When I applied to MIT, I listed my interests on the Educational Testing Service form as "experimental psychology" and "human factors engineering." Neither is a degree program at MIT. Seven years later I finished a degree in Cognitive Science - the program, newly created, had as its requirements most of the classes I had already taken (I think they were trying to get rid of me).
Since I still had a bunch of extra units, I decided to take advantage of a program within the Mechanical Engineering Department ("Course 2") that lets students design their own major ("Course 2A"). The requirements are simple: eight Mech. E. core courses, plus another six courses in your chosen area to be approved by the 2A administrator. I needed five more classes, but I figured it was doable. (After taking some time off to earn another year's tuition).
The 2A administrator at that time was a guy named Tom Sheridan. I walked into Sheridan's office, transcript in hand, and explained that I wanted to study human factors engineering. I had already taken physiology, perception, biochemistry, differential equations, several psych. classes, basic materials science, and undergraduate design. I was planning on taking mechanics, dynamics, manufacturing, engineering management, and graduate design. "Sounds OK" he said. "Do you think it's a real degree," I asked, "or does it just look kludged together?" He looked at me like he couldn't remember where he had parked. "Why are you talking to me?" "You're the 2A coordinator." (he's also former head of the Human Factors Society).
Not quite ten years later, I was taking a psycholinguistics class from Herb Clark in Stanford's Psychology Department. I was also planning on attending CHI (the conference on computer-human interaction) for the first time as a real person (as opposed to a student volunteer). To do so, I was going to miss Herb's class. About three weeks before the conference I stopped him after lecture, tickets in hand, and asked if there was a way I could make up a session. "I'm going to CHI. It's the conference on computer human interaction - mostly interface design stuff." He looked at me like I couldn't remember where I had parked. "Your assignment is to attend the keynote speech."
(Oh, and then there's Woody Flowers, who is best known for MIT's 2.70 contest, but who also teaches the graduate design class. Our first assignment was to design a system to drop an egg off the Green building (then the tallest building on campus). My solution was straight out of biologically-inspired design - I was going to emulate a coconut. "You need to protect against the impact," I explained to Woody, "but also against rapid deceleration. Coconut palms have already solved the problem - they have a hard but thin shell, and a softer sponge inside around the payload." What I didn't know is that Woody got his name from his thesis work, which studied how woodpeckers managed to drill into trees without giving themselves concussions (research that eventually led to the design of the Bell bicycle helmet). He looked at me like he really need to park. "Do you have any idea how much money I would have saved on woodpeckers?")