To compare the performance of different search engines, highlighting the overlap and rank differences.
Presents results of an overlap test search series among traditional CD‐ROM indexing/abstracting databases since the mid‐1980s, web databases, and authors' own polysearch engine, and reviews Google Scholar.
Finds that overlap is minimal among web‐wide search engines which crawl and index the mostly unstructured open web; and that overlap among Google Scholar and the native search engines is far less than the ideal 100 per cent in the optimal context of crawling and indexing highly‐structured and metadata‐rich collections.
Reinforces the existing view that for comprehensive searches one must search more than one database. Highlights and recommends several very good search engine sites.
