It was figured out much earlier than 1996 - CAML first appeared around mid 80ies, and the theory behind type inference is more than 30 years old. I strongly sus

It was figured out much earlier than 1996 - CAML first appeared around mid 80ies, and the theory behind type inference is more than 30 years old. I strongly suspect someone of Richard Gabriel caliber to be aware of this in 1996. Maybe the real difference is that computers became powerful enough to do this kind of compilation in near real time not so long ago ?