Modeling and Monitoring the Quality of Data by Integrity Constraints and Integrity Checking

Characteristic attributes for describing the quality of data, such as trustworthyness, soundness, riskiness, uncertainty, dependability, reliability and other semantic properties can be modeled and monitored by conventional database integrity technology. As opposed to traditional consistency constraints, occasional violations of some of the integrity conditions that describe quality aspects may be tolerable, even for extended periods of time. Traditional integrity checking methods are intolerant wrt. any constraint violation. They insist that all constraints are totally satisfied before updates can be checked for integrity preservation. Inconsistency-tolerant methods can waive that insistence. Thus, if data quality is modeled by constraints, it can be monitored by any integrity checking method that is inconsistency-tolerant. We illustrate that by an extended example, by which inconsistency-tolerant integrity checking is also compared to some alternative aproaches.