WHEN public sector computing hits the headlines, the news is seldom good. A report by the National Audit Office last month found that computerisation at the Passport Agency had produced tiny savings in staff instead of a predicted 20 per cent.
Last October, ambulances in London were thrown into disarray when the computer-aided dispatching system collapsed. Other embarrassments include: a Foreign Office accounting system that generated figures pounds 46m adrift; pounds 12m written off when changes in the Property Services Agency rendered a new system obsolete; and a Ministry of Defence accounting project abandoned at a cost of pounds 10m.
When lives have been put at risk, the recriminations have been far-reaching. Four months after the London Ambulance Service debacle, an official inquiry produced 80 pages of criticism covering every stage of the project. And where public investment in computing has had to be written off, the National Audit Office has been quick to investigate.
Time and again these reports have identified the same underlying problems. Projects fail for three main reasons: they are over-ambitious; the requirements they were intended to meet change or disappear; or they are contracted out to parties that cannot deliver.
The London ambulance project suffered mainly from the third problem, but many suffer from all three. The question we should be asking is: why does it go on happening?
Politicians reveal some strange instincts when confronted by information technology. They are eager to believe that computers offer more powerful controls for the ship of state and allow greater freedom to introduce rapid changes in policy. But they are reluctant to accept that information technology systems can cost a lot of money. For example, to move from a fairly reliable specification to an ultra- reliable one can be hugely expensive.
This driving optimism is due at least in part to the lobbying efforts of consultants and suppliers who, like over-enthusiastic surgeons, often omit to mention some of the risks and side-effects of their favourite therapies. One frequent side-effect is inflexibility. A National Audit Office review of eight defence projects rated half as inflexible and difficult to update. If a system has been designed to perform well according to a closely defined specification, it is not likely to be cheap or easy to adapt to a different role. And it is often just as difficult to introduce changes while a project is in progress; it is rather like deciding to move the stairway while the builders are working on the walls of the hall.
Some projects suffer from what might be described as Thunderbirdism. As a project gathers momentum, unforeseen technical problems emerge and, as in the children's television series Thunderbirds, they are promptly referred to 'Brains'. It is taken for granted that solutions can be found through outside brainpower. As a nation, we cherish the thought that anything technical can be cracked if subjected to enough spasms of deep thought. Thus, the London ambulance board was suffering from Thunderbirdism when it chose to believe assurances that its system could be fully operational in an impossibly short time.
Post-mortem examinations of projects rarely explore these fundamental issues. As with the ambulances, reports can apportion blame and suggest ways of avoiding repetition of mistakes, but they do not offer a long-term solution.
Two trends give special cause for concern. One is the increasing number of public sector computing contracts awarded to private companies; the other is the growing recourse to elaborate review and control procedures that may help to forestall some disasters but inflict considerable complexity and delay.
Both trends betray lack of confidence in the public services' ability to exploit information technology, and ensure that fewer public servants than ever will have experience of its sharp end.
Practical experience is a much better source of insight than reports from committees of inquiry. In any case, even if inquiries do produce recommendations, it becomes increasingly unclear who is expected to heed their advice. The individual contractors? The computer industry at large?
If we are going to rely on computer systems to support our ambulance services, courts, schools, hospitals and national defence, this seems a good moment to ask: your life in whose hands?
The author is a member of the Department of Accounting and Finance at the University of Birmingham.
Solo Electronic Systems
THE photograph that accompanied our article entitled 'Thunderbirds please go' on 10 May showed a terminal designed by Solo Electronic Systems Ltd as a part of the London Ambulance Service's computer- aided dispatching system.
We wish to make it clear that we did not intend to imply in any way, by use of the photograph, that Solo Electronic Systems may have been responsible for the failure of the computer-aided dispatching system of the London Ambulance Service in October 1992.
We apologise if any confusion was caused as a result of the publication of the article and its accompanying photograph.
(Photograph omitted)Reuse content