Re: [seul-edu] Discuss: How software educates.

[Ramin Miraftabi <ramin@cs.joensuu.fi>]

Finally an educational thread that I feel I can contribute
to. Teaching someone less than a University student is something that
I really do not know anything about.

I'll take this to remind you all that I'm researching Computer Aided
Instruction, especially from a Computer Science Education point of
view, but I have studied other uses as well.

* Kevin Turner (Kevin.Turner@oberlin.edu) [991130]:%y23Tue, Nov 30, 1999 at 01:51:13AM -0500]:
> Thinking about what sort of educational software one might write, I
> encountered two different categories: software that assists
> teachers, and software that takes the role of teacher.

Pitfalls in both applications. This is in many ways a traditional
division into two categories. Especially the software that takes the
role of the teacher is something that I'm especially allergic to (we
are talking about Intelligent Tutoring Systems here). Some of the
problems with ITS is that they rely too heavily on a student model
(and usually also a model of an expert knowledge) and try to create a
complete bug catalogue with which to compare the students errors. A
solution to this problem is seen in creating 'intelligent'
applications which do not rely on bug catalogues (at least try to make
them complete) but do try to analyze what the student does and how to
correct it. Another pitfall of most ITS is that they give their
'guidance' too strongly to the student, in many classical examples
(like the Geometry tutor) the student barely has the time to make a
mistake when the tutor kicks in. This is, IMHO, a very bad thing.

The main problem with software that assists the teacher is that if
used alone by the student, there is no way of ensuring that the
student achieves the correct understanding of the material.

Naturally there are many good examples of both uses, and especially
when used by a teacher to demonstrate a chemical reaction or an
algorithm's function (which is what I've worked on) the teacher can
ensure that the students have a chance to form a correct mental
representation of what is happening. With ITS the next step is
agent-based systems. In this case by an agent I mean a piece of
software that can with some degree of intelligence follow what the
student is doing and where the mistakes lie (in algorithm animations
or anything based on mathematics this is a relatively simple
task). In many cases the chance for the student to see that everything
did not go correctly will give the incentive to see where the mistake
was made. If the student cannot identify the mistake the advisor (not
tutor or instructor) can give a hint. Naturally, if needed, the
advisor will do more than hint, but at all time the control must stay
in the students hands. (This is getting long, so I'll stop this point,
if someone is interested, I'll try to see if I can publish a paper on
the topic on the Net. The problem is that we have some new research
going on with the topic that is yet to be published so we can't throw
all of our ideas our to the world, yet. The joys of the scientific
world.) 
> 
> An application for assisting teaching would be like a highly evolved
> blackboard.  It would have more knowledge and provide more complex
> feedback than a blackboard, but it would only come alive to students
> under the direction of a mind that knew how to address them.  Like the
> blackboard, the teacher uses it to illustrate a point...  Sure, you
> could use chalk, but chalk might not be the medium of choice for
> visualizing how the gravatational attraction between moving massive
> objects changes over time.

Visualizations have been proven to be a good thing. In many cases
creating a visualization is useful if you can justify the time
spent. Visualizing a three element array's bubblesort (or a simple
chemical reaction for that matter) can possibly be created much faster
and equally well on a blackboard, but more complex algorithms or
reactions need a computer to create the visualization for it to be
feasible. Think of drawing any sorting alogrithm on a 100-element
array :). 
> 
> I'm thinking here of relatively free-form simulators of various sorts.
> Where a wide range of interaction is possible, but it helps to have
> someone to point out where it is that the interesting relationships lie.
> Yes, students are encouraged to use the same tools after class to
> explore the world, but it would be of limited use to them without having
> at least gone through a tutorial.  Conway's "Game of Life" is a
> fascinating environment, but how many students would discover floaters
> on their own before they got tired of seeing cells die and left?

