Exploring the Composition of Unit Test Suites

TL;DR

This paper proposes visualization techniques to analyze the structure of unit test suites, helping developers understand their composition, dependencies, and relation to production code through case studies.

Contribution

It introduces visualizations for exploring test suite structures, aiding comprehension beyond traditional code reading, based on key design principles and case studies.

Findings

Visual patterns reveal test suite characteristics

Visualizations assist in locating and understanding test cases

Case studies demonstrate practical utility

Abstract

In agile software development, test code can considerably contribute to the overall source code size. Being a valuable asset both in terms of verification and documentation, the composition of a test suite needs to be well understood in order to identify opportunities as well as weaknesses for further evolution. In this paper, we argue that the visualization of structural characteristics is a viable means to support the exploration of test suites. Thanks to general agreement on a limited set of key test design principles, such visualizations are relatively easy to interpret. In particular, we present visualizations that support testers in (i) locating test cases; (ii) examining the relation between test code and production code; and (iii) studying the composition of and dependencies within test cases. By means of two case studies, we demonstrate how visual patterns help to identify key test suite characteristics. This approach forms the first step in assisting a developer to build up understanding about test suites beyond code reading.