Repeatedly swap neighbors that are out of order.
Walk the array pairwise and swap any two neighbors that are in the wrong order. Each full pass bubbles the largest remaining element to the end; after n−1 passes the array is sorted. An early-exit flag lets an already-ordered input finish in a single pass.
Teaching the idea of comparisons and swaps — it's the simplest sort that actually works. Also fine on tiny arrays or data that's already nearly sorted, where the early-exit flag kicks in immediately.
Anything production-sized. Quadratic time with constant-memory shuffling makes it strictly worse than insertion sort for the same niche, and dramatically worse than any O(n log n) option past a few dozen elements.