TITLE

Software Failure: Management Failure. Amazing stories and cautionary tales.

AUTHOR(S) SURNAME, FORENAME

Flowers, Stephen

Places & Countries of PUBLISHERS    YEAR
Publication

Chichester, UK        John Wiley & Sons    1996

PAGES    ISBN        BINDING    PRICE

x+197    0-471-95113-7    Pbk    £19.99

This book is an academic study heavily dependent on the failure of the London Ambulance Service System in October 1992. Much of the other cases reported are dependent on press reports. As the author says, most computer failures are settled out of court and the settlement includes a confidentiality agreement. However, the expert witness in these cases cannot be asked to sign confidentiality agreements on a settlement and so while on professional grounds they would not reveal client details they have accumulated a body of experience that properly handled would be worth publicising. Nevertheless the conclusions in the book are all sound, although there are a few points worth making. Contracts that provide for a supplier to produce a specification which when signed off is to form the specification of a system to be built have the potential for a disaster. It is just not necessary to commit to a system development supplier with or

without a sign off of the specification, prior to the sign off, prototyping and several other forms of check. Secondly an independent view of progress on the project must be formed. The easiest way to achieve this is for the purchaser to have a contractual right to a confidential interview with everyone who leaves a project when they leave. Few leave a project going well but the rats will leave a sinking ship and it may be that they were pushed because they pointed up problems or moved on from frustration because the problems they pointed up were ignored.

Project Management as such is only mentioned in the bibliography. Our observation is that on a building contract there will be at least half a dozen disciplines involved. On software contract we only have information systems engineers as generalists doing everything. There is some specialisation but no separate discipline. The books we review reveal this discrepancy. There is a wider view of computer systems development to be taken than at present envisaged.

            17.11.96    4563

Converted using Wp2Html from Andrew Scriven. Copyright Cliff Dilloway on the last date above. The Authors Moral Rights are asserted