A Unified Metamodel for Assessing and Predicting Software Evolvability Quality

• Main Author: Hmood, Aseel
• Format: Others
• Published: 2013
• Online Access: http://spectrum.library.concordia.ca/978104/1/Hmood_PhD_S2014.pdf; Hmood, Aseel <http://spectrum.library.concordia.ca/view/creators/Hmood=3AAseel=3A=3A.html> (2013) A Unified Metamodel for Assessing and Predicting Software Evolvability Quality. PhD thesis, Concordia University.

Software quality is a key assessment factor for organizations to determine the ability of software ecosystems to meet the constantly changing requirements. Many quality models exist that capture and assess the changing factors affecting the quality of a software product. Common to these models is that they, contrary to the software ecosystems they are assessing, are not evolvable or reusable. The thesis first defines what constitutes a unified, evolvable, and reusable quality metamodel. We then introduce SE-EQUAM, a novel, ontological, quality assessment metamodel that was designed from the ground up to support quality unification, reuse, and evolvability. We then validate the reus-ability of our metamodel through instantiating a domain specific quality assessment model called OntEQAM that assesses evolvability as a non-functional software quality based on product and com-munity dimensions. A fuzzy logic based assessment process that addresses uncertainties around score boundaries supports the evolvability quality assessment. The presented assessment process also uses the unified representation of the input knowledge artifacts, the metamodel, and the model to provide a fuzzy assessment score. Finally, we further interpret and predict the evolvability as-sessment scores using a novel, cross-disciplinary approach that re-applies financial technical analy-sis, which are indicators, and patterns typically used for price analysis and the forecasting of stocks in financial markets. We performed several case studies to illustrate and evaluate the applicability of our proposed evolvability score prediction approach.