Do internal software quality tools measure validated metrics?

TL;DR

This paper investigates whether internal software quality metrics provided by static analysis tools are scientifically validated, revealing that most tools offer metrics with little to no validation in academic research.

Contribution

It provides the first comprehensive analysis of the validation status of metrics offered by software quality tools, highlighting a significant gap between tool metrics and scientific validation.

Findings

Most tools offer unvalidated metrics

Only a small percentage of metrics are scientifically validated

Many metrics lack empirical support in literature

Abstract

Internal software quality determines the maintainability of the software product and influences the quality in use. There is a plethora of metrics which purport to measure the internal quality of software, and these metrics are offered by static software analysis tools. To date, a number of reports have assessed the validity of these metrics. No data are available, however, on whether metrics offered by the tools are somehow validated in scientific studies. The current study covers this gap by providing data on which tools and how many validated metrics are provided. The results show that a range of metrics that the tools provided do not seem to be validated in the literature and that only a small percentage of metrics are validated in the provided tools.