Those reasons are perfectly capable of being known, it's just not economically viable to answer them - would you pay 3x as much for an office suite that didn't crash? How about if it was flying your plane?
I don't quite agree. The returns, simply due to the size of the market of an Office Suite are many times greater than the miniscule market for advanced avionics. The reason software gets released with major bugs in it is pure greed, as well as the fact that the masses will accept being beta testers under the promise of new features, etc.
There is no room for this in aviation, thus the price value comparison is not relevant.
Perhaps the most salient point is that perfect, safe design of an aircraft capable of handling all the variables out there is akin to the perpetual wheel. Designers are still only dealing with the laws of physics and there are too many of them to design a perfect system.
There will never be a perfectly safe aircraft, only a compromise between what we know can happen and what we don't as mentioned earlier. The best and most reliable system we have is the human, IMO, this is the system we should spend a lot more energy on perfecting, training, educating and understanding.