[plt-scheme] Log "every change in" Definitions and Interactions...

From: "M. Fatih Köksal" (fkoksal at cs.bilgi.edu.tr)
Date: Mon Jan 26 08:58:27 EST 2009

Marco Morazan wrote:
> As much as I would be delighted for this to work, I am afraid that it
> is not that easy. Usually, students need to cycle through the steps of
> the DR. Errors in the data definition, for example, propagate
> throughout the DR process. Some students (many?) tend to fix such
> errors by first experimenting with the code they have written and that
> leads them to realize that the template and the data definition
> contain errors. At this point, they proceed to correct these errors.
> Other students (many?) have the right contract and template in their
> head, but incorrectly write it down. Their code, however, follows what
> is in their head. In short, there is an incongruence between what
> students are thinking and what they write down. Perhaps, this is the
> result of some type of dynamic correction while developing a program
> that they do not document? The point is that I do not see how
> recording a DrScheme session gives you enough data to reach a
> conclusion about the usefulness of the DR.

I'm thinking of tagging every submission by considering the order of the
Design Recipe applied. For Example:

DD, CC, PP, EX, TL, TS, CD -> this is what we expect (where
abbreviations are Data Definition, Contract, Purpose, Example, Template,
Tests and Code respectively)

CD, DD, PP, CC, EX, TL, TS -> such a sequence may be extracted from the
analysis data, since the code is written first and it does not imply
having the design steps will help for a fully functional and executable

TL, TS, CD, CC, PP, EX, DD -> such a sequences where some steps are done
after the code may be considered as just TL, TS, CD (skipping the steps
after code, if the code is not edited later)

CD, TL, EX, CC, EX, DD, TL, CD, TS, PP -> Such a sequence is rather more
interesting, since the student later edited some steps. Considering
until the latest CD in the sequence, we may see which steps help the
student to write the code.

These are some examples. We can come up with more cases. I think such an
analysis may lead us to a more meaningful analysis data.

> I am also
 curious about the development of what type of programs do
> you wish to study. Only those based on structural recursion? Are you
> also looking into generative recursion?

During the first semester we reach until part 5 of HTDP. Generative
recursion is not included. And it's not in scope of my thesis.

M. Fatih Köksal
İstanbul Bilgi University
Computer Science Department

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 260 bytes
Desc: OpenPGP digital signature
URL: <http://lists.racket-lang.org/users/archive/attachments/20090126/192670d3/attachment.sig>

Posted on the users mailing list.