Dependence on administrative regions [message #521065] |
Tue, 16 March 2010 11:25 |
Arik Kershenbaum Messages: 13 Registered: January 2010 Location: Israel |
Junior Member |
|
|
I've noticed that some elements of STEM (the infector, particularly) are tied to an ISO administrative region. This is slightly at odds with the model of a scenario-model-disease, which is independent of external property files. In particular, although it is possible to construct a simulation with a non-geographic model, it would be impossible to run it because the infector is tied to a country/administrative region.
This problem arose because I was trying to write an abstract lattice capability that could be used to test disease models in a controlled spatial grid. But there are other practical implications in any case where the fundamental geographical unit isn't an administrative region. For instance, we might (will!) want to simulate the spread of foot and mouth disease across a group of farms, or look at mosquito breeding pools. Or even the spread of disease between hospital beds. In all these cases, STEM would support the model, but not the infector.
It was suggested that the problem is with the infector wizard and not with the type itself, but I don't think this is the case. The infector implementation uses its targetISOKey to answer a getTargetURI(). Of course, I could define a new superclass for both old-style infectors and new, administrative-independent ones, but that's not a very nice solution.
I'm raising this issue here and not in the bug list, well partly because Jamie asked me to, but also because I like the scenario-model-disease architecture, and I think it's important to preserve it. But others may feel that STEM is and should be tied to clear geographical entities. So I'm hoping people will contribute their thoughts.
Thanks,
Arik
|
|
|
|
Powered by
FUDForum. Page generated in 0.03309 seconds