Linear Uniform Receptiveness in a Pi-Calculus with Location Failures

The notion of receptiveness arises in the π-calculus as a guarantee of determinacy in the behaviour of callable entities and was first investigated by Sangiorgi. The DπF process calculus, introduced by Francalanza and Hennessy, extends the π-calculus with located processes and location and link failures. In this paper we extend the notion of receptiveness to DπF and give sound characterizations of the property of linear uniform receptiveness in DπF in the form of two type systems. Our first type system ensures receptiveness, the property that no pending output will ever be left unattended. We achieve this by ensuring linearity and by ensuring that the input and output remain at the same location, such that location failure will effectively remove either both or none. Our second type system allows for migration but ensures that input capabilities remain within locations which are hidden from the context and thus not subject to failures.