Why Ads? 

The Mythical Man Month *****
- Essays on Software Engineering

Frederick P. Brooks Jr., 1995 (Anniversary Edition), Addison-Wesley

ISBN 0201835959

If you are into Project Management and have never read this book then shame on you! Joking a-side, this book was born out of the bitter experience of developing IBM's 360 Operating System back in the 60s. Brooks was the head of the development and is often known as the "Father of the 360". His observations and messages are as relevant today as then. It appears from reading this book that IBM came very close to delivering the new 360 architecture as hardware without an operating system.

The next paragraph is from memory. The details may not be quite right but I'm sure the underlying facts are.

Brooks analysed the project progress for each software release against the error reports and time to fix those errors. He extrapolated forward and came to the startling conclusion that within a few more releases the project would grind to a halt - the number of errors and time to fix them would escalate towards infinity. He then analysed why this was happening. The primary causes were unrealistic timescales and a demand for additional functionality, causing more people to be added to the project, that meant more separate components, which in turn created more interfaces, more dependencies, and more room for misunderstanding, all leading to more errors. Further, the there was an erroneous assumption that people were available and productive for 7 hours a day, 5 days a week, 48 weeks a year.

The only way to get the project back on course was to reduce the number of personnel, and to fix the outstanding errors before adding any more functionality.

Unfortunately, I never had my own copy of the original book, as it is one that must surely go down as a landmark book of the formative years of modern commercial computing. But now we can purchase this Anniversary edition that contains four new chapters, and which outline Brooks' most recent experiences, but the underlying message and lessons are the same.

POSTSCRIPT: A new Anniversary Edition has just been published, complete with an extra Chapter. The Amazon links point to this new edition.

Why not see how others have reviewed and rated this book? Click on the appropriate Amazon link to find out.

Strategic Data Planning Methodologies ****

James Martin, 1982, Prentice Hall.

ISBN 0-13-851113-6

I think James Martin was the first person to apply component based development to writing books. Often later books would have whole chunks from his previous books, as well as extracts from his course and videos. Perhaps this partly accounts for his prolific writing (28 books by 1982), his books being a staple diet of those of use brought up with computing in the 70s. I had a number of his earlier tomes, but being highly prized they went walk abouts.

By way of some contrast, this Strategic Data Planning Methodologies book is a slim volume of just 200 or so pages. It's all about data modelling and building enterprise models, and whilst the latter may be out of fashion, the former is still relevant, even in these days of object data modelling. Well structured and written, with plenty of diagrams (as always with JM), it is very much a practitioners book. It covers project organisation, and of course data analysis, both at a high-level coarse grain level and at a detailed fine grain modelling level. The former introduces the idea of subject data models (SDM) - collections of related entities and relations. SDM provide a far more manageable level of detail for undertaking the enterprise modelling and deriving a IT Architecture and Strategy. It does this by grouping the SDM according to data activity, and then determining implementation dependencies.

The chapter on IBM's Business Systems Planning methodology reminds us that business process analysis was a concept developed in the 70s. Further chapters cover Entity-Relationship analysis and Entity-Activity analysis. The chapter on Functional analysis makes the case for avoiding an over emphasis on the as-is situation, and engaging top management in the future to-be processes to achieve new corporate objectives - shades of Michael Hammer and BPR in 1982! Much is illustrated by using a mythical dialogue between a data analyst, team leader, executive, and strategy consultant. Again, thinking ahead, Martin has a chapter on distributed data based on both qualitative and quantitative analysis*.

The book concludes by outlining a Strategic Planning Process based on parallel streams of Process and Data analysis, something this author practices to this very day.

*This author was particularly pleased to see the latter, having just build his own case tool to perform dynamic process modelling as part of an enterprise wide modelling project. Written in dbase it ran on the company's first PC, an Osborne 1.

RAD + + ***

Bob Bowman, 1997, Antares Alliance Group.

ISBN none, private publication

Designing Human Systems for Health Care
- The ETHICS Method

Enid Mumford, 1993, Eights Associates

ISBN 0 9522576 0 2

This book is no longer available so the Amazon links refer to another Enid Mumford book on the ETHICS Method: Systems Design: Ethical Tools for Ethical Change (UK) and Effective Systems Design And Requirements Analysis (US).