We must remember however, that visualizations have not been proven to
really speed up the learning process. What they usually do is help
understand what is really happening which then helps them apply the
knowledge they already have when studying a new algorithm or chemical
reaction. Many visualizations also help motivate the student, who
wouldn't like watching whirling things on the screen? 

> 
> Yes, the presentation of tutorials for these tools could be computerized
> too, but that's still a sharp contrast from a piece of software that is
> expected to *be* the teacher.  With the computer as the teacher, it is
> not only responsible for presenting the material, if it is to be
> anything other than a glorified PowerPoint presentation, it has to
> assess and adjust for the student's level of mastery over the material
> as well.

See my points above...

> 
> That may be doable for simple stimulus-response style teaching, things
> like multiplication tables and the capital of Arkansas...  and that has
> its place, dull though it may be.  But I don't see software taking over
> for teachers where a deeper understanding is required...  For things
> like Gestalt-style teaching and teaching by discovery, the teacher needs
> a higher level of mastery over the material and a more intimite
> relationship with the students than I perceive today's machinery capable
> of.

Yes and no. I agree wholeheartedly with the fact that a computer
cannot take over the role of the teacher, but it can make the teachers
task a lot easier. What I see as the role of intelligent applications
(or agents if you will) in education is as a tool to help control the
students understanding while away from the teacher, i.e. not under the
eyes of the teacher. This would also free teachers (in limited
classroom settings) to help students with problems while enabling
faster students to go on to more challenging topics. In other cases
they can be used to facilitate information retrieval (help and filter)
when students are working on their own projects. I hope that I'll be
able to shed some more light on what I mean sometime later, if there
is any interest.

> 
> ...
> I then talked to a friend of mine who tutors math at a local elementary
> school.  She said that she saw the role of educational software as the
> same as her own: to convince students to spend time with the material.
> To make it fun.  Edutainment.  Presentation and assesment are Somebody
> Else's Problem to be dealt with inside the classroom, outside class you
> just want to keep their minds working on the subject for a while longer,
> be it through games, competition, or whatever...

Urf... Edutainment... (shivers) Why does the subject area need to be
made fun? My biggest irk with teaching nowadays is that learning
something has to be fun. That is all right, but everyone seems to
forget that the teachers task would be a lot easier, if the students
are instilled with the idea that learning in itself is fun. Not only
should a topic be fun to learn (whirly things and whiz-bang effects)
but learning something new should give the feeling of: hey, this *is*
fun. A good example of what edutainment can lead to is some history
books for elementary and junior high that I've seen lately in
Finland. Even my wife, who is a history major (MA level) had a hard
time in figuring out what the books where trying to teach. Mixing
legends with facts and superman flying aroung Acropolis does not lead
to a very positive learning experience, even if it was fun at the time
and held the kids attention for a while.

Learning isn't always fun, but we should try to remind the children of
the exhilerating feeling that comes when you've really understood
something. Think of coding as an example. Every application you
program presents a chance to learn. When you face a problem (a bug)
you have a hard time fixing it (at times, sometimes it's a breeze as
well) and the 'learning process' is not fun at all. But what do you
feel when the problem is solved? The feeling that comes is a good
one. This is what I feel we should concetrate on, not (only) making 
learning the content fun, but making the experience of learning
fun. And yes, I do know that it's a lot harder than making the content
fun (edutainment). Edutainment in itself is not necessarily bad, but
in many cases I feel that it has blurred the really important matters
out of sight.

> 
> Oops--  Thought quota reached.

;) This is something I've often thought about, so my brain isn't yet
working at full speed, I'm just ranting :)

The questions where good, I'll try to find some time to post comments
on them.

ramin
-- 
Ramin Miraftabi                         Student of Computer Science
email: ramin@cs.joensuu.fi              University of Joensuu
WWW: http://dawn.joensuu.fi/~ramin/     Joensuu, Finland

    - GPG public key ID 49C9CFF7 -- (c) Copyright 1999 -

		 Homepages *finally* updated!