in fairness, "void" strikes me as a type while "nil" strikes me as a representation. When I see "nil" in a language I expect it to have all the hideous C-like baggage of applying to any type and being a valid value for any variable. With "void," I would stop and read the documentation to see how it works.

If void behaves like nil in other languages, that's bad. But if it behaves like None in Haskell, that's good.

Seeing as how there is no not-Nothing ("Just") data constructor, it seems that it is more like ⊥ ("undefined").

But like most specifications for unimplemented software, he could have just forgotten to add this, and will realize he needs the other half when he starts implementing.