In the olden days before satellite imagery overlaid with Google Maps, cartographers had a rough job. They were supposed to draw an accurate reference from sketchy, incomplete information. While they did the best they could, there were always unexplored areas they couldn't know about. Since blank areas on maps don't inspire confidence (or sales), they filled them in with lovely renderings of horrific sea creatures. "Here be dragons." Of course we haven't explored here, they said. We know it's dangerous.
This didn't particularly encourage future explorers to venture there, either. If the mapmakers said there were dragons, well - there were dragons. You weren't supposed to go outside the known boundaries; you'd fall off the edges of the world.
However, when you fall of the edges of your world, you usually fall into a new (and larger) one. Take that metaphor and spin it around into systems thinking today. Instead of here be dragons, we have the much subtler factors beyond our control or even unexplained side effects. Can't do anything about dragons; they're just there, and if something flies into their realm, best let it go. We've roamed outside the known boundaries of our field.
Much has been said about the need to de-silo the working world. The words "interdisciplinary collaboration" are usually followed by the buzzword "innovation" somewhere. Executive books and management tomes (even engineering management tomes) are sprinkled with vague aphorisms like "listen to customer feedback" and "have a clear chain of accountability." The case study of the Hubble telescope, while impressive and clearly written, has one major flaw. It has no stories. As Boris said, "I agree with everything [the Hubble case study] said, but it doesn't help me."
It's like saying here be dragons, followed by instructions to be cautious around dragons and avoid getting your ship sunk by dragons. What we need are more specific stories: "...and there was the one time we found this beast of a dragon, purple wings and scaly tales, who flew in from the sky to attack our boat... except Jorge discovered they were vulnerable along the underside of their wings..." If a purple dragon is ramming at your ship, you want to know how people have gotten rid of them in the past, not that you "should avoid purple dragons."
Then there's dragon practice. As Andy said, "the more you can build complex projects that don't matter, the better. Maybe when you run a million-dollar project, then you won't mess up." (He was referring primarily to POE class, which is notorious as a learning experience in systems project failure - individual parts will work, but the whole will not.)
Me? I think the Hubble did a reasonable job of recognizing their dragons post-mortem. I'm impressed the Hubble worked in the first place. Precision parts and pre-internet distributed design? No matter how you slice it, that's a tall order, and they managed to pull it off.