One of the challenges in the Information Age is the management of
the enormous data available in the World Wide Web. This amount of
data demands a new generation of algorithms to retrieve information
in an effective way. This talk addresses the Top-k retrieval task,
that is, the procedure of obtaining the best k answers associated
to one query according with one given ranking or preference
criteria. Specifically we present a Top-k algorithm on relational
databases (SQL) associated to an arbitrary ranking function.