Effects of Interface Format, Feedback Style, and System Lag on the Usability of Hand-Held Internet Controllers

The increasing popularity of the World Wide Web (WWW) has created a new market: Web access through television to accommodate those who either cannot afford existing hardware or are intimidated by computers. Current efforts to combine the WWW and television have targeted potential novice users. One of the approaches for creating a WWW browsing system that is both simple to use and inexpensive is the utilization of the existing cable system to provide Web access through television. Some unique characteristics of this browsing system are fast access to the Web, the use of nine buttons on a universal remote control, and an index structure for reaching Web sites. Since browsing the Web through television is relatively new, many interface issues need to be examined. The purpose of this research was to investigate potential user interface designs for this WWW browser and to evaluate the usability of the nine-button interface.

Sixteen volunteers participated in the experiment and were asked to navigate to specific Web sites with two interface formats, five different system lag times (0.2, 0.7, 1.3, 2.0, and 3.0 seconds), and three feedback styles (active feedback, passive feedback and no feedback). Participants were prescreened for their experience with computers and browsing the Web. The experiment was conducted in a living room setting to simulate real life situations and participants were given a total of 42 tasks to complete throughout the experiment. Each task consisted of navigating through the tree structure with either one of the two interface formats until reaching a designated Web site. The number of errors committed and task completion times were recorded. In addition, participants were asked to rate the WWW browser system after each task as well as after the entire experiment.

Participants preferred the 0.2 second system lag and the active feedback style. Overall, they committed fewer errors and took less time to complete tasks with the tabbing interface than with the one-to-one mapping interface. Experienced participants committed more errors than did the inexperienced ones. Increase of system lag time was determined to have a greater adverse effect on the tabbing interface than on the one-to-one mapping interface.