[Haskell-cafe] Imagining a G-machine

[Simon Peyton-Jones]

| > But then GHC would be faster then JHC!  (Nobody cares about jhc,
| > certainly not enough to implement a recognizer for it...)
|
| Ah, but think of how much faster jhc development would be if it didn't
| take ghc 20 minutes to compile it every time I made a change :)

Oh!  A cruel jibe!

Simon

(Methinks I must attempt to recruit one J Meacham to GHC optimisation duties.)