[PRL] Type-directed partial evaluation

[Mitchell Wand]

> Take PCF instead of Scheme. Take good(tm) domains instead of 
> target-machine code (and read some of Mitch's old papers from the early 
> 80s). Then I can create source text for all infinite computable
> "compiled codes" using call/cc. Though I have to admit that the 
> algorithm is burried in two papers
> and I only know the one for finite computable "codes" reasonable well 
> to do it on the spot.

> -- Matthias

Umm, tell us more?

--Mitch