Browse Books

Reviewer:

Martin presents this material in three levels of detail: an “Executive Overview” of rapid application development (RAD) (chapter 1), an introduction to the main ideas (the remainder of Part 1), and a description of how to implement RAD (Part 2). Chapter 1 argues that developers can achieve tremendous reductions in the time and cost to deliver high-quality information systems by using modern methods that combine extensive end-user involvement with modern development methodologies supported by well-integrated CASE tools (I-CASE). Chapters 2 through 6 expand on the main ideas. The “Metrics” chapter argues that reporting practical quality and process metrics to the company president is an essential ingredient in improving productivity; it focuses on function points (explained in detail in Appendix 1) as the appropriate code size metric. The “Tools” chapter briefly describes several kinds of CASE tools, each supporting a different diagramming technique, and says that true rapid application development is only possible if the tools are integrated, using a common repository, with different kinds of diagrams linked together and cross-checked automatically. Similar points recur in the “Methodology” chapter, which emphasizes that software development methodologies must evolve through use and have support from integrated, automated tools. The “People” chapter talks about the roles different people (both users and developers) play in RAD. The chapter on “Management” focuses on the role of managers in introducing and continuing RAD. Part 2 talks about how to implement some of the ideas of Part 1. Chapters 7 and 8 explain how to run Joint Requirements Planning and Joint Applications Design (JAD) meetings, at which users and developers collaborate to capture requirements and design as inputs to an appropriate integrated CASE toolset. Chapter 9 advocates rapid prototyping, which Martin describes as continuous iterative enhancement of an initial subset of the system (created during the JAD meetings), evolving the software into a full system. He points out pitfalls of prototyping. Chapter 10, on software construction teams, raises many interesting points about making developers productive, including building teams that stay together over many projects, providing excellent tools and a comfortable working environment, and avoiding interruptions. Chapter 11 discusses the “timebox” approach: to apply an appropriate level of deadline pressure, management selects either an amount of functionality for the first version of a system or the time limit for implementing it, and the construction team chooses the other characteristic. Martin says that useful systems can be constructed in 60 days (after the JAD meetings) with RAD methods. Chapters 17 through 19 give overviews of usability testing and cutover to the new system. Chapter 23 argues that RAD systems are easier to maintain, primarily because the high-level representations in the I-CASE tools are smaller and easier to understand than the mass of lower-level code. Chapter 24 summarizes re-engineering old systems into I-CASE form. Chapters 13 through 16 and the appendices are more technical. Chapter 13 discusses information modeling and the desirability of getting relational data into fourth normal form. Chapter 14 describes process modeling, with is-composed-of, process dependency, and dataflow diagrams, and describes the cross-checks between diagrams that I-CASE tools must perform. Chapter 15 provides an overview of reusability. Chapter 16 details Paul Basset's “Frame” technology for reuse. Chapter 22 considers reuse during analysis. The appendices and methodology charts (about half the book) have more of the how-to: computing function points, using various diagramming techniques, and so on. They also include the original IBM paper on function points and Codd's 1972 paper on normalizing relational data. The book includes many summary tables and illustrations; although many of the tables seem useful (serving as checklists, for example), many of the illustrations are fairly trivial. Sprinkled throughout are interesting anecdotes about companies that have employed these methods, such as a Danish firm that was able to respond to a change in the banking laws over the weekend after the laws were passed. I found the book fairly readable in small doses, but quite tiring as a whole. Portions of the chapter organization made sense, as did some sections within chapters, but the book generally seemed to skip chaotically from topic to topic. The main body of the book is most ly advocacy and advice, aimed primarily at convincing information systems managers to introduce RAD techniques; using it requires a complementary analysis of the appropriate I-CASE tools to adopt. While it includes some interesting material, the book is too unwieldy to recommend.