Evaluación en ingeniería global de sofware de herramientas de ingeniería de requisitos y de la reutilización de requisitos en lenguaje natural basada en catálogos

Resumo

ABSTRACT Even though requirements reuse and globalisation are two relevant issues, and the usefulness of requirements reuse has already been confirmed in several studies, to the best of my knowledge, there are no proposals addressing both natural language (NL) requirements reuse and Global Software Engineering (GSE) together. I believe that if adequate reuse-based mechanisms for natural-language requirements specification are defined, effectiveness and productivity of GSE projects could be greatly favoured. The hypothesis of this doctoral dissertation is thus as follows: If adequate catalogue-based reuse techniques and an automated tool support were defined for natural-language requirements specification when stakeholders are globally distributed, then relevant gains of effectiveness and productivity in software development projects would be obtained. Goals In order to test the previous hypothesis, the following overall goal was defined for this dissertation: Propose catalogue-based reuse techniques and an automated tool support for natural-language requirements specification in GSE, and validate them empirically. This general objective is broken down into the following specific goals: " Goal 1. Devise reuse-based techniques for natural-language requirements specification in GSE. " Goal 2. Provide automated support for the requirements reuse techniques by using a prototypical tool. " Goal 3. Validate the techniques and the prototype empirically in a GSE scenario. Methodology Empirical methods have become an integral part of research and practice in Software Engineering (SE). Empirical studies are usually classified into three groups: survey, case study and experiment. Two of these techniques have been applied in this doctoral thesis, namely survey and experiment. The survey allows describing a situation or phenomenon using data obtained through interviews or questionnaires administered to a representative sample of the population under study. On the other hand, the experiment provides a high level of control, as it enables manipulation of variables and measurement of their effects. The Systematic Literature Review (SLR) has also been used in this dissertation. It is a means to rigorously synthesise the scientific literature in relation to a question asked by the researcher. This technique allows characterising the evidence on a particular issue and helps identify gaps in the state-of-the-art, in order to establish the basis of a new research activity. Other research methods are more specific and fit very well to the research carried out in certain fields of knowledge. DESMET is a method designed for evaluating SE methods and tools. For this reason, it has been also applied in this doctoral thesis so as to evaluate RE tools. Conclusions Knowledge management in globally distributed settings is a challenging task. This thesis primarily focuses on a reuse-based Requirements Engineering (RE) method for GSE that specifies knowledge in the form of natural language requirements. The use of a repository is proposed to arrange into catalogues sets of interrelated reusable requirements, with the purpose of managing requirements knowledge. To the best of my knowledge, there are no other proposals which tackle both issues, GSE and NL requirements reuse, jointly. My key contribution in the context of the previously mentioned method lies in: (1) devising the requirements specification techniques; (2) developing the prototypical automated tool support, based on an extensive study of the RE tools market; and (3) performing the empirical validation of the proposal.