The workshop^†, organised on the 31st August 1999 as part of INTERACT'99 on behalf of the IFIP WG 13.3, was very successful, with 18 participants coming from different domains: institutes for disability, computing departments, research centres in accessibility, telecommunication companies.

The morning session was devoted to the presentation of position papers whereas discussion and synthesis were achieved in the afternoon.

Position papers were classified according to four topics: (1) Elderly and Disabled Users, (2) Tools, (3) Evaluation and (4) Designers.

In the first group, M. Maguire (HUSAT Research Institute, UK) gave a broad overview of problems encountered by persons with special needs in respect of the new communication telecommunication facilities.

N. Alm (Dundee University, Scotland) and K. Nakamura (Kagawa University, Japan) advanced commercial arguments to convince designers of the need to take all potential users into account in the initial stages of any design process.

T. Verelst (IsdAC International Association, Belgium) explained through a videoconference the position of ISdAC (Information Society disAbilities Challenge) concerning accessibility guidelines. In particular, he defended the idea that people with disabilities must be involved in design and testing processes.

We express our thanks to A. Kilgoure from Heriot-Watt University and his team for providing a videoconference of such high technical quality.

Concerning the tools, different solutions were suggested.

J. Lindenberg (TNO Human Factors Research Institute, Netherlands) presented an engineering tool in the form of an electronic handbook. This tool has the advantage that it can present the information at the right time and hide irrelevant information.

A. de Baenst (University of Namur, Belgium) submitted the outcome of research sponsored by the Walloon Region in Belgium. It concerns the design and development of a Web site with methodology and guidelines for accessibility, intended for the designers and developers team of Internet services.

C. Nicolle (HUSAT Research Institute, UK) and J. Abascal (University of the Basque Country, Spain) explained their project for a source book for accessibility guidelines.

Finally, C. Jarrett (Effortmark Limited, UK) presented problems on interpreting guidelines for font size, interlines, type, etc. She is writing a ‘How-to’ manual on forms design.

In the category of evaluation, Ch. Colwell (University of Hertfordshire, UK) reported two experiments, which had been carried out, the first to investigate whether the WCA Guidelines developed by the WAI can be used effectively by Web page authors and the second to examine whether pages developed using the WCA Guidelines are accessible to visually impaired persons.

To close the session devoted to position papers, M. Tucker-Kinney (Nokia Mobile Phone, USA), from Nokia Mobile Phones, explained the design organisational process inside a big company and how guidelines are integrated in this process.

The general discussion started with the position of F. Windberg’s from the Royal Institute of Technology in Sweden on the necessity for a methodology using guidelines, and more than just guidelines. Other partners made reference to UserFit project. But the guidelines in the book were judged difficult to be used by non-specialists.

The following questions were then discussed:

• Do we have to whittle down guidelines, at the risk of a loss of richness?
• How can guidelines be incorporated in the design process inside an organisation? How can the changes be managed?
• Which tools are useful and usable? What tools for whom?
• How can elderly and disabled people be involved in the design process?
• How should cultural issues be dealt with?

To sum up the workshop, participants agreed on the diagram in Figure 1 for the poster presented to the conference.

There are a number of arguments which can be advanced to convince designers to take all potential users into account at the initial stages of any design process. In the US and to some extent in Europe, legislation requires that people with disabilities are not marginalised. This kind of legislation embodies the social and ethical values that all people matter, but also practical issues such as ensuring that a society benefits fully from the contributions which all of its citizens can make.

An economic argument for equal access would have the advantage of not depending on cumbersome enforcement procedures. It would in effect be self-enforcing, being driven by the forces of the marketplace. Thus far it has proved difficult to produce convincing economic arguments to back up the worthwhile goal of design for all. The proportion of people classified with disabilities remains small. One of the economic effects of this is that providing the marketplace with assistive technology is in the hands almost entirely of small to medium size enterprises. It is possible that the increasing proportion of elderly people in many countries will produce a gradual shift in this situation. Once a fairly large number of people with money to spend are affected by excluding environments and the lack of assistive technology, change may begin to happen.

Another economic argument is that design for people with disabilities can be considered in a number of ways as design for the rest of us.

Discussing any population in terms of those who are disabled and those who are not is misguided. Each human being has a range of cognitive and physical abilities. We all represent different points on a continuum, in fact, points in a multidimensional space representing all the possible ranges of human abilities. All people, as they grow older, move through a personal continuum of abilities and needs, as has been said. It is also true that all of us can be temporarily disabled from a variety of causes. Sickness, accident, tiredness, drunkenness, stress, panic, high workload, restrictive protective clothing, and many other factors can cause temporary disability. Environments can themselves be disabling, causing people to move within this continuum. An extreme example is a soldier in the middle of battle. A more everyday occurrence is the potentially dangerous visual impairment suffered by the driver who is backing up a large vehicle from within a cab with limited visibility.

Thus we are all potential beneficiaries of a policy of 'design for all', and we are all potential customers for differing versions of assistive technology. Such a wide market base should not be ignored by designers, manufacturers, and policy makers.

Another way in which all can benefit from technology designed to help people with disabilities results from the fact that we increase our understanding of processes by examining cases at the extremes. This is how medical science has advanced (and particle physics, for that matter). Consideration of the needs of 'extra-ordinary' as well as 'ordinary' users can highlight design problems that would not have been obvious. For instance, the introduction of kerb cuts to enable wheelchair users to get around cities more easily has also been of benefit to pushers of baby-strollers and delivery trolleys.

Design practices which consider users with a range of needs will lead to better design for everyone, and provide expanded markets for a wide range of goods and services.

Information technology is continually developing and in the last few years there has been a rapid growth in electronic telecommunications to provide Internet and other network-based services. Interest in using telecommunications to provide services to the public is growing, with a number of pilot services being set up across the country to explore market potential and/or stimulate demand.

Administrations across Europe are now using telecommunications technology to provide citizens with information (Hoare, 1998). The British Government, for instance, has issued a directive that 25% of all civil service communications must be on-line by 2002. This is intended to provide savings in paperwork and a streamlined service.

