From Fashion to Demographics: Issues in ICT Accessibility for Disabled People

Lecture given to the Division of Applied Computing, University of Dundee

Date: 27/10/2003
Venue: Dundee, Scotland

The excessive complexity of ICT design is an indulgence with which many users learn to live but it imposes massive additional disadvantages on disabled people in terms of cost, time and effectiveness. One potential solution lies with consumer electronics but in this area inaccessibility is increasing. The solution lies with user centred design which embraces the dual concepts of accessibility and usability but these will only become generic when demographics, rather than style, determine product design.

0. Introduction

People who ingest massive amounts of data and opinion from across a wide range of disciplines are as susceptible to prejudice, i.e. a theory not based on proper information, as a foolhardy child parachuted into a tropical jungle is susceptible to accident. The quest for novelty without the precaution of planning and the support of infrastructure is a dangerous enterprise. Or, put less dramatically, it takes about half a generation for knowledge from one discipline to be integrated into another.

So, I had a theory which was that the complexity of ICT products derives from an economic analysis which indicates that companies are much better off upgrading the top half of the market than supplying the bottom half; and that, in turn, the only way they can justify the upgrade cycle is if they keep on producing ever more features. I will return to this idea later in this lecture.

In the meantime, I want to look at four inter-related issues:

The complexity of ICT design in general and information design in particular
The burden complexity imposes on disabled people
The 'false dawn' of consumer electronics

The dependence of accessibility and usability on a demographic rather than a style base

I finally want to draw some conclusions which try to get us a little further than pious conference resolutions about dignity and rights, important though these are, and away from the conclusion of many more academic papers than this which call for more research.

Before doing this, however, I want to put the complexity/usability dichotomy into a socio-economic context because we are not talking about the severe problems a few disabled people have with ICT when they are looking for jobs in industrial or scientific establishments in the late 1970s, as was the case when "blind" and "computer programmer" went together like haggis and neeps. In history, economics and technological development:

The more rapid the change, the greater the benefit to those who control the socio-economic levers; and
if only temporarily, the greater the comparative advantage they enjoy.

Now of course there are exceptions to this general rule, the heads of French aristocrats dropping into baskets and the shooting of Russian aristocrats at the snowy margins of railway tracks come to mind, but even then most of the intelligentsia, nomenclature and plutocracies survived respectively into the age of Napoleon and Louis III and Lenin and Stalin. The idea is simple enough. Think of:

People living near and far away from new railway lines, electricity supply, motorways
People caught up in rapid economic growth (the industrial revolution, the 1990s) and decline (the 1860s, 1930s)

Now let us focus in more closely on ICT development. At some point technology, reading and writing, counting, automobiles, televisions, telephones, reach a social critical mass. Before this point there is a minority of beneficiaries, after this point those who do not have the technology become disadvantaged as the technology determines socio-economic norms. It was no disadvantage in 1950 to have no car, television or phone; but it is now. As technologies develop and fall in price the disadvantage narrows. In the field of ICT, however, as the poor or less well educated catch up with one technology, such as the mobile phone, another wave widens the gap. This is not, of course, uniform; the poor and less well educated were ahead of their more wealthy and powerful peers with mass television and television related technologies; the use of the VCR and multi-channel satellite television grew from social classed E and D up through C2; so we must be a little wary about using the phrase 'digital divide' as if it were a chasm.

All the same, in spite of any absolute advantage the bottom half of the market may gain from emerging technology, it's the comparative disadvantage which counts in terms of incomer and opportunity.

So, having set some kind of context, let me turn to the first major item on the agenda.

1. The complexity of ICT design in general and information design in particular

Most of you know this story already but I will rehearse it quickly for completeness. There was an engineer and information scientist who had committed himself in his professional life to the concept of user centred design. His organisation bid successfully to undertake some work which brought him to Dundee. Like many others, he was accustomed to opening his lap top, looking at the array of more than 100 live elements on the screen and mentally occluding the things he didn't need which then enabled him to focus on the task in hand. But when he presented the same screen full of data to some computer novices they baulked; they couldn't automatically occlude what was irrelevant to them because it wasn't clear what was relevant.

This is not a surprising nor a novel phenomenon; the court of Byzantium was Byzantine because the people inside it didn't have to think twice about the nuances of every word and gesture; in all cultures there is a method for focus and occlusion. The problem is that a there is no cultural settlement yet about the role of the computer screen. For many, including designers, it is a cultural phenomenon which quintessentially allows for focus and occlusion, for constructing and deconstructing; but for others, including the people being tested in Dundee, the computer screen was seen as a method of undertaking a simple task. It was easy for the researchers here to measure the functionality of the systems, to see that a radically simple interface was more popular with novices than proprietary software and search engines; but underneath this there is a much more fundamental set of issues which frequently separate the designer from the user.

