From: firstname.lastname@example.org (Jean-Francois Mezei) Organization: DECUServe Date: 22 Jan 96 04:40:39 References: 1 2 3 4 5 Followups: 1 2 3
View raw article or MIME structure
In article <4d9cvk$ntc@gsb-crown.Stanford.EDU>, rna@gsb-crown.Stanford.EDU (Robert Ashcroft) writes: > Beta testing in production with a product that carries 150 lives and moves > at 550 miles per hour? Oh yeah, this is a company _I want to support. > If what you say is true, Airbus is criminally negligent and ought to be > sued into oblivion. Lucky for them none of the A320 crashes happened in > the US. Air France was the basic beta tester or Airbus and they had quite a mouthfull of complaints to Airbus which, at first, did not wish to have its engineers talk to Air France pilots. So the 320 did have huge theeting problems, and nobody is denying this. You might wish to sue Airbus. But if I were you, I would sue the FAA and other accreditation agencies for having accredited the 320 for commercial service about 2 years before the plane was *really* ready for commercial service. > Or, possibly, it was a bug, or an something that the design engineers never > anticipated. Remember, humans ultimately build these things. Complicated > computer control software is just one more thing that can go wrong. 1- Most of the problems with the 320 WERE because of improper information being fed to the computers, thus generating alarms that resulted in unexpected actions. For instance, how is a pilot to react if the ground proximity alarm goes off at 20k feet over the ocean ? The landing gear incident was something which was probably unknown to all at that time and no engineer or pilot could have predicted that hydroplanning would be a problem during landing. In hindsight, it is easy to blame airbus, but they put in that extra check as a safety measure to make sure the plane was indeed in the ground. 2- If you know about the concept of knowledge base, you should know that it is possible to transform the knowledge and experience of experts into computer programmes which, when used, provide the user with a very good decision tool. If Airbus did things right, when it discovered the aqua-planning problem, it would have fixed the software to handle such cases since wheel brakes are basically inoperative. (I don't know if this actually happen, just mentioning the possibility) About enveloppe "breaking": If a manufacturer is going to find out what the envelopes of its aircraft are, perhaps it should have 2 numbers: a safe enveloppe and a maximum enveloppe. The whole point of these envelopes is to allow the pilot to push the button to the max without having to worry about breaking the plane. This is the same as ABS brakes in cars: you push the pedal to the max and you know that the computer will apply justy the right amount of braking to get you to stop the fastest. Users who are not used to ABS will defeat its purpose by "pumping" the brakes and won't get the most efficient braking because they don't use them properly. I think that the same applies to any new system.