The Importance of Digital Information and Tools Design in Enhancing Accessibility

Speech given at the Becta & Toshiba Seminar on the “Digital Divide (UK Perspective)”

Date: 19/02/2002
Venue: Coventry, UK

0. Introduction: Some Hard Questions

I want to begin with an extended prologue which raises a number of fundamental issues which will set some kind of context for a discussion of the design of digital information and tools and, in particular, the use of tools.

A few months ago I was asked to bid for a DfES contract to establish an evidence base for the proposition that adults acquire literacy skills more effectively through the use of ICT than they would through traditional methods. I thought that the proposition was so self evident that it would be a waste of public money to pursue it. Just before the deadline for bids I was telephoned by an official who asked why I had not submitted a bid. My answer was that it seemed obvious to me that somebody learning to read would clearly be better off using a screen and voice output simultaneously rather than struggling with a book. The response was that the research would only compare the use of a book with screen output. The comparison, in other words, was simply a matter of analysing the ability to comprehend symbolic, written language using the two methods. This was, in my view, a clear case of pretending that voice output did not exist.

Clearly, it is important to comprehend symbolic language where there is no voice output but such occurrences are bound to become ever more rare. I do not regard it as any more demeaning to listen to an audio book than to read a printed book.

At the heart of this small controversy there is a big question.

What balance are we to strike between learning a process and producing an output?

Do we want people to learn how to read print for its own sake or are we equally, or more, concerned with people being able to get hold of and understand information?

Important though this debate is for members of the general adult population who need to acquire literacy skills, however defined, these are vital questions when discussing the education of disabled children. How far are we justified in withholding a technique that improves output on the basis that we wish to emphasise process? It is, after all, difficult enough to function as a disabled child. There is a strong argument that you need all the technological help you can get. There has to be a very powerful argument for withholding technical support. That argument has to rest on some idea of centrality, on the notion that symbolic language is so fundamental that its acquisition must over-ride all other considerations. In a world of cinema, television, DVD and a multimodal, multimedia Internet, popularised through digital television, that assumption surely has to be questioned.

There is another set of factors on the horizon that also have to be taken into account. I can best sum these up by going back into the history of arithmetic. The Roman system of recording numbers made even simple addition quite difficult; you could neither add from the right nor from the left, dealing simply with two symbols that were one above the other; the convention, for example, of writing four as a V with an I to its left made this impossible. Equally impossible was adding from the left using the abacus technique. The articulation of the concept of zero and the introduction of Arabic numbers changed all that. Addition and subtraction became simple and long multiplication and long division became possible. Still, these were made easier after the development of tables of logarithms. These, in turn, were swept aside by the development of calculating machines.

What is interesting about this string of developments is that the skills needed for achieving an output have been reduced with each. There is, however, in our contemporary world, a new factor in the relationship between human beings, technologies and techniques, and output.

The simpler a process, the more amenable it is to automatic machine transaction.

Since the birth of the telephone exchange we have become accustomed to various levels of hybrid systems involving machinery and human beings, what we might call "hybrid systems". As digital information production becomes ever more sophisticated and modular and as computers become more powerful, an increasing number of formerly hybrid processes become totally machine transacted. Here is a familiar example. If you telephone a financial services company for a loan, you may be asked a long series of questions which the operator enters into a computer; they are then sent to an automatic evaluation system which gives a "yes/no" answer. If you fill in these questions using an Internet pro forma, you can cut out the operator at the end of a telephone; you have moved from a hybrid to a machine only system.

So the root question which this poses, which ends the rather extended prologue to my talk, is:

What is the point teaching people what will only ever be very low grade literacy and numeracy skills if these will soon no longer provide people with low grade office jobs?

Your answer may be that such skills are useful in life outside work - and that may be true - but the Government's policy towards the education system is that it is a pre vocational instrument.

You may hear these machine processing developments referred to as "The Semantic Web" or "Resource Description Frameworks". You might not need to know about them yourself but you surely need to know somebody who does.

