Sunwings Incident Writeup

[Gabriel]

I simply can't believe that a crew with their head in the game and a respect for procedure could make the same error at the same time like that.

And most likely you are right not believing that. It's by far most likely NOT_what happened. Again, even crew with their head NOT_in the game and a NO_respect for procedure would NOT_make THE SAME error at the same time like that.

[Evan]

And most likely you are right not believing that. It's by far most likely NOT_what happened. Again, even crew with their head NOT_in the game and a NO_respect for procedure would NOT_make THE SAME error at the same time like that.

You're really not following me. If one pilot makes the error and the other just copies his data entries, this is bound to happen. That's what I sort of suspect happened here (see disclaimer).

[Gabriel]

Your really not following me. If one pilot makes the error and the other just copies his data entries, this is bound to happen. That's what I sort of suspect happened here (see disclaimer).

Yes, I am following you and I agree that with such a method the possibility of this kind of mistake increases a lot.
Now, do you know what was the procedure that they were SUPPOSED to follow? Do you know if the procedure required them to make independent entries in the FMS and let the FMS compare? Do you even know if this particular version of the FMS and its software is even capable of that?

I am by no means discarding this possibility and I totally take it as a valid hypothesis. It is not the only plausible explanation though, and we don't even know if it is possible at this point. You cannot blame the pilots for not doing independent entries in the FMS and let the FMS compare if that was not what the procedure required, if that was not what they were trained to do, or if the FMS is not capable of such thing. Which I am not saying any of that is the case, but I cannot discard it either.

[3WE]

Blah, blah, blah, procedure, training, blah blah...

No.

The reckless, improvisational, jack leg, monkey, cowboy pilots screwed up and must be shamed on discussion fora and physically punished...period.

[Evan]

You cannot blame the pilots for not doing independent entries in the FMS and let the FMS compare if that was not what the procedure required, if that was not what they were trained to do, or if the FMS is not capable of such thing.

Correct. Even if it turns out to be pilot error, the pilots may not be the ones to blame. I would only blame them if they skipped over procedure or failed to concentrate on the dreadfully important task at hand (now where is that disclaimer...).

[Schwartz]

Slightly off topic- but I am wondering how you (and or Eric and or Schwartz) are going to design TOPMS to not take errant temperature entries (or other errant entries like weight) and correctly tell the pilots that the takeoff is progressing as it should (or shouldn't). Does TOPMS take it's own temperatures and calculate a nice, buffered takeoff and just give a yellow light that says you might want to go to 100% power...and then a red flashing one that says go to cook-the-engine-but-save-your-a$$ settings?

In a general sense, there are many strategies, field validation being in the easiest bunch. An outside temperature of -51C? A warning should have been issued right there on entry. The general strategy would be:
1) Reduce or eliminate all values selected by a pilot
2) Provide warnings or validations on bounds of expected values
3) Ensure that both pilots are forced to enter numbers independently, or one person having to do it twice DIFFERENTLY so they don't make the same mistake twice.
4) Make sure the context of the numbers on the screen are clear -- form design makes a difference, so at a glance or in a hurry, something out of whack is more likely to be noticed

Lots of strategies. I am guessing this old software was built like a spreadsheet with a long list of fields, little to no validation and no connection to actual measures. Given the number of times this software is used, even a small percentage will result in a good number of incidents.