readablility most likely

It's been a while but all the words are split, so you'd need to do it for every word separately.

I ran some benchmarks. This solution is faster than the "efficient" alternatives for "small" to "moderate" length strings; but as the input length increases, the difference between them shrinks until the efficient methods scale better for much larger strings. If you technically know you'll be dealing with short-ish strings, this can actually be faster.