Richard Griffiths - Lecture
Evaluating the Design
There are four approaches to this, though the methods may overlap:
The issues associated with each approach are discussed below.
The idea here is to introduce psychological theory into evaluation of a
design by checking identified psychological criteria during a "walkthrough".
The particular objective is to evaluate how well the designed software
supports the user in learning to use it. The walkthrough should be carried
out by someone with expertise in cognitive psychology as applied to interface
For a detailed description of the technique, see the tutorial paper
presented to CHI'95 by Rieman, Franzke & Redmiles, Usability
Evaluation with the Cognitive Walkthrough.
This technique, originally proposed by Nielsen and Molich, is not unlike
Cognitive Walkthrough, except that a set of usability heuristics ("rules
of thumb") rather than raw psychological theories are applied to the design.
Examples of high level heuristics are Schniederman's "Eight Golden Rules"
or Nielsen's "Ten
Usability Heuristics". Lower level guidelines are also available (and
may have been incorporated in the Design Style Guide produced or used on
a project). For a good example, see NASA's
HCI Style Guidelines.
These rules are due to Ben Schneiderman
For information on this topic from an expert in the field, see Jacob Neilson's
online material on Heuristic
Strive for Consistency
Enable Frequent Users to use shortcuts
Offer Informative Feedback
Design Dialogues to yield closure
Offer simple error handling
Permit easy reversal of actions
Support internal locus of control
Reduce Short term memory load
This approach attempts to prevent each designer "reinventing the wheel"
with each design. The cognitive psychology literature and interaction
design literature are searched to find analysis of similar designs, and
the results are then applied.
Review based evaluation
Building predictive models for evaluating interface design has been carried
out, for example, the GOMS
approach. Work is currently in progress to create more sophisticated user
models that interfaces can be tested against (a bit like a crash dummy
in automotive engineering!). An example of this approach is in the use
of the AI program SOAR as a programmable
user model (PUM).
The use of models
This page is maintained by Richard
Pages of individuals do not necessarily reflect the official
position of the University of Brighton.