The Unified Modelling Language (UML) is currently the big thing in object-oriented modelling and it worries me that people will leave this book on the shelf thinking that OMT is irrelevant now that Rumbaugh, in conjunction with others, has developed the UML. Don't make that mistake. Not only are the discussions relevant whatever notation and process you're using but the OMT notation in the book looks very like UML - the only major difference, as far as reading the book is concerned, is the method for indicating cardinality on associations.

The book contains selected articles from Rumbaugh's column in the Journal of Object-Oriented Programming. The first section deals with methods and the articles explain what a method is, what's involved in iterative development and how to handle development with several developers working in parallel.

The next few sections, the bulk of the book, describe how to use various modelling constructs, how to solve common modelling problems and how to avoid common modelling pitfalls. Many books introduce notation without really explaining how to use it. This book is very good at showing an initial model and then showing how more advanced notation can be used to improve the model - if you don't use association classes and qualified associations before reading this book you certainly will afterwards.

The penultimate section describes the development process at various stages of development. Amongst other things, Rumbaugh covers

The final section describes the OMT notation and process and the differences between the OMT notation and UML.

I wish I'd had this book as my second object-oriented analysis, design and modelling book but, even after reading a stack of other OO books, I'm glad to read it now. The book is full of practical advice on a wide range of topics and the clear writing and extensive use of examples make the material very accessible. I recommend this book very highly.