An Ontology-based Approach for Developing Systems to Identify Who Knows What in Software Organizations

Context: Software development is a knowledge-intensive process, and its success in an organization relies deeply on knowledge sharing. Knowledge management challenges are often increased in agile environments, which involve a lot of tacit knowledge, commonly acquired through experiences and hard to make explicit. Thus, knowledge sharing among practitioners is crucial. However, identifying suitable experts to share specific knowledge is not trivial. It involves not only discovering the individuals with the desired knowledge but also considering other factors that may improve the expert’s responsiveness, such as social connections and availability. Moreover, the data overload problem and the specific nature of the experts’ knowledge can hinder people from finding experts with the required knowledge. Expert-finding systems address this issue by identifying and ranking experts. Developing such systems is challenging and involves complex concepts such as expertise, skill, and knowledge, among others. Additionally, integrating diverse heterogeneous sources that offer evidence of expertise poses a difficulty. Ontologies can be helpful in this matter by providing semantics to data and addressing semantic conflicts that can lead to data misuse or misinterpretation. Objective: We aim to help organizations develop expert-finding systems to identify who knows what in a software organization considering data stored in the organization repositories and non-technical factors that may affect knowledge sharing. Method:We developed iKnow, an ontology-based approach that supports developing systems that use data on skill manifestations to identify who knows what. Results: iKnow was applied to develop the ExpertFY system, which was used by members of a software organization. ExpertFY helped identify experts the members had not identified themselves. Conclusion: the results suggest that iKnow is a useful and promising approach to developing expertfinding systems and ExpertFY improves the identification of who knows what.