Abstract:

Over the years, programming education at TU/e has evolved quite a bit.
In 2000, we had over 150 regular first-year students in Computer Science.
Five faculty members each took care of one group.  The students wrote
programs, and submitted them on paper for grading.  Often, these
programs could not compile (syntax errors), or would not have worked
due to other errors.  The administration of who had done what and with
what result was rather chaotic and never up-to-date.  As a teacher it
was next to impossible to browse the submitted work.

In the summer of 2001, Erik Scheffers and the speaker designed
and implemented a first version of PEACH/vs, which stood for

        Programming Education And Contest Hosting Verification System.

Peach is a web-based system where students submit their programs.
These programs are stored and automatically tested, and --where relevant--
can be graded manually by assistants.  Peach offers the teacher a uniform
overview of who has done what, when, and with what result.

An extra feature of Peach is that it can compare all submitted programs
to detect similarities, even across years for the same assignments.

Currently, Peach 3 is being used for various courses, not only in
programming.  It supports group assignments, deadlines (which are
invidividually extendible), flexible grade scales, etc.

I will provide some background information, give a brief demo, and in
particular discuss what it takes to prepare a programming assignment with
automatic evaluation.  The latter offers some interesting challenges.
In conclusion, I will summarize our experiences with the use of Peach
and look at future plans.

(End of Abstract)