Investigating past and present code reviewer recommendation systems

Abstract

Context: Selecting a code reviewer is an important aspect of software development and depends on several factors. Objectives: The aim is to understand existing solutions for code reviewer recommendation systems (CRRSs), factors to be considered when building them and various dimensions based on which they can be categorised. Our goal is to understand important features of CRRSs and what can be improved in existing CRRSs. Methods: A literature review study was conducted to understand the existing CRRSs. A survey of software development project members was conducted to understand the important and missing features in CRRSs. Results: We categorized the selected papers into two categories: based on the data type used to make recommendations and the kind of project used for evaluation. The survey helped us understand the features missing in CRRSs and observe some trends and patterns.