Search tools for scaling expert code review to the global classroom

Abstract

This thesis aims to answer the question "How can teachers of online classrooms give more qualitative feedback to students?" We narrow the scope of this question to an online software engineering class in which a major component is code review. We built two search tools that give teachers better coverage of student code. The first tool, Comment Search, allows students and staff to reuse any comment they previously wrote when reviewing another student's code. Staff can reuse any comment written by any staff member as well. After deploying Comment Search in a classroom for a full semester, we found that students and staff used this tool to write higher quality comments. We also found that many reused comments were about similar patterns in code. This inspired the second tool, Code Search, which allows teachers to search for sections of student code that contain a desired pattern. Preliminary results of Code Search are promising: for the queries that Code Search is built for, Code Search returns nearly all relevant results. Together, Comment Search and Code Search offer teachers the ability to give meaningful comments to many more students than otherwise possible.