[TYPES] type-sensitive CFA?

[Paul A. Steckler]

John Reppy asked:
> I'm looking at a program analysis problem for ML-like languages where
> I want to exploit type abstraction to improve the precision of a
> standard 0-CFA type analysis.  I was wondering if such analyses have
> been previously studied and/or implemented?  Any pointers would
> appreciated.  Thanks.

This paper by Jagannathan, Weeks, and Wright describes the 
use of types to control polyvariance:

 http://citeseer.csail.mit.edu/jagannathan97typedirected.html

-- Paul