5.1.2.3 Experiments in Epistemology
Over the course of the research, a more subtle type of task performed by developers emerged.
In pragmatic efforts to generalize structures created from particular courseware,
the tasks began to resemble exercises in "epistemology". For example, hypermedia
interface designers created generalized structures to represent the relationships
among knowledge that could be used across disciplines as demonstrated in the following
description by Schlusselberg. (E. Schlusselberg, personal interview, October 1, 1992)
You create a structure, which you think incorporates everything that
you would ever want to do, and then all of a sudden, you want to add one little element,
which you think is going to be absolutely trivial, and, instead, it disrupts your whole
organization. But it brings insight into how your old structure was wrong, and suggests
ways for improving or making a more flexible structure. I think the development process
is a spiral one, because you continue to go back to redefine and re-implement.
Initially, I just wanted to finish a project. Then I began to realize that there were
structures that repeated. And as I did more projects, it became really important to get
something that I could just plug into an environment. For one project it took me two
months to just build the templates. Then another project came in and in about a week's
time students were using the application. The quick turn around for this second project
was possible because I reused the structures I had created for the first project,
so all of a sudden production became much, much quicker.
When programmers of "microworlds" developed projects based on structures from other
projects, the structures also become abstracted to more generic structures.
This phenomena was elaborated in a conversation with Gregory Jackson, the current
Director of Academic Computing at MIT (personal interview, March 4, 1992):
When Athena was young, there wasn't much courseware for UNIX systems
much less X-Windows Systems, so when an instructor said "I'd like to do something,"
our staff here went to work and did it. There were several things that turned out
not to be interesting to students or faculty and they withered away. Then there
was a fairly large handful that really redefined how something was taught,
or redefined the notion of what you could do with computers. The things that
began to be notable, that won prizes outside, that kept getting used around here
tended to be simulations. For example, there's a number of fluid flow simulations.
If you pull down the courseware list, you will see TODOR. But if you go look at
several other things, you'll find things labeled differently, which are actually
exactly the same software. And they're exactly the same software, because fluid
flow does not just apply to air flow over a foil, it also applies to water flowing
through pipes, electrical fields flowing through space, and microwaves going through
tubes. All of these are analytically similar phenomena, and so one of the things that
happened was that once Earll Murman had done TODOR, other people said, "I can adapt
that to demonstrate this."
Another example is a big thing that happened about three years ago nation
wide. Someone realized that finite element analysis programs helped you
understand all sorts of energy transfer problems. All of the sudden there
were 20 finite element programs. And then the next year, there would be
one for civil engineering, one in thermodynamics, and the next year there
would be a generic finite element program with front ends for all these
different disciplines. There's been this evolution away from courseware as
the focus, and toward applications as what we use for courses That's what
we've predicted is going to keep happening. It is often talked about here,
there are a small number of things, which come up no matter which
technical field you end up majoring in. They have different flavors on
them, but you realize that if you can help kids learn these things that are
generically useful, as long as you put a front end on them, they are very
powerful. One of the keys is that it's a community of scholar's solutions.
Something that increases community. It's generally been true of things
that build community catch on like wildfire.
Over time, structures created through this process may begin to reflect common
knowledge structures across disciplines. This raises the possibility that designers
of courseware are creating laboratories of "epistemology" during their quest for
efficiency.