Empirical Studies conducted with software collections: a mapping study

JUAN ANDRÉS CARRUTHERS; JORGE ANDRÉS DIAZ-PACE; EMANUEL IRRAZÁBAL

SADIO Electronic Journal of Informatic and Operation Research

SOCIEDAD ARGENTINA DE INFORMÁTICA E INVESTIGACIÓN OPERATIVA

Lugar: Buenos Aires; Año: 2023 vol. 22 p. 82 - 107

1514-6774

Software projects are inputs in Evidence-Based Software Engineering, although they are selected without following a specific strategy, which decreases the generalization and replication of the results. One option is to use collections of existing projects, but these must have explicit construction and maintenance rules. The objective of this work was to perform a secondary study on software project selection strategies in empirical studies, and learn about: the rules consid-ered, project characteristics, code metrics, extraction tools and statistical analyses employed. A systematic mapping was used to identify articles from January 2013 to December 2021. We selected 150 studies in which 67% used their own rules for project selection and 31% worked with existing collections, and the majority (80%) used Java projects. Furthermore, there was no evidence of a standardized framework for project selection for empirical studies in Software Engineering.