Behavior: Basic Psychology of the User - Foundations for Designing User-Centered Systems - page 102

may be quite different. In the same way, our ability to perceive objects displayed
on computer screens, for example, whether they are text or graphics, 2D or 3D
representations, is a result of our prior knowledge or our expectations as to what
should appear as well as what lands on our retinas.
In the Ecological approach (e.g., Gibson 1979 ), the process of seeing is greatly
influenced by what uses (affordances) the object perceptions suggest. This
approach, which is also called direct perception, stems from work with aircraft
pilots. It takes as its starting point not a retinal image that is passively sensed, but
the ambient optical array that the observer samples. Perception and actions are
seen as tightly interlocking and mutually constraining.
In the Active Vision approach, Findlay and his colleagues (e.g., Findlay and
Gilchrist 2003 ) build on the Ecological approach, but emphasize the role of
cognition in controlling vision. This view opposes the idea of just taking scenes
and analyzing them, and instead incorporates the notion of interacting with scenes.
In interacting, the eye (and brain) chooses where to look next based on what has
been focused on before and what was seen in those locations.
4.5.6 Implications for System Design
If you are designing a system where the three-dimensional positioning of items on
the display is important, you will need to decide which perceptual depth cues you
should use. The final choice is likely to depend on the context in which the display
screen will be used, as different lighting conditions can affect the way that the
displayed image is perceived, so if you only use narrow and quite pale shadows,
for example, these may not be very obvious in brightly lit conditions.
For critical systems, if the user has to undertake a task which involves
responding to the number of target items that appear on the display screen, you
will need to consider how to display those target items. If the number of items is
very low, the users will be able to react very quickly. If the number is very large,
however, you may need to think about organizing them in some way (using the
Gestalt principles, perhaps), or even splitting them across displays to optimize the
response time.
At present, most display screens offer one level of resolution, such that all items
on the display are presented with that resolution irrespective of their importance. If
you are designing a system which requires an increased level of detail for items
that are currently being used (perceived) then you may want to consider the
possibility of using gaze-contingent multi-resolution displays (Reingold et al.
2003 ). These displays maximize the resolution at the part of the screen where the
user is currently looking. This approach can provide a much more detailed image
for a given bandwidth because more pixels (and information) are allocated to what
the user is currently looking at, rather than effectively wasting pixels by providing
an unnecessary level of detail for items that the user is not currently looking at.