A General Name Binding Mechanism

Abstract

Abstract. We study fusion and binding mechanisms in name passing process calculi. To this purpose, we introduce the U-Calculus, a process calculus with no I/O polarities and a unique form of binding. The latter can be used both to control the scope of fusions and to handle new name generation. This is achieved by means of a simple form of typing: each bound name x is annotated with a set of exceptions, that is names that cannot be fused to x. The new calculus is proven to be more expressive than pi-calculus and Fusion calculus separately. In U-Calculus, the syntactic nesting of name binders has a semantic meaning, which cannot be overcome by the ordering of name extrusions at runtime. Thanks to this mixture of static and dynamic ordering of names, U-Calculus admits a form of labelled bisimulation which is a congruence. This property yields a substantial improvement with respect to previous proposals by the same authors aimed at unifying the above two languages. The additional expressiveness of U-Calculus is also explored by providing a uniform encoding of mixed guarded choice into the choice-free sub-calculus.

JSCL: A Middleware for Service Coordination

“…However the calculus SC can be enriched by considering signals as tagged nested lists [2,14], which represent XML documents, and conversation schemata as abstractions for XML Schema types [1]. This extension of conversation schemata lead to a more general notion of reaction based on pattern matching or unification in the style of [6]. A further extension can provide component and schema name passing, modelling a more dynamic scenario.…”

A behavioural theory for a π-calculus with preorders

“…This example suggests that πP gives a better control on restricted names. (Note that a different approach is adopted in [3,4] to address this question.) In order to gain a better understanding of such phenomena, the main purpose of the present work is to deepen the study of the behavioural theory of πP, in an untyped setting.…”