So anyways, Knuth gave an interview maybe 10 years back where he lamented the lack of dexterity in math in American undergrads. He said even Stanford has to offer "remedial courses in math" for CS undergrads, because there is an "interest vs ability" mismatch. At that time ( during the 1999-2000 tech boom ), a lot of Stanford undergrads were indicating CS as their major. So they clearly had the "interest". But upon testing them, it was clear they didn't have the "ability". Hence this interest vs ability mismatch, and a need for remedial math courses at Stanford. At the time, I was studying CS with a bunch of Russian, European, Indian & American classmates. I found that Russians & Eastern Europeans tend to favor clever 1-line math code. They themselves were trained in physics and math & looked at CS problems as essentially a laborious function-mapping exercise. Whereas the Americans invariably wanted "readable code" and ended up writing verbose switch-cases and if-elif-endif clauses ie. essentially unrolling for-loops. Knuth's Concrete Math ( http://www-cs-faculty.stanford.edu/~uno/gkp.html ) was a sort of bible in my days. You had to know generating functions and recurrences and even stuff like complex powers. Like for instance i^i ( i raised to i ) leads to an infinite set of points with real and imaginary parts, scattered all over the 2D plane. You had to know cosets and Lagrange's theorem and set partitioning using cosets. Nowadays people simply enumerate everything as there is no penalty ( in performance terms ), so CS problems have a "select 1 answer from set of possible answers" flavor, rather than "contrive a function to compute the answer" flavor.