Over at Just Simple Enough, Amy Hurford compiles an interesting (and comprehensive?) list of all the reasons why one might do mathematical modeling. Too many people think the only reason for scientists to do modeling is “to make predictions”; this list is a useful corrective. Go read it–especially if you’re not a modeler.
Amy also notes that different references on models in ecology (e.g., Caswell 1988, Otto and Day’s 2007 textbook) offer different reasons to do modeling. None of the references touches on all the reasons in her list, and a couple of her reasons aren’t listed in any of the references she considers.
This list is a good complement to Bill Wimsatt’s discussion of all the various ways in which models can be false, and the various ways in which those falsehoods actually help models serve the various purposes they serve.