In the mid-1980s, Digital Equipment Corporation was among the first software companies to define methods for usability engineering. During development of the MicroVMS Workstation Software for the VAXstation I, the usability professionals on the team defined a set of measurable user performance objectives to guide the development process. A central benchmark task was designed in which representative users created, manipulated, and printed the contents of windows. The usability objective was to reduce performance time on the benchmark task by 20% between version 1 and version 2 of the system. As development proceeded, measurements of users' performance on several related subtasks were made to identify areas of greatest usability concern in the design of version 1. The usability problems having the largest impacts on performance were used to prioritize changes and guide the redesign of version 2. In the end the team exceeded their objective, improving performance time on the benchmark task by 37%, while staying within the originally allocated development resources. Interestingly, however, measured user satisfaction for version 2 declined by 25% relative to version 1. (See Good, et al. 1986).