Recursive typedef

[Karen Lanrap]

Markus Dangl wrote:

> but there must be a better (typesafe) way... Has
> anyone tried this before? 

If your approach would be possible, then also recursive types as 
formal parameters of functions would be possible. I tried for the 
latter case, but that was out of academic interest only. It did not 
work and I investigated that no further.

I do believe that there cannot be a typesafe way to define a type 
that is recursive in itself, but I have not tried to prove that.