This book does indeed cover what is promised in its title and much more besides. It is based on Graham's extensive experience and his well- informed and perceptive ideas. Indeed it ranges from some very detailed and practical advice for running RAD workshops thorough to comments based on such academic abstractions as category theory.

The audience of the book is doubtless restricted by this span: some would wish for a more extensive but purely practical cookbook; others might wish for a uniformly theoretical approach.

Recurring themes in the book are Graham's SOMA method and the SOMATiK CASE tool. At various points he justifies his approach and in particular where it departs from UML. Graham is a member of the OPEN consortium, but has kept his SOMA method separate. Given this isolation, there will be a lower readership for this book than one on UML. (Joe Wood gave a less than enthusiastic endorsement of Graham's book on the OPEN development process in C Vu, Vol 11 no 2.) Personally I found the text invaluable as a source of and trigger for, ideas, references and practical advice. Exposition of a single technique such as Graham's Mission Grid, for de-scoping large problems will excuse the very small number of typos and statements that I thought misleading (e.g. that OO COBOL was standardised in 1997).