MARITACA: From Textual Use Case Descriptions to Behavior Models

It is well known in Software Engineering that the cost of eliminating requirements faults increases when these faults are revealed. Therefore, it is important to express requirements in a way that it would be possible to validate them as early as possible in the development cycle. Use Cases (UC) are a popular format to represent requirements. On the other hand, states machines are a notation amenable for verification, either using for example, model-checking or model animation, and also are a popular notation for model-based test case derivation. However, building a state machine from pure textual requirements is not trivial for most practitioners, in particular, a model with properties that allow them to be handled by a tool. UCs are described in pure natural language, in general, and the manual extraction of state machine models from them can be time-consuming and error-prone. In this work, we present an approach that helps the extraction of state machine models from UC descriptions. These descriptions are in the form of a semi-structured language, instead of pure textual form. The goal is to provide practitioners with a preliminary version of a state model that can be further refined to become input for tools like test case generators. As a proof of concept, a prototyping tool, MARITACA, was developed that uses Natural Language Processing techniques to extract the state machines. The text also shows a validation of the model obtained from the tool, to demonstrate the applicability of the approach.