Re: Type names vs type structure

Phil Wadler wrote:
> For instance, I know of no paper that describes
> extending Hindley-Milner with unification on rational trees, although
> the idea seems to have been kicking around for a while.

I believe you can find an instance here

@inproceedings{Kaes-LFP-92,author="Stefan Kaes",
"Type Inference in the Presence of Overloading, Subtyping and Recursive Types",
booktitle="Proc.\ {ACM} Conf.\ {LISP} and Functional Programming",
publisher="ACM Press",year=1992,pages="193--204"}

but the idea has been around for much longer.