AP Test's Recursion Examples: An Exercise In Awkwardness

theodp writes "Yet another example of how AP exams are loaded with poor coding practices," quipped Alfred Thompson, referring to a recursive code example that prints the numbers 0 to 6, which was posted to the (closed) AP Computer Science Facebook group. "We are often forced to use code examples that are not ideal coding practice," Thompson notes. "We do that to make things clear and to demonstrate specific concepts in a sort of isolation that we might not normally use. We seem to do that a lot with recursion because the examples that require recursion tend to be fairly complex." So, while asking students to use recursion instead of a loop to print '0123456' serves the purpose of teaching recursion, Thompson opines that it's also a poor example of code practice. "Someone raised on functional programming where recursion is a pretty standard way of doing looping might disagree of course," he adds. "There is a saying that when all you have is a hammer all your problems look like nails. This seems, in a way, to be the case with recursion and loops. If your first tool of choice (or what you have learned first) for iteration is loops you tend not to think of recursion as a solution. Similarly if you start with recursion (as is common with functional programming) you are a lot more likely to look at recursion as a tool for iteration." So, do you tend to embrace or eschew recursion in your programming?

Comments

Re:What do you expect?

[vux984]

It's like teaching a kid long division using 6000 / 10, and disparaging the example saying, "yeah, but you would never use long division for that." Well, no shit, Sherlock. You're teaching mechanics.

But the mechanics of long division in that example are reduced to trivial busywork. That not a complicated enough example to even really see the mechanics properly.

It'd be like using Newton's method on a sample problem that converges to an absolute final answer after a single iteration, or teaching someone how to calculate the area of a trapezoid and using a square as the example problem to solve. Sure the formula works on a square but its not really instructive.

Application is for another lesson. Maybe even another class.

I'd argue that the solution to a problem is a lot easier to understand if you're given a context where the solution is needed FIRST. Starting with a degenerate problem that reduces to a trivial application serves to obscure the 'point' of the solution method.

Re:Never used recursion

[Anonymous Coward]

I use recursion in VHDL quite a lot, since functions in that language are completely pure, it is often the right thing to do.
Also when creating hardware there is a step called elaboration, where function are executed (optimised away) before a hardware netlist is created.

Yes, but...

[bradley13]

Yes, it is a test question. Yes, if you understand recursion, you can answer the question. However, it is poor code, because the recursion is not tail-recursive; anyone who uses recursion with unknown values ('n' in this case) will write a tail-recursive function.

Since it's a totally artificial question, there is no reason that they couldn't have used a tail-recursive function. Lots of students won't know the difference, but crappy code like this is a stumbling block exactly for the students who really do understand what's going on.