In your article (How a batch of tinned meat fostered fears of the millennium bug, 31 December) you quote Martyn Thomas, who led Deloitte Consulting’s Y2K work, rightly saying that heroic efforts successfully avoided major disasters arising from the Y2K software issues. At the time, I was a member of the advisory committee on the safety of nuclear installations. For nuclear systems, the major goal was not so much about finding and then fixing bugs – although that was done, of course. Rather, it was to do with re-establishing confidence in safety that might have been compromised by potential bugs.
It was expensive, but this expenditure was not wasted, as some still claim. In fact, the cost of assuring that systems are adequately safe can often be even greater than the cost of building them, but it needs to be done. This is a lesson that is often forgotten – or ignored – by system builders. Look at the problems faced by the gung-ho proponents of autonomous cars. Or the risks that we are running from the headlong rush to apply fallible AI systems.
Bev Littlewood
Emeritus professor of software engineering, City St George’s, University of London
