Defect Amplification and Removal
The defect amplification model [IBM81] can be used to describe the detection and generation of errors during preliminary design, detail design and coding steps of the software engineering procedure. The model is describing schematically in the given Figure 9.2 a box is represents a software development steps. In the duration of these steps the step, errors may be inadvertently occurring. Review may fail to uncover newly generated errors and errors from previous steps resulting in some number of errors which are passed by. Some cases errors passed by from earlier steps which are amplified (amplification factor, x) through current work. The box subdivisions represent each of these characteristics and the percent efficiency for detecting bugs a function of the thoroughness of review.
In the Figure describes a hypothetical example of defect amplification for a software development procedure in that no reviews are conducted. Describe in the figure each test step is supposed to uncover and correct 50 % of all incoming errors without introducing any new errors an optimistic assumption 10 preliminary design errors are amplified to 94 errors previous testing commences. 12 latent defects are released to the area. The Figure 9.4 considers the similar conditions except which code and design reviews are conducted as categorised of each development step. In this case ten initial preliminary design errors are amplified to 24 errors previously testing commences. Only 3 latent defects exist. Through recalling the relative costs related with the correction and discovery of errors overall cost with and without review for your hypothetical example can be build. In the Table it can be seen which total cost for maintenance and development when reviews are organized is 783 cost units when no reviews are conducted total cost is 2177 units-nearly 3 times more costly.
To conduct reviews a developer must expend effort and time and the development company must spend money. Moreover, the results of the preceding example leave little doubt which we have encountered a pay now or pay much more lately syndrome. The formal technical reviews for design and other technical activities give a demonstrable cost advantages. They should be organized.
Fig - Defect amplification model
Fig -Defect amplification no reviews