The generation of rules for intelligent help based on task analysis

Abstract

Users need a certain level of skill or knowledge in order to be successful in the performance of human computer interaction. Many software systems have been designed to assist in using an interface. These packages are not only expensive but restricted to specific applications, i.e., they provide help in the environment for which they were created. This paper deals with the design and development of a training tool, the Generic Rule Based Training System (GETS), which provided the proper connection with an application system, helps the user in the utilization of the interface. As input, the training system requires a User Action Notation (UAN) description of the steps to be performed in order to accomplish a given task. The descriptions are translated into rules and transferred to an expert system tool which supplies information on the user's state, including, among others, completed tasks, remaining steps to complete a task and recommendations for future steps. The system's abilities are demonstrated by its implementation on the Hopkins Ultraviolet Telescope's graphical interface.