I break this down into four sets of issues:

Fashion and style versus utility
The Wild West versus the desk in the library
The ephemeral bite and the permanent work
Analogue re-treads versus design for interactivity.

In my abstract I describe over complexity as an indulgence but it might be more accurate to think of it as inappropriate. For the designer the task is performed inside a context of choice and diversion, of multi tasking and, yes, adventure. We are past the age of the recurrent head crash which challenged young males but computing is still frontier territory.

This is where I return to one of my original points. Having talked to hundreds of ICT designers over the past ten years I have finally abandoned my economic theory of complexity. I used to think that the upgrade was an economic conspiracy but now I think it is the fulfilment of a design urge which puts ICT in the same realm of fashion rather than utility. Just as fashion designers are primarily interested in dressing young people even though there is more money in the fuller figure and expanding waist line of middle age, so the software designer is primarily designing for himself (and it almost always is him) rather than for the market. Instead of thinking of his product in the same class as graph paper, he thinks of it in the same class as television. Software is definitely part of the media sector, the creative industries and it is therefore a slave of the artistic self-image of high romanticism. The sector also suffers from arcane freemasonry.

Paradoxically, perhaps, this freemasonry is not carried out in grave and solid halls in ancient capitals but in the Wild West. There is no sheriff. The Dundee users might think of computing as something relatively orderly that they might undertake in a library or in the peace of their own spare bedrooms but behind the clutter of the desktop there is the clutter of the thrashing corporation; ICT companies are usually either in rapid periods of growth or decline; they rarely enjoy steady 5% year on year growth. So, to give you an example, I went to Nokia a couple of years ago to discuss the possibility of putting a small voice chip into mobile phones so that blind people could use the menus; the head of research told me: "At the moment we are making a phone every 14 seconds and our single priority is to bring that down to 13 seconds; come back when we have got our production in line with demand". Six months later I went back and the bubble had burst; Nokia was laying off staff and the same man told me: "We are still committed to accessibility but at the moment we have to concentrate on the essentials". It wasn't just the .com companies that didn't have an adequate economic model, Nokia could easily have forecast the production bottle-neck and the slow down.

Relative size is another complexity. In the software market the size of MicroSoft means that most of the rest are adventurers; this is not the kind of market, like the white goods sector where there are globally a dozen or so big, solid players. The giant behaves badly and the Pygmies behave badly, too, but the situation is so fluid and the competition so fierce that they can never assemble a united coalition. There is no international jurisdiction equal to Microsoft.

The third complicating factor is different assumptions about the idea of a desktop. When I think of a desk I think of Mr. Casaubon in Middlemarch slaving methodically over the great book he will never finish. The times may have damaged my ability to single task; there is the telephone and now email but I am still essentially a man with an empty desk. Things wait in an orderly in-tray at a decent distance from my writing space. Digital desk-toppers, on the other hand, seem to enforce the widely held view that children emerging from our education system have short attention spans; or, put round the other way, they live in a permanent state of multi tasking. This is connected in my mind with different ideas about output. One of the stunning, less remarked phenomena uncovered by the Hutton Enquiry was the extent to which civil servants committed to email the kind of thoughts they would never have committed to paper as if such communication was not only personally ephemeral but also institutionally ephemeral; did they really imagine that all the messages automatically destroyed themselves? The bite sized communications effected in a blizzard of multi tasking encourage the view that this is a very different process from deliberatively sitting down to write a letter whose every word must be weighed. If the designer has any kind of mental picture of the user it is, by and large, self referential, lacking social imagination.

The fourth dichotomy comes from a totally different source. All over the world governments have seen digital communication as a way of saving money and keeping tabs on the citizen. Most of the substance of this assertion is for another day but I want to say a few words about form filling with the introductory note that almost every government in the world that has embarked on e-Government launched with income tax forms. Anybody who has seen a major government form with up to 32 pages of 6-point print with the notes a mile away from its matching box will know how inappropriate it is to take these documents and dump them on the internet without modification.

Finally in this opening section on complexity, I want to say something about an obvious phenomenon which can't be classified as a designer/user dichotomy but which accounts for a great deal of the complexity which confronts the user. Here are some characteristics of most information available through the Internet:

Exponential growth
Unregulated
Untargeted
Heterogeneous
Genre unspecific.

