Comparison of GUI test automation strategies in a Clearing System : A case study at Nasdaq Stockholm AB

Abstract: The effectiveness of GUI-based automated tests is a topic that is often discussed in the testing community. As software systems and GUIs are getting more advanced, the testing of these systems may sometimes be both time-consuming and repetitive. Therefore, automated tests are used more frequently to create more agile working and to decrease the number of manual tests. There is an ongoing discussion about the benefits and shortcomings of implementing automated GUI test compared to performing manual GUI tests. Increasingly, companies today are implementing automated GUI tests, yet relatively few are analyzing enough when choosing their test strategy. The purpose of this thesis was to explore different test strategies in a system as complex as Nasdaq’s clearing system NRTC. The first strategy used manual GUI tests, while the second strategy used a test tool called Selenium WebDriver to automate the GUI tests. Hopefully, this research will contribute to the already accumulated knowledge about automated GUI tests and potentially help in choosing the right testing strategy. The findings of this research showed that the implementation time of manual tests was much faster, and that implementing automated test would be timeconsuming. The reason for this is that the GUI was not very testable, making it difficult to create the test scripts. In addition, the was a difference in runtime when using both strategies. The automated tests only took a couple of seconds to execute, compared to the manual tests that took a couple of minutes. After both strategies was investigated, an analysis was done by looking at the effectivity, maintainability, robustness, and the overall results to be able to find benefits and shortcomings. In conclusion, GUI automation cannot replace manual test completely. By not including the human capabilities, the testing process becomes very limited.