Hhmmm. Keyboard.C listing of this note tells, that P2.2 and P2.3 are configured as push-pulls. But doing this, then simultaneously pressing K1.0 and K2.0 will short circuit P2.2 by P2.3, or vice versa. This will nicely clean your switch contacts. ;-) What a shit!! How I love these stupid application notes...
A much better approach would be to configure P2.2 and P3.2 as quasi-bidirectionals.


Are you assumning that the guy who wrote this application note knows what a "dry contact problem" is??
It's true that quasi-bidirectional port topology provides an internal active pull-up, the weak one, which looks like a pull-up resistor of some dozens of kOhm. But this logical 0 input current (when the input gate has skipped), which is only "powered" by the very weak internal active pull-up, is so little (looking like some hundreds of kOhm!), that an old or cheap switch contact can have problems to maintain low impedance. As consequence, the input gate can skip back again, resulting in an oscillation of logical input state...

A stable and strong enough external pull-up is definitely the better solution than to trust this weak and very weak internal pull-up.

Kai