This excellent article calls for the unification of the information retrieval (IR) and database (DB) communities. The article provides context and background for their separate, yet converging, paths over the past 40 years of computer science (CS) research. Furthermore, the article frames the necessity of dovetailing IR and DB research as a grand challenge, based on the increasing data volumes, low cost of storage, and high computational ability, as well as demanding user expectations for a flexible and efficient search across large data corpora. In order to evaluate and provide meaningful answers to queries, such as “find young patients in central Europe, who have been reported in the past two weeks to have symptoms of tropical virus diseases and an indication of anomalies,” the IR and DB communities must cross-fertilize, according to the authors.

Weikum et al. list several low-hanging fruit areas: first, approximate matching and record linkage--for example, deciphering when queries for William J. Clinton are really the same as queries for Bill Clinton; second, schema relaxation and heterogeneity, dealing with unstructured or semi-structured data transparently and accessing the data in a variety of organizational structures, such as an entity-relationship (E-R) model, tuples, or the resource description framework (RDF); third, information extraction and uncertain data--that is, extracting structured information from text and evaluating it with confidence; and, finally, entity search and ranking, ensuring that the appropriate top N results returned from search are relevant to the asking users.

After the grand vision and collaboration areas, the authors illustrate recent successes in collaborations between the IR and DB communities, using their own experience in the context of the Yet Another Great Ontology (YAGO) project. YAGO is a growing ontology, represented in RDF and Web ontology language (Owl) Lite. It has been used to engender several recent modern search systems, including Microsoft’s Libra scholarly database. These sections present how some of these low-hanging fruit areas for DB and IR collaboration are represented in YAGO-driven projects, including the already-mentioned Libra, as well as Cimple/DBLife and KnowItAll/TextRunner. At the same time, this part of the article feels a tad self-referential and grandiose; so read on, but your mileage may vary.

I found the first part of the article (and the general feeling as a whole) fascinating. It challenges both the IR and DB communities to come together for the greater good of search, information representation, and retrieval in the modern world. Half of the article is a great overview of the challenge. While the other half is a bit wordy and technical, it does provide some existing context for how these challenges are being addressed today.