A utilização de testes automatizados no processo de garantia de qualidade de interfaces de aplicações web: um mapeamento sistemático

Abstract

Software testing plays a crucial role in ensuring quality during the development of soft- ware, as it aims to identify defects through program execution. However, manual testing has proven to be unfeasible for many projects due to its high financial cost and time consumption. To address this issue, various automated testing techniques have been de- veloped, which have shown significant results in reducing costs. Web applications have become an essential tool in society, present in the daily lives of millions of people, and the dependence on these systems has made the existence of a software quality assurance pro- cess indispensable. Web systems use graphical interfaces as a means of human-computer interaction, where all input and output data are managed by graphical elements, facil- itating and optimizing this interaction. Ensuring the quality of graphical interfaces is considered extremely important, as their failures can impact the entire functionality of the system. Numerous tools that automate the graphical interface testing process for web systems are available. Thus, this research conducted a systematic mapping of the use of automated tests in the quality assurance process for web interfaces. Using the search string, a total of 505 papers were found. After applying selection filters according to in- clusion and exclusion criteria, 37 relevant studies were considered and analyzed, revealing important characteristics about the use of technologies in the automation of web interface testing over the past 5 years. It was identified that approximately 64.8% of the analyzed studies proposed a new tool and/or technique for automated web interface testing. Among existing frameworks, Selenium received 62.5% of the citations within comparative studies of tools, making it the most used tool. Multiple limitations were reported, primarily the high budgetary cost of implementing existing frameworks and tools in the market, and in studies proposing new tools, difficulties in generalizing tests for different types of web platforms were reported. Therefore, this study allowed for the characterization of the use of automated testing in web interfaces, identifying tools and techniques proposed and de- veloped in the past 5 years, highlighting the multiple challenges that GUI test automation still faces, thus driving the need for further research to address efficient solutions for this activity.