We axiomatize the congruence relation for binding bigraphs and prove that the generated theory is complete. In doing so, we define a normal form for binding bigraphs, and prove that it is unique up to certain isomorphisms. Our work builds on Milner’s axioms for pure bigraphs. We have extended the set of axioms with five new axioms concerned with binding, and we have altered some of Milner’s axioms for ions, because ions in binding bigraphs have names on both their inner and outer faces. The resulting theory is a conservative extension of Milner’s for pure bigraphs.

Reproduction of all or part of this work is permitted for educational or research use on condition that this copyright notice is included in any copy.

Informatics bridges Turing-computation and interactive behaviour; examples of the latter include ubiquitous/pervasive and biological systems. But how does a model of computation fit within a model of less disciplined informatic behaviour? This paper offers a precise treatment of that relationship, identifying a class of calculational bigraphical reactive systems. We show how such a system contain a confluent calculation sub-model, and how calculation only ever enables, never prevents, informatic behaviour of the larger model. We submit these results as a modest but essential beginning of a unified informatic theory.

Bigraphs are emerging as a (meta-)model for concurrent calculi, like CCS, ambients, πcalculus, and Petri nets. They are built orthogonally on two structures: a hierarchical place graph for locations and a link (hyper-)graph for connections. Aiming at describing bigraphical structures, we introduce a general framework, BiLog, whose formulae describe arrows in monoidal categories. We then instantiate the framework to bigraphical structures and we obtain a logic that is a natural composition of a place graph logic and a link graph logic. We explore the concepts of separation and sharing in these logics and we prove that they generalise well known spatial logics for trees, graphs and tree contexts. As an application, we show how XML data with links and web services can be modelled by bigraphs and described by BiLog. The framework can be extended by introducing dynamics in the model and a standard temporal modality in the logic. However, in some cases, temporal modalities can be already expressed in the static framework. To testify this, we show how to encode a minimal spatial logic for CCS in an instance of BiLog.

We propose a novel and uniform approach to type systems for (process) calculi, which roughly pushes the challenge of designing type systems and proving properties about them to the meta-model of bigraphs. Concretely, we propose to define type systems for the term language for bigraphs, which is based on a fixed set of elementary bigraphs and operators on these. An essential elementary bigraph is an ion, to which a control can be attached modelling its kind (its ordered number of channels and whether it is a guard), e.g. an input prefix of π-calculus. A model of a calculus is then a set of controls and a set of reaction rules, collectively a bigraphical reactive system (BRS). Possible advantages of developing bigraphical type systems include: a deeper understanding of a type system itself and its properties; transfer of the type systems to the concrete family of calculi that the BRS models; and the possibility of modularly adapting the type systems to extensions of the BRS (with new controls). As proof of concept we present a model of a π-calculus, develop an i/o-type system with subtyping on this model, prove crucial properties (including subject reduction) for this type system, and transfer these properties to the (typed) π-calculus.

Abstract. Informatics bridges Turing-computation and interactive behaviour; examples of the latter include ubiquitous/pervasive and biological systems. But how does a model of computation fit within a model of less disciplined informatic behaviour? This paper offers a precise treatment of that relationship, identifying a class of calculational bigraphical reactive systems. We show how such a system contain a confluent calculation sub-model, and how calculation only ever enables, never prevents, informatic behaviour of the larger model. We submit these results as a modest but essential beginning of a unified informatic theory. 1

Abstract not found

This is a draft for publication. Comments are welcome.