Many large information systems written in COBOL, which were originally developed about two decades ago, have undergone numerous changes. As a result, these systems have become clogged with logic, and the majority are resisting any further change. However, with the changing needs of the MIS organizations that own such code, further system change is necessary, and restructuring this huge mass of clogged logic is difficult and demanding at best. Good methodologies are needed to decompose existing large systems and to fully understand their internal functionality so that program specifications for the existing system can be extracted and then combined with the specifications for the system changes needed.

In addition to this challenge, many MIS organizations find themselves wanting to move forward to object technology. This book is suitable for managers of such organizations, because it deals with two important issues: understanding the problems of rewriting older, complex COBOL systems and developing a de-engineering/re-engineering methodology to answer these problems; and having the methodology include the goal of implementing COBOL code using object-oriented design techniques. This book is well-written for managers and system development teams who want to convert older, complex COBOL systems into COBOL systems with objects using the two techniques mentioned above.

The book is divided into three parts, consisting of a total of ten well-organized chapters. The first three chapters fall into Part 1, which introduces the concept of objects and the problems of maintaining large COBOL systems. Chapters 4 through 8 fall into Part 2, which deals with the methodology of building COBOL systems with objects. The re-engineering methodology mentioned above is introduced in Part 2. Part 3 consists of chapters 9 and 10, which describe possibilities and present conclusions.

For readers unfamiliar with object-oriented techniques, chapter 3 offers a wealth of information, because it talks about COBOL objects and describes how objects appear inside a COBOL program. The reader gets an opportunity to use such objects within the context of a complete methodology, rather than just as standalone tools. The reader also gets the opportunity to practice loose coupling within modules.

Object-oriented design is a simple idea with great strength. I used to believe that for object-oriented design, you really need an object-oriented programming language, until Grady Booch disproved this notion with his book on Ada [1]. While reviewing this book, I accepted that COBOL programmers may take advantage of object-oriented programming. However, for hard-core object-oriented programmers, this approach may seem superficial (as Booch’s approach did to some critics in the mid-1980s). Perhaps a separate part consisting of two or three chapters at the beginning giving a hard-core, language-free introduction to objects would be a useful addition to the next edition. This would be of great service to managers and development teams who are facing the challenge of some day converting their applications from COBOL to C++ or a similar powerful tool.