The State of Development of CSE

Introduction

‘Computational Science and Engineering’ (CSE) is an emerging field of scientific activity. We stress the ‘emerging’ because, as will be seen, whilst there are strong developments producing parallel transformations in the state of scientific practice, these are often independent of each other, are uneven in the pace and level of their development, and are embedded in and distributed across a range of specialized environments. In our view understanding the current state and assessing the prospects of CSE requires attention to organizational issues which surround its practice and influence its form.

CSE is a relative newcomer on the academic scene but is bound to have increasing effects on the way in which science is conducted because of the (still) rapidly increasing power, lowering costs, advancing capacities, and ubiquitous availability of computing resources. Computational science enables a third new means of discovery in the natural sciences in addition to experimentation/observation and theory. Computation, through simulation, is increasingly used to explore complex systems that are (relatively) inaccessible to experiment or observation because of scale, cost or technological difficulties. Computational simulations are based on mathematical models embodying theoretical understanding from the scientific discipline and, where possible, results are validated against observational/experimental data. With valid computational models ‘virtual experiments’ can produce new findings. However, this practice is not consistent across all areas of science, for example in some areas of mathematical simulation such as astrophysical simulations which aim is to explore all possible universes the comparison to our universe through observational data is not always important (Leng, 2006).

Because CSE is pitched at the leading edge of scientific and engineering research and also at advanced levels of computing developments it is, by its nature, necessarily interdisciplinary. In a world of globalised internet access amongst researchers, CSE, from its beginnings, has been a means of inter- and multi-national research collaboration. Equally, by its very nature, CSE involves interdisciplinary collaboration between computational specialists, on the one hand, and specialists from one or more scientific research disciplines on the other in order to produce meaningful results. Other communities (such as vendors or service providers) are stakeholders in these activities and can have influence on them even though not normally active in the scientific part of the collaboration.

This can mean that the development of CSE as a recognizable and distinctive area of work does not yet have a strong institutional identity or organizational coherence. This is reflected in the fact that there is as yet little sound data on the extent to which CSE has developed as a distinct discipline in the UK (EPSRC & DFG, 2005) and there are as yet few degree programs or textbooks dedicated to the field, although there is no question that in the USA and in the UK¹ and Europe the presence of CSE is increasingly felt. Thus, although CSE is catching on, it has not yet become a strongly coordinated development since the exploitation of CSE and the conditions of its development has been very much a distributed process, with CSE researchers being located in existing research structures (university settings most prominently) and therefore tied to either the organization of university IT service arrangements or to IT service arrangements within specific research departments.