Process Support for Requirements Engineering

Requirements engineering (RE) tools are software tools which provide automated assistance during the RE process. The need for automated support varies in different projects. RE-tool support could clearly be useful if an organisation deals with requirements specifications containing many requirements which need to evolve over time. However, the mainstream RE practice relies on office tools (e.g. text editors and modelling tools) rather than targeted RE-tools (e.g., CaliberRM, RequisitePro, and DOORS) provided by various companies and research groups. Reasons for not using the RE-tools include financial causes, such as high tool price and perceived low return on investment. The part of the problem also lies in the difficulty to evaluate such tools before acquisition to support the RE process. Hence, to support the completeness and effectiveness of RE-tool evaluation, a sound framework providing methodological guidelines to the evaluators is needed. This work proposes an RE-tool evaluation approach (R-TEA), which provides a systematic way of the RE-tool assessment using two evaluation frameworks for the RE-tools. Both frameworks contain lists of features which provide a structure for the RE-tool comparison and assessment. The framework for the functional RE-tool requirements consists of three dimensions: representation, agreement, and specification. The representation dimension deals with the degree of formality, where requirements are described using informal, semiformal and formal languages. The agreement dimension deals with the degree of agreement among project participants through communication means. The specification dimension deals with the degree of requirements understanding and completeness at a given time moment. The second framework categorises the non-functional RE-tool features to process, product, and external requirements. Process requirements characterise constraints placed upon the user’s work practice. Product requirements specify the desired qualitative characteristics of RE-tools. External requirements are derived from the user’s internal and external environment. Both evaluation frameworks are applied to a specification exemplar which application initiates preparation of the requirements specification for the RE-tool selection. The requirements specification contains RE-tool requirements which specify what the RE-tool should do, what characteristics, constraints and properties it should have. Assessment of the RE-tools’ compatibility to the specified RE-tool requirements is performed using different evaluation techniques. Decision about RE-tool selection is made after summarising all the assessment results. In comparison to the existing tool assessment approaches, the R-TEA method targets only the RE-tool domain and guides the user of the evaluation frameworks through RE-tool assessment. In comparison to the existing RE-tool frameworks the proposed evaluation frameworks provide a more complete and consistent RE-tool assessment by combining both functional and non-functional RE-tool features and by providing a comprehensive explanation of the terminology. iv A prototype tool is developed supporting the frameworks and R-TEA in order to facilitate the RE-tool assessment. The R-TEA method is tested in a number of case studies. The findings report on positive trends of the frameworks, prototype and the RTEA method. Furthermore, a number of RE-tool weaknesses is highlighted during the case studies. A fragment-based RE-tool prototype is implemented in order to emphasis the research efforts of the automated RE process support, and to investigate the REtool weaknesses in a design experiment environment.