Maintaining desirable properties of information by inconsistency-tolerant integrity management

Abstract

We argue that several kinds of desirable semantic properties of data provided by information systems, such as their quality, trustworthiness, certainty, security or lack of riskiness, can be modeled and maintained by existing database integrity technology. As opposed to the usual requirement for conventional integrity constraints, formulas that model such properties may not always be perfectly satisfied. Therefore, it is necessary to be less rigorous about the satisfaction of such properties. That can be achieved by recently introduced techniques of inconsistency tolerance, in the fields of integrity checking, integrity repair and query answering. We support our argument by an extended example.