Practical programming assignments are critical parts of programming courses in Computer Science education. Students are expected to translate programming concepts learned from lectures into executable implementations that solve the tasks outlined in the assignments. These implementations are primarily assessed based on their functional correctness, ensuring that students' code produces the expected output when provided with specific inputs. However, functional correctness is not the only metric that evaluates the quality of programs. Runtime efficiency is a metric that is less frequently evaluated in programming courses, yet it holds significant importance in the context of professional software development. To investigate this gap and its potential ramifications, we conducted a large-scale empirical study on the time efficiency of 250 programming assignments that are evaluated solely on functional correctness. The results demonstrate that students' programming assignments exhibit significant variance in terms of execution time. We further identified 27 recurring inefficient code patterns from these assignments, and observed that most of the inefficient patterns can be optimized by automated tools such as PMD, IntelliJ IDEA and ChatGPT. Our findings provide actionable guidelines for educators to enhance the organization and integration of code performance topics throughout the programming course curriculum.

@inproceedings{TaoCYZ24,
  title = {Beyond Functional Correctness: An Exploratory Study on the Time Efficiency of Programming Assignments},
  author = {Yida Tao and Wenyan Chen and Qingyang Ye and Yao Zhao},
  booktitle = {ICSE},
  year = {2024},
  doi = {10.1145/3639474.3640065}
}