Method of context-dependent assistance for software user solving an applied task

Abstract

When using software to solve the applied tasks, the problem to implement the action in need by means of user interface arises. Partly, this problem is solved by studying reference manuals and consultations with the application developers. However, reference manuals are structured in the context of overall application functionality while difficulties arise in the data state context of the task being solved. Consultations lack this flaw, but they are costly and not always available in time. This fact stimulated the development of computer-based methods of contextual help. Yet the task of development of recommendations for performing the requested operation from the current state of the application data has not been solved so far. The research is aimed to reduce the time to get help by developing a knowledge representation model that determines user actions across the application data context, and a method for deriving model based recommendations. The model of the user action scenario is presented in the form of a colored Petri net. This decision content is based on the analogy between user action scenarios and workflow scenarios, for which Petri nets notation has been successfully used for years. For the topological analysis of the Petri net, the strategy of exhaustive depth-first search was applied. The method of the contextual recommendations is proposed to execute the operation requested by the user based on a scenario model in the form of a colored Petri net. The method novelty is application of topological analysis of the Petri net to construct a set of alternative scenarios for performing the operation, followed by filtering alternatives in the process of stepwise execution of the recommended actions. The suggested method provides context-dependent assistance in just one click. When using traditional reference manuals and files for the purpose, the number of clicks is determined by the number of options available to perform the operation.