Semantic Matching for Migrating Tests Across Similar Interactive Applications

Decanato - Facoltà di scienze informatiche

Data: 31 Maggio 2023 / 09:30 - 12:30

USI East Campus, Room D0.02

You are cordially invited to attend the PhD Dissertation Defence of Ali Mohebbi on Wednesday 31 May 2023 at 10:35 in room D0.03 (East Campus) and online.

Interactive mobile apps are extremely popular, and there is a strong demand for short time to market, high quality, and low costs. It is important to automate testing analysis to efficiently and effectively verify these applications, and comply with market requirements. In the last decades, many researchers have proposed several approaches to automatically generate test cases for interactive applications. Most approaches for testing interactive applications consider structural information, and ignore semantic information. As the result, they are not effective in revealing non-crashing semantic errors and wrong outputs. TEST REUSE is a recent promising research line that aim to generate semantically relevant test cases by migrating test cases across similar applications. TEST REUSE approaches leverages semantic information available in the GUI of applications to migrate test cases from a source application to a target application that shares similar functionalities. TEST REUSE approaches rely on semantic matching to find similar events between source and target applications based on the textual descriptor of GUI widgets. TEST REUSE approaches use semantic matching techniques based on simple considerations, and do not carefully investigate the different design choices and their effects. Understanding the design choices, that is, identifying commonalities and differences in semantic matching techniques in the context of test generation approaches, can help us develop effective migration strategies. In this thesis, we study the semantic matching of events in approaches that reuse test cases across interactive applications: We study the components of semantic matching of events available in the GUI, and identify the best choices for TEST REUSE among available solutions and those that we propose in this thesis. In the thesis, we (i) define a general context for TEST REUSE approaches that encompasses common components of all TEST REUSE approaches, (ii) introduce a framework to automatically evaluate semantic matching both in-isolation and in the context of test reuse, (iii) conduct a comprehensive empirical study, and (iv) provide important insights about the components of semantic matching and its effect on test reuse.

Dissertation Committee:

  • Prof. Mauro Pezzè, Università della Svizzera italiana, Switzerland (Research Advisor)
  • Prof. Valerio Terragni, University of Auckland, New Zealand (Research co-Advisor)
  • Prof. Antonio Carzaniga, Università della Svizzera italiana, Switzerland (Internal Member)
  • Prof. Fabio Crestani, Università della Svizzera italiana, Switzerland (Internal Member)
  • Prof. Mattia Fazzini, University of Minnesota (External Member)
  • Prof. Rui Maranhao Abreu, University of Porto, Portugal (External Member)