Automatic Comprehension of Textual User Requirements and their Static and Dynamic Modeling

Requirements engineering is the most important activity in software engineering, and is concerned with the gathering and understanding of user requirements written in natural language (NL). There is a gap between the textual description of the software to be developed, and the software UML models abstracting the static or dynamic views of the software. Our research is aimed at filling this gap by automatically extracting from text a more universal (graphical) hybrid model to present the variety of knowledge that can be found in textual requirements written in unlimited NL. The advantages of a hybrid diagram include its ability to illustrate, in a way which is graphical and simple to perceive, both the static and dynamic view of the system to be developed. The hybrid diagram serves as a source for an automatic inspection of the requirements text, and will provide timely feedback to the user on the actual content of the requirements he/she has written. Such feedback will enable correction and clarification of the text before it is delivered to the software developers, thereby helping them to avoid incorrect interpretation of the requirements. The hybrid diagram also allows for automatic generation of scenario-based black-box test cases for ensuring the final quality of the software.