Herm, that more or less confirms what I thought might be the case. I'ts really a shame too as it adds a fairly large tick back in the "don't use swift" column for the project I'm planning. I'll be managing a large team of developers working on the next generation of a complex robotics library with lots of different capabilities and domains, so it's pretty critical that we have a clean, consistent way of organizing code.
I read your Namespaces x submodules post from the beginning to try to get a handle on what the "current state of the art" is here. Did that ever resolve to anything elsewhere? It seems like solving this is a pretty critical aspect for swift to get any traction in the systems code space 
Do you have an opinion on the enum approach? Does anyone actually do that? I've tried to look at existing large-ish code bases like Vapor; it appears everyone just tries to carefully name everything to avoid namespace polluting and that sounds like the path to madness to me.