TITLE: SIGCSE Day 1: Media Computation BoF
AUTHOR: Eugene Wallingford
DATE: March 08, 2007 8:03 PM
DESC:
-----
BODY:
A BoF is a "birds of a feather" session. At many conferences,
BoFs are a way for people with similar interests to get together
during down time in the conference to share ideas, promote an
idea, or gather information as part of a bigger project.
Tonight I attended a BoF on the media computation approach
I used in CS I
last semester. The developers of this approach,
Mark Guzdial
and
Barbara Ericson,
organized the session, called "A Media Computation Art Gallery
and Discussion", to showcase work done by students in the many
college and high school courses that use their approach. You
can access the movies, sounds, and video shown at this
on-line gallery.
The picture to the right was an entry I submitted, an image
generated by a program written by my CS I student Keller
McBride. This picture demonstrates the sort of creativity
that our students have, just waiting to get out when given
a platform. I don't know how novel the assignment itself was,
but here's the idea. Throughout the course, students do almost
all of their work using a set of classes designed for them,
which hide many of the details of Java image and sound. In
one lab exercise, students played with standard Java graphics
programming using java.awt.Graphics objects. That
API gives programmers some power, but it has always seemed
more complicated than is truly necessary. My 8-year-old
daughter ought to be able to draw pictures, too! So, while
we were studying files and text processing, I decided to try
an assignment that blended files, text, and images. I asked
my students to write an interpreter for a simple straight-line
language with commands like this:
line 10 20 300 400
circle 100 200 50
The former draws a line from (10, 20) to (300, 400), and the
latter a circle whose center point is (100, 200) and whose
radius is 50.
This is the sort of assignment that lies right in my sweet
spot for encouraging students to think about programming
languages and the programs that process them. Even a CS I
student can begin to appreciate this central theme of computing!
Students were obligated to implement the line and
circle commands, and then to create and implement a
command of their own choosing. I expected squares and ovals,
which I received, and text, which I did not. Keller
implemented something I never expected: a colorSpray
command that takes a density argument and then produces four
sprays, one from each corner. I describe the effect as shaking
a paint brush full of color and watching the color disperse in
an arc from the brush, becoming less dense as the paint moves
away from the shaker.
This was a CS1 course, so I wasn't expecting anything too fancy.
Keller even discounted the complexity of his code in a program
comment:
* My Color Spray method can only be modified by how many arcs
it creates, not really fancy, but I did write it from scratch,
and I think it's cool.
I do, too. The code uses nested loops, one determinate and
one indeterminate, and does some neat arithmetic to generate
the spray effect. This is real programming, in the sense that
it requires discovering equations to build a model of something
the programmer understands at a particular level. It required
experimentation and thought. If all my students could get so
excited by an assignment or two each semester, our CS program
would be a much better place.
At the BoF, one attendee asked how he should respond to colleagues
at his university who ask "Why teach this? Why not just teach
Photoshop?" Mark's answer was right on mark for one audience.
Great artists understand their tools and their media at a deep
level. This approach helps future Photoshop users really
understand the tool they will be using. And, as another attendee
pointed out, many artists bump up against the edges of Photoshop
and want to learn a little programming precisely so they can
create filters and effects that Adobe didn't ship in their software.
The answer to this question for a CS or technical audience ought
to be obvious -- students can encounter so many of the great ideas
of computing in this approach; it motivates many students so much
more than "convert Celsius to Fahrenheit"; and it is relevant to
students' everyday experiences with media, computation, and data.
The CS education community owes Mark and Barb a hand for their
work developing this wonderful idea through to complete, flexible,
and usable instructional materials -- in two different
languages, no less! I'm looking forward to their next step, a
data structures course that builds on top of the CS 1 approach.
We may have a chance to try it out in Fall 2008, if our faculty
approve a third offering of media computation CS I next spring.
-----