That one's Haskell version asks for fib(1000). JS version only asks for fib(40), but that's still enough to prevent the truly naive implementation.

A good idea...

So is this issue.

I like your solution @ooflorent! It could be optimized with a cache :)

No need to argue, just benchmark and see ;-)

A couple things to note:

1. At the moment, most major browsers do not have TCO properly implemented (https://kangax.github.io/compat-table/es6/)
2. The wikipedia article for Levenshtein distance states (in regards to the recursive method):

This implementation is very inefficient because it recomputes the Levenshtein distance of the same substrings many times.

So, while @ooflorent's solution is very clearly readable, it is certainly not a best practice.

team is a collective noun and it is common (primarily among speakers of British English) to refer to them as a plural. not incorrect, just different usage.

Isn't there something like hidden in the test framework? Or is Haskell the only static language on CW where you can forbid modules/functions?

That was actually the first thing I noticed. But didn't expect anything better from people writing JavaScript. Garbage in, garbage out.

Readable would be nice (:

thanks for this valid number list, now it's much easier to write my own tests

While you're at it, what about the Python/Coffeescript version? (Haskell translation uses randomized tests)

Thanks!