Abstract
Programming ability is one of the most important abilities for undergraduates majoring in computer science. Taking Yunnan University as an example, this paper analyzes the necessity of improving programming skills and discusses a multi-dimensional reform strategy. The exploration covers curriculum restructuring, the construction of an online practice platform, third-party certification, and the organization of programming competitions. These reforms have yielded significant measurable improvements in student performance, providing a reference for practical teaching reform in computer science.
Introduction
In many Computer Science programs, a frustrating paradox exists: students graduate with high GPAs but struggle to write functional, production-ready code. This "high score, low ability" trap is often caused by traditional assessment methods that rely heavily on theoretical, paper-based tests. Such methods fail to foster engineering literacy—specifically logical thinking and debugging skills—and make manual code review inefficient and prone to plagiarism. To bridge this gap, Yunnan University implemented a "Positive Feedback" system designed to align academic training with industry standards.
Methodology
The researchers established a four-pillar strategy to create a continuous and rigorous training environment for students.
1. Curriculum Reform
The 8-semester program was restructured to ensure students code uninterrupted from their freshman to junior years. Theoretical courses were separated from experimental ones, with the latter receiving independent credits and higher weight in final grades to emphasize practical application.
2. Automated Practice Platform
A "Zero Installation" platform was built using GitLab Community Edition and Docker. By leveraging GitLab’s CI/CD runners, the system automatically executes and tests student code in isolated containers, providing real-time feedback and eliminating technical hurdles for beginners.
3. Third-Party Certification
The university integrated the CCF Certified Software Professional (CSP) exam into the curriculum. Reaching a specific score threshold is now a requirement for graduation credits, providing an objective, industry-standard benchmark that moves beyond potentially lenient internal grading.
4. Competition Culture
Specialized funds support student participation in national competitions like the Blue Bridge Cup and CCSP. These high-pressure environments serve as a "second classroom" to sharpen elite skills and foster a culture of excellence.
Results and Discussion
The reform demonstrated clear quantitative success. Between 2017 and 2019, the average CSP score for Yunnan University undergraduates rose from 92.3 to 163.5—a 77% increase. The university's average score consistently exceeded the national average in the latter half of the study.
Key insights from this practice include:
- Automated Feedback: Moving to a GitLab/Docker-based system eliminated the "plagiarism trap" and significantly improved grading efficiency.
- Continuous Engagement: Programming is a perishable skill; the "uninterrupted" curriculum prevents proficiency plateaus.
- Industry Alignment: Integrating professional internships with companies like Huawei ensures student skills remain relevant to trends in cloud computing and big data.
Conclusion
While the current system excels at teaching core programming and data structures, future work will focus on adapting to "Artificial Intelligence 2.0" and the "New Engineering" initiative. This includes updating the curriculum for AI-specific coding demands and further integrating engineering literacy to ensure the "Positive Feedback" loop remains effective as technology evolves.
Summary updated