Book Review: Applied Statistics for Software Managers

Katrina Maxwell
Prentice Hall, 2002

ISBN 0130417890 (paper)

Buy it from Amazon.com
Buy it from Amazon.co.uk

Other Reviews

 

 

This book presents clearly and concisely, with complete worked examples and instructions on how to do it yourself, pretty much anything a software manager might want to do with statistics.

The book is introduced by a chapter which steps through the methodology of data analysis, from selecting variables and models through quick preliminary analyses with graphs and tables to building and checking the model.

From then on there are chapters which present case studies on software productivity, time to market, a software cost model, software maintenance cost drivers, and 'What you need to know about statistics'. Finally the appendices present the data used.

The book will surely be a considerable help to hard-pressed software managers struggling to get their heads around the complex ins and outs of statistics.

My reservations centre entirely around the focus of the book. It is absolutely right that managers should obtain and compare metrics on the progress and performance of projects -- how else can organisations improve? Indeed, the higher levels of the CMM depend completely on the capture and application of metrics to enable development to be done better and better.

However it isn't true that just focusing on metrics and ever-better ways of comparing them will make things better. If projects don't have clear scope, well-organized and prioritized requirements, good and well-motivated staff, and interested, knowledgeable and focussed managers who can tell good engineering when they see it, no amount of statistical expertise will help.

Katrina Maxwell has gone to great lengths to present her mathematical knowledge in a practical way. Properly applied (to systems as well as to software projects) it should contribute to the toolkit of good and effective managers everywhere. Such people will be able to discern trends not only by crunching numbers, but by listening to people on projects, and dare I say it, to their customers.

© Ian Alexander 2002


You may also like: