Dictionary-based query recommendation for local code search.

Local code search tools help developers efficiently find code snippets in the code base under development. The quality of the retrieved code largely depends on the quality of queries provided by developers. Manually synthesizing high-quality queries is a non-trivial task, partially because it places a cognitive burden on developers by requiring them memorize words used in the code base under search. To help developers synthesize better queries, this paper proposes a recommendation technique called MultiD that uses multiple dictionaries. We also report an ongoing study to evaluate the effectiveness of MultiD.