An area of possible government assistance is to provide the public with on-line information about welfare benefit entitlement. This may have benefits for all members of society but could be of particular value for retired and older members of the public, many of whom do not always claim their entitlement. It is estimated in the UK, for example, that one million pensioners could be entitled to Income Support that they are not claiming (Benefits Agency, 1998). To provide information to the public, the Benefits Agency produces a website of over 1,000 pages for customers to browse through which can be viewed at http://www.dss.gov.uk/ba. Yet while many people are now connected to the Internet, and the EU average is 25 computers connected per thousand people, (Lennon, 1999) the number of older or retired people using it is still quite small. An indication of this is given by the results of the worldwide 10th Georgia Tech web survey (GVU, 1999). The survey, conducted in 1998, received 5,022 responses of which only 2.7% were from people who were 66 years of age and over.

This raises the key question of whether older people will be able to benefit from the promise of the connected future, with information available electronically on tap, or whether they will get left behind. This also represents a lost opportunity for suppliers as, according to Oftel (the UK telephone watchdog organisation) and disability campaign groups, "A growing grey market containing millions of potential customers is being ignored in the telecoms boom", Dawe (1998). This paper, therefore, discusses a number of issues relating to making telecommunication services accessible to the population as a whole.

When considering different ways of accessing public service information, various methods are possible. However for older people to make use of them, a number of barriers exist which must be overcome, as discussed below:

Few older people own a PC and may not be motivated to acquire one to gain information that may be of benefit to them. Connection of a PC to the Internet is fraught with difficulties that can defeat anyone who is not a technical expert. It was reported recently in a survey commissioned by COMPAQ, that nearly a quarter (23%) of respondents said that their work was interrupted daily due to computer crashes and other IT faults, which frequently causes computer rage. The ability to access information via the Internet (e.g. the World Wide Web) requires background knowledge and experience in order to navigate successfully through a series of web pages. Thus although many organisations and companies provide interactive web sites for purchasing items, or to help users calculate their benefit entitlement, the complexity of the interactive process would create many difficulties for inexperienced users without some level of basic knowledge of computer and Internet use. Web access for people with disabilities also causes problems although design guidelines for making web pages accessible are available such as the TRACE guidelines (TRACE, 1999). One technique that has also been developed at Oxford Brookes University is a speech synthesis system that can summarise Web pages by reading out headings, links and text paragraphs (Zajicek, 1999).

The next step in popularising the Internet is to develop web and E-mail access via the Television set. Digital TV companies are planning to offer this as part of their future digital TV interactive services and are developing browsers to support it. Some people see the World-Wide-Web browser providing the main interface with the user in the future, not only to the Internet but to other electronic household devices such as the video recorder (Price, 1999). Again, older or retired people may be reluctant to take up digital TV services which are often seen as unnecessary, and which require extra payments for the service, possibly a phone line connection and purchase of a set-top box. However currently in the UK free competition between digital TV service providers is strong and set top boxes are available free from Sky and ON-digital. To access the Web, the user will interact via the TV handset or a laptop keypad. This will require users to become familiar with indirect selection of items on screen with a selection bar, rather than direct pointing with a mouse. Operations such as scrolling are difficult via an indirect pointing device while the restructuring of a website so that scrolling is not required, may require a greater depth of structure and increased complexity. Other problems that may be faced are that text which is readable close up on a PC screen may be too small for viewing from an armchair at a typical viewing distance (10 to 15 feet).

The Henley Centre has completed a survey entitled Teleculture - The Citizen Speaks which showed that the public would prefer greater telephone accessibility to government services (Vernon, 1999). In order to be more cost effective in providing telephone access to support their customers, organisations are making more use of automatic speech recognition and speech synthesis interfaces, and menus operated by speech or by the telephone keypad, to guide the user to the correct person or information item. Customer records can be centralised and by channelling all enquiries through a central service, fewer staff are needed. Yet, in the eyes of the public, these automatic telephone services are often seen as remote, slow to access and de-personalised. A future concept is the use of intelligent agent technology to support a customer. The agent or assistant will elicit the users enquiry and will act autonomously on their behalf to perform the task or retrieve the information the user requires. This will allow the customer to interact with the service centre in an efficient manner. Yet again there are barriers to using such systems. People get frustrated navigating through several menus, many older people do not have tone phones to make menu selections, and the difficulty is often the first step of who to ring up to get the information or help they require.

Public kiosks have been present in cities and towns for many years, providing information and services, for example, about local amenities, public transport and services and events, as well as bank cash machines. Libraries as well as Internet cafes are also now providing PC access to the Internet which is an effective means for those without access from home or work to use the Internet on a casual, relatively inexpensive basis. Yet these services still tend to be used by young people, more confident about approaching an interactive terminal and working their way through it without any introduction or training. While some excellent guidelines exist for the design of accessible kiosks, (e.g. Gill, 1997) many kiosks fail to implement basic facilities for people with impairments such as large clearly labelled buttons, large clear lettering on the screen, or being placed at a suitable height for wheelchair access with a recess underneath (although banks now often do this). Further general advice on the design of kiosk based dialogues is given by Maguire (1999). However the reasons for non use of kiosks by older people are often basic and practical i.e. lack of mobility to reach the kiosk, being unsure about the procedure, concern that they will forget their PIN and lose their card, worry about being attacked and robbed etc.

Future research needs to be carried out to consider a range of delivery mechanisms for telecommunications services including public kiosk systems, terminals located within public buildings and Internet access from home. New approaches may also be considered such as loaning out a PC with a CD-ROM containing relevant information. Scenarios should be identified where benefits can be derived from particular delivery mechanisms e.g. home shopping for people who have mobility problems, providing welfare information in areas of social deprivation, accommodation information for tourists etc. For such scenarios, barriers may be identified and prototype applications can be developed to demonstrate how such problems can be addressed.

Specific questions that may be addressed include:

• What kinds of information can be delivered electronically?
• How should information be delivered in an acceptable way to a wide range of members of the public, particularly people who are older or disabled?
• How can such systems be made attractive, usable and consistent across applications and countries?
• What are the cost/benefits of such systems?
• How can such systems be developed to provide good performance and are reliable?
• What common user interface conventions across European countries can be agreed?

