[Noisebridge-discuss] Little red payphone

Seth David Schoen schoen at loyalty.org
Sat Dec 13 02:09:12 UTC 2008


esc writes:

> So seth and I spent way too much time last night on the phone (ha!),
> and made a lot of progress.
> 
> Check it: https://www.noisebridge.net/wiki/Payphone

Here is an interesting problem.  Someone proposed that we should put the
keys in the payphone in an unusual order -- I think ideally each key
would still be correctly labeled but it would be located in an unusual
place.

I kind of liked the order

150          (divisors of radix 10, sort of)
248          (powers of two)
369          (multiples of three)
#7*          (lucky number seven)

This has the nice property that 1 and 9 are in their original locations,
so that if someone wanted to dial 911 on the payphone without looking at
it (or without being aware whether the keys were correctly labeled), it
would still work.

Unfortunately, after our reverse engineering session yesterday, we
learned that the output of the keypad is in the form of three column
pins and four row pins; when a key is depressed, the corresponding
column pin becomes electrically connected to the corresponding row pin
until the key is released.  (We've also mapped which pin is which, and
where the corresponding pins are located on the keypad interface of a
regular land-line phone that I took apart.)  Ignoring the fact that
the real order of the pins on the circuit board is kind of scrambled
from this, you have something like

       keypad row           keypad col
  1st  2nd   3rd  4th       left   middle  right   (the fourth column
   |    |     |    |          |     |       |       with DTMF keys
   |    |     |    |          |     |       |       A, B, C, D is
   |    |     |    |          |     |       |       not present on the
                                                    keypad)

   |    |     |    |          |     |       |
   |    |     |    |          |     |       |
   |    |     |    |          |     |       |
  1st  2nd   3rd  4th       left   middle  right
   DTMF generator row         DTMF generator col

and normally you would connect them correctly so that the DTMF
generator actually knows what key you pressed on the keypad,
but you could connect them in some other order if you wanted to.

This means that connecting the pins in a different order can only
swap entire columns and/or swap entire rows.  (You could try to
transpose columns and rows but since there are more rows than columns
I think this would be a nuisance -- some of the keys would no longer
correspond to an addressable location, so you couldn't register when
they were pressed.)

We could still scramble the keys under this constraint, but if you
think about it a bit, it becomes a very severe constraint if you also
add the condition that 1 and 9 should remain in their original
locations.  Since 1 has to stay in the same spot, the leftmost
column and the topmost row have to stay where they are; since 9 has
to stay in the same spot too, the rightmost column and the third
row have to stay where they are.  That means that only the middle
column can move (but there's nowhere to move it to, since neither
of the other columns can be interchanged with it), and only the
second and fourth rows can move (but the only way to move them would
be to swap them).  Under these constraints the only two valid keypad
layouts would be

123
456
789
*0#

and

123
*0#
789
456

That's not a very exciting change, in my opinion.

Does anyone think it would be all right to move 9 and 1 out of their
original locations?  Then we could get a lot more possible modified
keypad layouts, but somebody who wanted to dial 911 on the phone would
have to think harder about the process (including being aware that
the locations of the keys were really moved, not just their labels).

-- 
Seth David Schoen <schoen at loyalty.org> | Wol dir, werlt, daz du bist
     http://www.loyalty.org/~schoen/   | also freudenriche!
     http://vitanuova.loyalty.org/     |           -- Carmina Burana



More information about the Noisebridge-discuss mailing list