Summary: Putting It All Together - Foundations for Designing User-Centered Systems - page 382

below), or Soar (Laird 2012 ), for example, to work your way through their
hypothesized modules to understand how behavior will be generated. At this
level of abstraction, there is a fair amount of agreement between the theories in
that they all include modules for input, memory, cognition, and output.
2. To summarize user behavior. In time, architectures may be used more often
during design as a way to conveniently capture theories of user behavior, and as
a teaching aid to help designers understand users (Pew and Mavor 2007 ).
3. To apply what we know about users. In some cases, the models only offer fairly
crude approximations of users, but they have been useful for populating computer
games and military simulations. In other cases, they are being used to test inter-
faces and make predictions for designs (for a review, see any of these reports:
Booher and Minninger 2003 ; Pew and Mavor 1998 , 2007 ; Ritter et al. 2003 ).
14.3.2 Types of User Models
There are now more than 100 cognitive (and user) architectures if you include
variants and different versions (Morrison 2003 ; Pew and Mavor 1998 ; Ritter et al.
2003 ). They can be categorized into four types:
1. Implicit descriptive models. When you look at a car, for example, you can
imagine some of the assumptions the designers have made about the driver.
These assumptions are captured in an implicit model of the driver in which
vision takes place in the top part of the body, and another part of the body
operates the controls on the floor.
2. Explicit declarative models. These models describe the components or a
structure in a system, but neither describe the mechanisms nor perform the task.
3. Explicit process models. The mechanisms and the way they operate are
described, but the models, while perhaps supported by software, do not process
information.
4. Explicit information processing models. These models include a full informa-
tion processing architecture that produces behavior, and can predict perfor-
mance times and the information processing steps that will be performed and
their results.
14.3.2.1 Implicit Descriptive Models
Implicit user models appear in many systems, and some people would claim that
all systems include a model of the user. For example, chairs assume a height of the
user, and many file systems assume users can read and write English. Taken
together, the set of assumptions used by the designer to create an artifact is an
implicit model of the user, but it may be a particularly impoverished, incomplete,
or incorrect model. There are certainly several tools and approaches that include