[comp.human-factors] Handicapped users

jones@pyrite.cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) (06/13/91)

At various points in my career, I've been involved with helping handicapped
users of computer systems, in one case, a blind user using a speech
synthesizer and Braille printer (built from scratch), and in other cases,
people with motor disabilities.

My general observation in all these cases was that for all these users,
fancy input devices and classy display management were major problems.
If software designers want to assure that their software is "handicapped
accessible," I have the following two suggestions:

  1) Make sure that everything can be done from the keyboard.  Do not
     require a mouse or a touch panel!

     This means that there should always be something equivalent to arrow
     keys for moving the cursor, and it means that menu items should
     always be accessible as keys on the keyboard. 

     For a person with motor disorders, moving the cursor by a sequence of
     arrow key presses can be painfully slow compared to pressing a single
     key corresponding to a menu item.

     For a blind person, finding the cursor on the screen is an awesomely
     difficult task, even with some of the marvelous tools now on the
     market.  It's far better to simply use keys on the keyboard to select
     menu items.

  2) Make sure that text is presented in a linear order.  Speech
     synthesizers are inexpensive, and it isn't hard to fix one up to
     intercept text as it comes to the display, but such interception is
     only really useful if the text is presented in reading order.  Many
     window managers are particular offenders here, with different subparts
     of a window hierarchy (for example, a menu) popping up in essentially
     random order.

     In any case, a decent speech synthesizer interface to a display
     manager should allow brousing of the display after it is built,
     stepping the cursor from text item to text item in two dimensions
     and playing back the text, optionally in continuous speech mode,
     word-at-a-time mode, or letter-at-a-time mode.  Peter Maggs built a
     speaking Plato terminal that did this 15 years ago, and it worked
     well with the Braille printer that Andrew Appel and I built.

In addition, the following keyboard characteristics are desirable:

  1) Limited use of shift keys.  People with only one hand and people with
     motor control disorders can have big problems dealing with more than
     one key at a time.  Locking shift keys solve the problem for case
     selection, but on some modern systems, there's only an alpha lock that
     has no effect on punctuation keys, and nobody I've seen makes a
     standard keyboard with a control-lock or meta-lock key.

     There are simple solutions we found to the locking shift problem for
     one user, we simply gave him a steel ball he could set on the shift
     or control keys to hold them down.

  2) Make it possible to disable auto-repeat!  Some people with motor
     disorders have to make separate and distinct efforts to press and
     release keys, and thus, they frequently hold a finger on a key long
     enough to activate the repeat feature.

  3) Avoid use of nonstandard key spacing!  The standard spacing between
     rows and columns of keys on a typewriter keyboard is 3/4 inch, and
     this spacing is big enough that it is possible to make a keyboard
     mask with one finger-sized hole over each key.  Such masks allow
     people with only spasmodic control over their hands to make quite
     effective use of a standard keyboard.  They just slide their hands
     over the mask until a finger is positioned over the right key, then
     shove the finger through the mask and bash that key, with no risk of
     hitting the neighboring keys.  Custom masks (in lots of 1) can be
     commercially machined for only a few hundred dollars.

					Doug Jones
					jones@cs.uiowa.edu