A field study of the Wheel—a usability engineering process model

Abstract

Interactive system developers are increasingly including usability engineering as an integral part of interactive system development. With recognition of the importance of usability come attempts to structure this new aspect of overall system development, leading to a variety of processes and methodologies. Unfortunately, these processes often lack flexibility, customizability, completeness, and breadth of coverage. This paper describes our development of a meta process or process model that we call the Wheel. This innovative approach to creating and tailoring usability engineering processes addresses these shortcomings, and describes an evaluation of its application in a real-world commercial development environment. The Wheel process model for usability engineering is not a process itself, but instead, it provides a general framework into which developers can fit specific existing or new techniques, methods, or activities to apply “best usability practices”. It grew out of our examination, adaptation, and extension of several existing usability engineering and software methodologies. The methods that most strongly guided creation of the Wheel were the LUCID framework of interaction design, the Star life cycle of usability engineering, and the waterfall and spiral models of software engineering. The resulting process model assumes the form of a sequence of distinct cycles (each of which produces a product form), allowing developers to focus on each cycle separately. Each cycle has the same four activity types: Analyze, Design, Implement, and Evaluate. Each activity type in a cycle is instantiated using an existing usability engineering technique. Working with an Internet technology company in northern Virginia under grant sponsorship from the Virginia Center for Innovative Technology (a research and development incubator for the Commonwealth of Virginia), we instantiated the Wheel process model and used it to develop a Web-based device management system. The process model performed remarkably well for this development environment, overcoming the tight constraints of budget and schedule cuts to produce an excellent process instance that resulted in a demonstration prototype of the company’s target system.