从工程工作流到学习闭环:氛围编程教育的内涵阐释与实践框架OA

Breakthroughs of generative artificial intelligence in natural language understanding,code generation,and multi-turn dialogue enable it to participate in the whole process from task comprehension to program execution and task fulfillment,giving rise to vibe coding,which adopts multi-turn dialogue as its primary interaction mode.Based on the theory of human-computer sym-biosis,this paper traces the evolution of programming interaction modes by examining technical participation links and human agency.It compares traditional programming education with vibe coding education,and defines vibe coding education as a human-AI collabo-rative programming practice wherein natural language serves as the primary medium for human-computer interaction,generative AI facilitates solution formulation and program implementation,and learners retain continuous dominance over task intent definition,pro-cess monitoring,and outcome evaluation.Building on this foundation,the paper proposes a four-step learning loop"Discuss,Plan,Ex-ecute,Iterate"(DPEI)for teaching practice and reveals its collaborative operating structure.Furthermore,drawing on a vibe coding teacher workshop case at Peking University,the paper illustrates the application of the DPEI loop in intention expression,task plan-ning,execution and iterative optimization,offering an actionable practical framework for vibe coding education.