Data-centric Testing for Pervasive Middleware-based Programs

Description

Pervasive computing means to compute anywhere and at anytime. Pervasive computing software applications are increasingly popular, and assuring quality by software testing is vital and practical. Many of these applications are middleware based and context aware. They often acquire environmental information (e.g., room temperature and light intensity), known as contexts, via the middleware and react to (changes of) situational contexts so that they can behave adaptively. Application logics thus span the application tier and the middleware tier. Existing software test data selection approaches however ignore the middleware tier in the testing of such applications.This project focuses on the code-based test data selection issues of context-aware middleware-based applications. The main tasks include the following:eliciting the impact of contexts on the data flow of such applications—the researchers expect to identify context-aware types of dataflow associations and hence significantly extend the standard dataflow approaches for the testing of these applications;formulating context-aware dataflow test data adequacy criteria—the researchers plan to extend the subsumption hierarchy of standard dataflow criteria to include their criteria, which are expected to be more effective for detecting failures for these applications; andconducting comprehensive empirical evaluations and case studies.