With the growing dependence on advanced telecommunications facilities to deliver information, unless these kinds of issues are addressed so that the facilities are accessible, then there is a danger that large sections of society may simply miss out on them.

I would like to thank my colleague at HUSAT, Colette Nicolle, for reviewing and commenting on this paper.

We are in the middle of a revolution similar to that of the industrial revolution of the 19^th century, the Information Society. This revolution is affecting the way we work, the way we study, the way we shop, and many other aspects of our daily life. Concepts such as teleworking, e–commerce, co-operative work, tele–cottage or tele–training are becoming part of our regular vocabulary.

The birth of this new society has many implications for everybody. It will open new opportunities, but it will close many others. Information is the keyword. Society might be split into two sectors: those who are "wired" and those who will stay "un–wired." Although while you are an active participant in the events, it is difficult to analyze the facts happening in your world with some perspective, we feel that people with disabilities cannot afford to stay out of the loop.

The concern of many people on these issues was the seed of ISdAC International Association. It was obvious that we had to inform our constituency about these changes. Conventional channels were no longer effective, and new approaches had to be found. In the middle of the fight, we must be aware of many accessibility issues. We realized that we had a twofold target: not only we had to reach our constituency, but we had to deal with content providers and software developers as well. We knew that we shall play the role of a bridge between the designers and a market sector which influence will not be negligible.

This presentation will not focus in a particular technical issue, but will present some snap-shots of our experience in relation with the topics of this workshop: Web design and software development.

The first challenge faced is the definition of Design For All or Accessible Design. There are many definitions circulating around, it is a widely discussed topic and we are sure that the audience is familiar with many of them [1, 2, 3, 4].

We think that the concept Design For All is misleading depending on what type of audience you are addressing. Generally, it gives to the designers the idea of a highly sophisticated and not cost–effective design process by which their products will reach a wider hypothetical market which strength has not been shown to them.

In the field of Web design and software development, our experience shows that outreach and education benefit when the message is simplified. The message to be sent is that we can solve simple problems with simple solutions. Of course, neither some of the problems nor of the solutions are that simple, but the designer will be overwhelmed if he/she must tackle two hundred issues simultaneously.

In previous papers [5, 6], it was exposed as well that, within this design process, designers can benefit from interaction with end–users whenever a group of those with the appropriate knowledge–base are ready to co–operate with or even drive the industry. As a short remark, we have found out during the latest years that many EU organizations in the sector –with the exception of a few well–known examples– did not have experience nor even interest in technical projects. Thus these organizations play a passive role in R&D projects, which are more often than not a technology–push instead of a response to a user need. We must stress that organizations with a stronger focus on accessibility issues are needed to help to send the message around.

This section will describe a training experience with a group of Web designers working for one of the biggest telephone companies in Spain. This group is used to design websites of great size and complexity, and they were "forced" to learn about web accessibility because of a disability– related Website.

The first issue was to figure out whether they were familiar with access issues to computers. Many of them did not even know that a person with visual impairments could access a computer, not to mention the internet. Therefore, they were exposed to a session where they were shown different types of disabilities, and how different access problems to computers are tackled. In particular, they found out about:

This introduction gave them some background on the type of problems people with disabilities face whenever accessing a computer. It helped to put a face behind the problem.

They browsed the net with these tools, and in particular, they examined their own designs. This gave them an idea about the consequences of ignoring accessibility guidelines for people with disabilities.

The following step was to show them how simple things such as adding ALT tags to images, including d–links or providing alternatives to scripts could improve accessibility dramatically. This gave the trainer the opportunity to discard the myth of cost–effectiveness for accessibility. An intensive training followed on how to implement the Web Content Accessibility Guidelines 1.0. They learnt –as remarked in the Guidelines– that following them will also make web content more available to all users, whatever user agent they are using (for example, they were not aware of the possibilities for mobile phone access) or constraints they may be operating under. They learnt as well that we were not discouraging content developers from using multimedia, but rather explain how to make multimedia content more accessible to a wider audience.

Shortly afterwards, this group of people began their design, and the results of the training were excellent. 80% of the designed pages conform to the Guidelines to level–A, and the rest required only simple changes. Even a level–AA was not difficult to acquire after a revision.

This experience was very positive, and despite it was not clear whether the management of the company will implement these procedures in other design projects, it will sure influence the future work of the designers. From this experience several conclusions were drawn.

1. Authoring tools are not very helpful when designing accessible websites

2. Training costs and development costs

3. The market

4. An Europeans with Disabilities Act?

Recently, ISdAC and Lernout & Hauspie, a leader company of speech and language products, technologies and services, including speech recognition, text to speech, compression, and transla-tion, initiated a collaboration. During the latest months, a group of ISdAC members collaborated with the company testing L&H VoiceXpress speech recognition software in Dutch and English.

Although this is a pilot under a non–disclosure agreement, we can mention that it has been a productive collaboration with a positive response to our feedback. This experience makes us think that it is a way to explore further with other software companies.

We will not like to finish this section on software development without mentioning the advance offered by Java^TM to develop accessible software. Java is a growing object–oriented language with applications not only in PC or workstations, but in many other devices, from mobile phones to control environment devices. Its portability together with its versatility can ensure a promising future for this technology.

From our point of view, the Sun Accessibility team has done an excellent job, facilitating to the designer the possibility to incorporate Assistive Technology input and output in the software. In fact, version 1.2.2 of the Java Development Kit released a few weeks ago, contains the Java Accessibility API, the Swing user–interface classes, and support for loading Assistive Technologies into the Java Virtual Machine. They developed as well an Access Bridge to the API of the Microsoft Windows operating system, to be incorporated in AT products. We cannot forget the efforts of the IBM Special Needs team who made available the IBM Guidelines for Writing Accessible Applications Using 100% Pure Java.

Again, these advances will not be present in commercial products unless awareness and training programs go hand–in–hand, and Authoring Tools implement these changes. Efforts within these lines must be pursued.

