The description does not really explain how the algorithm works. It doesn't just combine the last/first 1 character, but as many as are common from the beginning and end of the separate words.

For example, 'fat ata tape' => 'fatape'.

The description has no such example, even though it's made mostly clear by the test cases. This makes it a little more challenging.

I'm pretty disappointed to see everyone's solution running in (best case) Ω(n) time instead of (worst case) O(n) and average expected value of O(2) for large random tests.

pig_it("I agree!!!")