In most computer science research, “time” receives little attention. By dedicating a Research Theme to the topic, DIRC has been able to explore the full impact of timeliness issues on the development and use of computer-based systems.

The construction of large computer-based systems imposes a number of significant challenges, both technical and organisational. Their complexity makes all stages of their development (analysis, specification, design, implementation, deployment and maintenance) subject to failure and costly re-working. Even creating an unambiguous behavioural description of the timing aspects of an existing system is far from straightforward.

One characteristic of computer-based systems is that they are required to function at many different time scales (from microseconds or less to hours or more). Time is clearly a crucial notion in the specification (or behavioural description) of computer-based systems. However, in modelling schemes and informal descriptions time is usually represented as a single, flat, physical phenomenon. Such a representation limits understanding of the structural properties of the system; it also fails to support the separation of concerns that can come from thinking about the different time scales of the system.

A particularly important concern that has received little attention is understanding the heuristics and biases that develop when human operators make temporal decisions. Humans use a wide variety of temporal reference systems, and problems can arise when these information sources are incorrectly perceived or integrated. DIRC has developed a number of formal decision models and has compared the optimal decision strategies generated by these models with empirical data from microworld experiments.

To make better use of `time', with the aim of producing more dependable computer-based systems, we have proposed a framework that explicitly identifies a number of distinct time bands in which the system under study is situated. The framework enables the temporal properties of existing systems to be described and the requirement for new or modified systems to be specified. The framework identifies:

• key characteristic of all bands eg sampling, control, scheduling and
  planning, agreement, causality, timeliness, timing failures


• methods for describing, modelling and verifying behaviours within a band


• distinctive features of specific bands in the physiological and sociological domain (behaviour such as temporal affordance and emergence properties)


• vocabulary to adequately describe temporal activity within and across bands


• methods for describing, modelling and verifying mappings between bands

A number of case studies, including an extensive study of a neonatal ward in a major NHS hospital, have enabled us to identify timescales at which the system components operate. DIRC is making a real contribution to system analysis methods by emphasising how increased understanding of time bands and the use of time in the design process can lead to better comprehension of complex computer-based systems.

RECOMMENDED READING

Using Cognitive Task Analysis to facilitate the integration of decision support systems into the neonatal intensive care unit., Baxter, G.D., Monk, A.F., Tan, K., Dear, P.R.F., & Newell, S.J.
Artificial Intelligence in Medicine , Volume 35 , Issue 3 , pp. 243-257 , 2005.

Baxter, G.D., Filipe, J.K., Miguel, A., & Tan, K. (2005) The effects of timing and collaboration on dependability in the neonatal intensive care unit. In F. Redmill and T. Anderson (Eds.), Constituents of Modern
System-safety Thinking: Proceedings of the Thirteenth Safety-critical Systems Symposium. (pp. 195-210). London, UK: Springer-Verlag.

A. Burns and G. Baxter, Time Bands for Systems Strcture, in Structure for Dependability: Computer-Based Systems from an Interdisciplinary Perspective", (editors) D. Besnard, C . Gacek and C. B.
Jones, Springer, 2006, ISBN 1-84628-110-5.

Hildebrandt, M., and Meyer, J. (2005). When to act? Managing time-accuracy trade-offs in a dynamic belief updating task. In Proceedings of the 49th Annual Meeting of the Human Factors and Ergonomics Society.

Hildebrandt, M. & Rantanen, E. (2004). Time Design. Proceedings of the 48th Annual Meeting of the Human Factors and Ergonomics Society (pp. 703-707).

Loer, K.F., Hildebrandt, M. & Harrison, M.D. (2004). Analysing dynamic function scheduling decisions. Proceedings of IFIP 13.5 Working Conference on Human Error, Safety and Systems Development.