Continued from p. 3

The Challenge: Ensuring Usability

Because of the aggressive development schedule, EIRS was designed and assembled quickly, making heavy use of open-source software and ideas that team members brought into the project from their experience.

User interface and usability experts joined the project only after design and development had already begun.  This ruled out a conventional front-loaded user-centered process.  The early EIRS team did start with requirements analysis and use-cases, but developed minimal conceptual design or UI design specifications prior to starting implementation.

The UI and usability team usually found itself in the position of playing “catch-up”:  trying to ensure the usability of an application for which design was being done mainly by programmers, and on which implementation was proceeding rapidly.  Therefore, usability evaluation played a more prominent role in ensuring EIRS’ usability than did conventional up-front interaction design methods.

However, even conventional evaluation methods [4] were impractical.  This was due to: a) the frantic pace of the project, b) the low budget, and c) the fact that the only call centers set up before election day were 3000 miles from the usability team.  These factors ruled out even one round of formal usability testing prior to EIRS’ deployment in the election.  Normally, that would be regarded by HCI professionals as a recipe for disaster, especially for such an important, high-profile application.  However, the usability team understood that the methods we typically use in developing commercial systems were infeasible.  We adapted.  We took advantage of any opportunities to assess and improve EIRS’ usability.  We minimized our impact on the project by piggybacking on already-planned activities.

-> Continued on p. 5: Evaluation Methods Used