Effectiveness of software testing techniques in enterprise: a case study

Defect clustering
: "a small number of modules contains most of the defects discovered during pre-
release testing or shows the most operational failures". This means that the distribution of defects are 
not across the application but rather centralized in limited sections of the application, such as the small 
amount of particular modules or units that are used rarely or not fully testing because of different 
causes, as for instance, lack of people skills on product architecture and paradigms, incident handling, 
test levels, unclear test objectives and test process, testing obstacles, limited resources and time 
devoted for testing etc. Defect clustering in software testing is based on the Pareto principle, also 
known as the 80-20 rule, where it is stated that approximately 80% of defects are contained in 20% of 
software components (Boehm & Basili, 2001; Ostrand, Weyuker, & Bell, 2005). Studies by Boehm & 
Basili (2001), Gittens, Kim, & Godwin (2005), Li et al. (2009) and Ostrand et al. (2005) have proved 
that this principle fits defect distributions by components. The main benefit of defect clustering is that 
testing can be prioritized on focusing the same component rather than performing full testing of all 
components. Hence, the more number of defects will be found in shorter period. As a result, it 
increases testing efficiency and reduces the cost of software development. 
5) 
Pesticide paradox
: this paradox appears when the same kinds of tests are performed a certain 
number of iterations without any additional steps. In fact, this leads to more not detected defects 
because the same set test cases will no longer be able to find any new bugs. To overcome this 
“Pesticide Paradox”, it is essential to updated test cases regularly by adding some new and different 
tests to cover different units, components (modules) of the software or system in order to detect 
potential bugs. 
6)