There is a number of quality models for the evaluation of software products. The final objective for these models is to obtain a measure for the global quality of the software product and for each individual quality characteristic in order to select the more accurate solutions to improve the lasts and, through them, the whole software product. However, software quality attributes have positive and negative contribution relationships among them, which in fact restrict the selection of the best solutions. As one of the solutions to improve some of them could negatively influence another one, the global quality could have a small improvement, none improvement at all or what is worse, it could degrade. For this reason, we should measure the impact of the selected set of solutions in order to choose the set presenting the best results for improvement. With these metrics, it is possible to calculate the set of solutions that would improve a software product with the best quality gain according to the "system identity", composed by a set of quality attributes and its respective values, which therefore is one of the objectives of antifragility.
Pinciroli, F. (2016). Improving Software Applications Quality by Considering the Contribution Relationship among Quality Attributes. In Procedia Computer Science (Vol. 83, pp. 970–975). Elsevier. https://doi.org/10.1016/j.procs.2016.04.194