A colleague of mine (Christopher Anand, [EMAIL PROTECTED]) has an
automated, web-based program for automatically grading simple programming
assignements (in C, but theoretically that does not matter).  This is in
current use by a class of first-year computer science students, and has been
in use for ~3 years now.  He has been pondering the eventual open-sourcing
of this project.

The validation and checking is automated, but the testing methodology and
the feedback portions need to be designed-in by the author of a particular
problem.  Infrastructure is provided to help with all these steps.

On a somewhat related note, for mathematics (and physics and ...) problems,
there does exist such software.  See MapleTA from Maplesoft inc
(www.maplesoft.com).  The 'equivalence' problem for answers to differential
equations questions is on the same order of complexity as that of testing if
two ~20 line programs denote the same function (ie it is undecidable in
general, has many decidable sub-cases which are at least NP-complete, and
yet in practice can be effectively done on problems of interest).

Jacques

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of
Chris Douce
Sent: October 18, 2004 3:54 AM
To: [EMAIL PROTECTED]
Subject: PPIG discuss: Automatic marking of programs


Hi,

This is a two part question that is loosely related to the previous thread:

Lets say I have a class of students performing a group of simple progamming
assignments in Java (for sake of argument).  By assignments I'm thinking of
exercises that performs simple temperature conversion, calculate average
rainfall, calculates a factorial, or a set of classes to exhibit a
particular behaviour through published methods.

Does anybody know of any papers that try to address the problem of automated
(by computer) validation/checking of student programs, potentially providing
feedback regarding student assignments?  I appreciate that a lot of checking
and validation is performed by a compiler (and the interesting error
messages that are created), and that code can be automatically tested
through a mechanism like JUnit, providing they are integrated into the
software that is to be tested (and written properly)...  Do you know whether
anybody has investigated or implemented a more 'student friendly' mechanism
for automated checking/validation of submitted programs which could be
related to particular assignments or programming tasks?

Also, if we have a sample program, does anyone know of a tutoring system
that may assist students through the automatic generation of natural
language questions derived from a submitted program text?  I appreciate that
most of the time a compiler facilitates the generation of these questions,
through user interpretation...

All references and comments greatfully appreciated,

Many thanks.

Chris



 
----------------------------------------------------------------------
PPIG Discuss List ([EMAIL PROTECTED])
Discuss admin: http://limitlessmail.net/mailman/listinfo/discuss
Announce admin: http://limitlessmail.net/mailman/listinfo/announce
PPIG Discuss archive: http://www.mail-archive.com/discuss%40ppig.org/


----------------------------------------------------------------------
PPIG Discuss List ([EMAIL PROTECTED])
Discuss admin: http://limitlessmail.net/mailman/listinfo/discuss
Announce admin: http://limitlessmail.net/mailman/listinfo/announce
PPIG Discuss archive: http://www.mail-archive.com/discuss%40ppig.org/

Reply via email to