CloneChecker: A Software Plagiarism Detector

ROPAS/Programming Research Lab
School of Computer Science and Engineering
Seoul National University

originally written by
Sungsoon Jang and Sunae Seo
ROPAS
Computer Science
Korea Advanced Institute of Science and Technology

What is CloneChecker?

This program quantifies the similarity of OCaml, nML, C, Java, or Scheme programs and groups them into "cliques." This software has been used in SNU and KAIST's several undergraduate courses for finding plagiarism in program homeworks. It is written in nML.

How does it work?

Details cannot be open to everybody (in particular, to students). We will open the details only to professors/instructors who will use this software in their classes. Please contact us (yes@ropas.snu.ac.kr). However, we open our sources.

Dear students: If you want to spend some time to figure out how to avoid this clone-checker, we advise you to spend that precious time for doing your homeworks by yourself. Because we keep improving CloneChecker every semester based on our user's feedback and by means of our static analysis techniques we're afraid you can't beat it.

We bet that students love to counter:

"Dear CloneChecker authors: If you want to spend some time to figure out how to automatically detects homework clones, we advise you to spend that precious time for doing your ROPAS research. Because we keep improving our cloning skills every semester based on our friends' feedback and by means of our humane hacks we're afraid you can't beat us."

So, we are in an interesting game. Rock'n roll, guys.

Which languages does CloneChecker support?

• nML: We use the nML parser in the nML compiler system.
• C: The parser is based on ckit, which follows the ANSI C standards.
• Java: The parser is based on the grammar(JDK 1.2) rules from CUP, a parser generator for Java programming language.
• Scheme: Implementation adheres to the Revised(5) Report on the Algorithmic Language Scheme( R5RS ).
• OCaml: The parser and checker is written in OCaml. This is made by Jonghyun Park in PLLab of POSTECH and is based on our nML checker

Caveat

• CloneChecker compares individual files. It can't compare groups of files. For example, suppose that each homework consists of two files. CloneChecker has no idea that the two files in a single homework should not be compared.
• The programs to be compared must be of the same language.

Download

• Source (.tar.gz)
• Executable files of customized version for Java (.tar.gz)
• Source for OCaml (.tar.gz)
• Troubleshooting