1.1 Introduction

Database management systems (DBMSs) support individual applications and comprehensive information systems with modeling and long-term reliable data storage capabilities as well as with retrieval and manipulation facilities for persistent data by multiple concurrent users or transactions. The concept of data model (most notably the relational models, Codd 1970; and the object-oriented data models, Atkinson et al. 1989; Cattell & Barry 1997), the Structured Query Language (SQL, Melton & Simon 1996), and the concept of transaction (Gray & Reuter 1992) are crucial ingredients of successful data management in current enterprises. Nowadays DBMSs are well established and are, indeed, the cornerstones of virtually every enterprise.

Traditionally, data elements stored in databases have been simply structured (e.g., employee records, and product and stock information). Transactions have been of short duration and often needed to access only a few data items. Most traditional queries have been simple and the techniques used to answer them efficiently are well understood. Taking a broader view, DBMS-based information systems have been built in a rather database-centric way; that is, environment decisions such as the use of mainframe-based or client/server architectures have been typically based on what the DBMS itself requires or supports in this respect.

In the recent past, however, more and more new and demanding application domains have emerged that could also benefit from database technology, and new requirements have been posed to DBMSs. Many applications require the management of data types that are not handled well...