This document discusses ranking query results in databases to return the most relevant results. It addresses two common problems: empty answers, when a query returns no results, and many answers, when a query returns too many results. For empty answers, it proposes automated ranking functions to return approximately matching tuples without revising the query. For many answers, it adapts probabilistic information retrieval models to rank tuples based on global and conditional scores of specified and unspecified attributes. The document also describes implementing a ranking system with pre-processing, intermediate storage, and a query processing component.