dangerous software?
If some of the posters here would design computer controlled systems for aircraft, that would indeed be dangerous. Take a look at the example of the throttle encoding: angle encoders used for this kind of control are made in a way that precludes wrong readouts by design using an error tolerant coding technique (gray code or code with parity checking). They may not read out at all, or the link may be broken, but they most certainly do not go from 999 to 000. What rubbish. Nothing to do with software neither, by the way. There are other sensor techniques for command inputs like side sticks etc (like potentiometers or phase angel encoding), then they use redundant components or other fail safe designs.
We all agree that we would not trust an aircraft based on windows software, but you cannot compare an office system with an embedded and dedicated controller like FADEC or FMS or you name it, not be design and certainly not by price.
By the way, Captain Ernie Gann describes an uncommanded rollback on a flight from California to Hawaii in his book "fate is the hunter". It was no jet he was flying, it was a DC-4 running on 4 corncobs some 55 years before. Plenty of avgas on board. Now whatever the problem really was (Gann does not provide an answer, but resonance may have been a problem), a DC-4 was not known to be computer controlled. So such things did obviuosly happen without the help of software.