Originally Posted by
Ian W
Unfortunately, if all four have the same software version then all four could in theory crash and such faults do happen even on fully tested systems.
[/URL]
The type of error encountered in the F-22 squadron is the type of error that could be detected in well-designed testing regimes. Spherical navigation calculations have many well-known traps for software programmers. I wonder if they have flown an F-22 directly over the north pole yet?
Where it is difficult to fully test a piece of software tends to be in exploring the full state map of a system. This results in many possibilities for very, very subtle errors.
The current state of the piece of software in an engine control will depend on all that has happened to that engine in all preceding flights. Even for two identical engines bolted to the same aircraft at the same time they will both experience slightly different operating conditions throughout their service life. This may, though it is not guaranteed, result in a bug occurring in one engine but not another. Operating an aircraft with two engines of significantly different service times is a way of improving this chance (though of course there is still no guarantee).
Anyway, aren't we wildly speculating here? Just because someone has said that a software update might be one of the changes made does not mean that it was the software that went wrong. It is more likely that a software change is needed as a result of a change in the mechanical design of the engine.