Search in Book Reviews

The process described is an iterative, incremental one, based on Rational Software's Objectory process. A portion of what the author discusses is sound guidance for all types of project, but the majority of his guidance is aimed at OO projects that are primarily using the Unified Modelling Language (UML) throughout and he discusses the four standard phases of Objectory at length.

The breakdown of the book is as follows. Part one includes four chapters - how OO development makes projects and project management, 'easier'; how the UML can help in this regard; choosing a development methodology; and planning an OO project. The main thrusts of these chapters are respectively - reducing complexity by increasing modularity; documenting requirements and software design with use cases as well as other UML diagrams; why you should be using an iterative and incremental process rather than waterfall, spiral or RAD and how to research and write a software development plan and use it to set or defend your budget.

Part two also includes four chapters, covering project management issues through the four phases of the lifecycle - inception, elaboration, construction and transition. Each chapter considers development and process tasks, co-ordination and tracking, as well as artefacts required before you can exit each phase.

The final part of the book consists solely of a single short chapter on budget planning, development metrics and a proposed format for monthly project reviews. A couple of the spreadsheets described, a Microsoft Project file and a use-case Access database design are available at the book's web site.

I found this book quite useful and full of tips and advice on all sorts of topics from use-case design to team dynamics. It covers a lot in the space and applies OO and UML concepts to both software design and the people - and process-oriented management aspects quite well. The author seems very fond of the 80-20 rule and of the term 'dysfunctional.' Recent changes to the UML mean his version 1.1 descriptions are out of date in a couple of places and the limited example code snippets and class diagrams are hardly to be emulated (poor type choices, inconsistent naming conventions, etc.). I found a few typos and a few isolated but significant errors in the technical descriptions and UML examples, but I think the real value of this book is in its project management advice.