摘要：There are numerous of programming resources on the Internet, such as programming problems on online judge systems and program codes that solve these problems. Although these resources are valuable for students to practice programming, they are not effectively organized to facilitate students learning. Students and teachers may both hope that all these programming resources are organized as a tutoring sequence. For this purpose, an approach which is based on neural computing is proposed here to organize the programming resources automatically into a tutoring sequence. 2456 source codes were mined in our experiment, resulting in 97 mainstream solutions to 105 programming problems, respectively. These mainstream solutions were sorted by their complexities to form a tutoring sequence which organizes the problems together with their program codes from easy to difficult.
