Models and simulations are wonderful tools for helping us understand complex systems, and often show us relationships between components and actors that are not normally visible. It's possible to rely too heavily upon simulations, however, as even the best-constructed ones will leave out parts of the modeled system. One way of improving the capability of simulations to reflect reality is to open up their underlying engine, allowing users to get in and change the rules to better fit new information and previously-unrecognized conditions. In The Map Is Not The Terrain, The Sim Is Not The City, I took a look at how the popular computer game SimCity is used in the eduction of urban planners, and what might be accomplished if its proprietary code were opened up.
[While some] complaints arise from the fact that SimCity is built as a game -- the "God Mode," for example -- most derive from inability to modify the underlying model, whether to include mixed-use development (the ground-floor commercial/upper-floor residential buildings which help to make dense urban environments livable), to vary the demand ratings for various services, to make pedestrian travel more acceptable, or to alter the efficiency and availability of renewable power generation. As a result, some models of urban development, such as the "New Urbanism" movement of the mid-late 1990s, fall outside the scope of the simulation, and become invisible to developers-in-training. While a free/open-source version of the software would be the ideal (if highly unlikely) solution, format information and tools for altering the model would be sufficient. They have tools for changing the appearance of buildings and props, why not tools for the parts of the sim that really matter?