The information is out there. Thousands of gigabytes are available, but we know that they are not available for everyone. ISdAC International Association is willing to share the experience of its members with anyone interested to ensure the achievement of fully accessible products for greater social and economic benefit. Our aim is to ensure the highest levels of accessibility for maximal EU citizen inclusion.

The countries of the European Union generally want to enhance self-support for the elderly and disabled and consequently strive for accessibility of 'mainstream' Information and Communication Technology (ICT) for these inhabitants. The 'universal accessibility' approach is advocated as a possibility to include all user groups in the software development process.

The European Union supports research in this area, for example by funding the project ‘Assessing the Awareness of the Design for All Approach’ (AADAA consortium, 1998).

However, the 'universal accessibility' approach is not as widely spread in the design community, as it should be. Several reasons have been identified. First, the phenomenon that 'universal accessibility' products are perceived as dull by the designers creates serious problems for the acceptance of the 'universal accessibility' paradigm. This might, however, be a self-fulfilling prophecy. Second, the cost of involving elderly and disabled users in software development can be higher than for involving the rest of the population. On the other hand, a broader perspective comprises also the advantages of the 'universal accessibility' approach: The involvement of elderly in the design process often results in better products for all kind of users so that the overall utility of this approach is high (Gleiss et al., 1995). Furthermore, the number of potential users of the 'universal accessible' product also increases so it might very well be a good investment. Third, the extra constraints on design possibilities are another objection. However, it could be that the constraints force the designers to be creative, bringing about new ideas that enrich the total design space (Newell & Gregor, 1997). In sum, the three problems of dullness, design costs and design constraints can be solved by evoking positive attitudes, widened cost-benefit analyses and creative design solutions. However, to realise 'universal accessibility', the corresponding human factors knowledge should be incorporated into the design practice. Currently, designers take insufficiently account of human-factors knowledge on elderly and disabled (Neerincx, 1998). This paper provides a cognitive engineering approach to solve this problem.

Although there is a lot of human-factors knowledge on elderly and disabled (Craik & Salthouse 1992; Fisk & Rogers, 1997; Vanderheiden, 1997) current knowledge is insufficient to realise 'universal accessibility'. There is a vast variety of different guidelines but the status and origin of these guidelines is often unclear. Sometimes guidelines of different origin even contradict each other. Besides, in spite of the large amount of guidelines there are still blind spots in essential areas. It is clear that the human-factors foundations should be improved. Practical theory and empirical research is needed to create a complete and sound set of guidelines. A problem of current guidelines is that it is hard to translate the guidelines to specific user interface characteristics. How should a designer, for example, implement the guideline ‘minimise demands on spatial memory’ (Czaja, 1997)? A more supportive guideline would be for example: 'chunk information in a Web-site according to the user task hierarchy to minimise demands on spatial memory in order to improve navigation performance' (based on Miller, 1956; de Groot, 1966). To be able to provide such supportive guidelines practical theories are needed that comprise the mutual effects of user interfaces, human cognitive processes and HCI outcomes (Figure 1).

On the basis of a practical theory concepts for user interfaces can be developed e.g. for chunking of information. With these interface concepts the theory and corresponding guidelines can be tested. In the course of one of our projects we developed a support function for Web-based services which ‘chunks’ the information provided by a Web-based service based on human information processing theory. Currently we are testing the support function (Neerincx et al., 1999).

When a complete and sound foundation of human-factors knowledge has been established it is important to maintain and improve these foundations by adding design and use experiences. Once the foundations are laid it is important to focus on another issue. How do we make the acquired knowledge accessible to the designers?

The current body of human-factors knowledge is insufficiently addressed. One of the reasons is the vast variety of guidelines. A search on the Internet for guidelines applicable to web-design for example provides the designer with hundreds of hits. This problem will be addressed by creating the human-factors foundations mentioned in the previous section. The current lack of clear examples will also be taken care of in the foundations. Each guideline should be accompanied by a prototypical example of how the demand can be satisfied. In our view two important issues remain:

• the mismatch between guidelines and design procedures
• the designers lack of experience in the use of the guidelines

These issues indicate a need for techniques to address guidelines in specifications and assessment. Our proposal is to develop a universal accessibility engineering tool, which incorporates the guidelines, examples and experiences in a usability engineering framework. This framework will be used to indicate when and how a specific guideline should be applied.

Software development is an iterative process in which the artefact is specified in detail more-and-more and specifications are assessed more-or-less regularly to refine the specification, to test it, and to adjust or extend it. Cognitive engineering methods should provide guidance in this process in order to optimise usability. First, based on the users' goals and information needs, the system's functions and information provision are specified (i.e. the user interface task level). Subsequently, the control of the functions and the presentation (i.e. the user interface communication level) of the information is specified (Neerincx, 1998).

A cognitive engineering tool should be constructed to support such a human-factors contribution, for example in the form of an electronic handbook. The handbook should present the right guidelines and techniques at the right time. Clear examples and relevant experiences should accompany the guidelines. Specification and assessment techniques should also be provided by the electronic handbook. It should be easy to attune the handbook to a specific domain and it should provide an adaptive interface that enables hiding irrelevant information. Flensholt et al. (1999) provide an example of an electronic usability handbook that is based on these usability engineering principles for payload interfaces for space laboratories. In a similar way, a handbook will be developed to support the design of 'universal accessible' software.

The dull image, the economic cost and the constraints of 'universal accessibility' are problems that need to be resolved. We believe, however, that by improving the foundations and accessibility of 'universal accessibility' principles an important step can be taken in promoting it to designers. By offering them an electronic handbook with applicable guidelines, specific examples, specification techniques, assessment techniques and experiences, the foundations and accessibility of human-factors knowledge can be greatly improved. A sketch of a support tool for the 'universal accessibility' approach was presented at the workshop.

Position paper for the IFIP TC.13 INTERACT'99 Workshop: Making Designers Aware of Existing Guidelines for Accessibility

The access of people with disabilities to the New Technologies of Information and Communication (NTIC) is crucial on many grounds. Society is evolving in a direction where these technologies are part of everyday life and a two-speed society is not acceptable. At the same time, such technologies offer a substantial help in the process of improving the way of life of people with disabilities. In Belgium, the welfare department of the Walloon Region has considered the question important enough to encourage practical implementations and research in that direction.

