then the computational price we have paid may be too high for simply achieving " user friendliness. " we believe that KWS search should go beyond what SQL (or its superset) can achieve. Therefore, there is a need for re-examining DB KWS from a broader perspective.