Software Metric Framework

TL;DR

The paper introduces the Software Metric Framework (SMF), a tool designed to improve the testing, verification, and reproducibility of software complexity metrics through automation and standardization, initially for Java applications using Maven.

Contribution

It presents an extensible framework that facilitates implementation, testing, and reproducibility of software metrics, addressing key concerns in the field.

Findings

Supports automated data collection and analysis

Enhances reproducibility of software metric experiments

Initial prototype limited to Java Maven projects

Abstract

Many researchers have criticized the field of Software Complexity metrics for the lack of testing, verification, and reproducibility of many metrics and case studies that utilized those metrics. This document describes SMF, a tool that can help address some of these concerns, namely by enabling verification of metrics, reproducibility of experiments, and ease of implementation for new metrics. The tool in question is the Software Metric Framework; an extensible set of scripts, tools, and standards that allow others to implement metrics in a way that allows automated data collection and analysis. Because it is only a prototype, the framework has been limited to the analysis of Java applications that utilize Maven, a build system which greatly simplifies the task of compiling source code.