Excerpts from netbsd: 27-Jul-98 Re: [Q] Keyboard support fo.. Takuya
Koumoto@is.aist-n (1214)

>
http://sakura.aist-nara.ac.jp/Japanese/student/takuya-k/zsconsole6-netbsd-alph
a1.3F.tar.gz

I just had a look at it. Some random comments:
-The way the zs driver is layered could be improved.
  It's not optimal to require a <machine/z8530var.h>, there
  should rather be an <ic/z8530var.h> containing a generic
  "softc" which can be included by bus or machine specific 
  drivers - just like the com and the lance driver are done.
  (currently one can't write an mi driver for -let's say- a
  PCI 8530 multiport serial adapter)
-Would it help to define an event type "all keys up"?
  I don't know much about the lk201 protocol, but it looks
  as if zskbd_lk201_decode() could be made simpler
  if such a thing existed.
-zskbd_keydesc_us[] should go into a separate file
  (eg dev/dec/lk201_keymap.h). I'd also eliminate these
  useless 86 table slots at the beginning by some CPP
  macro (#define KC(n) (0xe000 | ((n) - 86))).
-What is the change to wsdisplaymmap() good for?
  For debugging only, or is there something wrong with
  the logics behind SCR_GRAPHICS?

best regards
Matthias