Abstract
The purpose of this study was to investigate the effects of the format used to display on-line help (full screen, split screen and windowed) on user performance and attitudes. Three prototypes of a programmer's editor were developed. The first prototype provided help in full screen format. The second prototype provided split screen help and the third windowed help. It was hypothesized that user performance would be best with windowed help, and that attitudes would be most positive toward windowed help. In addition it was believed that windowed help would fare best because it would allow users to see and work with about two-thirds of the product screen while the help was displayed. Forty-five application programmers participated. Each subject performed three editing tasks using one form of on-line help. No significant differences in performance or attitude were found between those who used full screen, split screen and windowed help. However, comments made by subjects during the assessment suggested that the nature of the help text itself may have affected the results. An examination of these comments indicated several ways in which the organization and writing style of the help text could be improved.