| Previous | [ 1] | [ 2] | [ 3] | [ 4] | [ 5] | [ 6] | [ 7] | [ 8] | [ 9] | [ 10] | [ 11] |
¡@
Wen-Chang Pai, Chi-Ming Chung1, Ching-Tang Hsieh2, Chun-Chia Wang3 and Ying-Hong Wang4
Department of Business Mathematics
Soochow University
Taipei, 100 Taiwan
E-mail: wencpai@ms1.hinet.net
1The Open University of Kaohsiung
Kaohsiung, 812 Taiwan
E-mail: cmchung@ms2.ouk.edu.tw
2Department of Electrical Engineering
4Department of Information Engineering
Tamkang University
Taipei, 251 Taiwan
E-mail: {hsieh@ee; inhon@mail}.tku.edu.tw
3Department of Information Management
Northern Taiwan Institute of Science and Technology
Taipei, 112 Taiwan
E-mail: gcwang@mail.ntist.edu.tw
Software quality is primarily determined by the quality of the software development
process. The goals of software testing are to assess and improve the quality of software.
Software testing has proven to be difficult in the absence of design information. Without
an adequate understanding of a program's structure, it is difficult to test it properly. Program
recognition is a technology that can help testers to recover a program's design and,
consequently, make software testing effective. Syntactically, a program is a sequence of
statements. If the flow of the program can be recovered and used to analyze the testing
paths automatically, then generating test data based on adequate testing criteria will help
testers to understand the program structure and efficiently improve the software quality.
This paper provides a method for analyzing the control-flow of a program and obtaining
to the original program structure. An approach to analyzing the testing paths
automatically to test every branch of a program is provided. The proposed method defines
a number of command types and test data generating rules. An algorithm to scan
program and analysis testing paths is also provided. This will allow testers to recover a
program's design and understand the software structure to assist software maintenance.
Received September 3, 2003; revised February 26, 2004; accepted April 28, 2004.
Communicated by H. Y. Mark Liao.