Web Testing for Reliability Improvement - Advances in Computers - page 179

In subsequent sections, we review basic concepts and testing techniques, exam-
ine the characteristics of the web, discuss the use of existing models and techniques
for web testing, and introduce an integrated web-testing strategy that ensures web
reliability. In particular, Section 2 defines some basic terms related to defect and
reliability to ensure a consistent interpretation of quality, and surveys major test-
ing techniques and reliability models; Section 3 analyzes the web environment and
its impact on web testing and quality assurance, particularly its challenges and op-
portunities, and focuses our attention on web problems closely identified with web
source contents and navigations; Section 4 examines strategies for testing individual
web elements and basic web navigation; Section 5 presents an integrated strategy for
large-scale web testing that uses statistical testing models to selectively test impor-
tant and frequently used web pages and to guide in-depth testing using traditional
testing techniques; Section 6 uses defect and usage data from web logs to assess
web site operational reliability and the potential for reliability growth under effec-
tive statistical testing, which also provide an external validation for our integrated
testing strategy; and finally, we present our summary, conclusions, and future work
in Section 7 .

2. Basic Concepts and Techniques

As noticed above, various terms related to problems or defects are commonly used
in discussing software quality and reliability. Several standard definitions [19] related
to these terms include:

Failure : The inability of a system or component to perform its required func-
tions within specified performance requirements. It is an observable behavioral
deviation from the user requirement or product specification.
Fault : An incorrect step, process, or data definition in a computer program,
which can cause certain failures.
Error : A human action that produces an incorrect result, such as the injection
of a fault into the software system.

Failures, faults, and errors are collectively referred to as defects . Testing plays a
central role in assuring product quality by running the software, observing its behav-
ior, detecting certain behavior deviations as failures, and helping us locate and fix the
underlying faults that caused the observed failures.
Depending on whether external functions or internal implementation details are
tested, we have two generic types of testing: Functional or black-box testing focuses
on the external behavior of a software system, while viewing the object to be tested