Most of us, even those who are computer literate, still do our business transactions in ways that have not changed much in the past 20 years. We use cash, we write checks, and we use credit cards. Each of those actions creates data, which are then processed, out of our sight, by vast information systems that have, in fact, changed dramatically. Transaction processing was first done by isolated local computers, then by centralized mainframes at the hubs of private networks of data collection points, and now by global webs of public networks connecting client/server processors.

Thanks to the Internet and the World Wide Web, the anonymity of this transaction processing is changing. The first step in that change is automating our manual steps. For example, we can use software that tells our bank to transfer money from our account to pay our bills, bypassing the artifact of the paper check. Some vendors are finding ways of making transactions even more automatic. One online bookstore, Amazon.com, has a system that “remembers” a customer’s credit card number and address and processes an order with one click of a mouse button.

The design and implementation of the information systems and databases that process these transactions have been and remain the meat and potatoes of the information processing industry. Personal productivity applications, such as word processing, spreadsheets, and graphics, have brought the computer visibly into each of our lives, but the majority of the world’s computing power still goes to the mundane and vital task of managing transactions.

Most books on transaction management are highly technical, focused on the details of data storage and data transmission. At the other extreme are books that urge the general public toward the “virtual future” where everything is connected to everything else. These books gloss over the issues of managing all these heterogeneous transactions so that they end up in the right places, with physical goods properly dispatched, and with all the accounting data properly recorded for the ultimate scrutiny of the SEC or IRS.

Chorafas has written a book for the rest of us. It has enough technical detail to be both credible and informative, but not so much that we need to brush off our computer science textbooks. It is useful for systems designers, for business users of transaction management systems, and for CIOs who want to know a little about their choices as they make investment decisions about new technologies. The book is written clearly. There is the inevitable peppering of arcane acronyms, but Chorafas decodes these at the beginning of each chapter.

The first three chapters give a good overview of the topic. Chorafas provides a history of transaction management, introducing issues related to key topics such as data storage, transaction boundaries over space and time, and heterogeneous databases. He talks about the fundamental need to guarantee consistency and integrity and how that can be accomplished in complex applications. He discusses levels of complexity, differentiating between short and long transactions.

Nontechnical readers might want to skim the middle of the book and engage again with the last chapter, which talks about the future of transaction and database management. Chorafas offers the notion of federated processing, which involves “islands of functional automation [that] must be interlinked in an efficient manner.” He talks about the feasibility of tight and loose coupling and suggests several approaches to improving database dependability using various mirror and clustering schemes.

The middle of the book gets deeper into details of specific problems and of the software available for dealing with them. Among the topics covered are transaction monitors, use of software objects, remote procedure calls, and locks and commitments. Again, the detail here is aimed at satisfying readers who want only an overview of the topics. At the beginning of each chapter, along with the acronym translations, are lists of keywords covered in the chapters. I am not at all sure why Chorafas included these lists. They are not in alphabetical order, and only some of the words appear in the index.

Chorafas increases the value of the book by naming specific software packages, but in an industry that changes as rapidly as this one, this makes the book misleading by the time it has been published. I suggest that readers use the book for its excellent descriptions of concepts and then go to more current sources for references to the tools that are available now. Likewise, there is one chapter on the impact of the Web on transaction management, but it is not enough. For example, it has several pages on security issues on the Web, but even today’s general reader needs much more than this.

Those few weaknesses notwithstanding, the book fills a niche in the literature on transaction and database management and serves the needs of its intended audience well.