A Financial Services client had been quoted over $250,000 to establish an automated testing suite for an enterprise application. This did not include licences or ongoing maintenance. The 4impact team was tasked with identifying other automation testing approaches.

Automation testing with open-source tools

The 4impact team was experienced in automation testing for software projects and brought with them a testing framework to underpin the tools and processes. The framework has two primary requirements, automation must be part of the project development deliverables and it must use data driven techniques.
Automation scripts must be maintainable by development team
This is important as most automation initiatives wither and die once they are handed over from a project to a support team. It was important that the creation and maintenance of the underlying scripts became part of the software development process. This means the automation scripts was checked-in to the source control system along with the actual source code.
Data driven framework
With the scripts being maintained by the development team a data driven framework would allow the testing analysts to construct testing scenarios based upon data, rather than screen capture methods. Data driven is accepted as the best long-term approach for automation testing
The team identified a number of open-source tools that were then short-listed according to their ability to meet the following criteria;
·        To provide an extensible, data-driven testing tool
·        To provide a non-programmatic, testing tool for Web Application testing
·        To provide a tester-centred environment with ability to interact with a Continuous Integration (CI) environment
·        To provide a simple to use tool for integration, system and
regression testing
·        Re-use open-source tools
Our commitment to deliver
In this instance, Webtest was chosen as the most suitable open-source tool due to its excellent reporting capability and internal experience within the client’s teams. The team was able to implement a page-model approach that allowed the abstraction of complexity from a dynamic user interface from the testing scripts. This allowed data-driven scripts to be written by the test analysts. In one instance, finance testing was brought down from three days of manual testing per run, to 30 minutes using automation.
The Webtest scripts were also used by the training teams to quickly setup large volumes of data to simulate production like environments and training scenarios. This saved approximately 20 days of effort in data setup throughout the project.
In all, the open-source tools saved over $250,000 in licence costs, and provided a robust and sustainable automation environment for the client.