A less "high falutin" term for cognitive issues in user interface design is "Human factors for Interactive software". The central issue in this course is the desing of interfaces for use by people who may (will) not be programmers. The goal is to make the computer transparent to the user, just as a pencil is transparent to its user.
The best way to achieve this transparency is to understand the user's own model of what the task requires and incorporate that into the user interface. One way to capture this idea is through the syntactic-semantic object action model. In a real sense this model captures the essence of a large part of symbolic reasoning. The goal of symbolic reasoning is to "drive the semantics into the syntax" (in quotes because this is borrowed from Shapiro's injunction to drive the tester into the generator). Computers work in the domain of syntax, but actions are about semantics. A useful system embeds correct actions in a syntax that is easy to grasp, precisely because it carries the semantics.
Understanding how the prospective user sees the task and the environment in which they work is therefor the beginning of the process. There is a great body of research on user modelling. It is primarily and necessarily empirical, although there is also a considerable body of speculative writing in the field. To vastly oversimplify we will work on the assumption that people organize their perceptions in object oriented terms. (See Lakoff's book, "Women, Fire and Dangerous things" for one view of how people conceptualize.) I am not claiming that this is congnitively accurate. I do claim that it is useful.
If we accept this view then it is natural to think in analogies. Indeed one is often introduced to an unfamiliar concept by having an explicit analogy drawn to a familair one. A favourite is to introduce the discussion of electricity by likening the flow of electricity through a wire to the flow of water through a pipe. Sometimes, as here the same word is used in the two contexts so as to reinforce the analogy. In the personal computer world analogies are drawn between directories and file folders and between file systems and file cabinets. Many PC users are not even ware that there is something called a file system. When wisely chosen and carefully used these analogies are helpful to users.
Sometimes the analogy that a user constructs serves to mislead her about the expected behaviour of the system. Dragging a disk icon to the trash can to eject the disk might suggest that the disk will be rendered useless (trash).
It is wise therefor to rely on familiar anlogies, rather than trying to construct special ad hoc ones for a particular application. One familiar analogy that is often useful is that of filling out a form. Less natural but often familiar to users who have more familiarity with computers (particularly from pre GUI days) is the command language.
In designing an interactive system it is important to have a method for understanding user's views of the task.
Return to GUI home page.
Last Changed: 17th April 1995