deSitter, #402
Well my animus for software "engineering" is hereby exhibited - all the
fancy laws and protection modes had the precise effect of sending 228
people to the bottom because they dealt with meaningless abstractions,
not a real world problem. -drl
Sorry, another unpalatable plate of tripe. If you were a professional in
any discipline, you would realise that such areas of work always have
unwritten rules that govern personal ethics and due diligence in all aspects
of the work. Such people are dedicated and to suggest otherwise, cast
aspersions etc, says more about you than any of the people that design
airbus products. Sure, they get it wrong sometimes, but that's what it
means to be human and yes, some aircraft will crash and result in loss
of life; that fact will never change. Check out actuarial data if you
want to see the reality, but also see how safe aviation is in comparison
to other modes of transport.
Try also to remember that software is mechanistic, stupid and at the bottom
of a very deep tree. For avionics, it is a machine reflection of the laws
that were designed into it by systems, aeronautics, airframe, power plant
and a host of other engineering disciplines, each of which will have a
team of specialists to decide the best approach.
As i've said before, it's easy to design a system where it's assumed
that there will never be faulty data or failures, but as
requirements get ever more complex and more systems must be connected
together, handling edge and failure cases in a predictable manner
becomes more and more difficult to provide for and to prove correct.
If you don't build capability to handle all the extreme cases into the
system, then there is no code to execute for that extreme case. By code
here, I mean overall system coordination and behaviour. Since there are
probably millions of possibilities in terms of the sequence of events
leading to an aircraft crash, it's not possible to design any system to
automatically handle all of them.