Shaun, thanks for taking the trouble to document all of this!
On modeling: yes, would be good to start working on a model that describes states/actions/invariants/etc. in f.realms and see if certain conditions (realm consistency issues such as overlapping landmasses, maybe implicit/intentional loopholes, etc.) can be verified to exist in the model.
Model checking is supposed to allow you to find such loopholes (analogous to finding flaws in security protocols), so would be great to try and take this approach to both verify and drive evolution of the intended semantics of f.realms.