An essential software development tool, code search engines are expected to provide superior accuracy, usability, and performance. However, prior research has neither (1) summarized, categorized, and compared representative code search engines, nor (2) analyzed the actual expectations that developers have for code search engines. This missing knowledge can empower developers to fully benefit from search engines, academic researchers to uncover promising research directions, and industry practitioners to properly marshal their efforts. This thesis fills the aforementioned gaps by drawing a comprehensive picture of code search engines, including their definition, standard processes, existing solutions, common alternatives, and developers' perspectives. We first study the state of the art in code search engines by analyzing academic papers, industry releases, and open-source projects. We then survey more than a 100 software developers to ascertain their usage of and preferences for code search engines. Finally, we juxtapose the results of our study and survey to synthesize a call-for-action for researchers and industry practitioners to better meet the demands software developers make on code search engines. We present the first comprehensive overview of state-of-the-art code search engines by categorizing and comparing them based on their respective search strategies, applicability, and performance. Our user survey revealed a surprising lack of awareness among many developers w.r.t. code search engines, with a high preference for using general-purpose search engines (e.g., Google) or code repositories (e.g., GitHub) to search for code. Our results also clearly identify typical usage scenarios and sought-after properties of code search engines. Our findings can guide software developers in selecting code search engines most suitable for their programming pursuits, suggest new research directions for researchers, and help programming tool builders in creating effective code search engine solutions. / Master of Science / When developing software, programmers rely on source code search engines to find code snippets related to the programming task at hand. Given their importance for software development, source code engines have become the focus of numerous research and industry projects. However, researchers and developers remain largely unaware of each other's efforts and expectations. As a consequence, developers find themselves struggling to determine which engine would best fit their needs, while researchers remain unaware what developers expect from search engines. This thesis address this problem via a three-pronged approach: (1) it provides a systematic review of the research literature and major engines; (2) it analyzes the results of surveying software developers about their experiences with and expectations for code search engines; (3) it presents actionable insights that can guide future research and industry efforts in code search engines to better meet the needs of software developers.
| Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/111266 |
| Date | 15 July 2022 |
| Creators | Li, Shuangyi |
| Contributors | Computer Science, Tilevich, Eli, Gulzar, Muhammad Ali, Servant Cortes, Francisco Javier |
| Publisher | Virginia Tech |
| Source Sets | Virginia Tech Theses and Dissertation |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | ETD, application/pdf |
| Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
Page generated in 0.0019 seconds