Re: [Toss-devel] No nominal subtyping with higher-order types
Status: Beta
Brought to you by:
lukaszkaiser
From: Lukasz S. <luk...@gm...> - 2012-06-16 21:18:27
|
On Sat, Jun 16, 2012 at 11:11 PM, Lukasz Stafiniak <luk...@gm...> wrote: > I've found a simple solution for higher-order types (so called > "funtype" in Speagram)! Nothing changes in the unification algorithm I > presented for hierarchical types. Higher-order types automatically *hierarchical terms > become invariant (one cannot be strictly speaking a strict subtype of > another, but they can unify). Here "unify" in the standard first-order sense, not in the hierarchical terms sense. |