Don't laugh. I worked on a multi-million dollar project that was derailed for a couple of weeks by an analyst who insisted that doing a bubble sort on unsorted results from an Oracle query would be faster than using a SQL ORDER BY. This was based on a half-remembered class that used bubble sort. Pointing out that bubble sort is usually given as an example of an expensive sort had no effect. Ultimately Oracle won, of course -- it's orders of magnitude faster. But many hours were wasted and tempers flared.

Just curious.. why didn't anyone run a quick benchmark of both methods?

Project politics. We did run benchmarks and Oracle's ORDER BY clearly came out ahead, but the analyst espousing bubble sort was entrenched and dismissed all benchmarks as special cases.