Let us just take a brief look at these ideas. We all know that the amount of information available on the Internet is growing so that it is now measured in LOCs (Libraries of Congress). This presents enormous problems for users, aggregators and information suppliers. To make things worse, there is no regulation. By this I do not mean censorship but, rather, the exercise of some measure of control over standards. Efforts have been made in all kinds of directions but the problem is too big.

A partial solution to this problem is the more judicious and granular use of addresses. Personally I think that pornographic sites should be forced to use .porn in their addresses; I am also an advocate of .me, .shop, .media and .ref. People can choose which class they think their information fits into but having made the choice they will be bound by a few rules, e.g. .ref will have some taxonomical implications. Without this the user will continue to receive a mass of undifferentiated information which was the fate awaiting the Dundee novices until a simple, laboratory situation was developed.

The solution to information searching lies partly in what I have proposed but it would be helped enormously if designers bore in mind the simple rule of (7 + or - 2).

Another part of the solution is the development of software tools which are centred on dialogue with the user; for example:

User searches for: dog

System answers: What kind of dog?

User: Scottish Terrier

System: What subject connected with the Scottish Terrier, e.g. buying breeding feeding grooming other?

User: Grooming

System: Do you want general articles from newspapers and pet magazines, specialist canine journals, vetinary or academic journals?

The trick here is to teach people to narrow searches until they are manageable. Such dialogues are immensely useful for all users but they are of particular value to disabled users, to whom I turn next.

2. The burden complexity imposes on disabled people

First of all, a couple of things I am not going to talk about: I am not going to define disability and associated concepts such as impairment. I am discussing disability and impairment as they relate to the interaction between people and digital information systems. In the fourth section I will talk about demographics but in the context of a market analysis of how the gaps can be narrowed between people and systems.

You will have gathered so far that I do not believe that those gaps exist wholly or even primarily because of user failings. At the moment much user frustration arises from poor design and a good part of that poor design imposes a cost shift from the producer to the consumer so that many training courses do not deal with the essentials of interaction with an information system but, rather, the design quirks of proprietary applications.

Here are the topics I will be dealing with under this heading:

Cost
Incompatibility
Idiosyncrasy
Incompleteness

Before we look at the consequences of complexity, let us start with a plain accessibility model which can handle simple data. The additional cost of accessibility technology varies enormously. In my case, one refreshable braille cell, made up of eight solenoids, costs £100. The device I usually work on a PDA with a CE operating system and a 32 cell braille display cost £4,200. Let us admit that that is the top end of the accessibility cost; but let me tell you a little about how I got here.

In the mid 1980s I came into computing on the end of C/PM with a 64kSony PX8 and a braille display which I tested but which never reached the market. I then moved onto a PC with a braille display with completely different controls, a different computer braille language, a different braille translation programme and, of course, I was learning DOS. I was getting used to this when Windows arrived. I changed braille display again and went to a lap top but I determinedly stuck to DOS until the middle of 2002 because the advantages far outweighed the disadvantages. Then I moved to the PDA. I spent precisely one month finding out that using yet another braille display with XP was impossible. Just to cite two problems: Jaws for Windows invariably puts the cursor for the braille display in the extreme right hand column so that some times you are typing a word but can't feel the end of it; the cursor moves from left to right until an arbitrary point when it stays still and the text moves from right to left so layout in columns is tedious. This experience isn't precisely incompatibility, I can use my PDA with braille display to communicate with the world, the process is just a deal more messy and tedious than it is for my peers.

Part of the reason for this is shown under the third factor which I termed "Idiosyncrasy"; you would be surprised at how many devices for the use of disabled people are put together with all the best of motives by people who think they know what is best for us. My favourite example is outside the ICT field, in the area of independent mobility devices; I know about dozens of these that have been developed over three decades but I have never seen one used outside evaluation and testing. That says something for the validity of the testing but it is a remarkably high failure rate.

These different irritants - cost, incompatibility and idiosyncrasy - are bad enough in themselves but the really depressing phenomenon is the poor outcome for all this expense and effort. Here I have more than £4,000-worth of equipment and I still can't access most web sites.

Now let me come on to the problems created for disabled people by complexity. We are getting close to a discussion which I want to avoid, for the time being, on accessibility and usability. All that I will say here is that a fully accessible web site may be unusable. The most common problem is clutter, too many elements on a page; the BBC site, the most popular in Europe, with its own special accessibility feature, is only usable if you have a phenomenal memory to match your phenomenal patience.

