A Measurement System On Debugging Behavior In Programming:Design And Implementation

The cultivation of program debugging ability is one of the difficulties in the current computer high-level language programming course.The traditional method of examination based on written test and computer test cannot describe the details of the debugging process,and it is difficult to make a targeted evaluation of the program debugging skills of students.It is of great significance to develop a measurement system for the program debugging process,to understand the actual process details of the students in the debugging,to evaluate the level and ability of the students to debug,and to achieve personalized teaching guidance.This thesis designs and implements a measurement system for program debugging capabilities.The system collects the line of sight gaze data and human-computer interaction data during the code reading and development process,and obtains the basic physical measurement data of student's debugging activities.This thesis further implements the synchronization and aggregation function of the multi-dimensional data measured by the same debugging process,and then forms the semantic annotation of different debugging activity segments,which provide the details of the activity for the analysis of the student's debugging.This thesis proposes a series of quantification metrics for student debugging.The analysis decomposes a single debugging complete sequence into multiple subsequences in time dimension.Experimental tests show that the measurement system developed in this thesis can effectively identify the behavior of students in the screen reading code and implement the mapping from the original eye movement trajectory to the source code line.The system can identify the code reading task and the editing,running activities of the students in the debugging task of the machine by integrating the multi-dimensional debugging process.Sub-task-based quantitative indicators can provide more details of the debugging proceed,providing a quantitative basis for characterizing and evaluating students' thinking cues.The research work in this thesis can provide a digital basis for teachers' post-evaluation in computer programming courses and self-improvement of students,and provide technical reference.