On the face of it, this is pretty depressing for many disabled children and for those whose achievements fall below national benchmarks. It is worse than that. I will discuss some major issues in digital information and tools design, pointing out that new products offer new opportunities before coming to the sad conclusion that the current system, far from mitigating the disadvantage of disabled children, is exacerbating it and that, without some radical re-thinking, new technologies will make that situation much worse.

I am going to split the topic into eight broad headings:

Holistic Systems
Demographics
The Accessible Information Matrix
Multimedia and Multimodal Information
Tools, Metadata and Data
Manipulation, Customisation and Heuristics
E-learning: Choice, Control and Collaboration.
Conclusion

1. Holistic Information Systems

Digital information systems can be divided into three basic components:

Hardware and Applications which receive, process and send information
Carriers that move information between senders and receivers
Information, systems and tools producers.

Since the advent of the personal computer we have thought of the chief barriers to access in terms of finance and hardware. I am not saying that these are not still important but I think it will be granted that the cost of ICT in real terms has fallen and continues to fall at a significant rate.

In the medium term, too, the cost of hardware per individual will drop significantly. At the moment you might have a PC, CD player, DVD, television, radio, land and mobile phone, not to mention the processors in other domestic appliances. Soon we will each have modular hardware for access, processing and transmission, one at home and one that travels with us. At the moment this hardware duplication results in a heavy focus on intelligence inside machines, on what are called client side services; but this de-duplication will free up resources for server side applications. These are better for most applications because they can be centrally updated and allow for a much more powerful nerve centre. This will mean that disabled people will be much more like their peers, designing and using customised hardware constructed from modules.

In my own case, I estimate that my consumption of refreshable braille cell technology will drop from its current 98 cells at £100 each to 40. I will also be able to use these cells to access equipment, such as my stereo, television and cooker which are currently inaccessible.

The second part of a holistic digital information system is the 'carrier', the telephone lines, wireless and satellite systems. The scarcity of bandwidth has caused considerable problems in the past decade; we have known that certain new developments, such as audio captioning, are possible but the technology has not been able to deliver again, the economics and availability of broadband are still in flux but the ultimate destination is not in doubt. Always remember, in the digital age two rules of forecasting have so far always held: first, new developments are hyped and, as a result of that, secondly, the medium term outcome is always grossly under-estimated.

What is not in doubt, on the basis of these rules, is the growth of digital television, particularly amongst what we used to call "Social Classes D and E". This development has been grossly under-estimated by the Government because it is still blinded by prejudice in favour of the PC and against the television on the basis of a neo-Cromwellian prejudice. While poor people are buying digital television services at their own expense the government is throwing clapped out PCs at them.

So, although there are still some problems with the first two elements of the whole system, we need to focus on the third element the manufacture of the information, the way that it is made and that is the subject of the majority of this presentation.

But before that, I must take a brief look at demographics.

2. Demographics

To a certain extent, disability is an epidemiological construct but at its base it is an administrative device. There is a line between those classified as disabled and the rest of the population which determines criteria for public and private expenditure on benefits, services, special provisions, privileges and even Christmas dinners. This may be necessary for certain classes of financial decision but it obscures the underlying demographics of functional limitation, not least as they relate to ICT. To understand the complexity of this you only have to imagine two examples: a wheelchair user who can function perfectly with a PC and a personal with no measurable functional limitation who freezes when presented with a PC|.

The best way to imagine the broad demographics is to think of

Four Major Clusters of Functional Limitation:

Cognitive
Physical
Hearing
Visual.

These are shown in descending order of magnitude and they do not relate necessarily to a medical condition. I am not attempting to separate those who have literacy problems because of a cognitive difficulty from those who are alienated from the educational process or those who simply do not know English.

Each of these clusters is represented by a spectrum running from severe to mild. The four clusters can each be represented by a wedge shaped piece of pie with the mildest cases at the outside and the severest nearest to the centre.

So, in our drawing, the population as a whole is represented by a circle: half of its are, represented by the outer band, consists of people with no measurable functional limitation; the inner 50% consists of people with some form of measurable functional limitation; within that, there is an inner core of 10% of the whole or 20% of the inner 50% which consists of people who would be Registerable as disabled. The approximate proportions of the inner 50% by cluster are: cognitive 40%, physical 28%, audio 18% and visual 14%.