It is under these auspices that the two teams are carrying out joint research into the design and the development of a methodological Web site with guidelines for accessibility intended for the designers and developers of Internet services. The purpose of this Web site is to encourage access to the Internet services for the majority and especially for people with disabilities and aged persons.

At the time of the INTERACT'99 conference, research is still going on and the Web site is in its construction phase. The workshop offers an opportunity to discuss with an expert audience some of the options taken up so far in the design process.

The first step in the research consisted of an analysis of the existing material in respect of ergonomical guidelines for accessibility on the part of disabled people, access conditions for them to Internet as well as the way to approach the issue of disability in the case of the services offered through Internet. The number of references in the field is vast as can be seen, for example, by looking at the following on-line lists of links [1] [2] [3] [4]. This study draw the attention to the problem of deciding which information has to be kept and what is the most efficient way of presenting it. It has also shown that the notion of "ergonomical guideline" is vague and ambiguous in the computer literature.

At the same time, a panel including developers was organised in order to note the way they are using guidelines and their expectations in this field. We can say that they were in favour of a tool that is easy to utilize.

The notion of disability has also been studied in the context of Internet. This speaks for a functional approach allowing aged and inexperienced users also to be taken into account. Economically, this appears to be a very worthwhile option. Indeed, in this way, the people presenting disabilities appear as a more valuable test population for improving the future Internet services than a marginal group of users.

As a result of the background study, we highlighted two main conditions for the methodology of the Web site:

• On the one hand, the information has to be presented not only in an attractive way but also in a usable way. This pleads in favour of a system with high interactivity and with various entry accesses.
• On the other hand, the information has to be understandable and approachable for people who are not necessarily experts in ergonomics or in the problems of disabled persons. This indicates that the system needs to be constructed that avoids over-information and that is adapted to the target audience of ordinary Internet designers or developers.

In order to meet the two above-mentioned conditions, it was decided to develop the Web site around a base of approximately 100 new consolidated guidelines implementing the knowledge and the principles of ergonomic science [5] [6]. The guidelines will be divided into three levels: the principles, the general recommendations and the specific recommendations concerning an accurate computer object such as equipment, telecommunications networks, software or Web design. There will also be some general information on disability, ergonomy and the technologies of information and communication.

The guidelines will be presented as a set of forms with a unique structure consisting of (1) the level of the recommendation, (2) a title, (3) a short description, (4) a more explicit statement, (5) possibly a connection with ergonomy, (6) an illustration, (7) the set of problems concerned by the recommendation with, if possible, a priority scale, (8) the place in the development process, (9) an optimal implementation, (10) possibly some complementary information, (11) possibly good or bad examples, (12) links to similar guidelines inside the Web site, (13) key-words and finally (14) references links.

To facilitate the access to the guidelines, various entry points will be provided allowing, for each guideline, access by impairment, by function, by contextual constraint, by the extent of the handicap in the aimed application, by the place in the development process, by the nature of the computing problem and by other approaches if necessary.

On the basis of a preliminary prototype, the interface presentation is being studied with the help of a graphics designer.

Among the originalities of the proposed approach, one can highlight: a self-consistent Web site, guidelines formulated following ergonomic criteria and principles, the use of the repetition principle in the formulation of the guidelines allowing understanding and verification of comprehension through multiple and detailed meaning description of each guideline, access by multiple keys (nature of the deficiency, computing object, integration stage in the application development, etc.), a presentation using the interactive possibilities offered by the Web site tool and the additional advantage of being written in the language of the region (French language).

We think that these points are fundamental if we are to meet the objective of making the design and development teams of Internet services aware of the accessibility problem. Only then will there be a chance of making the Internet services more accessible to the majority and especially to people with disabilities and aged people.

The paper presents some of the conflicting advice offered to designers on one particular topic in accessibility: the choice of font size for visually impaired people. This creates practical difficulties for designers when trying to apply the advice.

I am researching best practice in forms design for the Inland Revenue Forms Unit. The Inland Revenue is actively seeking to comply with both the letter and the spirit of the Disability Discrimination Act. A key issue in forms design is meeting the needs of people with a visual impairment. I had a view that at least one factor to consider is font size, so I set out to find out as much as possible on the topic.

The obvious source was the United Kingdom’s best known charitable organisation for blind and visually impaired people: the Royal National Institute for the Blind (RNIB).

The RNIB publishes a campaigning leaflet ‘See it Right’ and accompanying guidelines, called ‘Clear Print’. These guidelines have many advantages for the designer:

• They are published by a leading organisation, well known in the field and highly reputable.
• They are based on both academic research and consultation with people who have visual impairments.
• They are themselves accessible: clearly written and presented.
• The advice is straightforward and easy to understand.

However, RNIB does not represent just completely blind people, but has to consider a wide variety of visual abilities and impairments. Its guidance reflects their disparate needs:

"For the general reader type sizes between 8 point – 10 point (this means the height of a letter x is around 1mm – 1.5 mm) are frequently used. These print sizes are not legible enough for many readers, including, of course, blind and partially sighted people.

RNIB’s own aim is to produce documents for general readers using 12 point (to give an x height of approximately 2mm) and this is the size of print to which we believe others should also aspire. …

RNIB’s research has shown that a significant proportion of blind and partially signed people can read large print. RNIB recommends 14 point as the minimum print size for material intended for blind and partially sighted readers. However, RNIB sometimes uses 16 point when producing information for blind and partially sighted readers as, in our experience, many blind and partially sighted people need a typesize larger than 14 point.

It is difficult to be prescriptive in this area, as factors such as typeface and type weight will be relevant to any decision on type size. There appear to be no advantaged in enlarging type above about 20 point, though larger sizes may be necessary for headings." [1]

For the form designer faced with a form currently in 8.5 point, this leaves many questions unanswered. For example:

• How many people will find 9 point difficult? What about 10 point?
• How many partially sighted people are excluded by 12 point?
• Why does the RNIB not use 14 point itself, if it considers 14 point is the minimum?
• How many blind and partially sighted people need 18 point or higher?
• Why is there no advantage in increasing about 20 point?
• Why is there an advantage in a font over 20 point for headings alone? And if so, how big should the headings be?

