Abstract
We consider a model of querying remote databases, in which we compute results of queries not supported by the database system, by using sequences of supported queries and analysing locally their results. We study the expressiveness of this model of computation, as well as its complexity, measured in terms of the number of queries used.