M.E.T.A. - Modeling for More Effective Test Architecture

Testing and test automation are important parts of modern software development. Many techniques have been developed, including model-based testing (MBT). In MBT, the developer creates a model representing an abstraction of the system under test (SUT) behavior that can be executed to verify the SUT's conformance to that model. The technique has been widely studied in academia but is currently still underutilized in practice. Moreover, the technique is mainly used in embedded and hardware-driven domains, such as automotive systems, although research has presented many advantages of the technique.

Purpose of the project

The M.E.T.A. project is a collaboration between Blekinge Institute of Technology, Spotify, AddQ and ArcticBlue that aims to investigate MBT as a technology for practical use in industry. The project is divided into two phases where the first aims to evaluate the technology's use for GUI-based testing and the second part evaluate how MBT is used in practice and what best practice guidelines are available in industry.

Implementation

An MBT tool is developed for GUI-based testing that combines the established open source Graphwalker tool with the Eye2 image recognition library to create an MBT tool for visual testing.

A set of industrial best practices is being developed, elicited from industry how companies using MBT use it to its highest efficiency.