zlacker

All of these have the same worst-case algorithmic efficiency, O(n). The difference is the best-case efficiency. The "optimized" version in the article is O(1). Your solution is still O(n) best case.

The optimal solution will depend on the data. If most strings aren't palindromes then optimizing the best case is likely the better approach. (Example: You are adding an easter egg which will trigger on "random" user input.) If palindromes (or near-palendromes) are common than your solution will be faster as the slope is lower.