Towards a Taxonomy of Software Maintainability Predictors

Abstract

Software maintainability prediction has gained more attention in the last decade. Several studies have conducted empirical studies to look for models to predict software maintainability more accurately. In a previous work, a systematic mapping study (SMS) was performed in this context and a set of metrics used as predictors of software maintainability has been identified. But, unfortunately those metrics are not organized in a structured way. Moreover, some authors may raise the same metric with the same meaning, but with different wordings. Hence, it becomes a necessity to unify all the metrics in a taxonomy that will help researchers build maintainability models in an easy way. The proposed taxonomy is 3 levels with categories, subcategories and metrics. We expect that the use of this taxonomy by researchers can help us identify other options to both propose a more useful taxonomy and to perform its evaluation.