Cases where a single symbol can explicitly be declared as different things

[TorkelE]

rn = @reaction_network begin
    @species X(t)
    @variables X(t)
end

species(rn) # [X(t)]
nonspecies(rn) # []

When working on the inference update PR I figured it would make sense to have a couple of tests checking ensuring that this wasn't possible. Turns out it is...

I will sort this one out quite soon.

[TorkelE]

One question we have to decide. Should double declaration @species/@compounds be declared, e.g.

@reaction_network begin
        @species X(t) O(t)
        @compounds begin X(t) ~ 2O end
    end

There is no ambuguity what model this should generate. However, @compounds provide full declaration of metadata, defaults, etc. So if these are different it is not longer clear. It seems easier to not allow this kind of double declaration.

[isaacsas]

I can't see why a double declaration would ever be intentional, so I'd think it should give an error.