Toward Linking Declined Proposals and Source Code: An Exploratory Study on the Go Repository

A new study proposes the first LLM-driven pipeline to create traceability links between declined feature proposals and the related source code in the Go repository. The system achieved 83.6% accuracy in selecting the correct code granularity and a mean precision of 64.3% in generating correct links. This approach uncovers valuable design rationales and implicit decision-making knowledge previously lost when proposals are rejected after discussion.