Proper attribution in teaching and textbooks

In research papers, I'm very clear on the proper protocols for citing previous work and giving attribution to those whose ideas I'm building on. But I've always been enormously less clear on what the right moral and legal procedures are on the teaching side of my job.

For example, I usually write exam problems from scratch (or substantially altered from someone else's inspiration) but I have the impression that it's fairly common for people to simply grab problems out of other textbooks. Do the usual academic standards of proper attribution apply? (I've only rarely seen professors provide citations in their exam problems. Identifying the source could even invite integrity problems if it's a take-home exam.) In a research paper, I'd want to cite the source that gave me the idea even if I changed it completely when I used it, but when writing exams or homework that convention doesn't seem to hold at all.

I've been puzzled for years about similar issues with textbooks. How do you write a textbook for something like introductory physics without inevitably stealing ideas and approaches from the zillions of textbooks that you've used or read in the past? In research, most textbook-level facts aren't given citations at all, but what about specific analogies or ways of explaining a topic? What about a novel choice of order for the topics in the course? Did the second physics text that included chapter summary pages have to cite the first? (Or going farther, did its publisher risk a copyright lawsuit?)

I assume that major publishers have legal departments that are familiar with all the formal standards for that sort of thing. But if there are (or someday are) efforts to write freely available textbooks independently, how do the contributors deal with this sort of thing?

How many ounces are in a cup?

I'm really trying to figure out what happened here:

My best guess (which still seems crazy) is that the person designing these measuring cups looked up a conversion factor from ounces to metric and from cups to metric, and the two factors were rounded differently, so they decided to put both (different!) scales on the cup.

It's amazing how much less that makes me trust the thing. What else did they get horribly wrong? (And how did this ever get past any sort of quality control?)

Edit: Oh, hey, just ask Wikipedia. Apparently, the people making this measuring cup decided that the people using it in the US where it's being sold would surely intend to use "metric cups" (250 ml). A customary US cup is about 237 ml. Strangely, the legal definition (for nutrition labeling) of one cup in the US is 240 ml. Meanwhile, an Imperial cup is 284 ml. A Japanese cup is 200 ml, which for some reason differs from the traditional Japanese "gō" which measures 180 ml (all of which explains why the cups that come with rice cookers are so confusing).