This is a bit of lore in interface design: The goal of a particular factory position was to bend sheet metal. This was accomplished via a manually fed metal press. The sheet of metal would be placed on the machine, and the press would be activated. Now, this machine was potentially dangerous, so a good process would include the step “Take your arms out of the machine” between placing the sheet of metal and activating the press. The interface designers supported this process by placing two buttons about 3 feet apart. Both buttons had to be pressed to activate the press. The operator could place the sheet of metal and then would naturally have to remove both hands to activate the press. Unfortunately, sometimes the sheet of metal would be jostled out of alignment, and the result was some level of waste. To improve the yield, an operator had an idea: he could wedge something against one of the buttons, and use his free hand to guide the sheet of metal, pulling his hand out after activation. This would reduce the number of jostles. The results were, of course, tragic. The worker was thinking goal oriented. Perhaps he should have made the implicit goal of safety more explicit.
Crazy Nut Job (who has an amazing tumblog on economics you should definitely check out) made this fabulous comment on my post on Trusting Process (in fact all of the comments on that post are brilliant, if it’s at all of interest, you should check it out).
(via ninakix)
Definitely some great comments buried deep in the margins here — with more than enough threading to absolutely pummel Disqus. I’d add that CNJ’s description of his own process in modeling military airspace scheduling is also worth reading.
EDIT: The linking-to-specific comment may not work correctly, it didn’t for me.
