Thanks, fun kata!

that's a neat solution

Your foldl is crazy :P

Cool, I just published a quick fork :)

There is a merge conflict in the description. Can you fork this and save the fork? That should update the description to the latest. (You may need to make a trivial change to the code, like add some whitespace, in order to save)

Hmm, good point. I didn't expect people to hack into it this hard though lol!

I'll write something up when I've invalidated these kinds of solutions, dynamically generating predicates seems like a good idea.

Alright, I've changed the description to clarify what kind of predicates should work with iss/2, and added cases for testing expressions with other non-arithmetic predicates! Hopefully this makes the kata a bit clearer.

That's a fair point, I'll try adding new custom tests.

The only thing is I'm afraid of solutions passing right getting invalidated by some reasonable new sample usages though (I know X iss length([1,2,3,4,5]) invalidates all solutions right now).

I'm not sure what to do about that - if I should limit tests to cases that work for current solutions or not. Maybe it is better with these to make it more robust?

EDIT : Maybe I can keep the initial constraint that user defined arithmetic predicates take only numbers as inputs (like is/2) and it'd be fine for this kata

If the core part of the task is to implement iss/2, it should be tested separately. All the sample tests do is to test iss/2 along with fib/1 and fib/3, which couples everything together in the tests and makes it impossible to test the behaviour of iss/2. (There are already 2 lines of example usage of iss/2 in the description, they really should be put in the sample tests.)

Nice catch! I'll fix that right now, and re-read my desc just in case.

A minor description issue: iss/3 should be iss/2 in the first paragraph of Task Part 1 - meta-ish programming.