[racket-dev] Five feature/limitation interactions conspire to drive mad

On Dec 31, 2012, at 6:32 PM, Sam Tobin-Hochstadt wrote:

> I'm not sure that an ad-hoc naming solution is the right thing.  I
> really want to avoid adding more ad-hoc complexity to the Typed Racket
> type checker, and instead work on simplifying it into something that
> can be more precisely characterized, and thus be more sure that it's
> implemented correctly.
> 
> Is there a generalization of this solution that we can give a clean story for?


"Sometimes you have to make the internals quite complex to make it 
all appear elegant and simple." Guy Steele, as conveyed by Matthew
and confirmed by Sam