TITLE: Maybe We Just Need to Teach Better
AUTHOR: Eugene Wallingford
DATE: October 15, 2014  3:54 PM
DESC: 
-----
BODY:
Maybe We Just Need to Teach Better

A couple of weeks ago, I wrote
<A HREF="http://www.cs.uni.edu/~wallingf/blog/archives/monthly/2014-10.html#e2014-10-02T15_46_31.htm">
   Skills We Can Learn</A>
in response to a thread on the SIGCSE mailing list.  Mark Guzdial
has now written a series of posts in response to that thread,
most recently
<A HREF="http://computinged.wordpress.com/2014/10/15/we-need-to-fix-the-computer-science-teaching-problem/">
   Teaching Computer Science Better To Get Better Results</A>.
Here is one of the key paragraphs in his latest piece:

<BLOCKQUOTE><EM>
 I watch my children taking CS classes, along with English,
 Chemistry, Physics, and Biology classes.  In the CS classes,
 they code.  In the other classes, they do on-line interactive
 exercises, they write papers, they use simulations, they solve
 problems by-hand.  Back in CS, the only activity is coding with
 feedback.  <B>If we only have one technique for teaching, we
 shouldn't be surprised if it doesn't always work</B>.
</EM></BLOCKQUOTE>

Mark then offers a reasonable hypothesis: <B>We get poor results
because we use ineffective teaching methods</B>.

That's worthy of a new maxim of the sort found in
<A HREF="http://www.cs.uni.edu/~wallingf/blog/archives/monthly/2014-10.html#e2014-10-02T15_46_31.htm">
   my previous post</A>:
<EM>If things aren't going well in my course, it's probably my
fault</EM>.  Mark's hypothesis sounds more professional.

A skeptic might say that learning to program is like learning
to speak a new human language, and when we learn new human
languages we spend most of our time reading, writing, and
speaking, and getting feedback from these activities.  In an
introductory programming course, the programming exercises are
where students read, write, and get feedback.  Isn't that enough?

For some students, yes, but not for all.  This is also true in
introductory foreign language courses, which is why teachers
in those courses usually include games and other activities to
engage the students and provide different kinds of feedback.
Many of us do more than just programming exercises in computer
science courses, too.  In courses with theory and analysis, we
give homework that asks students to solve problems, compute
results, or give proofs for assertions about computation.

In my algorithms course, I open most days with a game.  Students
play the game for a while, and then we discuss strategies for
playing the game well.  I choose games whose playing strategies
illustrate some algorithm design technique we are studying.
This is a lot more fun than yet another <EM>Design an algorithm
to...</EM> exercise.  Some students seem to understand the ideas
better, or at least differently, when they experience the ideas
in a wider context.

I'm teaching our intro course right now, and over the last few
weeks I have come to appreciate the paucity of different
teaching techniques and methods used by a typical textbook.
This is my first time to teach the course in ten years, and
I'm creating a lot of my own materials from scratch.  The
quality and diversity of the materials are limited by my time
and recent experience, with the result being... a lot of
reading and writing of code.

What of the other kinds of activities that Mark mentions?  Some
code reading can be turned into problems that the students solve
by hand.  I have tried a couple of debugging exercises that
students seemed to find useful.  I'm only now beginning to see
the ways in which those exercises succeeded and failed, as the
students take on bigger tasks.

I can imagine all sorts of on-line interactive exercises and
simulations that would help in this course.  In particular, a
visual simulator for various types of loops could help students
see a program's repetitive behavior more immediately than
watching the output of a simple program.  Many of my students
would likely benefit from a Bret Victor-like interactive
document that exposes the internal working of, say, a
<TT>for</TT> loop.  Still others could use assistance with even
simpler concepts, such as sequences of statements, assignment
to variables, and choices.

In any case, I second Mark's calls to action.  We need to find
more and better methods for teaching CS topics.  We need to
find better ways to make proven methods available to CS
instructors.  Most importantly, we need to expect more of
ourselves and demand more from our profession.

When things go poorly in my classroom, it's usually my fault.
-----