VF>
  >                                    I suspect its probably some bad
  > implementation of keyboard reads that does something goofy with 
nterrupts
  > or polls that I, not being a programmer, haven't got a clue about.
VF>
It's not always software trouble.  In fact in the case of the AT keyboard
it is very rarely software trouble, and almost always hardware.  When
keyboard software bugs do happen, they don't happen intermittently and
"sometimes".  They happen consistently and always.  Also, like the "Caps
Lock" bug in IBMKBD.SYS for OS/2 Warp 3, they are fixed pretty smartly
too, because they happen to _everyone_.
Always bear in mind that the AT keyboard hardware design is not the best
in the world.  In fact, it's pretty awful (it's as if they took the worst
features of RS232 and Ethernet and combined them).  There's plenty of
scope for electrical failure and synchronisation problems, especially so
when manufacturers cut costs and corners.
Often the keyboard circuitry can become out of step with the motherboard
keyboard controller, and the only way to fix things is to unplug and
reconnect the keyboard (which resets the keyboard circuitry and gets it
talking again).
One motherboard manufacturer (who shall remain nameless) has obviously
heard from users often enough about this that it mentions in its technical
support documentation that keyboards can become unresponsive because of
poor electrical grounding of the keyboard connector causing static noise,
and recommends a "finger stock" on the keyboard connector to ground it to
the machine chassis in that case.
  > JdeBP <
___
 X MegaMail 2.10 #0:
--- Maximus/2 3.01
---------------