Common Code Quality Issues of Novice Java Programmers: A Comprehensive Analysis of Student Assignments

Starting to learn programming is often perceived as being quite tedious by students at the bachelor level. Many programming courses thus face high drop-out rates and moderate results for those who pass. This problem is exacerbated when teaching programming to students enrolled in non-computer science curricula. To overcome these issues, we have developed a novel didactic concept based on peer learning, tutoring, dedicated teaching, and learning material that supports individuality and competence-based learning. Our current focus lies on teaching basic programming principles, but to further support our students and foster a positive learning experience, we want to learn more about the difficulties they are facing, particularly with respect to best practices, coding conventions, and code quality. Therefore, we performed a static code analysis of homework assignments of students participating in our introductory programming course for two consecutive years. We analyzed over 13,000 Java files and more than 400,000 lines of Java code to identify common code quality issues faced by our students. Our analysis shows that the majority of rule violations are related to coding style. The violations do not differ much with respect to the topics covered in the homework assignments, and hardly change over time. The more lines of code the students write, the more rules are violated. Based on our findings we present concrete recommendations on how to support novice programmers in improving their code quality.