Smart systems for internet of things

Summary form only given. Since its early beginnings in the 1980s, much has been achieved in the research field of software architecture. Among other aspects, this research has produced foundational work on the specification, analysis and component configuration of software architectures, including the development of associated software tools. However, adoption of the research by industry has been largely methodological rather than based on precise specifications in architecture description languages (ADLs) or rigorously underpinned by formal models of behaviour and non-functional attributes. Why is this? Why were the actual formalisms and tools not more widely adopted? Can we draw any lessons from this? In this talk, I hope to explore this further, drawing on my personal experience as a researcher in distributed software architectures. I particularly hope to tickle the fancy of the younger members of our community, indicating the excitement of research, the benefits of belonging to a vibrant research community such as ours, and the benefits of being an active contributor. For the more mature researchers, there will be some nostalgic memories combined with some inevitable stepping on toes. For both young and old, there will be some thoughts for research opportunities as the need for self-managing adaptive software systems becomes more urgent.