A family of experiments to evaluate the understandability of TRiStar and i* for modeling teleo-reactive systems

•An i* extension for teleo-reactive (TR) systems named TRiStar.•A novel approach to modeling software requirements of TR systems using TRiStar.•An empirical proof of the higher efficiency of TRiStar vs i* for TR systems.•An empirical proof of the higher effectiveness of TRiStar vs i* for TR systems.

The teleo-reactive approach facilitates reactive system development without losing sight of the system goals.ObjectiveTo introduce TRiStar as an extension of i* notation to specify teleo-reactive systems. To evaluate whether the notational extension is an improvement in terms of effectiveness and efficiency over the original language when it is used to specify teleo-reactive systems.MethodA family of experiments was carried out with final-year engineering students and experienced software development professionals in which the participants were asked to fill in a form designed to evaluate the efficiency and effectiveness of each of the languages.ResultsBoth the statistical results of the experiments, analyzed separately, and the meta-analysis of the experiments as a whole, allow us to conclude that TRiStar notation is more effective and efficient than i* as a requirements specification language for modeling teleo-reactive systems.ConclusionThe extensions made on i* have led to TRiStar definition, a more effective and efficient goal-oriented notation than the original i* language.

Graphical abstractFigure optionsDownload full-size imageDownload as PowerPoint slide