Specifying in silico experiments as coordinations of coarse-grained processes

Abstract A paradigm shift has occurred from centralized to distributed computing as the baseline upon which information systems are designed and implemented. This has raised the prospect of cooperation by coordination of computational, rather than physical, processes becoming the prevailing mode for large-scale systems. Just as, in business, e-business, soon, in science, e-science. The analogue of a business process in scientific contexts is an in vitro experiment, insofar as a major output of scientific practice is the knowledge obtained by the experimental method. Hence the concept of in silico experiments as the core component of a computational model of scientific activity now being referred to as e-science. Surprisingly, though, while the challenges presented by the ambition to coordinate e-business processes at very-coarse grains (often entire supply chains) are being seriously addressed, no comparable activity levels are yet discernible in e-science. This paper describes a platform for in silico experiments that is centred around a process coordination language called ISXL. ISXL projects to working scientists a model of in silico experiments that approximates more closely their in vitro practices than comparable efforts. An ISXL-specified experiment is (1) cast in a conceptual model that helps enforce the essence of the research method as applied to the empirical sciences; (2) may include explicit hypothesis formulation and validation (rather than simply the core functionality of evidence gathering); and (3) is long-lived, in that there is built-in support for denoting experiment specifications that evolve, as well as the results of past runs, which are, in turn, annotated and kept in an underlying persistent store. Here the focus is on (1) and (2), (3) is discussed in [7].