Logical model of stepwise contextual help for CAD user

Automatic stepwise contextual help for CAD systems users reduces time to solve the application task since it saves time to search the prompt message in system documentation. Petri nets (PN) can be used to bind available actions of the user considering the state of application data (context). Application of the Petri net inversion method, which uses limited enumeration to construct chains of recommended actions is preferable than using the standard reachability analysis procedure based on exhaustive enumeration. However, the absence in the known implementations of an explicit separation of the axioms of inversion (knowledge) from the mechanism of their processing (inference) deprives the stepwise contextual help system of the necessary flexibility when changing the axioms to consider the assumptions associated with a particular model. Thus, the aim of this research is to provide the necessary flexibility of the contextual help system by separating the knowledge representation model from the inference engine. A colored PN is used as a model of user action scenarios. The inversion axioms are implemented in the PROLOG language. The standard inference engine of the PROLOG language is used as a tool to construct chains of recommended actions. The authors have proposed an axiomatic model of PN inversion and a method to construct a stepwise contextual help by the standard inference engine of the PROLOG language. The method differs by explicit separation of knowledge (inversion axioms) from the inference engine (stepwise recommendations). It reduces the computational costs of adapting the contextual help system when changing the inversion axioms. The proposed method allows to reduce the time spent on adapting the contextual help system, since the field of the changes is limited by the declarations of the inversion axioms. The reliability of the results is confirmed since the proposed method of contextual help is used for the CAD “Model and Archive” user of CSoft company. The results obtained allow creating contextual help services for existing applications with minimal changes to their code base.