I ham-fistedly deleted the original message, but:
On Sun, 23 Jul 1995, Paul Haas wrote:
> On Sat, 22 Jul 1995, jory bell wrote:
> > I came across a reference to a "2 of N" keyboard scanning method, which
> > uses less io thant the typical xy system.
>
> > Maybe I am in the dark ages here, but can anyone tell me the
> > hardware/software basis for this?
>
(Paul Haas give good idea)

I've seen and used this scheme before: both in a stone-age Motorola
remote-control transmitter chip and in a stone-age DTMF encoder chip.

Both chips required (any) two pins to be brought to ground to send the
corresponding code.  At the time, I assumed the Motorola chip did it as
a power-saving measure: no keyboard scan is required.  I assume the DTMF
chip did it that way because that's how ancient DTMF keypads were wired.
(back in the days when they used big inductors and capacitors, and
were expensive, heavy and huge: circa '73-75 or so)