A metrics proposal to evaluate software internal quality with scenarios

Software quality should be evaluated from different perspectives; we highlight the internal and external ones (ISO/IEC, 2002). Specially, internal quality evaluation depends on the software architecture (or design) and programming aspects rather than on the product behaviour. On the other hand, architectural evaluation methods tend to apply scenarios for assessing the architecture respect to quality requirements; however, mainly scenarios aren’t effective enough to determine the level of satisfaction of the quality attributes. In practice, each scenario could need more than one measurement. Also, we need a quantitative way of comparing and reporting results. The main objective of this article is presenting a proposal of metrics grouped by quality characteristics and sub-characteristics, according to ISO 9126 standard, which can be applied to assess software quality based on architecture. Once selected the most important quality requirements, these metrics can be used directly, or in combination with quality scenarios, into an architectural evaluation method. Metrics proposed also consider some particular technologies, such as OO, distributed and web systems.