From Phil Agre's Red Rock Eater News studying beginners is not new, and that user-interface people have
talked in terms of the user's "model" of the machine, as opposed to
the "correct" model that is embedded in the design of the machine.
The question then arises of what beginners' models tend to be like,
and how to make the beginner's model and the correct model align.
That alignment could come about in different ways: teaching strategies
that install the correct model in users' heads, or design strategies
that make the correct model transparent to users in the first place.
It's true that I am hardly the first person to think of investigating
users, and if I were writing an academic paper I would be responsible
for citing all of the relevant literature in useability and elsewhere.
My point was different. I want to suggest a naturalistic study of
users that places them in the full context of their lives. I also
want to encourage a polemical view of beginners as rational people,
and of experts as victims of brainwashing. And I am not exaggerating
this for rhetorical effect. I honestly believe that the computers
we have today incorporate ideas about people and their lives that are
radically false. We as experts have to gotten used to the pathologies
that result from these mistaken ideas, and the fine, naturalistic
detail of beginners' experience really is our best way of remembering
what we have lost.

I also have a problem with the word "model". I may not disagree with
it, but I think that it can be misleading. First of all, it suggests
something that is unified and coherent in people's minds. If so then
I am not sure that it makes sense to speak of beginners as having a
model of a computer at all. Beginners do exhibit very characteristic
forms of reasoning, but these can almost be defined as the absence
of the kind of model that experts are said to possess. Alright, you
might say, so beginners need to build a model. What's the problem?
The problem is that beginners are not empty. They come to computers,
for example, with elaborate expectations derived from other media
(on this topic see Jay David Bolter and Richard Grusin, Remediation:
Understanding New Media, MIT Press, 1998). They are also familiar,
unfortunately, with all manner of cultural constructions of computers
-- including all of the fiction that anthropomorphizes the computer
and treat *it* as unified and coherent, which it's not. The idea that
computers have all kinds of social contradictions running through them
is incomprehensible against that cultural background.

The notion of a "model" also focuses on the mental lives of individual
users, and thus distracts attention from the larger contexts in which
individual users are embedded. I enumerated several such contexts.
One is the microsociology of interactions between beginners and experts
-- for example, when the experts "help" the beginners by taking the
keyboard away from them, talking technical gibberish at them, changing
their configurations, and generally disempowering them in every way
they can. Another is the local ecology of knowledge and practice in a
home or office; this ecology can be supportive and constructive, or it
can undermine the beginner's capacity to learn. Each case will need
to be investigated on its own merits, but we will usually discover a
combination of both. Yet another context is the computer industry and
particularly the competitive dynamics of standards. Economic pressures
both for and against compatibility are easy to find, and every computer
is a historical conglomeration of different architectural choices and
competitive configurations over many years.

The most important intuition for the naturalistic study of users, in
my experience, is to look in the margins. Experts imagine themselves
to be inside the machine; they *do* have a model of how the machine
works, and they "see" the machine purely in terms of that model.
If the machine fails to behave according to the model, experts have
no problem simply restarting an application or rebooting the machine.
Beginners cannot distinguish between proper and improper functioning
of the machine, and they find this behavior on the part of experts
bewildering. Experts see only the picture; beginners see only the
frame. Experts cannot see all of the detailed work around the edges
of the system: getting an account, figuring out which machine you're
allowed to use, turning it on, logging in, password problems, getting
your hands registered the right way on the keyboard, learning what
you do with the keyboard versus what you do with the mouse, getting
help, knowing what things are called, knowing whether your work has
been saved, logging out, shutting the machine off (and whether you're
supposed to be shutting it off at all), and so on. Very few people
have ever investigated what really happens in college computer labs,
for example. One of them was Steve Strassman, who in the mid-1980s
wrote his senior thesis at MIT based on six weeks of observation in
the computer lab for the 6.001 introductory programming course. His
work is unpublished, to my knowledge. But if you're concerned with
access to the Internet, in my opinion the most important problems are
all in this zone of peripheral vision -- the area that is invisible to
the brainwashed experts. The cost of hardware and software is not the
big problem in my opinion, simply because the cost is dropping like a
rock. The problem, rather, is with knowledge and culture -- and with
the culture of expert knowledge that makes life harder than it has to
be for non-experts.

**
"