OLYMPIADS IN INFORMATICS, 2010, Vol. 4, 76-86
© Institute of Mathematics and Informatics,

ISSN 1822-7732

Testing of Programs with Random Generated Test Cases

Krassimir MANEVa, Biserka YOVCHEVAb, Milko YANKOVb, Petar PETROVb

aDepartment of Mathematics and Informatics, Sofia University J. Bourchier blvd. 5, 1164 Sofia, Bulgaria Institute of Mathematics and Informatics, Bulgarian Academy of Sciences G. Bontchev str. 8, 1113 Sofia, Bulgaria E-mail: manev@fmi.uni-sofia.bg

bDepartment of Mathematics and Informatics, Shumen University Universitetska str. 115, 9712 Shumen, Bulgaria E-mail: bissy_y@yahoo.com, m.yankov@f5bg.net, peshoto_bg@yahoo.com

Abstract

Testing of computer programs is an essential part of the evaluation process of a programming contest. It is a mix of functional and non functional testing and a specific case of the ``black box'' testing well known from the domain of Software engineering. The paper discusses one of the possible forms of creating test cases for a program that implement an unknown algorithm - the random generation of test cases - and the problems that could arise when random generated test cases are used for evaluation of programs submitted by participants in programming contests. Rooted trees were chosen for the start of the research because of their simplicity. More deep problems and more interesting results could be expected for general graphs and other combinatorial objects.

Keywords:

program testing, functional and non functional testing, ``black box'' testing, graphs, rooted trees, random generation, height, width and branching statistics of rooted trees


PDFTo preview full article text in PDF format click here

Get Free ReaderYou could obtain free Acrobat Reader from Adobe


Copyright © Olympiads in Informatics, Vilnius University Institute of Mathematics and Informatics, 2010