Let me say a few words about the averagely accessible web site. First, instead of being single task oriented or single feature oriented, a large part of a web page is repetitive. So, for example, if you elect to undertake a task and you are offered three choices and opt for one of them, you are usually taken to the top of a new page and have to creep down the metadata until you find the option you chose; every time you make a choice there is a new page and a repeated dump of metadata. Until recently the best way for a screen reader to deal with a web site was to climb a page from the bottom but now new data is usually the filling of a repeat sandwich. There is no quick way.

These problems, of course, are specifically described from the point of view of blind and visually impaired people using screen readers but they have their equivalents. People with low vision cannot use enlargement technology effectively because pages are too cluttered; people with learning and cognitive difficulties get swamped and lost; physically disabled people find cluttered pages unergonomic.

The simplest solution to these problems is to allow the system to customise according to the behaviour of the user. Let me give you an example in the field in which Alex Carmichael and I are working. We are currently engaged in helping blind and elderly people to make sense of digital television electronic programme guides (EPG) which carry one week's programme listings for just under 400 channels.

There has been a recent row between Sky, the platform owner, and the BBC which wanted its new Channels, 3 and 4 to be higher up the EPG. There is already a fair amount of data to show that viewing critically depends on EPG position. But, here's the clue; the EPG is not produced as a piece of self-perpetuating digital string, it's a database. So it would not be difficult to help viewers and get over the ranking dispute by setting up the system as reactive to user behaviour. If you are a Chinese immigrant, you might want the Chinese channels on your first page. As with the search narrowing, the objective should be to make painful encounters with the system count for more than the present occasion. Instead of an exclusive hang up with hardware and applications, what we need to think more about is the construction of dialogues to set preferences and the need of the system to monitor these for constancy and change.

The obvious drawbacks for disabled people with complex systems are cost, time and frustration. We have already said something about the hardware cost but on-line time must be taken into account, remembering that disabled people usually live in households with below average income. A screen reader will take minutes to find what a seeing person will spot instantly; a person with a learning difficulty may work hard, spend a lot of money and still have to give up.

The frustration is not just intrinsic to the process, it merges with concerns for equality and worries about comparative disadvantage. If I did not have an Admin Assistant I would have given up on the Internet altogether, even though I was a WAI legislator. Here I am, with state of the art technology, a clear understanding of how applications and navigation systems work, a good memory of the visual world; and I am still not autonomous. You can imagine how much worse it is for most other disabled people.

Never mind, the computer may be too difficult but I can always switch on the television. Or can I?

3. The 'false dawn' of consumer electronics

There was a time when television was easy. If you didn't like the programme you were watching you shouted: "turn over!". There were only two channels and it was either one or the other. The radio was a bit more complicated but not much. Manual tuning was from one end of a band to the other with a barrier at each end. Tape recorders used open reels. Putting a stylus onto a disc was always a bit tricky, I admit. The user interfaces for consumer electronics were usually robust, with mechanical switches.

Now let us see what it is like today. The first problem is the remote controller. I love to tell the story of a blind friend of mine who checked into a Miami hotel for a night and was charged for 24 hours of premium pornography. I find that I cannot use most hotel remote controllers without a great deal of struggle. Sometimes I find the channel I want and then in an attempt to adjust the volume I lose it. Television manufacturers and people from broadcasters like Sky say that the special features of a remote controller give them a special edge in the commercial market. I can understand why a few extra features might give commercial edge but surely not on/off, digits 0-9 and volume control. At the very least we need to undertake research to establish whether any features give an economic edge or whether the manufacturers just can't be bothered to reach a deal. The same goes for what should be the much simpler telephone. Many hand sets have the additional problem that they are too small, not just for people who lack fine motor skills but also for people who have fingers like a pound of pork sausages.

Music devices and radios also present problems with continuous tuning that requires screen access. In my house we have two remote controllers for the television and two different controllers for the two hi-fi systems. I can't find channels autonomously on television and can't pre-set channels autonomously on the radio.

I know this is all focusing rather heavily on me but I am supposed to be familiar with new technologies and highly adaptable. All I want is a small voice chip in the phone and each remote controller telling me which button I have pressed and where I am in a transaction. We also need clear screens in remote controllers.

The reason I entitled this section "The false dawn" was that I believed in what I had been told about the market, that it would ensure that consumers got what they wanted. It seems not to have done this. Here are some simple measures I expected from the market which have not fully come about yet:

Standardisation of the position of common features
Differentiation in shape of key features
Choice of size of handsets/remote controllers
Voice and screen in the control device
Voice in for standard instructions
Intelligence to monitor 'favourites'.

