In reference to Flight 235, there is a lot of speculation going on today, and since the flight recorders' data indicate that the running engine was shut down manually, everyone is suspecting that probably the pilots made the mistake to shut down the running engine instead of the faulty engine to attempt a restart.

If that was the case, why can't there be an alert to warn the pilots that they are shutting down the working engine?

In the software world, we attempt to create so many fail safe situations to avoid letting users mistakenly take an action which will kill the process, why can't that be done for a multi-million dollar aircraft where so many lives are at stake?

P.S: I know these are all theories and not any official crash investigation reports but I am curious as to how such suggestions can even be made? Even my radios for radio-controlled airplanes provide many fail safe features.

  • 4
    $\begingroup$ to be fair, in the software world, there are so many "are you sure?" questions that people rarely pay attention. $\endgroup$
    – falstro
    Feb 6, 2015 at 15:11
  • $\begingroup$ According to Aviation Herald - Taiwan's Aviation Safety Council (ASC) stated that the right engine failed and the crew shut down the left engine. $\endgroup$ Feb 6, 2015 at 15:25
  • $\begingroup$ Why can't it be done? It can! And it would be immediately followed by a multi-million dollar effort to get this system certified by the FAA. This sort of change to a currently certified system is very difficult and time-consuming. It will take accidents like this (if pilot error is indeed found to be the probable cause - we need to give the pilots the benefit of the doubt until the investigation concludes) to bring about the will to make a change. The old saying is that every regulation is written in blood.... $\endgroup$ Feb 6, 2015 at 15:28
  • $\begingroup$ Having looked at the AVHerald and Flight Global very early reports, it seems that even if they had a "click to confirm" or other warning, they still wouldn't have had much time to do anything about it. $\endgroup$
    – FreeMan
    Feb 6, 2015 at 17:00
  • $\begingroup$ I'm favorable to reopen this question. While additional answers should not bring something new, the duplicate reason used for closure doesn't seem fair, the other question is not answering this particular case. $\endgroup$
    – mins
    Dec 29, 2015 at 17:20

3 Answers 3


Because in an emergency situation you do not want to stand in the way of the pilot.(*)

Whoever designs the system knows that it could fail and it also knows that there are two pilots in the cabin. These pilots have been trained to deal with the emergency, have spent hours and hours in simulators to be ready when something happens, you do not want to treat them as clueless users, you want to treat them as professionals that are doing their jobs.

The system told them that there was a failure, if it would have been a black/white decision the system could have already shut off the engine, but it did not: it lets trained personnel to asses the situation and act accordingly.

*: as fooot correctly points out, you do not want to annoy the pilots with warnings (remember, you just told them an engine is not working!), they may be tempted to preemptively cancel them, causing them to miss other warnings that are actually important.

  • $\begingroup$ That makes sense but why cant there be a simple alert about the wrong move even if it does not override pilot's command. Even the best trained operators of a machine can provide a wrong input under duress and an alert to indicate that can't hurt? $\endgroup$ Feb 6, 2015 at 15:58
  • 2
    $\begingroup$ because it would need to assume a wrong and a right: if the system would know what is wrong and what is right it would do it, not ask the pilot to intervene. $\endgroup$
    – Federico
    Feb 6, 2015 at 16:01
  • 2
    $\begingroup$ @Iceman - the whole point is that something has gone wrong. The computer doesn't know if the engine has failed, or if it, the computer has failed... if it tries to ask not knowing which is true, it could make things worse. Better to alert the crew and allow them to make the decision based on all available evidence. $\endgroup$
    – Jon Story
    Feb 6, 2015 at 16:40

In a previous case where a pilot shut-down the wrong engine, the main recommendations concerned pilot training and clarity of cockpit information.

There doesn't seem to have been any desire to develop ever more complex systems that obstruct pilot actions in an emergency.

From AAIB report on 737 crash at Kegworth, UK 1989

enter image description here
enter image description here


Voice recorder information is not a conclusive way to tell what may have happened to the flight.

When one engine is lost on a twin, you normally reduce power to the other engine to mitigate yaw imbalance.

It is certainly possible to create every kind of "warning" system imaginable. The reason aircraft do not have this is because such systems add complexity, cost and weight. Complexity is a bad thing. The more wires and computers you have running around, the more scope there is for errors and problems. Designers generally try to include only things that have significant importance. Remember, the more time a designer spends on "warning systems", the less time they spend on solving problems like how to make the aircraft more efficient and reliable.

As far as the you-are-turning-off-a-running-engine warning is concerned. By the time the warning was received the engine would be off, so it would be post-facto warning--not very useful. Preventing the crew from turning off the engine and forcing them to "confirm" that is what they want to do would add more time to protocol--a very bad thing for that kind of protocol. Also, adding a confirmation system would involve a lot more complexity and probably another button (have you seen the inside of a cockpit?)

Just to describe a scenario. Let's imagine the aircraft has the old confirm-you-want-to-do-this button. The engine is on fire but still running. They need to turn it off. The system now asks them via what? a voice prompt? "Do you really want to turn off a running engine?" So, now they are looking through the hundreds of buttons in the cockpit for the yes-I-really-want-to-turn-the-engine-off button. While they are doing this the aircraft is hurtling towards the ground at 300 mph with an engine on fire. Then they finally find the button and push it, but nothing happens because the circuit to which the button is attached is defective because it only gets tested once every two years, but of course the flight crew does not know this. All they know is they have pushed the button and for some unknown reason they still can't turn off the engine. Meanwhile the aircraft is hurtling towards the ground at 300 mph with an engine on fire and they have seconds left to figure out what the hell is going on....

Just to re-iterate what Frederico said above: emergency systems should not get in the way of the pilot, so any system, no matter how "perfect" it is, that prevents the crew from taking control actions because it is second guessing that action is generally looked on as creating more risk, not less risk.

  • $\begingroup$ Voice recorder information actually there is already some FDR data, but no reports. $\endgroup$
    – Federico
    Feb 6, 2015 at 16:02
  • 1
    $\begingroup$ Yep I have :) And your answer makes sense too but it kinda over exaggerates the complexity that will be required for such an alert. $\endgroup$ Feb 6, 2015 at 16:03
  • $\begingroup$ @Iceman have you taken into consideration the idea that instead it might be you that are oversimplifying it? $\endgroup$
    – Federico
    Feb 6, 2015 at 16:04
  • $\begingroup$ Yes that's very well possible. But then that's the reason I asked because I'm just a student trying to learn more about it. Not trying to be authoritative nor do I intend to say that my idea is surely right. Trying to debate and learn $\endgroup$ Feb 6, 2015 at 16:07

Not the answer you're looking for? Browse other questions tagged .