Methodology I: Task Analysis - Foundations for Designing User-Centered Systems - page 306

11.2.3 Task and Interface Design
TA can be used in several ways to guide task and interface design as listed below.
1. To predict user task performance. These predictions of time, errors, or work-
load can be used to compute whether an interface can be used to perform a
time-sensitive task (e.g., interact with a video game), or how many copies of an
interface are needed to support a number of users (e.g., the minimum and the
optimal number of ATMs or ticket machines in an airport), or when a system
needs to be speeded up to keep pace with the user. This approach has been and
is being used to guide the requirements of how many sailors are needed to run a
ship (Chipman and Kieras 2004 ).
2. To suggest where users will make mistakes. Furthermore, we can take two
types of task analysis and look at their relationship—for example, is the rela-
tionship between goals/subgoals and the chronology of actions simple or
complex? Are similar tasks similar? Are common safe tasks easy to do and
expensive or dangerous tasks harder to do? Where there is a complex rela-
tionship between the users' goal structure and the interface's action structure,
then an interface may be hard to use and may lead to more errors than it should
or could. This approach has been applied to tasks in aviation. Models are built
of how ATC interacts with pilots (Freed and Remington 1998 ) or how ground
control navigates planes (Byrne and Kirlik 2005 ). The results suggest ways to
avoid errors and ways to mitigate the effects of errors.
3. To understand the relationship between old and new versions of a system.
Given a task analysis of how people currently do a task, or would conceive of
doing the task, we can ask how much that process and knowledge overlaps with
the formal analysis of how it should be done in the new system. Increased
similarity of how to use both systems can help with efficiency of learning, use,
and comfort, and can lead to greater acceptance of the new system. The original
work in this area (Bovair et al. 1990 ; Kieras and Polson 1985 ) noted the
knowledge to use an initial editor, and then compared that to the knowledge to
use a new editor. The results provide several suggestions, including: that
ignoring old knowledge is easy; that modifying existing knowledge is, of
course, slower; and that learning new knowledge slows the user down the most.
4. To compare different designs. A task analysis of two interfaces can predict
which interface will be faster and easier to use, as long as the two designs are
reasonably distinctive. In most cases, however, the comparisons are accurate
and useful. Gray et al. ( 1992 ), for example, used TA to compare the interface
for a new system that was to be used by telephone toll operators—the people
who used to respond when you dialed 0—against the old system's interface.
The new interface was predicted to be slower, and an empirical test confirmed
this. It also showed that the effect of the difference in performance would cost
the telephone company about $2 million per year.
TA has been used to investigate the design of menus using real and simulated
cell phones (St. Amant et al. 2004 , 2007 ). In this project a user model was