Hacker News Reader

Comment by jojobas | Hacker News Reader

Can you really do that though?

You have to trust your workers to not undo some random bolts here and there at some stage.

I worked on airplanes in the long ago. You have to trust that people document every bit of their work and get an inspector to sign off on it where required and get in process inspections sometimes too depending on the work. Nobody who does undocumented work on an airplane should be employed.

If you have a culture in a highly regulated industry where workers can violate those regulations and you dont notice by intention or by inaction, then yes you can be held responsible.

But what if there's absolutely no documented process involving those bolts at all? Its perfectly reasonable to ask what the process is and who is responsible. Boeing acknowledges they don't have an answer here. Is it because they never had a process for those bolts to begin with? You can't document following a process (or not following one) when there is no process at all.

The process to do something undocumented could be to plan and write down what you are about to do, record the before during and after states file the documents.

Alternately the process should be to do nothing until the system studies the action and defines the process. But even then you have a proofing run where you probably should have extra documentation. And arguably in a high safety environment- this would be preferred.

Back upstream to the source of the problem, too much rework was already being required by the time the components got to the Boeing assembly plant (i.e. the Spirit techs onsite doing repairs).

In an environment where process exceptions are the norm, people begin to deprioritize handling exceptions properly. Everything can't be critical or emergency all the time.

And once that deprioritization happens, safety critical process exceptions begin being treated like all other exceptions. Because they're all exceptions, and there are so many.

And eventually normalization of out of bounds situations via repeated team exposure leads to catastrophe.

NASA lost two shuttles via the same organizational problem.

That’s all very true, but just a casual opinion, I think Boeing's process problem is more one driven by prioritization of profits with management corroding the safety culture vs nasa procedural breakdown missing risks because of routine complacency of many previous launches.

Then Boeing management is responsible for not realizing the process was missing. Peoples lives were at stake here.

> Can you really do that though?

> You have to trust your workers to not undo some random bolts here and there at some stage.

Why? We don't trust surgeons to remove correct limbs, why we should trust some contractors thrice removed from actual accountability? There are checklists for a reason.

so its worth clarifying something because 'plug door' is a confusing term as it relates to the bolts. The issue here isn't about trusting workers to do things, its about systems and processes to ensure they know what the right thing to do is and they know how to do it correctly. You CAN (and SHOULD) use things like checklists to ensure people do each job correctly. There are tens of thousands of 'jobs' in an assembly like this - no one knows/remembers exactly how to do each one correctly, nor should a system expect rely on them doing so. Here's how that happens without anyone doing something 'random' and following their instructions as they understand them:

There are doors on an aircraft - you enter through them, they have inflatable slides attached. Then there are plugs - plugs that fill a space (this is one, there are a lot, basically any penetration of the pressure vessel). If there were a space where a door could be but isn't on this model, you need to plug it (hence, door plug or...plug door). In terms of assembly, maintenance, operation, etc. they have different functions and those functions have different documentation and traceability requirements. Doors are meant to be opened and closed - that is different from being removed (i.e., 'taking the door off its hinge'). That leaves us here:

* Doors can be opened and closed (this does not require documentation because it is part of the normal function)

* Doors can be removed, which is different from opened and closed (requires documentation because it is not part of its normal function, requires inspection after reinstallation)

* Plugs cannot be opened and closed (opening and closing them is not part of their normal function so...to whit...doing so is impossible within the vocabulary/concept of documentation rules, they are not designed for regular use)

* Plugs can only be REMOVED (documentation required, inspection required just like a door removal)

Where does that leave the door plug? is undoing the bolts and pushing it up off the stop blocks opening it or is it removal? In the minds of Boeing and the FAA - that is removal. In the minds of line workers, opened a door.

Suddenly, door plug and plug door become potentially different concepts. One is a plug for a door and is discussed and documented like a plug. The other is a door that serves the purpose of a plug...and is discussed and documented like a door. What seems to have happened is that workers opened the plug , which isn't 'possible' - meaning they 'removed' the plug and treated it like opening it. That meant it didn't trigger anything about a required set of procedures to do so, didn't trigger inspection, didn't trigger documentation, didn't trigger inspection, just got closed up and moved down the line.

I've dealt with a lot of these...a personal favorite was a company I worked with having someone who couldn't read be responsible for recording each serial number he stamped on each federally tracked part in our factory. He relied on his memory of the last serial number he did the day before.

> In the minds of Boeing and the FAA - that is removal. In the minds of line workers, opened a door.

If they really do allow line workers to interpret federal regulations themselves, that's quite a problem. But I doubt this was a decision made by someone that turns a wrench.

It’s not about interpretation it’s about ignorance. Where they trained in the difference between a door and a plug? Did they realize it was a door plug and not plug door? Interpretation requires awareness that you are interpreting and that is a surprisingly high bar.

I thought the plane eas delivered with the plugs missing because a contractor failed to install them.

It’s in the article (and every other article written about this debacle).

The plug was of course installed when the plane was delivered, otherwise there would have been a door-sized hole in the side of the plane when it was delivered.

The bolts of the plug were probably originally installed, but some rework was required in the area during final assembly, which required the plug to be removed and put back, and during that process the bolts were not reinstalled. It’s (lack of) documentation of that removal-and-reinstallation that are the subject of the article we are discussing here.

Lets be clear - I do not know the answer to this.

What I have read varies. Some say the door plug was installed and bolted. Some say the door plug was installed but not bolted. Some say the door plug wasn't installed - which I suspect from a lot of prior experience is a linguistic gap between engineers and reporters on installed but not bolted.

I remember reading but can't find at the moment something suggesting that inspecting the door plugs and the bolts was a required Boeing assembly line 'job' (think 'ticket'). But none of that matters if someone, as has been reported, later uninstalled the bolts to fix something else, without creating a record of doing so, because they misunderstood the specific procedure they were undertaking.

Reply on news.ycombinator.com