Arts et Métiers ParisTech / Laboratoire LCFC, Metz, France

---

Abstract—A large number of developed, acquired or purchased software tools do not respond to the users’ requirements and expectations which had often been at the origin of the project. This is mainly due to two reasons: firstly because the users’ requirements are not well identified or formalized as they should, secondly because the software and tool evaluation is not robust enough or does not have a minimum required quality [1]. We attempt to propose a new approach in order to assess and quantify the quality of the software evaluation process. The theoretical approach is based on elaborating a matrix (A_n,_m) of software functionalities versus user’s scenarios. The norm of the columns and lines vectors of this matrix may be considered as a quality indicator of the process. Performance metrics are also derived from such indicators to summarize the quality of software evaluation process from the users’ point of view. We have applied such an approach on one case study and derived indicators like task effectiveness and efficiency of the software product evaluation. This method helps in two stages of the development cycle of the software: during the design phase and during the verification and validation. The case study helps to identify software imperfections and insufficiencies in a practical manner early in the development lifecycle. Thus improving subsequent releases of the software product.

Index Terms—Operational Software Evaluation, Functional Matrix Decomposition, Verification and validation Process, Process Quality, Indicator and Metrics

[PDF]

Cite: Armin Azarian, Ali Siadat, "A Synthesis of Software Evaluation Methodologies and the Proposal of a New Practical Approach," Journal of Software vol. 6, no. 11, pp. 2271-2281, 2011.