I don't think my expectations were unreasonable. The population sector which watches the most hours of television is that over sixty; and most disabled people are a sub set of this population group. This group now has a substantial disposable income which it carries into later retirement. Those with the highest disposable income enjoy it from the last child leaving university until the first major medical or care bill kicks in. We are back in the territory of style and fashion when we should be in the realm of market penetration.

The classic inaccessibility problem in consumer electronics is, as you will know from the work of Alex Carmichael, the Electronic Programme Guide (EPG) to which I have already referred. It seems odd that the recent Communications Act had to specify accessibility to EPGs not only for disabled people who are frequently discounted in the market but also for elderly people. The Office of Communications (OfCom) has set up an Advisory Committee on matter of concern to disabled and elderly people. This is evidence, if any is needed, of the strange state of the market in consumer electronics and entertainment and leisure services.

At the moment I am Rapporteur on a Brussels Committee on Inclusive Communications (INCOM) which reports to the Communications Committee (COCOM) which regulates telecommunications and broadcasting. What is remarkable is that EPG access and research on the standardisation of remote controllers are highly controversial issues. I can see why the cost of signing versus the number of people who use it might be controversial; and I can even see why small channels might baulk at the cost of text sub titling and audio description but I can't see the problem with simple changes which will significantly boost market share. What we have here is a failure of demographics; and that is my final major topic.

4. The dependence of accessibility and usability on a demographic rather than a style base.

I suppose if I were pushed I would have to say that the main responsibility for the poor deal which disabled people receive from ICT lies with the disability sector itself. I want to explore this in some detail before moving on to more general points because too often the disability sector blames every other party but itself.

Here are a couple of simple statements which sum up the problem:

Disability is largely an ageing phenomenon but charity fund raisers prefer children
People with congenital or childhood disabilities formulate policy for the charitable sector
It is difficult to move money from care for children and those who acquire early disability towards the elderly.

It might be helpful to go a little deeper; the process is something like this:

Organisations in the disability sector raise money through images of severely disabled children
They tell us that perhaps 1 in 10 of the population is disabled
People in the community conflate the two propositions and think that 10% of the population is severely disabled.
So severely disabled that the sector and not general provision must look after them.

This is compounded when disability organisations present two sets of contradictions:

We want mainstreaming but the community is so prejudiced against us (the social model) that WE have to implement it
If people were more tolerant it would be easy to mainstream but it takes a lot of money and expertise.

I just want to take this to its final stage with two points:

The sector frequently excludes the plurality of disabled people who have learning/cognitive/developmental difficulties
The sector opposes dilution.

On the first point, I have struggled unsuccessfully to persuade the European Union Commission to include learning/cognitive/developmental difficulties in its definition of disability because this inclusion is opposed by the European Forum on Disability which is almost exclusively concerned with wheelchair users, sign language users and totally blind people. Which leads to the dilution problem. The market case depends upon expanding your constituency. Black power people in North America had a complete change of fortune when they decided in the mid 1950s that people of mixed race were black. In this field the legally disabled should be making alliances with people with mild impairments to build a market case but there is absolutely no evidence of this:

blindness should be in alliance with people who have problems with times new roman 10-point and with car drivers who want access to on line services without using their eyes;
deafness should be in alliance with all those who want clear signage and clear displays of public information;
physical disability should be in alliance with arthritis sufferers and the clumsy.

I have said before that if you gave everybody in the population a Pentium IV and an unlimited training voucher, half of the population would not be able to use the system effectively. The figures for consumer electronics and EPGs are not available but I doubt that more than 2/3 will be able to handle these with full effectiveness.

My simple conclusion is that the ICT sector has to break away from its obsession with fashion, style and excitement and take money making a little more seriously. Just as there is money in conservative suiting and pressed linen, so there is money in simple applications and simple appliances. It is sad that we will have to resort to legal remedies to obtain rights of access to the information society but that only goes to prove what I have known in my bones for a long time; that people whose job it is to make money are no better at it than we are at making ideas.

5. Conclusions

I now want to draw some very simple conclusions:

First, although there will be a market for radically simple interfaces which observe the (7 + or - 2) rule, the primary simplification mechanism for most applications will be intelligence within systems which monitors behaviour and adjusts default outputs accordingly. Once basic telephony becomes free, as it will by 2010, the use of intelligence will drop in price because it will all be server side
Secondly, we need to invest in blended communication in broadband which allows human contact simultaneously with digital information access
Thirdly, we need to look towards convergence to bring us modular customer interfaces, one for the home and one for travel, combined with a portable user profile
Finally, we need to get the demographics right, to forge an alliance between accessibility and usability, and escape from the disability ghetto.

Kevin Carey
Director, humanITy
20^th October 2003

Return to Accessibility General