Requirements specification iteratively combined with reverse engineering

A new concept of elicitation of requirements specification supported by an implemented CASE tool TESSI will be described in this paper. It concerns the problem of understanding, eliciting, and describing user requirements. The presented method consists of three steps. First, a semi-automatic transformation of the text of requirements into an object-oriented model will be supported. Second, a corresponding textual description will be automatically generated based on the just identified OO model. The generated text represents the analyst's understanding of the user's requirements. Third, the generated text will be read by the user and the domain experts in order to correct /complement it. A new version of requirements in the textual form will be created. This process will be iteratively repeated until there are no doubts about the analyst's good understanding of the problem. We argue that this approach which combines analysis and reengineering improves the quality of requirements specification.