Design-led approaches get it wrong.

They take an average and who is exactly average ?

In the picture a cycle path seems to have been designed by committee to fit the resources available and to suit everyone, but does it actually work for anyone ? Clearly it doesn't work for cyclists, unless they can simultaneously cycle and contort themselves into a shape that will pass the tree while balancing on the slope and not falling off!. Neither does it work for pedestrians nor people with children in pushchairs - in fact the only creatures it seems to work for are squirrels and male dogs. But, you may say, resources were limited - all that was available was a pot of paint. In that case, a better solution might have been to not do anything at all and allow cyclists and pedestrians to share what path is available in peace.

But what are the real problems and factors here and how do they relate to accessibility ?

There are several:

1. Firstly, there is the problem that as people we are all highly individual. What works for one person at one time may not work for another person and may not work for the same person at another time or in another context. What is optimal for one person at one time in one context is likely to be highly individual. Something that might work for several people in a general way may not be optimal for any one of them.

2. Secondly there is the problem that a designer of anything can not in general know all of the requirements of every potential user in every context at every time. The designer has in many circumstances to form a picture in her or his imagination of the customer or user that will use or enjoy the design. That picture will be wrong. At best it will be an average but often it will completely miss the mark for some user. There is a communication gap here - what is needed is some way for requirements to be communicated between users or customers and the designer.
   

3. There is a need to commoditise products and markets. This is not a business argument, it is one born of structural numeric constraints in the problem. If we take a physical product such as a suit of clothes, its not very practical to make clothes individually sized to fit every person in the world. It would be a waste of resources to try to do so and in any case the resources needed to do this would probably exceed what is available on the planet. Also, it is the case that though no two people are exactly the same size (if we measure to the millimetre say) some people are nearly the same size as others and we can use that fact to structure production for the clothes market to make optimal use of resources and still meet the need. This DOES NOT mean that we can make just ONE size of clothes for every person in the world - but by making a range of sizes we can use our resources reasonably, meet everyone's individual needs (nearly) and still have time to enjoy the flowers, Mahler and Argentinian Tango. BUT CLOTHES ARE PHYSICAL PRODUCTS. With some obvious exceptions, clothes cannot usually reconfigure or adapt themselves to an individual's size. Neither is re-use of clothes by a wearer who was not the first an entirely pain-free experience - a second user does not have the same experience as the first wearer. If we look at other products, the more resources tied up in a product, its manufacture and distribution the more this argument holds and the more we need to build generic designs not design individual products for each individual. From the standpoint of Jeremy Bentham's utilitarianism [1], by making and manufacturing generic product designs we can meet the needs of a greater number of people than by trying to meet every need individually because doing the latter would take more resources than we have. In effect we must choose - to meet the needs of a few people perfectly or many people less perfectly. In an ideal (utilitarian) world we would arrange that the degree of genericity in a design optimises our use of resources to produce the greatest match between product and users requirements. For accessibility this would mean:
   

   "Accessible (but maybe not optimally) to the greatest number of people."

4. So far, we have talked only of physical products, When we look at electronic products the picture is slightly different. Electronic products seem on the surface to offer much greater flexibility than physical products. Once a bicycle has been built we cannot change the shape of its frame to fit a different person without excessive costs. On the surface, it would appear that if we do it right, then electronic products might offer real adaptability to individual requirements without cost. There are after all little in the way of physical resources that are tied to any one copy of one software or eLearning product. Re-use of software is to all intents and purposes almost completely free - extra copying and storage costs are close to zero. So we ought to be able to build resources and resource delivery systems (such as Learning Objects in Learning Management Systems) that can adapt to the individual accessibility requirements of every individual user without extra cost. Straight away we need to say that THE ONLY WAY TO MAKE A RESOURCE OR SYSTEM ACCESSIBLE TO A PARTICULAR USER IS TO MEET THAT USER'S INDIVIDUAL ACCESS REQUIREMENTS. A generic design might be accessible to some user or it might not but its unlikely to be optimal and there will certainly be some users who are excluded by its design. We are all so different from one another in our access requirements that the only way to do it right for accessibility is to do it individually. But still something is missing in order to do this.
   
   

5. A missing piece is that if we are to make electronic products such as resources, interfaces and delivery systems that can adapt to individual access requirments we need mechanisms to communicate between product designer or product and those users. The product designer needs to ensure that the product is able to adapt to every individual requirement that it is asked to. Along with this, we need accompanying product technologies and delivery systems that are flexible enough to meet all required adaptation. There are several different ways to achieve the former. One such way, often called "Design for All" [2] is to arrange that individual users with extreme requirements work with designers at product design time. Another mechanism (albeit a very slow one) is to provide after-the-fact legal redress where users finding a website to be inaccessible to them might take legal action against the owners of the site to get it changed (for example UK Law [3]). In a third approach, universities often provide support staff that work with "disabled" persons and university services to ensure that what is provided is made accessible to those individual users, sometimes working with university teachers on their content for example or arranging assistive technology or alternative media.
   
   

6. Another mechanism to "bridge the communication gap" is that of personal preferences. The idea is that each user has one or more sets of personal preferences, each specifying technical requirements to which resources, interfaces or delivery systems can adapt. For example (highly simplified):
   

   "Large-fonts 1800-06.00"

   which might specify that a user requests that when the time is between 1800 and 0600 then the resources and interfaces should be delivered with large fonts. To meet the needs of a user a system might adapt itself to the preferences at delivery time. Preference sets may be of great use also at resource procurement time and system design time as examples of what requirements need to be met.

   This solution requires that we all agree on the form of preferences to adapt to and doing so requires common technical standards that all of our implementations use. Several promising pieces of work are underway in this area, notably IMS Access for All 3.0 [4 and ISO/IEC 24751 Individualised Adaptability and Accessibility for Learning, Education and Training [all three of 5, 6, 7] which is currently under revision [8] to meet the needs of the Global Public Inclusive Infrastructure (GPII) projects [9, 10].

These developing standards will be the topic of a future post.

References:

1. [http://en.wikipedia.org/wiki/Jeremy_Bentham
2. [http://www.designforall.org/
3. http://www.rnib.org.uk/professionals/webaccessibility/lawsandstandards/P...
4. http://www.imsglobal.org/community/forum/messageview.cfm?catid=88&thread...
5. http://standards.iso.org/ittf/PubliclyAvailableStandards/c041521_ISO_IEC...(Bil).zip
6. http://standards.iso.org/ittf/PubliclyAvailableStandards/c043603_ISO_IEC...
7. http://standards.iso.org/ittf/PubliclyAvailableStandards/c043604_ISO_IEC...
8. http://wiki.fluidproject.org/display/ISO24751/AccessForAll+Working+Group
9. http://gpii.net/
10. http://wiki.gpii.net/index.php/Main_Page