I want to describe one of the professors I know at the Grove School of Engineering at the City College of New York.
This man, in his sixties, teaches math-heavy courses in the Computer Science department. His course descriptions still mention Fortran, though I think he teaches mainly in C and C++. But mentioning Fortran is good for producing a certain effect, and my guess is he desires that effect. He doesn't hold with electronic bulletin boards and just hands out a multi-generationally photocopied one-page syllabus with the current year's changes marked in pen. It is usually not straight on the page. His lectures are mostly proofs, which he delivers looking alternately at the board or at a corner of the ceiling at the back of the room — though he does engage occasionally in a sort of Socratic baiting of students. ("What is an 'even number'?" Then shoot down all student answers on the way to the truth.) His exams are all problems from the book. He might as well be drawing in the sand with a stick in the time of Archimedes. But you can learn a lot from him if you work hard.
He seems very gruff, but I've spoken with him at length in private and I find him deeply concerned about whether students are learning the material or not. He is particularly unhappy about the fact that the distribution of student grades is bimodal — it is hard to "teach to the middle", he says, when there is none.