This book describes the approach to OO development used and advocated by the IBM OOTC. As you might expect from IBM, in essence it is a heavyweight, document-centred approach, designed to minimise risk and biased towards large system developments. This having been said, the authors recognise that smaller, more innovative, projects may have different requirements, and they do attempt to categorise under what circumstances the different types of documentation are relevant.

The methodology they describe is based on an incremental, iterative process model. In this model the design of the system evolves from simplicity and incompleteness to complexity and completeness. This is the most generally applicable approach and the description of it here is thorough and soundly based.

The central idea of the book is that of a Project Workbook, which is made up of various types of Work Product, each one being a type of document, such as Screen Layout, Analysis Object Model and so on. About 50 different Work Product types are described.

My main issue with this book is when it starts to deal with the details. In trying to avoid leaving anything out, a great number of topics are introduced and covered superficially, with lists of bullet points, in a style more suited to a consultancy report rather than a text book. For example, the topic 'Wrapping Legacy Code' is covered in just over 3 pages. Moreover, there are chunks of sample code included in some topics, which is inappropriate in this type of book, especially when they generally have little interest outside an IBM mainframe environment.

However, despite the above-mentioned reservations, this is likely to be a useful book for anyone starting up an OO development project of any size, especially where there is no existing methodology in place, as it does provide a lot of checklists and reference points to aid in deciding what forms of documentation to use.