This activity started in October 2000 with an original focus on design for dependability. The principal activities were focused on the use of diagrams in systems design, in investigating how to provide designers with knowledge of actual workplace practices, and with the notion of evidence-based design. Evidence-based design was concerned with the collection of qualitative and quantitative evidence about systems to help inform design decisions. The activity uses a series of ethnographic studies in medical settings whose aim is to understand how computer systems are configured and reconfigured after deployment and how this affects the dependability of the socio-technical system where they are used.

In a review of the activity about a year after it started, we decided that the deployment, reconfiguration and evolution of computer systems was of particular interest and that we should focus the activity on system deployment and evolution.

The overall aim of the activity now is to to study and characterize the relationships between dependability and evolution of the technical system, organization and environment, focusing on (re) configuration as the means to control the effects of evolution on dependability. Our goal is to provide conceptual and tool support for design for configuration, and configuration before and after deployment that will reduce the number of system or sub-system failures thereby improving the system¹s delivered quality of service.

Within this overall aim, we have the following objectives:

• To observe the long-term evolution and reconfiguration of socio-technical systems and their organisational settings and to analyse how user-led evolution and re-configuration both helps cope with and acts as a source of system failure.

• To devise and assess approaches to configuration in socio technical systems as a means to avoid, control and manage reasonably frequent non-catastrophic failures of sub-systems.

• To devise and assess evidence-driven approaches to support both developers and users in achieving dependability requirements across the deployment boundary. In particular to explore lightweight methods that exploit data gathering to direct the structure of configurations and how systems should be configured to fit a particular work setting.

Much of the Lancaster work in this activity has, until now, focused on providing information and evidence about work settings for designers. This has taken three forms:

• Patterns of interaction: Based on an extensive corpus of studies, we have develop a set of patterns that describe common workplace interactions and have identified the dependability implications of these patterns.

• Situation modelling: This work is based on the fact that system designers often have only second-hand knowledge of the working environment where a system is to be deployed. Because of this, they make design decisions that are inappropriate for that environment. The intention here is to provide a basis for modelling real situations where socio-technical systems are used and to provide a multi-media database that can inform designers about these situations.

• Configuration management: Focusing on the modelling, analysis and evolution control of system configurations. The aim being to provide support for the full range of socio-technical configuration components which form part of modern systems. All of the mechanisms which are being developed as part of this work are informed by results of ethnographic study.