Software Mining Studies: Goals, Approaches, Artifacts, and Replicability

摘要: The mining of software archives has enabled new ways for increasing the productivity in development: Analyzing quality, project evolution, investigating change patterns and evolution trends, models development processes, developing methods integrating mined data from various historical sources, or analyzing natural language artifacts repositories, are examples research topics. Software repositories include data, ranging source control systems, issue tracking artifact such as requirements, design architectural documentation, to archived communication between members. Practitioners researchers have recognized potential these sources support maintenance software, improve their architecture, empirically validate techniques processes. We revisited studies that were published recent years top venues engineering, ICSE, ESEC/FSE, MSR. In studies, we highlight different viewpoints: pursued goals, state-of-the-art approaches, artifacts, study replicability. To analyze (lexically) analyzed papers more than a decade. terms replicability looked at existing work field tools, platforms. address issues reproducibility shed light onto challenges large-scale would enable stronger conclusion stability.