What to do with empty systems with non-empty observed equations

[YingboMa]

Simulating an empty system leads to an error, so one has to do something like

julia> sys = structural_simplify(mediumModel)
Model mediumModel with 0 equations
States (0):
Parameters (0):
Incidence matrix:


julia> prob = ODEProblem(sys, nothing, (0, 10.0));

julia> prob.f.observed(medium.P, rand(100), rand(100), 0.0)
101325.0

Would be nice if we can "solve" an empty system.

[anandijain]

what is the solved system supposed to return?

[ChrisRackauckas]

Should this instead happen at the DiffEqBase level, where the generic higher solve checks some property and returns a solution before hitting a solver?