A domain–specific language for automated testing of iOS applications

Abstract

Society is increasingly relying on mobile device applications for a wide variety of purposes. This increase in demand, coupled with the high quality expectation amongst users, necessitates the efficient development of reliable applications. Currently, Automated Testing has become a dominant practice in order to ensure that software is functioning correctly. However, the process of automating mobile application testing frequently involves substantial difficulties, such as lengthy application setup configurations, as well as severely complex mechanisms that automate interactions (e.g. Tapping, Pinching, etc.). Adversely, test engineers are forced to divert focus from test creation, in order to fine-tune the extraneous sophisticated detail that is associated with these difficulties. Consequently, since the aforementioned challenges have the e↵ect of further complicating test suites, mobile application tests become increasingly difficult to understand and maintain. Often, in such circumstances, software engineers propose the development of a Domain–Specific Language (DSL), which enables users to express domain notions in a concise and unambiguous manner. The purpose of this Final Year Project (FYP) is to investigate the possibility of designing a DSL, which is intended to simplify the task of automating iOS application testing, whilst providing a layer of abstraction serving to suppress redundant detail. Consequently, the language design must capture core concepts within the relevant domains, namely the iOS platform and Automated Testing, to enable language users to concisely express application tests, which are simple to understand, as well as easy to maintain. In order to ensure that the developed language safeguards the properties of a well-defined DSL, the language is evaluated using various distinct strategies, including a case study, a critical analysis, as well as a feedback session with a number of candidates. Whilst substantial initial e↵ort is required to design and implement the language, the observations made during the language evaluation indicate that the DSL yields substantial gains in e↵ectiveness and considerably simplifies the test creation process. The concrete deliverables comprise of a formal description of the DSL, as well as a proof of concept tool, serving as a compiler for tests developed using the DSL.