You can probably do worse than to cut mathematics up into "algebra" (not just linear) and "analysis" (calculus). They're almost completely diametric points of v

You can probably do worse than to cut mathematics up into "algebra" (not just linear) and "analysis" (calculus). They're almost completely diametric points of view by the bulk modern mathematical canon [0]. There's a fair point to say that one shouldn't learn so much calculus as to lose your linear algebra, but the opposite is just as true. There are a lot of comments in this thread along the lines of saying that linear algebra is far more useful than calculus in CS. This statement feel not exactly unfair but certainly like it's a reactionary measure. Whenever you have a notion of change you have some kind of calculus underlying it. Even if you end up having to view that notion of change algebraically. For instance, there's a famous functional programming result that there is a well-defined "derivative of an algebraic data type" which allows you to talk about changes in data "over time". I'd also feel very bad for anyone who tried to learn too much probability without studying quite a bit of calculus first. [0] Sure there's discrete calculus and algebra underlies the mechanism of most calculus computation. Likewise there are continuous groups and you can't avoid how topology mixes the two concepts.