The final sentence, and the other factors quoted in the guidelines, alerted me to the importance of considering more than simply the size of type. Also, it was apparent that simply increasing font size introduces other accessibility problems. One form I investigated was originally standard A4 size (29.7mm x 21mm). Enlarging it so that all print was at 18 point or higher meant that the form became A2 size (59.4mm x 42mm). A form the size of a flip-chart poses insuperable problems for handling.

Faced with the complexity of this guideline, I started to research the factors that affect legibility – apart from size.

Even a short survey of some literature showed me that the only factor that is consistently listed as important is size.

Here is a table of factors listed by different authors. I could equally have chosen others.

Other factors I have seen mentioned as having an effect on legibility and/or accessibility for people with or without a visual impairment:

Gregory alerted me to appropriateness to audience:

"When designing for large print you should try to avoid making the document look as if it’s aimed at children – the ‘Janet and John’ design style." [6]

This plethora of factors also made me wonder about whether the research on which they are based is applicable to forms. Hartley, for example, warns us to be careful in our interpretation of this research:

"Taking a critical look at the validity and representativeness of the results, I think there is a good case for believing that they apply to the reading of continuous text but not necessarily to other special reading situations, such as reading signs and labels, maps and diagrams, or using "search" techniques" [5]

and Love had similar problems to mine:

"Unfortunately, much of the research in these areas is not very helpful to designers of instructional materials. This is principally because variables such as type-size, line-length and inter-line space have not been studied in the ‘real-life’ context of instructional text. Most researchers, for example, have considered issues of type-size in short, simple settings of continuous prose. Furthermore, they have usually used ‘justified’ text. " [4]

I also discovered that there is no one definition of visual impairment. Nor does the term 'visual impairment’ tell you how much someone can read.

"There are statutory or legal definitions of blindness and partial sight, but the mere statement of visual acuity – which is the basis of most attempts at defining visual handicaps – often gives a misleading idea of the a person’s functional acuity – that is the amount of eye-sight he actually makes use of" [5]

Because I had difficulty obtaining a coherent and defensible set of guidelines, I tried a different approach. I looked for models of the reading process, and tried to produce a model that the designers could relate to the choices they have to make. So far, this has not been successful. The designers want something that can be applied quickly and simply. They do not have time to explain the difference between visual and functional acuity if their font decision is challenged by a domain specialist in a meeting.

The straightforward advice in the RNIB’s guidelines is the first step, as it gives them some immediate rules to follow. What we need next is something that will help us decide what to do when we have to make compromises in the design, for example choosing between font size and paper size.

There is not yet a conclusion, but work in progress. The traditional usability issues of users, task and environment are as important here as in any other usability issue. These issues apply both for choice of font, where we need to understand the needs of the users when considering our choices.

But equally, they apply to the guidelines themselves: who will use them, for what task, and under what circumstances?

The situation today with the existing guidelines for making new technology and information accessible (for example World Wide Web Consortium, 1999) reminds much about the state of HCI or usability engineering in the late 1980. Jacob Nielsen speaks about something he calls "the intimidation barrier" (Nielsen, 1994a, 1994b). This barrier makes most people refrain from using any usability method since most methods require a lot of knowledge, are expensive, and are too complex.

Nielsen developed a method, or a set of techniques, that he called discount usability engineering (Nielsen 1994a, 1994b). The basis of this method could be stated as "don’t aim at perfection, rather the best one can do" or "it’s better to do a little than not doing anything at all". The foundation of this method is the four techniques User and task observation, Scenarios, Simplified thinking aloud, and Heuristic evaluation.

User and task observation is simply that you have an early focus on the users and observe the users while they are working and observe the user and the tasks.

A scenario is a cheap kind of prototyping where one simulates the interface with for example a paper mock-up while following a planned path of action.

Simplified thinking aloud is a variant of the more complex thinking aloud protocol that many psychologists use. This method requires the user to verbalize his or her thoughts while performing a task, to think aloud. The difference is simply that anyone could do this and that data analysis could be done with just notes taken during the session. No video recording equipment is needed.

Heuristic evaluation is a set of 10 usability principles that are used instead of guidelines with hundreds of rules. This work actually requires a usability expert, but different studies have showed that even a novice on these principles can use them and get good results, especially if more than one evaluator is used (Nielsen, 1994b, pp. 32-35).

For a good example of what I am talking about, one could compare with the somewhat classical collection of guidelines assembled by Smith and Mosier (1986). This looks very much like how the accessibility guidelines look today. Of course, one cannot expect to be able to sum up a complex set of accessibility guidelines in ten neat rules of thumb, but these rules of thumb could be useful for the designers and actually be used. The result might not be as good as if they had used the full guidelines, but would definitely be better than having no guidelines at all.

My proposal is that this approach is something that should be guiding in future work in this area. I don’t think that it is fruitful to believe that the extensive guidelines that exist today will have more impact than the ones that was formulated for general usability issues do. If one wants to influence the designers, the material that one uses must be suited for the designers and not for accessibility experts. Large collections of guidelines are not the answer.

Another misconception seems to be that to use guidelines is to apply a method. To discuss this issue one must first look into what a method is. Olson and Moran (1996) summarizes a method as "[…] a systematic, repeatable way to design" and as something that includes (1) a statement of the problem that the method addresses, (2) a device (a tool, technique or model), (3) a procedure for using the device, and (4) a result, or rather a statement of the nature of the result. According to this definition and to the intuitive image one might have of a method, just using guidelines doesn’t qualify as a method at all. What we need is a framework for using these guidelines, if we are to use them at all that is. Perhaps the guidelines should rather be used as a foundation for new methods and as a good way of validating their performance.