It is important to say, to avoid being sidetracked into a massive statistical dispute, that these figures are approximate and cannot easily be verified because of problems with transposing health, benefits, and ICT functionality data to form a coherent pattern.

Whatever the statistical quibbles may be, the broad picture to bear in mind is the 50/50 split. Admittedly the inner half is heavily biased towards the elderly. It is also important to be clear about this 'hinterland' of people between disability and full functionality.

To take an example from my previous specialist field: there are approximately 12,000 braille users in the UK but millions of people who cannot read Times New Roman 10-point on a standard PC screen. Between these two points there is a bewildering variety of functional limitation concerning the size of print, its font, contrast, ability to handle right hand justification, performance variants according to background and ambient light.

In the area of cognitive functional limitation the array of competences and barriers to optimal performance is even more bewildering. Here the problems even stray over into the blank half of the pie chart; there are Nobel Prize winners who do not know left from right and have problems with navigation, particularly in a disorganised, three-dimensional space.

I should also say, from the point of view of digital information and tools design, that these areas of difficulty do not include considerations of usability, i.e. whether the system accords with natural human information processing.

I raise this here because although the classification for benefits and even for Statementing may be justified, any understanding of the use of ICT must abandon the strict line between those who are legally classified as having "Special Needs" and those who are not. A line drawn for one purpose should not be used uncritically elsewhere. There are all kinds of reasons why children may have short attention span but no reason why one should have a special package which simplifies process and reinforces attention while her neighbour has to make do with the standard PC bundle even though she suffers, without an official piece of paper, from the same condition.

Unlike analogue material, digital material is capable of immense variation and manipulation. As I will come on to say later, as long as the initial design is correct, what is good for extreme cases will invariably be helpful to mild cases.

3. The Accessible Information Matrix

Having established that there are four broad categories of functional limitation, what kinds of capacities are limited or, to put the matter more positively, what must we consider when designing a system? I will briefly consider seven criteria:

Accessibility
Apprehension
Transparency
Navigation
Interaction
Expression
Fitness.

Accessibility

There are, of course, some quite basic physical arrangements which are often overlooked such as adjustability of the intensity and spread of ambient light, adjustability of the relationship between the hardware and the user and an choice of keyboards, mice and other hardware input devices. You would be surprised how often such simple things are overlooked in accessibility suites. Ask yourself the question, for example, how many PC screens have you seen where you can adjust the tilt as well as the orientation?

These physical factors, however, are only one sub section of a classification we call "Accessibility", the ability of the user to possess the intellectual content that is being offered. You have to remember in this context that hardware peripherals alone are no guarantee of accessibility; most require adjustments to configuration systems, new device drivers and proprietary software. What is more, many of the somewhat Heath/Robinson set-ups that are needed to drive special access peripherals throw up conflicts so, for example, if you run a refreshable braille display it is almost certain to disable the tracker-ball device. So accessibility depends crucially upon aspects of operating system and software design.

This, of course, is only the beginning. You will all be familiar with such aspects of information customisation as the ability to alter size of print, font, foreground and background colours and contrast. But - and here I am straying back into the process/output controversy - what about the ability to adjust the lexicographic range of a document or to invoke automatic grammar simplification, parsing, thesaurus-like functions, translation into another language? What about the ability to get rid of part of the file, such as a background picture or foreground text, in order to make better sense of what is being presented? Without such tools, many people will not be able to use their senses to get hold of the information.

What we are faced with here is a fundamental issue about the relationship between the creator of the intellectual property and the consumer where we have to define the optimal balance between what we might call entropy as the information passes from the creator to the consumer and the ability of the user or intermediaries to reduce the entropy through customisation. We are familiar with this in the analogue world where braille books are often produced without pictures and where sound tracks are sub titled for deaf people.

To what extent should we be allowed to alter the digital files of content creators? My answer, broadly, is that we should insist on the right of customisation and the use of tools as long as the output is in temporary files that self-destruct according to certain agreed criteria. We will return to this; but my key point is that "Accessibility", per se, is only the beginning, not the end, of the story.

