3 ms·
None with fatalities, however in 09 US Airways Flight 1549 ditched in the Hudson to much fanfare. There have also been accidents with fatalities involving feede
by Shivetya 13y ago
None with fatalities, however in 09 US Airways Flight 1549 ditched in the Hudson to much fanfare. There have also been accidents with fatalities involving feeder airlines to the majors within the US. Outside the US is another story entirely, but we tend to ignore most of the accidents unless the numbers are large.
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?
- a3n 13y agoLook at your bug database. Multiply that by the complexity of not only an airliner, but the huge and complex system in which it operates. Add weight for consequences in human fatalities. Now, would you prefer that system overriding the pilot, or the other way around? Humans can think outside the box. Software is the box.
- Kurtz79 13y ago"Look at your bug database. Multiply that by the complexity of not only an airliner, but the huge and complex system in which it operates." 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.
- a3n 13y ago"Seriously, you really think the software for an airliner is written in Ruby+Javascript and the developers use a cloud-based bug tracking app?" 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.
- omni 13y ago> 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. 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 http://en.wikipedia.org/wiki/List_of_software_bugs
- snom380 13y agoThat's pretty much the point. The safety required makes writing a complex, life critical system like this and getting it certified pretty much impossible with our current software methodologies. 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.
- notahacker 13y agoBesides which, I'd really hate to be the guy scoping out the software that decides when to ditch an aircraft in the Hudson river. Avionics software is considered reliable because it doesn't try to do too much.