Once tools and approaches have been developed that enable swapping in and out of model modules and model disaggregation, these approaches can be applied to enable collaborative development of a growing “family” of partially interoperable simulation models in the spirit of Linux and other open source software systems.
It is likely that the Collaboratorium Community will need to maintain at least some centralized control over the process of approving new modules, just as open source software efforts rely on a small group to select which proposed new modules are included in new releases. But the ability to propose and develop new modules could well be opened up to any parties with the interest and ability.
Similarly, once the Community has developed tools that enable the storing and comparison of multiple simulation runs, many people could also be involved in analyzing the results and drawing implications from the many simulations that have been run and then stored in the systems scenario library.
It is important to note that at the time of the Collaboratorium’s launch, this description of radically open modeling is currently a vision, and it is not certain that reaching this final stage is feasible. Simulation tools are more complex than other systems that have been developed using open source, collective intelligence approaches, such as operating system software like Linux or encyclopedia entries of the kind developed in Wikipedia.
But before Linux and Wikipedia were launched, many observers would have expressed skepticism about the feasibility of those projects. Our goal is to see how far along the path toward radically open modeling we can get.