Virtual Laboratories

At the frontier of most areas in science, computer simulations play a central role. The traditional division of natural science into experimental and theoretical investigations is now completely outdated. Instead, theory, simulation, and experimentation form three equally essential aspects, each with its own unique flavor and challenges. Yet, education in computational science is still lagging far behind, and the number of text books in this area is minuscule compared to the many text books on theoretical and experimental science. As a result, many researchers still carry out simulations in a haphazard way, without properly setting up the computational equivalent of a well equipped laboratory. The art of creating such a virtual laboratory, while providing proper extensibility and documentation, is still in its infancy. A new approach is described here, Open Knowledge, as an extension of the notion of Open Source software. Besides open source code, manuals, and primers, an open knowledge project provides simulated dialogues between code developers, thus sharing not only the code, but also the motivations behind the code. Computer simulations began more than fifty years ago as straightforward extensions of theoretical pen-and-paper calculations. Even in those days, computers were faster than human calculators by several orders of magnitude. This already created totally new opportunities, such as the possibility for astrophysicists to do virtual laboratory experiments. Whereas astronomy is the oldest of the physical sciences, it was the one branch of science that could not put its subject matter, stars, in a lab. Unlike other environmental sciences, like meteorology and geology, they could not even touch their subject matter, and the best they could do was to observe the ongoing experiments in the universe at enormous distances, without any form of control. But when Martin Schwarzschild and his collaborators used John von Neumann's computer in Prince-ton in the early fifties, suddenly it became possible to put a whole star in a virtual laboratory, and to watch it evolve over time scales of millions and billions of years. In the half century since then, computers have increased in speed by more than ten orders of magnitude. As a result, all areas of science have been affected. And no wonder: a quantitative increase in speed over human pen-and-paper abilities of fifteen orders of magnitude is reflected in an enormous qualitative advance as well. Just open any science journal and look at any article therein, and the chances are high that …