Thanks! We're really excited to see how they work out too, there's some exciting potential in there.

I can imagine having programs which are tree-like hierarchies of regions, but I'm starting to suspect we can have directed acyclic graphs, and perhaps even very small cycles in them...