Centre for Software
Reliability,
We have been exploring the problems of
representing and analysing complex socio-technical systems since the mid
eighties. About that time, the idea of projections
emerged and, while much attention was paid to computational and design views,
particularly in the area of information and communications systems, the idea
that you needed an enterprise model which represents the organisational and
commercial contexts of these systems became a part of received wisdom. Precisely how an enterprise should be represented and
analysed has remained a subject for speculation.
The material in these pages represents an
attempt to develop and apply a theory of architectural discourse. It is not easy reading: these are complicated and deep
issues, but we believe that they contain some progress and insights.
One of the problems with this area is
that it does not fit neatly into existing academic or sectoral silos. We are concerned
with issues of philosophy, language and communication, mathematics and formal
systems, economics and sociology as well as systems engineering and computing
science. It is precisely these compartmentalisations of
approach which seem to have become a barrier to innovative thinking.
|
|
|
First published in 1979, Godel, Escher,Bach, an
Eternal Golden Braid by Douglas R.
Hofstadter, was one of the inspirations for thinking about systems and
the way they are represented. The GEB cube is a graphical
illustration of the need for projections in representations. |
Theoretical
papers
This is a foundation paper. In it we explore the nature of architectural discourse, the
processes in which it is embedded and the different sorts of language that are
used.
What we are trying to do is explain the
relationships between rich pictures, "blobs and sticks", formal models and all
the other technical representations which are used to define the architectures
of complex socio-technical systems. We also try to
relate these different forms of expression to the cognitive processes involved
in exploring and negotiating in problem and solution space.
The
Ontology of Communicating Systems
The most important aspect of a projection
based approach to architecture is that it provides individual views which
represent a useful structuring of issues and concerns while, at the same time,
it is able to ensure that the representations within them are mutually
compatible and coherent. A common ontology across the
different projections provides this guarantee.
This ontology has proved its usefulness
in a wide range of application areas and has generated a number of insights and
clarification which we have found amazing. In the
second part of the paper, we present an overview of a model of health enterprise
in order to illustrate how enterprise projections are represented within this
ontological framework.
One of the most important jobs that is
done with an enterprise projection is to reason about the boundaries of
organisations and the composition of roles. These
processes generate conflicts and synergies of interest and structure the units
of success and failure of a proposed system.
Here we explore some ideas for a theory
of conversations and present an example of their application to explain some of
the problems that are commonly experienced in the procurement of complex
systems.
This paper is the most recent and
represents work in progress. It explores the concepts
of value and of the virtual.
This document is one of the final
deliverables of the COBRA project (ACTS AC203) which was undertaken between
August 1996 and September 1998.
It represents a relatively complete
application of our approach to modelling a complex application domain, in this
case the process of intermediation and brokerage. It
includes a high level definition of a computational model and we attempt to show
the relationship between the problem definition in a set of enterprise and
instrumental projections with the structure and behaviour of a computational
system.
The epilogue
contains some observations of the practicalities, benefits and challenges of
using this approach in the context of real world problems.