Djula Djarmati wrote:

> This is my first posting to the PICLIST so forgive me if something
> goes wrong...

  ... Nothing did!  Excellent!

> This might work, but the decoding is somewhat complicated.

  It's on a par with my suggestion.  It's *probably* easier than you
thought, just takes a little cogitation.  My cogitation time is severely
reserved for the next three days, so maybe (maybe!) I'll have further
comment/ suggestions later in the week.

  Perhaps more to the point, it doesn't match off-the-shelf keypads,
*most* of which are either matrix, matrix-with-common (old-style
telephone), or single-common.  All switches separate is an oddity.

>  Can't decode simultaneous keypresses but can detect (I think) and
> ignore them which is not so bad.

  It's on a par with my suggestion.  It obviously won't detect the key
with a diode when the corresponding key without is also pressed.

> Use proportional font when viewing.

  Hey, we're the PIClist!  Hmmm.  Mind you, some of us apparently work
for companies who require us to use brain-dead software at all times.
That's not promising...  :-(     ("All employees please check in
intelligence at cloak room upon entry.")

> 4 I/O
> 16 keys
> 6 x D
> 4 x R

  I *strongly* recommend the convention of always using pull-ups rather
than pull-downs on all switches.  PICs (and other comparable MCUs) have
built-in pull-ups although they are probably not suitable for user-
interface switches which should probably also be isolated from the PIC
by, say, 22 k ohm resistors.
--
  Cheers,
        Paul B.