From: Jean-Francois Mezei <firstname.lastname@example.org> Organization: Vaxination Informatique Date: 11 Aug 96 00:03:56 References: 1 2 3
View raw article or MIME structure
You can argue all you want about whether having multiple different systems cross check each other, or having identical systems cross check each other. But judging from what I have learned about the Airbus problems, many of the problems did not occur because of coding errors, they occured because of design errors (for instance checking for both squat switch AND wheels turning before allowing deployment of thrust reversers which caused a problem landing on a wet runway when wheels would skid and not turn). However, many of the problems were caused by computer flaws, notably with the alarm computers who generated many false alarms which resulted in the plane taking immediate corrective actions (confusing pilots). There were also cases of the adjustment of the cabin temperature affecting the engine thrust routines during a flight. With identical inputs and outputs, different programs can only differ so much. And if different systems are fed with the same erroneous data, the different systems will respond with erroneous outputs. I beleive that the different systems do have the following advantage: In the event of special yet-untested conditions arising that cause one computer system to crash, having different programs will reduce the likelyhood that all systems will crash (eg: division by zero, trying to access non-existant memory etc). As far as over-automation and imposed limits: Do you not think that the designers of the DC-3 imposed limits when they designed the rudders and elevators and decided just as much they could physically move ? Do you not think that the designers of new engines imposed limits when they added electronic fuel controls ? The argument that one should be allowed to "bend" the aircraft and cheat the limits in order to save the flight is made MANY MANY times. While I am sure that many examples can be brought up where such "cheating" saved lifes. However, I wonder if there are many RECENT proofs on RECENT aircrafts. Newer and lighter materials and alloys may not have that "bending" ability that older materials have, so claiming that it is safe to "bend" a A320 is not necessarily true even though it might be true for a similar claim for a vintage 727 or 737-200.