IP-UML: TOWARDS A METHODOLOGY FOR QUALITY IMPROVEMENT BASED ON THE IP-MAP FRAMEWORK (Research Paper) Monica Scannapieco Universita’ di Roma, “la Sapienza”, Rome, Italy IASI-CNR, Rome, Italy monscan@dis.uniroma1.it Barbara Pernici Politecnico di Milano, Milan, Italy barbara.pernici@polimi.it Elizabeth Pierce Indiana University of Pennsylvania, USA empierce@iup.edu Abstract. In this paper, a UML profile for data quality is proposed with the aim of supporting quality improvement inside an organization. The profile is based on the IP-MAP framework and has the advantage of giving a formal definition to the main concepts related to the management of quality improvement, as well as organizing such concepts into a set of models useful for realizing a software system. A methodology to improve data quality is also discussed. Specifically, a process to produce the UML artifacts designed by the data quality profile is specified and the initial steps of a pattern-based technique to design quality improvement are also presented. Key Words: Data Quality, Improvement, Methodology, UML, IP-MAP 1 INTRODUCTION Data quality improvement inside an organization is an issue inherently very complex. The proof of this complexity can be found in the various methodologies dealing with this issue that have been published both by researchers ([14], [10]) and practitioners ([3], [8]). When adopting one of these methodologies, one of the major problems is the lack of a modeling language that can help throughout all the methodology phases. Such a language should be formal enough to allow a unique interpretation of the language constructs, i.e. precise semantics should be defined for it. At the same time, people who do not have scientific and technical skills should easily understand it. This is because one of the most important activities when designing for quality improvement is the interaction with customers in order to find out the actual quality requirements. In the software engineering area, a language that has these characteristics is the Unified Modeling Language (UML). UML has defined semantics, and perhaps the principal reason for its success lies in its understandability even for non-technical people. However, the UML semantics are intentionally general. This is because different application domains and systems need different 279 Proceedings of the Seventh International Conference on Information Quality (ICIQ-02)