Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!olivea!mintaka!bloom-beacon!dont-send-mail-to-path-lines
From: dc@ncd.COM (Dave Cornelius)
Newsgroups: comp.windows.x
Subject: Re: Ncd keyboard num-lock
Message-ID: <9102150057.AA18369@pepper.com>
Date: 15 Feb 91 00:57:59 GMT
Sender: daemon@athena.mit.edu (Mr Background)
Organization: The Internet
Lines: 61

>From: parmelee@wayback.cs.cornell.edu (Larry Parmelee)
>Newsgroups: comp.windows.x
>Message-ID: <52027@cornell.UUCP>
>Date: 14 Feb 91 15:08:17 GMT
>References: <18034@memqa.uucp> <1991Feb12.193608.27897@convex.com> <18285@memqa.uucp>
>Sender: nobody@cornell.UUCP

>On the NCD keyboards, I don't believe you can use either of the lock
>keys as general purpose keys, due to the hardware design.

Incorrect.  There are no restrictions (other than the setup key)
regarding assignment of keysyms or attachments as modifiers.
All the translation is done in software.

>To see what I mean, use "xev", position the cursor in it's window, then
>try a few keypresses of the num lock key, and compare the events seen with 
>those seen when some other "ordinary" key is pressed.
>
>Generally, when you type an "A" (or any other key on the keyboard),
>two events should be seen: a "keypress" event when you first push
>the key down, followed by a "keyrelease" event when you let go of the
>key.  The exact timing of these events relative to the actual physical
>actions occuring vary with different hardware designs, in particular,
>key "auto repeat" can be handled in several different ways, "Shift"
>keys (like "shift", "control", "meta", "hyper", etc) may have special
>behaviour, "Lock" keys...

You're describing the default behaviour of the locking modifiers.
Via xmodmap, the modifiers can be changed.

>Most of the NCD keys generate both a keypress and keyrelease event,
>but the num lock (and I believe the caps lock) are "special" (broken :-)
>and generate only one event for each for each actual physical press/
>release cycle.  Apparently by design, the hardware contains a toggle that
>keeps track of wheither the last event generated by the press/release
>of a lock key was either a "keypress" or "keyrelease" event, and generates
>the other type of event when the key is pressed/released again.

The locking modifiers toggle press and release by design.  They are
special, as you note, but this operation is not broken.  Again, all
this happens to be in software on the NCD, but that's really academic.

>I suspect NCD was forced to use this design in order to have the nifty
>"caps Lock" and "Num Lock" LEDs on the keyboard.

Wrong again.  The LED functionality is completely under software control.
The setup menus (and remote configuration) allow you to change the
meaning of each LED.  Later releases of the NCD server generalize this
ad nauseum, allowing you to have *all* the LEDs act as NetActivity,
or whatever your peccadilloes stipulate.

>Attempts to redefine the NCD lock keys are probably doomed to fail,
>unless the new key is supposed to have a similar locking type of
>function.

RTFM on xmodmap :-)
-----------
Dave Cornelius				Network Computing Devices
					350 North Bernardo Ave
dc@ncd.com   -or-			Mountain View, CA, 94043
{uunet,mips}!lupine!dc			415-694-0675