🚀 go-pugleaf

>From Admin.JQJ@SU-SCORE Wed Mar 10 12:14:01 1982
Our discussion of keyboards on E-P seems to be focussing our attention away
from an equally important issue -- the (growing) literature on non-keyboard
input devices.  My cursory reading of that literature indicates little
agreement of late on what the best such input devices are (mouse?  light
pen?  knee pedal?  hand-heel pressure pad?  auxiliary function keypad?
ESP?).  This raises two issues I'd like to see discussed; can any of you
offer any comments on the following?

  [1]	What methodologies should we be applying to evaluate different sorts
of non-keyboard input devices?  Can we develop a cognitive model of their
use (analogous to the Card/Moran/Newell or Rumelhart/Norman models of
keyboard use) that would allow us to investigate comparisons?  Can we ever
reasonably measure their relative benefits outside of the context of a
particular editor?  If not, we're in trouble, since very few editors support
more than one type of pointing device, making direct experimental comparison
quite problematical.
  [2]	Imagine an extensible editor such as Emacs which supports a variety
of non-keyboard input devices.  What primitives should be available to the
Teco, Lisp, Mlisp, or whatever programmer to support this variety?  My own
bias is to hide from the programmer as many of the details of the interface
as possible, just as we hide differences in terminal output functions.
	A typical problem:  mouse-like pointers are normally implemented as
essentially half-duplex (local to the workstation) for speed of response;
even so, they are often quite slow (an Alto can't keep up with a rapidly
moving mouse).  How should we synchronize the mouse's cursor with the
ordinary character-cursor maintained by the editor on the remote host?
	Another problem:  such devices, in addition to providing a pointing
mechanism, often provide additional keys, but the number and placement of
such keys varies widely.  How to standardize on meanings across devices?
-------