Question

When the authors implement the mutex part of serializers, they use a list called cell. But the list only contains one element, so why not just use a variable?

Answer 1

Because a variable isn’t a first-class value that you can pass to another function. In 3.4, the authors implement a make-mutex function that uses clear! as a helper function, which takes a cell. If the cell were represented by a mutable variable, then clear! would have to be defined inside make-mutex! to close over that variable. The same goes for the test-and-set! helper function.

They also could have used, say, a box instead of a cons cell.