For a few years about 20 years ago, I advised the board of Nats, the air traffic control services provider, about software failures, including in NAS, its flight data processing system at that time (UK air traffic failure blamed on ‘extremely rare’ circumstances as CAA opens inquiry, 6 September). NAS was written in 1966 and had been heavily modified, but it retained an effective strategy for recovering from flight plan faults. When NAS failed, it restarted within seconds, discarding the most recently entered flight plan. If it failed again, this repeated up to 10 times. Only after 10 successive “startovers” did it admit defeat and stop. Might the current Nats software benefit from adopting a similar strategy?
Martyn Thomas
Emeritus professor of IT, Gresham College, London
• Interesting that the air traffic control system was stumped by two waypoints with the same name. I once shared flights to and from Cuba with another man of the same name. Boarding systems were baffled (“you’ve already checked in”), resulting in agitated staff and long queues at both Gatwick and Havana.
David Griffiths
Huddersfield, West Yorkshire
