Towards a Taxonomy of Software Maintainability Predictors: A Detailed View.

To help practitioners and researchers choose the most suitable predictors when selecting from existing Software Product Maintainability Prediction (SPMP) models or designing new ones, a literature review of empirical studies on SPMP identified a large number of metrics or factors used as predictors of maintainability. However, there is a redundancy and ambiguity in both the naming and meaning of these predictors. To address this terminology issue, a one-level taxonomy of the SPMP predictors identified in the literature review have been proposed. This paper now proposes a more detailed two-level taxonomy where the first level refers to four categories, namely, software design, software size, quality attributes (or factors), and software process, the second to sub-categories, and predictors inventoried from empirical studies on SPMP.