THE P ROJECTIT-S TUDIO/R EQUIREMENTS CASE TOOL A Practical Requirements Specification Case Study David Ferreira 1 , Alberto Rodrigues da Silva 1 david.ferreira@inesc-id.pt, alberto.silva@acm.org 1 INESC-ID, Instituto Superior Técnico, Rua Alves Redol 9, Lisbon, Portugal Abstract: Adopting a pragmatic attitude, regarding to the high rate of unsuccess verified among IT projects, and also bearing in mind the importance of the software development process’s early stages to overcome this problem, we embraced the goal of creating a CASE tool for requirements specification and validation. Our approach is based on a controlled natural language designated as ProjectIT-RSL and seeks to achieve a higher level of rigor and quality of the produced requirements specifications. The supporting tool, the ProjectIT-Studio/Requirements plugin, belongs to a broader workbench called ProjectIT-Studio, whose main goal is to cover the entire software development life-cycle, providing guidance during the process according to software engineering best practices and methods. This paper presents a small but illustrative requirements specification case study, which emphasizes the current tool’s support and its inherent usage advantages during the requirements specification process. Keywords: Requirements. Specification. CASE tool. Controlled NL and NLP. 1. Introduction Despite the resemblance with the conventional engineering areas and the efforts made during the last few decades, the software development process is still an immature activity, presenting several critical issues that compromise its success, namely in terms of scheduling, budget, and quality of the software product delivered. The main cause to this situation is the non adoption of a standard and systematic approach founded on the best practices and methods proclaimed by the software engineering community (Dutta et al., 1999). Independently of the degree of formality adopted, requirements in their essence describe what the system should do or at least a constraint on its behavior (Sommerville and Kontonya, 1998), which is obviously critical for the success of the whole development process. Several surveys and studies (Standish Group International, 1994) have emphasized the costs and quality problems that can result from mistakes in the early phases of system development, such as inadequate, inconsistent, incomplete, or ambiguous