It would be a combination look-up table and a simple mechanism to store the previous keystrokes and the number of times it has been pressed. If the current keystroke doesnt match the previous, the count is reset. If its the same inc the count and if it exceeds the number of possible values, it is reset. This count is used it to determine the next character from the "string" (the byte array) picked up from the look-up table.

This will be slightly more difficult if all keys don't have the same number of possible values (e.g. key1->ab, key2->cde, key3->f etc)

kundi