Minerando conhecimentos de projetos de software a partir dos registros de comunicação de desenvolvedores

Abstract

Software development teams use collaborative communication channels, such as instant messaging (IM) tools and forums, to support software development and management tasks. When teams use such channels, relevant software-related discussions stay in these tools' log files. However, such relevant software-related discussions and decisions can get "lost," unimplemented, forgotten, duplicated, or hard to find amidst the large volume of messages. Lost and duplicated information can compromise the sharing and reuse of project knowledge. Motivated by the mentioned context and the hypothesis that one can extract relevant software-related knowledge from the developer's communication logs, this work aims to answer the following research question (RQ): RQ: How can we use automatic approaches to identify relevant software-related discussions from the developers' communication log files? Based on the Design Science Research methodology, we developed the framework Miner4DevTeam to answer this question. The framework helps stakeholders determine software project knowledge from team communication log files and supports the development tasks. Such knowledge can reveal the project development context. Our findings show that the determined knowledge can support product evolution, aid companies' strategic decision-making process, and support project knowledge sharing and reuse. Future work brings opportunities to enhance the framework, explore different data repositories, and support the knowledge management field.