Automated Reporting of Code Quality Issues in Student Submissions

Abstract

Despite its importance in industry, code quality is often overlooked in academia. A number of automated tools to report code quality have been developed but many of them are impractical to use. They either are developed as a standalone tool, require the use of a particular IDE, and/or need historical data. This paper presents code quality issues reporter (CQIS), a tool that can be embedded in an assessment submission system; it identifies code quality issues for each student submission via static analysis, and reports those in an HTML page whose link is sent via email. The tool covers 52 code quality issues specifically curated for academia, 32 for Java and 20 for Python. According to four quasi-experiments with a total of 274 students, students with CQIS are likely to have fewer code quality issues so long as the expected solutions are long and complex and code quality is considered as part of the marking. These students are also more aware of code quality, and readability in particular.