TL;DR
This paper proposes an "overall optimization" strategy for computer science curricula by tightly integrating five core courses: Programming Language, Discrete Mathematics, Data Structures, Operating Systems, and Compiler Principle. By aligning teaching content and using case-based applications, the authors demonstrate significant improvements in student grades and their ability to apply abstract algorithms to complex, real-world systems.
What Problem Does This Paper Address?
In many computer science programs, core courses are taught as isolated "silos." Students might master the syntax of a Programming Language but struggle to implement a linked list. They might study Discrete Mathematics without realizing it provides the logical blueprint for Data Structures. Most critically, students often fail to see how these foundational topics enable advanced systems like Operating Systems (OS) and Compilers.
The authors identify a "disconnected" chain where students lack the ability to bridge the gap between theory and practice. This leads to a negative impact on a student's capacity for data abstraction and algorithm design—the very skills required for high-level engineering.
Methods
The researchers implemented a multi-pronged optimization method at Beijing Forestry University to synchronize the "strongly correlated chain" of five courses:
- Prerequisite Alignment: They reformed the "Programming Language" (C/C++) syllabus to emphasize the specific tools needed for Data Structures, such as pointer arithmetic, dynamic memory allocation, and templates.
- Mathematical Grounding: "Discrete Mathematics" was adjusted to explicitly link set theory and graph theory to their counterparts in Data Structures (e.g., linking tree-related algorithms to Huffman encoding).
- Case-Based Teaching: To make abstract concepts concrete, they introduced "Case-Based Teaching." For example, they used the analogy of washing and stacking dishes to explain the "First-In, Last-Out" (FILO) property of a stack.
- Computation Modules for Advanced Courses: They integrated specific "computation modules" into OS and Compiler courses. Students were shown exactly how the Banker’s algorithm (for deadlock avoidance) uses matrices or how LRU (Least Recently Used) page replacement relies on stack structures.
- Visualization: The team developed multimedia courseware using animations to help students visualize abstract algorithmic processes that are difficult to grasp through static text.
Datasets
As this is an educational research paper rather than a machine learning study, the "dataset" consists of academic performance metrics from two cohorts at Beijing Forestry University:
- Grade 2007: The control group (taught using the traditional, disconnected method).
- Grade 2008: The experimental group (taught using the new overall optimization method).
- Survey Data: A questionnaire was administered to the Grade 2008 students to gauge their approval of the reform.
Key Insights and Ideas
- The "Chain" Concept: The most important insight is viewing these five courses not as a list, but as a sequential chain where each link must be forged with the next in mind.
- Practical Contextualization: Students are more motivated when they see a data structure's "destination." Knowing that a stack is the engine behind a compiler’s expression evaluation makes the abstract data type feel essential rather than academic.
- Reducing Redundancy: By aligning the courses, the authors reduced the need to re-teach C++ basics within the Data Structures course, allowing more time for complex algorithm analysis.
- Inquiry-Based Learning: By posing design questions (e.g., "Should this stack use an array or a linked list for this specific OS task?"), teachers shifted the focus from rote memorization to engineering trade-offs.
Key Results
The optimization showed clear quantitative and qualitative benefits when comparing the 2007 and 2008 cohorts:
- Grade Improvements: The average score in "Data Structure" rose from 73.4 to 79.7, and the pass rate increased from 90.5% to 94.1%.
- Systems Knowledge: In "Operating System," the average score rose from 74.5 to 78.5.
- Competitive Success: Participation in programming contests saw a massive spike. ACM contest participation jumped from 14 students to 54 students, and "Creativity Cup" participation rose from 23 to 35.
- Student Sentiment: 98% of students in the reformed track approved of the new teaching method, noting that it made difficult algorithms easier to master.
Limitations and Future Work
The paper focuses primarily on the internal relationship between these five specific courses. While successful, the authors do not discuss how this optimization scales to other parts of the CS curriculum, such as Database Systems or Computer Networking, which also rely heavily on these foundations.
Additionally, the results are based on a relatively small sample size from a single university. Future work might involve longitudinal studies to see if these students maintain their edge in professional software engineering roles or graduate-level research. The authors also suggest that further development of "dynamic demonstrations" (animations) is needed to tackle the most difficult algorithms.
Summary updated