright, it was medium specific instead of advancing a general principle. As the number of technologies increases you also have an increasing problem with interoperability, legacy systems and updating.

7. Summary

You can now see that this is achievable but it is not at all simple. What I have done is to outline four basic steps:

First, there is the citizenship requirement for access to information and services. This is couched in terms of a general right of access. This right may be in conflict with other rights such as the right of the owner of intellectual property to forbid special access facilities such as text enlargement or transcription.
Secondly, there needs to be a definition of where responsibility lies for ensuring the enjoyment of the general right. In this case it is the provider of the information and services, such as the public sector or the bank, but the responsibility also lies with communications companies and device manufacturers to ensure that the accessibility is integral to the whole system, provider, carrier and receiver. There may be legitimate exemptions here where, for example, a severely disabled person needs an access device, which is provided by the public sector.
Thirdly, there are general principles, which express the way in which you describe enjoyment of the right of access. In my example I used two principles: one was the matching of the size of graphic to the screen size of the output device, the other was ensuring that the text strand does not depend on the graphic for its coherence.
Finally, there are technical benchmarks, which guarantee that the general intentions above can be realised. In this case I specified XML with Style Sheets.

Although it is difficult, this is not a unique position. There is a good parallel set of conditions in respect of accessibility to transport where all four of the above steps can be easily recognised.

What makes ICT so difficult is the changing nature of the technological capacity. When the WAI began its work people were still very fixated on the PC, access to text and the description of static graphics. I also think that there was an unspoken agenda about self- improvement, education and work. It has been very difficult over the years to get regulators and legislators to think about concrete services such as banking, shopping, broadcasting and access to health information. All we get are benchmarks, milestones and the replication of good practice. This is why any legal approach to the content of the Web needs to be much better grounded in the real world. The Communication on the information on public Web sites is encouraging but most of us who use the Web spend most of our time on sites that are not public sector. What is much more to the point, as Webcasting becomes more prevalent we will have to think about the massive entertainment industry as part of the accessibility remit.

8. Conclusion

This is, you will be pleased to know, not as bad as it looks. We already have voice in technology, heuristic middleware, intelligent agents and avatars. We have language engineering and précis and parsing engines. Research is still needed into hybrid systems, which involve computing and human beings. Ironically, the development of such processing devices as RDF will help disabled people at home but reduce their employment opportunities. The mistake we most often make in this rapidly changing world is to mistake current practice for science fiction; we tend to travel into the future with our backs to the engine, seeing the present in terms of the past rather than in terms of the future. For that reason we have to be particularly careful to distinguish the permanent from the transient.

The permanent requirement is for a medium-independent right of access to digital information and services. Theorists would argue that this is a necessary precondition for any discussion of general and technical standards but I think these have to be taken in parallel. Most officials have a much greater taste for rules than for rights so that may be the only place to start but, for all that, the right to information in the information age should not be a matter of social policy but of civil rights. As a citizen and a taxpayer I was deeply insulted at the Lisbon Meeting listening to a discussion on when and under what conditions I might be admitted to participation in e-Europe. I place great faith in the market's ultimate ability to recognise where its economic interest lies and I have great respect for the non- governmental sector but, ultimately, as a citizen, my legal right to information, from the Web as elsewhere, has to rely upon legislation.

9. References.

http://europa.eu.int/information_society/topics/citizens/accessibility/Communication_accessibility_EN.doc
European Commission: Green Paper: Living and Working in the Information Society: People First. European Commission. Brussels, 1996 http://www.hamburg.de/English/StadtPol/Europe/peopl1st.htm
http://europa.eu.int/comm/dgs/employment_social/lisbonconf.htm
http://www.eurorights.org/eudmca/CopyrightDirective.html
http://europa.eu.int/information_society/eeurope/news_library/documents/initiative/initiative_en
http://europa.eu.int/comm/internal_market/en/intprop/news/1100.htm
http://www.stakes.fi/includef/pam4.doc

Return to EU/UK Policy