TITLE: The Opposite of "Don't Do That"
AUTHOR: Eugene Wallingford
DATE: October 02, 2008  7:12 PM
DESC: 
-----
BODY:
I had one of those agile moments on Wednesday.  A colleague
stopped by my office to share his good feeling.  He had
just come from a CS 2 lab.  "I love it whenever I design
a lab in which students work in pairs.  There is such life
in the lab!"  He went on to explain the interactions within
pairs but also across pairs; one group would hear what
another was thinking or doing, and would ask about it.  So
much learning was in the air.

This reminded me of the old joke...  Patient: "Doctor, it
hurts when I do this."  (Demonstrates.)  "Can you help me?"
Doctor: "Sure.  Don't do that."

Of course, it reminded of the
<A HREF="http://www.cs.uni.edu/~wallingf/blog/archives/monthly/2005-04.html#e2005-04-07T07_49_17.htm">
   negative space</A>
around the joke.  Patient: "Doctor, life is great when I
do this."  (Demonstrates.)  "Can you help me?"  Doctor:
"Sure.  Do more of that."

"But..."  But.  We faculty are creatures of habit, both
in knowing and in doing.  We just know we can't teach all
of our material with students working in pairs, so we
don't.  I think we can, even when I don't follow my own
advice.  (Doctor, heal thyself!)  We design the labs, so
if we want students to work in pairs, we can have them
work in pairs.

I've had one or two successful experiences with all pair
programming all the time in closed labs.  Back when we
taught CS1 and CS2 in C++, in the mid-1990s, and I was
doing our first-year courses a lot, I designed all of my
labs for students working in pairs.  I wish I could say
I had bee visionary, but my motivation was extrinsic:
I had 25-30 students in class and 15 computers in the
lab.  Students worked with different students every week,
in pseudo-random assignments of my device.

My C++-based courses probably weren't very good -- I
was relatively new to teaching, and we were using C++
after all -- and the paired programming in our lab
sessions may have been one of the saving graces:
students shared their perplexity and helped each other
learn.  When they worked on outside programming assignments
for the course, they could call on a sturdy network
of friends they had built in lab sessions.  Without
the pairs, I fear that our course would have worked
well for very few students.

If something works well, let's try to understand the
context in which it works, and then do it more often
in those contexts.  That's an agile sentiment, whether
we apply it to pair programming or not.  Whether we
apply it at the university or in industry.  Whether
we apply it to software development or any other
practice in which we find ourselves engaged.
-----