A very effective way of understanding the users and the context in which they are using the technology is to pretend to be that user. When working with general usability questions it can be virtually impossible to pretend to be a novice user, not the expert that one is. But when it comes to accessibility this is sometimes much easier. I am not saying that it is possible to fully understand by doing this, but rather to get a notion about what it’s like and some of the implications. One could use the text based Web browser lynx to browse the Web, turn of the loudspeakers and try to use some multimedia application, disable custom colors when browsing the Web, try to use just one hand or just one index finger when using a word processor, or use a screen magnifying application when browsing the list of files and folders on the hard drive, just to mention a couple of examples. The power of "seeing is believing" is more striking than one hundred pages of guidelines.

How should this be accomplished? In the same manner as the discount usability engineering was once formulated, by extensive research on discount methods in comparison with the more formal and complex methods and guidelines on real cases of technology development.

This study evaluates the Web Content Accessibility (WCA) Guidelines, developed by the Web Accessibility Initiative (WAI). The first aim of this study was to investigate whether the Guidelines can be used effectively by Web page authors. To this end students of HTML (HyperText Mark-up Language) used the Guidelines to adapt Web pages. The second aim of the study was to examine whether pages developed using the Guidelines are accessible to visually impaired people. In order to achieve this visually impaired Web users assessed the accessibility of the pages developed by the Web page authors. The study found some difficulties in the use of the Guidelines that were similar to those identified by previous research. There were unexpected results regarding the accessibility of certain HTML elements.

The WCA Guidelines [6] are available on the Web and have been developed by the WAI, which is a program of the of the World Wide Web Consortium (W3C). Two other online documents are associated with the Guidelines, a Techniques document that contains examples of HTML code, and a Checklist, which can be used to evaluate Web pages. There are hypertext links between the three documents and to other documents on the Web, such as the HTML 4.0 specification.

Four aspects of the WCA Guidelines were identified which required evaluation to discover whether the Guidelines can be used effectively and efficiently. The current study investigated whether: (i) the organisation and presentation allows the discovery of the required advice; (ii) the content provides sufficient and relevant advice; (iii) the relevant advice can be identified, and applied appropriately; and whether (iv) the interface to the guidelines is usable.

The current study is set within the context of previous investigations of the use of guidelines. Such research has shown that there are limitations in their use, for example, in their selection [4]; in their practical application , and in their interpretation [1]. This study has found that these limitations also apply to the WCA Guidelines.

The use of the Guidelines was investigated using a group of students who were learning HTML. The 12 participants all had experience of creating Web pages, which ranged from having made a few personal pages to having been employed to create larger sites.

Prior to the experiment the participants were asked to read the Guidelines in order to familiarise themselves with the document. The task for each participant was to adapt an existing Web page, making it accessible by following the advice contained in the Guidelines. The following HTML elements were adapted by the participants: images, imagemaps, forms, frames, and tables. Each participant was given an HTML file that contained the element(s) they were to work with.

Participants attended the laboratory where they were observed while performing the task. The experimental method used was an adaptation of the 'think aloud protocol' [2]. This involved the participant telling the researcher what they were doing while they were performing the task, and describing any problems they encountered. Video recordings were made of the dialogue between the researcher and the participants, and of the computer screen at which the participants worked. The observations and the recordings provided useful insights into the experiences of participants, the problems they encountered, and the methods they employed.

From both observations of participants and their reports several aspects of the Guidelines were identified to which improvements could be made: their general structure and tone; navigation within and between the documents; the content and presentation of examples; and additional information to be provided. These are similar problems to those identified by previous research.

For example, many participants had difficulties in locating the information they required within the Guidelines. They also experienced other navigational problems, for example they were not always clear whether they were reading the Guidelines or the Techniques documents. An important implication of these navigational problems was that some participants missed important information regarding the implementation of the advice and general accessibility issues.

The pages developed in the first experiment were collated and made available on a Web site [5] which was evaluated by 20 visually impaired Web users. The evaluation was performed remotely, i.e. the participants did not attend the laboratory, but accessed the site from home or work, at a time convenient to them.

Participants were asked to view several examples of each HTML element, such as images, frames, tables etc. They were then asked to state which example they found to be more accessible, and why. In addition, the participants were asked to interrogate each table in order to find the answer to a specific question. Participants submitted their answers to the questionnaire via email. Participants chose which elements they wished to evaluate, so not all the elements were evaluated by all the participants.

The participants used a wide range of browsers and screenreaders to evaluate the Web pages. The browsers included mainstream browsers such as Internet Explorer and Lynx as well as specially designed browsers, such as WebSpeak. The participants had an average of 2.7 years’ experience of using the Web, ranging from a few months to 10 years.

The most interesting finding from this second experiment concerned the images. Although the page authors had given the images alternative text, this was not available to six of the fifteen visually impaired participants who evaluated the images. The accessibility seems to be related to the browser used and the participants’ experience of using the Web, rather than to the content of the alternative text. There is an apparent assumption in the Guidelines that assistive technology will automatically present alternative text to the visually impaired user. However, this finding suggests that this may not be a reasonable assumption.

The page authors involved in experiment 1 implemented various elements and attributes suggested by the Guidelines. In the tables IDs and Headers were used, in the forms Fieldset, Legend and Label were used, and for the frames, Noframes were used. Although the page authors followed the examples provided in the Techniques document, they each implemented the advice in a different way. The browsers used by the visually impaired participants do not yet support some of these elements. The participants were therefore unable to comment on them. Despite the lack of browser support for the accessibility techniques for tables, all but three of the participants were able to access them. This is in surprising contrast to the inaccessibility of the images.

The page authors also made other design decisions, such as changing the layout of the tables and of the frames. In some ways these decisions had a greater impact on the accessibility than the use of some elements intended to improve accessibility. This is of concern, because the use of such elements should have improved accessibility, and because the page authors spent time learning about and implementing these ‘special’ elements.

