Marciniak and Reifer, two hardened veterans of the software industry, have written a book for “the host of managers directly or indirectly involved in [software] acquisition projects.” Since “acquisition” involves both a buyer and a seller, their book appropriately addresses software acquisition management from two viewpoints--that of the software acquisition manager and that of the software development manager. The intended audience is any manager involved in any aspect of the acquisition process. In the book’s preface, this audience is defined as including “proposal managers who are responsible for the seller’s response to a buyer’s request, financial managers who plan and account for the cost of a program, contracting specialists who have to arrange the contract vehicle, administrative managers who have to provide services to support the acquisition, and so on.”

The authors offer the intended audience practical tips and warnings derived from bitter experience. For example, chapter 2 sets the stage for the remainder of the book by describing the software acquisition environment (the system and software life cycles and associated management processes). In this chapter, the authors offer the following warning that new players of the software game would particularly do well to heed, and maybe even carve in stone (p. 45):

[Software development] is an engineering process, and one cannot use a cookbook because the recipes for success are often too complex.

In modern systems, software provides 90 percent of the system’s features and functionality. That is why software is so important--it allows an unbelievable degree of flexibility. Of course flexibility does not come free. Added costs and headaches occur because software engineering today is largely a manual, personality-dominated team activity that takes skill, patience, and persistence to accomplish correctly.

The book consists of 12 chapters and 3 appendices. The first 5 chapters give an overview of the acquisition process. They address the following topics: the system/software development life cycle and associated processes, acquisition management as it pertains to software (for example, buyer/seller focus and associated acquisition management strategies such as competitive versus sole-source acquisition), statement of work (SOW) issues, and contracts (fixed price or cost reimbursable) and contract law (including Federal Acquisition Regulations (FAR)). Chapters 6 through 8 address buyer and seller software management models; topics include project planning, project management office roles and responsibilities, and building synergistic buyer/seller teams. The remaining four chapters cover special topics, such as cost issues, quality management, innovative contracting approaches, reliability warranties, data rights, artificial intelligence, and verification and validation. Appendix A is a one-page functional road map of the book’s contents. It shows in which sections various subjects (such as data rights, testing, and risk) are addressed and at what depth. This appendix complements the chapter road map that is given in chapter 1. Because the book is aimed at such a broad readership, these road maps make it easy for a reader to quickly locate the material of interest to her or him. Appendix B is a one-page list of IEEE software engineering standards issued between 1983 and 1987. Appendix C is a three-page listing showing the structure of the FAR. Following the appendices are a six-page bibliography (references are also listed after some chapters) and a ten-page index.

Estimating the cost of a software development effort is, at best, a black art. In chapter 9, “Cost Issues and Answers,” the authors do a credible job of dealing with this difficult, but critical, element of software acquisition management. They give an overview of cost estimating methods (Table 9-1 is particularly nice in that it compactly lists six methods and summarizes the strengths and weaknesses of each). They give some insight into how cost estimating is done in the real world. They also discuss the “art of pricing.” My only concern with this chapter is that the authors do not sufficiently stress that estimating the cost of software development is a high-risk endeavor for almost any software development effort. This point cannot be overemphasized--particularly for new arrivals to the software acquisition management game. I would make this point at the chapter outset and stress throughout the chapter the specific ways topics covered in the chapter serve to reduce this risk.

In fairness regarding chapter 9, the authors do point out difficulties associated with cost estimating and what can be done to alleviate these difficulties. It would have helped if they had taken the arguments one step further and indicated what could happen if such difficulties are not addressed, however. The advice given at the end of Section 9.1 on deciding what estimating method or models to use is a case in point. The advice given is basically sound, but it might have greater impact on the reader if the authors stressed what might happen if the advice were not followed. For example, the authors offer the following recommendation about estimating the size of a software job:

Invest the time to develop better size estimates. Technology exists to reduce the uncertainty in this variable around which cost estimates revolve. Remember, an error in size is considerably amplified when cost is estimated using nonlinear power laws (p. 201).

The authors should have stressed that sizing a software job (particularly for an application that has not been attempted before) is fraught with uncertainty. Because software development is an exercise in the development of understanding between the buyer and seller of what needs to be done, some range of uncertainty should be attached to any size estimate. This range of uncertainty should be factored into the cost (unless it is a “price to win” costing exercise) via, for example, a management reserve. In the absence of such a management reserve, a software project can slide into a cost overrun--frequently of 30 percent or more.

Chapter 12, “Future Directions in Acquisition Management,” is particularly noteworthy. This chapter is a nice, brief, and informative look at the way new software technologies are changing software acquisition management. It focuses on three topics--artificial intelligence (knowledge-based systems), reuse and standardization, and verification and validation. The authors do an especially good job of explaining the concept of a “knowledge-based system” in terms appropriate to the intended audience. They explain the special challenges the development of these systems presents.

If you are going to be involved with the acquisition of custom software systems, you will find this book helpful. First, it will give you a good idea of what you are getting yourself into. Second, it will offer you some tips that will help you reduce your acquisition risk. By “you,” I mean anyone in a buyer or seller organization (either commercial or noncommercial) who will have a hand in the acquisition process. If you are a first-time player in the custom software acquisition game, this book is mandatory reading.