Now, would you prefer that system overriding the pilot, or the other way around? Humans can think outside the box. Software is the box.
And take the n-root of the total based on the number of quality checks every minimal part of that software has been through before even being tried on a real plane, not considering the ones before it went into service.
Seriously, you really think the software for an airliner is written in Ruby+Javascript and the developers use a cloud-based bug tracking app? I work on systems used in public trasportation that are NOT safety-critical and the number of tests, checks and counterchecks they have to go through before being approved for service is soul crushing.
I'm pretty sure that the total count of in the bug database of any airline software currently in service basically amounts to zero.
> I'm pretty sure that the total count of in the bug database of any airline software currently in service basically amounts to zero.
The military, NASA, and a bunch of other critical organizations really wish that what you were saying was true. http://en.wikipedia.org/wiki/List_of_software_bugs
No. The point was to reflect on the nature of software and complexity.
I doubt if bug counts are near zero.
EDIT: And undiscovered bugs are definitely not zero.
There's been quite a few software bugs in certified aviation software. One in the control system for ABS brakes caused a runway overrun, another in the Airbus fly by wire system caused an in flight upset with multiple head and spinal injuries.
The interest I have in stories like this are, how many are preventable by having systems that can override a pilot and is there and preference to doing such?