Skip to Main Content
Engineering a Web search engine offering effective and efficient information retrieval is a challenging task. Mitos is a recently developed search engine that offers a wide spectrum of functionalities. A rather unusual design choice is that its index is based on an object-relational database system, instead of the classical inverted file. This paper discusses the benefits and the drawbacks of this choice (compared to inverted files), proposes three different database representations, and reports comparative experimental results. Two of these representations are one order of magnitude more space economical and two orders of magnitude faster in query evaluation, than the plain relational representation.