Inductive databases tightly integrate databases with data mining. The key ideas are that data and patterns (or models) are handled in the same way and that an inductive query language allows the user to query and manipulate the patterns (or models) of interest. This paper proposes a simple and abstract model for inductive databases. We describe the basic formalism, a simple but fairly powerful inductive query language, some basics of reasoning for query optimization, and discuss some memory organization and implementation issues. Keywords inductive databases, inductive querying, constraint-based mining