Sure, but why does it feel like we're going around in circles here? Labeling that an "asymmetric algorithm" is technically correct but it implies it's some rarer special case. Whereas it's literally just the common debouncing technique, at least when I was learning embedded development ~20 years ago.
And there's no reason for a keyboard to be using anything different. As I said, the real delay factor with keyboard is matrix scanning. If there is a keyboard that has 30ms of latency to register a keypress, I would guess that a ~400Hz (sqrt(104) -> 11 columns?) scanning frequency was as good as could be handled by whatever early cheap USB microcontroller they used, and its designers figured that was good enough for productivity use.
And there's no reason for a keyboard to be using anything different. As I said, the real delay factor with keyboard is matrix scanning. If there is a keyboard that has 30ms of latency to register a keypress, I would guess that a ~400Hz (sqrt(104) -> 11 columns?) scanning frequency was as good as could be handled by whatever early cheap USB microcontroller they used, and its designers figured that was good enough for productivity use.