<p>Reports of software a development projects that miss schedule, exceeds budget and deliver products with poor quality are abundant in the literature. Both researchers and the industry are seeking methods to counter these trends and improve software quality.</p><p>Software Process Improvement is a systematic approach to improve the capabilities and performance of software organizations. One basic idea is to assess the organizations’ current practice and improve their software process on the basis of the competencies and experiences of the practitioners working in the organization. A major challenge is to create strategies and mechanisms for managing relevant and updated knowledge about software development and maintenance. Insights from the field of knowledge management are therefore potentially useful in software process improvement efforts to facilitate the creation, modification, and sharing of software processes in any organization.</p><p>In the work presented in this thesis, we have made an overview of empirical studies on the effect of knowledge management in software engineering. We have categorized these studies according to a framework and we report findings on the major concepts that have been investigated empirically, as well as the research methods applied within the field. We have also investigated two main strategies for knowledge management, codification and personalization, through the application of four concrete methods in a software process improvement setting: Mentoring, Rational Unified Process, Process Workshops and Post Mortem Analysis.</p><p>We have classified the work in this thesis within three main themes:</p><p>RT1: Previous research on knowledge management in software engineering.</p><p>RT2: Application of knowledge management to improve the software process through codification of knowledge.</p><p>RT3: Application of knowledge management to improve the software process through personalization of knowledge.</p><p>The main contributions are:</p><p>C1: An overview of the research literature on empirical studies of knowledge management in software engineering.</p><p>C2: A method for tailoring the Rational Unified Process to the development process of a software consulting company.</p><p>C3: Improvements of the Process Workshops method by contextualization.</p><p>C4: Improvement of the root-cause analysis phase of the lightweight Post Mortem Analysis for more effective project retrospectives.</p><p>C5: Proposed methods to increase the learning effect of mentor programs in software engineering.</p>

Reports of software a development projects that miss schedule, exceeds budget and deliver products with poor quality are abundant in the literature. Both researchers and the industry are seeking methods to counter these trends and improve software quality. Software Process Improvement is a systematic approach to improve the capabilities and performance of software organizations. One basic idea is to assess the organizations’ current practice and improve their software process on the basis of the competencies and experiences of the practitioners working in the organization. A major challenge is to create strategies and mechanisms for managing relevant and updated knowledge about software development and maintenance. Insights from the field of knowledge management are therefore potentially useful in software process improvement efforts to facilitate the creation, modification, and sharing of software processes in any organization. In the work presented in this thesis, we have made an overview of empirical studies on the effect of knowledge management in software engineering. We have categorized these studies according to a framework and we report findings on the major concepts that have been investigated empirically, as well as the research methods applied within the field. We have also investigated two main strategies for knowledge management, codification and personalization, through the application of four concrete methods in a software process improvement setting: Mentoring, Rational Unified Process, Process Workshops and Post Mortem Analysis. We have classified the work in this thesis within three main themes: RT1: Previous research on knowledge management in software engineering. RT2: Application of knowledge management to improve the software process through codification of knowledge. RT3: Application of knowledge management to improve the software process through personalization of knowledge. The main contributions are: C1: An overview of the research literature on empirical studies of knowledge management in software engineering. C2: A method for tailoring the Rational Unified Process to the development process of a software consulting company. C3: Improvements of the Process Workshops method by contextualization. C4: Improvement of the root-cause analysis phase of the lightweight Post Mortem Analysis for more effective project retrospectives. C5: Proposed methods to increase the learning effect of mentor programs in software engineering.

Orientador: Antônio Francisco Magnoni / Resumo: A presente pesquisa tem como objetivo principal gerar um registro e uma análise do processo criativo da versão alpha do jogo digital independente (indie game) Lux Ex, que foi produzido quase que inteiramente por um único desenvolvedor, de forma autoral e adotando o modelo de produção independente conhecido como “Lone Wolf Developer” (Desenvolvedor Lobo Solitário). Tal modelo de produção possui particularidades que podem agregar ao produto final um caráter mais autoral, porque todas as suas etapas de realização são marcadas por um processo criativo pouco linear e entrelaçado com uma produção altamente interdisciplinar. Assim, o jogo apresenta uma complexa rede de significados para expressar conceitos e ideias emergentes da subjetividade de seu criador, valendo-se de sistemas simbólicos já estabelecidos. O objeto de análise da dissertação que aqui apresentamos, é o processo criativo completo de um jogo digital independente. / Abstract: The present research has as main objective to generate a registry and an analysis of the creative process of the alpha version of the indie game Lux Ex, that was produced almost entirely by a single developer, in an authorial way and adopting the model of production independently known as the "Lone Wolf Developer". Such a model of production has particularities that can add to the final product a more authorial character, because all its stages of realization are marked by a creative process not linear and intertwined with a highly interdisciplinary production. Thus, the game presents a complex network of meanings to express concepts and ideas emerging from the subjectivity of its creator, drawing on already established symbolic systems. The object of analysis of the dissertation presented here is the complete creative process of an independent digital game / Mestre