Apprehension

Apprehension is rather difficult to define but it is a much simpler idea than "comprehension" which is not a responsibility of a system. Apprehension is the ability to grasp what you are being told as opposed to what it means. The central problem here is the use of spatial devices for aesthetic purposes which cloud the author's intention. The classic case is skewing the sequence of information on a screen so that the user doesn't know the order in which elements should be read or dealt with. Usually, this involves breaking our traditional left to right and top to bottom sequences for aesthetic or "playful" reasons.

Transparency

Related to apprehension, transparency requires that there is a separation between style and content so that the user is clear about what is being said, as opposed to becoming involved in how it is being said or, more precisely, rendered. Failure to observe this simple rule is the most widespread problem in Web design. XML should minimise this problem.

There is also the dimension of separating the tangential from the central; this is an occidental concept but most of our users are being brought up in an occidental cultural context.

Navigation

Navigation is probably the most critical are to think about, particularly in respect of tools and options. Ironically, there has been an almost complete failure by Web designers to use hypertext properly which is, after all, the basis of the Web itself. The defining quality of hypertext is its ability to multi classify any node. Look at the average PC and you still see menu trees, the very device which hypertext was developed to overcome.

I would say that at the very least a good knowledge management system (KMS) should have the following navigation choices as appropriate to the subject:

Alphanumeric/chronological
Key word search/ voice in (with micro vocabulary)
Spatial/image mapping
Variable numerical taxonomy (2/4 for switches, 9 for numeric keypads)
Maximum display (for pointing).

Interaction

The term interaction is usually used in the narrow television sense of choosing a game or a camera angle but by it I mean the ability of the user to do what she is being asked by an interrogator. This means, for example, knowing how to fill out your tax form because the relevant note appears above the box you are trying to complete. One of the main techniques for bringing this about is that the system should be interrogative instead of being a simple digitisation of an analogue process.

I should also make a simple point here about voice in systems. It is commonly thought that these are particularly beneficial to people with poor writing skills. There are cases where this is true but, in general, because voice in systems are rigidly rule based, they are by no means a panacea. The key concept here is the definition of micro vocabularies for specific topics.

Expression

Most digital systems are simply sales or information systems designed to send material from big organisations to individuals; they are, in the jargon, one to all messaging systems. The chief delight of the new technology, however, is that it allows all to all-messaging systems. But this is only possible if the information system allows the user to express a view and add to the sum total of the system. This is partly a moral requirement but it dictates a technical requirement.

Fitness

Finally, there is the fit for purpose criterion which applies to the whole information system; is this the best configuration to do the job for the customer or child?

This is a long agenda but most of it is implicit in a multimodal, multimedia environment.

4. Multimedia and Multimodal Information

In general terms, the assumption that multimedia information is highly beneficial can only be right but, as we have observed in the previous Section, it can cause difficulties for apprehension and transparency. What some users will find helpful others will find distracting; some people will find that two media will be mutually reinforcing while others will find them mutually distracting.

When building multimedia information the simple rule is that

Each medium, no matter what its reinforcing characteristics for another medium, must be as individually comprehensive and coherent as possible.

So, for example, the visual track does not depend on audio nor the audio track on the visual. In other words, the information is built so that it is multimodal as well as multimedia with each mode as independent as it can be.

This is not a simple matter of invoking some basic, bare bones principle of interchange ability. If you cannot access a given mode, such as graphics, you do not just want to know the title of what you are missing but also some estimate of its significance. This involves either an extra effort on the part of the content creator or the facility of an intermediary.

Let us take an image of a soldier who is pictured against a line drawing of a public building with a slogan at the top saying: "END THE WAR!" Now without further detail the significance of this picture to the blind person or someone not schooled in visual imagery cannot be understood. It might be a recruitment poster with a fresh faced young lad in contemporary military garb standing outside the national assembly, with a slogan in straightforward bold type; and the implication of the slogan might be, as it was in the First World War, that the level of recruitment will determine the length of the war. Or it might be an ironic poster with a clearly camp soldier dressed in antique military uniform standing against the line drawing of a war memorial with the slogan in gothic letters associated with funeral invitations.

