The computational approach to doing science is inherently multi-disciplinary: it requires of its practitioners a firm grounding in applied mathematics and computer science in addition to a command of one or more scientific or engineering disciplines or in the high-tech arts. Thereby the role of computer science is similar to the role of mathematics in the mathematical sciences in that it provides the tools, ranging from networking and visualization tools to algorithms, that match modern computer architectures. Mathematics provides means to establish credibility of algorithms, such as error analysis, exact solutions and expansions, uniqueness proofs and theorems.
At this stage computational science may
be thought of as a methodology common to a variety of
sciences, which
makes use of the same kind of tools. It is conceivable that
computational science will one day be redefined as a discipline,
following significant breakthroughs and as some of its major
challenges are addressed.
The computational science community is very diverse and includes
researchers with a multitude
of area-specific terminology and research
methodologies. This community as a whole carries the
responsibility to define quality research in this area
and to set the standards
for publications. The community has the task to assess what
activities have value and to communicate results within this
very
diverse community. Furthermore, there is a need to develop
training and education of future practitioners of computational
science [3].
See
the white paper by Chuck Swanson at Cray Research, Inc.
for a list of institutions offering graduate and undergraduate
programs in computational science (1994).
