Entity search, a significant departure from page-based retrieval, finds data, i.e., entities, embedded in documents directly and holistically across the whole collection. This paper aims at distilling and abstracting the essential computation requirements of entity search. From the dual views of reasoning