I ought to make one aside here. That is that although we are eager enough to label pictures we are very poor at explaining their significance and we also do not understand well enough the huge variation in the ability to see and understand pictures. As the Internet becomes ever more graphics based this will become an increasingly significant challenge. On the other hand, many of us still struggle to do everything ourselves rather than using the vast resources of the Web. If I want a description of the Mona Lisa and associated links I go to the Louvre Catalogue.

There are some fairly simple tests that any designer can do in order to establish the degree of multimodality of a multimedia site, turn the sound off, turn the pictures off, turn the graphics off, turn the text off.

There is a basic philosophical issue here which we need to confront. No matter how good your multimodal site is it will not be equally good in all modes. You can't translate between media without considerable entropy; it's not easy to dance about architecture or write music about wine; nor is it easy to explain the significance of the smile of the Mona Lisa to a congenitally blind child. Here we need to be honest about limitations but that is quite different from not designing to the best of our ability having borne in mind different sets of user requirements for different modes.

As with so many other aspects of this topic, retro adaptation by intermediaries is far more expensive and inaccurate than placing a requirement on the author, as part of the brief, to produce multimodal material. Of course, this can't apply to a Damian Hurst, Jonathan Franzen or Thomas Ades but it can and must apply to authors of public sector material, particularly for the National curriculum.

5. Tools, Metadata and Data

The Internet brings into focus the symbiosis between data, metadata and tools. In the analogue world these were very separate. You might, for example, take a book and use a dictionary to simplify a passage from it. You might then write out the simplified passage and draw an illustration pointing out key features. You might consult some bibliographies to create cross-references, including, perhaps, some references to cultural artefacts other than printed books. At no point in this process, however, other than making pencil marks in the original text, would you in any way interfere with the product sold to you by the author; and even then you would not be altering the author's product in such a way that the alteration would not be immediately recognised.

I talked earlier about temporary files as a way of preserving the author's intellectual property rights but the subject of the relationship between the author and the consumer is far more complex than establishing rules for copyright and intermediary manipulation. Information designers should stop thinking of what they do as creating digital artefacts that work like books or celluloid films or pieces of music. The design of information should take place in the full and certain knowledge that there are tools which can assist users in benefiting from it and the information should be amenable to those tools.

The navigation issue also raises the acute problem of metadata. It is not enough for us to take huge quantities of information from a variety of sources and simply hope that users will find what they need. Authors and systems managers have a responsibility for metadata. In academic papers this is now a standard obligation met through the author's identification of key words but in most fields creators are lazy. Take, as a simple example, the total lack of a science of subject line creation in e-mails. I make a simple rule now; if I don't understand the subject line I automatically delete the message without reading it. Again, there are lexicographic issues here. Because we are working in a digital environment not only can the same node be classified in a variety of ways it can also be described in a variety of ways which will lead users to it.

The key points here are that:

Information design is not a purely aesthetic pursuit;
Taxonomy is highly significant; but that
Adopting the Dewey System is an inadequate response;
Data and metadata should be amenable to tools; and that
Tools should be integral to information systems rather than being clumsily invoked in parallel applications.

All of which leads neatly to the fundamentals of digital information and tools design.

6. Manipulation, Customisation and Heuristics

And now I come right to the heart of my theme. There are three central requirements which all information - metadata as well as data - must meet if it is to be optimally amenable by the population in general and disabled children and adults in particular; it must be (and please forgive this rather ugly shorthand):

Optimally Manipulable
Intuitively Customisable
Heuristically Amenable.
Optimally Manipulable. The fundamental criteria for manipulation are that:

Digital information should be

As granular as it possibly can be and
Each element should be labelled as precisely as possible.

These attributes relate closely to the use of the Semantic Web and to Scalable Vector Graphics but let me give you a simple illustration.

