An Analysis of the Software Components Quality in Use using Bayesian Networks

Proposals to improve the quality of software products have traditionally focused on improving the Internal or External Quality (regardless of the possible contexts of use) based on the idea that a good External Quality (within the meaning of ISO/IEC 9126) guarantees a good Quality in Use. This paper presents another way, moving the focus toward the Quality in Use as a significant/decisive element to design high quality software products for specific users or to select the product that better fulfills the client requirements. Our proposal is to analyze the relationships between External Quality and Quality in Use to determine the external quality subcharacteristics which are really significant to ensure the required quality level of a product in a specified context of use. The aim is to avoid unnecessary costs or irrelevant characteristics for the end user which unnecessarily raise the cost and effort of product development. We propose Bayesian Networks to model these relationships and provide a method to define them in a measurable way. As an example of using the proposed method we focused on the domain of software components to analyze their Quality in Use.