%0 Journal Article
%T A Granular Hierarchical Multiview Metrics Suite for Statecharts Quality
%A Mokhtar Beldjehem
%J Advances in Software Engineering
%D 2013
%I Hindawi Publishing Corporation
%R 10.1155/2013/952178
%X This paper presents a bottom-up approach for a multiview measurement of statechart size, topological properties, and internal structural complexity for understandability prediction and assurance purposes. It tackles the problem at different conceptual depths or equivalently at several abstraction levels. The main idea is to study and evaluate a statechart at different levels of granulation corresponding to different conceptual depth levels or levels of details. The higher level corresponds to a flat process view diagram (depth = 0), the adequate upper depth limit is determined by the modelers according to the inherent complexity of the problem under study and the level of detail required for the situation at hand (it corresponds to the all states view). For purposes of measurement, we proceed using bottom-up strategy starting with all state view diagram, identifying and measuring its deepest composite states constituent parts and then gradually collapsing them to obtain the next intermediate view (we decrement depth) while aggregating measures incrementally, until reaching the flat process view diagram. To this goal we first identify, define, and derive a relevant metrics suite useful to predict the level of understandability and other quality aspects of a statechart, and then we propose a fuzzy rule-based system prototype for understandability prediction, assurance, and for validation purposes. 1. Introduction The popular view of quality is still that of a subjective concept which perpetuates the idea that the more elaborate and complex product somehow offers a higher level of quality than its humbler counterpart. Whilst this misconception is well understood amongst ˇ°qualityˇ± professionals, the temptation remains to equate sophistication, instead of simplicity of function, with quality. The term software crisis was coined in the late 1960 (NATO conference held in Garmisch-Partenkirchen, Germany, in 1968) [1] to refer to problems associated with software projects. These included budget and schedule overruns and problems with the quality and reliability of the delivered software. Software quality has been described as a complex and multifaceted concept, which basically means that it means different things to different people. Current approaches to assuring and measuring software quality are predominantly process based rather than product based. There are two major wrong assumptions here: (1) if we look after the process the product will look after itself, (2) quality must be built in the product (can only be assessed when complete and difficult to
%U http://www.hindawi.com/journals/ase/2013/952178/