Experiment 1 evaluated the use of the Guidelines in the adaptation of Web pages. A number of problems with the Guidelines were identified, some of which were similar to those found in previous research. For example, navigational difficulties meant that important information was missed. The findings of experiment 1 were reported to the WAI WCA Guidelines Working Group. The suggestions were discussed and some changes were subsequently made to the Guidelines. In experiment 2 visually impaired Web users evaluated the accessibility of the pages developed in experiment 1. There were some unexpected results for example, tables seemed to be more accessible than expected and the alternative text for images less accessible. Some design decisions made by the page authors that were not based on the Guidelines seemed to increase accessibility more than some of the accessibility improvements, few of which were supported by the participants’ browsers. The findings of experiment 2 will also be reported to the WCA Guidelines Working Group for discussion.

Position paper for the IFIP TC.13 INTERACT'99 Workshop: Making Designers Aware of Existing Guidelines for Accessibility (31 August 1999)

The American Telecommunications Act of 1996 changed the meaning of access to technology. Suddenly, anyone could enter the telecommunications business. For those to whom even a phone call was an infrequent occasion, a new world was opened. Section 255 of the Telecommunications Act reflects the US Congress' recognition that individuals with disabilities need improved access to telecommunications technology. Additionally it has brought attention to a significant market.

Passage of this law presents a tremendous opportunity to Nokia and all telecommunications equipment providers. Beyond the significant merits of broadening access to telecommunications service and equipment for people with disabilities, enhancing our accessible product line also offers this industry the opportunity to reach a developing global market of almost 750 million people with disabilities. Because of the broad scope of this legislation, each of our employees is in a position to have a positive impact on this initiative. As a company, we believe it in the best interest of the industry to share what we have learned so far.

Nokia envisions a future network-rich communications environment that allows communication anytime, anyplace, and in any mode or combination of modes. The potential is there to use this increased functionality and flexibility to bring the benefits of the telecommunications revolution to all Americans, including those who have traditionally faced accessibility barriers to telecommunications products and services. And because Nokia's reach is global, these benefits could spread beyond US borders. Of course, we face challenges in realizing that potential.

Section 255 of the Telecommunications Act, along with Section 251(a)(2), mandated that manufacturers of telecommunications equipment consider accessibility during all phases of product design, development and fabrication. To implement its obligations under the '96 Act, the "Access Board" convened an advisory committee to develop recommended equipment accessibility guidelines for its consideration. Thereafter, the Access Board adopted the Telecommunications Act Accessibility Guidelines for equipment. In April 1998, the FCC adopted a Notice of Proposed Rulemaking (NPRM) dealing with its implementation of the regulations. Then, on July 14^th, the FCC issued a Final Report on the NPRM and finalized this process at the end of September 1999.

As it stands, Section 255 requires a product-by-product approach, accessibility on all phones when readily achievable. It also requires prompt and effective customer service and information options and imposes regulations on Internet telephony and voice mail accessibility.

For years, Nokia has been designing products with easy-to-use features such as ergonomic design, wireless text messaging, and large displays. The company also offers user manuals in alternate formats, including Braille, large print, cassette tape, and text on disk. Many of our phones and accessories make functions easier, but they are not directly developed for disabled consumers. Nokia, well known for it's intuitive user interface and ease of use, has been guided by a very close relationship with its customers. The core Nokia Values: Customer satisfaction, Achievement, Continuous learning and Respect for the individual have been adopted by designers and developers alike. But during the past 12 months, our company has accelerated its efforts to make its features known to the disabled community, as well as to further product and accessory development specifically for these users.

Nokia's global values, part of the "Nokia Way" are supported by a flat organizational structure and an atmosphere of openness that allows anyone with an idea to come forward and participate in problem solving. That atmosphere fostered the development of Mikko Haho's invention, the LPS-1 Loopset, produced at just the right time for the market. The Loopset almost single handedly put Nokia in the leadership position concerning Accessibility. But the path after the Loopset is less clear.

With the launch of the Nokia 6100 series, Nokia made the full transition to global platform development. Working from a global perspective is a significant challenge when considering regional or even national policy requirements. In order to determine the scope of this new policy, Nokia immediately convened a steering committee made of specialists in every imaginable function at Nokia.

At first, the committee was tempted to view Section 255 only as a legislative requirement to provide for people with disabilities. But as the committee realized the impact it could have, accessibility became a mission.

The committee took a bold grassroots approach announcing early both internally and to our customers our commitment to this issue. Once Nokia made its intentions known, the team got down to business with the full support of management.

One of the first things the project managers and members of the internal Accessibility Solutions team did was develop a manual for Nokia employees. The manual, "Meeting the needs of a diverse global marketplace," was the first step in explaining Nokia’s commitment to this market.

The second step was to get Nokia's core function, engineering, involved and informed. Very quickly, Nokia assembled a meeting for engineers from across the globe to examine the guidelines and create a technical translation of the Section 255 Guidelines. That translation would be used to rewrite product specifications.

The third step put the Accessibility team into networking mode, educating our designers and engineers about the needs of consumers with disabilities. With a personal perspective, our engineers could better understand if not what needed to be done, at least why. Finally, the new product requirements are in development, and with them, various support elements, such as a Website and technical manual.

One of the major challenges we face is the sheer speed at which these developments are occurring. Access must be planned and implemented on the same schedule as general products. New platforms must be designed, developed and fabricated at the outset to be accessible and usable by individuals with disabilities. To do otherwise means that people with disabilities will become more isolated rather than empowered by these advances. It will mean that they will be forced backward rather than propelled forward by the potential of this technological revolution. However, perfect development won't happen all at once. It will be a process of creation and assessment. That is how Nokia has gotten where we are today: by listening to the customer.

To conclude, here are the key measures Nokia put in place to make accessibility a solution rather than just a requirement. Get input and buy-in from key engineers and product support people before the guidelines go into effect. Take the steps needed to translate guidelines into design and engineering terms. Make it personal by sharing experiences. . Empower people with disabilities and offer accessibility solutions that encourage and help customers be more independent and less reliant on assistance from others. And, foster consistent communication

Professionals in the field must develop good engineering practices, find the means for exchanging information on best practices, find ways of informing themselves and their colleagues about developments in technology, regulations, public policy, business practices and resources useful in addressing accessibility concerns. Nokia's vision is to design and engineer future products that are just as usable and feature rich for people with disabilities as they are now for our other customers.