Predictive & Adaptive Software Language Engineering

点击翻译按钮获取中文摘要

Over the next years, key systems that significantly impact societies and everyone's everyday lives will be software-intensive, cyber-physical systems (CPS), with billions of connected smart objects and devices forecasted to exist in such systems (e.g., smart cities, buildings, and electrical grids). CPS include software as well as hardware components that need to interact with each other in increasingly complex ways and are described with a heterogeneous set of specification languages. CPS are often socio-technical systems, characterized by the complex interrelatedness of people with conflicting goals, technology, organizational processes, and the diverse social context in which these systems are deployed. A key aspect of such systems is a feedback loop that involves complex trade-off analyses and allows the CPS to self-adapt, leading to ever more dynamic systems. Such CPS will have to address many different concerns, some well-known (security and privacy), some having emerged over recent years (sustainability and human values), and unknown others that will emerge in the future. Just as natural languages evolve to communicate about new concerns and better describe new issues that mankind must face, software languages also need to continuously adapt to enable the ongoing specification of complex socio-technical systems and to cope with the profound impact that such systems have on humanity. Changes to software languages are not only triggered by emerging qualities such as sustainability or increasing system complexity, but also because problem domains are better understood over time, which needs to be reflected better by new or more concise constructs in domain-specific software languages. In essence, a software-intensive system, its users and their environments, and the software development environment form one interconnected, highly dynamic, heterogeneous, and adaptive eco-system with several feedback loops that is described with many specification languages. If required adaptations to software languages could be predicted as early as possible, software languages could be adapted more rapidly to provide the constructs needed to effectively describe CPS. Currently, the exploited feedback loop in adaptive socio-technical systems is limited to the users and their environment. This research program proposes to extend the feedback loop to the development environment including software languages, developers, and language engineers. By establishing predictive and adaptive software languages, it is possible to evolve software languages more effectively based on changes determined from the extended feedback loop, making such software languages a viable development choice for the Canadian software industry even in the context of rapidly evolving software development projects. The impact is improved productivity, leading to greater innovation in and increased competitiveness of the many organizations served by the Canadian software industry.