-or-

The instructor covered far more material than even he expected every student to master; however, a team of engineers out of one of those classes would collectively have mastered a lot more than if the course was designed so that everybody could get an A. (Likewise even for individual students who excelled in the course.)

I'm speaking from experience from a few classes taught by a physics professor at my university. Granted, he had a fixed grading scale, no after-the-fact curving, but the basic gist was that if you had a 70% average on his nasty tests and a 90% on homework, you would probably get an A. In his first-semester "Physics for Scientists and Engineers" course, the exam average was often under 50%. The distribution was typically bimodal in that course, whereas the Classical Mechanics course I took from him later had slightly higher averages and closer to normal grade distributions.

He wasn't afraid to fail people, either, and if you weren't willing to put in many hours a week figuring out the homework, you were likely to fail. Anyway, his choice to teach more material than he expected everyone to master (and the fact that he was a great teacher, with absurd office hours -- you could often find him in his office until midnight or later) makes me very glad to have taken those courses from him, rather than someone else in the department, despite that the work load pretty much sucked the life out of me those semesters.