Report on the 5 th International Workshop on the Design and Management of Data Warehouses (DMDW’03) Hans-J. Lenz 1 , Panos Vassiliadis 2 , Manfred Jeusfeld 3 , Martin Staudt 4 1 Freie Univ. Berlin, Germany, hjlenz@wiwiss.fu-berlin.de 2 Univ. of Ioannina, Greece, pvassil@cs.uoi.gr 3 Tilburg Univ., The Netherlands, Manfred.Jeusfeld@kub.nl 4 Univ. of Applied Sciences Solothurn, Switzerland, martin.staudt@fhso.ch 1 Introduction This article reports on the 5 th International Workshop on Design and Management of Data Warehouses (DMDW) held in conjunction with VLDB’03 in Berlin, Germany, on September 8 th , 2003. A short summary on the papers and the discussions held during the workshop is given. Data Warehousing embraces technology and industrial practice to systematically integrate data from multiple distributed data sources and to use that data in annotated and aggregated form to support business decision-making and enterprise management. Although many database techniques have been revisited or newly developed in the context of data warehouses, such as view maintenance and OLAP, little attention has been paid to the design, management and high quality service of the management of a given enterprise. Little attention is also paid to aspects that are intrinsic to the functionality and usage of data warehouses, either in the back-stage (like data cleansing or data extraction and loading) or in the front-end (like similar and uncertain queries, or what-if analysis). The DMDW workshop is intended as a forum to fill this gap. This year’s main theme for DMDW was “Business Intelligence for Data Warehouses: Functionality, Usability and Quality of Service”. In response to the call for papers, 21 papers were submitted. The Program Committee selected 9 papers based on their scientific contribution, novelty and relevance to the workshop. In a popular research topic, DMDW included three papers on the problem of data warehouse design. A second area of interest for this year’s workshop included novel technologies and best practices, possibly showing a movement of data warehousing research towards novel technological challenges, like for example, XML, grid computing and spatial data warehouses. Finally, following a tradition of keeping DMDW close to practitioner problems, three experience reports were also accepted for this year’s workshop. Apart from the DBLP server, the electronic edition of these proceedings is also available from the CEUR series: http://CEUR-WS.org/Vol-77 2 Data Warehouse Design Data warehouse design is a reoccurring theme to DMDW and appears to be one of the most popular fields for data warehouse research so far. Having matured from the era where data warehouse design was considered simply a selection of materialized views, the topics of this year’s DMDW on data warehouse design involve methodologies, correctness criteria and semi-automated support to the DW designer. The paper by S. Luján-Mora and J. Trujillo, entitled “A Comprehensive Method for Data Warehouse Design” presents a UML-based framework for the full span of DW design. The different phases and artifacts of a DW design process are identified and supported by the usage of UML. To this end, the authors start by identifying different kinds of important schemata for data warehouse design. Then, for each phase of the DW design the authors present different stereotypes for different uses. Packages are used to abstract schemata and their details are revealed through zoom out operations. In his paper entitled “Well-formed data warehouse structures”, M. Schneider models a DW as a graph with nodes being fact and dimension tables and edges representing the dependency relationships among the different nodes of the graph. Once the architecture of the graph has been established, constraints for the well-formedness of the graph are given. Furthermore, a mapping to relational structures is also given, both for star and snowflake schemata. Finally, the paper “Using Design Guidelines to SIGMOD Record, Vol. 32, No. 4, December 2003 113