Automation of test case generation from textual use cases

Textual use cases are often used to describe system requirements and are believed to be a good basis for system testing. However, due to the lacking of formal syntax and semantics of textual use cases, it is hard to build automating process to generate test cases from textual use cases. In this paper, we propose a new approach for this automating process. The use cases are written in a restricted yet acceptable language style. An Extended Finite State Machine(EFSM) are extracted from text use cases, then test cases are generated from the EFSM. Statements coverage is adopted as the criterion to generate test cases. A Market Information System(MIS) is employed to demonstrate our approach. This method can offer us at least three benefits: 1) test cases can be automatically generated from textual use cases. 2) new test cases can be generated if requirements are updated. 3) debugging information can be derived.