norman@sdics.ucsd.EDU (Donald A Norman-UCSD Cog Sci Dept) (11/22/88)
Marty Brilliant brings up two important issues (in his contribution to
the "focus window" problem).
First, he points out that the error of typing into the non-selected
window is more likely to occur with experts than with beginners. That
is a good point. I have found that slips, in general, occur mostly
once you have become expert enough at a task that you no longer focus
your attention on it but instead, do it automatically while
simultaneously attending to other things. And the focus problem is,
by definition, a slip. But this observation also means the design
effort should be focussed on the expert requirements, not the novice
ones.
Brilliant's second point is to argue for systems that infer your
motives. In his words:
Experienced users could work faster if they did not have to ``formally
select'' the window to be worked in. A smart system would know which
window the user is looking at, just as a person at a meeting can tell
which member or members of the group the talker is talking to.
Here, I wish to disagree. This is actually part of a long standing
debate (in the Norman & Draper UCSD book we called this the "tools
versus intelligence" argument). I won't review the issues here except
to say that it would be most difficult for a system to know in which
window I intend my operations to have effect. The place I am looking
at is NOT a valid clue, since I might be copying from the other
window, or just thinking about it. (I am just now looking out the
window at the surfers climbing down the cliff to the ocean while I
type this message -- what can you infer about my intended actions from
that observation?) Even people have great difficulty inferring the
intentions of one another. When they are correct, then, indeed, life
moves more smoothly. But when they are incorrect, the resulting
confusions and misunderstandings can be so serious as to far overwhelm
the occasional benefits.
The problem is that intentions are invisibly nestled within the mind,
and words, eye gaze, and other actions do not necessarily reflect the
contents of those intentions.
However, I agree with Brilliant that systems might do much more to
anticipate the user. Quite often the intended window is easily
determined by the form of input: typed text, for example, might be
appropriate only to one of the winodws, whereas mouse clicks might be
appropriate to only another one. In addition, the content of the
typed string can also be informative. Thus, shell commands are
usually only appropriate to the shell command listener. Thus, while
using a text editor and a drawing program in two different windows,
the form of input may often unambigously specify the target program.
It would also be valuable if the system would retain a history of the
inut, so if I discuvered that my actions got interpreted by the wrong
window, not only could I undo the unintended effects, but I could
replay those actions into the correct window.
etc. and etc.
don norman
Donald A. Norman [ danorman@ucsd.edu BITNET: danorman@ucsd ]
Department of Cognitive Science C-015
University of California, San Diego
La Jolla, California 92093 USA
UNIX: {gatech,rutgers,ucbvax,uunet}!ucsd!danorman
[e-mail paths often fail: please give postal address and all e-mail addresses.]