Background: Software engineering is a knowledge-intensive activity that requires engineers to manage information to collaborate efficiently and effectively. Within Software Engineering, the Requirements Engineering process bridges the knowledge gap between the customer and the development team by eliciting, managing, and communicating product requirements. The traceability of these requirements supports developers in producing higher-quality software that aligns with customer needs. In addition, traceability supports other activities, such as change impact analysis, software quality assurance, and requirements-based verification.

Problem: Despite decades of research on traceability, practical challenges still hinder the adoption of traceability in practice. This signals a need for new ways of practicing traceability that fit real-world needs. 

Goal: Building on previous work, this thesis instantiates, develops, and empirically evaluates Taxonomic Trace Links, a new way to trace requirements to various software artifacts through domain knowledge captured in a taxonomy. 

Method: The studies included in this theses follows mixed research methods, which are case study, systematic mapping studies, validation study, controlled experiments, and focus groups.

Results: The current state of practice in customer-supplier communication shows persistent challenges that we mapped to solutions in the literature. Our literature study shows that traceability through domain-specific taxonomies has not been empirically evaluated. Our development and evaluation of the technical solution for taxonomic trace links show that semi-automation of trace link creation and maintenance is possible. Finally, our empirical evaluation of taxonomic trace links shows that the solution is feasible in practice and can create trace links for multiple purposes.

Conclusion: Traceability between software artifacts has more benefits than currently realized by practitioners. However, current traceability solutions, based on direct trace links, do not appear to be easily adapted in different scenarios to trace different artifacts. Taxonomic trace links are an alternative approach that could overcome the shortcomings of direct trace links.