
This corpus consists of programs written by students from the Faculty of 
Mathematics, University of Belgrade. Programs were written on exams at an 
introductory programming course in the Fall Semester of 2011.

The corpus consists of 266 solutions to 15 different problems. These problems 
include numerical calculations, manipulations with arrays and matrices, 
manipulations with strings, and manipulations with data structures. Short 
descriptions of problems are given in file 

problem_description.pdf

In addition to each student's program p.c, there are p.o, 
p.cfg, and p.nod files (which are constructed from p.c file) :

--- p.o is produced by llvm-gcc tool (see http://llvm.org) and can be used as 
    an input file for the LAV tool (see http://argo.matf.bg.ac.rs/?content=lav)
--- p.cfg and p.nod are produced by CfgLAV and are input files for the
    similarity tool (see http://argo.matf.bg.ac.rs/?content=lav)

The corpus consists only of programs that successfully passed all manually 
generated test cases (other programs written by students, i.e. programs that 
did not pass the testing phase, were not included into this corpus). 

The final grades were calculated by manual inspection and are given in file

marks.xls

Finally, different solutions of these problems are given in folder 

solutions

This corpus was used for experiments described in the paper 

Software Verification and Graph Similarity for Automated Evaluation of 
Students' Assignments  

In these experiments, the analysis performed by the verification tool LAV 
was done with parameters given in the lav_parameters files.

Permission is granted to use and redistribute this corpus for research purposes.

--------------------------------------------------------------------
For any further questions send an email to 
milena@matf.bg.ac.rs
--------------------------------------------------------------------

