your solution is not robust because it uses a space separator that it then splits on to produce the pairs. that's assuming that the input string never contains spaces, which is not true in the random tests. not a kata issue.