re free type variables in SML

Henry Cejtin
Fri, 24 Aug 2001 14:40:12 -0500

In there has been a discussion going on about some compilers
(Moscow ML if I recall) which allow a top level declaration which has a type
which cannot be polymorphic, but where the mono-type it is isn't determined yet.
Some compilers seem to reject this while Moscow leaves the types free until
the first use.  Then they are fixed with what worked there.