by Vijay Kumar Gambhiraopet
Businesses view software testing as a physical activity of executing tests. Test execution, being the interfacing activity between software development and business, is most visible and often perceived as the primary reason for delivery delays. Hence, any solution to expedite this activity is readily accepted.
Run, run, run!
Automating test execution is seen as a silver bullet to address this issue. The tests can be generated by script-less tools, which do not require scripting and those with AI embedded promise to generate numerous tests in no time. Once the tests are generated, their execution automated, the tests can be run on-demand with no dependency on humans and at a fraction of time taken by them, eventually producing a defect-free product.
Going the open source & script less
Leaders are demanding solutions based on open source due to the ease of use, flexibility to integrate with enterprise solutions and availability of skills. Open source tools are often coupled with robotic solutions to offer a comprehensive solution for enterprise systems based on technologies accumulated over the years.
The race to automate in order to align to Agile practices has triggered a demand for plethora of script-less test automation tools. Percentage of tests automated is a high focus metric in governance reports. However, the automated scripts generated by these script-less tools are not portable – to protect the business of solution providers!
The catch-up game
Even as Agile is being embraced by customers, the automated execution of scripts is, often, limited to the regression test suite. The functionality developed in sprint is tested manually and upon successful testing a representative set is identified to be included in the regression suite, which are the candidates for automation. Hence, the automation is usually catching-up, spanning sprints.
Complete automation of a regression suite is targeted for the release schedule. While the velocity of a team is expected to improve over time, the effort to identify test scenarios, sourcing test data, test environment setup, defect cycle are ignored. The effort for these tasks accumulates over sprints leading to team burn-out and the eventual casualty is automation.
Testability is key to being Agile
An ideal automation strategy, thus, is to automate progressively, for a program to be truly agile. The strategy should include how to identify test cases for automation, adding testability a mandatory requisite to develop a functionality, procedures to ensure a dedicated and on-demand test environment and test data with commitment secured from the stakeholders.
“The last best experience that anyone has anywhere, becomes the minimum expectation for the experience they want everywhere”
Bridget van Kranlingen – IBM Leader
“NEW” expectations from QA
To meet this insatiable demand for quality, the responsibility on tester community is ever increasing. The testers must look beyond their confines of the team and ensure participation in product meetings, agile ceremonies and present a user perspective on the requirements, convey user priorities to make quality intrinsic to the requirements. Apart from looking deeper into functionality, a tester should start looking higher into the business objectives. Testers should make quality as everyone’s responsibility.
Automating test execution in isolation ends up being more of a problem than a solution. Any automation solution either to enhance quality and to improve test cycles should encompass the tasks across test discipline. Automation should be considered a lever to meet the business objectives and not an objective itself.
About the author
Vijay works at IBM as Test Automation leader for North America. He has been engaged with multiple clients across geographies and domains. His professional profile is at https://www.linkedin.com/in/gambhiraopet/