Measuring code comprehension effort using code reading pattern
SAYANI MONDAL PARTHA PRATIM DAS TITAS BHATTACHARJEE RUDRA
Click here to view fulltext PDF
In the software industry, 85% of development tasks involve maintenance. A developer’s typical job involves more code reading and less code writing. Since code reading is an integral part of comprehension, efficient code reading emphasizes a better understanding of code execution. The more efficiently a code is read, the more it reduces maintenance time. Thus, the developer’s reading pattern can help to estimate how efficiently a developer can read and comprehend code that can improve software productivity. Quantifying the comprehension effort spent while reading a code can help to explore the developer’s efficiency. The primary goal of this study is to quantify the comprehension effort using code reading patterns. An eye-tracking sensor tracks the developer’s eye movements while reading the code. The backward gaze transitions and internal navigation are identified from the reading pattern of the code and used to compute program comprehension effort. An experiment was to collect the data where 41 subjects were asked to read five short C programs. The findings of the study infer a significant correlation between comprehension effort and fixation points. This study explores the involvement of non-experts in higher comprehension efforts than expert ones.
SAYANI MONDAL1 PARTHA PRATIM DAS2 TITAS BHATTACHARJEE RUDRA1
Volume 48, 2023
Continuous Article Publishing mode
Click here for Editorial Note on CAP Mode