My request for discussion is prompted by this recent SciComp.SE post, Definition of incompressible flow, and relates to the scope of the forum.
As commented there I feel the topic falls just far enough outside the scope defined in our FAQ to vote to close the question. Once carefully parsed, it asks about how density variations in fluids due solely to temperature variations can be modelled while retains an essentially incompressible fluid formulation of flow.
The answers received were excellent, and while strictly speaking the question would be a better fit with Math.SE or MathOverflow, it's hard to argue with success. On the other hand the only connection to computation/programming is an implicit one that good computational modelling proceeds from tractable mathematical models.
Therefore I'm asking for community reflection on how strict one ought to be when voting to close "edge cases" of this type (topics of mathematical modelling that do not explicitly raise computational issues).