Paris Kanellakis's pioneering paper in 1990 provided a framework for constraint databases by combining concepts from constraint logic programming and rela-tional databases. The principal idea is to generalize a tuple (or record) data type to a conjunction of constraints from an appropriate language; for example, order constraints or linear arithmetic constraints. Such a tuple can be seen as representing a large, possibly even infinite, set of points in a compact way (e.g., for spatial databases and GIS). Constraint databases have since become a very active area of database research.
展开▼