What I recall from studying LISP some quarter-century ago: For any recursive solution to a problem, there also exists an iterative solution. Because of the overhead of subroutine calls, the iterative solution usually runs faster and with less memory. Indeed, tail-recursion is essentially a way of turning a recursive solution into an iterative one.

BUT!

The most important measure of speed is the programmer's speed in writing the code, the reviewer's speed in deciding if it's correct, and the maintainer's speed of understanding it. In the cases where they're appropriate, recursive solutions are often much more compact and clear than the iterative solution.

When is a recursive solution more appropriate than an iterative one? If you've got a set of objects and you need to solve a case for each one, and the solutions aren't all linked together, iterate. But if your data structure looks more like a tree or a graph, look for a recursive solution. Given a hard case, can you shave off part of the problem and turn it into a very similar problem, only smaller? Recurse.

throop