The message I'm trying to get across is what you refer to as "'gleaf' on 'programming' issues" is actually design feature that, if you rely solely on it, will keep you out of trouble when single probe goes haywire or kill you in extremely low probability event of two out of three measurements become synchronously false in the absolutely same manner. Its normal operation is based on assumption everyone sticks to reasonable, established and approved procedures so Perpignan doesn't cut it as proof something is wrong with the design.
Sorry if you feel that your sentence is misunderstood because I've taken it out of context but It seems to me the context just reinforces your notion that "voting" in Airbus FCS is fundamentally wrong. It is all very nice to wish for:
Originally Posted by BOAC
Each false or failed input that enables you normally to identify 'it is not flying' has to be 'protected' against generating a false warning on its own and in an 'and' scenario with other units.
Problem is that at our level of technology such a feat is simply impossible. It is easy to detect unequivocal fault but If you take the mass of data which don't contradict each other outright but paint a picture which looks wrong, the computer that would first recognize something is wrong and then point to it would need near-human intelligence. Sure you can try to program the computer compare this to that, that and that and make conclusions of its own, but you would inevitably end up with program code so complex it would be extremely unreliable.
One can blame a computer for not being able to do pilot's work if it gives him emotional satisfaction but anyway it's ignorant and otherwise useless to do so.
Originally Posted by BOAC
you should remind yourself of the saying about something 'foolproof' - the world can always find a better fool, as is often evidenced here. Human history is stuffed with demonstrations of Murphy's Law, and if you don't cater for that. in aircraft and software design...................
It's fundamentally incaterable for to design aircraft or system resilient to human ignorance so it is very basic requirement anyone working on or around aeroplanes know them very well, and is aware how dangerous they can be so he doesn't take his job lightly. When we're talking about that specific occurrence, why would anyone in the world design alpha probes to withstand the water jet from pressure washer when there are covers for them available with their use prescribed when washing the aeroplane?