Let us return to the picture of our soldier, the image is made up of a figure of a soldier, a line drawing of a building and a slogan. Let us say that there is also a piece of explanatory text at the bottom to the effect that this is a pacifist picture which explains the dress of the solder, the choice of building and the choice of the Gothic font. We now have four discrete elements to which must be added the background on top of which all these elements have been placed; let us call this dove grey. So, we have:

Dove grey uniform background
Fine line drawing in navy blue of war memorial
Multi-coloured figure of soldier
Slogan in 16-point red Gothic script
Caption in 10-point times New Roman.

Forget the picture for a minute and think of a piece of recorded music. Anybody born since 1950 will be totally at home with the idea of multi-track recording which allows engineers to manipulate individual voices and instruments to alter the balance of the output. Nowadays pip bands routinely issue re-mixes of songs for different media and environments.

Now try to imagine the same technique being used for our picture. Instead of a bound up piece of celluloid that can only be manipulated uniformly, so that any enlargement affects all elements equally or so that darkening or lightening affects the whole picture, with this stratification technique a user could manipulate the five layers of the picture totally separately; and could, for example:

Change the dove grey background to white or another colour or intensity of colour
Embolden the line drawing without altering the slogan or caption
'Lift' the soldier into another window and magnify features of his uniform such as the brass buttons
Change the font of the slogan from Gothic to a simple upper case sans serif
Alter the size of the caption without adjusting the size of the slogan.

The central point here is that you cannot do any of this unless the whole collage, which was probably built up from separate elements anyway, is presented as a series of precisely labelled layers or strata. Some of this manipulation is beneficial particularly to people with a visual impairment but much of it is of great benefit to people with cognitive problems, it has the added beneficial side effect of allowing the originator to update one element of the collage without having to make a totally new product.

Intuitively Customisable

Almost all disabled children need some for of customisation for their information system; that is, in a sense, what defines their condition; they cannot use a standard system optimally. But you cannot customise a system that has not been designed to be manipulable. Even then, however, there are some important criteria; put briefly, any system must allow the user to:

Identify features and optimally customise
Store characteristics/profile
Invoke characteristics/profile from anywhere in the system
Restore the system to default from anywhere in the system.

The first of these four, the initial customisation, is the most important. Even with subjects who know exactly what colours, fonts, size, combination of media they want, the process of customisation is tedious. Systems have certainly improved, particularly since the advent of NT but the process is still very complex. Again, reverting to my own field, visually impaired children have a wide variety of needs and devices so you cannot make any uniform provision based on a rough and ready assessment. A child that uses a screen reader will want text set to the left, a child using screen magnification may want the text spread evenly across the page; a child with retinal damage may want all the text bunched tightly in the middle of the screen whereas this would be troublesome for a child with retinitis pigmentosa. You can't make any rule on the basis of visual acuity, not least because all our eye tests are designed for the agricultural and industrial age, not for a life spent largely inside a metre.

Obviously, once you have got through this process you want to be able to store the characteristics and invoke them immediately from anywhere in the system and also put the system back to its default; this is vitally important where different children are using the same machine. Again, this has been much easier since the advent of NT. Options for storage to be considered are: within the system; disc; smart card; web site.

Heuristically Amenable

The extent to which any system should be heuristically amenable reverts to my key questions about process and output. There are, of course, some aspects of heuristics that are not controversial. If a system can learn to adjust to the behaviour of individual clients then it will take care of a good deal of the customisation we have just discussed. A system will also help to sort out priorities and make searching more fitted to the needs of the client. Based on past behaviour it will steadily shift priorities.

Do we want this? Or do we really think that the whole essence of education is that if it's not hurting it's not working? Having asked the questions, I want to put the whole issue of heuristics onto one side while I look at some e-learning issues but I will come back to it in my conclusion.

7. E-learning: Choice, Control and Collaboration

Before reaching my conclusions I want to register the key advantages to users in general and disabled people in particular of the move from analogue to digital learning tools and environments.

Of course, whatever these tools and environments may be, we need to remember that any learning system is a hybrid between artefacts and human beings as facilitators, so I am not advocating that disabled children should operate in a totally virtual e-learning environment any more than I would advocate a traditional environment without teachers and mentors but there are three key areas of advantage in e-learning which ought to be noted; these are:

Choice
Control
Collaboration.

