Qunits: queried units in database search

Keyword search against structured databases has become a popular topic of investigation, since many users find structured queries too hard to express, and enjoy the freedom of a ``Google-like'' query box into which search terms can be entered. Attempts to address this problem face a fundamental dilemma. Database querying is based on the logic of predicate evaluation, with a precisely defined answer set for a given query. On the other hand, in an information retrieval approach, ranked query results have long been accepted as far superior to results based on boolean query evaluation. As a consequence, when keyword queries are attempted against databases, relatively ad-hoc ranking mechanisms are invented (if ranking is used at all), and there is little leverage from the large body of IR literature regarding how to rank query results. Our proposal is to create a clear separation between ranking and database querying. This divides the problem into two parts, and allows us to address these separately. The first task is to represent the database, conceptually, as a collection of independent ``queried units'', or ``qunits'', each of which represents the desired result for some query against the database. The second task is to evaluate keyword queries against a collection of qunits, which can be treated as independent documents for query purposes, thereby permitting the use of standard IR techniques. We provide insights that encourage the use of this query paradigm, and discuss preliminary investigations into the efficacy of a qunits-based framework based on a prototype implementation.