Another thing to note is that homework at Caltech can be rather challenging. It is common (or at least when I was there, around the same Walter was, and I believe it still is so) for people to collaborate on it. You get a pretty good idea from this how your classmates are doing, and you can make a pretty good prediction by the end of the term how they are going to do on the exam. In my experience, pretty much everyone came out about where I expected them to. If there was significant cheating going on, there should have been noticeable instances of people doing unexpectedly well on exams.
> I don't remember ever locking my dorm room door.
Did you ever meet the guy in the bushes on California Blvd? He was a mentally ill college-aged kid who hid in the bushes in front of a building around half way between campus and Lake Avenue. He would ask people if they had seen "Sandy". (I know some people who managed to have a long talk with him, but they were not able to figure out if Sandy was a girl he actually knew, or someone imaginary).
One day, some of us were hanging out in my room, which was at the top of the stairs that lead to the basement of the South houses. Someone started joking about funny things we could say when bushes guy asked about Sandy. I suggested something involving a crude sex act.
Apparently bushes guy happened to be lurking in the South houses basement and was listening. That night, I woke up in the middle of the night with him standing over me, rather agitated, asking why I would say such terrible things about Sandy.
I started locking my door after that.
Doing the homework was how I learned the material. Things came too thick and fast during lecture to learn it, it was all I could do to keep up taking notes.
I discovered that if I knew how to solve all the homework problems front to back, I'd do ok on the exams. I also rarely could do the homework without some help from other students, who were always happy to coach me through my difficulties. Having a "troll" session with your friends to work through the homework was commonplace, followed by a "wretch" session with the TA for the stuff none of us could figure out :-)
For me personally, I did not attend Caltech in order to get good grades. What I learned there has given me a huge advantage in my career. Interestingly, my experience with Caltech's honor system has also shaped my interactions with others for the better. (For example, I've never used DRM on any products I've built, despite sometimes strong pressure to.)
Yet cheating is nearly non-existent. How is this possible?
I think it is because the honor system is a pretty explicit statement that the Institute implicitly trusts its students. It views professors and students as collaborators, rather than adversaries.
One consequence is that the students like the honor system very much, and do not wish to see it change. Hence, there is intense pressure from the students to themselves to not cheat - and someone who does cheat is pretty much ostracized.
(There are other aspects to the Caltech honor system - professors are not allowed to take attendance or base grades on attendance in any way. The Institute also does not attempt to control student behavior on or off campus. I don't remember ever locking my dorm room door.)