Using Problem-Domain and Artefact-Domain Archtectural Modelling to Understand System Evolution

EPSRC Funding: £155,317

Abstract:

It is a common experience that building an extension to a house does not always give you more space. This apparent contradiction occurs because the total dynamics of the house is altered by what seems to be an independent addition to the structure. Each room may have a clearly define function but this function can only be carried out if it is correctly connected to the total architecture of the building. Similar experiences are found with computer systems where apparently innocent additions are made to respond to the changes in a system's environment only to find that the modified system becomes less efficient or less useful. In addition to this 'gestalt' system response to change there are constructional layers in a design that predetermines the potential for change of a building or computer system. The most difficult to change for a building is the site and structure and the easiest is the space plan and the decorations. Drawing from the concept of plate tectonics in geology the term architectonic is used to describe the movement of these layers over time. The design decisions that go to make a long term successful building has to take into account the extendibility of the building with respect to its potential future use; to understand such use requires insight into the culture within which it is to be constructed.

The project is to uncover the architectonic nature of computer systems and see how these may be related to high-level, but also grounded, model of the original problem domain. We thus propose both a General Systems Architectural Model (GSAM) that reflects all the patterns of connectivity that may be found within any program and a Grounded Systems Engineering Method (GSEM) that captures the architecture of a problem domain (the culture). We hypothesis that there will be an isomorphism between the architecture of an artefact and the problem domain as perceived by the designer. We also hypothesis that for an artefact to be continually modified it has to conform to certain architectural constraints. The purpose of the study is to examine a wide range of real programs (those designed for a job) and the problem domain that they address and from this examination determine the validity of the model and hypotheses by uncovering features and measures that define the architecture of the system. It is proposed that these features and measures could be used to make predictions about the flexibility, robustness and reliability of software. Such predictions would be used to formulate and close the design control loop. We hope that this work will lead to a universal constructional theory for systems.

GO TO .... | ideo home | SII home | EPSRC SI programme home | SII call for proposals | SII proposers' guidance | SII contacts