Seminar: Understanding why systems fail


Understanding why systems fail

Monday 25 June, 2007 - Edinburgh International Confrence Centre, Edinburgh, UK

Contingency Planning and Longer Term Perspectives on Learning from Failure in Safety-Critical Systems.

There have been significant technical advances in the dependability of many complex systems over the last twenty years. Partly in consequence, accident and incident rates have fallen across a range of industries. However, these advances are often offset by increasing exposure. For example, reductions in the accident rate may be offset by rising numbers of departures in the aviation industries. Improvements in the engineering of safety-critical software are offset by the increasing use of programmable systems within application areas that would not have been considered ten or twenty years ago. It seems likely that these changes in both the supply and demand of safety-critical systems do not occur by chance. Innovations in dependability are linked to the increasing need to deploy these applications.

In other areas, new threats and hazards are placing increasing demands on safety and security engineering. It seems clear that existing methods of risk assessment cannot easily be applied to analyse the potential consequences of climate change, especially where political influence and public opinion must inform engineering decisions. Similarly, concerns over terrorist attacks across Europe, Asia and North America are revealing new vulnerabilities. These areas of concern are growing as ICT and SCADA systems combine with market deregulation to increase dependencies between many different infrastructures.

Given these broad, long-term changes it is critical that we find ways of learning from any adverse events that occur. Incident and accident reporting systems are only one way in which we can inform safety management systems with direct operational experience. Conventional forms of risk assessment and resilience engineering must be supported by longer term contingency planning. However, there are many areas that cannot easily be addressed using existing techniques;

these include but are not limited to:

  • political and social factors affecting incident and accident reporting;
  • learning from failure in critical infrastructure development and protection;
  • the human factors contribution to long term risk;
  • impact of insurance and financial services on engineering development;
  • the validation of long-term contingency planning;
  • the consideration of 'external factors' (meteorological changes, pandemics etc) on long-term engineering decisions;
  • cross-overs between security and safety in contingency planning.

The seminar will take a broad multi-disciplinary view, corresponding to the range of factors and issues of relevance.


This Safety-Critical Systems Club event was co-located in Edinburgh with the Dependable Systems and Networks Conference (DSN 2007), which was the premier international conference on the dependability, safety and security of complex computer based systems. For full details of DSN 2007, including details of costs and registration, please visit the DSN web site here.

Potential delegates are invited to consider attending both events. To facilitate international participation at this Safety Club seminar, the charge to all delegates has been set at £195, which is the standard member rate applicable to subscribing members of the Club.



09:30   Coffee and Registration
10:00   Introduction
10:10   Detection and Recovery from Clinical Laboratory Error Resulting in False but Plausible Test Results
Nick Chozos, Univ. of Glasgow; Bob Wears, Univ. of Florida; Jim Barnes, Clarkston Medical Centre; and Frank Finlay, Southern General Hospital Glasgow
10:50   Classes of Socio-technical Failures: Microscopic and Macroscopic Scales
Stuart Anderson and Massimo Felici, University of Edinburgh
11:30   Break
11:50   Towards Sustainable Policy Coordination: A Simulation-based approach
Jorge Hollman and José Marti­, University of British Columbia
12:30   Lunch
13:30   Responsibility Modelling for Contingency Planning
Ian Somerville, Tim Storer and Russell Lock, University of St. Andrews
14:10   Understanding the Interaction between Coping Strategies and Degraded Modes of Operation
Chris Johnson, University of Glasgow; and Chris Shea, ESR Technologies
15:00   Break
15:20   "Oh what a tangled web": exploring the role of networks in the incubation and escalation of failures within the context of an organisational fitness landscape
Denis Smith and Moira Fischbacher, Department of Management, University of Glasgow
16:20   Panel session
Workshop participants will have the opportunity to raise issues from the day's presentations in a discussion intended to develop an agenda for future research into longer term risk assessment and contingency planning.
17:00   Close

Cost and registration

The cost of this one-day event is £195, which includes lunch and refreshments during the day.

A copy of the programme outline and registration form is available here.

SCSC.UK uses anonymous session cookies please see Privacy policy

SCSC 06-03-2018 [V4e]