Geoscience Reference
In-Depth Information
Finally, note that in order to promote reproducibility, the e-Research community has adopted the
concept of a workflow as a sequence of computational steps which can be repeated, refined, updated
or perturbed. Further examples of GC workflows will be provided in Section 10.3. Workflows pro-
vide a powerful means not just for preserving and repeating experiments as a foundation for sharing
knowledge but as a means for sharing the very means by which that knowledge was created in the
first place. They are a tangible manifestation of the argument that computation now lies at the heart
of scientific scholarship (cf. Donoho et al., 2011, quoted by Brunsdon, 2014).
10.3 NATIONALE-INFRASTRUCTURE FOR SOCIAL SIMULATION
In a recent project funded by JISC through its information environment programme, we have set
about articulating the components of a virtual organisation for spatial analysis and geosimulation
and have begun to connect specific software and hardware components for the exercise. In sympathy
with scientific views of a virtual laboratory or virtual observatory, it is argued that data-intensive
research architectures are potentially crucial tools for social research and applied problem-solving.
A slightly simplistic but evocative representation of this is in the idea of a
SimCity for Real
. Is it
possible to characterise the development of a city not in the abstract as a
game
but as either a peda-
gogic device which reveals important truths about urban processes and structure or even as a
what
if
modelling device which is capable of assisting in the impact evaluation of various policies or
planning options? Although the National e-Infrastructure for Social Simulation (NeISS) (see next
paragraph) is perhaps the first project to characterise this class of problems in e-Research terms,
the challenge is by no means completely novel or unique. For example, work on the UrbanSim
project (http://www.urbansim.org) over many years has striven to create a reusable architecture for
urban planning and design, which indeed has many collaborative elements including applications in
64 countries across 6 continents and software contributions from many participating organisations.
The fruits of our own endeavours are characterised as a NeISS (http://www.neiss.org). As in the
model of the research life cycle presented in Figure 10.1, the architecture for NeISS comprises not
just computation but also data, analysis, publication and preservation and ultimately uptake and util-
isation. A series of
simulation services
provide the foundations of the approach. Indeed, the notion
of social models is right at the heart of the whole exercise, as may already be clear. For example, we
will introduce below (Figure 10.4 in Section 10.5) a
use case
of this technology which centres on the
relation between congestion charging and transport behaviour in a major city. This modelling tool is
fed by data inputs and outputs from these procedures, which need to be represented and interpreted
through a process that is rendered here as
visualisation
.
The Taverna scientific workflow management system (Oinn et al., 2004) has been developed by
bioinformaticians as a software suite for designing and implementing data-intensive workflows.
The application of workflows in the NeISS architecture facilitates the linkage of data, models and
visualisations to address specific research challenges and applied problems. As well as technical
integration, some form of social curation of experiments is needed if they are to be shared and
reused. For this reason, wikis, blogs and other mechanisms and platforms were provided through
which researchers can annotate, discuss and combine intelligence about scientific knowledge and
experiments. For bioinformatic applications, a virtual research environment (VRE) called myEx-
periment has been created to facilitate the publication of scientific workflows and the associated
numerical experiments. The more conventional notion of
publishing
is a reasonable way to describe
this activity in everyday terms - the objective is to make available individual services (data, models,
visualisation) or composites (i.e. workflows) to collaborators seeking to repurpose, extend, verify or
otherwise embellish these research elements.
An exchange of workflows could easily be the preferred method for sharing of resources between
experienced or specialised research users of a simulation e-infrastructure. In everyday experience,
such users would be comfortable with designing, writing or editing software codes, with complex
data manipulation or with addressing difficult questions using statistical packages (e.g. Statistical
