Querying Design and Planning Databases

We identify some essential features or functionalities involved in querying databases with incomplete designs/plans and study the intrinsic complexity of such operations when presented in a declarative setting. These results thus form a kind of “lower bound” on the complexity of query evaluation of any language incorporating these features. We also identify syntactic restrictions that yield tractable subclasses and show that these restrictions are meaningful in the real world. Finally we study a query evaluation procedure for languages incorporating these features. We show that the usual SLD-resolution procedure can be extended straight-forwardly. Since our language involves computing with sets, such a proof procedure is of intrinsic interest in itself.