Originally Posted by bsieker
A/THR is a lot more complex than I imagined.
... and you are the boffin, Bernd! What chance do mere pilots have of fully understanding it before they go to work today? If there was any sense in this world, the type would be grounded for a day while all the pilots on it went to school to be signed off on the latest wisdom. That isn't intended to imply fault with the aircraft or the pilots but for sure the two must communicate/interface better than we have been discussing here.
Instead, the big picture will
be handled differently, and the safety net of statistics "there is no evidence to suggest that..." will be relied upon while everyone gets up to speed to varying degrees. Hidden fingers will be kept on pulses, lawyers will have a superbly lucrative couple of years and the patient will be declared to be in reasonable continuing health I guess.
I am not dismissing those thoroughly experienced pilots who
do understand all the foibles, nor the statistics of hundreds of thousands of hours of relative normality, but you Bernd are a member of a world leading working group of scientists, are you not, whose main field of expertise is Safety of Computers and Automation in Commercial Aircraft Systems? And you have been studying the factors in this a full 7 days a week lately, yet you too are confused by this particular A/THR system?
And to PBL, thank you for coming back here to respond. Does the A/THR logic confuse you too?
Professional Pilots don't have to understand realtime embedded safety critical systems theory in any real depth, but they are supposed to understand Boolean logic. That said, I am pretty certain they can pass ATPL exams and AB type ratings without really understanding more than just some of the rudiments. If I am right, when it comes to the logic in these Thrust Lever circuits those pilots are trusting somewhat blindly on a whole raft of non-absolutes in their knowledge i.e. latest FCOM not Warsaw FCOM as PBL points out, instruction, their own observations, a comprehensive set of system warnings of error states, reputation and 'safety in numbers' of the particular type, and finally that virtual safety net of statistics of hundreds of thousands of hours of relative normality. Fair enough, if their trust is well placed. We can't all be rocket scientists, brain surgeons or top-flight engineers who "know" what they are doing versus "know" what they were told had to be learned.
So is this culture where we carry on with the same rules, but just market/change the emphasis on the product small print a little bit, really conducive to the sort of aviation industry we all want right now? Or do we want money spent on more highly qualified pilots or on more gracefully failsafe systems interfaces, or on less unforgiving runways ? So far, it seems we can't even afford a day on the ground to compare notes, such is the need for the flywheel of progress to keep turning.
PBL introduces the concept of prioritisation or ranking of uncertainties. It wasn't of course intended to be an answer to my "what is the desired culture" question just now, but I would say that I think the Frank Ramsay concept plays too soon into the hands of the accountants and lawyers when used in summarising this kind of investigation, at this stage anyway.
Earlier I indicated some small common ground between systems development (or rather 'developers') in the financial world and in aviation. I would hate to see the two cultures converge even further. To use the gambling terminology, I used to say I could without fear or stress bet my life on all my activities in one and that I respected the other, but now I am not quite as certain as I once was. I shall read on with interest.