Taking A Stand Against User Errors

This happened in the eighties or early nineties. The scene is a factory making heavy machinery. The factory floor had terminals connected to a mainframe for tracking parts and whatever else they needed it for.

I get a call from the factory floor, and after the usual pleasantries, the user says:

User: “I can’t log in when I stand up.”

I think that it’s one of THOSE calls again, and goes through the usual: Is the power on? What do you see on the terminal? Have you forgotten your password? The user interrupts:

User: “I know what I’m doing. When I sit down, I can log in, and everything works, but I can’t log in when I stand up.”

Me: “There can be no possible connection between the chair and the terminal, and sitting or standing should in no way affect the ability to log in.”

After a long back and forth on the phone, I finally gave up and walked to the factory floor to show the user that standing can’t affect logging in.

I sit down at the terminal, get the password from the user, log in, and everything is fine. I turn to the user and say:

Me: “See? It works; your password is fine.”

User: “Yeah, told you, now log out, stand up, and try again.”

I oblige, log out, stand up, type the password, and… invalid password.

Okay, that’s just bad luck. I try again: invalid password.

And again: invalid password.

Baffled by this, I try my own mainframe account, standing: invalid password.

I sit down and manage to log in just fine. This has now turned from a crazy user to a really fascinating debugging problem.

The word spreads about the terminal with the chair as an input device, and other people start flocking around it. These are technical people in a relatively high-tech factory; they are all interested in fun debugging. Production grinds to a halt. Everyone wants to see if they are affected. It turns out that most people can log in just fine, but there are certain people who can’t log in standing, and there are quite a few who can’t log in regardless of standing or sitting. 

After a long debugging session, they find it.

Turns out that some joker pulled out two keys from the keyboard and switched their places. Both the user and I had one of those letters in the password. We were both relatively good at typing and didn’t look down at the keyboard when typing while sitting. But typing when standing is something we weren’t used to, and we had to look down at the keyboard, which made us press the wrong keys. Some users couldn’t type properly and never managed to log in. While others didn’t have those letters in their passwords, the switched keys didn’t bother them at all.