Enhancing static source code search with dynamic data

  • Authors:
  • Reid Holmes;David Notkin

  • Affiliations:
  • University of Washington, Seattle, WA;University of Washington, Seattle, WA

  • Venue:
  • Proceedings of 2010 ICSE Workshop on Search-driven Development: Users, Infrastructure, Tools and Evaluation
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Developers frequently try to locate references to particular program elements within their systems; however, these queries often return an overwhelming number of results. The result sets for these queries tend to be large because integrated development environments locate matches using static search approaches; however, the developer may be more interested in which references actually happened for a particular execution, instead of which references could happen in a hypothetical execution. We posit that dynamic search approaches can complement customary static search approaches in the same ways dynamic analysis complements static analysis. Specifically, in this paper, we hypothesize that filtering static reference queries with dynamic trace data can reduce the number of results a developer must consider when performing a query, helping them to focus on a subset of the static query results. To test our hypothesis, we filtered the results of the Eclipse find references query with dynamic trace data for three different projects; our preliminary evidence demonstrates that dynamic trace data can be used to effectively filter the result sets of static source code queries.