The emergence of web collective intelligence is an interdisciplinary research topic involving information science and social science. The topic is mapped to the problem of article edit evolution in Wikipedia based on mapping idea, then the framework of article edit evolution is proposed, the article edit interaction networks are constructed based on the history data of the featured articles in Wikipedia; in the networks, a node is editor and a link is the edit interaction connection between editors. The topology potential entropy is developed to study the article edit evolution. Results show that networks structure and articles evolve toward the orderly direction; the evolution has experienced three development stages from low to high with the emphasis on integrity, accuracy, and readability. The collective structure gradually becomes stable, and it has a scale-free property, both article quality and collective intelligence reach at a high level. There is a process from seesaw-like complementarity to dynamic balance between word quantity and word meaning, the critical point of balance is closed to the golden section point; article edit evolution follows the golden section law. The framework is effective, and the research deepens the knowledge of article edit evolution, web collective intelligence, and social computing.