Commentary: Stock Exchange learns the hard way

Click to follow
If the Stock Exchange board were to find itself in a court of law on a charge of gross incompetence, the possible cancellation of the pounds 70m-plus Taurus paperless transactions project would provide a priori evidence for a conviction. The judge would be donning his black cap. The accused - Peter Rawlins, the chief executive - would be preparing his plea for mitigation, but his shifty gaze would show that he had few hopes of a lenient sentence.

As this sorry saga of raised hopes, fumbled deadlines and broken promises has wended its weary way, Taurus has graduated into one of the outstanding examples of technological ineptitude of our times. It is a black hole down which brokers' money has been shovelled with a rare exuberance.

Yes, Taurus is complicated. Yes, the Stock Exchange has asked it to do too much. Yes, the security standards have proved demanding. Yes, the Government's legislative changes have not helped. Yes, the participants who have been involved in testing it have added curlicues and ornaments. Like a party of revellers enjoying a night out, they have ordered the most expensive items on the bill because they know it will be shared.

But all the excuses in the world cannot disguise the brutal fact that the council gave the go-ahead for Taurus in 1989, aiming at its introduction in October 1991; then May 1992; then November 1992; then spring 1994. Perhaps the board should now settle for a workaday runabout that gets it most of the way to its destination, rather than a concept car whose wheels will never spin.

The Stock Exchange needs to learn the lessons that so many other organisations have learned the hard way. There was the famous incompatibility between the computers at the Inland Revenue and the Department of Social Security. There is the London Ambulance Service, whose computerised call- out system crashed. British Nuclear Fuels has had its computer embarrassments at Sellafield, where radiation safety doors were accidentally opened. The moral with computers is usually to adopt as many tested routes as possible - and keep it simple.