A database system should return more information to the user in response to a query than just the query's answer set itself. While database systems today respond with correct answers, the answer set alone is often misleading to the user. (This is especially true whenever the answer set is empty.) Users may have to expend considerable effort, usually in the form of superfluous follow-up queries, to arrive at the information they initially wanted. Still, systems today do little to alleviate such misconceptions that occur.
There have been a number of cooperative answering techniques explored by researchers, and we have worked on relevant techniques too, to help address these problems. The CDBS that we are developing, Carmin, integrates a number of these cooperative techniques together in a uniform manner. The development of the Carmin System serves two purposes:
In current work, we have developed a suitable architecture for a CDBS. In our research, we are devising a uniform computational basis, and a uniform semantics, for the set of general cooperative techniques. We are also paying particular attention to the computational (efficiency) issues that are involved in producing cooperative responses, for both the deductive and relational database domains.