Novel method to cope with complex Web searches
Answers to complex search queries such as 'Where can I attend an interesting conference in my field close to a sunny beach?' cannot be found, even though the Web contains all the required information. Search engines can only answer generic or domain-specific queries. The EU-funded SECO (Search computing) project proposed search computing as a new multidisciplinary science to provide the abstractions, foundations, methods and tools needed to answer these and many similar queries. Project partners utilised search service characteristics to develop query languages, query execution plans, plan optimisation techniques, query configuration tools and exploratory user interfaces. These cover a wide range of features, from theory to applications in data management, human interaction and software engineering. Specifically, novel methods were developed to extend the rank-join theory. The methods were supported by an open infrastructure to efficiently execute search computing queries. Researchers examined several ways to present complex results to users within the liquid query framework where query results adapt to changing user needs. By focusing on search service modelling, they introduced the service mart pattern. Outcomes were presented in about 100 publications, leading international journals and conference proceedings. In addition, SECO promoted eight international workshops and produced four books. Lastly, the SECO consortium obtained a United States patent for a method to extract, merge and rank search engine results. SECO introduced search computing as a viable solution to manage complex queries through dynamically selected, cooperating search services. Ultimately, users will be able to extract information from their preferred search engines in response to complex queries.
