Two insights on teaching computer programming

Peter Van Roy pvr at
Mon Jan 2 04:54:42 EST 2006

Henderson Peter wrote:

>>> On Dec 29, 2005, at 7:38 AM, Peter Van Roy wrote:
>>> Java or whatever the CFPL is, it doesn't much matter in the first year.
>>> On Dec 29, 2005, at 1:16 PM, Peter Van Roy wrote:
>>> I will check out Holt's work.  But I presume
>>> that his basic premises are not exactly the same as those we would
>>> choose today. The CFPL he picks is certainly different, for one thing.
>> Peter, how should I reconcile these two statements of yours?
>> According to your first statement, the choice of CFPL doesn't
>> matter and according to your second one, you are allowed to
>> reject Holt et al [1] because he chose a _different_ CFPL
>> (PL/I) in the 1970s than what is the, eh, _current_ CFPL?
>> Which one is it? -- Matthias
Dear Matthias,

I think you are inferring too much from my statements.  My comment on
Holt is just one element in a discussion on his work and how it would
apply today.  I didn't say that one could reject Holt.  How could I make
such a statement without reading his work?  Thanks for the reference,
I will look it up.

BTW, I have the feeling that we don't disagree so much on the core.
Let me explain a bit more of my context and you will see better where
I am coming from.  At UCL, I did not decide how the first year course
is to be given.  I did have a say in the second year course.  The first
year course (which is not taught by me but by Baudouin Le Charlier)
is not so bad: it takes subsets of Java with simple libraries, to teach
programming, some program design, some algorithmics, and also a little
bit of semantics (such as how to write specifications of programs).  It is
quite rigorous, within the limitations of Java.  But I am not saying it
could not be done better -- that's why I am looking, e.g., at your work,
as one approach in which the first year was carefully looked at and
worked out.  My course is a second year course, and in some sense I am
doing a "course correction" with respect to the first year course.  My first
"thesis" says that this course correction seems to work well, and I 
think it is because the students are mature enough in the second year to
understand and not yet too rigid in their thinking.  This is my 
experience at
UCL, but YMMV on this property of the students.  I was surprised by the
positive reaction of the students regarding the course and its approach.
That's why I posted the message to math-thinking, to get comments from
the community.

After teaching programming in the second and third years for some time
now, I am now starting to get interested in the first year course.  I 
very early on that the first year is quite different from the second year.
Andres Becerra and I are looking carefully at the first year.  Your work on
the first year is one of the most thorough that has been done.  But some of
our basic principles are different, so maybe we will come up with a somewhat
different first year course.

A Happy New Year to you and all the people on the math-thinking list.


More information about the Math-thinking-l mailing list