Recomendação de desenvolvedores externos para projetos de software baseada na análise de contribuições prévias

Abstract

The software development industry has evolved in the recent years and new challenges have emerged. Among these changes came Software Ecosystems, a new development paradigm, where external contributors support software production by providing solutions that complement a common platform for these developers. Due to the large number of technologies, frameworks and domains that an ecosystem can host, an equally large number of contributors acquainted with varied topics of their knowledge and skills have also emerged. However, recruiting collaborators with desired characteristics becomes a complex task due to the varying degrees of knowledge and skill that each developer has in their various competencies. Given this, we present a architecture of a recommendation system (RS) supported by an ontology capable of recommending collaborators who have shown expertise in the topics of interest. In order to do so, the RS uses retrieval expertise techniques to score the developers´ level of knowledge about topics represented in a query. The architecture is then able to provide the contextual information of the recommendation, i.e., a visualization of where one can find the knowledge topics that led to the recommendation of each contributor. Proof of Concepts were conducted on two software ecosystems to demonstrate feasibility of the architecture, which have shown evidence that the architecture is able to perform recommendations and still offers context information, important to the decision-making process over the recommendations made.