The aim of the visit was to work towards completion of task T2.7 on “Inferring behavioural descriptions for b-APIs (O2.3)”.
During this visit, initial discussions took place with António Ravara (Nova) and Carlos Mão de Ferro, a PhD student of Francisco Martins (Green by Web) on how theory and applications of behavioural type systems in object-oriented language, in particular StMungo and Mungo tools for Java designed and developed at the University of Glasgow, can be applied in the context of sensor networks, the latter used by Green by Web.
More specifically, António Ravara is now working on correct usage of linear objects, object having behavioural types associated to them, also known as typestate specifications, to avoid null pointer exceptions in the Mungo language. Laura on the other hand, is currently working on type-theoretic techniques for generalising access control beyond strict linearity. This work will be a foundation for better type systems for access control in object-oriented languages and in particular Mungo. This will thus add to the usability of the Mungo tool, going beyond strict linearity of behavioural types, which forms part of the proposed work for T2.7.
The second and final part of this secondment, will look into the application of these concepts, and in particular jointly with Nova on the inference of typestates for the Mungo tool.