Browse Books

Reviewer: Wilfried Lemahieu

While most vendors are adding object-oriented features to their established relational databases, the market for fully object-oriented databases is just beginning to outgrow the prototype stage. The concept of an object-oriented database management system (ODBMS) that unifies application and database development into a single environment is certainly attractive, combining support for concurrency control, recovery, querying, transaction management, and other database capabilities with transparent persistence of complex objects. However, an immeasurable disadvantage of such systems in comparison to relational DBMSs has been a complete lack of standardization. The first chapter of this book introduces an ODBMS standard issued by the Object Database Management Group (ODMG). ODMG is a consortium with members from almost the entire ODBMS industry, with the ambitious aim of developing a standard that “provides for object databases what SQL provided for relational databases.” The ODMG 2.0 standard synthesizes the existing SQL-92, Object Management Group (OMG), and object programming language standards into a standard for building portable object DBMS-based applications. Its functional components include an object model, an object definition language, an object query language, and language bindings to C++, Smalltalk, and Java. Each of these is covered in a subsequent chapter. Chapter 2 defines the unifying basis of the ODMG standard; this ODMG object model is based on the OMG object model, enhancing it with database functionality. It includes the description of a type hierarchy, modeling constructs for states (attributes and relationships) and behavior (operations), meta-objects, locking, concurrency control, and transaction management. Chapter 3 deals with the Object Definition Language (ODL), a specification language that allows for defining database schemes, object types, attributes, relationships, and operations. ODL is based on the OMG Interface Definition Language (IDL) . Because it is restricted to object type definitions, without regard to the implementation of the methods behind them, ODL makes it possible to describe a database schema that is independent of both the programming language and the particular ODMG-compliant DBMS. Chapter 4 introduces the Object Query Language (OQL) as a declarative language for querying and updating database objects. OQL combines SQL-92 properties with object-oriented concepts. It includes object extensions to support object identity, polymorphism, querying over complex objects, navigation along path expressions, operation invocation, and inheritance. The last three chapters describe the C++, Smalltalk, and Java bindings, respectively, as libraries that implement the ODMG object model concepts. These bindings define object manipulation languages (OMLs) that extend the ANSI standards to support manipulation of persistent objects, navigation, transactions, and database operations. The authors show how OQL and OML can be interspersed in such a way that OQL queries can be embedded into application code and, conversely, OQL can invoke OML operations. In two appendices, the ODMG model and the OMG object model are compared, and the ORB binding, which enables ODBMS objects to participate as objects through an object request broker (ORB), is explained. Readers will need a thorough knowledge of object orientation to understand this book. It is aimed at engineers, managers, and students interested in object database systems. However, its enumerative approach makes it more suitable as a reference guide than as a textbook.