Methodology II: Cognitive Dimensions and the Gulfs - Foundations for Designing User-Centered Systems - page 334

difficult—most people can usually handle one level of indirection, but multiple
levels of pointers tend to be more difficult to follow. Another type of hard mental
operation is boundary problems, for example, counting fence posts—If you have a
ten-foot fence to put up, and fence pieces are one foot long, how many posts do
you require? In this situation, the boundaries of an object have to be explicitly and
carefully represented and manipulated. If you thought there should be 10 posts,
you have left out the first post, because there needs to be 11 (you can see this much
more easily if you actually draw the fence).
An important thing to remember about these kinds of hard mental operations is
that they can be easy to implement computationally, but can be especially trou-
blesome for users. Again, these problems can be solved at several levels, including
either by avoiding the problem by understanding the relative difficulty of opera-
tions, or by providing tools to assist in these operations (e.g., graphic displays of
indirection in C programs) and providing representations designed to be shared
between types of system implementers.
12.3 Turning Cognitive Dimensions into a Methodology
The CDs can be used as a method for informing design during the early or middle
stages of development (e.g., Cohen et al. 2012 ). The designer can ask questions
based on the appropriate dimensions for the system, such as:
1. Are there hidden dependencies and one-way links in the structure?
2. Is the system viscous, or can users and designers reconstruct structures easily
and fluidly?
3. Is the order of generation and action the same as the order in which users will
want to do things and have the appropriate information? Are there multiple,
appropriate orders of behavior and strategies supported?
4. Is the system role-expressive? Are the functions of the components obvious and
discriminable?
5. Does the system require any hard mental operations?
6. Are there conceptual gaps between the user and the system? (Blandford et al.
2008 )
The answers to these questions are shared among the design team. They can
then be used to moderate the design of the system, or may lead to other design
activities such as studying the relative costs of mental operations or other tasks that
help reduce the risks associated with the design.
Designers might wish to order the dimensions they use, and they may also wish
to add other dimensions. A new dimension could be added, for example, to con-
sider the role of cognitive load (Sweller 1988 ) describing how much the user has to
know, do, and learn at any given point in a task.
The CDs have been used to develop extensions to the Microsoft Excel
spreadsheet that integrate user defined functions into the worksheet grid (Peyton