Below is an English translation of the notes:

When analyzing only the inner while (excluding the outer for loop):

• Starting from both ends (left = i + 1, right = n - 1): left only increases and right only decreases, so each pointer moves at most n times, giving O(n) for the inner loop. (They never "chase" each other; movement is strictly one direction.)
• Starting with right = left + 1: To adjust the sum, right must be pushed far, and every time left changes, right needs to move a lot again. So even the inner part tends toward O(n^2)-level movement.

Therefore, looking only at the inner loop:

• Both-ends start: O(n)
• right = left + 1: near O(n^2) of unnecessary movement

Notes on sorting:

Arrays.sort for int[] uses Dual-Pivot Quicksort, with average/worst-case O(n log n). Overall time complexity is O(n^2), and including sorting it becomes O(n log n + n^2) = O(n^2).