Choice

Making analogue learning materials amenable to individual needs is costly: transcribing books into braille; making different kinds of print; altering colours and contrast; simplifying images; removing background sound from audio drama; isolating individual elements; multi-tasking between different products. All of these are made much simpler in a digital environment but, as I have indicated throughout, this is only possible if the whole information system from creator to consumer, via the system, the tools and the organisation, make this possible.

These requirements highlight one of the most important aspects of granularity and that is the separation of tools from content. To give a simple example, if a file contains a piece of text and the page making is integral to the file it is more difficult to enlarge the text without producing corrupt pages at the page making process. If the text and associated images can be manipulated and then the page making tools applied, adjustments are much easier. If we can establish what I call the 'flat file' and define a set of manipulation tools it is much easier to alter formats and outputs.

Control

All the above devices help individuals to control the environment in which they learn. This does not simply apply to the microenvironment of the learning system, it applies to the wider environment of where people learn. If information systems are designed properly it will facilitate much more remote or distance learning which may be particularly useful for students who find mobility difficult. Teleconferencing and advances in surveillance systems provide a great deal of flexibility if used sensitively. However, one area which needs much sharper definition and design is that of procedural tools: "clock here" and "Are you sure" are somewhat primitive.

The additional area of choice which these systems offer is the ability to work outside conventional school hours but, again, this does require high quality procedural tools. One emerging phenomenon in the e-learning environment is the apparent failure of users to notice that on-line assistance from a tutor cannot be available on a 7/24 basis even though the basic materials are always available. Tutors cannot be available on that basis; but the better the procedural tools the less need there will be for assistance on routine matters. There ought, also, to be much better heuristic tools so that a system recognises when a user is confronted with difficulties he cannot solve and where he is referred to a tutor or, if it's three in the morning, advised that he would be better off taking a break because he is getting nowhere.

Collaboration

Last but by no means least, there is the issue of collaboration. It is my contention that the development of a National Curriculum, largely based on solo achievement, has disproportionately disadvantaged disabled children whose life chances are best enhanced through collaboration. It is unfair to place such a high value on solo achievement in the testing process when that simply disadvantages people in seeking opportunities in largely collaborative environments. I recognise that this is a profoundly political issue but with the new digital learning environment we are confronted with the question; are we going to take advantage of collaborative potential in digital learning systems or are we going to pretend they do not exist, as some have done with pocket calculators?

On-line collaborative learning requires highly sophisticated procedural tools and the design of these is only at a primitive stage; but if there is a case to be made for allowing children to behave at school as they are likely to behave at work then that case is even stronger for disabled children who start with severe disadvantages and finish school facing a competitive and largely hostile world.

We have come full circle. Employers are much more interested in the ability to output rather than the processes people use to get there. If I am puzzled by a problem I usually resort to my e-mail list just as others frequently send me enquiries. That is what collaborative learning and output is all about. We are looking at a real as opposed to a metaphorical network.

8. Conclusion

Periods of rapid change, economic, social or technological, always widen the gap between the best and worst off. Even if the worst off benefit absolutely, they suffer widening comparative disadvantage, that is the situation we currently face with the Information Society. I do not think it is difficult to argue that in the learning environment disabled children can readily be classified as sufferers from that comparative disadvantage. As the skills ratchet continues to operate, as machine processing advances, as learning opportunities proliferate, as new techniques are developed, it is not difficult to understand why disabled children are in danger of being left behind. You only have to look at the pathetic state of the digital games market for children with special needs to understand the depth of the problem. While the major manufacturers turn out vast numbers of products for Play Station and Game Boy, a small voluntary organisation or a mad professor makes a little game for a disabled child. I am not asking for a dedicated niche market industry of special needs games suppliers but I am asking for a better understanding of the importance of digital information design as a way of broadening and deepening access to the Information Society at school and at home. We can crack the hardware problems, we can even, with some persistence, extract funding from the public purse, but the central task now is to understand information systems as holistic entities and concentrate on basic digital information design. If we do this, we will have to answer some important philosophical questions but without a great advance in design those questions will only be theoretically interesting.

Return to Accessibility Criteria