The Science of Large-Scale Information Retrieval

Information scientists have investigated potentially useful methods for Web search engines and archives, yet relatively few of these methods are in active use. These scientists tend to operate without the imperative for large-scale speed and performance, and may not fully evaluate their innovations. This work presents some promising approaches to information retrieval with suggestions for the practicality of applying them to contemporary large-scale retrieval systems. It is suggested that Web search engines and archives may benefit from incorporating these approaches. For information scientists, a shift to increased concern for scalability and practicality is called for.