Information Technology Reference
In-Depth Information
parameters, and a set of test cases were generated to achieve pairwise coverage.
Several system errors were found.
Lott et al. [ 29 ] uses an example, which is a basic billing system. The system
processes telephone call data, and has the following four call properties:
Access (which can be Loop, ISDN, or PBX);
Billing (which can be Caller, Collect or 800);
Call type (which can be Local, Long Distance, or International);
Status (which can be Success, Busy or Blocked).
Kuhn and Okum [ 25 ] reports an experiment with a module of the Traffic Col-
lision Avoidance System (TCAS). TCAS is a well-known benchmark in software
testing and analysis. In the CT model, there are 12 parameters, which include two
10-value parameters, one 4-value parameter, two 3-value parameters and seven
2-value parameters.
Cohen et al. [ 13 ] used a simplified mobile phone product line as an example. The
product line is artificial, but its structure reflects portions of the Nokia 6000 series
phones. The model includes five components (two 2-value components and three 3-
value components): Display , Email viewer , Camera , video camera and
video ringtones . There are some constraints like, a graphical email viewer
requires color display.
Yuan et al. [ 45 ] use CT in the testing of graphical user interfaces (GUIs). Each test
case is a sequence of events, and the events may be repeated. The goal is to generate
a number of test cases, with certain coverage.
Justin Hunter [ 24 ] performed an empirical study involving 10 projects at six
companies. The experiments show that “CT can detect hard-to-find software faults
more efficiently than manual test case selection methods”.
Kuhn et al. [ 22 ] described the testing of smart phone applications. They studied
the resource configuration file for Android apps, and found dozens of options. In
the CT model, there can be nine parameters (like TOUCHSCREEN , KEYBOARD ,
SCREENLAYOUT_SIZE , etc.).
Garvin et al. [ 18 ] use a software product line for a media player as an example.
Wang et al. [ 42 ] described a black-box testing approach to detecting buffer over-
flow vulnerabilities. They noted that, in many cases, the attacker can influence the
behavior of a target system only by controlling the values of its external parame-
ters (e.g., input parameters, configuration options, and environment variables). They
adapted the CT techniques to generate a test suite, so as to steer program execution
to reach some vulnerable point.
Zhang et al. [ 46 ] conducted a case study on applying combinatorial testing to
audio players to see whether they can correctly recognize and display information
described in ID3v2 tags of audio files. Two CT models have been built for the ID3v2
tags based on two different test goals. It turns out that in some situations, a few audio
players cannot display audio information correctly, and they may even crash.
In the topic [ 23 ] (Chap. 5 , “Test Parameter Analysis”, by Eduardo Miranda), sev-
eral case studies are given:
Search WWH ::




Custom Search