General concurrent systems modelling
The links on this page reference old MSc material which have been mounted in their original form as an empublic project (see emfcsBeynon1997) [Charlie, November 2005]
EM treats the notion of a concurrent system as a special form of construal. In making such a construal, the modeller stands (as if) in the role of an external observer. This makes particular demands on the character of the observables, dependencies and agency in terms of which the modeller accounts for the system. The potential contexts for such construal are very diverse: the modeller may be one of several designers participating in the concurrent design of a product, and be projecting herself into the role of an external observer trying to make coherent sense of potentially divergent views of what form the product should take; the modeller may be observing a natural phenomenon and trying to contrue it as an experimental scientist might; the modeller may be a participant in a well-established and routine business process who wishes to understand the process from the manager's perspective. Concurrent interaction is involved in all these scenarios, but these interactions will not necessarily be such that they define systematic predictable behaviours that achieve identifiable goals.
The above perspective on concurrent systems is developed in three lectures [M2, M5 and T1] from the MSc module EMfCS. These motivate the application of EM to two kinds of activity:
Identifying and constructing a concurrent system: This entails assembling physical objects / agents, establishing a regime for observation and protocols for interaction. Subject to reliable operation (an assumption about the environment), we then have generic patterns of state change (behaviours) that fulfil expectations.
Construing a situation as a concurrent system: This involves accounting for system behaviours by identifying agents / objects and the observables that mediate their interaction (cf. an LSD account). It typically draws on a vast amount of experience: personal, of natural phenomena, of social interaction. The issues raised concern knowledge that migrates in many different ways: from subjective to objective, particular to general, provisional to assured.
The knowledge representation entailed in these activities cannot be propositional in character. Formal specification of intended system behaviour and/or system components - by itself - has only a very limited role to play. EM advocates knowledge representation based on the open and interactive development and use of (typically computer-based) artefacts. The principles and techniques involved in this development resemble those of experimental science and engineering. In EM, interaction with computer-based artefacts is used to generate experiences that metaphorically represent other experiences of our own, and the "experiences" we project on to other agents.
The three key techniques associated with the above activities are introduced in lecture T1 in EMfCS under the alliterative banner of agentification, artefacts, animation. These are respectively addressed with reference to LSD accounts [T2], the use of definitive principles for visualisation [T4], and modelling and simulation within the framework of the Abstract Definitive Machine [Th1, Th2].
A number of EM models to illustrate issues in concurrent systems modelling feature prominently in the lectures and tutorials on the EMfCS module. These include cruisecontrolBridge1991, digitalwatchCartwright1995, linesBeynon1991. The manner in which railway technologies and practices have developed over time is used in EMfCS to illustrate the subtlety of the progression from preliminary experimental interaction in concurrent systems to the reliable automated concurrent behaviour to which we aspire in modern reactive systems. Two models that are related to this theme are railwayYung1995 and claytontunnelSun1999. The latter model makes use of a distributed variant of the EDEN interpreter to animate the roles of participants in the general setting for a historic railway accident that occurred in 1861.