where any time” part will require sys-temwide infrastructure of the kind pursued by the GPII. Ability-aware operating systems infused with SUP-PLE-like user-interface generators could help create personalized applications. Improved sensing and modeling of users’ abilities and contexts, as in walking user interfaces, could enable mobile and wearable systems to better support diverse contexts of use. One challenge is to avoid explicit task-based training and calibration in favor of implicit observation and modeling from everyday use, as in Evans and Wobbrock5 and Gajos et al. 8

To date, ability-based design has focused primarily on single-user experiences, but the social lives of users could also lend themselves to collaborative support. How should the abilities of a pair, group, team, crowd, or organization be considered? For service arrangements, what would it look like to have an ability-based design for services?

Moreover, abilities exist on many levels, from low-level sensorimotor and cognitive abilities, to mid-level abilities for daily living, to high-level social, occupational, professional, and creative abilities. Such abilities form a hierarchy paralleling Maslow’s hierarchy of needs, 20 whereby each need corresponds to an ability to meet it. Ability-based design seems applicable throughout such a hierarchy, but the range has yet to be explored.

Concerning “adaptivity,” providing each individual with a unique user interface raises several pragmatic issues, as in, say, authoring help documentation, provision of customer support, and making the design process of personalized experiences consistent with accepted design practice. These challenges are real but, as we discuss elsewhere, 9 solvable.

With the vast range of human
abilities from which to draw, adaptiv-
ity based on sensing and modeling
is a powerful way to realize custom
designs that, while inevitably imper-
fect, nonetheless provide good user-
system fits at scale. Adaptive interfac-
es can remember users’ abilities and
preferences and draw on them when
generating interfaces for both famil-
iar and unfamiliar systems, providing
more satisfying and effective access
to achieve each best-fit interface. For
example, accessibility features locat-
ed in five layers—operating system
features, installed AT, browser fea-
tures, cloud AT, and Web app fea-
tures—can be configured to work to-
gether to provide best-fit user
interfaces, with features at each level
being invoked (or not) in order to meet
the user’s needs and preferences.

GPII auto-personalization supports interfaces that self-adapt, as well as configuration of interfaces and adaptations, to match a user’s needs. By combining auto-adjusting interfaces, preference-configured interfaces, and user-selected-and-configured AT, the GPII can function as a bridge among these approaches, maximizing the utility of each one for an individual at any point in time. The GPII also supports auto-configuration based on contextual changes. 40 The GPII thus meets all seven principles of ability-based design.

Taking Up the Challenge

Pursuing these and other projects, some patterns have emerged for us. For example, we noticed a perspective shift as we began to actively seek out the abilities people have, inspiring an openness to consider how we could create or change technologies to suit different abilities. We also noticed a seamlessness between designing for people with limited abilities and designing for people in ability-limiting situations. We realized accessibility is indeed a worthy goal for all users. Because we were looking to modify systems, not users, we deem-phasized assistive hardware add-ons. Customization arose from a powerful sequence of sensing, modeling, and adapting; it also arose from support for end-user configurability, as with the U.S. Air Force cockpits mentioned earlier. We thus made our interactive systems more aware of their users and contexts.

Where does ability-based design go next? One way to answer is to treat the vision of ability-based design as a grand challenge and ask what it would take to create a world in which anyone, anywhere, at any time could interact with technologies that are ideally suited to his or her situated abilities. Achieving the “anyone any-

What would it take
to create a world
in which anyone,
anywhere,
at any time
could interact
with technologies
that are ideally
suited to
his or her
situated abilities?