The impact of native language on use case modeling: A controlled experiment

Software development is a collaborative activity where the quality of the end product depends to a great extent on the quality of the requirements engineering process. Hence, the requirements engineering phases are crucial during the development of a software system. If done incorrectly, it may result in incomplete and inadequate system-to-be functionalities. The requirements engineering processes use natural language for communication with system stakeholders. Requirements ambiguity has been a key area of concern in overall software development process. An attempt to bridge the gap between a pure natural language approach to specifying software requirements and a formal language is the use case modeling approach. To the best of our knowledge, there is no reported evaluation of the impact of stakeholder requirements written in ones native language on the use case modeling approach. In this paper, we present a controlled experiment to investigate the usefulness of using native language (Arabic in this paper) system description on the use case modeling. The results show that using a native language for system description improves the functional correctness and reduces ambiguity related errors in a use case model. However, the time required to perform use case modeling is not affected by using either native or English (as a second language).