Metric and Tool Support for Instant Feedback of Source Code Readability

Abstract

In the software maintenance phase, comprehending the legacy source code is inevitable, which consumes most of the time of the phase. The better the code is readable, the easier it is for code readers to comprehend the system based on the source code. This paper proposes an enhanced source code readability metric to quantitatively measure the extent of code readability. In addition, we developed a tool support named Instant R. Gauge to update the code on the fly based on the readability feedback of the current code. The tool also provides the history of the readability change so that developers recognize the more readable code and gradually change their coding habit without any annoying advice. The suggested readability metric achieves 75.74% of explanatory power, and our experiment showed that readability of most of the methods authored in our tool is higher than that of the methods without our approach.

Keywords

• multiple linear regression
• readability
• source code analysis