SUPER is an exploratory project into the next generation of user-DBMS interfaces. Its main objective is to demonstrate that a visual paradigm can lead to powerful and user-friendly interfaces supporting all phases of the database life cycle (i.e. creation, manipulation and evolution). Visual interaction in SUPER is based on direct manipulation of objects and functions, with a special focus on providing users with maximum flexibility and independence from database technicalities. The set of tools offers facilities to meet the varied demands from categories of users with different levels of skill. Diagrammatic representations and a basic set of functions are better suited for novice and occasional users, while menus and dialog boxes speed up the dialog for expert users. At the same times a consistent interaction style over the various functions and tools has been emphasized. SUPER has been designed as a front end to a relational or an object-oriented DBMS, i.e. the persistence of data, consistency and concurrency problems are delegated to an off-the-shelf database management system. The current prototype supports schema definition, query formulation and browsing using a powerful data model based on objects and relationships.