In lecture #2.6 Professor Leonard, in his proof of the chain rule ends up doing something like this:
dy du
-- * --
du dx
He then cancels the 'du' top and bottom to complete his proof. But the 'dy/dx' form of indicating a derivative is surely not really a fraction is it? If we use the Newtonian notation there is no 'fraction' at all. The chain rule certainly works and you can see that the answer is the same as if you just distribute a polynomial raised to whatever power, however the above 'cancellation' seems very strange to me. It it were a fraction we'd end up with 'something over something' but we don't -- there is no fraction so how can canceling work? He just glosses over this so I hafta ask you guys.