CS SEMINAR: Rethinking Testing of Complex Software Systems: Beyond Traditional Approaches

Abstract: Software systems now sit at the heart of modern life, shaping the infrastructures that support communication, finance, healthcare, transport, and security. Yet as these systems grow in scale, connectivity, and dynamism, testing them becomes an increasingly difficult scientific and engineering challenge. Many of the assumptions that informed traditional testing approaches were developed for systems that were more stable, more bounded, and more architecturally transparent than the software ecosystems we face today.

 

In this talk, I argue that the testing of complex software systems must be rethought beyond traditional approaches. Drawing on ideas from model-based testing and formal methods, I will discuss how behavioural abstraction can still provide a disciplined way of reasoning about software whose internal structure may be hidden, evolving, or only partially accessible. At the same time, I will highlight why many established testing strategies struggle when confronted with the scale, heterogeneity, and continual evolution of contemporary software-intensive systems.

 

Looking ahead, I will outline a broader research vision for the future of software testing  that seeks not merely incremental improvement, but a deeper rethinking of how assurance can be achieved in the presence of complexity. In this vision, the rigour of formal reasoning is complemented by the emerging potential of artificial intelligence and quantum computing, opening new possibilities for testing systems whose complexity is no longer exceptional, but fundamental.