Bill Thompson talks about Unasking the Right Questions in the latest E-Access Bulletin:

…we keep thinking about ‘accessibility’ and ‘usability’ as separate, almost orthogonal aspects of design. Unfortunately, this remains the dominant model, and it has now become a barrier to future progress because it encourages designers to think about creating tools and services for the ‘normal’ population before considering accessibility [...] Instead of thinking about ‘access’ at one end and ‘usability’ at another, we should attempt to recast our debate in terms of what technology does for all of us, not just those whose have ‘special’ requirements.

Fairly standard argument for properly integrating accessibility into design practices, but where I think the article gets more interesting is with this:

So how should we frame our debate if we move beyond what I think is a dangerous attempt to retain the distinction between ‘usability’ and ‘accessibility’? I think it is time to explore the idea of affordance’, as it could offer us a way forward. [...] Bill Gaver [...] wrote: “Affordances go beyond value-free physical descriptions of the environment by expressing environmental attributes relative to humans. For example, the physical measure of height, which has no inherent meaning, can be recast in terms of the affordance of accessibility, which does. Because accessibility emerges from the relation between elevation and people’s physical characteristics, it is an objective fact about a situation.”

[...] If we start to frame the issues facing users whose capabilities deviate from the norm in terms of affordances rather than simply of accessibility, this might free us from the ‘modal totalitarianism’ that infects so much design, whether in products like screens and keyboards or on-screen in websites, widgets and services.

Affordances matter equally to the ‘abled’ as to the ‘disabled’, and so the same design methods can be used, and outcomes can be evaluated in a much broader way. This allows us to start to move away from the current model, in which we have ‘assistive’ technologies to overcome ‘deficits’ that make some users ‘abnormal’, to one in which we all have skills and abilities that vary along a large number of axes.

Has anybody used this approach to reframe their requirements in terms of affordances instead of accessibility?