The multi-objective additive value rankings are displayed as shown in Figure 6. We could see that F1 and F2’s priorities are still far ahead of others. Because this prioritization is conducted before system testing, it is the same that we should test F1 and F2 in the test round 1 as we use multiplicative value function form. That means that in both ways, we could get the similar prioritizations that maximize the ROI of the testing investment. It should also be noted that after using the exponential value function, decision makers’ preferences, the emphasis on high business importance (e.g. F1, F2) and avoidance of
high cost (e.g. F4) are obviously displayed in the stacked bar ranking graph.
Because this prioritization is conducted before system testing, the quality risk probability for all features is low as shown in Table 2, all the 9 features shows narrow value bars on Quality Risk Probability. For Testing Cost, we could see that except F4, all others’ value bars have no obvious differences, this shows that testing manager’s tolerance of testing cost, however, if the testing cost is very high, e.g. F4 with score 9, the value decreases dramatically. We think that exponential single value functions can more realistically reflect the decision maker’s value.
Replaced/Superseded by document(s)
Testing is one way of validating that the customers have specified the right product and verifying that the developers have built the product right. Traditional testing methodologies such as: path, branch, instruction, mutation, scenario, or requirement testing usually treat all aspects of software as equally important , and is often reduced to a purely technical issue leaving the close relationship between testing and business decisions unlinked and the potential value contribution of testing unexploited . Value based software testing serves to bridge the gap between the internal software testing process and business value that comes from customers and market . The essential of Value-based software testing is feature prioritization that brings business importance that comes from CRACK key customers, software quality risk from developing team, software testing cost, process estimation and controlling from the testing team, and the
external commercial factors, e.g., market pressure from the market into consideration. This decision making process could be seen as a multiple-objective decision making process that tries to maximize the business importance, identify the most risky features, minimize the testing cost and minimize the market erosion. The prioritization strategy could be reflected and influenced by different multiple-objective decision value functions.
In , a multiplicative multiple-objective value function is used to generate the testing priority ranking. In this paper, we will use additive multiple-objective value function to generate testing priority ranking and compare the result in the end.