As Seriph is misunderstanding autopilots, so Desertia is having his shot at misunderstanding the whole scope and breadth of modern control technology. Too much sand up the nose, mayhap?
The short answer is that - contrary to Desertia's assertion - we currently know how to make (at considerable cost) software, computers, and composite systems that are provably more reliable and more nearly 'correct' than any other component one finds on an aircraft.
To the extent that autopilots, control systems, etc. are not perfect in practical use, the problem in well-engineered designs is almost always one of incomplete SPECIFICATION of the circumstances to be dealt with and/or UNDERSTANDING of the appropriate solution. In less completely engineered ones, a lack of complete TESTING can also be an issue.
To the extent these issues exist, all are basically matters of as- yet incomplete communication between the people flying and the people designing. The full automation of aircraft is a 125-year problem, with only 75 years so far under the belt. When it's done, there will still be pilots and they'll still be saving the bacon now and then, so don't worry about that.
In a PC-dominated world it is an easy mistake to confuse the motives and abilities of those who design truly critical systems for health and safety with the products of highly visible companies, like Microsoft, that aim to profit greatly from a perpetuum of faulty consumer products requiring frequent replacement.
Last edited by arcniz; 31st Jul 2002 at 02:19.